Practising Human Geography

P R A C T I S I N GH U M A N

G E O G R A P H Y

CLOKE-PRELIMS.qxd 3/3/04 12:06 PM Page i

CLOKE-PRELIMS.qxd 3/3/04 12:06 PM Page ii

P R A C T I S I N GH U M A N

G E O G R A P H Y

Paul ClokeIan Cook

Philip CrangMark Goodwin

Joe PainterChris Philo

SAGE PublicationsLondon • Thousand Oaks • New Delhi

CLOKE-PRELIMS.qxd 3/3/04 12:06 PM Page iii

© Paul Cloke, Ian Cook, Philip Crang, Mark Goodwin, Joe Painter, Chris Philo 2004

First Published 2004

Apart from any fair dealing for the purposes of research or private study, or criticism orreview, as permitted under the Copyright, Designs and Patents Act, 1988, this publicationmay be reproduced, stored or transmitted in any form, or by any means, only with theprior permission in writing of the publishers, or in the case of reprographic reproduction,in accordance with the terms of licences issued by the Copyright Licensing Agency.Inquiries concerning reproduction outside those terms should be sent to the publishers.

SAGE Publications Ltd1 Oliver’s Yard55 City RoadLondon EC1Y 1SP

SAGE Publications Inc2455 Teller RoadThousand Oaks, California 91320

SAGE Publications India Pvt LtdB-42, Panchsheel EnclavePost Box 4109New Delhi 100 017

British Library Cataloguing in Publication data

A catalogue record for this book is availablefrom the British Library

ISBN 0 7619 7325 7ISBN 0 7619 7300 1 (pbk)

Library of Congress Control Number 2003108066

Typeset by C&M Digitals (P) Ltd., Chennai, IndiaPrinted in Great Britain by The Cromwell Press Ltd, Trowbridge, Wiltshire

CLOKE-PRELIMS.qxd 3/3/04 12:06 PM Page iv

1 Changing practices of human geography: an introduction 1

PART I CONSTRUCTING GEOGRAPHICAL DATA 35

2 Official sources 41

3 Non-official sources 62

4 Imaginative sources 93

5 Talking to people 123

6 Observing, participating and ethnographies 169

PART II CONSTRUCTING GEOGRAPHICAL INTERPRETATIONS 207

7 Sifting and sorting 215

8 Enumerating 247

9 Explaining 285

10 Understanding 307

11 Representing human geographies 336

12 The politics of practising human geography 364

Summary of Contents

CLOKE-PRELIMS.qxd 3/3/04 12:06 PM Page v

CLOKE-PRELIMS.qxd 3/3/04 12:06 PM Page vi

Preface x

Acknowledgements xvi

1 Changing practices of human geography: an introduction 1Practising human geography? 1A thumbnail history of practising human geography 7Conclusion 31Notes 32

PART I CONSTRUCTING GEOGRAPHICAL DATA 35

2 Official sources 41Introduction 41Types of official information 42Information and state formation 43The contemporary informational state 49Understanding the construction of official information 53Conclusion 61Notes 61

3 Non-official sources 62Introduction 62Non-official sources in geographical research 63Critical issues in the use of non-official data sources 68Conclusion 91Note 92

4 Imaginative sources 93Introduction 93Understanding the construction of imaginative sources 94Imaginative sources in geographical research 104Case studies 115Conclusion 122Notes

Contents

CLOKE-PRELIMS.qxd 3/3/04 12:06 PM Page vii

Contents

viii

5 Talking to people 123Introduction 123The practices of talking to people 126Questionnairing 130Interviewing 148Discussion groups 159Ethics: an important end note 164Notes 168

6 Observing, participating and ethnographies 169Introduction: what is ethnography and how can it be geographical? 169Geography’s humanistic ethnographies 171The black inner city as Frontier outpost 173Geography’s ‘new’ ethnographies 182Top tips for prospective researchers 195Conclusion: field-noting 196Notes 204

PART II CONSTRUCTING GEOGRAPHICAL INTERPRETATIONS 207

7 Sifting and sorting 215Sitting down with your data 215What happens when we put things into boxes and make lists 223A geographical detour into set theory 227Alternatives and recommendations 240Notes 245

8 Enumerating 247Enumeration and human geography 247Describing, exploring, inferring 253Modelling spatial processes 264Geocomputation 271GIS and spatial analysis 272The authority of numbers? 277Notes 284

9 Explaining 285The complexity of explanation 285Explanation through laws: geography as spatial science 286Explanation as causation 288Intensive and extensive research 289The search for a revolution in geographical explanation 290Explanation through abstraction 291Explanation and subjectivity 295Explanation and practice 299Concluding comments: from the explanation of geographyto the geographies within explanation 305Note 306

CLOKE-PRELIMS.qxd 3/3/04 12:06 PM Page viii

CONTENTS

ix

10 Understanding 307Introduction 307Seven modes of understanding 310Conclusion: between understanding and explanation 335Notes 335

11 Representing human geographies 336Introduction 336The work of writing 337Presenting research 342Representation and rhetoric 347Representation in practice 358Beyond the book 362Conclusion 363Notes 363

12 The politics of practising human geography 364The ‘personal’ politics of geographical practice 364The politics of research practice 367The politics of the academy 370Ethics, morality and geographical research 374Notes 375

References 376

Index 409

CLOKE-PRELIMS.qxd 3/3/04 12:06 PM Page ix

x

This book is inspired by these observations fromLinda McDowell, about whom we will say morein Chapter 1. While written over a decade ago,they can still be mobilized to frame the concernsof the present book, recognizing as they do a bodyof writing that was starting to accumulate from theearly 1990s as an explicit commentary upon theresearch methods deployed by human geographers.Over the intervening years ‘a greater self-con-sciousness about research methods’ has thereforebecome more commonplace in the literature, lead-ing to a small industry of textbooks and essays,some very good, on this theme. Even so, there isarguably still much more to be done in this respect,not least to make accessible to a wider audiencemany of the complex issues bound up in the veryacts of ‘doing’ or, as we like to term it, the practis-ing of human geography.

Our purpose in the present book is hence is totake seriously the many tasks entailed in conduct-ing research on given processes and problems, cer-tain types of societies and spaces, and nameablepeoples and places. We wish to ask about what isinvolved in such research, how it happens ‘in thefield’, whether a village under an African sky, ahousing estate drenched by Glasgow drizzle or theseeming safety of a local planning department ordusty historical archive.We will on occasion querywhat precisely is meant by this so-called ‘field’, butmore significantly we will ask about what exactlyit is that we do in field locations: what sources arewe after, what methods are we using in the process,

and how exactly do we manage to extract ‘rawdata’ ready to be taken away for subsequentdetailed interpretation? What kind of interactionsoccur here, particularly with the people fromwhom we are often trying to obtain information,whether they be the gatekeepers of sites and doc-uments that we wish to access or, more signifi-cantly, the people whose lives in specific spaces andplaces we are hoping to research (our ‘research sub-jects’).And what goes on once we do get the databack to our office, library or front room: how dowe endeavour to ‘make sense’ of these materials, tobring a measure of order to them, to begin manip-ulating them to describe and to explain the situa-tions under study, to arrive at the cherished goal ofunderstanding? Moreover, if we get this far, what isthen at stake as we try to write through our find-ings, to offer our interpretations, and as we pro-duce accounts which purport to represent peoplesand places more or less different from our own?

Can all this really be as simple as many earlierhuman geographers tended to assume? Can it allbe taken for granted, subsumed under headingssuch as ‘intuition’ or the enacting of the ‘scientificmethod’, and is it genuinely free from any rela-tionship to the researcher’s own values and beliefs,ethics and politics? We would want to answer thelatter questions with a definitive no and, as thebook unfolds, to indicate why we suppose this tobe the case by striving to provide answers to manyof the previous questions just raised. In line withMcDowell’s observations, therefore, we are

Preface

In the last few years, there has been an exciting growth of interest in questions about what wedo as human geographers and how we do it. Reflecting the general shift within the socialsciences towards a reflexive notion of knowledge, geographers have begun to question the con-stitution of the discipline – what we know, how we know it and what difference this makes bothto the type of research we do and who participates in it with us either as colleagues or researchsubjects … An intrinsic part of these debates has been a greater self-consciousness about researchmethods. (McDowell, 1992a: 399–400)

CLOKE-PRELIMS.qxd 3/3/04 12:06 PM Page x

convinced that human geographers do need tobecome still more self-conscious, more reflexiveand more willing to ‘reflect back upon’ all aspectsof their research practices.

We should underline the specific contributionthat we are aiming to make in this book, then, inthat it is designed to provide human geographyresearchers – from the undergraduate to the moreseasoned academic – with an introduction to themany and varied considerations integral to thepractising of human geography.We are not review-ing all the near-infinity of possible data sourcesopen to human geographers, since such an attempthas already been made (Goddard, 1983) and thereare a number of specialist texts dealing with par-ticular sources such as censuses, inventories andpublished surveys that may be of use to the humangeographer (see various papers in the CATMOGand HGRG series and also the contributions inFlowerdew and Martin, 1997). Neither are weoffering a complete ‘cookbook’ of methods, goingsystematically through a range of methods in turnand outlining how to do them, although there willbe moments in what follows where we dwell onspecific methods available for both (as we will say)constructing geographical data and constructing geograph-ical interpretations. There are already many such‘how-to-do’ manuals in the general social scienceliterature, and there is also something of this char-acter in several human geography texts.1 We woulddefinitely see such texts as complementary to ourown, but having a different feel in their focusedattention on the nuts and bolts of specific methodsthrough which human geographers both gatherand process data.

It might be noted that numerous older geogra-phy textbooks lead school children and undergrad-uate students through hands-on methods of fieldsurvey, land surveying and mapping (e.g.Dickinson, 1963), but we must admit to regardingsuch an interest in what has been termed ‘practicalgeography’ (Bygott, 1934) or ‘mathematical geog-raphy’ (Jameson and Ormsby, 1934) as beyond ourpurview.This is less the case with more recent con-tributions to the use of statistical, mathematical andGIS procedures in geography, which we do touchupon in Chapters 7 and 8, but we do not assessthese in technical detail because such treatmentsare provided elsewhere by specialist quantitativeand GIS geographers (for references, see Chapters7 and 8).We would also have to acknowledge, per-haps controversially, that we do not think that suchdevelopments are central to contemporary human

geography. They undoubtedly generate useful‘tools’ to be deployed on occasion, and we cer-tainly applaud the notion of combining quantita-tive and qualitative methods (see also Hodge, 1995;Philo et al., 1998), but we do not see how what arebasically technical exercises can be taken as morethan a small part of the larger whole which is thepractising of human geography. Our own prefer-ences, and maybe prejudices, are of course hintedat by such a statement (and see below). What weshould also underline at this point is our rejectionof the oft-made assumption that utilising qualita-tive methods entails a loss of rigour in the researchprocess, a forsaking of the objective, analytical andreplicable attributes supposedly integral to deploy-ing quantitative methods within the strictures ofthe conventional ‘scientific method’ (see Chapter 9).We resist the accusation that the route to qualita-tive methods amounts to a ‘softening’ of humangeography, where ‘softening’ is understood asweakening and making things easier (e.g.Openshaw, 1998).We regard such views as flawedbecause quantitative work (and science more gen-erally) is just as shot through with human frailtiesand social conditioning as is qualitative work, aclaim increasingly borne out by social studies ofscience and technology (e.g. Demeritt, 1996), andalso because it is possible – as we hope to demon-strate – for qualitative human geography to bepractised in a manner combining its own versionof intellectual rigour with a responsibility to the real-ities (not merely the academic’s inventions) of thefield beyond the academy.

Our intention is also not to provide a compen-dium of ‘stories from the field’, relating the experi-ences of particular human geographers as they havesought to operationalize substantive research pro-jects, although we do recognize the great value ofsuch personalized accounts.2 We therefore makesome use of such experiential materials at variouspoints in what follows. Furthermore, our intention isnot to rehearse the complex arguments about eitherabstract moral stances or ‘ethical philosophies’ whichmight be brought to bear in the discipline (see Sayerand Storper, 1997; Proctor, 1998; Smith, D., 1994;1997; 1998), although we are attuned to more specificissues rooted in the ethics and politics of research prac-tice (e.g.Mitchell and Draper,1983a; see also Mitchelland Draper, 1983b). This is especially true withrespect to claims about the ‘positionality’ of theresearcher, ones which talk about researchers need-ing to reflect self-critically on the power relationsrunning between them and their research

Preface

xi

CLOKE-PRELIMS.qxd 3/3/04 12:06 PM Page xi

Preface

xii

subjects, and we introduce such claims in Chapter 1and then throughout many of the followingchapters.

Neither is our immediate aim to show howpractical dimensions of research link with moreconceptual orientations, the latter being identifiedby such daunting terms as ‘positivism’, ‘Marxism’,‘humanism’, ‘feminism’, ‘postcolonialism’, ‘post-structuralism’, ‘postmodernism’ and the like. Ahandful of works do prioritize this linkage, notablyHoggart et al. (2002) in their attempt to tease outthe role of epistemological differences – meaningvariations in the concepts adopted by differentresearchers in their attempt to arrive at reliableknowledge about the world – as key influences onwhat can be achieved in the researching of humangeography. Something similar certainly is offeredon occasion in the book (giving a link back to aprior text by Cloke et al., 1991; see also Robinson,1998; Winchester, 2000; Shurmer-Smith, 2002a),and we would insist that some familiarity with thediscipline’s recent theoretical concerns is essentialto the self-critical practice of research being urgedhere. None the less, we will seek to introduce con-cepts in as gentle a fashion as possible, trying not toassume too much prior knowledge on the part ofthe reader, and we should make clear that the bookarises from six different authors who are themselvesdifferently persuaded (or dissuaded) by the merits ofdifferent conceptual orientations. If there is a com-monality between us in this regard, it lies in a con-viction that the overall conceptual landscape ofcontemporary human geography is a healthy one,and that there are now available all manner of excit-ing concepts to guide researchers in their practicalenquiries. Such concepts must remain as ‘guides’,since integral to the whole ethos of this book is thesuggestion that researchers should always be reflectingself-critically on every component of their research,concepts included. If given concepts do not appearto ‘perform’ well in helping the researcher to get togrips with the particularities of an issue, situation,space,place or whatever, then they should be revisedor abandoned. (We do realize, though, that recog-nizing whether or not a concept ‘performs’ well inthis respect is not quite as simple as such a remarkmight imply.)

From the above listing of what the book is not,we hope that an impression may be emerging ofwhat it actually is. It is, then, a book that – draw-ing upon experiences of research, contemplatingthe ethics and politics of research, and insisting onseeing research practices in the context of wider

conceptual orientations – does aim to stand back alittle from the hurly-burly of getting actual researchprojects planned and executed. It wants to inject apause in the battling with logistics, officials, respon-dents, tape-recorders, statistical tests, software pro-grams and the like, even though such everydaythings at the ‘coal face’ of research will be men-tioned repeatedly. It wants instead to ponderarguably deeper questions about the why, how,what, when and where of the research process,probing more fully into precisely what it is thatresearchers are searching for and aiming to workupon (this ‘stuff ’ called data), and then discussing thediffering strategies for ‘making sense’ of these data(for forcing them to ‘make sense’) and for convert-ing them into written-through accounts, findingsand conclusions for various audiences (from the dis-sertation marker to the academic conferenceparticipant). It is indeed to explore at some lengththe dynamics of practising human geography and, inso doing, to offer something distinctive which isneither a treatise on theory in the discipline nor a‘how to do it’ manual for disciplinary practitioners,but rather a theoretically informed reflection on themany different twists and turns unavoidably presentin the everyday ‘doing of it’. Moreover, in partexploring these dynamics should help to inform us,and our readers, when critically reading the writ-ten-through research of others.

The chapters that follow take different cuts atthis goal, and they differ in how they do thisaccording to the specific concerns of the chapter,the contents of the relevant pre-existing literature,and the particular competencies, interests and the-oretical predispositions of the author(s) who havehad prime responsibility for each chapter(although we are not going to tell you who didwhat!). We willingly acknowledge that there issome unevenness between the chapters, someoverlaps, doubtless some omissions, a few varia-tions in emphasis, even a few differences of opin-ion, but such inconsistency is also very much partof both the ‘real’ human geography (of the world)under study and the ‘real’ human geographers(from the academy) who struggle to study it.

A ‘map’ of the book

Following this Preface, Chapter 1 sets the scene forthe rest of the book by sensitizing readers to thedifferent ways in which human geography can be

CLOKE-PRELIMS.qxd 3/3/04 12:06 PM Page xii

practised, first by contrasting the extremes ofpractice by two semi-fictional human geographers(‘Carl’ and ‘Linda’) and, secondly, by sketching outa thumbnail history of changing practices withinthe ‘doing’ of (human) geography.While not wish-ing to imply that older practices were entirely mis-taken and have nothing still to teach us, we arecritical of the extent to which earlier geographerstended to regard the research process as relativelyunproblematic, as either the ‘natural’ way of pro-ceeding for individuals gifted with geographicalinsight or the ‘logical’ way of proceeding if obeyingthe basic rules of conventional science. In the courseof the chapter we introduce a number of key termswhich are utilized throughout the book – termssuch as ‘research’,‘field’,‘data’,‘methodology’ and thelike – and we also introduce several key themes todo with the heightened attention which humangeographers are now showing (rightly, in our view)to the complexities involved in both the ‘field’ andthe ‘work’ elements of ‘fieldwork’.

The heart of the book is divided into two parts,and we propose a division between two funda-mentally different sets of practices relating to thetreatment of human geographical data. By this phrasewe simply mean data (in the plural) which havebeen, are now and could in future be used byhuman geographers, and it does not necessarilymean data which are obviously related to space,place, environment and landscape (the staple bigconcerns of academic geography: see Hubbardet al., 2002; Holloway et al., 2003), although in thevast majority of cases such geographical referenceswill be involved somewhere (see Chapter 7). Itshould be noted too that we will often speak of‘geographical data’ rather than of ‘human geo-graphical data’, but this is merely to avoid thesomewhat cumbersome appearance of the latterterm, and throughout the book we will almostalways mean specifically data pertaining to humanor social dimensions of the world. Our discussionwill be of little immediate relevance to physicalgeographers, except in so far that there are overlapsin Chapter 1 with the history of changing practiceswithin the ‘doing’ of physical geography.We shouldadditionally emphasize that at every turn in thebook our concern is for processes of construction:accenting that data, how we come by them, and allthe many procedures which we then operate uponthem, from the most basic of sorting to the mostcomplex of representations, are all in one way oranother constructions found, created and enactedby ‘us’ (human geography researchers) as people

living and working within specific economic,political, social and cultural contexts.There is noth-ing natural here; nothing straightforwardly pre-given, preordained or untouched by humanemotions, identities, relations and struggles.

Part I is entitled ‘Constructing geographical data’,and here we tackle two very different varieties ofdata: first, those data which are ‘preconstructed’ byother agencies (governments, companies, journalists,poets and many more: see Chapters 2–4) and fromwhich human geographers can then extract materi-als relevant to their own projects; and, secondly,those data which are ‘self-constructed’ through theactive field-based research of human geographersthemselves (when using methods such as question-nairing, interviewing, observing and participating:see Chapters 5 and 6). In the first instance, the focusis very much on how these sources are constructed,on the many different contexts, influences andforces embroiled in the putting together of suchsources, which can be both purportedly ‘factual’ (asin parliamentary reports or news footage) and seem-ingly ‘fictional’ (as in novels or films). (We are cer-tainly aware that there is no clear separationbetween paying attention to how sources are con-structed and then trying to interpret what they aretelling us, underlining that the boundary betweenour Parts I and II is not a hard and fast one.) In thesecond instance, the focus is much more on the rolesplayed by researchers when in the field, and beginsto ask about the methods which can be utilized togain a window on the lives of research subjects, par-ticularly by talking or spending time with them.Quite specific questions hence arise about suchmethods, about how practically to put them intooperation, but so too do a host of questions aboutthe relations which inevitably run betweenresearchers and the researched, thus prompting care-ful reflection on matters of power, trust, responsibil-ity and ethics (in short, on the micro-politics ofengaged research). In addition, we appreciate thateven at this stage of research human geographers inthe field will begin to write about that field, notingdown preliminary findings, their personal experi-ences and their thoughts on how the research isgoing and on their interactions with research sub-jects. In Chapter 6 we briefly discuss this oftenunremarked feature of the initial research process. Itis in the course of such jottings that the seeds ofmore developed interpretations start to emerge, andit is also at this stage that the ‘textualization’ of ourresearch – the conversion of it into written formsfor wider audiences – begins to occur.

PREFACE

xiii

CLOKE-PRELIMS.qxd 3/3/04 12:06 PM Page xiii

Preface

xiv

Part II is entitled ‘Constructing geographicalinterpretations’, so as to stress that what we are talk-ing about now is indeed still very much a creativeact of construction, of making something, and isassuredly not some magical process whereby inter-pretations drop fully formed out of data.What weare terming ‘interpretation’ covers various strategiesthrough which human geographers endeavour to‘make sense’ of their data, the so-called ‘raw data’which they have collected or generated in one wayor another, and thereby to provide accounts, arriveat findings and posit conclusions. Somewhat hesi-tantly, we distinguish between five interpretativestrategies that could be conceived of as being com-plementary, but some of which in practice oftenend up being positioned as antagonistic byresearchers with particular investments in claimingone strategy as fundamentally superior to another.The ‘wars’ between such strategies is an underlyingtheme of these chapters, although in large measurewe feel such wars to be unhelpful and even anunnecessary distraction from what ought to be thehigher goals of arriving at good interpretations.Thefirst of our strategies, which we simply term ‘siftingand sorting’ (Chapter 7), cannot avoid being presentin any study, even if it is rarely given explicit con-sideration, and it entails the basic activities throughwhich a measure of order is imposed on raw databy the identifying of relevant entities in, to use ashorthand, ‘lists and boxes’. The second strategy,which we term ‘enumerating’ (Chapter 8), isinevitably a follow-on from sifting and sorting, andit entails the whole panorama of numerical meth-ods which are commonly utilized to measure prop-erties of distributions and to detect patterns withindata sets. In this chapter we deal with techniques ofnumerical analysis which have until recently beentaken as the core of geographical interpretation, butwhich we reckon warrant less special considerationthan has usually been the case.Although many willnot agree, we regard enumeration as essentially adescriptive activity, describing quantitative data sets,their differences from one another and possiblerelationships, in a manner that requires further stepsto be taken in translating back from the formalvocabularies of statistics, mathematics and comput-ing into the ordinary languages familiar to mostreaders.Only when such translation occurs can it besaid that the research has shifted from description tosomething that we, and many if not all otherphilosophers of science, would accept as somethingmore clearly explanatory.

Our third strategy is therefore what we term‘explaining’ (Chapter 9), where we consider whathas been the prevailing objective of so muchhuman geography raised in a ‘scientific’ vein,whether the science be that of a Newton, a Freudor a Marx, wherein successful interpretation (andhere the term ‘analysis’ is often heard) involvesbeing able to answer ‘why’ questions by specifyingthe causal processes combining to generate partic-ular human geographical phenomena. Our fourthstrategy,which we term ‘understanding’ (Chapter 10),points to what has recently become a more popu-lar approach with human geographers inspired bythe humanities and cultural studies, wherein suc-cessful interpretation involves being able to eluci-date the meanings that situated human beings holdin regard to their own lives and that inform theiractions within their own worldly places. In addi-tion, we appreciate the range of debates whichhave recently raged over the so-called ‘crisis ofrepresentation’, the deep worries about whatexactly happens when academics start to write orto lecture about peoples and places other thantheir own, and in Chapter 11 we review some ofthe conventions, rhetorics and other considerationsarising as human geographers seek to representtheir interpretations to different audiences. Just asall studies cannot avoid containing a moment ofsifting and sorting, so all studies, unless never writ-ten through in any form, cannot but include andusually culminate in acts of representation.We canregard representation as an interpretative strategyin its own right, if being itself fractured by manydifferent assumptions about the relations runningbetween ‘word and world’, but in other respectswe prefer to regard it as a moment that indeedcrops up as an adjunct to the various practices out-lined in all of the earlier chapters (including thosein Part I of our book).

Finally, we bring things to a close in a chapter(Chapter 12) picking out a particular strain ofthemes which have been present, albeit not allthat expressly tackled, within the precedingchapters. Building upon comments in Chapter 1about the values of the researcher, notably thosewith an obvious political edge, this chapterexplores the politics of practising human geogra-phy, reviewing in particular the thorny questionswhich surface once researchers begin to reflectupon the entangled politics influencing theirdecisions about topics to study, data sources toconsult and field methods to deploy, and then

CLOKE-PRELIMS.qxd 3/3/04 12:06 PM Page xiv

interpretative strategies to bring to bear on thedata derived. Here we pay attention to the poli-tics of the research process itself, notably withrespect to the often highly uneven power rela-tions traversing the social realm, the academy andeveryday lives that cannot but contextualize thisprocess, energizing but also sometimes constrain-ing a researcher as he or she initiates and seeks topursue his or her preferred practices of data con-struction and interpretation.While not suggestingthat all researchers should nail their politicsclearly to the mast – many of us may not be socertain about our politics and may prefer to allowthem to change according to context – we are inno doubt that practising human geographersshould offer at least some self-critical reflectionon their own, if we can put it this way, ‘politicalinvestment’ in the projects which they conduct.

Paul ClokeIan Cook

Phil CrangMark Goodwin

Joe PainterChris Philo

All over the place, 2003

Notes

1 In the general social science literature see,for example, Blunt et al. (2003), Burgess(1984) and May (1997). For human geogra-phy texts see, in particular, Lee (1992), Rogerset al. (1992), Cook and Crang (1995),Flowerdew and Martin (1997), Lindsay(1997), Robinson (1998), Hay (2000), Kitchinand Tate (2000), Limb and Dwyer (2001) andHoggart et al. (2002).

2 See, for example, Eyles (1988a); see also sev-eral pieces in Eyles and Smith (1988), Nast(1994), Farrow et al. (1997), Flowerdew andMartin (1997) and Limb and Dwyer (2001).

PREFACE

xv

CLOKE-PRELIMS.qxd 3/3/04 12:06 PM Page xv

We have accumulated a mass of debts over the (far too) many years that this book has been ingestation and preparation, and we cannot begin to acknowledge all or even many of them. What wewill do, though, is to say a massive thank you to Robert Rojek at Sage for the patience of Job (andthen some) in waiting for the disorganized rabble that is the authorial team of this book to get its acttogether. More particularly, we want to thank David Kershaw for his great contribution to the copy-editing process: his efforts have certainly helped to bring some more discipline to proceedings, andthe book has been much improved as a result. Finally, we would like to thank all those at Sage whohave been involved in the final production of the book for their hard work in getting the final productto look as attractive (and coherent!) as (we hope that) it does.

Every effort was made to obtain permission for Figures 1.1 and 1.3 and for Box 1.6.

The following illustrations were used with the kind permission of:Allyn & Bacon (Boxes 5.5, 8.5)Andrew Sayer (Table 5.1, Figures 9.2 and 9.3)Association of American Geographers (Figures 9.6 and 9.7)Blackwell Publishers (Box 7.3 and Figure 7.3)Cambridge University Library (Figure 4.3)Cambridge University Press (Figures 7.1 and 7.2)Continuum International Publishing (Box 5.2)Meghan Cope (Box 10.1)The Countryside Agency (Table 8.1)Paul du Gay (Figure 4.1)Elsevier Science (Figures 9.8, 9.9)Hodder Headline plc (Figures 8.3, 8.4, Box 8.9, Figure 9.1)John Wiley & Sons (Box 8.7)Mansell Collection (Figure 3.6)Linda McDowell (Figure 1.2)The National Gallery (Figure 4.5)Nelson Thornes (Box 7.2)Open University Press (Box 5.1)Pearson Education Ltd (Box 8.6, Figure 8.2)Philips Maps (Figure 1.4)Pion Ltd (Box 8.3)Syracuse University Press (Figure 5.1)Taylor & Francis Books Ltd (Figure 4.2, Table 5.2, Figure 5.3, Boxes 5.3, 5.4, 5.7, 8.8)Verso (Figure 3.7)W.W. Norton & Co. (Box 8.4)

xvi

Acknowledgements

CLOKE-PRELIMS.qxd 3/3/04 12:06 PM Page xvi

Practisinghuman geography?

Let us begin by imagining two differenthuman geographers, one we will call Carl andthe other Linda. Both these are human geo-graphers who believe that at least part of whathuman geography entails is the actual ‘practis-ing’ of human geography: the practical ‘doing’of it in the sense of leaving the office, thelibrary and the lecture hall for the far less cosy‘real world’ beyond and, in seeking toencounter this world in all its complexity, tofind out new things about the many peoplesand places found there, to make sense of whatmay be going on in the lives of these peoplesand places and, subsequently, to develop waysof representing their findings back to otheraudiences who may not have enjoyed thesame first-hand experience. Both of them areenthralled, albeit sometimes also a littledaunted, by everything that is involved in thispractical activity. Both of them are convincedthere is an important purpose in such activity,both because it enriches their own accountsand because it can produce new ‘knowledge’which will be eye-opening, thought-provokingand perhaps useful to other people and agen-cies (whether these be other academics,students, policy-makers or the wider public).For both of them, too, this practical activity

is something they usually find enjoyable,fun even, and both of them would wish tocommunicate this importance and enjoymentof practising human geography to others.YetCarl and Linda go about things in rather dif-ferent ways, and it is instructive at the outsetof our book to consider something of thesedifferences.

For Carl, the approach is one which doesvery much involve packing his bags, leavinghis home, locking the office door and headingout into the ‘wilds’ of regions probably atsome distance from where he normally livesand works. In so doing he tries, for the mostpart, to forget about all the aspects of his lifeand work tied up with the home and theoffice: to forget about his social and institu-tional status as a respected member of thecommunity and senior academic, to forgetabout his relationships with family, friends andcolleagues, to forget about the books, reportsand newspapers which he has been reading,and to forget about the concerns, troubles,opinions, politics, beliefs and the like whichusually nag at him on a day-to-day basis. Inaddition, he is determined to leave with anopen mind, with as few expectations as pos-sible, and even with no specified questions toask other than some highly generalizednotions about what ought to interest geogra-phers on their travels. Instead, his ambition is

The Changing Practices of HumanGeography: An Introduction1

3151-Ch-01.qxd 3/2/04 2:36 PM Page 1

to become immersed in a whole new collectionof peoples and places, and to spend timesimply wandering around, gazing upon andparticipating in the scenes of unfamiliarenvironments and landscapes. He mightoccasionally be a little more proactive in chat-ting to people, perhaps farmers in the fields ashe passes, and sometimes he might even countand measure things (counting up the numbersof houses in a settlement or fields of terracedcultivation, measuring the lengths of streets orthe dimensions of fields). From this engage-ment, as Carl might himself say, the regionsvisited begin to ‘get into his bones’: he startsto develop a sense of what the peoples andplaces concerned ‘are all about’, a feel whichis very much intuitive about how everythinghere ‘fits together’ (notably about how theaspects of the natural world shape the rhythmsof its cultural counterpart or overlay), and anunderstanding of how the local environmentswork and of why the local landscapes end uplooking like they do.The impression is almostof a ‘magical’ translation whereby, for Carl,meaningful geographical knowledge aboutthese regions is conjured from simply being inthe places concerned, formulated by him asthe receptive human geographer from activi-ties which are often no more active than astroll, the drawing of a sketch, the taking of aphotograph and the pencilling of a few notes.And the magical translation then continues,perhaps on return to his office, when Carlbegins to convert his thoughts into writtentexts for the edification and education ofothers, and through which his particular feelfor the given peoples and places is laid outeither quite factually or more evocatively.Taken as whole, this is Carl’s practising ofhuman geography.

For Linda, the approach is arguably rathermore complicated. She is much less certainabout being able to manage a clean breakfrom her everyday world as anchored in her

home, her office and her own social roles andresponsibilities, nor from her prior academicreading, and nor from the accumulated bag-gage of assumptions,motivations, commitmentsand formalized intellectual ideas which swirlaround in her head. Moreover, her researchpractice, her fieldwork, may not take herphysically all that far away from the home orher office: she might end up researching peo-ples and places that are almost literally justnext door, or at least located in the estates,shopping centres, business premises and so on,in a nearby city. The separation of everydaylife from the field, the regions under study,which Carl can achieve, is not possible forLinda: indeed, it is also a separation aboutwhich she might be critical. And, whereasCarl aims to go into the field as a kind of‘blank sheet’, Linda’s approach depends onhaving a much more defined research agendain advance, not one that entirely prefiguresher findings but one that will incorporate def-inite research questions based around a num-ber of key issues (perhaps connected to priortheoretical reading). Like Carl, though, shedoes wish to become deeply involved withparticular peoples and particular places(which might be very specific sites such as ‘theCity’, London’s financial centre and its com-ponent buildings, rather than the much largerregions visited by Carl). She does want to getto know the goings-on in these micro-worlds,to become acquainted with many of the indi-viduals found in these worlds and to try ashard as possible to tease out the actions, expe-riences and self-understandings of these indi-viduals in the course of her research. Theimplication is that what she does is very much‘hard graft’ research, since she has to beextremely proactive in deploying specificresearch tools – perhaps questionnaires, butmore likely a mixture of documentary work,interviewing and participant observation (allof which will be covered in our book) – so as

Practising human geography

2

3151-Ch-01.qxd 3/2/04 2:36 PM Page 2

This term describes the overall process of investigation which is undertaken on particu-lar objects, issues, problems and so on. To talk of someone conducting research inhuman geography is to say he or she is ‘practising’ or ‘doing’ his or her discipline, butit also carries with it the more specific sense of a sustained ‘course of critical investiga-tion’ (POD, 1969: 703) designed to answer specific research questions through thedeployment of appropriate methods. The ambition is to generate findings which canbe evaluated to provide conclusions, and usually for the whole exercise to be reportedto interested audiences both verbally and in writing. It contains, too, the suggestionthat the exercise will be conducted in a manner critical of its own objectives, achieve-ments and limitations, although we will argue that, by and large, human geographershave been insufficiently self-critical in this respect. The term ‘research’ is now verywidely used in the discipline (e.g. Eyles, 1988a), and its relative absence from earliergeographical writing suggests that geographers prior to c. the 1950s and 1960s wereless attuned to the notion of producing geographical knowledge through premedi-tated and structured procedures.

Box 1.1: Research

to generate a wealth of data which will enableher to arrive at specific interpretations per-taining to the issues (or, to put in another way,at clear answers to her initial research ques-tions).There is perhaps less the magical qual-ity of Carl’s approach, therefore, in that thelabour allowing Linda to complete herresearch is much more evident and probablyrather more bothersome, wearisome and evenupsetting. The labour also continues to beapparent at the writing-up stage in that Lindareckons it vital to include sections explicitlyon the methodology of the research, includingnotes on its pitfalls as well as its advantages,alongside debating at various points the extentto which someone like her – given just whoshe is, her social being and academic status –can ever genuinely find out about, let alonearrive at legitimate conclusions regarding, theissues, peoples and places under study.Taken asa whole, this is Linda’s practising of humangeography: it differs enormously from Carl’s.

You should notice that several terms in thelast paragraph are italicized, and Boxes 1.1–1.4define and expand upon the meanings ofthese terms.They are crucial to the book, and

you should ensure that you understand thembefore proceeding. They are also crucial toour introduction, which will now continueby making Carl and Linda more real.We havetalked about them so far as fictional charactersthrough which we could illustrate differentapproaches to the practising of human geo-graphy, but we should also admit to having inmind two real human geographers, one pastand one living, who are Carl Sauer and LindaMcDowell. Carl Sauer (see Figure 1.1) was ageographer based for virtually all his career inthe Berkeley Department in California, andhis chief interests lay in the ‘cultural history’ oflong-term inter-relationships between whathe termed the ‘natural landscape’ and the ‘cul-tural landscape’, and in teasing out distinctivepatternings of human culture as revealed inthe mosaic of different material landscapesproduced by different human activities (agri-cultural practices, settlement planning, reli-gious propensities).1 For the most part, Sauerdisliked statements about both theory andmethodology (although see Sauer, 1956), andhe tended to regard the practising of humangeography (and indeed of geography more

CHANGING PRACTICES OF HUMAN GEOGRAPHY

3

3151-Ch-01.qxd 3/2/04 2:36 PM Page 3

This deceptively simple term – the field – normally refers to the particular locationwhere research is undertaken, which could be a named region, settlement, neigh-bourhood or even a building, although it can also reference what is sometimes calledthe ‘expanded field’ (as accessed in a few studies) comprising many different locationsspread across the world (see also Driver, 2000a; Powell, 2002). We would include here,too, the libraries and archives wherein some researchers consult documentary sources,which means that we are also prepared to speak of historical geographers researching‘in the field’. In addition, we suggest that the field should be taken to include not onlythe material attributes of a location, its topography, buildings, transport links and thelike, but also the people occupying and utilizing these locations (who will often be theresearch subjects of a project). As such, the human geographer’s field is not only a‘physical assignation’, but it is also a thoroughly ‘social terrain’ (Nast, 1994: 56–7), andsome feminist geographers (e.g. Katz, 1994) have extended this reasoning to insist thata clean break should not be seen between the sites of active research and the othersites within the researcher’s world (a claim elaborated at the close of this chapter). Thisbeing said, we do wish to retain some notion of the field as where research is practi-cally undertaken, but we fully agree that fieldwork must now be regarded as muchmore than just a matter of logistics. Instead, fieldwork should be thought of as encom-passing the whole range of human encounters occurring within the uneven social ter-rain of the field, in which case it is marked as much by social ‘work’ as by thepracticalities of getting there, setting up and travelling around.

Box 1.2: Field


4

‘Data are the materials from which academic work is built. As such they are ubiquitous.From passenger counts on transport systems to the constructs used in the most abstractdiscussion, data always have a place’ (Lindsay, 1997: 21). Data (in the plural) hencecomprise numberless ‘bits’ of information which can be distilled from the worldaround us and, in this book, we tend to think of data, or perhaps ‘raw data’, as thischaos of information which we come by in our research projects (whether from thephysical locations before us, the words and pictures of documentary sources, the state-ments made in interviews and recorded in transcripts, the observations and anecdotespenned in field diaries, or whatever). As we will argue, a process of constructionnecessarily occurs as these data are extracted from the field through active research,ready for a further process of interpretation designed to ‘make sense’ of these data (tosubstitute their ‘rawness’ with a more finished quality). Various kinds of distinction aremade between different types of data (see also Chapter 7), the most common of whichis that between primary and secondary data. The former is usually taken as data gen-erated by the researcher, while the latter is usually taken as data generated by anotherperson or agency, but we restate this particular distinction in terms of self-constructedand preconstructed data (see also the Preface and below). For us, therefore, primarydata should be taken to include everything which forms a ‘primary’ input from thefield into a researcher’s project (i.e. anything which he or she has not him- or herselfyet interpreted). These data can include highly developed claims made in a govern-ment report or well-thought-out opinions expressed by an interviewee, in effect inter-pretations provided by others, but they remain primary data for us because theresearcher has not yet begun to interpret them. We do not really operate with a notionof secondary data, therefore, except in so far as we might reserve this term for theinterpretations of primary data contained in the scholarly writings of other academics.

Box 1.3: Data

3151-Ch-01.qxd 3/2/04 2:36 PM Page 4

‘In the narrowest sense, [methodology is] the study or description of the methods orprocedures used in some activity. The word is normally used in a wider sense to includea general investigation of the aims, concepts and principles of reasoning of some dis-cipline’ (Sloman, 1988: 525). On the one hand, then, there are the specific methodswhich a discipline such as human geography deploys in both the construction and theinterpretation phases of research (including such specific techniques as measuring,interviewing, statistical testing and coding). On the other hand, there is the method-ology of a discipline such as human geography that entails the broader reflections anddebates concerning the overall ‘principles of reasoning’ which specify both how ques-tions are to be posed (linking into the concepts of the discipline) and answers are tobe determined (pertaining to how specific methods can be mobilized to provide find-ings which can meaningfully relate back to prior concepts). For some writers (includinggeographers: e.g. Schaefer, 1953; Harvey, 1969) there is little distinction betweenmethodological discussion and what we might term ‘philosophizing’ about the basicspirit and purpose of disciplinary endeavour, but we prefer to regard methodology inthe sense just noted, and hence as a standing back from the details of specific meth-ods in order to see how they might ‘fit together’ and do the job required of them. Inthis sense, our book is most definitely a treatise on methodology.

Box 1.4: Methodology

generally) as something fairly obvious,coming ‘naturally’ to those who happened tobe gifted in this respect. Linda McDowell (seeFigure 1.2) is a geographer presently based atUniversity College London, and her chiefinterests lie in the insights that feminist geo-graphy can bring to studies of ‘gender divisionsof labour’ as these both influence the spatialstructure of the city and enter into the day-to-day gendered routines of paid employment,in the latter connection paying specific atten-tion to senior women employed in the London-based financial sector.2 While McDowell hasnot written extensively about methodology,she has contributed significantly to thedebates currently arising in this connection(see 1988; 1992a; 1999: ch. 9), and it is appar-ent that for her the practising of human geo-graphy is something necessitating considerable‘blood, sweat and tears’.

Our reasons for now fleshing out thehuman geographers who are ‘Carl’ Sauer and‘Linda’ McDowell are various and, at onelevel, simply emerge from a wish to emphasize

that human geography is always produced byindividual, flesh-and-blood nameable peoplewhom you can see and perhaps meet. Theycould be you! But at another level, the differ-ences between ‘Carl’ Sauer and ‘Linda’McDowell are highly relevant to the broaderarguments which we are developing in thisintroductory chapter. Indeed, in what followswe take Sauer and McDowell as exemplars oftwo very different ways of practising humangeography which ‘map’ on to, respectively,older and newer versions of human geo-graphical endeavour that can be identifiedwithin the history of the discipline.We mustbe circumspect about such a mapping: aSauer-esque approach is still very much withus today, partly in the continuing works ofregional synthesis and description carried outby many who regard this as the highest expres-sion of the ‘geographer’s art’ (Hart,1982;Meinig,1983; Lewis, 1985); while a McDowell-esqueapproach does have its historical antecedentsin the use of certain clearly defined methods,such as questionnaires and interviews, long


5

3151-Ch-01.qxd 3/2/04 2:36 PM Page 5

before the current eruption of interest inputting such methods at the heart of humangeographical research (see below). Yet, webelieve that there is still some truth in theproposed mapping, and that a profoundchange has occurred in how human geogra-phers envisage and proceed with their practis-ing of academic research: a change which canbe indexed by contrasting the likes of Sauer

and McDowell. By the same token, we wishto resist the impression that older approachesare bad whereas newer approaches are good,an impression readily conveyed by ‘presentist’accounts which project a narrative of thingssteadily improving, progressing even, from aworse state before to a better state now.Thismeans that we still find there is much of valuein an older Sauer-esque orientation, in the


6

Figure 1.1 Carl Ortwin Sauer

Source: From Leighley (1963: frontispiece)

3151-Ch-01.qxd 3/2/04 2:36 PM Page 6

ideal of suspending one’s everyday and acade-mic concerns in the process of becomingimmersed in the worlds of very differentpeoples and places, and in no sense are weseeking to encourage an ‘armchair geography’unaffected by the wonderment, hunches andideas which strike the human geographer inthe field. Yet it would be wrong to deny thatwe are more persuaded by McDowell than weare by Sauer, and that the basic purpose of ourbook is very much inspired by the likes ofMcDowell – complete with her insistence onthe labour, messiness and myriad implicationsof actual research practices, all of which mustbe carefully planned, monitored, evaluatedand perhaps openly reported – than it is bythe more intuitive, magical,‘just let it happen’stance adopted by the likes of Sauer.

A thumbnail historyof practisinghuman geography

Leading from the above, and to frame whatfollows in our book, we now want to chart

something of the history of changes in thepractising of human geography. It is onlyrecently that serious attention has been paidto ‘aspects of disciplinary practice that tend tobe portrayed as mundane or localised, but thatrepresent the very routines of what we do’(Lorimer and Spedding, 2002: 227, emphasisin original). Various authors are now claimingthat we fail to appreciate much about our dis-cipline without recognizing that ‘geographicalknowledge [is] constituted through a range ofembodied practices – practices of travelling,dwelling, seeing, collecting, recording andnarrating’ (Driver, 2000a: 267). They furtherworry that many of our ‘knowledge-producing activities’, old and new, remainlargely absent from how we represent ourresearch, suggesting that ‘our products ofknowledge (our texts and even our emphasesin conversations of recollection) could domore to make available this tension of thepresent tense of the world’ (Dewsbury andNaylor, 2002: 254): meaning precisely thefraughtness of our actual practices as we dothem. It is in the spirit of trying to make morevisible such practices, and in so doing to assessthem critically, that we now turn to ourthumbnail history.

The history relayed here is not intended tobe a comprehensive one, particularly sincemore historiographic research is required toclarify the details of how human geography(and also geography more generally) has beenpractised by different practitioners duringdifferent periods and in different places. (Andnote that active research is required to findout about this history, even if it be researchwhose ‘field’ is the archive and whose ‘data’chiefly comprise the yellowing pages of writings,maps and diagrams produced by past adven-turers, explorers and academic geographers:see Boxes 1.2 and 1.3.) Our history shouldalso be read in conjunction with other worksmore specifically concerned with the historyof geographical inquiry (Cloke et al., 1991;


7

Figure 1.2 Linda McDowell

Source: Courtesy of Linda McDowell

3151-Ch-01.qxd 3/2/04 2:36 PM Page 7

Livingstone, 1992; Peet, 1998). The historythat we tell will be somewhat arbitrarily sep-arated into three different, roughly chrono-logical, phases: we focus chiefly on the first ofthese, for which Sauer is an exemplar in hispreference for immersed observation; andthen on the third of these, for whichMcDowell is an exemplar in her preferencefor what we will term reflexive practice basedas much on listening as on looking.Reference will also be made to a ‘middle’phase in which the practising of humangeography did begin to be problematized,rather than regarded as intuitively given, andhere we will mention the rise of a ‘survey’impulse which ended up being hitched to aparticular (and we will argue narrowly)scientific orientation. For each phase, we willoutline the basic details of what the geogra-phers involved were doing and arguing,before switching to offer some more evalua-tive comments about pluses and minuses thatwe perceive in their practices.

‘Being there’ and‘an eye for country’

Probably the most longstanding traditionwithin the practising of (human) geography,albeit one rarely considered all that explicitly,has been one which makes a virtue out of thegeographer being personally present in agiven place and thereby able to observe itdirectly through his or her own eyes. Thereare two interlocking dimensions to this tradi-tion: the travelling to places within whichthe geographer can become immersed, sur-rounded by the sights, smells and other sensa-tions of the places involved; and then theactual act of observation, the gazing uponthese places and their many components, thepeoples included.

Taking the first dimension of ‘being there’,few would dispute that the very origins of

something called ‘geography’ lie in the earliesttravels which people from particular localitiesbegan to make to visit other places andpeoples further away, and in how such peoplesubsequently returned to convey their ‘geo-graphical’ discoveries of these distant placesand people to their own kinsfolk. H.F. Tozer(1897) duly suggests that the origins of‘ancient geography’ must be found here, andhe stresses the impetus for particular societies –notably Ancient Greece – to ‘trace theincrease of the knowledge which they pos-sessed of various countries – of their outlineand surface, their mountains and rivers, theirproducts and commodities’ (Tozer, 1897: 1–2).Although it is unlikely that the ancient geo-graphers such as Strabo would have reportedentirely on the basis of what they ‘saw takingplace before their eyes’ (Tozer, 1897: 2), theyprobably aimed to witness as much as possibleand then to base the rest of their work on thefirst-hand observations of other travellers.Indeed, it is probably not too fanciful to pro-pose that a fairly direct lineage can be tracedfrom these earliest geographers, many ofwhom must have been intrepid adventurers,through to the vaguely ‘heroic’ figure – almosta kind of ‘Indiana Jones’ character constantlyjourneying to distant lands – which may stillbe associated with the role of the geographerin the popular imagination.

Even in more academic circles such anotion is not entirely absent, most notably inthe powerful motif of the geographer as‘explorer-scientist’ which many (especiallyStoddart, 1986; 1987) see as capturing theessence of academic geography’s origins andcontinuing purpose. Leading from the ‘Ageof Reconnaissance’ (c. 1400–1800) whenvoyages of discovery were attended by a grad-ual recovery of the lost navigational skills ofthe ancients, an academic ‘geographicalscience’ or ‘scientific geography’ began totake shape (Kimble, 1938; Bowen, 1981;


8

3151-Ch-01.qxd 3/2/04 2:36 PM Page 8

Livingstone,1992:chs 2 and 3).By the eighteenthand nineteenth centuries, European explorerssuch as Captain Cook were regularly takingscientists who talked of ‘geography’ on theirexcursions, while ‘geographers’ such asHumboldt were themselves mounting remark-able expeditions to the likes of Middle andSouth America. Through the endeavours ofsuch individuals, specifically their attempts ataccurate scientific description, measurementand specimen collection, the field-based pro-duction of geographical knowledge becamemore systematized, rigorous and the herald ofa formally instituted academic discipline(taught in universities and boasting its ownsocieties and journals: see Bowen, 1981;Capel, 1981; Stoddart, 1986: chs 2, 7–10;Livingstone, 1992: chs 4–7). Furthermore,organizations such as Britain’s RGS (RoyalGeographical Society) began to providedetailed guidance to explorer-geographers,offering more than just ‘hints to travellers’ inspecifying procedures of description, mea-surement and mapping which would enablereliable geographical findings to be procuredfrom their sojourns overseas (Driver, 1998;2000b). Many controversies attached to thisphase in geography’s history, however, andconsiderable debate surrounded the extentto which geographical knowledge derivedfrom the explorations could be trusted.Arguments duly raged both then and morerecently over issues such as the value ofwritings by ‘lady-travellers’ (Domosh,1991a; 1991b; Stoddart, 1991) and as to howto regard the bellicose activities of explorerssuch as Stanley who appeared to be moreagents of empire (and European conquest)than exponents of geographical science(Driver, 1991; 1992; Godlewska and Smith,1994). None the less, the undisputed core ofthis growing body of knowledge which wasincreasingly identified as academic geogra-phy remained the simple fact of ‘being

there’, of being present in the places, oftenfar-flung, under examination.

Such a notion has continued to be central toacademic geography, and to give one instanceit is interesting to read Robert Platt’s 1930sespousal of a ‘field approach to regions’ whichwilfully set its face against those in NorthAmerican geography who were then propos-ing some system to the practising of geography(see below). Sparked by a strong feeling thatone should ‘[g]o to the field when the oppor-tunity arises without worrying over lack ofpreparation’ (Platt, 1935: 170), he recounted anexpedition with students to the regions betweenJames Bay and Lake Ontario in Canada whichyielded impressionistic senses of these regionsrather than guaranteed accurate findings. Hethereby produced a species of regional geogra-phy organized as a narrative of the journey,reporting on what had been encountered enroute as a window on phenomena such asforestry and trading patterns, and in so doinghe offered an almost anecdotal evocationclearly spurred by personal experience of thesites and sights encountered:

There were no signs of human occupancenor animals of respectable size. The air wasbright and warm, and the scene pleasantexcept for one item which spoiled anotherwise agreeable environment: swarmsof insects from which we had no means ofescape, a few mosquitos and innumerablevicious flies. (1935: 153–4)

In this context it is appropriate to return toSauer, since he too evidently supposed that‘being there’ was essential for the good geo-grapher, a claim that surfaced in an early piecewhen declaring that ‘less trustworthy’ aresources ‘which have not been scrutinised geo-graphically at first hand’ (1924: 20). HereSauer’s favouring of field-based study, onepredicated on being immersed in the peoplesand places under study, was loudly exhorted,and precisely the same sentiments resurfaced


9

3151-Ch-01.qxd 3/2/04 2:36 PM Page 9

In 1945 Charlotte Simpson published a paper entitled ‘A venture in field geography’,summarizing ten years of ‘local geography’ fieldwork undertaken by school childrenand undergraduate students in one particular Gloucestershire village. She stressed therole of ‘observational work’, based on a field walk taking in a ‘viewpoint commandinga … larger area’ (1945: 35), and she outlined her sense of the discipline as ‘an intenselypractical subject, dealing with realities which can be experienced at first hand’ (1945: 43).This paper indexed a whole tradition of running locally based fieldwork for youngergeographers, and the mid-century British geographical literature is awash with notesand guides regarding fieldwork in schools.3 The establishment in 1943 of somethingcalled the Field Studies Council (Jensen, 1946) was important here in promoting theability of ‘reading a landscape’ (Morgan, 1967: 145), initially publishing a series ofcountryside Field Study Books (Ennion, 1949–52) and from 1959 sponsoring a specialistjournal called Field Studies (wherein geographers have often published papers). Whilesomeone like Coleman was seemingly obsessed by the need to make small childrentake long walks in the countryside, other writers had a clear sense akin to that of aSauer or a Wooldridge about why such activity keyed into the core concerns of thediscipline: ‘the landscape is our subject matter, so we must look at it first hand as wellas through the media of books, films and maps … The need is simple and should notbe expressed in quasi-philosophical terms’ (E.M.Y., 1967: 228).

Box 1.5: ‘Field geography’ and ‘local geography’

over 30 years later when he stressed the needto be ‘intimate’ with regions being researched‘in the course of walking, seeing andexchange of observation’ (1956: 296).

We will revisit the point about observationshortly, but for the moment it is revealing toadd that Sauer echoed Platt in proposing amore informal engagement with places freefrom too many of the trappings of formalregional ‘surveying’ (see below):

To some, such see-what-you-can-find field-work is irritating and disorderly since onemay not know beforehand all that one willfind. The more energy goes into recordingpredetermined categories the less likelihoodis there of exploration. I like to think of anyyoung field group as on a journey of discov-ery, not as a surveying party. (1956: 296)

The ideal, he added, was to be in the fieldachieving ‘a peripatetic form of Socraticdialogue about qualities of and in thelandscape’ (Sauer, 1956: 296). It is also

intriguing that, while noting the traditionwhereby the geographer ‘goes forth alone tofar and strange places to become a participantobserver of an unknown land and life’ (1956:296), he insisted as well on ‘the dignity ofstudy in the superficially familiar scene’(1924: 32) much closer to home.This line ofreasoning, which found a wonderfully Britishinflection in the stress on student studies of‘field geography’ and ‘local geography’ (seeBox 1.5), has since led to the emphasis inmany undergraduate geography courses onrunning field classes and field days within animmediate region or country (while overseasfield trips may be seen as inheriting the more‘global’ aspects of claims made by Sauer andhis like).Another result has been papers con-sidering the different forms of locomotionthat geographers might employ when onactive fieldwork, as in Salter’s (1969)neglected note about the value of ‘thebicycle as a field aid’!


10

3151-Ch-01.qxd 3/2/04 2:36 PM Page 10

The second dimension mentioned above,to do with observation, is obviously tightlylinked to the theme of ‘being there’. It, too,has certainly been a feature of geographicalinquiry down the ages, given that the wholestress on the witnessing of distant lands whichbecame codified in the RGS’s ‘how toobserve’ field manuals (Driver, 1998; 2000b)hinges upon the expectation that the individ-uals involved – whether lay folk, professionalvoyagers or academic geographers – will beable to see, to look, to gaze upon the peoplesand places visited. Most of the more method-ological remarks which can be found in theearlier literature of academic geographyhence concentrate on the observation issue,and it is telling to recall Platt’s simple state-ment that, once in the field near James Bay,‘we opened our eyes and looked around’(1935: 153). Sauer is again a sure litmus for theprevailing wisdom: ‘Geographic knowledgerests upon disciplined observation and it is abody of inferences drawn from classified andproperly correlated observations … We areconcerned here simply with the relevance ofthe observations and the manner in whichthey are made’ (1924: 19).

We will shortly review Sauer’s reference toboth classification and ‘properly correlatedobservations’, but at this point let us move tosimilar statements in his 1950s paper pivotingaround the remark that the ‘morphologic eye’allows the geographer to evince ‘a sponta-neous and critical attention to form andpattern’ (1956: 290):

The geographic bent rests on seeing andthinking about what is in the landscape,what has technically been called the con-tent of the earth’s surface. By this we donot limit ourselves to what is visually con-spicuous, but we do try to register both ondetail and composition of scene, findingin it questions, confirmations, items orelements that are new and such as aremissing. (1956: 289)

Underlying what I am trying to say is theconviction that geography is first of allknowledge gained by observation, that oneorders by reflection and reinspection thethings one has been looking at, and thatfrom what one has experienced by intimatesight comes comparison and synthesis.(1956: 295–6)

Sauer also described the propensity for geog-raphers ‘to start by observing the near scenes’(1956: 296), before making the above-mentioned comment about going forth ‘tobecome a participant observer of an unknownland and life’ (1956: 296). More recently, andquoting one of the passages above from Sauer,C.L. Salter and P. Meserve have advocatedgeographers compiling ‘life lists’ of their accu-mulated field visits and the like, concluding inthe process that:

To be a real geographer, one must observe.There is great power in observation. For ageographer, there are few skills moreimportant in intellectual growth than thedevelopment of the ability to ‘see what’sthere’. The act sounds so very innocent, yetbeing able to discern patterns on the hori-zons, what components make up the whole,significance in details, and whole from itsparts, represents a critical geographic skill.(1991: 522–3)

We will also return to Salter and Meserve ina moment.

The British literature is full of similarclaims advocating the centrality of observa-tion to the geographer’s craft, notably in thewriting of Sidney Wooldridge, where hecelebrates what he describes as ‘an eye forcountry’, which should be encouraged ingeographers from an early age:

I submit that the object of field teaching, atleast in the elementary stage, is to develop‘an eye for country’ – ie. to build up thepower to read a piece of country. This isdistinct from, though plainly not unrelatedto, ‘map-reading’. The fundamental princi-ple is that the ground not the map is the


11

3151-Ch-01.qxd 3/2/04 2:36 PM Page 11

primary document, in the sense in whichhistorians use that term. From this first prin-ciple I pass to a second, that the essence oftraining in geographical fieldwork is thecomparison of the ground with the map,recognising that the latter, at its best, is avery partial and imperfect picture of theground, leaving it as our chief stimulus toobserve the wide range of phenomenawhich the map ignores or at which it barelyhints. (1955: 78–9)

A class of young geographers, taken to aviewpoint in the field, should not be made topore over a map ‘instead of concentrating onthe work of looking and seeing’ (Wooldridge,1955: 79), and the unequivocal message wasthat ‘eye and mind must … be trained byfieldwork of laboratory intensity’ (1955: 82).In arguing this way, Wooldridge also insistedthat it was vital for refined powers of observa-tion to be inculcated in young geographersthrough fieldwork in the ‘little lands’ close tohome, and that the transmission of coregeographical skills ‘lies in the development ofthe laboratory spirit and the careful, indeedminute, study of limited areas’ (1955: 80).Such beliefs clearly urged the value of ‘beingthere’, and provided an even more forcefulassertion than did Sauer of the need forobservation-based fieldwork in the geogra-pher’s immediate locality.These were dominantthemes in mid-century British geography,informing the ‘field’ and ‘local geography’ ini-tiatives which emerged in schools (see Box 1.5),and they also featured in the efforts of some-thing called the ‘Le Play Society’, alongside itsinitially student offshoot called the ‘Geogra-phical Field Group’, which sought to encour-age British professional geographers in theconduct of rigorous fieldwork (Beaver, 1962;Wheeler, 1967). One ambition of the lattersociety was to get geographers out oflibraries, to curtail the practice of many whichinvolved little more than synthesizing factsabout regions from second-hand librarysources, and to foster in them an imitation of

geologists and botanists in achieving ‘thehighest qualities of observation and faithfulrecording in the field’ (Beaver, 1962: 226).Moreover, much was made of the role ofobserving landscapes in the field by an indi-vidual called G.E. Hutchings, who wished toblend into geography the skills of the ‘fieldnaturalist’, and also sought to provide somerigour to geographers in the oft-promotedbut rarely discussed art of landscape drawing(Hutchings, 1949; 1960; 1962: see Box 1.6).

Having laid out something of this highly per-vasive emphasis on ‘being there’ and its associ-ated ‘eye for country’, we should nowacknowledge that we see many problems withsuch a stance on the practising of humangeography (our criticisms would not neces-sarily apply to such a stance on the practisingof physical geography). This being said, weshould emphasize again that in no way wouldwe wish to deny the importance of spendingtime in the field, immersed in the worlds ofparticular peoples and places, and neitherwould we want to underplay the importanceof careful observation. Indeed, at variouspoints in the book we will have much to sayabout such matters, albeit expanding on themin ways which Sauer,Wooldridge and the likeprobably would not recognize.There are sig-nificant criticisms to be made, however, inpart to anticipate the alternative proposals of amore recent turn in the practising of humangeography.

Thinking first about ‘being there’ and,while not wishing to commit us to staying inour armchairs, there is perhaps a certain arro-gance in the assumptions of many older geo-graphers about their supposed right to be ableto travel widely, to visit wherever they wantedand to do their geographical research wher-ever they alighted. Such an arrogance alsoarises when the likes of Stoddart (1986)contemptously dismiss the likes of Wooldridgefor suggesting that much fieldwork should


12

3151-Ch-01.qxd 3/2/04 2:36 PM Page 12

G.E. Hutchings (1962: 1) once declared that he sought ‘to relieve the bookishness ofeducation with practice in observation and exploring out of doors’, thereby explaininghis preference for combining geography and natural science through the medium ofboots-on field study. He emphasized the importance of landscape as ‘something thathas to be viewed, whether scientifically or aesthetically’, and insisted that ‘it is verynecessary for the geographer to acquire by training in the field what Prof. Wooldridgecalls the “eye for country’’’ (1949: 34). And again, he stated that geography ‘is a kindof learning arising in the first place from curiosity about the visible and tangible world,and requiring a capacity for looking beyond the superficial appearance of things’(1962: 2). Revealingly, given Hutchings’s clear belief in geography as an observationalpursuit, he published a book on landscape drawing which was directed particularly atthe needs of geographers, in the course of which he gave technical hints about howto produce a sketch which ‘is an honest picture of a piece of country, drawn with closeattention to the form of its parts and the appearance of the various objects in itaccording to the effects of light and perspective’ (1960: 1).

A redrawn field sketch of the Conway Valley and the Afon Llugwy

Source: Hutchings (1960: 18–19)

Box 1.6: Field studies and landscape drawing

take place near to home, asserting instead thatthe wider world should be the geographer’sprovince. For many geographers the beliefthat the world is their ‘oyster’ has never beenquestioned, and the possibility that large por-tions of it are really somebody else’s world isnot one that is often addressed.We are not somuch talking here about the complicationgiven by national borders or legal ‘ownership’of land, although these can both be pertinentin some human geographical research, but,rather,we are talking about how the field – thespecific places to be visited, including thehuman settlements, homes, workplaces andsites of recreation – is somewhere that weshould perhaps be more hesitant to enter than

we have often been in the past. These areplaces where other people do live, striving toscrape a living and to make a life, and thesepeople may pursue all sorts of activities whichthey would want to keep private from theprying intrusions of outsiders (and a host ofconsiderations duly arise about the preserva-tion of cultural difference, the guarding ofspiritual and religious mores, the keepingsecret of illegalities and so on). In addition, theplaces involved might be ones which holddeep meanings for those people who occupyand make use of them, and the presence thereof outsiders, particularly intrusive ones takingphotographs and writing notes, could begreatly resented.


13

3151-Ch-01.qxd 3/2/04 2:36 PM Page 13

Objections on these counts to the geogra-pher as intrusive alien are increasingly comingfrom development geographers, persuaded inpart by the criticism that geographers work-ing in Africa, Asia, South America and else-where effectively reproduce the samestructural relationship with ‘native’ peoples ashad arisen in the expeditions of the colonialexplorer-geographers from earlier centuries: arelationship in which power, influence andassumptions of superiority lie with the whitegeographers appropriating knowledge, labourand skills from the peoples of colour in theseplaces (e.g. Sidaway, 1992; Madge, 1993;Powell, 2002).The relationship in question isneatly illustrated in Figure 1.3, which shows awhite explorer-geographer, perhaps Stanley,being carried across a river on the shoulders ofblack bearers. We guess that nothing like thishappens in the research of today’s develop-ment geographers, but is the presence of (say)Anglo-American researchers in the places oftheir black research subjects so completely freeof all the inequalities and embedded assump-tions which are coded into this illustration?

The seeming innocence of just ‘being there’can also be questioned in situations where geo-graphers are researching closer to home, sincethe activity of strolling into (say) a Cornishfishing village or an Alpine skiing resort to com-mence the work of immersed observation is

surely one that many people in these places –whether local villagers or people on holiday –might regard as an unwanted imposition.Moreover, while some human geographershave now started to study marginal groupingssuch as ethnic minorities, children and theelderly,‘Gypsies’ and other travellers and so on,it is certainly not obvious that the researcherarriving in the places of such groupings is agood thing for them. It does comprise anintrusion and an imposition, one that may bedeeply disturbing to the individuals and fami-lies involved, and one which could have direconsequences if reseachers made public certaininformation about their precise locations,movements and place-related activities (a con-cern constantly expressed by David Sibley inhis research on travellers: 1981a;1985).We real-ize that human geographers will want to con-tinue doing engaged work that requires themto be present in the situations of other peoplesand places, and we fully support this importantaspect of research, but we note too that –following the examples of critical developmentgeographers, McDowell, Sibley and manyothers – the apparent rightness of such researchpractice can no longer be straightforwardlyassumed. Rather, the picture must become oneof researchers negotiating access to peoples andtheir places, both by formally liaising with thepeoples concerned and by thinking much


14

Figure 1.3 White ‘explorer-geographers’ being carried across a river by black ‘native’bearers

Source: From Brice and Bain (no date, c. 1918: 34)

3151-Ch-01.qxd 3/2/04 2:36 PM Page 14

more carefully than hitherto about the politicsof ‘being there’ as bound up with the differingorigins, backgrounds, attributes and socialstandings of the human geographer relative tothese peoples and places.We will return to suchaccess issues again in this chapter, and thenagain later in the book.

Turning now to the issue of ‘an eye forcountry’, it should be explained that thereis now a sustained critique of the pervasive‘occularcentricism’ of much conventionalgeographical inquiry. Acknowledging thiscritique forces a thorough-going reappraisalof the obsessive advocacy of observationwhich figures in many of the statementsquoted above.There are various strands of thiscritique, but all of them converge on what isentailed in geographers setting themselves upas privileged observers able to gaze upon –and, more dubiously, to gaze down upon –peoples and places laid out before their eyeslike so many exhibition entries. Several histo-rians of the discipline have begun to examinethe observational technologies which geogra-phers have deployed, highlighting the extentto which visual images of landscapes arethemselves not so much innocent factualrecords as fabricated or ‘staged’ representa-tions. David Livingstone (1992: 130–3)assesses the ‘artistic vision’ which served tocompose many of the observations taken byeighteenth-century explorer-geographers,discussing the tensions which existed for bothscientific illustrators on the voyages andengravers back in Europe when trying tobalance the need for a faithful (empirical)rendition with the prevailing aesthetic tastesof the age:

Banks always felt a tension between the callof taste and that of pictorial reproduction,and so his painters did devote some of theirenergies to romantic topics like grottoes,exotic rituals and so on because thesesuited the then fashionable rococo style.Moreover, even when accurate depictions ofnative peoples … were provided, it just was

very hard to bring an objective account ofthem before the British public. Engraverswould dress up the original paintings tokeep them in line with their own philosophi-cal predilections. (1992: 131, emphasis inoriginal)

Tackling the role of photography in the ‘imag-inative geography’ of the British Empire, andas linked to the production of geographicalknowledge by nineteenth-century Britishexplorer-geographers (many of whom wereassociated with the RGS), James Ryan (1997:17) exposes the limits to the Victorian (and stillprevailing) assumption that photographycomprises ‘a mechanical means of allowingnature to copy herself with total accuracyand intricate exactitude’. Alternatively, Ryanfinds a suite of cultural constructions runningthroughout the photographic observations of‘distant places’ throughout this period, teasingout the ‘symbolic codes’ which structured thecomposing and the framing of the images, andalso hinting at the effects of these images ontheir audiences:

Through various rhetorical and pictorialdevices, from ideas of the picturesque toschemes of scientific classification, and dif-ferent visual themes, from landscape to‘racial types’, photographers representedthe imaginative geographies of Empire.Indeed, as a practice, photography did morethan merely familiarise Victorians withforeign views: it enabled them symbolicallyto travel through, explore and even possessthose spaces. (1997: 214)

The reference here to ‘possessing’ spacesthrough observation will be recalled in amoment, and it also suggests a key claim thatmight be pursued in critical accounts ofwhat is entailed in the much more recent useby geographers of technologies such asremote sensing (with its self-evident linksto military and commercial uses of suchtechnologies).

A second but related dimension of thecritique centres on the more metaphorical sense


15

3151-Ch-01.qxd 3/2/04 2:36 PM Page 15

in which many geographers have configuredthe world as an ‘exhibition’ for them to wanderaround, as it were, gazing upon the exhibits(the diverse collections of peoples and placesthere displayed) and making judgementsabout them. Indeed, Derek Gregory (1994;see also Mitchell, 1989) borrows the phrase‘world-as-exhibition’ when tracing thistendency through different phases andapproaches to geographical study, relating it aswell to the ‘cartographic’ impulse which hasled many geographers to conceive of them-selves flying over landscapes of nature andsociety laid out below them. Revealingly,some writers have even drawn connectionshere to the strangely detached sensationwhich arises when looking down from anaeroplane, and (more worryingly) from thebasket of a World War One balloon or thecockpit of a World War Two bomber (Bayliss-Smith and Owens, 1990). Gillian Rose(1993b; 1995), meanwhile, has developed apowerful argument that this version of an aca-demic gaze reflects a distinctively ‘masculinist’way of looking at the world, one predicatedon an assumed mastery which allows theviewer to see into all corners of the world –all of which are reckoned to be available andamenable to the gaze, transparent to the piercingintellectual eye – and one which also carrieswith it an inherent desire to possess, to sub-due, the phenomena under the gaze. Rose’sargument is difficult, hinging on a combiningof psychoanalytic ideas with a historicalaccount of how male intellectuals haveeffectively constructed science (geographyincluded) in their own (presumed) self-image.More simply, though, she outlines the mas-culinist propensities of fieldwork: both theheroic encounter of rugged individuals with achallenging field which is coded into the‘being there’ approach, notably of someonelike Stoddart (1986), and the associated figureof the male researcher observing, describing,

measuring and thereby capturing this field forhimself (see also Sparke, 1996; Powell, 2002:esp. 263).

Central to the objections raised by Gregory,Rose and others to the prominence of obser-vation is the suggestion that the faculty ofsight should not be accorded such a masterstatus in geographical work, whether in actualpractices or in how we conceptualize thewider projects of the discipline. One implica-tion is that other senses through which theworld is knowable by us, notably hearing andmore particularly still the practice of listeningclosely to what people say, should be broughtmore fully into our practice as human geogra-phers (see also Rodaway, 1994).4 A secondimplication is that we should resist thetoo glib deployment of terms saturated withassumptions about observation and sight inour thinking about the discipline, and in theprocess to resist a general orientation whichconceives of the discipline as trying to maketransparently visible all facets of the subject-matters under study.This is not the occasionto expand further upon such lines of criti-cism, nor upon their implications, althoughwe hope that readers will be able to appreci-ate how the third approach to practisinghuman geography described below does takethem on board.

Surveys andscientific detachment

It should not be thought that the practising ofhuman geography prior to recent years hasonly been about immersed observation, how-ever, and it is actually the case that efforts toprovide a more systematic basis for geographi-cal research – one going beyond intuitivefieldwork to develop a definite technique of‘survey’ – did figure in the history of thediscipline earlier in the century. Sauer himselfwas instrumental in starting this ball rolling


16

3151-Ch-01.qxd 3/2/04 2:36 PM Page 16

This distinction has become rather sedimented in the thinking of many geographers,perhaps to the point where it becomes unhelpful and overloaded with misunder-standing and prejudice (see Philo et al., 1998; see also Demeritt and Dyer, 2002).Qualitative data are data that reveal the ‘qualities’ of certain phenomena, events andaspects of the world under study, chiefly through the medium of verbal descriptionswhich try to convey in words what are the characteristics of those data. These can bethe words of the researcher, describing a given people and place in his or her fielddiary, or they can be the words found in a planning document, a historical report, aninterview transcript or whatever (in which case the words are in effect themselves thequalitative data). Sometimes these data can be visual, as in the appearance of a land-scape observed in the field (see Box 1.6), or as in paintings, photographs, videos andfilms. Quantitative data are data that express the ‘quantities’ of those phenomena,events and aspects of the world amenable to being counted, measured and therebygiven numerical values, and the suggestion is that things which are so amenable willtend to be ones which are immediately tangible, distinguishable and hence readilycounted (1, 2, 3, …) or measured (an area of 200 m2; a population of 20 000 people livethere; a per capita earning of £100 000). It should be underlined that such counts andmeasurements are still only descriptions of the things concerned, albeit descriptionswhich are arguably more accurate and certain than are qualitative attributions (chieflybecause they allow a common standard of comparing different items of data, and alsothe possibility of repeating this form of describing data: i.e. other researchers wouldcount the same number of things or measure the same areas, population levels, percapita incomes, etc.). The use of quantitative data is hence commonly reckoned to bemore objective (allowing researchers to deal with data in an accurate, certain andtherefore unbiased manner), while qualitative data are commonly reckoned to bemore subjective (leaving researchers prone to injecting too much of their own ‘biases’in their dealings with data). As should be evident from much of this chapter, and of therest of the book, we do not agree with such a conclusion because it forgets about thecountless other issues which militate against the possibility of complete objectivity(which means that being quantitative is no convincing guarantee of objectivity).

Box 1.7: Qualitative and quantitative

with his 1924 paper (see also Jones and Sauer,1915) which was entitled ‘The survey methodin geography and its objectives’, in the courseof which he urged geographers to developregimented and replicable methods of geo-graphical inquiry. The similarities betweenthis paper and his later writings notwithstand-ing, there are also key differences whichreflect the enthusiasm of the younger Sauerfor developing systematic principles of areallybased ‘geographic survey’ incorporating not

only qualitative materials but also quantitativeinformation, the latter being derived fromboth ‘statistical tables of state and nationalagencies’ (Sauer, 1924: 20) and ‘local statisticalarchives’ (1924: 30). (See Box 1.7 for a pre-liminary note on the distinction betweenqualitative and quantiative data: this distinc-tion, and its limitations, will be explored fur-ther in later chapters, especially Chapter 8.) Itis perhaps surprising to hear the youngerSauer’s own words in this regard:


17

3151-Ch-01.qxd 3/2/04 2:36 PM Page 17

The purpose … is not to make fieldworkmechanical, but to increase its precision.The choice of things to be observed mustremain a matter of individual judgement asto the significant relationship between areaand population. Out of such field measure-ments will come the … ideal of statisticalcoefficients. From them the geographer willdetermine ultimately the extent to whichthe theory of mathematical correlation is tobe introduced into geography. (1924: 31)

It is telling that Sauer linked this version offield survey to the possibility of a more statis-tically minded geography, one which by the1960s was regarded as the province of a fullyscientific discipline, and we will return to thislinkage shortly. The thrust of his reasoninghere was echoed and extended a year later inD.H. Davis’s (1926: esp. 102–3) rejection of‘superficial observations’ when calling insteadfor geographers to evolve a ‘mechanical qual-ity’ to their ‘system of recording essential dataaccurately’, one suitable for ‘establishing cor-relations’, which would then lead ‘geography …to be entitled to rank as a science’. Similar dis-cussions of survey as a scientific methodologyfor geographers can be found elsewhere inthe early- to mid-century literature, and it wasthese discussions, with their thinly veiled criti-cism of those who favoured a more impres-sionistic field style, that prompted both Platt’s(1935) reactions and certain reservations fromthe older Sauer (1956), as already mentioned.

More practically, several papers (e.g. Jones,1931; 1934) appeared in the North Americanliterature which began to itemize the kinds ofthings which needed to be recorded in acomprehensive field-based geographical sur-vey, the forms and functions of land uses to bemapped, as well as specifying the specific sur-vey methods which might be employed tocreate this record (field walks and drive-bys,complete with their counting and mapping ofphenomena, along with collecting statisticaldata from ‘local depositories’).A review of ‘fieldtechniques’ available for use by geographers in

their surveys of areal units was provided byC.M. Davis (1954), and a feel for the groundcovered by this remarkably thorough earlystatement of survey methodology can begained from these claims in the paper’s open-ing paragraph:

There are four sources of factual informa-tion: 1) documents such as maps, groundphotographs, statistics and written materials;2) air photographs; 3) direct observation;and 4) interviews with informants. Andthere are four ways of analysing factualinformation for the purpose of identifyingand measuring areal, functional or causalrelations, each requiring the use of symbols:1) analysis by expository methods, usingword symbols; 2) analysis by statistical meth-ods, using mathematical symbols; 3) analysisby cartographic methods, using map sym-bols; and 4) analysis by photo-interpretationmethods, using photo-interpretation keys.(1954: 497)

Davis began to suggest a distinction betweenthe data to be collected (constructed in ourterms: see below) and the procedures throughwhich those data can be analysed (interpretedin our terms), and he also indicated that thedata collected could be qualitative or quanti-tative, with implications for the sorts of ana-lytical techniques to be deployed on thesedata from the field. Equivalent practical state-ments also appeared in the British literature,many of which effectively hovered betweenthe celebration of an unsystematized ‘beingthere’ stance and providing systematic guid-ance about what should be found out inlocally based field surveys (see Box 1.5).Theemerging ‘field studies’ movement whichhooked into academic geography in variousways (e.g. Hutchings, 1962; Morgan, 1967;Yates and Robertson, 1968) articulated avision not far from the survey orientation ofsome North American geographers, and theGeographical Field Group (descended fromthe Le Play Society) was expressly committedto ensuring that ‘[o]bservation, direct inquiryand documentation, including statistical


18

3151-Ch-01.qxd 3/2/04 2:36 PM Page 18

material, all contribute to the data-collectingprocess’ (Edwards, 1970: 314; and note thatthis group conducted a ‘series of “regionalsurvey” type excursions’ described inWheeler, 1967: 188). In an edited collectionon the geography of Greater London, A.E.Smailes described ‘urban survey’ as thedetailed recording of information about townsites from field-based observation (‘reconnais-sance survey’) gleaned from ‘traversing thestreets’ (1964: 221).Additionally, in a manoeuvreparalleling the survey cataloguing recom-mended somewhat earlier in North Americaby Wellington Jones (1931; 1934), Smailesproposed a specific ‘urban survey notation’which produced a pseudo-quantitative formof data logging ready, presumably, for moresophisticated statistical analysis (see Figure 1.4).Whatever the precise details of the survey sys-tems developed by these scholars, however,what we would immediately emphasize istheir list-like, box-filling, counting and map-ping ambitions: ones reflecting the primaryambition of the geographers concerned toaccumulate data through which they couldcharacterize the areas and sub-areas understudy.

It is true that there were some qualitativeelements here, as in the significance occasion-ally placed on talking to field ‘informants’ (seeChapter 5), but the basic trajectory of thesurvey approach was none the less towards aself-proclaimed scientific orientation. Theprime ambition was to conduct systematicsurveys which would produce comprehensiveand reliable quantitative data representativeof areal units (whether these be regions aslarge say, the Paris Basin or as small as, say,Glasgow’s West End). There was also thebeginning of suggestions about being able toconduct statistical analyses on these quantita-tive data, perhaps by using standard statisticaltests to establish the strength of correlationsbetween different sets of data (i.e. to showthat certain areas are marked by high values

on different variables, say income levels,occupancy rates and car ownership).We havealready noted D.H. Davis’s (1926) explicitlinking of such surveys to a version of geo-graphy which could claim the name of‘science’, while James Anderson (1961) playedup the role of survey work in the context ofland-use classifications as enhanced by ‘statis-tical probability sampling procedures’ and theuse of computers. In such a vision surveywork would contribute to providing the dataon spatial patterns, chiefly data which mightthen be deployed in the process known as‘regionalization’, the supposedly scientificdelimitation of regions or areas fundamentallydifferent from one another, and then in aprocess of classifying different types of regionsaccording to certain distinctive clusters ofattributes (e.g. Philbrick, 1957; Grigg, 1965;1967; see also Chapter 7).

There is a complicated story to tell in thisregard, but at bottom a continuity can bedetected from the quantitative impulse in thissurvey work and the rise of geography as spa-tial science from the 1950s onwards (seeLivingstone, 1992: chs 8 and 9; see alsoBarnes,2001a).Spatial science,as is well known,entailed a fusion of quantitative techniques


19

Roof material

PeriodWall material

Function No. of floors

Detail of function

N U M ERAT O R

DENO MIN ATOR

Figure 1.4 A.E. Smailes’s pseudo-scientific‘urban survey’ notation

Source: From Smailes (1964: Figure 41, 204)

3151-Ch-01.qxd 3/2/04 2:36 PM Page 19

with a form of locational analysis aiming toexplicate the basic ‘spatial laws’ governing theorganization of phenomena (human behav-iour and productions included) across theearth’s surface.5 In Chapter 8 we say muchmore about enumeration in geography, butfor the moment it is sufficient to emphasize theextent to which quantification became thefavoured way of going about things in geogra-phy as spatial science, linked into a particularmodel of how statistical tests (and then moreformalized mathematical modelling) could allaid in the explanation of revealed spatial pat-terns (and see also Chapter 9).

All manner of claims were made for thesuperior merits of spatial science, completewith its quantitative sophistication, and at thenub of such claims was the assertion that aproperly ‘scientific’ approach to research wasone which overcame the potential ‘bias’ ofresearchers in ensuring the completelydetached (and hence certain, accurate andtrustworthy) cast of the research undertaken.In particular, the use of numerical values asmeasures of quantity, distance, position andthe like was reckoned to provide an objectiverepresentation of what was actually happeningin the ‘real world’, in contradistinction to themuch less reliable data obtained throughthe subjective understandings integral to boththe intuitive stance of a Sauer or the conver-sational elements of some survey work. (SeeBox 1.7 for a summary of the tangled debatesabout objectivity and subjectivity.) Such werethe assumed advantages of a spatial-scientificpractising of human geography, one whichgrew out of the above-mentioned surveytradition, but which came to embrace amuch wider set of procedural, technical andexplanatory goals. As a coda, and anticipatingsome of our arguments in Chapters 7 and 8,this version of human geography continuestoday, notably in the development and appli-cation of geographical information systems(GIS) and various forms of geocomputation.

Arguably, there is a level of sophisticationabout these more recent approaches to quan-titative geography that was absent in the earlydays of spatial science.

Having laid out something of this scientificand survey approach to practising humangeography, we should acknowledge that heretoo we see many drawbacks with what wasbeing proposed. Many of these drawbackswere bound up with the overall philosophicaldifficulties attaching to spatial science, parti-cularly as have been rehearsed throughexposing the somewhat narrow ‘positivist’philosophical assumptions which can be said –certainly in retrospect (Gregory, D., 1978a;Hill, 1981; Barnes, 2001a; 2001b) – to haveframed this scientific turn within the disci-pline. At various points in our book we willengage with these problems, demonstratingthe ways in which they arguably hamper thepractising of human geography, although – aswith the ‘being there’ and ‘eye for country’stances – we are not denying that someaspects of the scientific and quantitative turnsstill have much to offer in the doing of humangeography (and in a similar vein, see Sayer,1984; Philo et al., 1998). None the less, we areconcerned at the extent to which method-ological treatises in human geography con-tinue to be dominated by an exposition ofspatial-scientific techniques (e.g. Lindsay,1997; Robinson, 1998): a reduction ofmethodology to matters of technique.6

Similarly,we are concerned about a rather nar-row sense of what ‘geographical enumeration’can entail which fails to look much beyondstandard parametric tests. Such considerationsremain pertinent to the practising of humangeography, to be sure, but what concerns us isthe lack of serious engagement with moreconceptual questions about the limitations ofwhat a self-professed scientific and quantitativehuman geography can and cannot achieve.Such questions are raised throughout this


20

3151-Ch-01.qxd 3/2/04 2:36 PM Page 20

book, even if not always being convenientlylabelled as such (to reiterate: Chapters 7–9 alldebate these questions in one way or another).

What we will specifically underline now,since it is so relevant to the third phase in thehistory being recounted, is that we are deeplysuspicious of the claims about the detachmentof the researcher which are celebrated in theliteratures of scientific, quantitative andsurvey-based human geography. We are per-turbed by the determined erasure of the ‘I’,the researching individual or group, whichserves (in our minds) only to occlude the reali-ties of the active research progress throughwhich flesh-and-blood geographers such as‘Carl’ and ‘Linda’ actually get their feet, heartsand minds muddied in the places and peopleunder study. Spatial science, along with bothits antecedents and its derivatives, thus closesoff the possibility of debating the practising ofhuman geography in the fashion of this book.As just remarked, spatial scientists tend toreduce methodology to technique, beingbothered about the correct running of anappropriate statistical test but less about any-thing entailed in the deriving of the data onwhich the test is conducted (unless relevant todeciding on which particular test is suitable),nor about anything following conceptually,politically, ethically or otherwise from choos-ing to tackle the data statistically rather thanin some other way.There is a further and pos-sibly simpler objection to raise in relation tothe appearance of spatial science, moreover, inthat it evidently led many human geographersto lose interest in field-based primary data,given that they rapidly became far moreinterested in the enticing array of statistical-mathematical techniques available (and beingrefined) for analysing secondary data (see Box 1.3).As Robert Rundstrom and Martin Kenzerneatly put it:

Although quantitative human geographerswere primarily concerned with abstract theorydevelopment [specifying the spatial laws of

location theory], many of the early spatialanalysis papers … were based on fieldwork.The pattern changed by the middle of the1970s. Continuing progress in spatial analy-sis was marked by theoretical developmentsrelying on pre-existing data. Primary databecame superfluous. Ackerman (1965)already considered fieldwork a mere chore,only occasionally necessary to validate theanalytic, theoretical work of spatial science.James and Mather (1977) noted that somehuman geographers questioned whetherfieldwork was still a necessary part of thediscipline. (1989: 296)

Instead of going out into the field to collectdata, many spatial-scientific human geogra-phers started to spend the bulk of their timesitting in their offices and laboratories, punch-ing in data found in library sources (e.g.Census surveys), reworking their own olderdata or even inventing data sets, as a preludeto the real work (for them) of using computerfacilities to effect statistical-mathematicalinterrogations, simulations and model-building.To put it another way, ‘economic geography[and human geography more widely] movedfrom a field-based, craft form of inquiry to adesk-bound technical one in which placeswere often analysed from afar’ (Barnes, 2001b:553). The practising dimension of theirinquiries therefore collapsed into the tech-niques, reinforcing our previous argument,with a loss of concern for nuances of data, thecomposition of the field or the overallresearch process. In the terms of this book (seethe Preface), this meant a loss of interest in theconstruction of data in preference for focusing,albeit very narrowly, on the interpretation of(quantitative) data. It is an exaggeration, butperhaps not too great a one, to state that thisversion of human geographical inquiry ceasedto practise human geography except in themost minimal of senses. Neither a Sauer nor aMcDowell would find much here to satisfythem, and the same is probably true ofmany human geographers today who con-tinue to use numerical data and sophisticated


21

3151-Ch-01.qxd 3/2/04 2:36 PM Page 21

statistical-mathematical procedures but alwayswith an alertness to the origins, meanings andlimits of the numbers and their manipulation(e.g. Dorling, 1998).

‘Being reflexive’and ‘listening to voices’

From about 1970 onwards, many humangeographers, unhappy about both olderapproaches to the discipline and the spatial-scientific version, began to seek for new pos-sibilities.There were numerous bases to theirquarrel with how human geography wasbeing practised at the time, centring chieflyon the limited conception of how humanbeings entered into the ‘making’ of their ownworlds, but also on the almost completeabsence of what might be termed a ‘political’vision of why research was undertaken in thefirst place (who was it supposed to benefit andwhy?). While somewhat oversimplifying thepicture, it can be argued that these twinobjections to previous approaches, and mostespecially to spatial science, fed into tworather different alternative varieties of humangeography – to be referred to here respec-tively as ‘humanistic geography’ and ‘radicalgeography’ (see Cloke et al., 1991;Peet, 1998) –which both demanded new ways of practisinghuman geography. Indeed, their emergenceand subsequent elaboration, particularly whenmixed in with the insights from ‘feministgeography’ from the mid-1980s onwards, haveeffectively called forward a sensibility almostwholly unheard of before in the discipline. Inshort, this sensibility can be described as‘being reflexive’, which means that humangeographers are now called upon to reflectmuch more explicitly upon their own researchendeavours than hitherto, giving careful con-sideration to precisely what it is that they aredoing in their own projects: the conceptual,practical, political and ethical implicationsarising for these projects, for themselves, for

the people and places under study, and perhapseven for society more generally.

We would argue that, while not often giventhis credit, humanistic geography was decisivein prompting the developments leading to thenew sensibility just mentioned. Humanisticgeography was an umbrella term which arosein the 1970s (esp. Entrikin, 1976;Tuan, 1976a;Ley and Samuels, 1978) to denote a range ofperspectives highly critical of how mosthuman geographers, but most obviously spa-tial scientists, tended to conceptualize humanbeings. Extending an earlier ‘behavioural’ turnin the discipline and drawing upon variousso-called ‘philosophies of meaning’ (Ley,1981a; for summary details, see Cloke et al.,1991: ch. 3), the humanistic geographers com-plained bitterly about the ‘pallid’ view ofhuman beings present in existing scholarship(Ley, 1980): one that portrayed human beingsas little more than mere objects or at bestrobots with no interior sense of themselves,no intentions, no hopes or fears and no cre-ative role to play in shaping their surround-ings. Instead, so they insisted, the disciplineneeded to be dramatically reforged around avery different conception of humanity, avision which recognized humans in all theirflawed ambiguities as experiencing, perceiv-ing, feeling, thinking and acting beings. Sucha vision sought to enlarge the ‘space’ forhuman beings within the discipline, to grantthem a measure of dignity, to ‘people’ humangeography; in fact, to foster a new emphasison the human part of human geography.Theintellectual terrain here was uneven, but oneover-riding outcome leading from thisexpanded conception of the human beingwas the need to find ways of accessing thehuman qualities, the sheer humanness, nowreckoned to be central to disciplinary con-cerns. Spurred by a changed appreciation ofwhat is important in the world under study –people and their inner lives, rather than spatialpatterns and supposed spatial laws – the


22

3151-Ch-01.qxd 3/2/04 2:36 PM Page 22

humanistic geographers had to consider freshresearch practices, novel stances before theirsubject-matters and unfamiliar methods forgetting close to people and their everydayapprehensions, understandings, routines andactivities. It meant starting to use methodswhich provided some structure to the tasks ofmeeting with people, perhaps interacting withthem on an everyday basis, perhaps talkingwith them in depth and certainly ‘listening totheir voices’. It meant rediscovering the ques-tionnairing and interviewing techniques ofthe earlier survey tradition but, more signifi-cantly, it meant bringing into human geo-graphical research the ‘ethnographic’ practicesof in-depth interviewing, participant observa-tion and the excavation of meaning whichwere much more the province of other acad-emics such as anthropologists and sociologists.It meant returning to a measure of immersedobservation in the vein of Sauer, but it alsomeant a much more sustained encounter withthe peoples in the places visited. It requiredthe thoroughly involved people-centred field-work which led the likes of John Eyles(1985), Michael Godkin (1980) and GrahamRowles (1978a; 1978b; 1980) to spend days,weeks and months in the company of indi-viduals, witnessing the grain of their livedworlds, discussing with them the meaningsattached to places, environments and land-scapes comprising the spatial contexts of thesesmall worlds.

A key figure in all this was undoubtedlyDavid Ley, whose famous exploration ofinner-city Philadelphia, including an attemptto recover the existential meanings of placeand ‘turf ’ held by black street gangs, blendedthe ideas of a humanistic geographer, theinterests of a social geographer and the prac-tices of an ‘urban ethnographer’ (Ley, 1974).Following in part the example of ethnogra-phers associated with the Chicago School ofurban sociology (Jackson, 1985), but growingas well from his response to ‘the relentless barrage

of everyday pressures in the inner city’ (Ley,1988: 132), Ley created his own distinctiveversion of what he later came to term ‘inter-pret(at)ive social research’:

With limited experience to fall back on,aside from the intuition gained froma British field tradition [presumablyWooldridge’s ‘eye for country’] and knowl-edge derived from a number of ethnogra-phies, devising a method was in part amatter of learning on the job. The principalmethod was participant observation. Theperiod from January to July was set out asthe length of continuous residence in theneighbourhood … It is essential to establisha systematic procedure for recording fielddata in any ethnographic research, and mypractice was to write up field notes eachevening, ranging in size from a paragraphto (occasionally) 1000 words [see alsoChapter 6]. These notes were records ofimpressions, events and conversations,sometimes reconstructed from brief phrasesor sentences scribbled down during thecourse of the day. (1988: 130)

The ‘unstructured, everyday encounters’which generated this data were then supple-mented with more formal face-to-face ‘ques-tionnaire interviews’, some taping of publicmeetings and some reading of ‘agency dataand documents from police, planning andschool board authorities’, as well as extensivefield observation on phenomena such as ‘graf-fiti, vandalised cars or abandoned properties’(Ley, 1988: 130–1).The result was an eclecticmix of data sources and methods of both datacollection (construction) and analysis (interpreta-tion), containing both quantitative and quali-tative moments, and it broke new ground increating a model for practising human geo-graphy which has since been widely emulated.The majority of later researchers would prob-ably not call themselves humanistic geogra-phers, preferring instead labels such as socialor cultural geographer (see Jackson andSmith, 1984, or the various studies reported inJackson, 1989; Anderson and Gale, 1992), butthe basic procedures deployed by them did


23

3151-Ch-01.qxd 3/2/04 2:36 PM Page 23

arrive in the discipline with the humanisticexperiments of someone like Ley (and notethat participant observation was explicitlyclaimed as a prime method for humanisticgeography by Smith, S.J., 1981; 1984; see alsoJackson, 1983). It might be added that the ver-sion of inquiry being progressively refined inthis vein has since also been termed ‘interpre-tative geography’ (Eyles, 1988b) or ‘interpre-tative human geography’ (Smith, D.M.,1988b), and several fine examples of practicehave been collected together in books editedby John Eyles and David Smith (1988) andmore recently by Melanie Limb and ClaireDwyer (2001).

Ley has remarked that ‘[e]thical issues arefar more conspicuous in ethnographicresearch because of the close relationshipbetween the researcher and the community’(1988: 132), thereby indicating that the newforms of research initiated by humanisticgeography have forced into the open ques-tions about the researcher’s personal involve-ment with a project, people and place. Whatare the ‘biases’ of the researcher? How dothese influence how he or she conducts theresearch, how he or she represents the peoplesand places studied in a write-up, and theinformal ‘contract’ which he or she strikes upwith a community about what is done, saidand finally given back? All these questionsstart to concern the researcher in a mannerwhich had never been the case for previousgenerations of geographers, and certainly notfor spatial scientists, and the ‘ethics and values’of doing geography thus become a subject fordebate as never before (Mitchell and Draper,1983a; 1983b). On an operational level, anextreme instance is reported in Rowles’s(1978b: esp. 179) study of the geographicalexperiences of elderly people, and entailedhim sitting at the deathbed of one of his agedrespondents, Stan. Here, as the only ‘friend’whom Stan had left in the world, the onlyperson remaining who cared enough to sit in

that hospital room, Rowles found himselfurging Stan not to die because the researchwas still incomplete. But he hated himself forthinking in this way: was this all that it hadbeen about, getting the research done, andhow hollow, how intrusive but meaninglesshad been his ‘befriending’ of Stan for the pur-poses of academic research? The ethics – theturmoil, stress and guilt – attendant upon sucha moment were light-years away from any-thing experienced by, say, spatial scientistsworking on impersonal numerical data sets atthe computer terminal.This is not to suggestthat there are any simple guidelines for howresearchers should pick their way through theethical minefield which can confront them inresearch of this character. As Ley (1988: 133)acknowledges, ‘[t]here are no set pieces inanswering those questions, and indeedanswers will vary according to the circum-stances of the community’, but what can beinsisted is that in specific studies ‘the questionsmust be asked and answered in good faith’.This is not to demand that researchers alwaysput down in writing their thoughts, worriesand responses on this count, but it is to pro-pose that they should be able to offer somerelevant reflections if ever challenged to do so.

The researcher’s presence as an ‘I’, a creativeand reflexive figure in the research processwho is not erased as a non-issue (as might aSauer) or cloaked behind a veil of claimedobjectivity (as might a spatial scientist), istherefore as much a part of this approach tohuman geography as are theories, data,methodsand so on. On the still deeper level of theresearcher’s underlying interests, convictionsand motivations, furthermore, the appearanceof humanistic geography was also crucial inforegrounding such values in a fashion rarelyif ever seen before in the discipline. To putthings another way, humanistic geographyprompted attention not just to the subjectivityof the researched but also to the subjectivityof the researcher. Writing in a determinedly


24

3151-Ch-01.qxd 3/2/04 2:36 PM Page 24

scientific manifesto for human geography,Ronald Abler et al. (1971: 24) had declaredthat the scientific ‘way of life’ should be ‘total’,and that it should be completely divorcedfrom how scientists might ‘let their hair downemotionally and theologically … during theiroff hours’.The latter aspects of their lives, ofwho they are and of what they feel or believe,should hence be roped off from their efforts asprofessional human geographers.

But writing only three years later, AnneButtimer, a humanistic geographer who wasalso then in religious orders, advocated some-thing entirely different when insisting thatsuch a compartmentalization of the humangeographer’s personal and professional faces ismischievous because it just cannot happen,since personal values can never be systemati-cally erased from the framing, conduct andwrite-up of research. And for Buttimer suchan erasure should never be attempted anyway,being unnecessarily restrictive because itdenies many of the well-springs of genuinehuman concern and creativity, and alsodepriving us of crucial grounds for sensibleethical judgement. Science alone cannot pro-vide those grounds, so she argued, and thedangers of a science without ethics nowbecome increasingly obvious. Her alternativeproposal was quite clear:

Each reader [each human geographer]should endeavour to explore the valueswhich guide/influence his [or her] mode ofbeing in the world, for it is the contentionof this paper that one’s geography cannotbe considered a separate domain of one’slife but is influenced by many personal, cul-tural and political ‘values’ surrounding thatwork. (1974: 5)

Buttimer duly reflected upon many of thevalues which shaped her own geography,pointing out that they were ‘strongly influ-enced by Christian, and especially existentialthought’ (1974: 5), and she thereby counteredAbler et al.’s (1971: 24) pronouncement that‘God … is not permitted’ as a ‘concept’

informing human geographical research. Shesupposed that all human geographers wouldentertain different, idiosyncratic assemblagesof values, making generalizations difficult, butshe also recognized the importance of supra-individual intellectual, ‘cultural and political’values which can themselves become the focusof careful ‘sociological’ scrutiny (1974: Part II).While many have disagreed with the specificsof her arguments here (e.g. see the four com-mentaries appended to Buttimer, 1974), ourview is that her insistence on human geogra-phers being reflexive about the diverse valuesshaping their own work is one which contin-ues to resonate loudly with more recent effortsat practising human geography.

Alongside humanistic geography, a self-professed radical geography emerged duringthe 1970s (Peet, 1977; 1978), anchored ini-tially in the pages of Antipode: A RadicalJournal of Geography, and subsequently diffusingto become a wide-ranging critical windowon social and spatial inequalities of manydifferent shades. Taking as its focus environ-mental and social problems with a clear geo-graphical expression, from the devastation ofrainforests to poverty, deprivation and dis-advantage, a radical-geographical perspectivearose which sought to expose the systematicstructuring of injustice which leads to a worldfragmented into spaces of plenty (occupied by‘the haves’) and spaces of deficit (occupied bythe ‘have-nots’). Starting with approacheswhich did little more than document, tableand map this polarity at various scales fromthe international to the intra-urban, radicalgeographers gradually evolved a conceptualbasis for explaining these inequalities whichincluded (sometimes contradictory) inputsfrom welfare economics, anarchist theory anddifferent strands of Marxism.7 Some of theresearch undertaken in this vein retained asurvey feel, albeit utilizing survey techniquesto expose inequalities in phenomena suchas income levels, housing conditions and


25

3151-Ch-01.qxd 3/2/04 2:36 PM Page 25

ill-health indicators (this was particularly trueof something called ‘welfare geography’:Smith, D.M., 1977; 1979; 1988b). Indeed,much of the research continued in a fashionnot wholly different from the more scientificand quantitative cast of previous work, even ifputting the data and techniques to a radicaluse critical of the social status quo, and even iffuelled by a commitment to radical (evenrevolutionary) social change wholly absentfrom the more ‘establishment’, often policy-orientated studies of previous generations,notably of spatial scientists.We will return toexamine this commitment presently.

It may be claimed, then, that radical geogra-phy did not usher in as dramatic a change inthe routine practices of human geographers asdid humanistic geography. Its immediatemethodological implications were not sogreat, even if conceptually and politically itwas probably more unnerving to establishedmodes of inquiry. None the less, mentionmight be made of the ‘advocacy geography’experiment associated with William Bunge’sattempt to shift the orbit of professionalgeographical research out of the universitycampus – together with the geographersthemselves, and also their students – and intothe streets of the inner city, chiefly the blackinner city of Detroit, where the aims ofstudies should be directed by the articulatedneeds of poor inner-city residents themselves(see Colenutt, 1971; Horvarth, 1971; Bunge,1971; 1975; Merrifield, 1995). Rather thanoffering yet another calibration of a ‘centralplace model’, for instance, advocacy geogra-phers should be uncovering the geographiesof slum processes, traffic accidents affectingchildren, diseases of babies and the like, andacting as advocates able to demonstrate thecontours of problems to city authorities whomight be sufficiently convinced by academicevidence to respond positively. Failing that, thegeographers involved should be themselvesinvolved in grassroots projects like building a

children’s playground. This species of radicalgeography thus urged an action-basedresearch, predicated on full involvement in aresearch activity designed to achieve highlypractical ends: a policy-orientated researchfrom below,on behalf of those who might withjustification be referred to as ‘the oppressed’.

Intriguingly, such research did have things incommon with humanistic geography in that itdepended upon a sustained participation inthe lives and struggles of certain inner-citycommunities – Amaral and Wisner (1970)spoke of ‘participant immersion’ instead of aless involved ‘participant observation’ – andalso because it forced researchers to deal withconcrete ethical issues rooted in their respon-sibilities towards the relatively powerlesspeople whom they were supposed to be serving.David Campbell expressly reflected upon ‘rolerelationships in advocacy geography’, underlin-ing the virtues of a thoroughly democraticpractice resistant to the ‘elitism’ common inmost other work by human geographers, andstriving instead to empower the researchsubjects who should ‘become problematisers oftheir [own] situations and … active creators oftheir [own] environment’ (1974:103).Moreover,and echoing still further the ethical charge ofhumanistic geography:

Constant self-criticism and re-evaluation inan attitude of humility and respect forothers is … a vital and healthy componentof advocacy activity … ‘Humanising socialchange’ [to borrow a phrase from Harvey] isdependent … upon the ability of advocatesand academics to create humanising rela-tionships with those whom they work …Radical science must be based upon a per-sonal commitment to genuine communica-tion with others in an attitude of mutualrespect. (Campbell, 1974: 104–5)

The radical, politicized overtones of theseremarks must have been anathema to themore ‘conservative’ geographers of the era,but they are ones with which many humangeographers today would have great sympathy,


26

3151-Ch-01.qxd 3/2/04 2:36 PM Page 26

and the notion of democratic, empoweringand respectful research practice will certainlyreappear at various points in the chapters thatfollow.

More generally, the rise of radical geogra-phy, particularly in a guise which turned toMarxist critiques of the inequalities integral toa globalizing capitalist economic order, carriedwith it explicit commitments to a coherentpolitical programme: one which oscillatedbetween a reformist line, supposing that theexisting state of society can be improvedthrough the standard democratic process, anda revolutionary call for complete social trans-formation (at its crudest a call for the workersto seize control of the ‘means of production’from the capitalists). David Harvey (1973) ledthe way when self-consciously shifting from abasically reformist line, associated with a wel-fare position, to an assertively revolutionaryline convinced that the only way to create true‘social justice in the city’ would require anoverturning of capitalist forms of urbanism.This latter way forward would also necessitatean input from, if not necessarily a Leninistintellectual vanguard, then certainly a corpusof Marxist academics, geographers included,prepared to undertake the theoretical andpractical work of planning revolutionarychange. Radical geographers ever since havebeen wrestling with this tension betweenreformist and revolutionary ambitions, as isclear from recent debates played out in thepages of the journal Society and Space (Blomley,1994; Chouinard, 1994; Tickell, 1995), and afurther feature of debate has been the seeminggulf between radical theorizing in the acad-emy and radical activism on the streets (seealso Routledge, 1996; Farrow et al., 1997;Kitchin and Hubbard, 1999).

The principal point for us here, though, isthat – just as Buttimer insisted on humanisticgeographers incorporating explicit reflectionson their basic values – radical geographershave often entertained some self-interrogation

about political values, objectives andinvolvements. Perhaps the most rigorousformulations in this respect have emergedfrom Jürgen Habermas (esp. 1972), a famousGerman Marxist intellectual, who has pro-posed the refining of an explicitly criticalscience or theory fitted to achieve ‘the realisationof a[n] … emancipatory interest’ (Gregory, D.,1994: 107, emphasis in original) which wouldfree all peoples of the world from the yoke of(capitalist) oppression. Habermas’s visionexplains how all varieties of intellectuallabour are determined by ‘cognitive interests’which turn their practices of knowledge pro-duction to particular ends, usually ‘technical’or ‘practical’ ones functional to the main-tenance of the social status quo (and an exten-sion of his argument would include allvarieties of human geography, including bothspatial science and humanistic geography, asessentially ‘reactionary’ in this sense: seeGregory, D., 1978a). Following from such arecognition, however, the argument is that itshould then be possible to frame a new ver-sion of intellectual endeavour predicated onan emancipatory cognitive interest whichwould be at once critical (of an inherentlyunjust world) and self-critical (constantly eval-uating the extent to which the academic’sown practices are emancipatory in both over-all design and specific interventions). Whilerarely presented in such obviously Habermasianterms, except in Gregory’s (1978a) statementsabout ‘committed explanation in geography’,it is arguably the case that this notion of beingsimultaneously critical and self-critical hasenergized the efforts of most radical geogra-phers over the last two decades or so.We willpick up on the arguments about the politicsof human geography, returning to some of thematerials just outlined, in our final chapter(Chapter 12).

It would be possible to say more here aboutthe burgeoning twists and turns in humangeography which have, more recently, built


27

3151-Ch-01.qxd 3/2/04 2:36 PM Page 27

upon the twin pillars of humanistic and radicalgeography to forge further dimensions forpractising human geography. But for the sakeof brevity, and yet to cover what have beenpivotal new claims relevant to our practisingtheme, it will suffice now to mention certainaspects of the interlocking contributionsmade by both ‘feminist geography’ and ‘post-colonial geography’. Feminist geographyinitially arose to provide an explicit examina-tion of the specific spatial experiences, con-straints and worlds of women, that ‘other half ’of humanity almost never considered by pre-vious generations of male geographers(Tivers, 1978), and it quickly developed as amore fundamental critique of how unequalgender relations shape countless sociospatialstructures endemic to a diversity of ‘patri-archal’ human societies past and present(McDowell, 1983; Foord and Gregson, 1986;WGSG, 1984; 1997).

In the process questions of how to do fem-inist geography inevitably came to the fore,particularly in the matter of thinking abouthow research could be carried out whichwould enable the voices of women to beheard, notably when recounting their experi-ences of an everyday male superiority, harass-ment and even abuse accepted by many ofthem as sadly ‘natural’. The task also becameone of finding methods which would besufficiently sensitive to tease out often verysubtle dimensions to how women’s perceptionand use of space differs from that of men,whether in terms of a phenomenon like the‘gender division of urban space’ (McDowell,1983) or something like women’s fear ofpublic spaces such as parks and subways(Valentine, 1989). It has been argued bothwithin (esp. Rose, 1993b) and beyond thediscipline that conventional models of academicinquiry, with their scientific and quantitativeemphases, display an inherent ‘masculinism’which militates against the kind of groundedresearch which is probably essential in this

context. The debates are tricky, but we canallow McDowell to be our guide in a passagewhich also signals the character of the specificmethods that feminist academics, geographersincluded, have tended to favour in their ownresearch:

Certain feminists … not only reject thequantitative, ‘scientific’ approach toresearch, but argue that it is specifically apatriarchal model as it denies the signifi-cance of women’s experience of oppression,classifies their concerns as private ratherthan shared, and embodies the values oftraditional views of women’s and men’sexpected positions in society. They haveargued that feminist research should recog-nise and challenge the everyday experiencesof women. In order to excavate women’sexperiences, feminist methods should valuesubjectivity, personal involvement, the qual-itative and unquantifiable, complexity anduniqueness, and an awareness of the con-text within which the specific [issue] underinvestigation takes place. (1988: 166–7)

The onus shifts towards being a highly person-alized research encounter, in which the mostqualitative of methods such as in-depth inter-viewing and the taking of ‘oral histories’ of ‘lifestories’ (see Chapter 5) are pursued in a man-ner which necessarily entails a sustainedexchange – potentially one dealing with emo-tional materials – wherein the personalities ofboth researcher and researched cannot be arbi-trarily suspended. For both parties involved,such an exchange is potentially draining, aswell as being fraught with the dangers atten-dant upon the release of emotions, oftenresentments and angers, which will usually dofar more to illuminate the realities of a givenissue than could any other data source.8 Wemay be stressing the more extreme end of suchfeminist research here, and we acknowledge(and hope) that research encounters will not allbe of this intensity, but we do wish to under-score just how much the researcher as a wholeperson – as an embodied individual with his orher own worries and frailties9 – enters into


28

3151-Ch-01.qxd 3/2/04 2:36 PM Page 28

the feminist research frame.We have thereforetravelled a very long way from the mostlyfact-finding questions asked of, say, local farmersby a Sauer-esque geographer strolling oneevening through a pleasant valley.

Feminist geographers do not only deploysuch intensive methods, of course, asMcDowell (1988) makes clear and as Hodgeet al. (1995) also insist when assessing the pos-sible use of quantitative techniques by femi-nists conducting geographical research.Yet wewill stick with this picture of intense inter-subjective research encounters – ones demand-ing an intimate meeting of two or moresubjectivities: that of the researcher and thoseof the researched (see also Cook and Crang,1995; see Chapter 10) – since such a pictureis helpful in clarifying an additional set ofclaims. And what will also be useful in thisrespect is briefly to acknowledge the influ-ence of postcolonial geography.10 If feministgeography confronts the axis of gender,problematizing its constitution as well as itseffects, postcolonial geography confronts theaxis of ‘race’, problematizing the inequalitiesbetween white people and people of colourwhich feature today in so many differentsituations under study by geographers (fromthe relations between ‘developed’ and ‘lessdeveloped’ countries to the circumstances ofracial minorities in predominantly whiteWestern cities). Given their acute sensitivityto axes of social difference, it is feminist andpostcolonial geographers who have donemost to reflect upon the problematic powerrelations which can arise in the researchencounter, most starkly when men areresearching their ‘other’ (women) or whenwhite people are researching their ‘other’(people of colour), but also in many otherways when differences of class, education, sex-uality, age, (dis)ability and so on potentiallydrive a wedge between the world of theresearcher and that of the researched. Thereare many thorny considerations here, but we

would suggest that the consensus emergingfrom recent texts such as Jackson (1993), Nastet al. (1994) and Cook and Crang (1995) isone insisting upon a ‘reflexive notion ofknowledge’ (McDowell, 1992a: 399), the cruxof which necessitates researchers reflectingcritically upon their own ‘position(ality)’ –their own backgrounds, attributes and values,as bound up with their own personal geogra-phies (the sites, localities and networks oftheir own biographies) – in relation to the‘position(alitie)s’ of those peoples and placesunder study. Such a stance on the doing ofqualitative human geography underlinesmuch of the recent Limb and Dwyer (2001)collection, where four chapters explicitlydebate matters of ‘positionality’.11 We try tovisualize this emerging model of intersectingpositions in Figure 1.5, the implication beingthat the researcher should aim to clarify his orher own position in a wider societal hierarchy ofpower, status and influence, thereby ascertainingthe different sorts of relationships – completewith the many differing roles, responsibilitiesand possible limitations to what can and shouldbe ‘exposed’ about the researched – which maysurface in a given research project.

From such a model it becomes apparent whyit is impossible for a geographer like McDowellto leave behind her personal world in the samefashion as can a Sauer: this is not only becauseshe has personal duties which she cannot forsakeas easily as can most male academics,but it is alsobecause she firmly believes that it is wrong inresearch terms to do so, since who she is (all thebaggage of her own position) is so very perti-nent to what she can achieve in her research. Itshapes her gaining of access to particularresearch situations rather than others; it shapesher ability (and willingness) to build ‘researchalliances’of empathy, trust and dialogue betweenher and the people whom she researches (seealso Pile, 1991); it shapes what findings she canobtain, the ways in which she will interpretthese findings,and her sense of what is and is not


29

3151-Ch-01.qxd 3/2/04 2:36 PM Page 29

appropriate to reveal in final write-ups of pro-jects undertaken.More particularly, it means thateverything which she does in this regard cannothelp but be influenced by her feminist experi-ences, values and politics, but she is reflexivelyaware of these feminist influences on herresearch and is self-critical about what theyenable to be seen and what they might alsoocclude. She is thereby arguably more objectiveabout the determinants of her research practicesthan are the likes of a Sauer or a spatial scien-tist.12 While recognizing drawbacks with a visu-alization such as that provided in Figure 1.5, wedo regard it as one which usefully pulls together

many of the themes which first surfaced in bothhumanistic and radical geography, but whichhave now been recast most effectively in thelight of both feminist and postcolonial geogra-phy.This visualization is also one which readersmight find useful to revisit at various points inthe chapters which follow.

We should acknowledge that some humangeographers may be unhappy about our nar-rative above, particularly given that we attachpriority to ‘being reflexive’ and ‘listening tovoices’ as the key recent developments in thepractising of human geography, and in so


30

*Conventional, if contestable, senses of the relative positions of the individuals concerned in a status hierarchy

• personal• cultural• political• professional, etc. ...

• gender• race• class• age• education, etc. ...

• suburban home• university• libraries• theatres• foreign holidays

‘Positions’

Values

‘Old boys network’,power, status,money, respect (?)...

A judgeWhat are hisvalues, ‘positions’ and‘geographies’?...

Soup kitchen andshelters and parks;no power, no status,no money, norespect?...

A ‘bag-lady’What are hervalues, ‘positions’ and‘geographies’ ?...

‘Geographies’ (indicative)

RESEARCHER

RESEARCHED

‘Inferior’*

‘Superior’*

Figure 1.5 Our visualization of the encounter between the differing‘position(alitie)s’ of researcher and researched (the ‘research subjects’)

3151-Ch-01.qxd 3/2/04 2:36 PM Page 30

doing put less store by more technicalinnovations such as GIS or computationalapproaches. None the less, and as should beapparent from what we said above, we do feelthat the latter innovations – while undoubt-edly of great utility in certain projects – areless significant as contributions to a genuinelyhuman geography than is the emergence of aself-critical reflexivity which begins ‘toquestion … what we know,how we know it andwhat difference this makes both to the type ofresearch that we do and who participates in itwith us as either colleagues or researchsubjects’ (McDowell, 1992a: 399–400). Thisbeing said, we appreciate that there are stillcriticisms to be levelled at a reflexive humangeography which claims to be good at listen-ing to the voices of others, and which therebysets itself up as (striving to be) both ethicallysound and politically empowering in relationto (less privileged) peoples and places understudy. In particular, Clive Barnett (1997) hassuggested that there may be problems withthe notion of ‘giving voice’ to others, in thatthere are many others in the world for whomsilence may actually be a preferred, even moreempowering, strategy. Similarly, Gillian Rose(1997a) has suggested that there are problemswith the impression of ‘transparent reflexivity’which is conveyed by the debate about posi-tion(ality) – the assumption that researcherscan somehow lay bare the many dimensionswhich comprise their position(ality) –because, as a psychoanalytic perspectiveindicates, many of the impulses, desires andpassions which feed through into our academicwork are ultimately lodged in realms of theunconscious inaccessible to conscious reflec-tion. Additionally, Rose criticizes notions ofempowerment which operate with a ‘map’ ofpower such as that implied in Figure 1.5,given that it hints at the possibility ofresearchers being able to redistribute stores ofpower from position to position (from thepowerful to the powerless, from themselves to

the people under study).As she rightly pointsout, the notion of power here is perhaps toosimplistic, in that power arguably operatesmore relationally than both the map andclaims about redistribution imply, as indeedhas been claimed in various recent texts onthe messy geographies of power (e.g. Hannah,1997; Sharp et al., 2000). The arguments byBarnett and Rose are very much set withinthe horizons of thinking previously rehearsedin this subsection, however, and they comprisea gloss (albeit an important gloss) on recentdebates about practising rather than the criticaldemolitions with which we concluded thetwo previous subsections of the chapter.

Conclusion

In this chapter we have introduced many ofthe themes relevant to the practising ofhuman geography, initially by contrasting twoextreme examples of how different humangeographers (‘Carl’ and ‘Linda’) go about theresearch process, and then by providing amore sustained review of changing ways inwhich human geography has been practisedover the years.While wishing to suggest thatthere are still things of value to take from thefirst two approaches assessed here, in that theirrespective attributes of immersed observationand systematic rigour do contain much ofmerit, we are more firmly persuaded by theclaims integral to the ‘being reflexive’ and ‘lis-tening to voices’ orientation. Indeed, thislatter orientation is now highlighting all mannerof complications with the practising ofhuman geography, countless issues whichwere either ignored in the past or were notcalled into play because different (arguablysimpler) research questions were being asked,and in the course of this chapter we havesought gradually to draw out these complica-tions for closer inspection. They are all oneswhich feature at various points in the book.


31

3151-Ch-01.qxd 3/2/04 2:36 PM Page 31

To conclude this chapter, though, we willprovide a summary listing of the major themesgrowing out of the above narrative. In terms ofthe construction of data, we have shown how pastgenerations of human geographers have tendedto regard the construction of data as a fairlyunproblematic matter, something that ‘simplyhappens’ in the field or occurs as packets of sta-tistics are sent to you in the post. Instead, it isnow argued that much more attention reallydoes need to be paid to precisely how thesedata are come by. Although this has not reallybeen a theme above, consideration must begiven to the composition of preconstructed data, asderived from sources other than theresearcher’s own primary research (seeChapters 2–4). Rather more has been said hereabout self-constructed data, those that have beenpieced together through the researcher’s ownendeavours (see Chapters 5 and 6), and henceabout the precise methods which need to bedeployed. Rather than simply ‘being there’,having an intuitive ‘eye for country’, conduct-ing list-like surveys or seeking out suitablelarge-scale numerical data sets, it has becomevital to ponder more carefully than hithertothe researcher’s methods. In particular, thenecessity for formalizing and extending qualita-tive methods has become increasingly evidentwith the heightened concern for what peopleunder study think, feel and do in their everydaylives. Questionnaires and, more especially, inter-views and participant observation have thusbecome popular, with humanistic and feministgeographers being at the forefront of this sus-tained qualitative turn. Moreover, and particu-larly as these geographers have made plain, theresearcher’s own underlying values and ethicalviews cannot be discounted as an influence onhow data are constructed,notably in the contextof the uneven power relations running between(the positions of ) the researcher and the resear-ched. The many implications of this researchrelationship can never again be regarded asunimportant.

In terms of the interpretation of data, we haveshown how past generations of human geo-graphers have tended to regard the interpretationof data as equally unproblematic, somethingthat involves gifted intuition or batteries ofquantitative analysis. Instead, it is argued thata broader span of attention is now required tothe overall interpretation of data, demandingmuch more than just the learning and refine-ment of new statistical-mathematical procedures.No longer is it assumed self-evident howresearchers move from data to conclusions,with the whole terrain of interpretationbecoming something requiring consideration,and different possibilities for interpretationneeding to be explicitly weighed up byresearchers (see Chapters 7–11). Moreover,and particularly as the humanistic, radical andfeminist geographers have made plain, theresearcher’s own underlying values and politi-cal commitments cannot be discounted as aninfluence on how data are interpreted, and theclear message from such geographers is thatwe should be fully aware of – and prepared toreflect explicitly on – how the whole cast ofour research is shaped by such values andcommitments (and see also our Chapter 12).

Notes

1 Some of Sauer’s key writings are collected inLeighley (1963).

2 Key writings by McDowell include (1983),(1989), (1999); and, for the women in theCity work, see McDowell and Court (1994;McDowell, 1997).

3 See, for example, French (1940), Coleman(1954), Dilke (1965), Wheeler and Harding(1967), Archer and Dalton (1968), Yates andRobertson (1968) and Coleman and Lukehurst(1974).

4 A related issue is that the unthinking occu-larcentricism of the discipline is insensitive togeographers who are visually impaired, andit is rare to come across a proposal such asKingman’s (1969) regarding ‘field study forthe non-sighted’. More generally, those whocelebrate the being there of fieldwork tend


32

3151-Ch-01.qxd 3/2/04 2:36 PM Page 32

to assume the able-bodied status of allgeographers, another aspect of neglectingthe inaccessibility of many fieldwork sites tomany individuals who are differently abled:see Hall et al. (2002).

5 Prime statements at the time about this vari-ety of geography included Bunge (1962),Burton (1963), Harvey (1969), Abler et al.(!971) and Amedeo and Golledge (1975).

6 In passing, it is worth repeating Barnes’s(2001b: 552) point about the extent to whichthe early spatial scientists were fixated ontheir new ‘machines’, IBM mainframe com-puters and the like, and on the numericalanalyses of large quantitative data sets nowmade possible by such technology.

7 Key texts included Peet (1978), Smith, D.M.(1977; 1979), Harvey (1973; 1982); for sum-mary details, see Cloke et al. (1991: ch. 2).

8 See also recent claims about taking seriouslythe emotional registers of the researcher:Widdowfield (2000), Anderson and Smith(2001).

9 See, for example, Pile (1991), Nast (1998), Parr(1998a), Dewsbury and Naylor (2002: esp. 256–7).

10 See Crush (1994), Radcliffe (1994), Jacobs(1996) and Nash (2003).

11 Butler; Ley and Mountz; Mohammad;Skelton.

12 See also the claims in Philip (1998) drawingin part on Wright’s (1947) notion of ‘objec-tive subjectivity’; see, too, Box 1.7.


33

3151-Ch-01.qxd 3/2/04 2:36 PM Page 33

3151-Ch-01.qxd 3/2/04 2:36 PM Page 34

Geographical data

Human geographers are almost always using data (see Box 1.3) in theirresearch. We may sometimes conduct highly abstract theoreticalmeditations on, for instance, the nature of social space, the character ofhuman places or the deeper philosophy behind how we struggle to knowwhat is happening in the world of people. In these cases the activeengagement with data may be quite limited, but it is unlikely to be entirelyabsent. Alternatively, we may occasionally conduct more synthetic workthat seeks to pull together, perhaps to find new ways of understanding,the findings from previously completed research projects by geographersand other academics. In these cases there is an active engagement withdata, even if these data have been collected and processed beforehand byscholars other than the person doing the synthesizing, but it remainsimperative that the synthesizer is aware of how the relevant data havebeen come by and worked upon in these prior research projects.

In probably the majority of our studies as human geographers, however,we are deeply preoccupied with data from first to last. Whether theundergraduate student undertaking a dissertation or the establishedprofessor conducting funded research, we are largely dependent upondata for the successful completion of our research projects: we need toacquire data, to appreciate both their possibilities and their limits (whatthey can and cannot tell us) and to be able to deal with data, often in avariety of ways, so that we can arrive at findings, conclusions and furtherspeculations. And yet, despite the recent raft of textbooks on matters ofmethodology that have appeared in the geographical literature,referenced in our Preface, there is still little systematic reflection on thedata we use, nor on the means by which we endeavour to get data toreveal things to us. The two parts of this book that follow, therefore, striveto tackle both these issues in turn, beginning with five chapters exploringwhat is involved in the different sorts of data which we routinely bring tothe table of research. Our key claim spread across these five chapters is adeceptively simple one: that all these data, whatever their source, areconstructed in one way or another. They do not magically or spontaneouslyarise in the world just waiting for the geographer to come along; rather,they are ‘made’ by somebody for given reasons and in specific ways.

PART I CONSTRUCTING GEOGRAPHICAL DATA

3151-part-01.qxd 3/2/04 2:51 PM Page 35

Preconstructed geographical data

Most human geographers have to put a lot of effort themselves into‘getting hold’ of data with which to work. For a few of us this is aseemingly quite straightforward process. Perhaps we photocopy a fewpages from the British Census, or download these from relevant websites,or perhaps we buy some data from another organization, say profiles heldby a company on its customers and their home neighbourhoods, whicharrive as either hard copy or on a computer disk ready to be processed. Forsome of us there is more effort involved, since it entails visiting a national orlocal archive to hunt out materials, perhaps a Parliamentary SelectCommittee report, perhaps an inventory of people resident in anineteenth-century institution, the minutes of city council meetings,someone’s diary or letters. For others of us it might require trawlingthrough back issues of magazines or copies of old television programmes,which might necessitate buying these from some source or travelling to aspecialist archive. Sometimes we will be lucky, easily locating and beingable to purchase, borrow or access the relevant data. On other occasionswe will not be so lucky and may experience considerable difficulty intracing data appropriate to the research questions being asked.Alternatively, we may find ourselves unable to access the data once they arelocated because they are confidential to an organization or individual, orbecause they are simply too expensive to buy, or because they are housedin an archive that is closed, private or too far away. It is not unknown forresearch projects to hit a dead-end because it is simply impossible to trackdown relevant data or to get a sight of relevant data sources if they do exist.

The crucial point to notice here is that in all the instances listed abovethe data may have the appearance of ‘simply sitting there’ waiting for usto locate and to access them, but such an appearance must not disguise thefact that all the data sources involved have, indeed, been constructed.These data sources have all been ‘made’ by somebody for reasons whichneed to be ascertained and in particular formats and using particularmethods which need to be uncovered. In our vocabulary, these data havebeen preconstructed, which means they have not been constructed or‘made’ by the researcher but by somebody else at a prior moment in time(whether long ago or last week). These data have not been constructed bythe researcher and have been put together for entirely other reasons thanthe purposes of a researcher carrying out a research project. A census is puttogether because a national government wants to know things about itssubject population, their age and sex composition, their occupations andthe like, chiefly because such information helps it to formulatesocioeconomic policies, and it is only as a by-product that researchers canstart to ask their own questions of the census data so gathered andtabulated. Minutes of committee meetings are taken because it isimportant that members of the committee have a record of proceedings,

PART I Constructing geographical data

36

3151-part-01.qxd 3/2/04 2:51 PM Page 36

of decisions taken and arguments won or lost, and perhaps, too, so thatthe committee can be accountable to non-members. A diary is kept by anindividual or letters are written for a multitude of very personal reasons,while a film or a documentary is made for reasons of entertainment andinstruction.

If human geographers are going to utilize these data sources, to extractthings from them for interpretation (see Part II), it is vital that we are asfamiliar as possible with the subtleties of these sources: that we do knowthe why, the how and also the when and the where of their construction,since it is only on the basis of such knowledge that we can sensibly assesswhat such sources can and cannot tell us. It is only by having a clearunderstanding of these data that we can decide the extent to which we‘trust’ them to be revealing an accurate picture of the ‘reality’ beyond (e.g.the population structure, a committee’s deliberations, a diary-writer’sexperiences), or can appreciate the basis for effects, messages and the like –we might say ‘biases’ – more or less wittingly installed into particular kindsof data (e.g. the political content of a report or the emotional charge of adocumentary, both of which might themselves become the focus of aresearcher’s attention). In the first three chapters that follow, we strive tolay bare many of the issues bound up in the making of three different ifoverlapping varieties of preconstructed data, suggesting along the waynumerous things the human geographer must keep in mind when drawingupon such data. First, in Chapter 2, we look at ‘official sources’ (e.g. agovernment report); secondly, in Chapter 3, we look at ‘unofficial sources’(e.g. a newspaper article); and, thirdly, in Chapter 4, we look at ‘imaginativesources’ (e.g. a novel). In each case, we focus on the data involved, thedocuments that are part of what some writers now term a wider ‘thingworld’ of documents, devices and artefacts of all kinds, asking about howthese data are constructed and made available for us as humangeographers to labour upon in our own research.

Self-constructed geographical data

For many human geographers, however, the data required are arguably rathermore difficult to come by than it is for those who utilize preconstructedsources. For many of us, then, there are no pre-existing data sourcesavailable with which we can answer our research questions. We may wantto know things about the characteristics, activities, worldviews and the likeof people living in a given locality, or we may want to gain detailedinformation about the thoughts and actions of a very particular group ofpeople (residents of a gentrifying neighbourhood, for instance, or thesenior executives of a multinational company). There may be some extantdata sources which we could consult in these respects – censuses, opinionpoll surveys, petition leaflets, company reports and so on – but it is unlikely

CONSTRUCTING GEOGRAPHICAL DATA

37

3151-part-01.qxd 3/2/04 2:51 PM Page 37

that such sources will contain precisely the information that is required, forthe very reason that these sources have not been compiled with the specificresearch aims of the human geographer in mind. Sometimes we have nochoice but to make do with these sources, notably if we are historicalgeographers dealing with eras beyond living memory whose only access tothe people, places and events of the past is through documentary remainsin dusty archives.

Yet for many of us studying more contemporary subject-matters, we dohave the option of trying to generate appropriate data for ourselves. Wehave the opportunity to ‘go out into the field’ to conduct what is normallycalled fieldwork – although note our wish, as stated in Chapter 1 (see Box 1.2),to keep an expanded rather than a narrowed conception of both the fieldand fieldwork – in the course of which we seek to generate data throughour own active engagement with, or questioning of, all manner of relevantpeople going about their business in the everyday world (whether they bea planner, a salesperson, a musician, a refugee or whatever). There arenumerous methods deployed by human geographers in this attempt togenerate data through their own endeavours, and some of these areperhaps all too familiar, such as the questionnaire survey, while others ofthem are probably not so familiar (certainly to undergraduate students),such as using focus groups or conducting participant observation. All thesemethods are relatively well known to social scientists of differentpersuasions, including historians and anthropologists, and there areestablished theories about their strengths, guidelines about thepracticalities of their conduct and debates about the ethics wrapped up intheir operation. Human geographers must be aware of these theories,guidelines, ideas and the like, and should also contribute to reflections onthe methods involved (see Box 1.4) on the basis of their own specialinsights about, for instance, what it means to be trying to generate dataabout ordinary people occupying specific spaces or possessing feelingsabout particular places.

In this regard, therefore, we as human geographers strive to make ourown data, to produce our own sources of data: in a word, we are nowtalking about data that are not preconstructed but self-constructed by usthrough our active fieldwork. If we conduct a questionnaire survey, it is wewho are deciding upon the questions to ask, the sample of people to beasked, the procedures for administering the questionnaire, the tactics fordealing with respondents and so on. Any data arrived at in the process areindelibly shaped by the decisions we have taken, and the reliability of thesedata, however exactly we might conceive of what is or is not ‘reliable’, isdeeply dependent upon the manner in which we have gone about thequestionnaire exercise from the first moment of question formulation tothe last moment of questionnaire delivery in the street, on the doorstep orover the telephone. If we conduct interviews, it is similarly we who areresponsible for every aspect of how schedules of interviews are drawn up,


38

3151-part-01.qxd 3/2/04 2:51 PM Page 38

who is interviewed and how interviews run in the flesh, and the dataproduced are unavoidably a measure of how successful or otherwise wehave been in topic selection, interviewee choice and interviewmanagement. If we deploy other methods, such as different versions ofparticipant observation, the data that are embodied by entries in our fielddiaries cannot but reflect our own abilities and inabilities in playing a rolein field settings, our confidence or inhibitions when it comes to directinginterpersonal interactions and the quality or otherwise of our observationalskills, among other things.

It is essential that we do think very carefully about these methods, thatwe stand back from the logistics of undertaking them to consider thebroader picture of exactly how they work, how what we do affects theirexecution and how relations with people under study are fostered,maintained or perhaps threatened by their utilization, to name but a fewsalient considerations. In so doing, both the possibilities and the limitationsof data constructed by these methods should become apparent, and whatis appropriate and what is inappropriate to demand of such data shouldalso become much clearer to us. In Chapters 5 and 6 we strive to lay baremany of the issues bound up in the making of several different ifoverlapping varieties of self-constructed data, suggesting along the waynumerous things that the human geographer must keep in mind whengenerating such data. First, in Chapter 5, we look at different methods forwhat we refer to quite simply as ‘talking to people’, examining questionnaires,interviews and focus groups; and, secondly, in Chapter 6, we look at doingethnographies’, tackling those methods that return in some respects butcertainly not others to the ‘being there’ tradition of fieldwork examined inChapter 1. In both cases, we focus on the practices involved as weencounter a ‘world of people’, stressing that such practices are indeliblysocial, marked by the complexities of negotiating social relations bothdirectly with our research subjects and more broadly in the social settingsunder study. Moreover, because of these social dimensions, it is inevitablythe case that things may not always run smoothly and that our best effortsto structure research encounters will sometimes go awry – people will notturn up or be around, will answer our questions ‘oddly’, will be hostile tous or will simply introduce a host of wholly unexpected issues – and henceleave us realizing that the self-constructing of data has to be flexible, opento disasters and prepared to change in the process. Self-constructed dataare hence unlikely to be neatly self-controlled data, precisely because themethods involved are social practices rather than akin to laboratorytechniques.

CONSTRUCTING GEOGRAPHICAL DATA

39

3151-part-01.qxd 3/2/04 2:51 PM Page 39

3151-part-01.qxd 3/2/04 2:51 PM Page 40

The centralization of knowledge requiresfacts – and the legitimization of some facts,and the methods used to collect them,against other facts – to justify features andforms of policy. (Corrigan and Sayer, 1985:124, emphasis in original)

Introduction

One of the central arguments of this book isthat in all human geographical research it isvery important for the researcher to considerthe processes through which data sources areconstructed. In some cases the researcher isdirectly involved in the processes of data con-struction. In others he or she is using data thathave been produced by others. In this and thenext two chapters, we consider sources ofgeographical data that are not produced bythe researcher him or herself but constructedby others before the research takes place. Suchsources may have been produced privately byindividuals, social groups, voluntary organiza-tions or firms (these types of data are consid-ered in Chapters 3 and 4). Alternatively theymay have been produced by governmentagencies or public authorities; that is, by thestate. Such ‘official data’ are the subject of thischapter. We are treating ‘official information’separately from other types because it is pro-duced in distinctive ways for particular pur-poses and these need to be understood whenusing the information in research.

Official information is of enormousimportance for research in human geography.The population geographer’s migration flows,the economic geographer’s local labour mar-kets, the social geographer’s crime patternsand the political geographer’s election analy-ses (among many others) all usually depend, atleast in part, on official information.Historical geographers often rely heavily,though by no means exclusively, on officialrecords of past events (for examples, seeBaker et al., 1970). Given the importance ofhistorical sources to their work, they haveoften also been particularly sensitive to theimpact of the circumstances under whichthose sources were produced, an issue that isof particular significance to the arguments ofthis chapter. Census data are of great value toboth historical and contemporary geogra-phers and it is interesting that geographers arenow at the forefront of technical and intellec-tual developments in census design and datacollection (Martin, 2000). Official voting dataprovide the raw material for numerous stud-ies of electoral geography.1 Official docu-ments are also textual, as well as statistical, andgeographers make good use of these too. Forexample, Moon and Brown (2000) examinethe ‘spatialization’ of policy discourse throughan analysis of health policy documents inBritain.

In addition to the sheer number of studieswhich use it, official information is important

Official Sources2

3151-Ch-02.qxd 3/2/04 2:36 PM Page 41

because of its particular authority. When weread official statistics or government reports,we often assume they are reliable and accuratesimply because they are official. Furthermore,some of this authority rubs off on academicresearch based on such data, making it seemsound, well founded and valid.There are cer-tainly some good grounds for these assump-tions. First, governments (especially inindustrialized countries) have very largeresources at their disposal. These resources(staff, finance and so on) are usually much largerthan those to which academic researchers haveaccess. Governments can thus draw largersamples, undertake more thorough analyses,use longer timescales and cover wider areasthan other researchers and organizations.Secondly, governments frequently give them-selves legal powers to support their researchactivities. For example, in most countries,members of the public are obliged by law tocomplete a census return on a regular basis,often every ten years. In many cases informa-tion may be a useful by-product of legislationprimarily enacted for other purposes. Forinstance, the collection of taxes may also gen-erate information about employment andincome, while vehicle licensing leads to infor-mation about car-ownership patterns.Thirdly,in the modern world, governments are inter-ested and involved in much of the everydaylife of the population.Through the breadth ofits activities the government knows some-thing about what we are doing: whether weare shopping, working, earning, marrying,having children, travelling, suffering illness,claiming welfare benefit or finally dying. Ofcourse we are not always identifiable as indivi-duals within this information, but our activi-ties contribute to it, none the less.

Yet despite the resources, legal powers andbreadth of involvement of governments thereare also good grounds for supposing that offi-cial information is not as reliable as is fre-quently assumed and that its special authority

is undeserved. Governments are not neutralreferees overlooking society but players activelyinvolved in the game. Like other organizationsin society, they have particular objectives inobtaining, processing and presenting informa-tion and particular interests at stake in its con-tent. Official information should therefore betreated with the same healthy scepticism whichmost good researchers bring to the study ofunofficial sources.

Throughout this book we will be stressingthat no source of research material may betaken at face value, even, or perhaps especially,where it carries the cachet associated with‘official’ information. In all cases, data used byresearchers in human geography are con-structed in specific cultural, political andeconomic contexts which influence theircharacter and content. If we are to constructvalid interpretations of data, therefore, weneed to understand, and to take account of,those contexts.Thus, in this chapter we con-sider how, why and in which contexts gov-ernments construct the data they do and theconsequent implications for their use byhuman geographers engaged in research.

Types of officialinformation

Governments produce information in a vari-ety of forms. A simple typology might showthat information can be textual, graphical andcartographical, aural or numerical in form. Agovernment report is a textual source, officialphotographs are graphical, government radiobroadcasts are aural and government statisticsare numerical.These differences are importantand they affect how the data can be inter-preted. In this chapter, however, we are focus-ing on data construction rather than datainterpretation. As far as construction is con-cerned, three broad categories of data can bedistinguished.


42

3151-Ch-02.qxd 3/2/04 2:36 PM Page 42

First, governments are large bureaucraciesand produce information as part of thebureaucratic process.A minister may commis-sion a report before reaching a decision.Parliament produces bills and Acts in theprocess of legislating. Local councils publishthe agendas and minutes of their meetings.Official inquiries absorb written and oral evi-dence and produce reports. The courts pro-duce written judgments. In all these cases,documents and information are the by-productsof the process of governing and of the opera-tion of large bureaucratic organizations.

Secondly, governments monitor the soci-eties they govern. As we shall see shortly,monitoring and surveillance are very impor-tant in shaping the nature of the modernstate. They also produce huge amounts ofinformation. Taxation records, populationrecords, censuses, health records, financial sta-tistics and a whole host of other sources ofinformation are generated simply because thestate and the government need or want tokeep a check on what is going on.

Thirdly, governments communicate. Frompublic health announcements to political pro-paganda and from communiqués at interna-tional summits to the contents of the schoolcurriculum, governments address their citi-zens (and the wider world) in a variety ofways. Such communication may involve socialengineering of a mild kind, such as trying topersuade parents to immunize their children,or it may involve giving direct orders pro-hibiting certain activities. It may involve therelatively straightforward provision of infor-mation (about welfare benefits or health care,for example) or it may be sophisticated polit-ical publicity intended to change people’sminds.

These three categories may overlap. Forexample, official statistics are often producedthrough the bureaucratic process of govern-ing, and may then be used in public informa-tion campaigns.

Information andstate formation

Record-keeping and theorigins of states

All organized human societies, past andpresent (including tribal societies based onkinship relations), have some form of govern-ment or collective rule. Following the rise ofsettled agriculture, early urban centres anddivisions of labour by about 5000 years ago,government increasingly became institution-alized in states of various forms (Mann, 1986).According to the sociologist Anthony Giddens,the origins of states are closely related to theinvention of writing. Giddens (1985: 41–9)challenges the idea that writing is straightfor-wardly a representation of speech. He arguesinstead that writing initially took the form ofrecord-keeping:

Many linguists have regarded writing as nomore than an extension of speech, the tran-scription of utterances to transcriptions onstone, paper or other material substancesthat can be marked. But neither the first ori-gins of writing in ancient civilisations nor aphilosophical characterisation of languagebears out such a view. Writing did not origi-nate as an isomorphic representation ofspeech, but as a mode of administrativenotation, used to keep records or tallies.(1985: 41)

There are examples of non-modern stateswhich apparently used no form of writing. Inthe Andean Mountains of what is now SouthAmerica, for example:

About A.D. 1400–30, one ‘tribal’ groupingand chiefdom, the Inca, conquered the rest.By 1475, the Inca had used massive corvéelabor gangs to build cities, roads, and large-scale irrigation projects. They had created acentralized theocratic state with their ownchief as god. They had taken land into stateownership and had put economic, political,and military administration into the handsof the Inca nobility. They had either devisedor extended the quipu system whereby

OFFICIAL SOURCES

43

3151-Ch-02.qxd 3/2/04 2:36 PM Page 43

bundles of knotted strings could conveymessages around the empire. This was notexactly ‘literacy’ … Yet it was as advanced aform of administrative communication as anyfound in early empires. (Mann, 1986: 122)

What seems crucial, therefore, is not writingper se but the resources and ability to keeprecords. The significance of this developmentshould not be underestimated since, withoutthe capacity for record-keeping, an organiza-tion would lack a key means of what Giddensterms ‘administrative power’ (1981: 94–5; 1985:19). Indeed it is doubtful whether it could bean organization at all.The exercise of admin-istrative power based on the collection andorganization of information is thus a definingfeature of all states. It is both an outcome ofstate activity and constitutive of state power:states generate information through theirfunctions but also require information inorder to undertake those functions in the firstplace. Generating information, therefore, isnot an optional extra for state organizationsbut seems to be central to the very possibilityof states existing at all.

Official informationand the development ofthe modern state

While record-keeping is a necessary feature ofnon-modern states like the Inca Empire, it isin modern states that it reaches its most devel-oped form. Drawing on the work of the clas-sical sociologist Max Weber, ChristopherDandeker argues that ‘bureaucratic decisionsand calculations depend on knowledge of thefiles, that is on a mastery of the informationstored centrally in the organization’ (1990: 9,emphasis in original). For Weber, the institu-tions of the state are exactly such archetypalbureaucratic organizations. The constructionand processing of very large quantities ofinformation are defining features of govern-ment in the modern world. The amount of

information routinely collected by stateorganizations is vast. The UK government’sannual summary of official statistics (theAnnual Abstract) contains over 300 separatetables of data dealing with topics as diverse aspopulation, iron and steel production, govern-ment debt and academic research. The USgovernment’s Statistical Abstract of the UnitedStates runs to over 1000 pages. By contrast, inmedieval England the famous Domesday sur-vey of 1085 contained just 11 questions relat-ing to the ownership of property (Darby,1977: 4–5). Furthermore, medieval govern-ments undertook such data collection veryinfrequently: no more often than once everyseveral decades. By contrast, modern govern-ments are engaged in an almost continuousprocess of monitoring the various indicatorswhich are the basis of ‘official statistics’.

Furthermore, statistical information formsonly a part (albeit an important one) of thetotal information produced by the state.Policy decisions depend on ‘knowledge of thefiles’ but the contents of the files can takemany forms. Records concerning individuals,for example, may include reports, photographs,transcripts and forms. The governing processitself involves the production of large quanti-ties of textual and graphical, as well as numer-ical and statistical, information. For example,governments often establish commissions ofinquiry relating to particular areas of policy.The reports of such bodies are often invalu-able sources of research data. Other kinds ofrecords include financial accounts and regis-ters of various kinds. An example of a geo-graphical study that makes use of this kind ofsource is Driver’s account of the geography ofthe workhouse system in Victorian Englandand Wales, which draws on the unpub-lished Registers of Authorised WorkhouseExpenditure (1989; 1993: 73–94).

The ‘outputs’ of the policy-making processcan also be a rich source of data for research.Modern states produce copious quantities of


44

3151-Ch-02.qxd 3/2/04 2:36 PM Page 44

planning and policy documents which bothpresent government policy and enable itsimplementation. They also contain very par-ticular representation of the social and eco-nomic world (the objects of governance, as itwere). Governments have also significantlyexpanded the quantity and range of informa-tion they publish under the umbrella ofgovernment publicity or public information.In this case states are addressing their populationsdirectly, and examples include the educationalcurriculums, health advice, edicts compellingpeople to do certain things and to refrainfrom doing others and so on.Whichever theform in which state information is produced,one thing is clear: the rate of its productionhas increased dramatically over time and con-tinues to do so.

How should we account for the extensivegrowth in government activity in this area?One important reason concerns technology.Today governments have the technologicalcapacity to conduct data-gathering exerciseson a scale which their medieval predecessorscould scarcely have imagined. Printing, trans-port, telecommunications and storage facili-ties are all involved.Yet while this is clearly anecessary condition of the growth, it is not, initself, a sufficient one. It does not tell us whygovernments might want to gather informa-tion on this scale.To answer this question weneed to consider the ways in which statesthemselves change over time and how thosechanges are related to changes in civil society.

All official information is gathered for spe-cific government purposes. However, thosepurposes vary through time and across space.The worries and concerns of the Normanmonarchy in medieval England were notthe same as those of the drafters of theConstitution of the USA or of the present-day European Commission. A good exampleof this historical development is provided bythe English (subsequently the British) state.According to Philip Corrigan and Derek

Sayer in their book The Great Arch (1985), theEnglish state underwent a succession of phasesof formation. Each phase involved what theyterm a process of ‘cultural revolution’ (Corriganand Sayer, 1985: 1–2): a transformation in theways in which the agents and agencies of thestate made sense of and represented civilsociety. Part and parcel of this process was agrowth and extension of the generation anduse by the English state of official informationof various sorts.

The purpose of the Domesday survey, forexample, has been the subject of academicdebate, but whatever its purpose it is clear thatofficial information in medieval Europe wascollected, when it was collected at all, for alimited number of reasons. By modern stan-dards, medieval states had a small number offunctions centred on the security and financeof the royal household, the defence of therealm and the regulation of feudal ownershiprights. Beyond these, there was relatively littlecontact between the state and its subjects.Most of what are regarded as state functionstoday in the fields of education, health care,support for economic activity and even thepreservation of law and order were carriedout, in so far as they took place at all, by otheragencies, particularly the Church and thefeudal manor.

The state did need information in somespheres of activity, notably taxation, but feudalties of rights and obligations and religiousconviction ensured that for the most partgovernment was carried on and social orderwas maintained without the state needing toknow in detail about most of the activities ofits population.

In the 1530s, during the so-called Tudorrevolution in government,Thomas Cromwell,an official at the Court of Henry VIII, wasresponsible for a significant increase in theinformation-gathering activities of the Englishstate (Corrigan and Sayer, 1985: 48). Much ofthis was aimed at countering threats to the

OFFICIAL SOURCES

45

3151-Ch-02.qxd 3/2/04 2:36 PM Page 45

King, but it included making compulsory, from1538, the keeping of parish registers ofbaptisms, marriages and burials. According toCorrigan and Sayer these registers:

provoked widespread resentment, andpeople feared their use for taxation purposes.Cromwell’s own justification, responding tosuch fears, is interesting. The registers werebeing instituted, he said, ‘for the avoidingof sundry strifes, processes and contentionsrising upon age, lineal descents, title ofinheritance, legitimation of bastardy, andfor knoledge of whether any person is oursubject or no’; ‘and also’ he added, ‘forsundry other causes’ (Elton, 1972: 259–60).Concern with property and concern withrule are equally witnessed here. Registersare not merely a technical device, theymaterialize new sorts of claims of a stateover its subjects. The Elizabethan Poor Law,for example, was to rely for its administra-tion on the kind of detailed knowledge ofindividuals the registers supplied. (1985: 50)

Here then was a definite extension of thefunctions of government accompanied andenabled by an extension in the information-gathering activities of the state.Tudor Englandalso saw the systematic mapping of thecountry with the county maps of ChristopherSaxton. According to Derek Gregory, ‘thesepaper landscapes represented the administra-tive apparatus of the state with an apparentlyindelible authority, revealing the steadyencroachment of surveillance into still morespheres of social life’ (1989c: 376). Such piece-meal encroachment continued during theseventeenth and eighteenth centuries (Brewer,1989), but it was not until the nineteenthcentury that the information-gathering activ-ities of the state took on anything like theirmodern form in terms of both quantity andquality.

This development both reflected and partlyinfluenced the changing relationship betweenstate institutions and the society they gov-erned. In this, the twin processes of urbaniza-tion and industrialization were central. InBritain, the agricultural ‘improvements’ of the

country during the seventeenth andeighteenth centuries had created a large classof landless people. These people, separatedfrom the traditional religious, cultural andeconomic ties of rural life, and with no meansof support other than the sale of their labourpower, made their way to urban areas.There,new, factory-based systems of productionwere hungry for the labour that they couldprovide and the industrial towns and citiesquickly expanded. In 1801 (the year of thefirst Census of Population in Britain) 33.8%of the population of England and Wales livedin urban centres. By 1901 the urban popula-tion had expanded eightfold and its propor-tion had increased to 78% (Carter, 1990: 403).This startling transformation in the geographyof the country caused and made visible socialproblems on a scale never previously imagi-ned. Working-class districts were regarded asovercrowded and the people who lived inthem uneducated and at risk of disease andcrime. In addition they were uprooted fromestablished ties of kinship and community,which had often served to mask, if not to pre-vent or solve, similar problems in rural areas.In the new industrial areas, however, therewas no disguising the suffering and povertywhich existed and which was exacerbated byexploitative conditions of employment in thefactory system.

It was not long before the mass of the urbanpoor became a focus for crusading social inves-tigators and thence for government policy.Seduced by its own rhetoric of progress, bour-geois society in Victorian Britain was hardlyabout to blame industrialization itself for theplight of the poor. Instead it preferred to locatethe causes of social distress in the moral failingsof individual sufferers themselves and the over-crowding and dirt of their immediate environ-ments.According to Felix Driver:

The association of moral and medical con-cerns was a central feature of nineteenth-century social science. The question of


46

3151-Ch-02.qxd 3/2/04 2:36 PM Page 46

density was a special focus of interest. Therewere many attempts to relate populationdensity to crime and other indices of‘demoralization’, at various ecologicalscales. The most common foci were theovercrowded, unsupervised and disorgani-zed spaces of the nineteenth-century city …What seems to have most concernedmiddle-class commentators was their ownlack of control over and within such areas;indeed, their obsession with hiddenrecesses, narrow turnings, dark alleys andshadowy corners was quite overwhelming.The literature on the rookeries of London,for example, was predicted [sic] on theassumption that they were located beyondthe public gaze, outside the ambit of officialsurveillance. (1988: 280–1, emphasis added)

A mixture of fear, loathing and social concernof the middle class and the government forthe mass of the urban poor prompted a desireto know precisely what were the socialconditions of the population, to enable bothreform and social control. The rapid growthin the collection and interpretation of socialstatistics in nineteenth-century Britain canthus be related directly to the urban-industrialtransformation. Similar processes took placein other countries which experienced similarindustrial development, albeit sometimes withdifferent emphases. In France, where thedevelopment was related particularly to con-cerns about public health,‘a professionalizationand regularization of population statistics,linked to a range of reformist projects,occurred during the first decades of thenineteenth century’ (Rabinow, 1989: 60). Inthe USA the focus of concern was frequentlyimmigration and the formation of ethnicallysegregated ‘ghettos’ (Ward, 1978).What unitesall these transitions is the use made of statisti-cal data to try to understand new, radicallydifferent and much larger social worlds:‘with-out the combination of statistical theory …and arrangements for the collection of statis-tical data, … the society that was emergingout of the industrial revolution was literallyunknowable’ (Williams, 1979: 171).

These desires to make knowable theunknown both to ‘improve’ and to control itled to an upsurge in the collection of statisti-cal information. In Britain, the governmentwas accompanied, and frequently assisted, inthis task by interested members of theVictorian middle class. The establishment of‘statistical societies’ in Manchester (1833) andLondon (1834) (and later in other provincialcities) laid the foundation of a long nationaltradition of empirical social research (Cullen,1975).The studies undertaken by these soci-eties and individuals were the forerunners ofnon-governmental research, which is the sub-ject of the next chapter. They culminated inthe 1890s with the publication of CharlesBooth’s huge survey of Life and Labour of thePeople in London.

For the government, statistics were com-piled with increasing frequency throughoutthe nineteenth century. The StatisticalDepartment of the Board of Trade wasfounded in 1832.The Registration Act 1836led to the establishment in 1837 of theGeneral Register Office for England andWales to record births, deaths and marriagesand to undertake the decennial Census ofPopulation. A General Register Office forScotland was set up in 1855.

The close links between the governmentand the growth of statistical information areexpressed in the very word ‘statistics’ itself.Originally the word was simply an adjectiveassociated with the noun ‘state’ (statist + ic).According to one nineteenth-century defini-tion, statistics is ‘that department of politicalscience which is concerned in collecting andarranging facts illustrative of the conditionsand resources of a State’ (Corrigan and Sayer,1985: 134).

The purpose of this information gathering,particularly in the first half of the nineteenthcentury, was to enable the development ofstate policy. Policies of moral and environ-mental improvement were introduced on the

OFFICIAL SOURCES

47

3151-Ch-02.qxd 3/2/04 2:36 PM Page 47

back of statistical inquiries. The workhousesystem, the development of public healthmeasures, medical activities and educationwere all central to the government ofVictorian Britain and all depended on infor-mation. Furthermore, all were intended totreat social ills by extending state influence,regulation and control into all parts of civilsociety.

Effective public policy was not the onlyconsequence of the growth in state informa-tion gathering. The development of statisticsalso affected the object of study (i.e. society)itself.This is a difficult notion to grasp at firstas we tend to assume that statistics are first andforemost a simple description of social ‘facts’.However, the production of official statistics isnot the neutral, technical and scientific exer-cise it appears to be. As we have suggested,official data do not provide complete or trans-parent pictures of social reality. Rather theyare influenced and conditioned by the inter-ests at stake in their production.The demandsof government policy or the moral concernsof the middle classes are reflected in thestatistics that are constructed.At the very leastthey require that certain topics of inquiry areselected as relevant over others.At worst theremay be active manipulation of figures to pro-vide justifications for particular governmentactivities.But we can go further than this. Notonly are statistics sometimes an inaccuratepicture of ‘reality’ but, by being collected atall, they also change the nature of that ‘reality’.Catherine Hakim (1980) has analysed thecensus reports accompanying the BritishCensus since it began in 1801, and her find-ings help to show what we mean.

As she points out, the census has been animportant source of information on occupa-tions.Yet the particular choice of occupationcategories included in the survey has played apart in making some occupations respectableand others illegitimate. Estimates suggest that20% of unmarried women in London in 1817

were involved in prostitution (Hakim, 1980:563). Yet the category of ‘prostitute’ is notlisted in the occupational classification of theBritish Census. This refusal to give formalrecognition to the economic activity of asubstantial proportion of the city’s populationhad the effect of making prostitution appearless socially acceptable, as well as calling intoquestion the accuracy of the data.

In the early years of the census the full sep-aration of paid employment from home life,and the resulting partial exclusion of womenfrom the labour market, had yet to occur.Theeconomic unit was considered to be the fam-ily as a whole, rather than an individual malebreadwinner. By the middle of the nineteenthcentury, however, the census had begun to askfor information on the occupation of individ-uals and this led to numerous women beingclassified as economically inactive, despitecontinuing to be engaged in domestic labourin the home. The change in the form inwhich these figures were collected partlyreflected, but also partly produced, these devel-opments in the labour market. By categoriz-ing women as ‘inactive’ the official censuscontributed to the exclusion of women fromthe labour market and to their confinement inthe home.

The power of such official ‘discourses’ and‘representations’ to affect as well as reflectsociety is recognized by Corrigan and Sayer:

Whilst we should reject states’ claims asdescription – they are, precisely, claims – weneed equally to recognize them as materialempowerments, as emphases which car-tographize and condition the relations theyhelp organize. Here, as elsewhere, to adaptan old sociological maxim, what is definedas real (which is not to say the definitionsare uncontested) is real in its consequences.(1985: 142)

Commenting on Hakim’s research, Corriganand Sayer point out that the British Census isabove all an inquiry into who is a subject ofthe Crown,and is thus itself centrally implicated


48

3151-Ch-02.qxd 3/2/04 2:36 PM Page 48

in ‘the formation of social identities’ (p. 132).Giddens makes the point forcefully:

it might well be accepted that, given cer-tain reservations about the manner of theircollection, official statistics are an invalu-able source of data for social research.But they are not just ‘about’ an indepen-dently given universe of social objects andevents, they are in part constitutive of it.The administrative power generated bythe nation-state could not exist without theinformation base that is the means of itsreflexive self-regulation. (1985: 180, empha-sis original)

Recent writings by human geographers andother social scientists have drawn on MichelFoucault’s (1991) concept of ‘governmental-ity’ to capture the relationship between statepower, information (or knowledge) and theconstitution of the objects of governance.2

‘Governmentality’ refers to the ways in whichknowledge, information and understandingoperate as tools (or ‘technologies’) of govern-ment. For Foucault, knowledge, includingthe kinds of knowledge produced by officialinformation, is not a neutral, transparentwindow on to a pre-existing social world, inwhich social problems are somehow simplyrevealed by objective fact-finding on the partof state organizations. On the contrary,Foucault’s approach implies that the socialproblems that form the objects of governmentpolicy in some sense come into beingthrough the production of official knowledgeof them. Take the example of ‘regionaleconomic development’ which has beenof longstanding concern to both economicgeographers and governments. Official eco-nomic statistics are used to identify regionswith economic problems, such as high unem-ployment or an overdependence on a narrowrange of (perhaps declining) industrial sectors.In order to do this the state must first definethe boundaries of each economic region(or ‘regional economy’).Yet as recent work bygeographers has shown, the very idea of

coherent regional economies is at bestcomplex and problematic and, at worst, down-right misleading (Allen et al., 1998). Regionalterritories are certainly not natural or self-evident containers for economic processes. Infact, the definition of regions is itself as mucha contestable political process as the develop-ment of regional economic developmentpolicy. From the perspective of the literatureon governmentality, ‘problem regions’ asobjects of state policy-making exist onlybecause of the state’s acquisition and mobiliza-tion of particular kinds of economic knowl-edge and information (Painter, 2002). This isnot to deny the existence of high levels ofpoverty or other kinds of economic distress inparticular places, but it does call into questionthe way in which economic problems cometo be framed as such and, in this example, howthey come to be framed as ‘regional problems’requiring ‘regional’ solutions.

The contemporaryinformational state

One of the important insights of the Italianpolitical theorist, Antonio Gramsci, concernedthe relationship between force and consent inthe maintenance of political power. ForGramsci, in the modern world power is notsolely a question of control over armed forcebut also involves the active co-operation of thepopulation. Both the creation and managementof this ‘consent’ by the state and its use of ‘coer-cion’ (police, prisons, etc.) require informationto be gathered on the population. As Giddensputs it,‘surveillance as the mobilizing of admin-istrative power – through the storage and con-trol of information – is the primary means ofthe concentration of authoritative resourcesinvolved in the formation of the nation-state’(1985: 181).

During the twentieth century informationcollection, storage and use by states have

OFFICIAL SOURCES

49

3151-Ch-02.qxd 3/2/04 2:36 PM Page 49

become increasingly sophisticated anddifferentiated.After the Second World War, inparticular, the role of states has expanded con-siderably, with a concomitant increase in theirinformational activities. First, in the decadessince the 1940s (at least until recently), mostdeveloped or industrialized states have beenwelfare states. This has involved governmentsin the widespread provision of public servicesin the fields of social security (e.g. old-agepensions), health care and education. Whilethese expenditures have seen a degree ofretrenchment in the 1980s and 1990s, theyremain substantially above the level of stateprovision typical of the nineteenth and earlytwentieth centuries. Planning and providingthese services require substantial amounts ofinformation relating to the geographical dis-tribution, age and health of the population.

Secondly, over the same period, state claimsto legitimacy have shifted in primary focusfrom ‘high’ politics (national security and themaintenance of state borders) to ‘low’ politics(trade, economic growth, social welfare). Thishas involved a relative decrease in thecollection of ‘intelligence’ information (i.e. spy-ing) and diplomacy and a relative increase in thecollection of information relating to economicperformance, industry and employment.

Thirdly, information is now held not onlyon populations or geographical areas as awhole but also to a very great extent onmembers of the population as individuals.In contemporary Britain every adult isassumed to be registered with the NationalHealth Service, the national insurance systemand the electoral registration officer.Most willalso be recorded by the Inland Revenue andmany by the Driver and Vehicle LicensingAuthority and by the Passport Office. Otherspecialist agencies such as the social servicesand educational organizations have files onparticular indviduals.The police maintain thePolice National Computer which records notonly the details of convicted criminals but

also details of police contacts with othermembers of the population, including manywho may be completely innocent of anywrongdoing.This type of information reachesits most developed form in so-called totalitariancountries, in which large parts of the stateapparatus are given over to the surveillanceand monitoring of individuals. As Giddenspoints out, however, the difference between‘totalitarian’ states and ‘democratic’ ones is adifference of degree, rather than kind.Accordingto Giddens, all modern states have totalitariantendencies (1985: 310). Indeed, since totalitari-anism depends upon surveillance, a keyprecondition of its emergence exists by defi-nition in all modern states.

Information collected at the level of theindividual is undoubtedly highly significant interms of states’ social control activities, especiallyat the more ‘coercive’ end of the consent–coercion spectrum. However it is not a signifi-cant source of data for geographical research.The reason for this is, of course, its confiden-tiality. Information on individuals is confiden-tial to the state, for two main reasons: first, forthe protection of the individual concernedand, secondly, for the protection of the inter-ests and activities of the state.Thus informa-tion relating to crimes and criminals, inparticular, is kept secret in order to maintainits power. If suspected criminals were aware ofthe information held about them, it is argued,they would be better able to evade the effortsof the state to police them.

New technology has added a further twistto debates about the secrecy of state informa-tion.The use of microelectronic informationtechnology has vastly increased the capacityof both the state and other organizations tostore, sort, process, duplicate and analyseinformation, including information aboutindividuals. This has led to concern amongcivil rights organizations about the potentialuses to which such readily available largequantities of data might be put. As a result of


50

3151-Ch-02.qxd 3/2/04 2:36 PM Page 50

such pressure, some countries now havestatutory protection for those about whomrecords are kept. In the USA the Freedom ofInformation Act 1966 guarantees (within limits)the rights of individual citizens to have accessto large parts of the state’s information archive.In Britain the provisions of the Data ProtectionAct are much weaker and relate only to accessto records on oneself, and until recently onlywhere these are held on computer.The Act wasintended primarily to allow individuals to cor-rect errors of fact in records relating to them.There are many exclusions, including, notably,police records. Historically there has been nogeneral freedom of information in Britain,maintaining the reputation of the British stateas one of the most secretive in the Westernworld (Michael, 1982).

Through the highly restrictive operation ofthe Official Secrets Act, most informationheld by the British government is confidentialunless expressly released to the public. Not-withstanding some recent amendments toremove some of the more absurd examples ofWhitehall secrecy, there is still much informa-tion gathered by the British state which isreserved for the confidential use of govern-ment alone.This confidentiality extends wellbeyond information on particular individualsto cover government-commissioned reportson a whole variety of subjects of potentialinterest to academic researchers.Again, in theUSA there is considerably more accessthrough the Freedom of Information Act. Insome cases this results in ‘secret’ British infor-mation being publicly available in the USA.This picture may change in the future, as theBritish government has introduced a Freedomof Information Act. However, the legislationhas been heavily criticized by civil rightsorganizations as likely to restrict further,rather than to open up, access to governmentinformation.

For the most part, therefore, the officialinformation most used by academic researchers

is that which is actively made public by thestate, usually through publication through agovernment publishing house, such as HerMajesty’s Stationery Office in Britain (nowmostly privatized under the trading namethe Stationery Office Limited) or the USGovernment Printing Office, or through amultilateral agency such as the EuropeanCommission or the United Nations. Theremainder of this chapter considers this subsetof government information which is mostwidely available, even though it may representonly a small proportion of the information inthe state’s possession.

Government informationorganizations

Most contemporary governments publishlarge amounts of statistical and other infor-mation relating to their populations andeconomies and social systems.This informa-tion is made available as a public service toresearchers, journalists and other interestedindividuals through the publication of statis-tical reports on a regular (e.g. annual ormonthly) basis. Such reports are normallyavailable in large public libraries, in mostuniversity libraries and can be purchasedfrom government publishing houses. Typicalexamples include the British government’sAnnual Abstract of Statistics and the USgovernment’s yearly Statistical Abstract of theUnited States.

These materials are very useful to humangeographers undertaking research. A widerange of topics is covered and the figures areusually detailed and can be broken down intosmaller geographical areas and/or time peri-ods. Frequently even more detail, and the pos-sibility of cross-tabulation, is available fromthe government department concerned.Increasingly this material is being, or will be,made available on various electronic media,

OFFICIAL SOURCES

51

3151-Ch-02.qxd 3/2/04 2:36 PM Page 51

such as computer-readable magnetic tapes ordisks. In the future more and more of thesedata will be available for use in computermapping and geographical informationsystems, allowing rapid cartographical repre-sentations to be constructed from officialstatistics. In many countries the state has itsown cartographic service (in Britain, theOrdnance Survey), reflecting the importantrole of territory in the formation and securityof modern states.

The data contained in official statistics areoften population (rather than sample) data.This makes them, at least in principle, morereliable than similarly constructed sample dataas there will be no sampling errors. Samplingerrors arise because the characteristics of asample of the population are unlikely to beprecisely the same as those of the wholepopulation. However, there may be errorsassociated with survey design and implemen-tation. Where population data are available,descriptive statistics for the population can bederived, and it is not necessary to infer popu-lation characteristics from a sample. Wheresamples are used in official statistics, they arelikely to be large (or at least larger than in mostacademic research) and this will tend to reducethe significance of sampling errors. In manycases government agencies have been at theforefront of the development of social surveytechniques, and the reliabilty of the data theyproduce should thus, again in principle, begood. (See Chapter 5 for further informationon the use of samples and surveys, and Chapter8 on the interpretation of numerical data.)

For all these reasons (availability, coverage,detail and reliability) official statistics are veryuseful in human geographical research.However, as has been discussed above, officialstatistics are not first and foremost producedas a public service. The principal aim oftheir construction is to inform governmentpolicy. Furthermore, as we saw in the work ofHakim and Giddens, official statistics influence

the society they represent by defining andcategorizing society in particular ways ratherthan others.At the end of this chapter we willconsider some of the technical questionswhich arise from these theoretical insights andlook at some specific examples to illustrate theproblems and possibilities of using official sta-tistics in geographical research. Before doingso, however, we will briefly discuss the changingrole of government statistical agencies, againusing the British state as an example.

In Britain, the modern data-constructionactivities of the state date from 1801, the yearof the first national Census of Population.Since then, a Census of Population has beentaken every ten years, with the exception of1941. In the twentieth century this basic sur-vey was joined by an ever-increasing output ofofficial information.The first four censuses (in1801, 1811, 1821 and 1831) were undertakenin England and Wales by the Overseers of thePoor under the guidance of John Rickman,the Clerk to the House of Commons. AGeneral Register Office (GRO) was estab-lished in England and Wales in 1837 and inScotland in 1855. The Registrar General ineach case was responsible not only for the reg-istration of births, marriages and deaths butalso for undertaking the census (Mills, 1987).Initially the GRO’s brief was limited to thesetwo main functions. Social survey research wasin its infancy and the usual means, apart fromthe census, for gathering information for thegovernment were inquiries carried outby royal commissions on the basis of subjectiveevidence from public officials aroundthe country (Burton and Carlen, 1979;Whitehead, 1987). In the late nineteenth andearly twentieth century the investigations ofCharles Booth and Beatrice and Sidney Webbestablished the usefulness of social surveys, butit was not until the Second World War that thepopulation census was joined by regularsample surveys carried out on behalf of thegovernment.


52

3151-Ch-02.qxd 3/2/04 2:36 PM Page 52

The wartime government created theGovernment Social Survey to ‘gather quanti-tative information on, among other things,the state of public morale in wartime condi-tions’ (Whitehead, 1987: 45).After the war thesurvey organization, part of the CentralOffice of Information (formerly the Ministryof Information), was retained and in 1957 itwas given the job of undertaking a regularFamily Expenditure Survey, which continuesto the present day. From 1967 to 1970 it wasmade an independent department, beforebeing merged with the GRO to form theOffice of Population Censuses and Surveys(OPCS). The OPCS was responsible for thecensus, social surveys, health statistics from theNational Health Service and populationregistration, and came under the responsibilityof the Secretary of State for Health.

The OPCS was complemented by a sepa-rate Central Statistical Office (CSO). Like theGovernment Social Survey it started life in1941 to provide better statistics for the man-agement of the wartime economy. It remainedpart of the Cabinet Office until 1989 when itbecame a separate government departmentresponsible to the Chancellor of the Exche-quer. Its main responsibility lay in the collec-tion and publication of trade, financial andbusiness statistics, and the national accounts. Italso housed the government Statistical Serviceand published a wide range of official statisticson behalf of other government departments.

In 1996, the OPCS and CSO merged toform a new ‘Office for National Statistics’(ONS), an independent government depart-ment responsible to the Chancellor of theExchequer. The government claimed that thenew organization would ‘meet a perceived needfor greater coherence and compatibility in gov-ernment statistics, for improved presentationand for easier public access’ (ONS website).Further changes took place following the elec-tion of the Labour government of Tony Blair in1997. A new organization, National Statistics,

was launched in June 2000, headed by thecountry’s first National Statistician. A StatisticsCommission was also established to oversee thequality and integrity of official statistical data.

In many cases the statistics produced by thegovernment reflect information which is col-lected in the course of other activities. Theunemployment figures, for example, are aproduct of the operation of the unemploy-ment benefits system. Sometimes statistics willbe collected specifically to inform new policyinitiatives. In many cases today this work iscontracted out to external, non-governmentalresearchers. These contractors may be dedi-cated survey and research organizations, suchas the market research companies, or they maybe academics working in universities. For aca-demics there are substantial attractions to thistype of work. It represents a source of researchfunding at a time when other funds are diffi-cult to come by and it appears to provide away for academic researchers to influence thepolicy-making process. Very often, however,there are unforeseen difficulties.There are fre-quently restrictions placed on the publicationof results, and often the information requiredby the government takes a different form fromthat suggested by the academic judgement ofthe researcher. It is also doubtful how far gov-ernment policy is changed as a result of suchresearch. Frequently, broad policies have beenagreed and research is conducted either to jus-tify the policy after the event or to prepare forits implementation. Where results contradictassumptions of government policy-makers it islikely that the research will be suppressedrather than that the policy will be changed.

Understandingthe construction ofofficial information

Government statistical organizations are per-haps the most organized and focused examples

OFFICIAL SOURCES

53

3151-Ch-02.qxd 3/2/04 2:36 PM Page 53

of the information and knowledge productionactivities of modern states.As we have alreadynoted, however, governments also producehuge quantities of non-numerical informa-tion. In both cases, however, similar motiva-tion is at work. The process of governmentboth requires and produces information. Atthe same time, the ways in which informationis constructed and presented help to shape the‘policy landscape’ in particular ways; somepolicy approaches are made possible and legiti-mated, others are blocked off and discredited.In this sense no information is ‘neutral’ andthis is certainly true of government informa-tion, whether statistical or textual. Using offi-cial data sources in geographical research thuseffectively requires careful attention to theprocess of data construction and means thatthe researcher should ask some hard questionsabout the information he or she is proposingto use:

• Why was the information constructed?• To which government policies does it

relate?• Have policy concerns influenced which

data were constructed and how?• In what ways?

The question ‘What policy concerns or politi-cal ideas motivated the construction of theinformation?’ is important even where theanswer seems obvious. Take inflation forexample.We take it for granted that informa-tion about price increases will be publishedon a regular basis. But why is this important?One could argue that collecting retail priceinformation at all sets ‘inflation’ up as a policyproblem and that it leads to particular policyresponses. Some economists, for example,reject the idea that a goal of low inflationshould be as absolutely central to governmentpolicy as it has been in recent years, but thehigh profile of the inflation statistics tends toreinforce the idea that inflation is a very bad

thing. Until recently in Britain there was notseen to be a need to compile information ona national basis about the educational attain-ment of school children while they were stillat school.Today, testing of pupils on a nationalbasis at 7, 11 and 14 provides informationallowing ‘league tables’ of school performanceto be constructed. Leaving aside the veryimportant issue of whether the information isa good measure of school performance, thereare profound implications from the fact theinformation is constructed at all. There is aclose relationship here between the compila-tion of the information on a national basisand the former Conservative government’spolitical aim of increasing competition amongschools for the best pupils and scarceresources. Whether one agrees with this aimor not, the data-construction activity herecannot be thought of as a neutral activity sep-arate from the political controversy. Rather itis part of the political process.

These questions about motivation and therelationship between information and policiescan be asked of data sources of all types(statistical, textual and graphical). In addition,there are some further questions that need tobe considered in the case of textual data:

• Which ‘voices’ are present in the text?Who is ‘speaking’ – a supposedly disinter-ested expert? A professional politician? Aninterested organization or individual?Perhaps several voices are present, as maybe the case in a report of public inquiry,for example.

• By contrast, whose voices are absent? Arethere ‘silences’ in the text representingissues or points of view that have beenneglected or deliberately left out?

• What mechanisms led to the productionof the report? How was the text puttogether – verbatim transcription? Sum-maries of longer drafts? Single or jointauthorship? And so on.


54

3151-Ch-02.qxd 3/2/04 2:36 PM Page 54

• What rhetorical devices and figures ofspeech are used to convey the message?What metaphors are used? What effect dothese have on the content of the document?

There is also a set of questions that relatespecifically to statistical data:

• Which categories have been used andwhy?

• What would be the effect of using differ-ent categories?

• What and who are included and excludedfrom the count and why?

• What would be the effect of includingother variables or groups?

• Which sampling and survey techniqueshave been used?

• What are the likely errors and biases asso-ciated with these techniques?

• What corrections and adjustments havebeen made to the results to allow forerrors, and what effects have these had?

None of these questions (with the partialexception of those concerning the mathemat-ics of sampling error) can be answered in apurely technical way. They are matters ofjudgement and debate. Much will depend onthe extent to which information is availableon the way data have been constructed andon the political and theoretical stance takenby the researcher (see Chapter 12).3

We will now consider these factors in moredetail in relation to specific examples.

Domesday Book

On the face of it, it might be thought thatDomesday Book would be of interest only tothose studying the historical geography ofearly medieval England. While this may betrue of the evidence it supplies of the eco-nomic and political landscape of the time, italso provides a fascinating illustration of someof the more general issues that arise in using

official sources of data. Many of the problemsand difficulties of using Domesday Book as aresearch source are also present in using con-temporary official sources, albeit in moremuted ways. Domesday Book, therefore, givesus a way of highlighting in acute form manyof the factors that need to be considered inusing any official data.

As every British schoolchild knows, SaxonEngland was invaded in 1066 by William,Duke of Normandy, who established a muchstronger system of government in place of therelatively weak and divided Saxon rule. Aspart of this process, a fully fledged feudalsystem was put in place, under which all theland belonged to the sovereign who divided itamong the nobles as the chief tenants. Thechief tenants could divide up their lands andsublet them to other tenants, who could sub-let again and so on all the way down the socialand political hierarchy. Each feudal tenancyinvolved sets of mutual rights and obligations.The lord provided protection to his tenants(vassals) and they in turn provided militaryservice and paid certain dues. After his mili-tary victory, William abolished most of theexisting Saxon earldoms and parcelled out theland afresh, much of it to his Norman sup-porters.The smallest unit of feudal organiza-tion was the manor which was the seat of alocal feudal lord to whom the local popula-tion owed allegiance. Manors possessed theirown land (known as the demesne) and alsosublet land to the lowest tier of tenants.

To understand the construction of the doc-ument, is necessary to appreciate the adminis-trative geography of Norman England. Themajor territorial divisions of the country werethe shires (from Old English scir), also knownas counties (from Old French conté ). Countieswere divided in turn into ‘hundreds’ so-calledbecause they corresponded nominally to anarea containing one hundred units of taxationknown as ‘hides’. In those parts of the countrythat had been heavily affected by the Danish

OFFICIAL SOURCES

55

3151-Ch-02.qxd 3/2/04 2:36 PM Page 55

invasions the Scandinavian terms wapentake(instead of ‘hundred’) and carucate (instead of‘hide’) were used. The word ‘hide’ derivesfrom the Old English word for household andrefers to a unit of land large enough to sup-port a family and its dependants. Hundredswere subdivided into vills (feudal townships).However, the geography of the vills did notcorrespond directly to that of the feudalmanors because a manor could possess land inmore than one vill.

The historical geographer, H.C. Darby,made the interpretation of Domesday Bookhis life’s work and he was well aware of thecomplexities involved in the data-collectionprocess.The precise circumstances that led tothe survey that produced Domesday Book areobscure. However, the Anglo-Saxon Chroniclerecords that, in 1085, following discussion at acouncil at Gloucester, ‘King William sent hismen into each shire to enquire in great detailabout all its resources and who held them’(Darby, 1977: 3) and that all the surveys weresubsequently brought together. There is verylittle in the way of direct evidence about howthe surveys were undertaken. However, it ispossible to make some deductions about theprocess of data construction from DomesdayBook itself and, especially, from a number ofsubsidiary documents associated with it(Darby, 1977: 3–9).

Some evidence of the operation of theinformation-collection process is provided bya document called Inquisitio Eliensis, which isa survey of the estates of the abbey of Ely.According to Darby:

The Inquisitio tell us that the king’s commis-sioners heard the evidence ‘on the oath ofthe sheriff of the shire, and of the baronsand of their Frenchmen, and of the wholehundred – priests, reeves and six villeinsfrom each vill’. There was a separate jury foreach hundred, consisting of eight men,whose function ‘was apparently to approveand check the information variously assem-bled’ … half the jurors were English and the

other half Norman … We cannot saywhether the commissioners themselvesattended every hundred court … orwhether … they merely held one session inthe county town. A number of entries makeit clear that they sometimes heard conflict-ing evidence … It has been suggested ‘thatfiscal documents already in existence weredrawn upon to help with the compilation ofa partly feudal and partly fiscal enquiry’.The making of the inquest may have been afar more complicated process than was atone time thought. (1977: 5)

While it is impossible to be certain that thisprocedure was followed in all the localinquiries that produced Domesday Book, theoverall picture is fairly clear and suggests thatcommissioners appointed by the king weresent out around the country to conduct thesurvey, and that they did so, not through ahouse-to-house inquiry as might happentoday, but by gathering together ‘juries’ oflocal representatives including a cross-sectionof the population and questioning themabout the local situation.

The Inquisitio Eliensis also lists 11 questionsto be asked in each hundred (see Box 2.1).AsDarby (1977: 5) notes, ‘whether these werethe “official instructions” for all counties, wecannot say, but they, or at any rate a similar setof questions, must also have been askedelsewhere’.

This ‘fieldwork’ seems to have been con-ducted on a county-by-county and hundred-by-hundred basis and provided the initialsurvey returns. It appears that these initialreturns were then summarized in documentsthat grouped together several counties, stillorganizing the information on an area-by-area basis. From these summaries the finalDomesday Book proper was prepared. In thefinal version some of the detailed informationwas edited out but, more significantly, theareal organization of information wasreplaced by a structure that mirrored the feu-dal system. That is, for each county the datawere:


56

3151-Ch-02.qxd 3/2/04 2:36 PM Page 56

1 What is the name of the manor?2 Who held it in the time of King Edward?3 Who holds it now?4 How many hides are there?5 How many teams, in demesne and among the men?6 How many villeins? How many cottars? How many slaves? How many freemen?

How many sokemen?7 How much wood? How much meadow? How much pasture? How many mills? How

many fisheries?8 How much as been added or taken away?9 How much was the whole worth? How much is it worth now?

10 How much had or has each freeman or each sokeman? All this is to be given in trip-licate; that is in the time of King Edward, when King William gave it, and at thepresent time.

11 And whether more can be had than is had?

Explanation of terms not defined elsewhere:

‘In the time of King Edward’ refers to the reign of Edward the Confessor (1042–66).‘Teams’ means ploughteams made up of (usually eight) oxen. ‘Villeins’ were tenantstied to the land on which they worked. ‘Cottars’ were tenants who occupied a cottagein return for labour. ‘Freemen’ (unlike ‘villeins’) were allowed to leave the land onwhich they worked. ‘Sokemen’ were a class of personally free peasants attached to alord rather than to the land.

Box 2.1 The Domesday questions

rearranged under the headings of the mainlandholders, beginning with the king him-self and continuing with the ecclesiasticallords, the bishops followed by the abbots,then with the great lay lords, and finallywith the lesser landholders in descendingorder … It follows that if two or more lordsheld land in a village, the different sets ofinformation must be assembled from theirrespective folios in order to obtain a pictureof the village as a whole. (Darby, 1977: 8)

There are two documents that are thought tobe surviving regional summaries. These arethe so-called Exeter Domesday Book cover-ing the south west of England and the surveyof the eastern counties of Norfolk, Suffolkand Essex. The information in the Exetersummary was edited and incorporated intothe main Domesday Book, but that for theeastern counties, was, for some reason, never

incorporated. Domesday Book today, therefore,consists of two volumes, the Great Domesday(also known as the Exchequer Domesday)that was compiled in Winchester at the king’sTreasury and the Little Domesday that coversNorfolk, Suffolk and Essex and which repre-sents an earlier stage in the data-constructionprocess.

Darby (1977: 8) provides an example oftypical entry from the Great Domesday Bookin which all the information relating to thevillage of Buckden is to be found in oneentry, because all the land was held by onelandlord, the Bishop of Lincoln:

In Buckden the bishop of Lincoln had 20hides that paid geld [i.e. tax]. Land for20 ploughteams. There, now on thedemesne 5 ploughteams, and 37 villeins and20 bordars having 14 ploughteams. There, a

OFFICIAL SOURCES

57

3151-Ch-02.qxd 3/2/04 2:36 PM Page 57

church and a prieset and one mill yielding30s (a year), and 84 acres of meadow. Woodfor pannage one league long and one broad.In the time of King Edward (i.e. in 1066) itwas worth £20 (a year), and now £16 10s.

From the hundreds of entries like the onefor Buckden, historical geographers suchas Darby have been able to construct detailedaccounts of the geography of NormanEngland (Darby, 1950–1; 1951; 1977). In thisprocess of interpretation, all the questionsraised above about the construction of dataare relevant.

In the case of Domesday Book, the moti-vation for the gathering of the information isa matter of some debate.Traditionally it wasviewed as a taxation assessment (Round,1895; Maitland, 1897). More recently it hasbeen suggested that it provides a record ofthe patterns of feudal ownership and of theKing’s relations with his senior feudal tenants(Galbraith, 1961). Its role could also havebeen primarily symbolic, serving to securethe Conqueror’s authority over his new sub-jects (Clanchy, 1979: 18).To some extent, theview that is taken of the purpose of the doc-ument will affect the interpretation. If it istreated as a fiscal document, for example,issues of accuracy become very important.On the other hand, if its role was largely sym-bolic, then whether the Bishop of Lincoln’sholding in Buckden was precisely 20 hides orsome other figure arguably becomes lessimportant than the insight provided into therelationship between Church and state by theallocation of lands to bishops in the firstplace. The questions about the character ofthe information gathered are also pertinent.Livestock, for example, are largely ignoredby the survey, while the information onindustry and towns is very limited.While thisreflects the rural and agricultural basis ofthe Norman economy, it also influencesthe interpretations that can be made of thesource.

The UK unemploymentfigures

Unemployment has been one of the mostimportant social and political problems sincethe mid-1970s. Politicians of all parties recog-nize this, although proposed solutions varyradically. For those on the free-market rightunemployment is said to be caused either bythe unwillingness of individuals to work or bythe effects of state intervention in the labourmarket producing disincentives to do so.Thetraditional left-wing view has been thatunemployment is a product of unfetteredfree-market capitalism and that state regula-tion or economic intervention is required torespond to market-generated economic crisesand to more general problems of market fail-ure. Whichever view is taken, it is clear thatmeasuring the level of unemployment accu-rately provides an important indication ofeconomic success and, by implication, of thesuccess of social and economic policies ofgovernments. However, this raises the spectreof manipulation; if the measure of unemploy-ment falls, governments can claim success fortheir policies even if unemployment itself hasremained stable or increased. According tomany commentators, this was precisely whathappened in Britain during the 1980s and1990s.

The geography of unemployment andemployment has also long been of interest toeconomic geographers. In order accurately tointerpret the geographical distribution ofemployment changes and trends and to mapand analyse the pattern of employmentopportunity and labour market activity, geog-raphers use official statistics.Accuracy, and theprocess of data construction, is therefore notjust an issue for assessing policy success; it alsohas profound implications for academicanalyses, including those on which policiesare partly based.Writing at the end of a longperiod of government by the Conservative


58

3151-Ch-02.qxd 3/2/04 2:36 PM Page 58

Party (1979–97) during which the integrityof official British unemployment data had fre-quently been a matter of political debate,Levitas (1996) provided a detailed review ofthe measurement of unemployment in theUK.This section is based on her analysis.

At the heart of the problem is the question ofdefinition. Prior to 1982, the official measure ofunemployment was based on a count of thoseregistered as seeking work. Those who wereclaiming welfare benefits were required toregister, but others could also register if theywished, so that the count included those whowanted to find work but who were for onereason or another ineligible for benefits. In 1982the official measure was changed so that itincluded only those claiming benefits.This ledto a reduction in the measure of unemploymentof about 200 000 people (out of a total of over3 million).As Levitas points out (1996: 48), thischange was highly significant because it turnedthe unemployment count into an administrativemeasure. This meant that changes in the rulesrelating to benefit entitlement would have animmediate effect on the unemployment mea-sure which was not produced by any real changein the economy. By the time of Levitas’s study,30 further changes had been made to the unem-ployment count. This meant that the officialheadline figure for unemployment in the UK in1996 was based on a very different definitionfrom that used in 1979. This, according toLevitas, had three consequences:

First, there is the effect upon unemployedpeople of a benefit system using increas-ingly stringent criteria for the receipt ofdeclining amounts of benefit. Secondly, thecount cannot be treated as a continuousseries, as its coverage becomes narrowerover time. Thirdly, it cannot be regarded asa measure of unemployment. (1996: 48)

The changes included the following:

• 1986: introduction of more stringentavailability for work criteria, especially

affecting women with child-care responsi-bilities (107 000 claims for benefit dis-allowed under this rule in 1987).

• 1988: removal of benefit entitlement from16 and 17-year-olds (reduction in thecount: over 100 000).

• 1989: introduction of requirement thatclaimants must be able to prove they are‘actively seeking work’.

• 1996: introduction of Jobseekers Allowance( JSA) in place of unemployment benefitinvolving means testing of benefit claimsafter only six months, instead of a year.

All these changes affected the unemploymentcount regardless of the real state of the econ-omy. Whatever their merits in terms of theoperation of the social security system, theyproduced entirely administrative reductions inthe measure of unemployment. For manycommentators, therefore, the official countbecame largely useless as a measure of unem-ployment. Even the retiring head of the gov-ernment statistical service asserted in 1995that ‘nobody believes’ the claimant count(Levitas, 1996: 62).

However, as Levitas pointed out, the prob-lems with the claimant count do not meanthat official statistics could not be used to pro-vide better measures of unemployment. Shereviewed a number of alternative approaches,including that provided by the Labour ForceSurvey (LFS). The LFS is a sample survey inwhich 60 000 households are interviewedabout their labour market activity.

Although it produces a better measure ofunemployment than the claimant count, andone that can be improved further by academicresearchers undertaking secondary analysis ofthe data, the LFS does have some drawbacksof its own. Because it is a sample survey ratherthan a count it is subject to sampling errors.There are also problems with the sampleframe. Because it is based on addresses, itexcludes those living in hostels and with no

OFFICIAL SOURCES

59

3151-Ch-02.qxd 3/2/04 2:36 PM Page 59

fixed abode (two groups whose numbers haveincreased with the rise in homelessness andwho are disproportionately unemployed).Finally, where individuals are absent at thetime of the interview, proxy responses arerecorded by asking other members of thehousehold.The accuracy of some of this infor-mation may be rather questionable, and it mayalso be subject to gender bias. For example, awoman seeking work who spends her timeundertaking domestic chores might describeherself as unemployed but be described by hermale partner as a ‘housewife’ (Levitas, 1996:53–4). ‘Housewife’ is a particularly genderedterm and was itself popularized, in part at least,by official policy after the Second World Warwhen governments were faced with largenumbers of men returning from militaryservice, and wanted to encourage women toreturn to domestic labour and give up the paidemployment many had undertaken during thewar. Despite these drawbacks, the LFS doesprovide significantly more reliable data onunemployment rates, has the twin advantagesof following international definitions (allow-ing for international comparison) and allowingdetailed analysis to examine relationships amongother social variables, such as gender, ethnicityand age.

While the LFS serves much better than theclaimant count for measuring unemploymentaccording to international definitions, and ina more consistent way, there are some widerissues of definition that call into questionconventional measures of unemploymentaltogether. The population is conventionallycategorized not into two groups of employedand unemployed but into three: employed,unemployed and ‘economically inactive’.Theexistence and definition of this third groupraise further problems for the measurement ofunemployment and as a consequence for thedefinition of unemployment as a politicalproblem. The LFS definition of unemploy-ment is explained by Levitas as follows:

To be counted as unemployed in the LFS,respondents must have:

(a) done no paid work in the week inquestion;

(b) wish to work;(c) be available for work within the next

fortnight;(d) have made some effort to seek work in

the past four weeks, or be waiting tostart a job already obtained. (1996: 55)

Take the example of a middle-aged man whohas been out of work for two years and whobelieves there are no jobs available for him. Itis not unlikely that he will have made noeffort to seek work in the past four weeks but,at the same time, be available for work andwish to work. On the LFS criteria he will berecorded as economically inactive, rather thanunemployed. Similarly, many married womenare defined as economically inactive, ratherthan unemployed, because, while they maywish to undertake paid work, they are notactively seeking employment because they areoccupied with domestic responsibilities.Anomalies such as these have led to employ-ment and unemployment falling simultane-ously in some parts of the country at sometimes. If the population is static but employ-ment is falling, this is hardly a picture of eco-nomic dynamism and success, and yet changesin ‘economic activity’ rates might allow a gov-ernment to claim as much. At bottom theproblem is precisely the one that forms thebasis of the arguments in this chapter, namely,that statistics are a social construction. AsLevitas (1996: 60) puts it: ‘The line between“unemployment” and “non-employment” or“economic inactivity” is imposed upon a real-ity that is far more complex and fluid.’

Far from reflecting reality in a neutral andobjective way, official statistics depend uponparticular understandings of reality, and theseunderstandings affect (often in profoundways) the picture that is painted of social andeconomic life by official sources of data. In afurther shift in the official statistical view of


60

3151-Ch-02.qxd 3/2/04 2:36 PM Page 60

unemployment since Levitas’s analysis waspublished, the New Labour governmentelected in 1997 decided to adopt the LabourForce Survey definition as the source ofBritain’s official unemployment figures.

Conclusion

In this chapter we have deliberately not pre-sented detailed case studies of geographers’use of state information as to do so would beto focus more attention on the interpretation ofofficial sources than on their construction. Byconsidering the issue of construction in detailwe have tried to demystify official sources –to knock them off their pedestal, as it were –and to show that just because informationcomes with an official seal of approval doesnot mean that researchers can take it simply atface value. In fact, as we have suggested, offi-cial sources can be just as partial and subjectto errors and evasions as any other kind ofdata. Indeed, when it comes to political influ-ence, official sources are probably moreaffected than others. First, official informationis always collected for a governmental (andhence political) purpose. Secondly, partlybecause of the cachet of the ‘official’ label,official sources can appear to be more neutral,

reliable, comprehensive and authoritative thanother kinds so that their political charactercan be disguised. For this reason, paying atten-tion to the process through which they areconstructed is particularly important, as thisshould highlight the purposes for which theywere produced and allow the researcher totake these into account in their interpreta-tion. None of this should lead to an outrightdismissal of official data, though. For manygeographers they will continue to be aninvaluable source, partly because of theircomprehensiveness and quality. Handled care-fully, official information can be expected tofuel the practising of human geography for along time to come.

Notes

1 Prominent examples include Johnston (1979),Taylor and Johnston (1979), Johnston et al.(1988), Shelley et al. (1996) and Johnston andPattie (2000).

2 Murdoch and Ward (1997), Blake (1999),Dean (1999), MacKinnon (2000) and Raco(2003).

3 Those interested in considering this latterpoint in more detail should also see the com-panion volume to this book, ApproachingHuman Geography – Cloke et al. (1991).

OFFICIAL SOURCES

61

3151-Ch-02.qxd 3/2/04 2:36 PM Page 61

Introduction

The state, in all its guises, is not the onlyagency to provide the human geographer withrich sources of preconstructed data. Indeed, intandem with the extensive growth in govern-ment record-keeping, which we detailed inthe previous chapter, there has been a dramaticincrease in the extent and scope of ‘non-official’data. The inexorable rise in informationtechnology and electronic communication(including the Internet and World Wide Web),the ease of desktop publishing and printing,and the continued expansion of the technicaland bureaucratic nature of capitalist societyhave all combined to accelerate this process.The trend towards increased bureaucracy,administration and record-keeping in the pri-vate sphere was first noted by Weber (1978) atthe beginning of the last century. Since then,of course we have witnessed a veritable ‘infor-mation explosion’ as companies, news agen-cies, research organizations, political partiesand voluntary groups, to name but a few, haveall struggled with the collection, storage andcommunication of an ever-increasing tide ofinformation. In this chapter we will be con-cerned with those aspects of this informationwhich purport to tell the truth, throughunderstanding the construction of supposedly‘factual’ data. In contrast, the next chapter willexplore the construction of self-proclaimed

imaginative data through an examination ofliterature, music and the visual arts.

As we will maintain over the next twochapters, this distinction between ‘factual’ and‘imaginative’ sources should not be held toorigidly. Both types of data are socially con-structed, and both are produced for specificpurposes in specific social contexts.Moreover,all attempts to describe the world in a factualmanner involve particular ways of imaginingit, and all such accounts will contain someelement of subjectivity. Indeed, any construc-tion of factual or documentary knowledge isbased on a particular view of what is real orwhat constitutes the truth. However, becausewe are concentrating on the ways in whichthese sources are constructed, and because theprocesses behind the production of ‘imagina-tive’ and ‘factual’ sources do differ (to varyingdegrees), it is useful to consider the two typesseparately.

In this chapter we will examine the pro-duction of supposedly ‘factual’ data by themultifarious agencies and institutions whichlay outside the official realm of the state. Indoing so we will be concentrating on thosesources which are of greatest use to thehuman geographer. These vary enormously.The archetypal source, and the one that weare all most familiar with in our academicinquiry, is the text printed or handwritten onpaper, usually produced in the form of books,

Non-official Sources3

3151-Ch-03.qxd 3/3/2004 3:34 PM Page 62

journals and magazines. But the potential fieldof ‘factual’ documentary sources is muchbroader than this. The historical geographer,for instance, might make use of sources carvedin stone; those working with geographicalinformation systems will use electronic andmagnetic sources; the urban geographer inter-ested in housing change may well use displayboards in estate agents’ windows. It is also thecase that the text of a factual source does nothave to be written. Promotional or campaign-ing videos, documentary radio programmes,television news broadcasts and audio tapes oforal histories are all valuable sources of factualinformation about social phenomena, butnone of them are written documents.

In short, the possible scope of non-officialsources is extremely wide and can range fromthe most specialized company report costingthousands of pounds to a torn and tatterednotice in a shop window. Useful sources arealso produced by individuals just as much asby organizations and institutions. A geogra-pher looking at a person’s ‘sense of place’, forinstance, might use diaries and letters contain-ing snippets of information about the placewhere the individual lives.The message here isthat potential researchers need not feel boundby written sources, or by those which areovertly geographical in nature, or by thosewhich appear formal and thereby authorita-tive. Indeed, we would stress that the humangeographer can find a rich and rewardingrange of material in a whole variety of non-official sources – published and unpublished,public or private, written or spoken, wittinglyor unwittingly provided. Later in the chapterwe will explore some of this potential materialby examining the production, and use, ofselected examples in more depth. But beforewe do so we will outline some general advan-tages in using non-official sources and thenconsider the main types of such data youmight use.

Non-official sources ingeographical research

There are a number of reasons why geogra-phers might wish to turn to non-official ‘fac-tual’ data as a source of research material.Wehave already hinted at the breadth and scopeof the sources available and by looking acrossthese we can identify five potential advantagesin using such material:

1 Non-official sources can open up socialworlds which are inaccessible and relativelyclosed. Davis (1990) has used a variety ofnewspaper reports, police documents andverbatim records of televised interviews tolook at the military-style policing of con-temporary Los Angeles, with particularreference to the so-called ‘war on drugs’.In this instance both the police and thedrug gang members represent relativelyclosed worlds, but Davis is able to con-struct a vibrant picture of current eventsfrom his use of contemporaneous newsmaterial. From this he is then able to con-struct vivid accounts of the changingsocial and cultural geographies of urbanLos Angeles. In a similar manner, Jackson(1988a) has used unpublished archivalmaterial from voluntary and communitygroups, school reports and newspapers toexamine the experience of neighbour-hood change in Chicago.

2 Documentary sources of ‘factual’ dataoffer a major chance for the geographerto conduct historical and longitudinalstudies. Thus Massey (1984), in her pio-neering work on the geography of eco-nomic change, uses company records andreports to sketch out the historical perfor-mance of particular industrial sectors overthe last century, and Martin et al. (1993)use OECD sources to chart the declinein trade union membership over the last

NON-OFFICIAL SOURCES

63

3151-Ch-03.qxd 3/3/2004 3:34 PM Page 63

30 years, before drawing on currentmembership records to assess the geogra-phy of contemporary unionization.

3 In a similar manner, just as they can beused to compare events across time, non-official ‘factual’ sources be used to analysecross-cultural processes. Because similarrecords tend to be kept by similar organi-zations across national boundaries, theyoffer a reasonsably safe way of comparinglike with like (although see Dickens et al.,1985, for a comment on the pitfalls ofcomparative research, in their study ofhousing provision in Britain andSweden). Moreover, some agencies andinstitutions are international in character,and they provide comparative precon-structed data as part of their own research.The World Bank, for instance, publishes ayearly report of development indicators,and the International Labour Organizationproduces regular comparative material onissues concerning employment change.In a similar way, the World HealthOrganization produces internationalcomparative data on health issues. Allthese will be of potential use to thehuman geographer.

4 Another reason why ‘factual’ data fromnon-official sources are popular is becausethey are potentially very cheap and ofteneasy to access.The sources are often avail-able in reference libraries, and they areoften collected together in a usable form.One of the advantages of using any formof preconstructed material is the very factthat it is already constructed – someoneelse has spent the time and money tocollect the research data and put themtogether. A collection of these sourcesmay well have to be drawn together orused selectively in a manner which suitsthe individual researcher, but the bulk ofresearch expenditure has usually beenundertaken by someone else. Allied to

this is the fact that the researcher can drawon a potentially vast sample size, withoutbeing limited too much by considerationsof cost and time. To return to an earlierexample, by using the membershiprecords of just eight trade unions, Martinet al. (1993) were able to document thegeography of some 3 million individualunion members.

5 Finally, the researcher can use documen-tary material for ideas as well as for data.An examination, for instance, of the datacontained in the reports of the Anti-slavery Campaign will raise a number ofissues concerning child labour, the pro-duction techniques of multinational com-panies, gender relations within the ThirdWorld, the consumption habits of those inthe West, the international debt problemand so on. Researchers should not feelthey can consult these materials only insearch of ready-made data but, instead,non-official sources can offer a rich veinof ideas and stimuli for possible areas ofstudy.

As we have already noted, the range of poten-tial non-official data sources for the humangeographer is vast. Some are intended specifi-cally for research purposes, but most are not.Some are meant for public consumption inone way or another, but others are intendedto be private. Some are the product of inter-national organizations and agencies, whileothers are produced by individuals. All, how-ever, are the end result of deliberate socialconstruction, and all purport to tell the truthabout social events. Because we are stressingthat such sources can be understood onlywithin the context of their construction, wewill briefly outline the major data sourcesavailable to the researcher, according to thetype and intention of production. We willthen look in more depth at a smaller selectionof these, at their detailed construction and


64

3151-Ch-03.qxd 3/3/2004 3:34 PM Page 64

at the potential pitfalls in their use bygeographers.

Reference material

Some source material is designed specificallyfor reference purposes. This ranges fromlarge-scale electronic databases to individualresearch papers. Often such material is col-lected together and published in the form ofdirectories, manuals, almanacs, guides, regis-ters, yearbooks and calenders (Scott, 1990:156). Economic and historical geographershave long made use of commercial directo-ries, for instance, which list industrial andcommercial activities on an area-by-areabasis. Other commercial manuals are sectorbased and can offer technical informationregarding a particular industry or financialinformation regarding investment and share-holding. Social and urban geographers regu-larly use reference material on housing costssupplied by building societies to monitor thechanging nature of housing markets(see Hamnett, 1998).

Academic books, journals and articles area major source of reference material whichcan get overlooked simply because they areso familiar to researchers. In fact, a potentialproblem here is the sheer diversity of acade-mic material, but bibliographic databases,abstracts and digests can be used as tools forliterature searches. Some primary academicmaterial used in the writing of books andpapers is also available to the researcher.TheEconomic and Social Research Council helpsto fund two data archives at the University ofEssex, one concerned with qualitative dataand one concerned with computer-readablequantitative data.The former, QUALIDATA,aims to locate, assess and document qualita-tive data and arrange for their deposit insuitable public archives; to disseminate infor-mation about such data; and to encouragethe reuse of these data. A major function of

the centre is to maintain an informationdatabase about the extent and availability ofqualitative research material from a widerange of social science disciplines in general,whether deposited in public repositories orremaining with the researcher. The centrehas surveyed a huge range of ESRC-fundedqualitative social research projects datingback to 1970 and is cataloguing the datafrom classic postwar sociological studies aswell as monitoring current ESRC and otherfunded projects. The Qualidata catalogue(Qualicat) is available via the Internet(www.essex.ac.uk/Qualidata), from whichresearchers can search and obtain descrip-tions of qualitative research material, its loca-tion and accessibility. Quantitative data areavailable from the Data Archive, originallyestablished in 1967 and now holding some4000 data sets. While many of these sets arebased on official data, others are not and useresearchers’ own data in addition to thosefrom a range of other organizations. Thematerial can also be searched and accessedvia the Internet (www.data-archive.ac.uk). Aquick look at these two archives will givesome idea of the range and scope of ‘factual’material available to the researcher.

Research reports

Recent years have seen an explosion in semi-autonomous research centres, sometimes linkedto universities but more usually funded throughcharitable or private sources. These produceextensive series of research reports, coveringareas of interest to the geographer such ashousing, urban policy, economic development,environmental change, planning and health.

Other organizations and institutions produceresearch reports which, although not aimedprimarily at the academic researcher, containa range of useful materials. Many of thesereports are now produced for lobbying pur-poses, and the rise of official state information


65

3151-Ch-03.qxd 3/3/2004 3:34 PM Page 65

which we considered in the previous chapterhas been paralleled by the rise of oppositionaland lobbying material. Pressure groups such asShelter, the Low Pay Unit, Labour Researchor the Child Poverty Action Group produceregular publications and reports on housing,employment, trade union and poverty issues,respectively. An interesting collaborationbetween the Child Poverty Action Group andthe Social and Cultural Geography ResearchGroup of the Royal Geographical Society–Institute of British Geographers resulted in apublication on the social geography ofpoverty in the UK. This contains severalexamples of geographers using researchreports by unofficial agencies to build up apicture of spatial inequalities in housing,health, employment, educational achievementand income, and the volume itself can also beused as a rich documentary source (see Philo,1995). Aside from pressure groups, organiza-tions such as the Economist produce regularbriefings on financial issues, and ‘think-tanks’such as the Institute for Public PolicyResearch and specialized research institutes,such as the Joseph Rowntree Foundation, thePolicy Studies Institute or the TavistockInstitute, all have extensive publication lists.Political parties also have research depart-ments which publish their findings, as dotrade unions and employer organizations suchas the CBI and the Institute of Directors.

Reports, memosand minutes

Commercial and non-commercial institutionsalso produce a wealth of internal material inthe form of reports, memos, messages andminutes which, while not based on primaryresearch itself, can provide a rich source ofresearch material. Large-scale companies areconstantly producing reports and strategies,and any large institution will normally

produce at least annual institutional plans andforecasts. Sadler (1997), for instance, has usedcompany reports from a variety of car manu-facturers and suppliers to study the changinggeography of the European automobileindustry. Although most of this material willnot be aimed at the public market it is oftenpossible for the researcher to negotiate indi-vidual access, especially if constraints on theuse of the material are accepted. Throughsuch access one can glean realms of informa-tion on employment patterns, labourprocesses and productivity, investment deci-sions, marketing strategies, production targets,management techniques and so on.

Documentary media

A whole range of purportedly factualmaterial is produced by the news media.Broadcast news, documentary programmes,photographs, newspaper and press reportsand specialist magazines between them pro-vide many hours and column inches of‘factual’ material. Indeed, there are nowcable TV stations, radio channels and satel-lite broadcasting networks solely devoted tothe production of news. Much of thewritten material is available on microficheor via the Internet, and CD-ROMs are nowable to hold several years’ worth of newspaperreports on one disk. These can be searchedvia keywords so that ten years of pressreports can now be ‘filleted’ by the researcherin a matter of hours, if not minutes.Geographers regularly make use of suchmaterial in their work. Cresswell (2001), inhis work on the US tramp, has drawn onnewspaper reports, magazine articles anddocumentary photography in order toexamine how tramps were both perceivedand treated by the media. Interestingly, hiswork shows how the news media, far from


66

3151-Ch-03.qxd 3/3/2004 3:34 PM Page 66

simply documenting the ‘truth’ abouttramps, actually played a critical role inconstructing our knowledge about tramps,and was instumental in the making up ofthe tramp as a figure on the social and geo-graphical margins of society. In a similarvein, Beauregard (1993), in his book onpostwar cities in the USA, shows how thosewho were reporting and reflecting on thestate of urban America for newspapers andmagazines were central to the constructionof a discourse which promoted the idea ofurban decline – and which reshaped boththe public understanding of, and an expla-nation for, the fate of the postwar US city.

Publicity andpromotional material

This is a source produced for externalconsumption, but not primarily for researchpurposes. Commercial organizations, andincreasingly their non-commercial counter-parts,will market themselves and their productsvia a host of promotional and publicity devices.These can again be a good source of informa-tion for the geographer. Goss (1999), forinstance, shows how the marketing and sellingof products in the contemporary shoppingmall are based on the consumption of particu-lar spatial and temporal archetypes, eachdepicting a particular geography. In this case,examining publicity and promotional devicesallows us to see how the retail environmentpartly operates through the specific consump-tion of an idealized geography – based aroundarchetypes such as the marketplace, the festivalsetting, nature, heritage, childhood and primi-tiveness. On a larger scale, Molotch (1996), in achapter entitled ‘L.A. as design product’, hasexplored how the ‘local aesthetics’ of art, designand advertisement have contributed to thechanging economy and urban form of Los

Angeles. Indeed, he claims these aesthetics haveinfluenced modes of expression and economicproduction on a global scale.

Personal documents

These are usually not intended for researchpurposes, although politicians and mediapersonalities are now producing personaldocuments with one eye on the Sundaynewspapers and the other on their bankaccounts. In the main, though, by personaldocuments we mean private material such asletters, diaries, household accounts, addressbooks and personal memos – although therange can be extended to cover oral record-ings of life histories, photographs and familyportraits, and personal possessions andbelongings.These are a potentially fascinatingsource of material on community change,gender relations, consumption habits, employ-ment trajectories, local politics, generationalchange and so on.This type of material can beespecially helpful in historical research whereit can be used to piece together a portrait of atime and a place which have long passed.Blunt (1994a) and Bell (1993), for instance, usethe letters and diaries of Victorian womentravellers to investigate issues of gender andexploration, and gender and health, respec-tively, in colonial Africa.

Maps

These have provided a key source of geo-graphical material which, in the past, hasgone largely uninterrogated. But like otherfactual sources, maps help to construct, aswell as represent, our view of the world andas such they are useful to the researcher in avariety of ways.They are singled out here asa particular type of factual document preciselybecause they have been so privileged as a


67

3151-Ch-03.qxd 3/3/2004 3:34 PM Page 67

source of information within the discipline ofgeography. Geographers have tended to con-struct and use maps rather uncritically – theyhave been seen as relatively straightforwardrepresentations of landscape and of othergeographical ‘facts’. Recently, however,authors such as Harley (1988) and Wood(1993) have explored how maps act as partic-ular representations of the world, servingsome interests over others.Thus they can beread as statements about the world ratherthan as simple reflections of it. In this sensethey are the same as other ‘factual’ sources,and to appreciate fully their use as a researchsource requires an understanding of the issuesbehind their construction. We will move onto explore this more fully later in thischapter, where the use of maps provides oneof our set of case studies. First, though, beforeexamining different types of non-officialsources, we will chart some of the moregeneric issues in their use which are commonto all types of factual data.

Critical issues in the use ofnon-official data sources

Scott (1990: 19–35) sets out a number ofconcerns which are common to the use ofall types of documents. He groups themunder the four areas of authenticity, credi-bility, representativeness and meaning. Wewill keep to these headings as they havebecome established in accounts of docu-mentary data. It should be noted, however,that Scott’s work is mainly centred onwritten textual sources, but his frameworkcan be drawn on to encompass other typesof data.

The question of a document’s authenticityis concerned with the two related issues ofsoundness and authorship. The former refersto whether the document in question is

complete or reliable in terms of its content ifit is a copy or damaged in any way.The latterrefers to reliability of authorship.With regardto soundness, a source may have words, sectionsor pages missing, or have contents which havebeen distorted through copying, whethermanually or mechanically. Visual text may beunfocused or unclear in other ways. Audiotext may be difficult to hear clearly or mayhave been edited since the original produc-tion. In each instance the researcher shouldattempt to establish the extent of the problemand decide whether it is sufficient to invali-date the source’s use. It is not enough merelyto assume that everything is in its place andthat the source is sound and complete.

What is also common is that a range ofrecords covering a variety of documents willbe incomplete in some way. A particular yearmay be missing from a historical run ofrecords or the contribution of one person toa meeting or discussion may be missing ordistorted. Documents are often lost or mis-placed when institutions and agencies close,move or amalgamate. Some are kept whileothers are not, and the source will be incom-plete in some way or another.A set of diariesmay have a month missing or a year.A collec-tion of letters may have those from one par-ticular correspondent destroyed. Memos andminutes will often have sections deliberatelyblanked out. Again it will be up to theresearcher to decide whether these missingpieces are vital or necessary to the research inquestion, and thus whether the source issound enough to be used for his or herparticular purpose.

In many instances the veracity of theauthorship of a source will not be an issue. Insome instances, though, geographers will needto establish the authorship of the documentor record. Sometimes, those signed in thename of one person have been compiled andwritten by another. Letters or papers which


68

3151-Ch-03.qxd 3/3/2004 3:34 PM Page 68

appear in the name of politicians, anorganization’s director or a company managermay well have been authored by one or moreof those responsible for handling their corres-pondence. Again, this need not necessarilymatter and may produce interesting researchquestions in its own right, but the full inter-pretation of any text cannot be separated fromissues of its production. If the researcher isaware of the organizational structure or politi-cal circumstances in which the document waswritten, this will inform us not only about itsproduction but also about its content andmeaning.

The notion of credibility involves assessinghow distorted the contents of a document arelikely to be. As Scott points out:

All accounts of social events are of course‘distorted’, as there is always an element ofselective accentuation in the attempt todescribe social reality. The question of cred-ibility concerns the extent to which anobserver is sincere in the choice of a point ofview and in the attempt to record an accu-rate account from that chosen standpoint.(1990: 22)

Accepting that there will be an inevitablebias in any document, then, we are left toexamine the issues of sincerity and accuracy.The former refers to how much the authorof the document believes what he or sherecorded; the latter to how accurate theaccount is likely to be. There are many rea-sons as to why we may doubt the sincerity ofa source. One way of assessing this is to askwhat interest, material or otherwise, theauthor has in its content. Politicians will havean obvious interest in putting a particulargloss on to a document – whether for self-promotion and aggrandizement or for pur-poses of party propaganda. In an era wheresenior civil servants admit to being ‘econom-ical with the truth’ when giving written evi-dence to a court, the ethical standards of

those producing political sources makeinsincerity more rather than less likely. Thesame can be said of the financial induce-ments offered by much of our press for‘scooped’ news stories and, in any event, thecirculation battles between titles willinevitably affect the ways in which news isreported. There are now well-documentedcases of ‘factual’ news and documentarystories, in newspapers and on television,simply being invented. Business records mayalso be written deliberately to deceivewhere, for instance, tax avoidance is an aim.

Where the author acts with the best ofsincerity, the accuracy of the source can stillbe questioned. Although the accuracy of any‘factual’ source can never be precisely judged,we can examine those circumstances whichmight have led to inaccuracies. Most notablythis involves assessing the conditions underwhich the source was produced and, in par-ticular, the proximity of the compiler to theevents described.Was the account first hand,whether made from contemporaneous notesor those made later, or was it produced fromsecondary sources, or using memory alone? Ifthe latter, how long after the event was theaccount produced? An interesting exchangeof letters took place in the Independent news-paper in August 1993 concerning the accu-racy or otherwise of the political diaries ofTony Benn, a Labour Cabinet ministerduring the 1960s and 1970s. When theiraccuracy was questioned, Benn’s biographerwrote of times spent in Benn’s archive withthe diary editor, who constantly checkedback to contemporary documents, includingthe minutes and agendas on which Benntook his notes. He also stated that commentsmade when the hardback version was initiallypublished were collated and checked and, ifnecessary, the paperback text was altered(Independent, 31 August 1993: 25). This mayadd credibility to the paperback version but


69

3151-Ch-03.qxd 3/3/2004 3:34 PM Page 69

raises interesting questions of researchmethodology for the contemporary politicalgeographer. The same event would in thiscase be described differently (by the sameauthor) depending on which edition he orshe consulted.

There will also be an inevitable samplingbias in many of the sources we consult,whatever their sincerity and accuracy. Thebias will be towards the literate middleclasses or those with access to documentaryproduction facilities. Only recently, forinstance, through history workshop schemes,have working-class people started to pro-duce and document their own versions oftheir own histories on any kind of scale.Rose (1990) has used some of these oralhistories in her account of community pol-itics in Poplar, and deliberately draws a con-trast with accounts given in the ‘official’histories of working-class politics, writtenmainly by male middle-class politicalactivists to suit party political purposes. Inthe same manner as material from tele-phone polls will never represent thosewithout a telephone, or that produced by ahouse-to-house survey based on the elec-toral register will never reflect the views ofthe homeless, each ‘factual’ source used bythe researcher will be biased towards thosewho have the means to produce it. Thisbrings us on to the issue of representative-ness more generally. We may be forced touse certain data because these are quiteliterally the only ones available to us,whether for reasons of selective survival,denial of access, lack of knowledge or pres-sures of time and resources. The fact thatthe source is perhaps not representative ofits type is again not necessarily a problem,as long as the researcher is aware of this andadjusts the claims that can be made, andinferences that can be drawn, accordingly.

The fourth issue identified by Scott in theuse of factual sources concerns their mean-ing. In some cases this will be quite literal –our understanding may be prevented by theuse of technical terminology, acronyms orshorthand notes of one kind or another, aswell as by the use of an unfamiliar language.In other instances we will be confrontedwith the problem of interpretative under-standing. The essential idea here is that thetext of any document will hold differentlayers of meaning, which go beyond thewords themselves. For now we will simplystate that an understanding of the practices,styles and definitions employed by thoseproducing the source will help us understandits meaning. (We return to explore the dif-ferent ways of addressing these interpretativeunderstandings in Chapter 10.) It is impor-tant that we investigate these differentunderstandings because very few sources areactually intended solely for research. Mostsources are generated – whether by a com-pany, a voluntary group or an individual –for reasons wholly unconnected with theresearch needs of the geographer. It is there-fore important to acknowledge the condi-tions under which the sources wereproduced and the purposes for which theywere intended.As Hammersley and Atkinsonput it (1983: 137):

rather than being viewed as a (more or lessbiased) source of data…documents…shouldbe treated as social products: they must beexamined, not simply used as a resource. Totreat them as a resource and not a topic isto…treat as a reflection or document of theworld phenomena that are actually pro-duced by it.

We should therefore seek to establish the pur-pose behind the production of the sourcerather than just accept the record as somehow‘given’. Factual sources are not impartial and


70

3151-Ch-03.qxd 3/3/2004 3:34 PM Page 70

CASE STUDY 3.1: THE WRITTEN TEXTS OF NEWS MEDIA

Newspaper articles and stories form a major part of our lives, alongside audio and visualnews media. They tell us what is going on in the world; and they tell us about importantchanges taking place in different cities, regions and nations; and they explain to us whysuch changes are occurring and their likely implications. In short, our geographicalimaginations of places, people-groups and individuals are influenced significantly by ‘thenews’, which can form a very important source of information on which to base ourgeographical understanding of the world around us.

The news is often presented as a neutral account in which facts are collected, analysedand reported in an objective fashion, without bias, and in unambiguous and undistortedlanguage. Take, for example, the following account by Andrew Neil (the then editor of theSunday Times) of his newspaper’s stance on the 1984–5 miners’ strike:

From the start, the Sunday Times took a firm editorial line: for the sake of liberaldemocracy, economic recovery and the rolling back of Union power, and for the sakeof the sensible voices in the Labour Party and the TUC, Scargill and his forces had tobe defeated, and would be. It was a position from which we never waveredthroughout this long, brutal dispute. Our views, however, were kept to where theybelong in a quality newspaper: the editorial column. For us the miners strike wasabove all a massive reporting and analysing task to give our readers an impartial andwell-informed picture of what was really happening. (cited in Fowler, 1991: 1–2)

The desired code for interpreting text in newspapers, then, is to search the editorialcolumn for views and to find unbiased, factual news in the remainder. However, in orderto make critical use of newspapers as sources of information, it is necessary todeconstruct this code, realizing that language in the news is used to suggest ideas andbeliefs as well as ‘facts’. In order so to do, it is necessary further to recognize how thenews is socially constructed (see Thompson, 1997).

In short, all news is reported from some particular angle (see Hall et al., 1980;Franklin and Murphy, 1990). This case study deals with the language of news asexpressed in written text, but the principles of socially constructed news also apply tovisual and verbal presentations. The structure of the written news medium encodesparticular significances which stem from the positions within society of the publishingorganizations concerned. We therefore have to understand both how and why the news ismade. The ‘how’ involves a selection of what is newsworthy, and thus a series ofjudgements in which a partial view of the world is expressed. Different newspapers willhave different partialities – in Britain, the Mirror is more likely than The Times to report ona Lottery win, but The Times is more likely than the Mirror to cover famine in Rwanda.


71

autonomous accounts of particular eventsand processes.We have already shown that thisis the case with regard to state documents.

We now turn to examine some of theconditions of production of non-official datasources.

3151-Ch-03.qxd 3/3/2004 3:34 PM Page 71

CASE STUDY 3.1 (Continued)

Moreover, the selection of news is likely to be accompanied by a transformation ofmaterial, such that the same story will be given differential treatment by differentnewspapers. Particular transformations can fuel specific geographical myths. For example,Burgess (1985) shows how tabloid coverage of the urban riots in England in 1981 helpedto create a ‘myth of the inner city’, constructed through very selective representations ofthe physical environment, along with sometimes extreme commentaries on the nature ofblack and white working-class cultures. The myth was, in Burgess’s view, createddeliberately for ideological reasons in order to distance the rioters (and the reasons whythey were rioting) from ‘normal’ society. Hence the events of the riots, and the social andpolitical challenges which they represented, could be packaged up in a mythological innercity, which was conveyed as discrete, hostile and alien so as not to allow them to disturbexisting political, economic and social orders.

The danger for human geographers seeking to interpret such partialities is that bias willbe recognized in all but our own favourite outlet. It is important, therefore, to recognizewhy the news is made in a partial manner. Here, the goals of maximizing readership andadvertising, and therefore revenue and profit, are obvious considerations. In addition,however, there are deeper-seated goals relating to the strategic support given togovernments or political parties which are most likely to favour the other commercialinterests of newspaper magnates such as Murdoch, Packer and Thompson. In thatnewspaper publication is an industry and a business, with a particular position in thenation’s and world’s economic and social affairs, it is to be anticipated that the output ofnewspapers will be at least partially determined by that position. This is particularlyimportant given the concentration of newspaper ownership in recent years. In 1965 therewere 11 companies owning 19 newspapers in Britain. Some 30 years later there were7 companies owning 21 titles.

In some ways these partialities are straightforward. For example, the press isoften seen to be preoccupied with money, fame and royalty, each of which servesto symbolize hierarchy and privilege (Fowler, 1991). To interpret the world in sucha way as to naturalize hierarchy and privilege will serve the interests of capitalistconcerns such as newspaper empires. Other partialities are perhaps less obvious,although equally important. Box 3.1 presents some conclusions from an interestingstudy by Reah (1998) about how newspapers interpret social groups. From theseexamples it is clear that the language of the news is used not only to stereotype particularsocial groups – as being other to male, white heterosexual orthodoxy – but also todiscredit, belittle and misrepresent them for the purposes of reinforcing dominant culturalnorms and the power vested therein. Thus geographies of gender, race and sexuality, forexample, will of necessity be most interested in the ways in which news media presentparticularly partial views in these respects (Figure 3.1 shows a montage of the originalarticles).


72

3151-Ch-03.qxd 3/3/2004 3:34 PM Page 72


73

Women

A random selection of stories suggests that certain stereotypes are operating inthe press in relation to women. There is a tendency to depict them as existing pri-marily in relation to their families – their children, their husbands or partnersrather than as individuals in their own right. Younger women are frequentlydescribed in relation to their physical appearance. Women are often depicted asweaker – they are victims, they are on the receiving end of action rather than theperformers of it. If this selection is representative, then the newspaper-readingpublic receives a series of images depicting women according to a very limitingstereotype that values them in only a narrow set of roles. (1998: 69)

Ethnic groups

In this example Aboriginals and white Australians:

The Aboriginals are described as being ‘treacherous and brutal’. The qualities‘gentle’ and ‘trusting’ are mentioned as being falsely attributed to them byothers. The descriptions contained in the naming compare the Aboriginals to‘children’, and ‘Ancient Brits who painted themselves blue’. These descriptionsportray either the concept of the savage – a sub-human species; or the noblesavage – a sub-human species with a child-like innocence. Both of these repre-sentations are caricatures that are rooted in early colonialism.

The white Australians, on the other hand, are ‘tough’ and ‘bloody-minded’. Theyare named as ‘individuals’, a description that contrasts strongly with the por-trayal of the Aboriginals.

Therefore, closely linked to naming are the attributes and qualities that groupsor individual members of groups are credited (or discredited) with when they arediscussed in newspaper texts. Newspapers are artefacts of the dominant culturalnorm…The attitudes of the dominant culture tend to be reflected in the lan-guage of news stories, particularly those about minority power groups. (1998: 62)

Gay homosexuality

This concept of homosexual behaviour as abnormal is carried through the storyvia the word choice. A love poem is ‘slushy’. They are said to have ‘committed’,a verb normally associated with criminal activity, ‘a lewd sex act’ together. Thisword choice carries strong connotations of illegal and abnormal behaviour.There is also an element of belittlement in the use of idioms such as ‘caughtwith their trousers down’, ‘caught with his pants down’, that have connotationsof a stage farce. These, combined with words such as ‘slushy’ and ‘toyboy’,refuse to recognise the possibility of serious emotional involvement in such arelationship. (1998: 73)

Box 3.1 Reah’s (1998) conclusions on hownewspapers portray particular social groups

3151-Ch-03.qxd 3/3/2004 3:34 PM Page 73


74

THE fierce battle for cir-culation between mid-

market tabloid newspaperstook a fresh twist yesterdaywhen police confirmed theyhad arrested the wife of asenior executive at ExpressNewspapers on suspicionof theft.

Anita Monk, married tothe Express’s deputy editorIan Monk, was detained atan hotel at Heathrow airportin possession of a pre-publication copy of a damn-ing biography of theDuchess of York.

A second copy of thebook, Fergie: Her SecretLife, which the rival DailyMail was serialising afterpaying a reported £170,000to the publisher, was laterfound at her home.

The duchess tried tohave the book banned buthad to give up after shewas ordered to lodge£500,000 before her legalaction could go ahead.When she failed, a fiercebattle broke out over itbetween, the Express andthe Daily Mail. The Mail ranextracts of book, whichincluded claims that the

duchess would have killedherself but for her daugh-ters, as ‘the book she triedto ban’.

But on the first day ofthe Mail’s serialisation,November 2, the Expressalso ran a story with manyof the same revelations,which, the Express told itsreaders, had been pub-lished by a US magazine.

The previous day MrsMonk had been arrested atHeathrow.

A police spokesmansaid: ‘A woman aged 52was arrested on suspicionof theft and handling stolengoods.

‘Officers seized from thewoman a copy of a book yetto be published and anothercopy was seized at herhome. It was Fergie: HerSecret Story. The womanwas taken to Uxbridgepolice station and bailed toreturn on 18 Novemberpending further inquiries.’

A spokeswoman forExpress Newspapers said:‘Ian said he was not avail-able to comment on this.’

Guardian

Express chief’s wife arrested onsuspicion of stealing Fergie book

A STUNNING pal of comedy star JohnCleese miraculously escaped death yes-terday when her plane careered throughan airfield fence at 100mph and wasrammed by a van.

The £1million Lear jet carrying 27-year-oldLisa Hogan broke in two as it shot across apacked main road.

Terrified drivers braked and swerved, fear-ing a fireball.

But as a waterfall of fuel poured from theplane’s shattered wing tanks, 6ft 2in Lisacooly climbed from a smashed window.

A stunned witness to the crash on the A40 atNortholt, West London, said: ‘We found her sit-ting on a wall. All she wanted was a cigarette.’

The Irish-born beauty, who stars withCleese in his new film Fierce Creatures, wastreated in hospital for a cut leg.

Sun

GLENDA TO FIGHTON THE BEACHESLABOUR will be fighting for votes onthe beaches of Spain tomorrow.

Shadow Transport Minister GlendaJackson will hand out sticks of rockto holidaymakers in Benidorm andAlicante.

Mirror

I lay wideawake asdocs cutme openMum June’s op hell

A MUM who had a heart attack during major surgeryclaimed yesterday that she was WIDE AWAKE whendoctors operated.

June Blacker said she suffered agonising pain fromthe scalpel because the anaesthetic hadn’t worked.

Now June, whose heart stopped for THREEMINUTES is suing health chiefs over her horror ordeal.‘It was every patient’s worst nightmare’, she said.

The 43-year-old catering worker, who was in hospi-tal to be sterilised, couldn’t raise the alarm becauseshe had been given muscle-relaxing drugs

Mirror

Figure 3.1 Newspaper articles and the geography of gender, race and sexuality

Source: Reah (1998)

3151-Ch-03.qxd 3/3/2004 3:34 PM Page 74


75

Women wantedfor buildingsite workAnd it’s not justso we can whistleat ’em, sayBRITAIN’S most sexistindustry has been told torecruit women – or risk amanpower crisis.

Construction chiefs arebeing urged to shed theirmacho image to make roomfor female brickies.

And wolf-whistling at prettygirls could be a thing of thepast among tomorrow’sfemale scaffolders.

Many firms already report ashortage of skilled electri-cians and joiners. Now thereare fears that a drop in thebirth rate 20 years ago meansthat were will not be enoughyoung men to fill the jobs.

Industry bosses will be toldto make construction moreattractive to women by pro-viding all-girl trainingcourses, flexi-time andproper childcare facilities.

The move could lead tocreches on building sites oncedominated by hairy, beer-swilling men with tattoos.Sexy pin-ups would almostcertainly be banned to makewomen feel more ‘comfort-able’ in the workplace.

And women’s supportgroups to rival the old boy’snetwork could be introduced.

The shake-up will be for-mally recommended by theGovernment -sponsoredConstruction Industry Boardnext month.

A CIB survey has foundthat schoolchildren think theindustry is ‘dirty, dangerousand demoralised’.

A CIB source warned: ‘Ifthe industry fails to respondto the report’s recommenda-tions to recruit more womenit will face increasing skillshortages and problems withrecruitment’.

‘Demographic changesand the increasing successof women in both educationand employment will meanthat employers will face ashrinking number of applica-tions from young men’.

‘We are not calling forwomen to be hired to bepolitically correct’.

‘It is a question of produc-tivity and necessity in anindustry increasingly hit byskill shortages’.

‘The industry shouldrealise it is ignoring half thecountry’s workforce by notrecruiting women. But whyshouldn’t women work asbricklayers or electricians?’

The construction industryhas an unenviable reputationfor sexism. CIB figures showit has the lowest proportion ofwomen employees of anymainline British industry.

Women make up 46 percent of Britain’s workingpopulation but only 10 percent of the constructionindustry – and most of themare secretaries or clerks.

Not only are fewer womenrecruited but their drop-outrate is higher than the men’s– possibly because of themacho lifestyle.

The battered huts of build-ing sites are notorious for theirexplicit pin-ups and the crudejokes of their occupants.

The CIB wants ‘mentoringand support’ schemes forwomen to stop them feelingisolated.

Daily Express

Figure 3.1 Continued

3151-Ch-03.qxd 3/3/2004 3:34 PM Page 75


76

Salute fromthe PomsDOWN UNDER they are having the

biggest and longest birthday party in

history.

It is exactly 200 years since the firstBritish fleet, with its mixed cargo of sheepstealers, pickpockets and naughty ladies,dropped anchor in Botany Bay.

However, there are skeletons at thefeast.

‘What is there to celebrate?’ demand athousand placards.

The Aussies are being asked to tear outtheir hearts over the plight of the poor oldAbos.

They are asked to believe that, before thewhite man stole their land, Australia was aparadise inhabited by gentle, trusting, child-ren of nature living on the fat of the land.

In fact, the Aboriginals were treacherousand brutal.

They had acquired none of the skills orthe arts of civilisation.

They were nomads who in 40,000 years

left no permanent settlements.The history of mankind is made up of

migrations. Australia no more belongs tothem than England does to the AncientBrits who painted themselves blue.

Left alone, the Abos would have wipedthemselves out.

Certainly, they have suffered fromthe crimes (and the diseases) of theEuropeans.

TamedThat is inevitable when there is a collision

between peoples at different stages ofdevelopment.

The Aussies tamed a continent.They have built a race of tough, bloody-

minded individualists.We have had out differences – over poli-

tics and more serious things like cricket –but when the chips were down they havealways proved true friends.

Not bad at all for a bunch of ex-cons!

We hope that they enjoy their long, long

shinding and crack a few tubes for the

Poms.

Sun


THE SUN SAYS

3151-Ch-03.qxd 3/3/2004 3:34 PM Page 76


77

PERVYJERRY’SPOEM OFLOVE TOGAY BOY‘I’ll give you my best’

TORY MP Jerry Hayes sent hisgay toyboy a Valentine’s cardcontaining a slushy poem whichpledged: ‘I promise to love you ingood times and in bad.’

Hayes, 43, signed the card hegave to teenager Paul Stone with asingle kiss.

The poem, by Dorothy R Colgan,is called My Promise Of Love ToYou and contains the line: ‘Ipromise to give you the best ofmyself.’

It adds: ‘I fell in love with you forthe qualities, abilities and outlook onlife that you have.’

Hayes sent the card to Stone in1992. It was in an envelopefranked in the House of Commons.The lad was then 18 – at a timewhen the legal age of consent for

homosexuals was still 21.Their 16-month affair took place

when the MP, who is married withtwo children, was on an IRA hitlist.

Security chiefs found the listduring a raid in Ulster.

ThreatHayes’ name appeared together

with other prominent Tories andsenior members of the ArmedForces.

The fling with former YoungConservative Paul could have leftHayes vulnerable to blackmailattempts by the Provos. The MPwas targeted by the IRA after hebecame a Parliamentary aide toNorthern Ireland minister Robert

Eight police marks-men guarded his familyhome in Wenden’sAmbo, Essex, roundthe clock for threemonths.

His garage doorswere bomb-proofed.And behind themarmed cops secretlyset up a control roomwatching all approa-ches to the property.

An ex-cop who wasresponsible for theMP’s security saidyesterday: ‘We tookthe threat very seri-ously indeed.’

Paul told The Sun’ssister paper the Newsof the World howHayes got him a Houseof Commons SecurityPass by claiming hewas his unpaidresearcher.

He said the MP oftenchatted to him aboutUlster. Paul added:‘When I joked to himthat I could blow up theHouse of Commons hejust ignored me.’

Paul, now 24, saidhe first met Hayes at agay rights group meet-ing.

Hours later they com-mitted a lewd sex act ina park, it was claimed.

Paul, who wasstudying for his A-levelsat the time, said Hayesalso wined and dinedhim at expensiverestaurants.

Magic

Hayes, vice chair-man of the CommonsAIDS committee, alsosent Paul intimate let-ters on Commonsnotepaper.

One read: ‘I loveyou. I miss the magicof your hugs andwatching you whenyou are asleep.’

Another declared:‘Just remember, 1 Welove each other VERYmuch. 2 We miss eachother VERY much. 3We have fun togetherand can help eachother.’

Paul had access tovirtually every area ofWestminster.

He said: ‘Jerry likedhaving me around forwhen the mood tookhim, which was quitein lot. I was beingintroduced to peoplelike Michael Portilloand VirginiaBottomley.’

He said Hayesalways professed tobe devoted to his wifeAlison, 40, and theirchildren aged sevenand nine.

Hayes’ politicalcareer has been col-oured by a string ofgame-for-a-laugh stuntson TV.

He has dressed upas a bumble bee, sky-dived in a chickencostume – and waseven whipped in kinkybondage gear on alate-night show


3151-Ch-03.qxd 3/3/2004 3:34 PM Page 77


78

CASE STUDY 3.2: THE VISUAL TEXTS OF TOURIST ADVERTISING

Visual images are an integral part of the means by which we communicate incontemporary society. These images can be still or moving (see Case Study 3.4), but ineither form the visual portrayal of information represents a vital and culturally relevantsource of data for human geographers. In this case study we focus on still images, ofwhich there is a wide range – for example, those used to supplement words innewspapers or magazines; those used to advertise products on billboards or in leaflets;those used to portray places on postcards; and those used to convey the activities of aparticular business, government department or voluntary agency on the front cover of areport. In these and other cases, photographic and other forms of image seem to presentus with reality – a mirror image which shows the ‘truth’ about people and places. Afterall, we are told, the camera never lies. However, just as with written text, visual mediaimages are like a language, and human geographers need to be aware of how thatlanguage has been constructed, both technically and socially, before they can interpretthe meanings locked therein (Gamson et al., 1992). Visual texts then make a form ofreality rather than simply mirroring reality. They are carefully constructed and edited topresent a particular view of the subject they portray.

The technical construction of visual images consists of a number of formalcharacteristics and conventions (Wells, 1997). The different elements of the image areformally organized; like grammar and syntax in a sentence, elements are mademeaningful by their organization. Such organization will involve the relationship betweenwords and pictures, any narrative styling – for example, when a number of sub-imagesare to be read in a particular order so as to produce an unfolding story like a comic bookor a Peanuts cartoon – and the precise framing of any photographic visual text.Photographs will be composed so as to include some things and to exclude others, aprocess which inevitably results in partialities because of the choices involved. Within thecontent, some characteristics of places of the people in those places will be located in theforeground or the background. Horizons will be set, lengths of field will be selected andfocal points will emerge. Thus the ‘reality’ of places and people will be restricted to theinclusive, exclusive, focus/non-focus and foreground/background decisions of theimage-maker.


Human geographers have been taking a sharper interest in news media as a source ofgeographical information and imagination over recent years (see Burgess and Gold, 1986).Particularly effective studies include the investigation of the production and consumptionof environmental meanings in the media (Burgess, 1990; Burgess et al., 1991); theexamination of newspapers as sites of regional definition and regional promotion (Myers-Jones and Brooker-Gross, 1994); and the analysis of newspaper stories to highlightdifferent discourses on the Falklands/Malvinas conflicts (Dodds, 1993). Each of thesestudies demonstrates how the socially constructed text of newspapers can bedeconstructed in order to provide nuanced accounts of important geographical issues.

3151-Ch-03.qxd 3/3/2004 3:34 PM Page 78

In presenting the products they want to sell, the advertisers connect the tractor toimages of masculinity both in text and pictures. What they anticipate will help sellthe tractor is seen from the abundant use of words like horsepower, supremeenginepower, hydraulic power and pullpower, great capacity, position con-trol, reactions control, driving control. Keywords are power, precision, control.

The tractor is pictured large and strong. Big, heavy, greasy, dirty and noisymachines are symbols of a type of masculinity which often is communicatedthrough the male body, its muscles and strength, and what is ‘big, hard andpowerful’ (Lie, 1992). The analogy to male potency is quite clear, and visualizedby the phallus, the exhaust pipe standing up in front. In an illustration from oneof the ads, the machine is not pictured out in the field, as we first perceive it tobe. The tractor is placed upon a landscape consisting of hills, valleys, forests andmountains…It is placed on top of nature, ‘virgin land’, and photographed frombelow – an angle which makes anything else, including the potential buyer,seem/feel small in comparison unless, of course, the buyer fancies himself asbeing on top of the machine (and that is probably the intention). (1995: 126,emphasis in original)

Box 3.2 Brandth’s (1995) analysis of tractor advertisements


79


Many of these technical decisions are organized around recognizable culturalconventions of how to ‘read’ visual texts (see Panofsky, 1970). They make sense becausewe understand how they work. For example, the close-up view of a place or person drawsus in to study the emotional or place-related state of the subject in detail. By contrast thelandscape image offers a convention for viewing a wide-angle depiction of the scale andgrandeur of a wider area, an idea extended first by aerial photography and more lately bysatellite images.

The context of visual images is similarly constructed. Carefully manufactured visualswill often seek to represent multiple meanings, some of which will be at the surface andothers more hidden. The image-maker may employ myths or ideologies (Gold, 1994) soas to convey a scene from a particular viewpoint. Iconographic symbols representing awider whole are important here. Thus images of the ‘Wild West’ of America will conveyliteral meanings about the movement west and the conversion of wilderness into citiesand farmland, but they will also suggest symbolic meanings about the freedom andadventure inherent in the making of America. They are far less likely to convey theexploitative nature of such processes with regard to American Indians or the profitmotives involved.

Just as the written texts described in Case Study 3.1 were shown to reflect biasesagainst ‘other’ social groups, so visual texts can be equally discriminatory. An example ofthe ways in which visual images can convey gendered myths and ideologies can be foundin Brandth’s (1995) analysis of advertisements for tractors (Figure 3.2). She argues thatthese advertisements serve to inform male farmers of what positive characteristics ofmasculinity they should be striving for, and thus that they participate in the social shapingof masculinity by reinforcing it and reshaping it (see Box 3.2).

3151-Ch-03.qxd 3/3/2004 3:34 PM Page 79

Figure 3.2 John Deere tractor advertisement

Source: Brandth (1995: 129)


80

3151-Ch-03.qxd 3/3/2004 3:34 PM Page 80


81


The use of visual images as sources of geographical text and data is on the increase.Some of the best-known examples are found in accounts of visual texts in geographiesof tourist places. One of the problems inherent in this kind of work is the very subjectivityof analysis of images (see Selwyn, 1996). Put bluntly, what one geographer interpretsfrom an image may well be different from that of another. Nevertheless, in studies oftourism it is now recognized that the visuality of brochures has a significant role in thecultural practice of tourism, and so image subjectivity becomes a necessary part ofunderstanding the anticipatory expectations of tourists (see Cloke and Perkins, 1998).

A good example here is Goss’s (1993) study of the textural strategies employed inmagazine advertisements to see Hawaii as a tourist destination for mainland Americans.He identifies five aspects of the imagined geographies of Hawaii presented in theseimages, which differentiate Hawaii both from the US mainland and from other touristdestinations.

ParadiseThe advertisements consistently reference paradisal icons such as beaches, palms,waterfalls, tropical gardens and exotic flowers. They also valorize the unspoiled,unhurried way of life of the natives and a sense of their unrepressed sexuality. Hesuggests that some advertisements implicitly offer the good prospects of sexual liaisonswith natives (who ‘wore little more than a smile’ – Figure 3.3), and that others emphasizethe renewal of passion for tourist couples by depicting them in intimate or preintimatesettings (Figure 3.4).

MarginalityThe advertisements construct Hawaii as a place on the geographical, temporal and socialfrontier of North America, offering an exotic experience within the security of the familiar.The location of Hawaii is mystified and presented as halfway between East and West, andthe life of Hawaii is often represented nostalgically as a precapitalist village economy,where the past meets the present.

LiminalityVisual images from the advertisements suggest a liminoid tourism which provides abreak from the everyday life of the modern world. Tourists will discover not only theauthentic world of the native but they will also (re)discover their authentic self.

FeminizationGoss sees the advertisements as constructing the Hawaiian people and environment asfemale, and constructing tourism as an act of masculine possession. The visual texts aredominated by female characters (Figure 3.5), and he suggests that the viewers of the visualtexts are likely to be men, whereas the readers of the verbal texts are likely to be women.

3151-Ch-03.qxd 3/3/2004 3:34 PM Page 81

Figure 3.3 Hawaiian tourist advertisement, 1976

Source: Goss (1993: 677)

3151-Ch-03.qxd 3/3/2004 3:34 PM Page 82


83

AlohaThe word ‘aloha’ is used in most of the advertisements and seems to guarantee thehospitality of Hawaii (Figure 3.3). It defines the relationship between the tourist andworkers in the tourist industry and, in Goss’s view, systematically exploits andcommercializes the feelings and emotions of the Hawaiian people for the profit of thetourist industry.



3151-Ch-03.qxd 3/3/2004 3:34 PM Page 83


84



In this example Goss demonstrates how visual images are able to reconstruct thegeography of a place – Hawaii – replacing it by a series of signifiers which representdifference and symbolize the ‘other’. Similar analyses of visual texts in many other areasof human geography will add beneficial perspective to how we imagine the worlds inwhich we live and the people who live either alongside us or in far-off exotic places.

3151-Ch-03.qxd 3/3/2004 3:34 PM Page 84


85

CASE STUDY 3.3: THE VISUAL TEXTS OF MAPS

‘The map is not the terrain’, the skinny black man said.

‘Oh, yes it is’, Valerie said. With her right hand she tapped the map on the attachécase on her lap, while waving with her left at the hilly green unpopulated countrysidebucketing by: ‘This map is that terrain.’

‘It is a quote’, the skinny black man said, steering almost around a pothole. ‘Itmeans, there are differences between reality and the descriptions of reality.’

‘Nevertheless’, Valerie said, holding on amid bumps, ‘we should have turned leftback there.’

‘What your map does not show’, the skinny black man told her, ‘is that the floods inDecember washed away a part of the road. I see the floods didn’t affect your map.’(Donald Westlake, High Adventure, cited in Wood, 1993: 26–7)

The brochures discussed in Case Study 3.2 have a straightforward and usuallyrecognizable purpose. Although they will use all kinds of visual devices in support of theirtasks, tourist brochures are inescapably designed to promote or sell particular places astourism destinations. Maps, on the other hand, will often take on a more authoritativecharacteristic. In general, they are recognized, and used, as sources of information,something to be read for the purpose of seeing what is in a place, where things are orhow to get there. Maps have been central to the study of geography, even to the extent offorming half of such unfortunate defining epithets as ‘maps and chops’.

Yet, as Harley (1988) has shown, maps are very seldom recognized as sociallyconstructed forms of knowledge, which through the decision of map-makers withregard to ideological intent, the use of cartographic devices and the choices of what toinclude and exclude should not be read as straightforward ‘thin’ descriptions. Rather,maps should be scrutinized in terms of the conditions in which they are made andthe social constructions which they employ and embody. Only then do we begin torealize how appropriate a medium they are for manipulation by powerful elements insociety.

Harley offers three sets of insights on the ideological power of maps. First, he seesmaps as a kind of language (Mitchell, 1986), thereby opening up ideas of translating andinterpreting the ‘language’ of maps, as well as ideas of writing for particular readerships,the conditions of authorship, the potential for secrecy and censorship and so on.Secondly, maps employ iconology which imports deeper-level symbolic meanings as wellas superficial, literal meanings. It is at this iconological level which power is mosteffectively conveyed. Thirdly, cartography is a form of knowledge. As Harley (1989: 279)points out: ‘Whether a map is produced under the banner of cartographic science – asmost official maps have been – or whether it is an overt propaganda exercise, it cannotescape involvement in the processes by which power is deployed.’

Maps are almost always produced for a strategic purpose. For example, most officialmapping by the state is connected to defence and military purposes and is oftenconstructed from satellite imaging techniques that were specifically developed for themilitary for surveillance objectives. It is inevitable, then, that maps as a form of

3151-Ch-03.qxd 3/3/2004 3:34 PM Page 85


86


knowledge will carry with them at least some of the characteristics from the sociopoliticalarenas in which they have been constructed.

Human geographers continue to be fascinated by maps and analogous visualizations assources of both information and representation (Dorling and Fairbairn, 1997). However, anessential corollary of this fascination is to understand how maps in the past haveconstructed their own history at different scales. For example, at the scale of empire mapshave served as weapons of imperialism, with land often being claimed on paper before itwas occupied by armies. Surveyors served alongside soldiers, and maps served tonaturalize the exploitation of colonial civilizations (Driver, 1992). Maps also served tolegitimize the politics of empire, as Figure 3.6 illustrates in terms of the portrayal, bothgraphically and symbolically, of the British Empire in the late nineteenth century. Maps havealso served as weapons of nationalism and of national security. Thus the post-Soviet erahas involved, both metaphorically and literally, the rewriting of the map of Eastern Europe.Similarly the Bosnian and Albanian crises of the 1990s have focused on the perceivedmismatch between legitimate ethnic groupings and the national boundaries of states asconveyed by maps.

One interesting perspective on the national boundaries conveyed by maps is to befound in Bunge’s (1988) Nuclear War Atlas, in which he graphically illustrates the need tointerpret maps in terms of more than one-dimensional boundary lines:

If the United States rolled a nuclear missile across the border from Minnesota intoOntario, Canadian customs, among others, would have a sovereign fit. If thesame bomb were hoisted up 10 feet, Canada would be equally upset. If it were lifted100 miles, it would not even be noticed. How high (in feet) is Canadian Sovereignty?(p. 71)

The flat map of boundaries would suggest national sovereignty in this case, but itwould obscure the power of the US military to roam across ‘international’ skies in orderto carry out its purposes. Equally, the need for such power can also be illustrated bythe notion of how flat-map boundaries are transgressed. To quote from Bunge again:‘To state the problem in the exciting militaristic outlook of the 1990’s, “The Russiansare not coming. They are already here”. They are not “ninety miles off our shores” inCuba or sitting in missile silos in the Soviet Union. They are up in the sky’ (1988:83–5). Boundary transgression can be used to create fear, and therefore legitimizedefence, as well as to hide notional strategic power from the purview of traditionalmapping.

Alongside these scales of empire and nation, maps have also been used in morelocalized situations – for example, to convey the power of property rights. Thus in Britain,the much-loved Ordnance Survey maps of (among other areas) the great countrysideoutdoors are an essential waymarking guide to desirable footpaths and routewaysthrough scenic areas. However, they have embedded within them the notion of ‘right ofway’ through privately owned land, and therefore might also be considered as means oflegitimizing and upholding the power of private landowners over clearly demarcated

3151-Ch-03.qxd 3/3/2004 3:34 PM Page 86

Figure 3.6 World map of the British Empire, 1886

Source: Courtesy of the Mansell Collection

3151-Ch-03.qxd 3/3/2004 3:34 PM Page 87


88

private scapes. Leisure maps are also increasingly an agent of directing the ‘gaze’ of thetourist (Urry, 1990), pointing out ‘viewpoints’ and thereby organizing both thedestinations and the perspectives from which landscapes are seen or gazed upon bymany visitors (Wilson, 1992).

Local maps can also be used to convey alternative constructions of power.A ‘conventional’ street map of Los Angeles might convey key lines of communication,housing and business areas, civic buildings and so on. However, Figure 3.7 is MikeDavis’s map from his account of Los Angeles in City of Quartz (1990). His perspective isto map out the territories of power in south central Los Angeles’ sociologically distinctgang culture – a mapping which is usually unseen and indeed even hidden in officialaccounts of the city, but which represents a fundamental street-level knowledge forresidents, and researchers, of the area.

Figure 3.7 Gang territories, Los Angeles, 1972

Source: Davis (1990: 321)

3151-Ch-03.qxd 3/3/2004 3:34 PM Page 88


89


Maps, then, require very careful interpretation from human geographers in order tounderstand the ‘thick’ descriptions as well as the more obvious ‘thin’ descriptions beingconveyed. Some maps will carry deliberate distortions for political or strategic purposes.Others will be influenced much more subtly by the values of the map-producing agencyand society. Some distortions will, therefore, be subliminal (for example in the use ofgeometry). Others will occur because a particular map is ‘silent’ on particular charactersof what is being represented. Maps exert an influence through omission as well asthrough inclusion. Careful interpretation is particularly required because the socialhistory of maps is predominantly one of power not protest, and even in the current ageof more communication, map-making continues to be dominated by powerful groups insociety.

CASE STUDY 3.4: THE MOVING VISUAL TEXTS OF ‘NATURE’ DOCUMENTARIES

Moving visual texts are an essential facet of the media age. Morley and Robins (1995)talk of the existence of global media in contemporary society and suggest that, in theseterms at least, national frontiers are a thing of the past. On the one hand, the onset ofglobal media raises optimism for bringing the world closer together, with one-worldcitizenship responsibilities, interconnecting cultures and international democracy. On theother hand, global media are part of private giant economic interests pursuing traditionalcapitalist objectives of profit-making. The struggle for power and profit between suchconcerns of Time-Warner, Sony and News International is therefore being waged on thisglobal scale.

So it is that television programmes and films have now crossed many boundariesand have achieved world awareness of their plot, characters and merchandise.These imaginative moving visual images are dealt with in the next chapter, buthere we recognize that such images are also used for documentary purposes. Newsprogramming and documentary output from the global media companies arescreened worldwide, allowing us to experience the world through our televisionsets, and increasingly also our personal computers. These filmic texts are crucialto any contemporary theoretical concern for space (Cloke, 1997). Weare much more likely to know about spaces and places and the people whoinhabit them through moving visual text than either through direct experience orby reading written text. As Aitken (1994) suggests: ‘I’d rather watch the moviethan read the book.’ Indeed, it is not just that filmic texts represent the real worldto us, it is that such representations become our real world. It might even besuggested that the real world is increasingly being produced to matchthe representation.

Once again, we want to stress that documentary visual images are technically andsocially constructed rather than being mirror images of ‘real’ life. It follows that in order

3151-Ch-03.qxd 3/3/2004 3:34 PM Page 89


90


to understand the messages and meanings conveyed therein, we need to deconstructthe codes and devices being employed by the image-makers. Many of these devices aresimilar to those discussed in Case Study 3.2 in the context of still, visual images. Theformation and conventions of the image, its use of myth and ideology, its capacity toconvey discriminating impressions of particular social groups, its use of iconographyand signification, are all important in the understanding of filmic images. There are,however, other conventions which are specific to the interpretation of movingvisual images.

Filmic texts employ a number of techniques which audiences will have come toaccept, subconsciously, as normal (Aitken, 1994). For example, we have learnt tosuspend our disbelief and accept the impression of reality given by the rapidsuccession of static serial images in motion over time. We also find it unproblematicto consume filmic images which transpose unreal timescales, through the use offlashbacks, action replays and slow motion. Camera techniques of close-up,panning, tracking and zooming are also entirely acceptable signifiers of scale,intensity and interconnections between places and people. Documentaryand news techniques which pose a foregrounded reporter against a backgroundwhich represents an issue or place are viewed as an uncritical signification of theproximity of the production of the film to the topic itself. The background thussomehow signifies the newsworthiness of the place or issue. Documentary or newsinterviews present an ‘in-touchness’ with the other. These are just some examples ofthe filmic conventions which we take for granted in our viewing of filmicrepresentations of the ‘real world’.

Geographical research on film (Kennedy and Lukinbeal, 1997), however, also has tocontend with a further series of interpretative claims borrowed from film criticism (seehooks, 1997). First, filmic images often constitute voyeurism. Much voyeurism isassociated with the sexuality embedded within imaginative filmic texts. Yetdocumentary films can also be constituted around the voyeuristic impulse. News anddocumentary programming often deals with death – human massacres, famine andwarfare, as well as powerful animals (lions, sharks and so on) devouring their prey –and injury and illness. Attention is also often focused on problematic life experiencessuch as homelessness, drug addiction, squalor, hunger, orphaned children, mad cowsand so on. In these kinds of ways, filmic text organizes the gaze of the watcher as that ofmomentary tourist, gaping at the pain and distress of other people and animals.

Secondly, filmic images will often construct ‘otherness’ as the direction of news ordocumentary establishes through pictures on dialogue what is the mainstream self andwhat is the marginalized other. Filmic text is therefore used to exoticize ‘foreigners’ orindigenous populations, to regulate what the viewer will regard as ‘normal’ practice (andtherefore what is dysfunctional) and to regularize particular cultural or political biases.Even in news and documentary formats, hidden ideological biases can pervade theproduction, and therefore consumption, of filmic text.

Thirdly, filmic texts can distort particular spaces, places and the activities of peoplewithin them. There is often a sense of accepting the authenticity of places, people andevents in news and documentary coverage. However, the very presence of cameras and

3151-Ch-03.qxd 3/3/2004 3:34 PM Page 90

Conclusion

What each of these sources have in commonis that they purport to tell, or represent, the

truth as accurately as possible.Yet what theyalso have in common, as Harley says of maps(1988: 278), is that ‘in the selectivity oftheir content and in their signs and styles of


91


reporters may itself ignite a smouldering situation, with the actions and place meaningscaught on film only being ‘real’ because the film was there ready to catch them. Equally,there are many different stories and images which can be conveyed about a particularplace, and news and documentary coverage will often choose a dramatic portrayal, even ifdifferent voices are incorporated within the narrative.

There are still relatively few examples of research which brings human geographyinterests to bear on ‘factual’ filmic texts.1 One interesting example of this genre ofinterpretative research is Wilson’s (1992) discussion of Disney nature films, in which henotes a number of key conventions of filmic narrative. For example, the stories are alwaystold starting at the beginning, and ending at a new beginning. Thus in The Lion King(admittedly not a documentary!) the picture starts with the birth of Simba, and ends withthe birth of Simba’s first male cub. Here, according to Wilson, is a transparent allegory of‘progress’, symbolizing the forward-looking notion of modernity, with better standards ofliving in the next circle of life.

Wilson also captures the changes in the locational authenticity of Disney’s natureprogrammes. Early films were short on location and often in the West. However, bythe 1950s and 1960s the West had become tainted both environmentally and politicallyas the virgin landscapes of previous films were transformed by resource extractionindustries and the industrialization of timber production and water power. Theselandscapes thus became too laden with contradictions for the shooting of furthernature films, and so the Disney wildlife features were relocated, either into the BurbankStudio lots, using trained animals, or into more remote locations, often in deepestAfrica. Such relocations changed the focus of ‘wildlife’ and ‘wildlife locations’ in manyways. More recently, Disney’s wildlife programmes have switched to highlytechnological shots of the private lives of animals, insects and plants, using trickphotography, extremely intrusive camera techniques and ‘trained’ creatures in non-natural conditions. These changes have often been mirrored by other natureprogrammes.

Wilson’s analysis demonstrates how filmic interpretation of wildlife constituted theaddressing of nature as a social realm. Animals were inserted into human storiesand usually treated as social beings. In so doing Disney precursed what McKibben(1989) has termed ‘the end of nature’. Nature as filmed became a culturalconstruction on to which human stories were told and projected. It is this kind ofstory-telling through the filmic images of news and documentary programming whichprovides an exciting and very significant source of information for human geographyresearchers.

3151-Ch-03.qxd 3/3/2004 3:34 PM Page 91

representation, [they] are a way of conceiving,articulating, and structuring the human worldwhich is biased towards, promoted by andexerts influence upon particular sets of socialrelations’.An understanding of this selectivityand bias will enable the researcher to appreciatethe issues involved in the use of these sources.In the next chapter we will examine the con-struction and use of imaginative sources ofdata, which we would expect to promote par-tial and particular viewpoints. However, thenon-official sources we have considered here

also do this, but under the banner of faithfullyreporting the truth. As we maintained at theoutset of the chapter, imaginative and factualsources do have a lot in common. This issuewill be explored further in the next chapter.

Note

1 But see Burgess and Gold (1986), Burgess(1987), Rose (1994) and Benton (1995).


92

3151-Ch-03.qxd 3/3/2004 3:34 PM Page 92

Introduction

As we have shown in Chapters 2 and 3, inrecent years geographers have adopted anincreasingly critical approach to conventionalsources of data such as official statistics andother ‘factual’ documentary sources. Thesesources are no longer seen as providing directand unmediated access to geographical realitybut as social constructions produced for par-ticular purposes in specific social contexts. Atthe same time there has been a growingacceptance among geographers of the legiti-macy of using sources of information whichare overtly products of human imaginationand whose primary original purpose is not tomake factual statements about the world but,instead, to entertain, provoke, inspire or movethe reader, listener or viewer; in short, toengage the emotions and, indeed, the imagi-nation. Such ‘imaginative sources’ are thefocus of this chapter, but before consideringtheir construction in detail there are a coupleof ambiguities to address.

In some ways, the distinction between ‘fac-tual’ and ‘imaginative’ sources is more appar-ent than real. Geographers are not alone incoming to realize that all attempts to describe,explain and understand the world involveparticular ways of imagining it, and that thereis no way of presenting a purely factualaccount devoid of imaginative content.Nevertheless, because we want to emphasize

the ways in which sources are produced, wefind it helpful to deal with ‘imaginative’ and‘factual’ works separately because the conven-tions governing their construction are (at leastto some extent) distinct from one another.

Another ambiguity in this chapter con-cerns the concept of ‘data source’. Frequently,geographers are interested in imaginativeworks not as sources of data which will pro-vide information about something else but asresearch topics in their own right.For example,one geographer might read the novels ofCharles Dickens in order to learn about thegeography of nineteenth-century London,while another might use the same books in astudy of the rise of the novel as a literary formand its relationship to nineteenth-centuryindustrialization. In keeping with the struc-ture of the book, the main emphasis in thischapter will be on the use of imaginativeworks as sources for research, rather thanobjects of research, not least because in thelatter case the sources for the research maywell include the statistical, documentary andinterview materials discussed in otherchapters. However, since studying imaginativeworks as objects of research often involves aninvestigation of the conventions and condi-tions of their production, we cannot draw arigid distinction between data sources andresearch topics. Thus, in writing this chapterwe have often found it useful and necessary toconsider geographical studies which treat

Imaginative Sources4

3151-Ch-04.qxd 3/2/04 2:42 PM Page 93

imaginative works as research objects as wellas research data.

Geographers are turning to imaginativeworks with increasing frequency. There aremany reasons for this and, to some extent, theydepend on the particular nature of the workinvolved and the issues under investigation.At a more general level, however, we mayidentify five claims about imaginative workswhich are commonly used by geographers tojustify their use:

1. The production of imaginative works(‘artistic creation’) is a significant socialactivity, that has a geography to it andshould therefore be one of the activitiesgeographers study.

2. Artistic works embody imaginative andimaginary geographies. Geographersbring a sensitivity to issues of space, place,landscape and environment to their analy-ses and can therefore add an additionalperspective to the interpretations of arthistorians, literary critics, film theoristsand so on.

3. The products of creative people provide awindow on to the human condition andthereby illuminate issues of interest to geo-graphers, such as human feelings for place.

4. Imaginative works are themselves theproducts of particular geographical con-texts and thus help geographers to under-stand those contexts: Renaissance dramatells us something about ElizabethanEngland and Hollywood movies give usinsights into contemporary America.

5. The form and content of imaginativeworks are related to wider social and geo-graphical relations and processes; theyboth reflect and affect these wider socialrelations.

As will become clear, these claims are by nomeans unproblematic but they provide auseful starting point for thinking about the

relationships between imaginative sources andgeographical research.

Understanding theconstruction ofimaginative sources

Just as in the cases of official and non-official‘factual’ sources, it is particularly important tounderstand the conditions under which imag-inative works are produced when using themas research sources in human geography.These conditions are of two related sorts.First, there are the changing conventions andregimes in which creative people work andthrough which the meanings of their productsare conveyed to and understood by the reader,listener or viewer. These include such thingsas the technique of perspective drawing inWestern art (which was developed in theRenaissance and which governed the con-struction of most painting and drawing for500 years), the conventions of narrative con-struction in literature or the use of particularstyles of music in film soundtracks to conveyemotions or to support the action. Secondly,there are the social contexts within whichimaginative works are produced.These mightinclude the organization of the publishingindustry, the relations of patronage between apainter and a client, or the wider social rela-tions of class, gender and ethnic differencethat influence (unconsciously or otherwise)both the form and the content of imaginativeworks.

There are two main reasons why it isimportant to understand the processesthrough which imaginative sources were pro-duced when seeking to use them in geo-graphical research. First, we would argue thatit is important to ‘demystify’ the process ofartistic creation. Writers, artists, poets, archi-tects should not, we feel, be seen as specialbeings endowed with powers of insight and


94

3151-Ch-04.qxd 3/2/04 2:42 PM Page 94

revelation denied to others. This is not tosuggest that emotions and what is often called‘spirituality’ are unimportant in imaginativeworks. Rather it implies that emotionalresponse and spiritual feeling are not theunique preserve of specially creative peoplebut are available to all. By understanding artis-tic creation as a social process which takesplace in an economic, political and culturalcontext, we hope to show that while imagi-native works do possess certain distinctivequalities, these arise from particular culturalnorms, conventions and practices rather thanfrom some mysterious capacity of theirproducers.

Secondly, we would argue that imaginativeworks cannot be adequately interpreted with-out an understanding of their construction.For one thing, in some cases interpretation ofa source and analysis of its construction blurinto a single process. Thus the meaning of apainting frequently depends on the conven-tions and context of its production. However,even where interpretation can be separatedfrom an understanding of its construction, theconditions of production still make a differ-ence.Thus one can interpret a film like BladeRunner for its symbolic content without adetailed knowledge of the Hollywood filmindustry, but such a knowledge would expandthe range of questions one could ask.Why, forexample, are action films like Blade Runner themainstay of so many companies’ productionschedules? Why was the later ‘Director’s Cut’version of the film seen as ‘more authentic’than the initial release?

It is easy to stress the importance in principleof understanding the conditions of produc-tion of imaginative sources but rather moredifficult to describe how it should be donegiven the apparently huge differencesbetween forms like poetry, architecture andvirtual-reality simulations. One way would beto undertake an in-depth study of eachtype.This would involve studying in turn the

television industry, the media, the publishingindustry, the conventions adopted by painters,poets, architects and film-makers and so on.However, not only would each study requirea chapter, or even a whole book, to itself butsuch an approach also neglects what is com-mon to all imaginative productions. At firstsight is seems difficult to see what the com-mon threads might be, but it is possible tooutline an approach to the construction ofimaginative sources that can be applied to allor most of the forms we have discussed above.There are a number of ways of doing this (foran overview see du Gay, 1997, and Hall,1997). Our approach, set out below, drawsmainly on the work of Raymond Williams(1981). Williams’s writings have providedmuch inspiration for subsequent accounts ofthe production of culture and for culturalstudies more generally. Williams’s approachcame to be known as ‘cultural materialism’because of the emphasis he placed on the roleof material interests and material processes inthe production of culture. Although the ‘cul-tural turn’ in geography marked something ofa shift away from avowedly materialistapproaches towards a focus on textuality, dis-course and rhetoric, we want to emphasizethe continuing relevance of cultural material-ism for geographical research. For one thingwe consider the distinction between the dis-cursive and the material to be overdrawn. AsWilliams shows in his work, even somethingas apparently wholly ‘discursive’ as speaking orsinging entails all kinds of material transfor-mations involving the body (lungs, vocalchords, tongue and teeth) and the environ-ment (air and the acoustic properties of theperformance space). Intriguingly, these kindsof concerns have recently come into the heartof geographical theory with the growth ofinterest in the radically materialist actor-network theory of writers such as Bruno Latourand Michel Serres (Serres and Latour, 1995;Bingham and Thrift, 1999;Whatmore, 2002).

IMAGINATIVE SOURCES

95

3151-Ch-04.qxd 3/2/04 2:42 PM Page 95

Moreover, despite the passage of time sinceWilliams’s ideas about the production ofculture were first published in the 1960s and1970s, they continue to exert a substantialinfluence on cultural studies and culturalgeography (Milner, 2002).

The process of culturalproduction

Conventional cultural and literary theoryaccords a central place to the role of theauthor or creator as the prime or even onlysource of a work’s meaning. Some morerecent writers have proclaimed the ‘death ofthe author’ and have asserted that the mean-ings of texts or cultural representation escapeentirely from their authors’ control. In ourview authorial or creative intention has somerole in the production of culture but, and it isa big but, it is rarely, if ever, the most signifi-cant aspect of the production process. At thesame time the input and impact of an authoron his or her work are not limited to con-scious intention. Conscious intentions arestructured and influenced by unconsciousmotivations and desires and constrained bycultural conventions and material and discur-sive contexts.

All cultural and artistic works are necessarilysurrounded by material processes and prac-tices which influence, constrain and enablethe development of particular culturalphenomena in specific ways in different timesand places. For example the institutions of theartist’s patron in the Renaissance periodaffected the nature of the art produced, just asthe art market does today. The different meansof cultural production (from the human voiceand body to the television camera and thecompact disc) influence what is produced. Inaddition to these material conditions, non-material (or ‘discursive’) practices also play animportant part. For example, the ‘genre’ of thesoap opera or the detective novel helps to set

the conventions for new soap operas orthrillers, while variations in symbolic relationsare crucial in differentiating Greek tragedyfrom Renaissance drama from the realistnovel (Williams, 1981). All these aspects(material and discursive) of the productionprocess are also related in complex ways tothe wider society of which they are a part.Thus revolutions and changes in dominantartistic forms often coincide, and are inter-twined, with major transformations in social,political and economic relations.This is not tosay that the content of the cultural change ismechanically determined by social change,but neither is it wholly independent.

There are seven sets of relations that,together, explain what, we are calling the con-struction of imaginative sources of data(Williams, 1981): cultural institutions, culturalformations, the means of cultural production, cul-tural identifications, cultural forms, cultural repro-duction and cultural organization.

INSTITUTIONS

Virtually all artists work in some relationshipto social institutions of various kinds (institu-tions comprising artists themselves – forexample, artistic ‘movements’ such as the Pre-Raphaelite Brotherhood considered in thenext section on ‘formations’). These institu-tion vary historically and geographically.Thedominant institutions in medieval Englandwere very different from those in twentieth-century America. Nevertheless it is possible toboil down the enormous variety of institu-tions that have existed to four main types(Williams, 1981). These are the institutedartist, patronage, markets and post-marketinstitutions.

The instituted artist was particularly a fea-ture of early societies and involved the artist(usually a poet) taking a key part in the centralsocial organization.The archetype is probablythe Celtic bard. Despite apparent similaritieswith patronage systems, this situation is distinct


96

3151-Ch-04.qxd 3/2/04 2:42 PM Page 96

because patronage relations depend on adeliberate decision to recognize or patronize aparticular artist. In the bardic case, by contrast,‘the social position of this kind of social pro-ducer was instituted as such, and as an integralpart of general social organization’ (Williams,1981: 38).

Patronage in some cases evolved from thecase of the instituted artist. In Wales, forexample, there was a gradual transition inwhich the role of bard became increasinglydetached from the court and began to beattached to individual noble households, orsometimes to be peripatetic between house-holds, seeking support, shelter and commis-sions (Williams, 1981: 39). In many societies,however, there was no institution of ‘courtpoet’ and the more normal form was fullpatronage. For many centuries and in manycountries artists have been retained by indi-vidual or institutional patrons and have pro-duced their work to order on a commissionbasis. This means, of course, that the contentof these works has been determined notsolely, or even mainly, by the conscious inten-tions of the author but by the instructions andspecifications of the patron.A further distinctform of patronage involves protection, sup-port and recognition for the artist but littledirect commissioning. Here the artist mayhave considerable control, within limits, overcontent and form. Williams identifies thetheatrical companies of Elizabethan Englandas examples. Sponsorship represents anothertype of patronage, which sees the survival ofsome form of patronage relation into the ageof the market and the general commodifica-tion of cultural products. In eighteenth-century England this often took the form of asubscription list in which a number of spon-sors would support the work of an artist at anearly stage before it had found general successin the market. This form continues as com-mercial sponsorship of the arts where itseeks a commercial return through the use of

advertising or financial investment. Theseforms also influence the character of artworks. For example, it is frequently assertedby artists that large companies prefer to spon-sor ‘safe’, ‘mainstream’ and ‘conventional’work, as opposed to innovative and avant-garde material. Finally, patronage can alsocome from the state through the publicexpenditure of taxation. In these cases there iscontinuity with earlier forms but also thescope for extending and developing the arts asa matter of public policy.This can also lead tosome works being favoured over others forpolitical reasons. For example, in Britain inthe 1980s the Conservative government wasmotivated by a right-wing political agenda torestrict the public funding of the culturalactivities of gay and lesbian groups by limitingthe range of things on which local authoritiescould spend money.

In all relations of patronage, the patron(whether an individual or an organization)has considerable power, and it is this whichunites these very varied social forms. Thepatron can withhold or provide support atwill, and such support is withheld wheneverthe activities of the cultural producer areunacceptable to the patron. For the humangeographer, therefore, using artistic sources asresearch data requires that close attention bepaid to the nature of any relations of patron-age involved in the production of the work.

Wherever and whenever cultural productsare produced for sale and exchange, they arecommodities and the artist becomes a com-modity producer (Williams, 1981: 44). Thisinstitutional relationship is qualitatively differ-ent from that of patronage.The form and con-tent of the work are determined not by thewhims of the patron but, at least in part, alsoby the perception of actual or projectedmarket demand. However, this can take manyforms.Production may be artisanal. In this casethe artist is producing under his or her owncontrol and offering the work for direct sale

IMAGINATIVE SOURCES

97

3151-Ch-04.qxd 3/2/04 2:42 PM Page 97

on the immediate market. A contemporaryexample would be a fashion designer with hisor her own shop.There are also two forms ofpost-artisanal production. The first, in whichthe artist sells to a distributive intermediary, isakin to some modern commercial art galleriesin which the gallery owner buys from theartist.The second, in which the artist sells to aproductive intermediary, is akin to the modernbook publisher in which the publisher notonly buys the artistic work (such as a novel)but also produces (by printing and binding)the finished book. In both cases the artist’srelation with the market becomes indirect.

A third type of market relation is themarket professional.This phase sees the artistgain specific professional rights, such as thoseassociated with copyright and royalties, overthe work. This is associated with a new typeof intermediary, the literary agent. In thisphase the artist does not have direct relationswith individual buyers, or completely indirectrelations with the market through intermedi-aries, but has ‘relations with the market as awhole’ (Williams, 1981: 48).

The fourth and so far final phase of themarketization of cultural production is thecorporate phase, and this is particularly associ-ated with the development of new mediasuch as film and television. Here the creativeartist acts not as an individual author but aspart of a production process undertaken by anorganization, such as a magazine publisher, atelevision station or a record company. Thecultural product is the outcome of the activi-ties and interactions of dozens of individuals,contracted and employed by the corporation.The input of the ‘artist’ (pop star, film direc-tor, etc.) is now just one component in theproduction process which depends to a largeextent on the labour and creativity of others(technicians, designers, actors, musicians, edi-tors and producers).

All the ‘phases’ of marketization (artisanal,post-artisanal, professional and corporate) can

and do coexist in time and space, and marketrelations can and do coexist with other insti-tutional arrangements, such as patronage.

In interpreting such sources in geographicalresearch, the impact of different forms ofmarket relations on the nature of these worksmust be analysed and understood. The bottomline (literally) for market production is that itmust, by definition, make a profit, at least inthe medium term.This means that the manyartistic forms that are not ‘profitable’ couldnot exist in modern societies without otherforms of support.

These other forms of support are providedby a final group of institutions: ‘post-market’institutions.These in turn can be divided intothe modern patronal, the ‘intermediate’ andthe governmental (Williams, 1981).The first iscommon and involves foundations, subscribersand private individuals supporting worksfinancially in a continuation of the traditionalforms of patronage discussed above.The inter-mediate category accounts for organizationswhich are not private in the patronal sense butnot fully part of the state either. The ArtsCouncil and the BBC are examples of organi-zations which are publicly funded butwhich ‘direct their own production’ (Williams,1981: 55). Finally, governments produce artthemselves.This situation is most common instate socialist countries such as China. Here,artists may be direct employees of the stateand, to a greater or lesser extent, their outputmay be controlled by state officials.

FORMATIONS

The second group of relations comprises thegrouping of cultural producers themselves intoformations such as schools, movements andtendencies.The social institution of the Celticbards has already been mentioned, and thebards acted together to sustain and to policethe rules of their art. This was an early exampleof a formation. In the Middle Ages, the craftguilds were influential and in the English


98

3151-Ch-04.qxd 3/2/04 2:42 PM Page 98

mystery plays each guild took responsibilityfor a particular aspect of the production.TheItalian Renaissance gave birth to the academy,which marked an increasing distinction of artsfrom crafts and a growing concern with theeducation and training of artists. The acade-mies, in turn, gave rise to exhibitions.The riseof the market professional documented abovewas associated with the development of pro-fessional societies, particularly societies ofwriters, who came together to negotiate withpublishers and governments over issues ofcopyright and so on.

More diffuse and less institutionalized thanall these are the artistic ‘movements’, ‘schools’and ‘-isms’ which are so much a feature ofthe modern (post-eighteenth-century) culturallandscape.These can be grouped according toboth their external and internal relations.

Internally, artistic movements can be basedon formal membership, a collective publicmanifestation or on a conscious associationand group identification (Williams, 1981: 68).The first of these may well have a formalconstitution and forms of internal authorityand is exemplified by the German Order ofSt Luke. The second, exemplified by theEnglish Pre-Raphaelite Brotherhood, involvesan exhibition, a publishing imprint, a period-ical (e.g. the Pre-Raphaelites’ The Germ) or amanifesto. The third category, such as theFrench Nabis, has no formal structure orpublic manifestation but, none the less,involves group meetings and a shared identity.

External relations are also of three types.These are the specializing formations, thealternative formations and the oppositionalformations. Specializing formations work tosustain and promote work in a particularmedium or branch of the arts or in a particu-lar style. Alternative formations aim to providefacilities for the production and circulation ofcertain kinds of work for which there isno outlet elsewhere. Oppositional forma-tions (like the surrealists, the dadaists or the

situationists) raise this alternative activity toactive opposition to the existing institutionsand practices. Clearly, different forms of inter-nal relation can go together with differentexternal ones. Oppositional groups might behighly internally structured or fairly diffuse.

For human geographers using artistic datasources the question of formations is impor-tant because the character of particular works,and their meanings, are partly determined bytheir formational heritage. Moreover, as wasthe case for institutions, formations themselveshave a geography: they vary in their form andpurpose from place to place and from timeto time, and they are themselves internallyheterogeneous (Williams, 1981: 85–6).

MEANS OF PRODUCTION

The means of cultural production can bedivided into those that inhere in the humanbody and those that involve the use and trans-formation of non-human materials. The useof the inherent capabilities of the humanbody allows, principally, dance, song andspeech as forms of artistic expression. Thesemay be traditional forms or may involve longperiods of professional training. In either casethe physiology of the human body placeslimits on what can be danced, sung or spoken.It should not be supposed, however, that theseforms of production have been surpassed bytechnology; rather, both coexist today(Williams, 1981).

In the use of non-human means of produc-tion the social relations involved are muchmore complex. External objects may be com-bined with the human body, as in the use ofmasks and costume in dance. Instrumentsof performance may be developed, notablymusical instruments of various types.Then thereare the ‘selection, transformation and produc-tion of separable objects’ (Williams, 1981: 90)such as the use of clay or metal in sculptureand pigment in painting, and systems of signi-fication separate from human gesture or

IMAGINATIVE SOURCES

99

3151-Ch-04.qxd 3/2/04 2:42 PM Page 99

speech, notably writing. Finally, there is thedevelopment of ‘complex amplificatory, extend-ing and technical systems’ (Williams, 1981: 90)from printing, through photography andrecording to computer graphics and ‘virtualreality’.The first three of these five are essen-tially extensions of the resources inherent inthe human body, at least as far as the socialrelations involved are concerned. The fourthand fifth, however, involve qualitatively newsocial relations of production.

IDENTIFICATIONS

The further set of social relations concernsthe ways in which particular works and cul-tural activities are identified and categorizedsocially. Put more simply, what is art? ForWilliams there can be no eternal or universalanswer to this question. Defining and catego-rizing cultural products is itself a socialprocess, and conventional universal definitionsof ‘art’ fall down because they fail to recognizethe implications of this insight.

For example, it might be argued that ‘art’ isconsciously performed or exhibited for anaudience yet this may leave out works likecave paintings, commonly thought of as anearly form of art but situated in dark and inac-cessible areas which are hardly the first choicefor a site selected deliberately with a view toexhibition. Perhaps art is distinguished by thespecial skill with which it is executed. But awide variety of other activities involve highlevels of skill and are not usually regarded as‘art’. Finally,‘art’ might be defined as possessing‘aesthetic’ qualities such as beauty, harmonyand so on. However, this definition fails too, asmany things have aesthetic appeal withoutbeing counted as ‘art’.

All these problems suggest that the processof defining art and categorizing activities andobjects into ‘art’ and ‘not art’ is a social processand there are no universal definitions.Williams’saccount emphasizes changes in categoriza-tion through time, but he rather neglects an

issue of central interest to geographicalresearch – namely, the ways in which theprocess varies over space.

Williams argues that ‘art’ must be seen as asocial form and that society identifies certainactivities and works as art through systems ofsignals. Locating something in an art gallery,for example, identifies it as a work of art.Drama is conventionally signalled by locationin a theatre. On television, where imaginativeworks are joined by factual programmes, newsbroadcasts and so on, ‘art’ is signalled by con-scious titles and introductions which clarifyfor the viewer that the programme is dramarather than documentary, for example.

There are also systems of signals internal toimaginative works. Specific devices and tech-niques (such as the soliloquy in drama or thesonnet in poetry) provide cues to the audi-ence or reader through which to identify theactivity. These devices themselves can beunderstood as social practices and analysed interms of the social relations they embody.Thus in drama the use of the soliloquy (anindividual performer talking to him or her-self ) depends on (and perhaps contributes toproducing) a particular conception of ‘self-hood’ as, for example, self-conscious individ-uality, which means in turn that the devicewould have been simply unthinkable in socialcontexts where different conception of theself were found. In this sense it is significantthat the soliloquy was strongly developed inRenaissance drama in early modern Europe.

FORMS

At the start of the chapter we outlined differenttypes of imaginative works used by geogra-phers. There are, however, further differenceswithin each type. Epics are different fromdetective stories, religious icons are differentfrom landscape paintings and the silent come-dies of the early years of cinema are distinctfrom modern adventure movies.Williams callsthese different subcategories ‘forms’ and he


100

3151-Ch-04.qxd 3/2/04 2:42 PM Page 100

uses the example of drama to illustrate hisargument because of the long historicaldevelopment of drama as an activity and thevariety of forms through which it haspassed.

The earliest historical example of drama isclassical Greek tragedy. This form emergedfrom earlier predramatic forms such as thedithyramb or choric hymn.There was then aseries of innovations which together pro-duced the emergence of drama as a culturalphenomenon.These included dialogue betweena single figure and the chorus, then dialoguebetween two figures, and then the addition ofa third actor. Over time, more actors wereadded and the chorus gradually became moremarginal. At this stage drama was fully dis-tinctive and would have been recognized bymodern audiences as drama. Similar processesof development could be traced for othercultural activities – the emergence of formaldance from folk dance or of individual paint-ings, for example.

There were distinct limits to the Greekdramatic form, however, which mean thatwhile it was clearly drama it was differentfrom modern drama. Thus the modern con-vention of having different roles spoken bydifferent actors was not part of the Greekform, and it also had its own distinctive modesof production combining singing, speakingand recitative,which mark it out from modernforms defined by spoken dialogue.

Another example is English Renaissancedrama (the theatre of Shakespeare and hiscontemporaries) which, while still ‘drama’, isquite distinct from the Greek case:

Meanwhile, within the different social orderof Renaissance England, quite other formalinnovations were made. The ‘dramatic’, bythe late sixteenth century, was a highlyspecific combination of acted dialoguebetween individuals and developed spectacle.Making its way in popular rather than pri-marily aristocratic theatres, it drew heavilyon those arts of visual representation

(in both acting and scene) which had beencentral in a popular pre-literate culture. Actsof violence, for example, were now directlystaged, rather than narrated or reported.Drama as visible action, without words, wasavailable in the simple form of the dumbshow or in the highly developed forms ofstaged processions, battles or visions …Music and song were also used, but withrare exceptions not integrally, but as iso-lated elements of performance. (Williams,1981: 154–5)

What really distinguished the new drama,however, was a new form of dramatic speechthat is particularly interesting to the socialresearcher because of its diversity. For the firsttime the speech of all social groups was repre-sented, from the most educated to the most‘vulgar’. However, this endured for only arelatively brief period before an increasingsocial exclusiveness of the theatre as a culturalexperience was linked to a correspondingreduction in the linguistic range of Englishdrama.As we have stressed before, the form aswell as the content of imaginative works can’tbe separated from the social context in whichthey take place.

Williams used drama to illustrate his argu-ment partly because it was his own specialismand partly because of the long historicalrecord, which allows the changes to be docu-mented. The same principles could, though,be applied to other imaginative works, such aspainting, the cinema or the novel.The centralpoint remains the same: that new forms anddevelopments from one form to another can’tbe understood in abstraction from the histor-ical (and we would add geographical) settingin which they occurred. Where imaginativeworks are concerned, ‘what’ and ‘how’ areintimately related to ‘when’ and ‘where’.

REPRODUCTION

‘Reproduction’ includes both the ways inwhich cultural practices are transmitted fromgeneration to generation and the ways inwhich particular cultural forms are reproduced.

IMAGINATIVE SOURCES

101

3151-Ch-04.qxd 3/2/04 2:42 PM Page 101

The issue of reproduction is importantbecause it is related to social and culturalchange. As societies change and develop, thecharacter, content, role and status of culturalactivities change too. Two aspects to repro-duction may be identified. First, there is thereproduction of cultural practices throughsociety. Here education and tradition are ofcentral importance but need to be treatedwith care. In both cases a process of selectionoccurs. In an education system particular cul-tural practices are valued over others and areincluded formally or informally in the educa-tional curriculum.A ‘tradition’ can be thoughtof as ‘the process of reproduction in action; …a selection and reselection of those significantreceived and recovered elements of the pastwhich represent not a necessary but a desiredcontinuity … [This] “desire” is not abstractbut is effectively defined by existing generalsocial relations’ (Williams, 1981: 187, emphasisin original).Traditions are struggled over, andwhat gets to count as tradition (such as the‘canon’ of great authors in English literature)depends on the relationships between socialgroups.

This is not to say, however, that culture issimply the direct expression of more funda-mental social processes. The ‘closeness’between cultural practice and the conditionsof cultural practice varies widely. In somecases, such as television production, the con-nection is a close one; in others, such as sculp-ture and poetry, there is a significant distancebetween the reproduction of the culturalactivity and wider social relations.

The second aspect of reproduction is inter-nal reproduction, or reproduction within cul-tural forms. The concept of form may berefined by considering two different levels.‘Modes’ include the dramatic mode, the lyri-cal mode, the narrative mode and so on.Within each mode traditions are establishedand practitioners learn from them and thusreproduce the mode through time. Within

modes there are ‘genres’, such as ‘tragedy’,‘comedy’, ‘epic’ or ‘romance’.Williams arguesthat modes tend to survive major changes inthe social order while genres either die out orare radically redefined.

For the geographer interested in imagina-tive works, issues of reproduction are impor-tant because they enable particular works tobe located in relation to traditions, modesand genres which in turn relate to changesin social relations. For example, as we notedabove, landscape painting as a genre wasclosely connected in the eighteenth centuryto the power of the landed gentry and theirdesire to display their wealth and to main-tain the social relations on which it wasbased.

ORGANIZATION

Finally, there is ‘organization’. Organizationhas already been considered in part in thediscussion of institutions and formations, offorms and kinds of art, and of the processesof cultural production and reproduction.However it is possible to go further than thisand to consider the organization of culturein more general terms. Here Williams out-lines an approach to culture which manygeographers using imaginative works haveadopted and adapted. He argues that cultureis organized as a signifying system, a definitionwhich he contrasts with both the too spe-cific notion of culture as ‘works of art’ andthe too general notion of ‘culture as a wholeway of life’.

Individual systems, such as the economicsystem, the political system and the kinshipand family system, each has its own signifyingsystem and these are part of a wider ‘generalsignifying system’. While economics, politicsand so on cannot be reduced to processes ofsignification alone ‘it would also be wrong tosuppose that we can ever usefully discuss asocial system without including, as a centralpart of its practice, its signifying systems, on


102

3151-Ch-04.qxd 3/2/04 2:42 PM Page 102

which, as a system, it fundamentally depends’(Williams, 1981: 207).

For example, a national currency is both acrucial factor in economic activity and,simultaneously, a signifying system whichconveys meanings such as economic valueand political control over a delimited terri-tory. (If you doubt the latter, think of the dis-pute over the European single currency;clearly for some a national currency like thepound carries meanings of national identityand political authority as well as economicvalue.)

A signifying system is one in which mean-ings are produced, transmitted and inter-preted in a cycle of ‘encoding’,‘decoding’ and‘recoding’.This can be thought of as a ‘circuitof culture’ (Johnson, 1986) involving the pro-duction of meanings, which are then codedand embodied in texts of various sorts.Textsare then read (‘decoded’ and interpreted) andthereby affect lived culture which in turninfluences the production of further mean-ings, new texts and so on. Both productionand reading are conditioned by thesocial relations in which they take place(Figure 4.1).

This ‘circuit of culture’ is part of everydaylife but it is also significant in geographicalresearch. Thus geographers interpret textsand, as we have been suggesting throughoutthis chapter, in order to do this effectivelythey need to understand the process of theproduction of texts as well. For a geographerconsidering an imaginative source, therefore,it is important to consider not only the mate-rial and institutional conditions of produc-tion and the conventions and forms whichgovern the creation of artistic work, but alsothe signifying system (the cultural circuit) inwhich the work is embedded and fromwhich it acquires meaning. Geography mattershere because signifying systems vary throughtime and across space. The field of meaningin which eighteenth-century landscape

paintings were produced was very differentfrom that which gave rise to twentieth-centurynovels.

Two caveats need to be entered, however.First, the circuit of culture involves changes tomeaning. The embodiment of meaning intexts detaches it from its original context andallows it to be read in another, and the mean-ings decoded by the reader will differ fromthose encoded in the production of the text.This would only be a ‘problem’ if the purposeof interpretation was to recover the originalintentions of the author. However, as we haveshown above, this cannot be the purpose ofinterpretation, in part because authorialintention influences the meaning of texts inonly a limited way. The transmission of mean-ings in signifying systems is always mediated(hence ‘media’) and thereby changed.The cir-cuit of capital is thus a dynamic and evolvingprocess, not a wheel spinning in perpetualequilibrium.

Secondly, there is no guarantee that twoanalysts will both decode a text in the sameway. For example, the film Blade Runner wasinterpreted by David Harvey (1989a) as con-veying the phenomenon of time–space com-pression and the domination of corporatecapital, and by Doreen Massey (1991a) asconveying the inequalities of patriarchalgender relations. Again, this would be a‘problem’ only if it was assumed that all textshave a ‘true’ meaning which it is the job ofthe analyst to uncover. In fact, texts usuallyconvey multiple and often conflicting mean-ings, reflecting the complexities of the signi-fying systems in which they are produced,and it is perhaps therefore no surprise if dif-ferent readers and viewers derive conflictinginterpretations as a result. On the otherhand, it would be a mistake to assume that‘anything goes’ where interpretation is con-cerned. Just because a text will bear manyinterpretations does not mean it will bearany interpretation you like!

IMAGINATIVE SOURCES

103

3151-Ch-04.qxd 3/2/04 2:42 PM Page 103

Imaginative sources ingeographical research

We have relied heavily on Williams’s approachbecause the emphasis he places on the processand context of cultural production is in keep-ing with our concern in this part of the bookwith the ways in which sources of data aregenerated, either directly by the researcher orby other actors whether institutional or indi-vidual. In the case studies at the end of thischapter we look at the way this kind ofapproach can be adapted to help to under-stand the genesis and nature of the some ofthe kinds of imaginative sources used in geo-graphical research. First, though, let us take alook at the range of sources that geographersmight use.

The range is large and growing andincludes literature, music, the performing arts,the visual arts, film, photography, architectureand electronic media. Some of these, notably

film, still photography and electronic media,are also used for ‘factual’ documentary pur-poses, although the boundary between docu-mentary and imaginative uses is often blurred(sometimes deliberately, as in drama-documentary films). In general, however,where the primary impulse behind the cre-ation of the work is imaginative or artistic it iscovered by this chapter, and where it is docu-mentary or factual it is covered by Chapter 3.Let us now consider each type in turn.

Literature

Literature (in which we include novels,poetry and plays) is probably the commonesttype of imaginative source used in humangeography, and geographers have been enga-ging with literary works for longer than mostother forms. During the so-called ‘quantita-tive revolution’ in human geography in the1950s and 1960s, the only sources of data


104

Representation Regulation

Consumption

Production

Identity

Figure 4.1 The circuit of culture

Source: From du Gay (1997: vi)

3151-Ch-04.qxd 3/2/04 2:42 PM Page 104

which were considered valid by many humangeographers were ‘factual’, and particularlynumerical, ones. In the 1970s the positivistbases of the ‘new’ human geography werecriticized from a number of positions.Foremost among these was humanistic geog-raphy, which was also one of the first strandsin contemporary human geography to useliterary and artistic sources of data.1

In the 1970s and early 1980s, humanisticgeographers began to turn to literature (espe-cially novels) as a source of insights intohuman experience, most notably the experi-ence of place (see Figure 4.2).2 In part thiswork was motivated by the belief that novel-ists and poets possess special skills of evocationand an aesthetic sensibility that allow them todevelop such insights and convey them to thereader.The implication is that literature pro-vides an alternative, richer ‘truth’ about thenature of place and landscape that is not avail-able from more conventional data sources:

The truth of fiction is a truth beyond merefacts. Fictive reality may transcend or con-tain more truth than the physical or every-day reality. And herein lies the paradox ofliterature. Although different in essence,and therefore a poor documentary sourcefor material on places, people or organisa-tions, literature yet possesses a peculiarsuperiority over the reporting of the socialscientist … Literary truth has a universality:it evokes a response in Everyman’s breastwhile apparently concerned with the parti-cular … [It] is a truth that is more humanlysignificant. (Pocock, 1981b: 11)

While not doubting the skills of writers orthe evocative nature of their work, other geo-graphers argued that literature should not beseen as a mysterious, transcendent or other-worldly activity, but should be understoodand analysed as a social practice in its ownright, embedded in networks of economic,political and cultural relations (Thrift, 1983a;Silk, 1984; Daniels, 1985). This perspectiveimplies that before literature can be used as asource of evidence in geographical research

we need to consider the geography of itsproduction and consumption. In other words,we need to understand the contexts in whichit is written and read and the impact thesecontexts have on the meanings which areconveyed. In line with the overall frameworkof this book, this is the approach which wewill adopt.

The study of literature from a geographicalperspective continued during the 1980s3 andhas since begun to move beyond the novel toconsider other literary forms such as folk tales(Kong and Goh, 1995). More recent work canto some extent be distinguished from earlierstudies in three ways. First, it places moreweight on the social context of the productionand consumption of literary works and on theliterary conventions used by writers. Secondly,it is increasingly linked to a wider resurgencein cultural geography, which has seen a grow-ing emphasis in all kinds of geographicalresearch on issues of textuality, meaning andrepresentation.Thirdly, in line with an ongo-ing reconfiguration of the social sciences moregenerally, contemporary work is marked by itsinterdisciplinary character.Today the relation-ship between geography and literature isstudied by sociologists, cultural theorists andliterary critics, as well as geographers, whohave themselves drawn more and more on theinsights of other disciplines such as literarytheory and cultural studies.This more diverseapproach is evident in the collection edited byPeter Preston and Paul Simpson-Houseley(1994) entitled Writing the City. Finally, whilegeographers have become more sophisticatedin their treatment of literature, literary theo-rists and literary critics have also begun toexplore geographical approaches to their work(Page and Preston, 1993).

Travel writing

In one sense travel writing is just another typeof literature along with novels, poems and so

IMAGINATIVE SOURCES

105

3151-Ch-04.qxd 3/2/04 2:42 PM Page 105


106

Figure 4.2 Cover of Humanistic Geography and Literature

Source: Pocock (1981a)

3151-Ch-04.qxd 3/2/04 2:42 PM Page 106

on. However, we are treating it as a separatecategory for two reasons. First, travel writinghas made a particular contribution to thegeographical literature over a long period oftime. Initially, travellers’ accounts of theirjourneys added directly to that literature andwere seen by geographers and others as validcontributions to geographical knowledge. Asgeography became dominated by the view ofknowledge common in nineteenth-centurynatural science, travel writing was increasinglyseen as ‘unscientific’ and dismissed by the geo-graphical establishment (Domosh, 1991a).More recently there has been a resurgence ofinterest in travel writing among geographersand others (e.g. Blunt, 1994b; Gregory, 1995).Contemporary geographers do not simplytake travellers’ accounts at face value as earliergenerations might have done, but nor do theydismiss them on the grounds that they are‘unscientific’ (not least because scientific writ-ing is itself no longer thought to be a straight-forward and transparent activity). Rather,travel writing, such as that of Mary Kingsley(Figure 4.3), is interpreted for the insights itcan provide into the ways in which the worldwas understood by the writer and the societyin which he or she lived, and the ways inwhich those understandings in turn affectedthat society and its relationship with theplaces described in the writing.

Secondly, we distinguish travel writingfrom other literary forms because of the dis-tinctive context in which it is written. Unlikenovels which are intended as fiction, travelwriting involves the blurring of genres.Travelwriting is part reportage, part documentary,part factual description and part literaryembellishment. The precise mix depends onthe author and the stylistic conventions of thetime and also varies across a range of stylesfrom guidebooks to much more literaryaccounts. As we have suggested above, wewish to emphasize the role played by the con-ditions of production and consumption of the

works in informing their interpretation.Therefore the fact that travel writing involvesconditions of production distinct from thoseof more strictly literary forms means that weshould treat it as a separate category whilerecognizing the links that do exist betweenthe two.

Music and theperforming arts

By contrast with their now well-establishedstudy of literature, geographers have onlyrecently begun to express a professional inter-est in music in any sustained way.4 Even lessattention has been paid to dance and otherperforming arts, despite the integral role ofspace and spatiality in those activities.However, the current strong interest beingshown by cultural geographers and others inthe human body suggests there is much scopehere for future research.

Given the extraordinary importance ofmusic in most Western and non-Westerncultures, the lack of attention paid to it bygeographers may seem strange.This may havebeen due in part to the emphasis in traditionalcultural geography on material culture and tothe apparent difficulty of defining a distinc-tively geographical perspective on music.Nowadays, however, geographers work withless constraining definitions of both ‘culture’and ‘geography’ and there is in any case agrowing recognition of the intellectual valueof interdisciplinary work which refuses to beconstrained by the traditional narrow acade-mic division of labour.

As with the more recent work on literatureand travel writing, and in keeping with theperspective adopted in this book, contem-porary research undertaken by geographersusing musical sources does emphasize theimportance of the context in which(and the conventions through which) it isproduced and consumed. As Leyshon et al.

IMAGINATIVE SOURCES

107

3151-Ch-04.qxd 3/2/04 2:42 PM Page 107

(1995:424) put it:‘key issues of geography andmusic [are]: the nature of soundscapes, defini-tions of music and cultural value, the geogra-phies of different musical genres, the place ofmusic in local national and global cultures.’

Moreover, the role of geography is central:‘Space and place are … not simply the siteswhere or about which music happens to bemade, or over which music has diffused, butrather different spatialities are suggested as being


108

Figure 4.3 Mary Kingsley in 1897

Source: From the Royal Commonwealth Society Collections, Cambridge University Library

3151-Ch-04.qxd 3/2/04 2:42 PM Page 108

formative of the sounding and resounding ofmusic (1995: 424–5). Leyshon et al. go on tooutline how an understanding of geographicalcontext and of the role that space and placeplay in constituting music and musical activ-ity might illuminate studies of both theWestern classical tradition and the popularmusic industry (1995: 425–31).

Painting and the visual arts

The resurgence of interest in cultural geogra-phy has also led to much greater attentionbeing paid to the visual arts as sources of geo-graphical insight (e.g. Bonnett, 1992; Daniels,1993). Within this trend, painting has pro-vided most of the inspiration. Because of itsobviously geographical theme, many geogra-phers have become particularly interested inlandscape painting, though in principle manydifferent types of subject-matter could beconsidered. The interpretation of landscapehas been one of the most enduring themes ingeography, and many geographers are inter-ested equally in landscapes on the ground andin the representation of landscapes in paint-ings and other media. Recently, there hasbeen a growing realization in geography thatphysical landscapes (trees, fields, parks, gar-dens, buildings and so on) are frequently asmuch products of the human imagination aspaintings, and that the techniques involved ininterpreting paintings (especially for theirsymbolic content) can be applied to greateffect in the interpretation of the physicallandscape too (Cosgrove and Daniels, 1988).This process of describing and interpretingvisual imagery for its symbolic meanings isknown as iconography. While painting is themost widely used medium among the visualarts for geographical research, other formshave started to generate some interest. DavidMatless and George Revill (1995), for example,have examined the ‘land sculpture’ of AndyGoldsworthy.

As with music and literature, our view isthat a full understanding of a visual sourcerequires a knowledge of the conditions of itsproduction and consumption, including theconventions through which its meanings areconveyed. Landscape painting in eighteenth-century England, for example, was not simply(or even mainly) about the innocent andrealistic depiction of the countryside. On thecontrary, landscape artists were frequentlycommissioned by major country landownersto represent visually the power and wealth oftheir estates. Such depictions then in turnhelped to shore up that wealth and power bymaking them appear to be part of the ‘natural’order of things (Daniels, 1989; see also Rose,1993b: 91–3).

Photography

Although geographers often approach photo-graphy in a similar fashion to the other visualarts, it is worth distinguishing as a separate cate-gory because of the distinctive conditions inwhich photographs are produced and con-sumed. ‘Art’ photography, such as the land-scapes of Ansel Adams, are often very ‘painterly’of course, but photography also includespopular ‘snapshot’ forms as well as a variety ofapproaches which depend for their effect onthe qualities of the technology itself, such asthe mosaics produced by artist David Hockney.

Landscape, field and aerial photography hasalways played an important, if mundane, role inconventional geographical research and carto-graphy. Although these uses of the mediummay seem straightforward and descriptive theyare by no means a merely technical matter.The use of aerial photography in map-making, for example, implies a particular viewof the role and status of maps. Although itappears to reveal the whole landscape it hasthe effect of excluding from the map every-thing which is invisible to the camera. In land-scape and field photography, apparently

IMAGINATIVE SOURCES

109

3151-Ch-04.qxd 3/2/04 2:42 PM Page 109

neutral and truthful representations depend onconsciously or subconsciously selected view-points, framing and composition which inturn convey particular sets of meanings andimply particular values. Even the commonappeal to photography as a guarantor of truth(‘the camera never lies’) privileges the visualover other means of perceiving the world.

In this chapter, however, we are not mainlyconcerned with the geographer’s own use ofphotography in geographical research butwith the production of photographs by otherswhich are then used as a source in geography.Cultural geographers have begun to examinephotography for the meanings and symbols itconveys and for its social content. In doing so,the context within which photographs aremade is of central importance. For example,Phil Kinsman (1995) has shown how the land-scape photography of Ingrid Pollard cannot beunderstood without reference to the photo-grapher’s identity as a black woman or to therole of the countryside and ideas about ‘race’in the construction of British national identity.In David Harvey’s (1989a) discussion of thephotographs of Cindy Sherman,he argues thatthe unsettling approach to appearance andidentity which they reveal is closely bound upwith a contemporary shift from modernity topostmodernity, which in turn needs to beinterpreted in the context of the restructuringof capitalism. Doreen Massey (1991a) and RosDeutsche (1991) also argue that Sherman’sphotographs can’t be understood separatelyfrom their cultural context but, in sharp con-trast to Harvey, they argue that the genderrelations in which their production and con-sumption are entwined are just as important astheir relationship to capitalism.

Cinema and television

Related both to photography and to theperforming arts, cinema and television arenevertheless distinct because of their particular

conventions, production technologies andorganization. They are also widely used fordocumentary work, and this form hasreceived some attention from geographers.Gillian Rose, for example, has studied the useof film by community groups in east London(1994).

Cinema and television are hugely impor-tant in daily life and in the economy and arethus potentially of considerable interest togeographers. Although a central element inthe discipline of cultural studies, fictionalwork in these media has only recently begunto be studied by cultural geographers.This isdespite the importance of space and distancein the construction of films and of represen-tations of place in their content.

This past neglect is now being addressed bygeographers working from a variety of per-spectives. Gold (1985) looked at the futuristicvisions of the city in early cinema. Tomaselli(1988) used Afrikaans cinema to investigatethe role of the myths of the trek, the outsideand Eden in the production of Afrikaanernational identity. In his analysis of TheCondition of Postmodernity, David Harvey(1989a) draws on two films, Ridley Scott’sBlade Runner and David Lynch’s Blue Velvet,to show how the experience of time–spacecompression is exhibited in cultural phenomena(see also Deutsche, 1991; Massey, 1991a).Stuart Aitken and Leo Zonn (1993) have dis-cussed the connections between gender rela-tions and the environment in the films ofPeter Weir, while Matthew Gandy (1996)examines the films of Werner Herzog toinvestigate the portrayal of nature in Westernthought.The breadth of potential for geogra-phers in studying film is demonstrated by therange of essays published in a collection editedby Aitken and Zonn (1994a).

While professional painters work mainlyalone, either for themselves or to a commission,professional cinema and television productionis a highly organized activity involving dozens


110

3151-Ch-04.qxd 3/2/04 2:42 PM Page 110

or even hundreds of people, large amounts ofcapital equipment, and big organizations suchas film companies and commercial or publicservice broadcasters.These contexts form theconditions of production of most filmsand television programmes and have beenanalysed from geographical perspectives.Michael Storper (1994), for example, arguesthat the US film industry has moved fromearly craft production through the mass-production era of the big studios in the 1940sand 1950s, to the increasingly vertically dis-integrated forms of organization of today, inwhich particular parts of the productionprocess, from screenwriting to lighting andfrom editing to catering, are contracted outto specialized companies (see also Scott, 1984;Storper and Christopherson, 1985). Thesestudies are important because, as withother media, the conditions of productionneed to be understood in constructinginterpretations.

As well as the impact of industrial organiza-tion and the material processes of film and tele-vision production, however, it is necessary toconsider the cultural and aesthetic context inwhich movies and programmes are produced.For example, Aitken and Zonn (1994b) notethat the conventions governing the portrayal ofplace in film have changed over the course ofcinema history. Larry Ford (1994) shows howthe representation of cities in film has developedfrom the passive backdrops of early cinema,through the shadowy, menacing settings of filmnoir, to contemporary experiments with colour.Furthermore, the geography of film is not onlyabout the ways in which places are shown butalso about the space of the medium itself. Farfrom being a direct portrayal of life, films arecarefully constructed through the composition,framing and lighting of particular shots, throughtechniques such as panning, zooming and jump-cutting and through special effects. The con-struction and manipulation of space are centralto many of these aspects of film-making.

Finally, grassroots and amateur traditions offilm-making and video production also existand these too are potential sources for geo-graphical research. Mike Crang, for example,has argued that geographers’ growing interestin the use of video by police and the localstate in attempting to control urban crimeand disorder should be matched by an equalattention to popular and domestic uses of(increasingly accessible) video technology(Crang, 1996). Video is also becoming a directtool of geographical research. Adopting prac-tices developed in visual anthropology(Turner, 1991; 1992), geographers are begin-ning to make videos as part of the process ofresearch itself, often in collaboration withthose being studied and sometimes handingcontrol of the camera to the ‘objects’ of theresearch (Byron, 1993; Crang, 1995).

Architecture

Geographers have long been concerned withthe built environment, with the planning oftowns and cities and with urban morphology.However, it is only recently that they begun tomove on from a concern with the structuresand functions of whole urban areas or neigh-bourhoods to cultural and political analysis ofthe buildings which make them up. Lily Kong(1993), for example, has analysed the relation-ships between religious buildings in Singaporeand the conceptions of ‘sacred space’ associatedwith them. David Harvey (1985) has written adetailed narrative of the construction of thebasilica of Sacré-Coeur in Montmartre, Paris(Figure 4.4), revealing the political conflictsand suffering which surrounded its develop-ment. In more contemporary vein he argues(1989a) that the transition from modernismto postmodernism in architecture must beunderstood in relation to the changing char-acter of the capitalist economy.Derek Gregory(1994) has discussed the work of Walter

IMAGINATIVE SOURCES

111

3151-Ch-04.qxd 3/2/04 2:42 PM Page 111

Benjamin, whose writings on the passages(arcades) of Paris provide a powerful evocationof the cultural dynamics of modernity.

Paris again provides the source of anotherexample. Woolf (1988) interprets the ParisOpera House in the context of the opulenceand conspicuous consumption of the Parisianelite during the Second Empire, but challenges

the conventional view that the flamboyantbuilding accurately symbolized Parisian life atthe time. By setting her analysis against thepicture of a divided and conflictual urbanenvironment, Woolf shows that the lifestylerepresented by the Opera House was limitedto a tiny proportion of the population andthat the building symbolizes aspirations,


112

Figure 4.4 The basilica of Sacré-Coeur

Source: Harvey (1985: 201)

3151-Ch-04.qxd 3/2/04 2:42 PM Page 112

desires and political power rather thaneveryday life.

Mike Davis, in his account of 1980sLos Angeles as the City of Quartz (1990), showshow the increasingly fortified architecture ofthe city reveals new social divisions betweenthe wealthy, security-conscious residents of theprivately guarded housing developments andthe increasingly marginalized sections of thepopulation whose physical exclusion from‘Fortress LA’ mirrors their growing socialexclusion from American society.

As with other forms of imaginative expres-sion, architecture is used as a research sourceby geographers in a number of ways. Whatthese examples suggest, though, is that issuesof context are as important here as they arein interpreting other media such as film andliterature.

Objects andmaterial culture

Buildings are expressive of the human imagi-nation but they are also functional objects thatare used for all kinds of purposes by peoplewho may rarely if ever think about the imagi-native meanings bound up in their construc-tion.The same is also true of all kinds of othereveryday objects (from clothes to cars) thatsimultaneously express meaning and servefunctional purposes. The interpretation of‘material culture’, as the study of such objectsis commonly known, has become increasinglyimportant in geography.Work on the geogra-phy of consumption, for example, has blurredthe distinction between cultural and eco-nomic geography to a remarkable extent,recognizing the ways in which functionalcommodities become loaded with culturalmeanings. For example, a study two of usundertook with Mark Thorpe examined therelationship between food products andmulticultural identities in Britain (Cook et al.,1999). Other geographers have focused on

fashion and clothing retailing in both the‘designer’ and ‘charity shop’ segments of thatmarket (Crewe and Goodrum, 2000; Gregsonet al., 2000), the iconography of banknotes(Gilbert, 1998), lifestyle magazines (Jacksonet al., 1999) and household furniture (Leslieand Reimer, 2003).

This focus within geography on materialculture or, more prosaically, ‘things’, has beengiven impetus by related developments. Firstthere is geographers’ increasingly close rela-tionship with anthropologists such as DanielMiller, whose work on shopping, consump-tion and material culture has been particularlyinfluential. Secondly, economic geography hasdeveloped in new directions, including afocus on commodities and commoditychains. Thirdly, the popularity of actor-network theory in geography has drawnattention in new ways to the transformationof objects as they are produced, circulated andconsumed.

Electronic media

With the rapid development of new informa-tion and communication technologies basedon microelectronics, fibre optics and the stor-age and transmission of information in digitalform, a whole new range of potential sourcesfor geographical research has emerged. Mostgeographers are now familiar with computersused for data analysis, wordprocessing, cartog-raphy and graphics. In addition, more andmore researchers rely on local networks ofcomputers for communicating with colleaguesand on the Internet to send and receive elec-tronic mail, to search library catalogues, data-bases and archives, and to view, download anddisseminate a wide range of multimediaresources (incorporating some combination oftext, graphics, photographs, video and sound)through the World Wide Web.

In addition, electronic media are potentiallyas much an object of study for geographers as

IMAGINATIVE SOURCES

113

3151-Ch-04.qxd 3/2/04 2:42 PM Page 113

a tool for geographical research. In some cases,electronic technology is used simply to storeand transmit sources which have been dis-cussed above, such as literature, photographsand video. Increasingly, however, the distinc-tive capabilities of electronic media are beingexploited by creative artists, writers anddesigners to generate new forms whichcannot be thought of merely as electronic ver-sions of old forms, and which need to beunderstood on their own terms (Gilbert, 1995).These qualitatively different forms include thefollowing.

MULTIMEDIA PRODUCTIONS

In a multimedia production, a musical work ispresented, for example, simultaneously as sound,video of the musicians playing, text relating tothe composer, the work or the performance,graphical display of the score, and electroniclinks to related documents or presentations.

HYPERTEXT

Hypertext involves the creation of electroniclinks between parts of the text of one documentand a part or all of another document. Similarlinks can be made between photographs, videoclips and so on.At its simplest this may involveproviding hidden definitions of technical termsor links from, say, the name of a character in anovel to a short biographical note to remind thereader of who the character is.More elaborately,it has been suggested that hypertext potentiallyinvolves a new relationship between writer,reader and the text.Text is no longer constrainedto the linear form it usually takes on paper butcan become layered or multidimensional.Thereare clear possibilities here for creative writers ofall sorts but so far there are relatively few ‘hyper-novels’ available to the public.

COMPUTER GAMES

Although the market for computer games maybe dominated by ‘shoot-em-up’ productions,there are examples which engage the

imaginations of both the programmer and theplayer in more complex ways, including elab-orate role-playing and simulations of socialactivities, such as Sim City.

GRAPHICS AND ANIMATION

Although they appear to mimic conventionaldrawing and cartoon films, respectively, com-puter graphics and computer animation oftenuse the qualitatively distinctive features of elec-tronic technology to produce their effects.These capabilities make it much easier, forexample, to incorporate real actors with ani-mation in films like Who Framed Roger Rabbit?or to provide dramatic special effects like thosein The Mask. Increasingly, full-length featurefilms, such as Bug’s Life, are being producedentirely using computer-generated animation.

INTERACTION

Electronic media provide for new forms ofinteraction between the ‘author’ and the ‘audi-ence’.At its simplest, readers can provide rapidfeedback through electronic mail, but moresophisticated uses will allow readers to assist inscripting plots or directing characters, for exam-ple. Increasingly, however, mass access to theInternet is challenging conventional notions ofwriter and reader. Bulletin boards, discussiongroups and self-publishing are all growing aschannels of expression and debate and provide,some argue, more democratic and participatoryaccess to ‘the media’ than ever before.5

VIRTUAL REALITY

Virtual-reality (VR) technology combinesinteraction with special video and sound facil-ities and attempts to create the illusion thatthe viewer/user is participating directly inexperiences and environments which arecreated or recreated by a computer or trans-mitted electronically from elsewhere. WhileVR is still developing it may offer the potentialto engage in radically new ways with imagi-native sources. Instead of sitting in a cinema towatch a film, for example, it may be possible to


114

3151-Ch-04.qxd 3/2/04 2:42 PM Page 114

IMAGINATIVE SOURCES

115

‘walk through’ the screen and ‘take part in theaction’. Here, too, the traditional distinctionbetween reader and author and between thereal and the imaginary is blurred or eliminated.

Geographers have begun to study the newelectronic media in a variety of ways. Fromthe perspective of political economy, ManuelCastells (1989) has investigated the relation-ships between cities and the growing role ofinformation and communication technology.Economic geographers have made manystudies of the geography of the computing,software and electronics industries.6 Morerecently the imaginative aspects and culturalimplications of the use of new media havebegun to be explored.7

Case studies

To conclude the chapter we will consider threecase studies drawn from the geographical liter-ature to show how understanding imaginativeworks in geographical research requires anappreciation of their conditions of production.As we shall show, the cases apply some of theprinciples we have outlined. However, the

framework above cannot be used as a mechanicaltemplate when considering imaginative works.Rather, it should be seen as a set of guidelineswhich need to be applied to particular cases inan appropriate way.

The case studies illustrate the overall argu-ment of this chapter, which is that in orderto use imaginative sources in geographicalresearch it is necessary to understand theconditions in which they were produced. In thefirst case study we can see how an adequateinterpretation of the role of landscapemetaphors in Elizabethan drama depends on anunderstanding of the Elizabethan worldview,which provided the discursive context in whichthe plays were written. Similarly, the secondcase study demonstrates that interpreting popu-lar music (in this case in Singapore) involvesmuch more than examining the geography inthe lyrics. Issues of identity, the structure of themusic industry and processes of ‘transcultura-tion’ all need to be taken into account in con-sidering the context in which such work isproduced. In the final case study the genderrelations of eighteenth-century England arecentral to interpreting the landscape painting ofThomas Gainsborough.

CASE STUDY 4.1: LANDSCAPE METAPHORS IN THE WORKOF WILLIAM SHAKESPEARE

The first case study is provided by a paper by Paul Chamberlain published in TheCanadian Geographer in 1995. It presents an interpretation of the role of metaphor inShakespeare’s plays, with particular reference to the portrayal of the landscape and thehuman body. Chamberlain is interested in ‘“the metaphorical vision” or the use ofmetaphor to understand the human environment relationship’ (1995: 309).

Chamberlain identifies five (overlapping) types of metaphor:

1. Dead metaphors such as ‘table leg’. In dead metaphors, the figure of speech hasbecome so much a routine part of the language that most of the time it is not seen asmetaphorical at all.

2. Metonymy, in which one word is used to stand for another, as in the use of ‘bottle’ torefer to strong spirits.

3. Personification, in which inanimate objects are given life-like features. Chamberlain’sexample is the use of the word ‘angry’ to describe the sea.

3151-Ch-04.qxd 3/2/04 2:42 PM Page 115


116


4. Metaphors in which there is ambiguity between the object of the description(the metaphrand) and the word doing the describing (the metaphier).

5. Extended metaphors, in which an author makes continuing use of a metaphor throughthe whole or a large part of a work.

An analysis of Shakespeare’s work is interesting geographically, according toChamberlain, because it means that ‘we can not only explore an intriguing geographicalaspect of the literary text of a highly creative Elizabethan, but can also improve ourunderstanding of the relationship that human beings had with their environment duringthe Renaissance’ (1995: 318–19).

Chamberlain then discusses two sets of metaphors which appear in Shakespeare’splays. First there are those in which the human body is described in terms of thelandscape. These are ‘landscape:body metaphors’. Secondly there are those in which thelandscape is described in terms of the human body: ‘body:landscape metaphors’. In eachcase the landscape includes the celestial landscape of the stars and planets as well as theterrestrial landscape. This may seem odd to modern readers (we might use the term‘skyscape’ for the former case) but in the context of Elizabethan literature it is importantto stress the unity of the earth and the heavens for reasons which will be apparent later.

Chamberlain shows how Shakespeare uses a range of metaphors in each of these twocategories. These include simple uses such as likening a human being to a mountain:‘Mountains are frequently superimposed on a variety of characters such as Luce in The Comedyof Errors, whom Dromio of Syracuse describes as being a “mountain of mad flesh”’ (1995: 313).Celestial images are also used: ‘in Antony and Cleopatra when Mark Antony commits suicide byfalling on his sword, the guard exclaim[s] that “The star is fall’n”’ (1995: 313).

More elaborate metaphors include allusions to bodies as gardens:

Some of the most important metaphors in Shakespeare’s work involve the garden.This element of the cultural landscape is likened to the body by Iago in Othello whenhe tells Roderigo: ‘Our bodies are our gardens, to the which our wills are gardeners;so that if we will plant nettles or sow lettuce, set hyssop and weed up thyme, supplyit with one gender of herbs or distract it with many – either to have it sterile withidleness or manured with industry – why, the power and corrigible authority of thislies in our wills.’ (1995: 314)

In Richard II Shakespeare uses the idea of a disorderly garden to convey the king’s(gardener’s) lack of effective governance over the nation (garden). Body:landscapemetaphors also abound. In As You Like, ‘the exiled duke fondly likens the chilly gusts ofwind to “counselors” (II.i.10) and later finds “tongues in trees” and “books in runningbrooks” (II.i.16)’ (1995: 317).

Similarly, cities are given human properties and celestial imagery recurs, as in thefollowing speech by Henry in Henry V:

I know you all, and will a while uphold.The unyoked humor of your idleness.Yet herein will I imitate the sun,Who doth permit the base contagious clouds

3151-Ch-04.qxd 3/2/04 2:42 PM Page 116

IMAGINATIVE SOURCES

117


To smother up his beauty from the world,That, when he please again to be himself,Being wanted, he may be more wond’red atBy breaking through the foul and ugly mistsOf vapours that did seem to strangle him(I.ii.192–200) (1995: 316)

In using works such as Shakespeare’s plays in geographical research, geographers seekto interpret the texts and their meanings (in this example, especially the use of metaphor).However, as Chamberlain shows and as we have been arguing in this chapter, suchinterpretations, if they are to be adequate, must be informed by an understanding of theconditions of the production of the texts. Chamberlain argues that this involves givingattention to the textual, the extra-textual and the inter-textual (references to other texts).

A full analysis of the conditions of production of Shakespeare’s texts would include,among other things, a discussion of the organization of Elizabethan theatre, of theconventions of Renaissance drama, of the political conditions of sixteenth-centuryEngland and a host of other ‘extra-textual’ factors. To inform his interpretation of thelandscape/body metaphors, Chamberlain identifies three sets of conditions as of particularimportance. These are the technical aspects of metaphor which were outlined above, therole and importance of astrology and alchemy in Renaissance systems of thought, andElizabethan cosmology based on the doctrine of the ‘Great Chain of Being’

The importance of astrology to the knowledge systems of late medieval Europe clarifiesthe significance of Shakespeare’s use of celestial imagery. Elizabethans took it for grantedthat the movements of the sun, the moon and the planets influenced the affairs ofhumans, and such ideas are frequently used by Shakespeare including, as we have seen,in his use of metaphor. The idea of the Great Chain of Being in which all creation is seenas a linked hierarchy from God at the top to the humblest inanimate object at the bottom,helps to explain the power and apparent naturalness with which Shakespeare links people,the earth and the heavens in his use of metaphor. In order adequately to understand thegeography of Shakespeare, therefore, it is essential also to understand what the literarycritic E.M.W. Tillyard called the ‘Elizabethan world picture’ (1972).

CASE STUDY 4.2: POPULAR MUSIC IN SINGAPORE

Our second case study is drawn from the work of geographer Lily Kong. In a researchpaper published in the journal Society and Space in 1996, she presents an interpretationof the work of the Singaporean pop star, Dick Lee. While separated from ElizabethanEngland by 400 years and 7000 miles, here, too, we find that relations between texts,contexts and other texts are central to the analysis (Kong, 1996: 279). Indeed Kong’swork shows just how fruitless is any attempt to separate rigidly the interpretation of textsfrom an understanding of their production.

Dick Lee was born in 1956 and was brought up in Singapore surrounded by Englishculture. It was only when he travelled to England for the first time that his Asian identitybecame clear to him (Kong, 1996: 278). Lee grew up in the 1960s and 1970s and was

3151-Ch-04.qxd 3/2/04 2:42 PM Page 117


118


heavily influenced by a range of Western popular musicians of the time. Since then hehas pursued his own successful recording career and has developed a distinctive musicalstyle which combines local Singaporean and wider Asian influences with Western idioms.Kong elaborates on the three geographies present in the work of Dick Lee.

First, she argues that ‘a sense of the local is strong in Lee’s music’ and ‘that this is evidentin two ways: the injection of distinctive local contexts, creating a sense of place, and therecovery of personal, communal, and national heritages, in which a combined sense ofhistory and nostalgia is often drawn upon’ (1996: 279). Kong shows that Lee’s lyrics containfrequent references to local places in Singapore and to particular aspects of Singapore’scultural, economic and political milieu. In addition, Lee draws on his own roots to emphasizethe importance of local identity. Thus a local market, ‘Beauty World’, remembered from hischildhood, provides the title of one of his albums and a musical. As Kong (280) puts it, ‘forLee, his sense of place and feeling of nostalgia are all rolled into one and find expression inhis music’. She also shows how he incorporates a mixture of traditional musical genres intohis output, reflecting the rich mix of Singapore’s multiracial society.

Secondly, Kong illustrates the global influences on Lee’s music. This produces ‘localmusic with a transnational flavour’ through a process of ‘transculturation’ (1996: 281):

What are the evidences of such transculturation? Instrumentation, styles, language,costumes, and guest artistes can all be called into play here. In Life in the Lion City(1984) Lee mixed traditional instruments with synthesisers, and claimed this to be hisfirst introduction to the search for a new identity via music. This has become arecognisable feature of his music, described as blending pop funk beats with Asianinstruments … and a fusion of different idioms … In addition, Lee’s music has oftenbeen described as ‘anglicised’ versions of local songs … For example, a Chinese folksong, ‘Little White Boat’, is given English lyrics; a traditional Malay folk song, ‘RasaSayang’, is infused with a modern rap; and traditional folk tunes (‘Alisham’, ‘Springtime’,and ‘Cockatoo’) and English pop styles are cleverly incorporated. (1996: 282)

Thirdly, Kong describes how Lee has begun to search for a regional Asian identitythrough his music. This, she says, was for three reasons:

First, he wanted to change what he saw as the Western image of Asians … Second,he recognised that there was a sense of shared identity crisis. It was no longer acase of ‘Who am I?’ or ‘Where am I going?’, but one of ‘What role do I play as anAsian?’ – a question he felt many modern Asians were asking. Third, he wanted torecover what he says to be the spirit of the 1960s when Asian pop enjoyed asignificant degree of popularity. (1996: 284)

These personal motivations are only part of the story, however. As an artiste undercontract to the recording company WEA, Lee is bound into economic circuits whichstretch beyond the confines of his native Singapore. Noting that Singapore has apopulation of only 3 million, Kong argues that the fact:

That Lee was signed up with WEA which, like all record companies, is first andforemost a large business enterprise governed by laws of profit and loss, meant thatit was imperative to reach out to as large an audience as possible. Given that the

3151-Ch-04.qxd 3/2/04 2:42 PM Page 118

IMAGINATIVE SOURCES

119


local market is very small, to sign up any local artiste would necessitate marketingand distribution in the regional market. Music that is produced is therefore guided bythis need and it is of little surprise that Lee’s music addresses a question of Asianimport. (1996: 285)

In Williams’s terms, therefore, Lee’s attempts to stretch the genre of his work to developan Asian rather than just a Singaporean popular music is matched by institutional factors,such as the structure of the record industry and the size of markets. According to Kong,however, Lee’s attempts to link his work to a pan-Asian identity is only partially sucessfulin musical terms, in part because it is unclear whether the notion of a pan-Asian identity,bringing together very diverse traditions, makes sense. Nevertheless, even if a unifyingidentity is elusive, the music can still have wide regional appeal (1996: 287).

Kong shows how these three geographies of Lee’s music (the local and Singaporean,the global and Western, and the Asian and regional) can be related to broader processesof international cultural change. She outlines two contrasting interpretations of change.First is the globalization thesis, which posits that cultures are becoming more and morealike, as global networks of communication and transnational media corporations operateto homogenize cultural production. Secondly is the ‘music is local’ thesis, which suggeststhat ‘music is necessarily local in its production, consumption, and effects’ (1996: 275)because music is about giving expression to ethnic, cultural and other local differencesso that ‘unique local sounds are produced because of the unique local milieux in whichthey are produced’ (1996: 275).

For Kong, however, both these theses are inadequate:

Although there are clearly those who advocate global and local positions, and argumentcan be made that global and local cultures are in fact relational rather than oppositional… As Friedman (1990, page 311) argues, ‘Ethnic and cultural fragmentation andmodernist homogenization are not two arguments, two opposing views of what ishappening in the world today, but two constitutive trends of global reality’. (1996: 276)

Instead of globalization or localization, what is happening, Kong suggests, istransculturation:

It could be argued that the process at work is actually one of transculturation, a‘two-way process that both dilutes and streamlines culture, but also provides newopportunities for cultural enrichment’ (Wallis and Malm 1987: 128). Transculturationis apparent when musicians are influenced dually by their own local culturaltraditions and by the transnational standards of the music industry. The result islocal music with a transnational flavour or transnational music with a local flavour …Some … have argued that transculturation in fact gives rise to ‘third cultures’ whichdraw upon the culture of the parent country but also take into account local culturesand local practices. (1996: 276–7)

To interpret the musical works of Dick Lee, therefore, Kong finds it necessary to considernot only the immediate context of their production in Singapore and in Lee’s own lifestory, and the economic relations of the international music industry, but also the widerprocesses of cultural change by which Lee, Singaporean culture and the music industryare all affected but which they, in their turn, are affecting as well.

3151-Ch-04.qxd 3/2/04 2:42 PM Page 119


120

CASE STUDY 4.3: MR AND MRS ANDREWS

Our final case study is a painting that has become so widely cited by human geographersthat we feel it has become the one cultural artefact no self-respecting commentary on thepractice of human geography can afford to ignore. Thomas Gainsborough (1727–88) wasan English portrait and landscape painter. Mr and Mrs Andrews was painted around 1750,probably to celebrate the marriage of Robert Andrews and Frances Carter in 1748. Thepicture (see Figure 4.5) shows the couple in an English country landscape, part of theirestate. Mr Andrews carries a gun and a hunting dog is at his feet. It seems he has been outhunting. Mrs Andrews is seated under an oak tree. In front of the couple the sheaves ofcorn imply that it is autumn. The painting, though unfinished, is brilliantly executed and istestament to Gainsborough’s status as one of the great artists of eighteenth-century Europe.

Landscape has long been a central category in geographical analysis and in the last20 years landscape painting has also become a focus of great interest. By and large,geographers studying landscape painting take it for granted that no painting providesunmediated access to the reality of the landscape depicted, no matter how realistic thestyle of expression chosen. Landscape painting, as is the case with other art forms suchas Shakespearean drama or popular songs, is heavily conditioned by conventions and theprocess of production. With this in mind, Mr and Mrs Andrews has attracted a certainamount of interest from geographers as well as art critics.

In his influential book (1972) and television series, Ways of Seeing, the critic JohnBerger argues that Mr and Mrs Andrews exemplifies the relationship between oil paintingand property. Mr and Mrs Andrews are not simply a couple taking delight in their naturalsurroundings. Rather, ‘they are landowners and the proprietary attitude towards whatsurrounds them is visible in their stance and expressions’ (Berger, 1972: 107). Thepleasure the couple took from the painting was not purely aesthetic but included ‘thepleasure of seeing themselves depicted as landowners and this pleasure was enhanced bythe ability of oil paint to render their land in all its substantiality’ (1972: 108).

The painting adopts realist conventions, in particular the convention of perspective,which ‘centres everything on the eye of the beholder … the visible world is arranged forthe spectator as the universe was once thought to be arranged for God’ (1972: 16).Moreover, Berger argues that oil painting specifically, more than other arts, provided anparticularly appropriate medium for the expression of the emerging social relations ofcapitalism: ‘oil painting did to appearances what capital did to social relations. It reducedeverything to the equality of objects. Everything became exchangeable becauseeverything became a commodity’ (1972: 87).

Berger’s analysis of oil painting and of the Mr and Mrs Andrews picture is taken up by thegeographer Stephen Daniels. Daniels (1989: 213–14) points out that Berger himself impliesthat landscape painting cannot be wholly complicit in bourgeois values because the sky, whichis central to the landscape genre, ‘cannot be turned into a thing or given a quantity’ (Berger,1972: 105). This, for Daniels, is an aspect of what he calls the ‘duplicity’ of landscape.

Another geographer, Gillian Rose, takes the interpretation of Mr and Mrs Andrews furtherstill and, in doing so, is critical of what she sees as Daniels’s unwillingness to offer asufficiently critical analysis of the pleasurable aspects of looking at landscape painting. Rosepoints out that there are important differences between Mr and Mrs Andrews in the painting:

They are given different relationships to the land around them. Mr Andrews stands,gun in arm, ready to leave his pose and go shooting; his hunting dog is at his feetalready urging him away. Mrs Andrews meanwhile sits impassively, rooted to herseat with its wrought iron branches and tendrils, her upright stance echoing that ofthe tree directly behind her. If Mr Andrews seems at any moment able to stride offinto the vista, Mrs Andrews looks planted to the spot. (1992: 14)

3151-Ch-04.qxd 3/2/04 2:42 PM Page 120

Figure 4.5 Thomas Gainsborough, Mr and Mrs Andrews, c. 1750, oil on canvas

Source: From the National Gallery Picture Library

3151-Ch-04.qxd 3/2/04 2:42 PM Page 121

Conclusion

In this chapter we have considered thegeographer’s engagement with imaginativesources from a perspective that stresses theimportance of the conditions of productionof cultural phenomena. An enormous andever-increasing variety of imaginative sourcesis available for geographers to analyse andinterpret, from classical myths to the latestelectronic simulations. These texts andartefacts take many different forms andcarry a huge range of meanings, but they allarose through a process of cultural produc-tion (du Gay, 1997). By drawing attention tothe materiality of culture and to the impor-tance of institutional and other socialrelations,Williams’s approach provides a use-ful framework (though not the only one)within which to interpret that process of pro-duction. Of course the conditions of produc-tion also shape the data that are produced by

geographical researchers themselves, and itis to those kinds of sources that we turnin the next two chapters.

Notes

1 For a discussion of humanistic geography, seeCloke et al. (1991).

2 See, for example, Seamon (1976), Tuan(1976b) and Pocock (1979; 1981a; 1981b).

3 For example, Barrell (1982), Lutwack (1984),Mallory and Simpson-Houseley (1987),Sandberg and Marsh (1988), Shortridge (1991),Cresswell (1993) and Daniels and Rycroft (1993).

4 See, for example, Smith, S.J. (1994), Cohen(1995), Hudson (1995), Kong (1995a; 1995b),Leyshon et al. (1995), Revill (1995) andValentine (1995).

5 For a less optimistic view of the ‘informationage’, see Roszak (1994), Stoll (1995).

6 For example, Hall and Markusen (1985), Scottand Angel (1987), Hall and Preston (1988)and Morgan and Sayer (1988).

7 For example, Benedikt (1991), Hillis (1994),Shields (1996) and Crang et al. (1999).


122


For Rose, this distinction between the two figures reflects the dominant association ofwomen and femininity with nature and masculinity with culture. Thus it is Mr Andrewswho is in fact the landowner, while Mrs Andrews’s function in eighteenth-century Englandwas primarily to provide children, and ideally a male heir, to carry on her husband’sdynasty. Rose suggests that Mrs Andrews’s seated position under the oak tree and besidethe field of crops reflects the assumption that women were closer to nature. Effectivelyshe becomes part of the landscape over which Mr Andrews holds sway.

In addition, Rose suggests, the passive attitude of Mrs Andrews and her status as animplicitly sexual being (because of the association between women and fertility) areconnected with the visual pleasure afforded to the spectator. In this view the active gazeof the spectator is associated with masculinity, while the observed landscape isassociated with femininity because of both its apparent passivity and its closeness tonature and fertility. Thus the pleasures of looking at landscape are not only dependentupon a particular set of property relations to do with the ownership of land but are alsoconnected with highly unequal gender relations in which ‘the gaze’ involves a masculineappropriation of a sexualized and naturalized femininity (Rose, 1992: 14–18).

These geographical readings of Gainsborough’s painting suggest that, as with the other twocase studies, effective interpretation will depend heavily on an understanding of the context inwhich the work was produced, and also consumed. The qualities of oil paint, the conventionsof perspective, property relations in eighteenth-century England and the gendering of natureand of the spectator’s gaze all feed into and influence the meaning of the picture.

3151-Ch-04.qxd 3/2/04 2:42 PM Page 122

Introduction

The history of human geography is oftenconceived of as a series of blockbuster eras ineach of which an all-encompassing mix ofphilosophy, theory and method dominates thelandscape of scholarship until the next ‘revo-lution’ heralds the next era with its differentconcepts and toolkits.As we have discussed inChapter 1, such a view is not only historicallyerroneous but it also creates misleadingassumptions about the heritage of particularmethodologies. As a prime example, recenttexts on human geography’s methods havetended to convey the impression that anyformalization of qualitative methods is anentirely recent occurrence, to be understoodin term of supposedly ‘postquantitative’ culturalturns in the subject.Yet it is important to beginthis chapter on ‘talking to people’ with a clearacknowledgement that some considerableconsideration was given to more qualitativeprocedures throughout the twentieth-centuryhistory of human geography.

In North America, for example, there isevidence of a tradition of ‘talking to’ field‘informants’ that stretches back to some ofSauer’s earliest work. For example, in his 1924piece, Sauer recognized the value of using ‘thetally sheet or questionnaire’ to gain and torecord information derived from questions tolocal people about ‘the standard of living andthe movement of population’ (1924: 30).

Around the same time, Whittlesey (1927)discusses the importance of conducting ques-tionnaires alongside observation and thecollection of statistical information. Heacknowledges that ‘systematic questioning is anecessary supplement to observation, becausecertain facts required for an understanding ofthe geography of any area are not observableby the most skilled worker’ (1927: 75).Whittlesey outlines a research methodologywhich includes ‘conversations with well-postedpeople’ (1927: 77–8), including officials such asfarm agents and school superintendents, andindividuals such as clergy or bankers. Theassumption here is that those unobservablefacts which are necessary for understandingthe geography of an area are to be found inthe knowledges and attitudes of these keylocal figures. C.M. Davis (1954: 523–7) wentso far as to include several pages of discussiondevoted to ‘the interviewing of informants’,distinguishing between ‘interviewing byquestionnaire’ (when asking specific sets ofquestions, preferably face to face) and ‘inter-viewing by informal conversation’. In the caseof questionnairing, Davis stressed that chieflyfactual information would be the main infor-mation sought, but added that attempts couldbe made to discern more attitudinal informa-tion, perhaps using the (somewhat dubious)kind of question format shown in Figure 5.1.In the case of what we might regard as trueinterviewing, he stated that the objective

Talking to People5

3151-Ch-05.qxd 3/2/04 2:43 PM Page 123

would be chiefly to ‘gather qualitativeinformation, such as the seasonal variationsof farm practice, or individual preferencesfor competing market centres’ (1954: 525).Revealingly, he offered this further remark:

Information obtained from informants is anancient source of geographical field data.Since early Greek times, if not long before,geographers learned how people lived byasking them questions. It seems probablethat more of the information included incurrent geographical writings is derivedfrom informants than from direct fieldobservation. Yet the techniques used arenot often described. Formal descriptionsand experiment with the techniques of theinterview are found in the literature ofanthropology or sociology but not of geog-raphy. (1954: 527)

We begin to get the impression, then, of a richhistory of geographers asking questions and ofgeographical information being heavilyderived from ‘informants’. There are evenglimpses of how such information was beinggathered. In a paper gloriously titled ‘Thebicycle as a field aid’, Salter (1969) discussesthe role that the bicycle plays in ‘openingroutes of information which might otherwise

be unfound’ (p. 361), notably because thebicycle provides an excuse for a ‘chat’ to localpeople – it gets the fieldworker into a positionwhere ‘informants’ can readily be accessed.

In Britain, the evidence of an emergentmethodology of ‘talking to people’ is moresparse, with few explicit acknowledgementsof the role of qualitative interviewing untilrecent times. However, it does seem clear thatamong the ‘field studies’ favoured by manyBritish geographers from at least the 1930sonwards there could not but have been someengagement with field ‘informants’. So, forexample, Beaver’s (1962) discussion of the LePlay Society – a field-based organizationencouraging practical exercises in geography –outlines typical field meetings, involvingacademics and lay folk, which ran accordingto a clear ‘scheme of study’ in a given locality.For instance, under the heading of climatology,there was supposed to be the accessing oflong-term statistical data and verbal inquirieswere to be made about local peculiarities,including oddly enough ‘extraordinary mete-orological phenomena, e.g. fireballs’ (1962:235, emphasis added). Social data, meanwhile,were to include attention to ‘social organisation


124

‘How would you feel if a person in any one of the following groups should ask to marry your daughter? Marka cross in the appropriate column for each group.’

Table I

Welcome with Accept Accept but Refuseenthusiasm willingly not happily

Physician

Retail merchant

Army officer

Porter

Farm laborer

Lawyer

Figure 5.1 C.M. Davis’s formatting of questions

Source: James and Jones (1954: 42)

3151-Ch-05.qxd 3/2/04 2:43 PM Page 124

(local government, education, religiousorganisation, household life), social psycho-logy – customs, folk-lore, art’ (1962: 235). It isdifficult to know how many of these subject-matters could be explored without talking tolocal people, but nothing is said about howthis might be done in practice. Intriguingly,though, a window on what might have beenhappening arises where Beaver comments on‘those in the world of university geographywho stood aloof, who scoffed at middle-agedschoolmarms prying into peasant cottages’(1962: 239).

Around the same time, Hutchings presentsan account of ‘village studies’ and wondersabout the extent to which pure observationwill be sufficient to understand what is reallyhappening there:

Students quickly discover that the story of avillage, and especially of its social and eco-nomic working, is by no means legible in vis-ible signs. Indeed, their observation, farfrom explaining the place, raises a lot ofquestions, the answers to which they canonly obtain by talking to the inhabitants orfrom local literature. Whatever the qualityor quantity of this harvest of knowledge,it is pretty safe to say that not much of itwill be geography as we know it. (1962: 10,emphasis added)

There is an interesting sense, or even fear,here that ‘talking to people’ may end upgenerating information that was not normallyconceived of then (in the early 1960s) as‘proper’ geography.

Hence Anglo-American geography, doesinvolve a tradition of talking to informants ‘inthe field’, with both questionnairing andinterviewing being part of emerging tech-niques of regional survey. The impression,though, is that by the 1950s this tradition wasbeing deflected into a kind of a ‘regionalscience’mould, thus leaving behind the Saueriancultural-geographical roots and perhaps toobecoming more interested in the urbanregions of modernity than the rural-agricultural

regions of premodernity. Interestingly, thereferences to ‘talking to people’ in these earlyliteratures tend to envisage researchers inrural-agricultural backwaters talking to farm-ers and rural communities. With attentionturning away from these ‘premodern’ regionalworlds (see Entrikin, 1981), there seems alsoto have been a methodological shift from amore qualitative survey approach, in whichquestionnaires and interviews might figure, toa more quantitative approach preoccupiedwith statistical indicators of input–outputflows through urban (modal) regions. Thisshift was not uncontested, however. In thecontext of a discussion of the use of fieldresearch in place-name studies, Berleant-Schiller (1991) suggests the importance ofinformation given orally by field informants.She concludes that, at least since the early1950s, geographers had been failing todevelop these methods to their full advantage,not least because of a general retreat fromfieldwork in human geography until thecoming of spatial science (see Chapter 1). Shewrites:

Few American geographers of the last 30years have imitated the model of fieldresearch that Cassidy (1947) established inhis Place-Names of Dane County, Wisconsin,which incorporated materials from manyinformants as well as from maps andhistorical sources. Robert C. West (1954:67) corresponded with local informants,but his work on the term ‘bayou’ marksthe beginning of the movement awayfrom fieldwork in place name research.Miller (1969: 245) denied vigorously anyneed for or advantages in field research,dismissing it as ‘slow’, ‘unrewarding’ andsubject to ‘bias and misinformation’.(1991: 92)

Berleant-Schiller is hostile to this view andwishes to oppose such a writing out of courtas ‘an irreplaceable source of primary data inhuman geography’ (1991: 92). Here she raisesa claim which has far greater generality thanthe specific context of place-name research.

TALKING TO PEOPLE

125

3151-Ch-05.qxd 3/2/04 2:43 PM Page 125

She argues that ‘the implications of regardinglocal residents and native speakers as mis-informed about their own culture are deep,especially since Miller’s judgements were notso much part idiosyncratic as part of a largertrend’ (1991: 92).The loss of interest in talk-ing to people, then, in Berleant-Schiller’s viewis not simply a matter of human geography’s‘progress’ towards spatial/regional science. Itmarks a significant junction in philosophicaland methodological pathways – namely, thatthe preference for scientific ‘objectivity’ wasconstructing negative arguments about the‘bias’ and ‘misinformation’ inherent in meth-ods involving talking to people. Such argu-ments were even infecting more ‘peripheral’fields of cultural-geographical inquiry such asplace-name research.Yet even then there werealternative arguments about the ‘irreplaceabil-ity’ of allowing local residents and nativespeakers to speak for themselves and thus pre-sent essentially informed accounts of theirown situations.

So, what are the implications of these tradi-tions of talking to people in human geogra-phy? We should recognize that a use of moreformalized qualitative methods does have arather longer history in geography than wenormally realize, and that qualitative pro-cedures prior to the last decade or so couldinvolve more than just intuitive ‘hanging out’.We should also take seriously the suggestionthat many data deployed by human geogra-phers have always come via conversationalmeans and that we are not the first generationof geographers to appreciate this nor to urgesome explicit reflection upon how such aconversational approach operates and mightbe improved in practice. This being said, itmust be acknowledged that in earlier years aqualitative method such as interviewing wasmostly deployed as a supplement to the cata-loguing mentality of the surveys discussedabove, and so the purpose to which it wasput does differ considerably from the more

issue-based probing characteristic of morerecent studies.Yet even in these rather differ-ent contexts important arguments werealready being rehearsed about the inherentphilosophical advantages of allowing peopleto present informed qualitative accounts oftheir own circumstances. These are themeswhich reverberate throughout this chapter.

The practices of talkingto people

We now turn to a discussion of where ‘talkingto people’ fits into contemporary methodo-logical strategies in human geography.Continuing with the themes established inprevious chapters, our emphasis here is on theconstruction of data by researchers them-selves, and our assumption is that we are nowdealing with situations where preconstructedsources of data will have already been consid-ered as possible bases for practising humangeographical research but rejected becausethey are inappropriate for the research ques-tions which are being pursued. Here, there-fore, we look at different ways in whichhuman geographers can ask questions ofother people in order to construct informa-tion on issues which are important to theresearcher and/or the researched. Our aim isto avoid what Gouldner (1967) has calledmethodolatry, in which there is a danger thatwe become ‘compulsively preoccupied with amethod of knowing, which is exalted withoutserious consideration of how successful it is inproducing knowledge’ (Eyles, 1988a: 3).Rather we will attempt to show how someforms of questioning are best suited to particu-lar research circumstances, and these circum-stances will include the political and theethical as well as the technical.

There is considerable scope for confusionin the way in which we consider question-naires and interviews. People delivering


126

3151-Ch-05.qxd 3/2/04 2:43 PM Page 126

face-to-face questionnaires are often labelledas ‘interviewers’, and the resulting question-naire findings can thus easily be regarded asthe product of ‘interviewing’. Moreover, thereare grey areas between the more unstructuredend of questionnairing (for example, the useof open-ended questions) and the more struc-tured end of interviewing (for example, theuse of precisely worded and structured inter-view schedules). In this context we want tomake clear at the outset what we see as animportant distinction between questionnair-ing and interviewing. Questionnairing is usu-ally part of a wider quantitatively drivenstrategy of social survey where representativesamples of people can be questioned in orderto produce numeric measures of behaviour,attitude, attribute and so on. Interviewing isusually a qualitative exercise aimed at teasingout the deeper well-springs of meaning withwhich attributes, attitudes and behaviour areendowed.

Choosing an ‘appropriate’ practice for ask-ing questions should therefore be consideredas part of a much wider research strategy, fullyinterconnected both with careful formulationof research questions and with a clear idea ofhow the constructed data are to be inter-preted. One framework for beginning tomake these interconnections is the choice ofwhat Harré (1979) has termed ‘extensive’ and‘intensive’ research. Sayer (1984; 1992) hassummarized what each of these can mean interms of typical research strategies (Table 5.1),with extensive work focusing on patterns andregularities found among a large-scale repre-sentative group of people using methods ofdata construction such as formal question-naires and standardized interviews, and datainterpretation typically involving statisticalanalysis. By contrast, intensive work pursuesspecific processes with a small number ofpeople using interactive interviews andethnographies for data construction, andqualitative analysis as the strategy for data

interpretation. This distinction betweenextensive (== questionnaires and statistics) andintensive (== ethnography and qualitativeanalysis) can tend to be over-rigorous in‘pigeon-holing’ particular practices of con-structing and interpreting data, a pointacknowledged by Sayer (1992) in the secondedition of Method in Social Science where headds the following note to his previousaccount of extensive and intensive research:‘Note that the extensive/intensive distinctionis not identical to the more familiar distinc-tion between survey analysis and ethnogra-phy. Intensive research need not always useethnographic methods to establish the natureof causal groups, and surveys need not bedevoid of attempts to understand the socialconstruction of meaning’ (p. 244).This is a keypoint. It is, therefore, also important to notethat different methods of asking questions canincorporate attempts to identify differentdepths or layers of meaning, and the dangersof ‘methodolatry’ are as applicable to thoseresearchers favouring ethnography as to thosepreferring social scientific methods.

In order to move away from these impor-tant but sometimes partisan categorizations,we prefer to think of the range of differentquestionnairing and interviewing methods asreflecting a changing set of relationshipsbetween the researcher and the researched.This chapter thereby discusses a range ofsocial situations in research in which particu-lar researcher/researched relationships signaldifferent kinds of social construction of data.To illustrate the range of these relationships,we can briefly consider four types of socialsituation commonly encountered in humangeography research. First, there is a series ofrelationships which might be regarded asrobotic. Here the researcher will be making useof other people’s surveys, as in the use, forexample, of opinion polls. Data in this case areeffectively preconstructed – there has been nocontact or direct interaction between the

TALKING TO PEOPLE

127

3151-Ch-05.qxd 3/2/04 2:43 PM Page 127

researcher and the researched.The researcherhas had no freedom to pose the questions andcertainly no opportunity to follow up onthe answers given.The nature and form of thedata are fixed and cannot be influenced by theresearcher’s own subjectivity. Significant epis-temological issues arise in the interpretationof information which has been precon-structed in this way and, although the

researcher will have some scope to imposepersonal and positional interests during theuse of such data for interpretative analysis,much of the relational power of the researchis vested in those who set the questions andco-constructed the answers with respondents.

Secondly, we can envisage remote relation-ships between the researcher and the researched(for example, as occur in the deployment of


128

Table 5.1 Intensive and extensive research: a summary

Intensive Extensive

Research How does a process work What are the regularities, question in a particular case or common patterns, distinguishing

small number of cases? features of a population?What produces a certain How widely are certain characteristicschange? or processes distributed or What did the agents represented?actually do?

Relations Substantial relations Formal relations ofof connection similarity

Type of Causal groups Taxonomic groupsgroupsstudied

Type of Causal explanation of Descriptive ‘representative’ account the production of certain generalizations, lacking in produced objects or events, though explanatory penetration

not necessarilyrepresentative ones

Typical Study of individual Large-scale survey of populationmethods agents in their causal or representative sample, formal

contexts, interactive questionnaires, standardized interviews, ethnography. interviews.Qualitative analysis Statistical analysis

Limitations Actual concrete patterns Although representative of a and contingent relations whole population, they are are unlikely to be unlikely to be generalizable ‘representative’, ‘average’ to other populations at or generalizable. different times and places.Necessary relations Problem of ecological fallacy discovered will exist wher- in making inferences aboutever their relata are present, individuals.e.g. causal powers of objects Limited explanatory power.are generalizable to othercontexts as they arenecessary featuresof these objects

Appropriate tests Corroboration Replication

Source: Sayer (1984: 222)

3151-Ch-05.qxd 3/2/04 2:43 PM Page 128

postal questionnaires).The involvement of theresearcher in the direct framing of questions isa significantly powerful process in the con-struction of data. However, the lack of face-to-face contact with the respondent meansthat the researcher has no interpretativehistory to draw on when seeking to makesense of the answers. Issues relating to theinappropriateness of the categories providedfor answering questions, the potential ambigu-ity of the questions themselves, the unantici-pated special circumstances of the respondent,the different layers of meaning which arevariously revealed or cloaked by answers andso on will be unavailable to the remoteresearcher whose co-constitutive role is ham-pered by the lack of direct interaction withthe respondent. Naturally, however, suchremoteness of the research relationshipappeals to researchers who are seeking toobjectify their research methods.To pose thesame questions in the same way to all surveyrespondents can be positioned tactically as amanoeuvre which screens out the perceivedsubjectivity, or ‘bias’, of the interviewer.

Thirdly, we can consider more interactiverelationships (for example, those which occurduring semi-structured face-to-face encoun-ters). Here, data are co-constructed as inter-viewer and interviewee work their waythrough questions which begin as the ‘prop-erty’ of the researcher but which become co-owned and co-shaped in the unfoldinginteractivity of questioning, answering, listen-ing and conversing. Here, the original schemeof intended data construction can be divertedor even subverted by both the researcher andthe researched: the former as he or she followsup what appear to be interesting conversa-tional angles; and the latter as he or shedemarcates consciously or subconsciously theboundaries of what he or she will reveal inthe interview. Interactive research relation-ships will differ markedly according to thepower relations involved – interaction with a

high-flying business executive, for example,will be very different from that with a home-less person or a child. Interactivity, then, isreplete with social, political and ethical differ-ences, and the resultant co-construction ofdata will reflect these issues.

Fourthly, research relationships can beinvolved as in the use of reflective interviews ordiscussion groups. In these situations, theresearcher is yet further immersed in the socialsituation of conversation, leading to a higherprofile and potential impact for the subjectiv-ities of both the researcher and the researched.In these contexts the notion of getting ‘true’answers to ‘straightforward’ questions isincreasingly replaced by the potential for dia-logue, to sponsor an unfolding performance inwhich roles and topics are fashioned in thegive-and-take of the event. Involved researchrelationships represent a deliberate departurefrom ideas about neutrality and observational‘distance’ so as to avoid treating people likeobjects and to find ways of treating people likepeople.The search for essentialist truth is thusreplaced by a desire to see the world throughdifferent windows and to hear the world via apolyphony of different voices (Crang, 1992).

In these kinds of ways, the judgement ofhow appropriate a particular approach is to adefined question will depend on the role of theresearcher, as well as of the researched, in con-structing data.A number of underlying episte-mological issues will be crucial in thisjudgement.We take the view that all knowledgeis both situated and struggled over (Smith, S.J.,2001). The researcher’s own positionality – hisor her subjectivity and positioning – will rep-resent a significant contextualization ofhis or her role in co-constructing and theninterpreting interview data. However, as Rose(1997a) has pointed out, we need to be awarethat we cannot ever fully recognize or repre-sent our own positionality. So while it isimportant to engage with the values and sub-jectivities of the researcher, because these will

TALKING TO PEOPLE

129

3151-Ch-05.qxd 3/2/04 2:43 PM Page 129

be important factors in the co-construction ofknowledges in research encounters, it is alsoimportant to realize the limits to knowing sub-jectivity. Equally, the intersubjective nature of theresearch encounter is a necessary reflection ofboth the researcher as co-constituent of result-ing knowledges and the researcher becoming(however briefly) part of the life of theresearched. Intersubjective encounters can bedesigned as a means to understand the life-worlds, experiences and meanings of theresearched (Buttimer, 1976), but the conse-quences of such encounters can never be fullypredicted or known. Finally, the power relationsof the encounter are crucial. Although therewill be times when power is exercised uponthe researcher (as with the high-flying businessexecutive mentioned above), it is far morecommon for power to be exercised by theresearcher in his or her investigation of, andinterpretation of, the lives of others.The extentto which we as researchers are able to representother lives is a crucial and tricky epistemolog-ical question, given that all such representationswill be in some senses fictional in that they gobeyond actual utterances of the researchencounter (see Bennett and Shurmer-Smith,2001).The choice of appropriate approaches to‘talking to people’ will, therefore, necessitatecritical appraisal of these epistemological issueswhich underpin the relationship betweenresearcher and researched.

Questionnairing

We begin our discussion at that end of therange of interviewing methods in which theresearcher is most remote from his or herrespondent. Johnston (2000b: 668) defines thequestionnaire as an instrument of data con-struction ‘comprising a carefully structuredand ordered set of questions designed toobtain the needed information without eitherambiguity or bias… Every respondent

answers the same questions, asked in the sameway and in the same sequence’. As such, ques-tionnaires are an essential tool of social surveyanalysis, potentially delivering research datawhich represent particular attributes of thesample. Questionnaires allow researchers tocount up differing kinds of responses to ques-tions, particularly where the questions are‘closed’ (that is, referring to a fixed range ofpotential answers).This survey technique alsoallows the counting up of scorings from ques-tions involving sematic differentials (such asvery good, good, satisfactory, bad, very bad).The practices of counting up producenumeric measurements of what people thinkand how they behave, alongside informationabout their gender, age, occupation and so on.This information can then be cross-tabulatedand used to make quantifiable inferencesabout the wider sample from which the sampleis drawn.

Questionnaire surveys are a familiar fact oflife in the developed world. Not only dohouseholds have the regular state census tocontend with, but people walking down thehigh street will also often be accosted by sur-veyors requesting information for politicalopinion polls or market research, and a knockat the front door can mean an encounter withquestionnaire surveys used by anyone fromthe Jehovah’s Witnesses to human geographyresearchers. This very familiarity has in turnled some people to develop their own respon-sive strategies, such as a polite but quickrefusal to participate – perhaps a reflexiveresponse of ‘survey fatigue’ to the intrusivenature of surveys in society.

Overfamiliarity with the idea of a ques-tionnaire survey can also pervade the practiceof research in human geography. Questionnairesurveys have become part of the standard bag-gage of human geography researchers, beingused in the main to produce quantitativeinformation. Indeed, they have become sostandard that there is sometimes a sense in


130

3151-Ch-05.qxd 3/2/04 2:43 PM Page 130

which the thought ‘I’ll do a questionnaire’often precedes the arrival at clearly thought-out research questions, and research cansometimes, therefore, be reduced to a ques-tion of ‘what can I do with this informationfrom the questionnaire?’ (see Bridge, 1992).This is not to suggest that the questionnairecannot perform a perfectly legitimate role inpractising human geography, either as a solemethod of constructing data or in tandemwith other methods. Indeed there are manyrecent instances in which combinations ofdifferent data constructions are positivelyadvocated. For example, Jayaratne (1993: 109),writing on issues of methodology in feministresearch, states:

I also advocate the use of qualitative data,in conjunction with quantitative data todevelop, support and explicate theory. Myapproach to this issue is political; that is, Ibelieve the appropriate use of both quanti-tative methods and qualitative methods inthe social sciences can help the feministcommunity in achieving its goals moreeffectively than the use of either qualitativeor quantitative methods alone.

The important guiding principle in consider-ing a questionnaire survey is thereforewhether it is the most appropriate means ofconstructing data with which to address aparticular research issue. Such a judgementwill in turn necessitate a recognition of boththe technical capabilities and restrictions ofquestionnaires, as well as the potential forvariations in data construction because ofunspoken assumptions in the questions asked,difficult-to-interpret differences in theanswers given and the potentially differentoutcomes of questionnaires in different socialsituations.

When is a questionnaireappropriate?

A very substantial literature exists on surveyresearch using questionnaire techniques.1

Issues of what form a questionnaire shouldtake, what questions should be included andhow the survey should be implemented arethe subject of exhaustive commentary.Much less, however, is said about when aquestionnaire is or is not appropriate. Forsome, the questionnaire survey is an integralpart of the application of scientific methodto social science, and its function is thereforeone of permitting enumeration to take place(see Chapter 8). Thus Oppenheim (1992:101) maintains that ‘the detailed specifica-tion of measurement aims must be preciselyand logically related to the overall planand objectives. For each issue or topic tobe investigated, and for each hypothesis tobe explored, a precise operational state-ment is required about the variables to bemeasured’.

In cases where questions are used as vehi-cles for ‘scientific’ hypothesis-testing, prob-lems of obtaining a statistically representativesample will often be paramount (see Chapter 8).For others, the usefulness of a questionnairesurvey will be less tied to a research frame-work of hypothesis-testing. Here, the empha-sis is on the requirement for a highlystructured research technique which can beused to construct a body of information con-sisting of the answers of a large sample ofrespondents to a consistent series of questions.As such it can form part of Sayer’s extensiveresearch strategy (see Table 5.1) although amore flexible mix of structured and less struc-tured information can be constructed duringone face-to-face interview, and so intensiveinformation may also be available using thesemethods. Perhaps Bell’s (1987: 58) advice ismost appropriate here:

Ask yourself whether a questionnaire is likelyto be a better way of collecting informationthan interviews or observation, for example.If it is, then you will need to ensure you pro-duce a well-designed questionnaire thatwill give you the information you need, thatwill be acceptable to your subjects and

TALKING TO PEOPLE

131

3151-Ch-05.qxd 3/2/04 2:43 PM Page 131

which will give you no problems at theanalysis and interpretation stage.

Questionnaires, then, can be useful in a num-ber of situations: where you are interested inconstructing data based on the responses of alarge number of people; where you have onlytime-constrained, face-to-face access to yourrespondents and therefore only limited inter-action is possible; where you are unable forone of a number of reasons to gain face-to-face access to your potential respondents andtherefore cannot interview them personally;where you require data to feed into a precon-ceived statistical form of interpretation; wherethe presentation of your research findings willachieve greater impact (for example as withsome government agencies) if they are basedon a recognizable (and politically acceptable)questionnaire survey methodology; and so on.

What form can aquestionnaire take?

There is a range of potential forms of ques-tionnaire which should be carefully consid-ered before deciding on a survey researchstrategy, although these boil down to threebasic formats:

1. The questionnaire which is posted orhanded to the respondent, who completesthe survey and then posts or hands it backto the researcher (contemporary use ofInternet or email surveys also conform tothis format).

2. The questionnaire which is carried outover the telephone, with the researcherasking questions and recording answers.

3. The questionnaire which is carried outface to face, in which the researcher againasks questions and records the answers.

These formats can be mixed, however.Respondents can be asked to read and recordanswers to particularly sensitive areas of

questionnaires during face-to-face surveys;combinations of telephone and face-to-faceinterviews might be necessitated in circum-stances where incomplete telephone numberinformation is available; and respondentsmight be asked to complete and return asecond element of a survey having completeda face-to-face questionnaire – this sometimeshappens when a ‘diary’ of activities and trips isrequested for the following week or month.

It is possible to outline some typical prosand cons for these different survey formats(see Table 5.2), although the various areas ofadvantage and disadvantage are more arguablethan this kind of tabulation suggests.To take aglaring example, the response rates frompostal questionnaires are often thought to bemuch lower than those conducted face to faceor by telephone. For example, Bridge (1992:198) notes that ‘response rates of 30–40 percent from a survey of residents in an ordinaryneighbourhood is considered good’. Perhapsless obvious but equally important, given ourcomments at the beginning of this chapterabout interviews as sets of relationshipsbetween the researcher and the respondent, isthe suggestion in the table that the quality ofanswers is problematic in face-to-face surveys.The assumption here is that questionnairesallow researchers to avoid some of the ‘distor-tions’ which are seemingly the inevitableproduct of the face-to-face social interactioninvolved in less structured interviews. These‘distortions’, so-called, represent the fear thatin an interview questions can be asked in a‘leading’ way, with the likelihood that therewill be too many forceful prompts and toofrequent misunderstandings of responses.Questionnaires, on the other hand, are some-times thought to be able to provide consis-tent, representative and bias-free answersacross a large sample.

This question of bias is an interesting one.One approach is to create a ‘bias-free’ ques-tionnaire by avoiding issues relating to attitudes,


132

3151-Ch-05.qxd 3/2/04 2:43 PM Page 132

beliefs or prejudices altogether, therebyexcluding respondents’ biases. Such a strategymay be suitable in some situations wherestraightforward factual data are required on anextensive scale.Another, and often more fruit-ful, approach may be to use questionnairesprecisely to tease out these abilities, beliefsand prejudices. In many contexts researchersdo want to know about the ‘biases’ of theirrespondents and, by restricting a questionnaireto what seems like objective, factual accountsof human geographical situations, the poten-tial richness of the encounter may be lost. It is

this perspective which often informs thechoice between remotely delivered question-naires and undertaking face-to-face encoun-ters in which the structured questioning ofthe questionnaire can be mixed with, or evenreplaced by, less structured interview tech-niques. Face-to-face encounters with respon-dents will greatly enhance the researcher’sability to interpret some of the context inwhich questionnaire answers are set. What islost in the consistency of questioning neces-sary for hypothesis-testing may represent asignificant gain in the questionnaire as a

TALKING TO PEOPLE

133

Table 5.2 Advantages and disadvantages of face-to-face, telephone and mailmethods of administering questions

Face to face Telephone Mail

Response ratesGeneral samples Good Good GoodSpecialized samples Good Good Good

Representative samplesAvoidance of refusal bias Good Good PoorControl over who completes the Good Satisfactory GoodquestionnaireGaining access to the selected Satisfactory Good GoodpersonLocating the selected person Satisfactory Good Good

Effects on questionnaire designAbility to handle:Long questionnaires Good Satisfactory SatisfactoryComplex questions Good Poor SatisfactoryBoring questions Good Satisfactory PoorItem non-response Good Good SatisfactoryFilter questions Good Good SatisfactoryQuestion sequence control Good Good PoorOpen-ended questions Good Good Poor

Quality of answersMinimize social desirability Poor Satisfactory GoodresponsesAbility to avoid distortion due to:Interviewer characteristics Poor Satisfactory GoodInterviewer’s opinions Satisfactory Satisfactory GoodInfluence of other people Satisfactory Good PoorAllows opportunities to consult Satisfactory Poor GoodAvoids subversion Poor Satisfactory Good

Implementing the surveyEase of finding suitable staff Poor Good GoodSpeed Poor Good SatisfactoryCost Poor Satisfactory Good

Source: de Vaus (1991: 113), adapted from Dillman (1978)

3151-Ch-05.qxd 3/2/04 2:43 PM Page 133

hybrid interview method which is able tocombine extensive information with someintensive insight.

Thus far it has been implicit in our discus-sion that the questionnaire is carried outbetween one researcher and a series of indi-vidual respondents.This can frequently not bethe case. For example it is possible to admin-ister a questionnaire to a group of people atthe same time, such as a class of schoolchildren or people attending a public meet-ing, although the use of focus groups isoften more appropriate in these situations.Sometimes group questionnaires permitvisual texts – video, film clips, slides, photo-graphs, etc. – to form the basis of questions,or the questionnaire can be completed sub-sequent to a particular ‘live’ debate (forexample, between politicians or between localcouncil representatives and local peopleprotesting over a proposed development).Considerable care is required in these con-texts because respondents will tend to talk toeach other about questions, ask the inter-viewer questions about how they shouldanswer and even copy each other’s answers.Again, the use of focus groups could be amore appropriate choice in such situations.Perhaps a more common form of ‘group’questionnaire is when the researcher requiresinformation about a ‘household’. Often onemember of the household answers on behalfof others, giving a particular view of thelifestyles and opinions of others which wouldbe different if the person concerned wereanswering. For example, if questions are askedof the ‘head of the household’ who will oftenbe assumed to be male, the responses arehighly likely to involve gender-biased repre-sentations of that household. It is better(although much more difficult to organize) toundertake the questionnaire when allmembers of the household are present tospeak for themselves. A similar dilemma(though often impossible to resolve) is where

one individual responds to a questionnaire onbehalf of an agency. The difference betweenindividual and ‘corporate’ attitudes in thesecases is often very difficult to discern, andmuch will depend on the status and role ofthe individual within the agency. In additionthere may be crucial differences betweenagency information for external consumption,which often constitutes a well-rehearsedpublic relations narrative, and informationwhich reveals some of the (often political)debates going on inside the agency.

Constructing thequestionnaire

Those people who have experience of usingquestionnaire methods will almost invariablyconclude that it is enormously difficult toconstruct a good series of questions. Wewould suggest that the questionnaire survey isby no means the ‘easy option’ that it is some-times assumed to be by human geographersmaking their methodological choices. Wewould reiterate that the key factor here is tothink through very carefully the preciseresearch questions that are to be focused onand, if a questionnaire survey seems appropri-ate, to apply specific parts of the questionnaireschedule to specific elements of the overallresearch questions. The idea will be to askpurposeful questions in as unambiguous amanner as possible, with a view to providinguseful information of a type suitable for theinterpretative strategy you will have alreadychosen. The literature on questionnaire sur-veys abounds with illustrations of goodpractice – indeed,whole books have been writ-ten on the subject of how to construct ques-tions (see Foddy, 1993). It will often be usefulto supplement a reading of these manuals ofgood practice with scrutiny of questionnairesthat have actually been used by otherresearchers. For example, Forbes (1988: 112)writes candidly about how he got together a


134

3151-Ch-05.qxd 3/2/04 2:43 PM Page 134

questionnaire of informal economic activityin urban districts (kampungs) of Indonesia:

Socio-economic surveys were the mainstayof the work being done on the informalsector at the time, due in large part to theinadequacy of official data sources when itcame to the marginal groups in the informalsector. I found a number of interview sched-ules that had been used in Indonesia forkampung surveys, informal sector surveysand migration surveys, and drafted out aschedule of nearly 50 questions.

A good questionnaire will often be a combi-nation of previously used and innovativequestions which are closely focused on theoverall aims of the research.

The types of question employed willdepend very much on what constructions ofdata will be suitable for the interpretativestrategies to be used. If interpretation is to bein numeric form using descriptive or otherstatistics (see Chapter 8), questions will nor-mally be closely structured in a ‘closed’ form.Bell (1987), following Youngman (1986), sug-gests six types of questions which can be usedto construct structured data (see Box 5.1). If,however, the chosen interpretative strategypermits the construction of written text asdata, a seventh more open-ended type ofquestion can be used. When using question-naires for the first time, it is often tempting todeploy a large number of open questionswithout having thought through how theresultant chunks of texts can be used effec-tively in providing answers to the big researchquestions which are the focus of the project.These passages of text will then need to be‘coded’ and interpreted (see Chapter 10).Open questions can bear much interpretativefruit, however, and may open up some fasci-nating lines of intensive inquiry in situationseither where extensive research is the mainpriority (thus necessitating questionnairerather than less formal interviews) or wherean ethnographic approach is not possible forone of a variety of reasons. An example of a

research report where qualitative text(constructed by using open-ended questions) hasbeen used in conjunction with more numerictabulated findings is the report on ‘Rurallifestyles in England’ (see Cloke et al., 1994).Here the research was designed in part toreplicate a previous survey of rural householdsin England to investigate changes in housing,employment, income, accessibility, serviceprovision and other key factors of rural life.The methodological constraints of having touse an extended version of a previous surveyinstrument were partially freed up by openingout spaces for respondents to add commen-tary on aspects of their lifestyles. The result,though imperfect when measured against thedemands of rigorous qualitative research, didat least allow the ‘voices’ of rural people to beincluded in the published output.

Having decided on the types of questionrequired, it will be necessary to give someconsiderable thought to the precise wordingof particular questions. It is all too easy forquestions to be ambiguous, imprecise orreplete with inherent assumptions. Althoughpiloting the questionnaire (see below) willhelp to expose any difficulties with wording,it is worth a considerable amount of prepara-tory time in sorting and sifting particularquestions. Oppenheim (1992) provides 13basic rules for the careful framing of questions(Box 5.2).To these we would add the warn-ing to be fastidious in ensuring that the lan-guage used in questions does not makeunacceptable assumptions about gender, sexu-ality, ethnicity, age, localism, religious faith andthe like.

These rules will not guarantee a successfulquestionnaire but, given a well-thought-outseries of research aims and an awareness of theinterpretative manoeuvres to be carried outon the constructed data, the wording of ques-tions will be important. Many students whodo give due care and attention to the pitfallsof questionnaire technique become somewhat

TALKING TO PEOPLE

135

3151-Ch-05.qxd 3/2/04 2:43 PM Page 135

Box 5.1 Different types of questions

More structured questionsList: A list of items is offered, any of which may be selected. For example, a questionmay ask about qualifications and the respondent may have several of the qualificationslisted.

Category: The response is one only of a given set of categories. For example, if age cat-egories are provided (20–29, 30–39, etc.), the respondent can fit into only one category.

Ranking: In ranking questions, the respondent is asked to place something in rankorder. For example, the respondent might be asked to place qualities or characteristicsin order (e.g. punctuality, accuracy, pleasant manner). Number ‘1’ usually indicates thehighest priority.

Scale: There are various stages of scaling devices (nominal, ordinal, interval, ratio)which may be used in questionnaires, but they require careful handling.

Grid: A table or grid is provided to record answers to two or more questions at thesame time. For example, ‘How many years have you taught in the following types ofschool?’ Here there are two dimensions – type of school and number of years.

1 or 2 years 3 or 4 years 5 or 6 years More than 6 years

Grammar

Comprehensive

Middle

Independent

Other (please specify)

Less structured questionsVerbal or open: The expected response is a word, a phrase or an extended comment.Responses to verbal questions can produce useful information but analysis canpresent problems. Some form of content analysis is required for verbal materialunless the information obtained is to be used for special purposes. For example, youmight feel it necessary to give respondents the opportunity to give their own viewson the topic being researched – or to raise a grievance. You might wish to use ver-bal questions as an introduction to a follow-up interview or in pilot interviewswhere it is important to know which aspects of the topic are of particular impor-tance to the respondents.

Source: Bell (1987: 59–60), following Youngman (1986)


136

3151-Ch-05.qxd 3/2/04 2:43 PM Page 136

Box 5.2 Oppenheim’s rules for question design

TALKING TO PEOPLE

137

Length: Questions should not be too long; they should not contain sentences of morethan, say, 20 words.

Avoid double-barrelled questions: ‘Do you own a bicycle or a motorbike?’ In suchinstances the respondents are left in a quandary if they want to say ‘yes’ to one partof the question but ‘no’ to the other.

Avoid proverbs and other popular sayings, especially when measuring attitudes, forsuch sayings tend to provoke unthinking agreement. Instead, try to make the respon-dent think afresh about the issue by putting it in other words.

Avoid double negatives: For example, ‘Cremation of the dead should not be allowed’,followed by agree/disagree. Respondents who favour cremation have to engage in adouble negative (that is, they do not agree that cremation should not be allowed). Tryto put the issue in a positive way.

Don’t know and not applicable: Categories are too often left out, both in factual andin non-factual questions. It has been argued that some people give a ‘don’t know’response in order to avoid thinking or committing themselves, but do we really wantto obtain ‘forced’ responses which are virtually meaningless?

Use simple words, avoid acronyms, abbreviations, jargon and technical terms – or elseexplain them. When in doubt, use pilot work.

Beware the dangers of alternative usage: This applies to even the simplest words which‘could not possibly be misunderstood’! To some people, the word ‘dinner’ indicates anevening meal but to others it means a cooked meal, as in ‘Usually I give him his dinnerat lunch time’ [sic].

Some words are notorious for their ambiguity and are best avoided or else defined – forexample, the words ‘you’ and ‘have’ as in ‘do you have a car?’. Other notoriouslyambiguous words are ‘family’, ‘bought’, and ‘neighbourhood’.

All closed questions should start their lives as open ones so that the answer categorieswill be based as much as possible on pilot work. Even then, always consider the inclu-sion of an ‘Other (please specify)’ category.

Beware leading questions: ‘When did you last borrow a video tape?’ This assumes(1) that all respondents have access to a videotape player; (2) that they generally bor-row tapes; and (3) that it is not done by someone else on their behalf.

Beware loaded words such as democratic, free, healthy, natural, regular, unfaithful,modern.

3151-Ch-05.qxd 3/2/04 2:43 PM Page 137

Box 5.2 (Continued)

Don’t overtax the respondents’ memories by asking questions such as ‘in the past 13weeks, which of the following branded goods have you purchased?’ For recurrentitems of behaviour it is risky to go back more than a few days. For more importantevents such as major acquisitions, periods in hospital or being mugged, the respondentfirst has to be taken through the preceding few months with the aid of ‘anchoringdates’ such as public holidays, family birthdays, school holidays and the like, followedby an attempt to locate the event(s) in question in an individual time frame.

Pay due attention to detail such as layout, show cards and prompts (note especially thedanger of overlapping categories, as with ‘age 25–30, 30–35’), the wording of probes(follow-on questions), routing or skip-to instructions and even question numbering.

Source: Abridged and adopted from Oppenheim (1992: 128–30)

overwhelmed at the prospect of meeting allthese various requirements. With care, how-ever, a questionnaire can be constructed andimplemented effectively. Figure 5.2 shows aquestionnaire which was designed and carriedout by a small group of first-year geographystudents at Lampeter. Their research aimsrelated to the diversification of agriculturalenterprise and, although they would not wanttheir work to be represented as any kind ofmodel of best practice, it does indicate that asimple questionnaire can be put together rathernicely so as to construct useful data with whichto discuss manageable research issues.

Thus far we have presented only a rathersimple account of the kinds of questions thatcan be asked in a questionnaire. In practicethere is a vast array of possible devices whichcan be used in this context, and there is plentyof scope for innovation in questionnairework. To illustrate some of these differentavenues we will briefly discuss the construc-tion of data on attitudes using questionnairetechniques. The instinctive reaction whenasking someone about his or her attitudetowards a particular event, circumstance orcharacteristic of his or her lifestyle is simply toask ‘what do you think …’. Indeed this may

be a productive method of constructingopen-ended qualitative text on a particularissue, and delivered consistently it may alsoproduce attitude statements capable of com-parison. There are, however, other ways ofapproaching this task. For example, thoseresearchers seeking numerical data for statisticalinterpretation will view attitude statements asmere raw materials for attitude-scaling tech-niques. In a uni-dimensional form, attitudescaling often means asking respondents toindicate their feeling about something on anumeric scale, say between 1 and 5, where 1indicates strong agreement and 5 indicatesstrong disagreement. Such scales can also beused to indicate a semantic differential. Hereparticular adjectives – for example, ‘periph-eral’ and ‘core’ – can be chosen to representthe two extremes of the continuum andrespondents are asked to mark their attitudeon a scale between the two extremes.2

Attitude scales can also, however, beconstructed in a multidimensional form.According to Oppenheim (1992: 174, empha-sis in original), this requires that attitude scales

should be phrased so that respondents canagree or disagree with the statement.Usually, after the pilot work, a large number


138

3151-Ch-05.qxd 3/2/04 2:43 PM Page 138

TALKING TO PEOPLE

139

Please note: ALL INFORMATION VOLUNTEERED WILL BE USED IN THE STRICTESTCONFIDENCE.

Instructions: Please tick boxes where relevant. Where you are asked to specify your answer, pleaseuse a brief statement or sentence. Sufficient room has been left for the questions that may requireslightly longer answers.

1 Name: . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 Address: . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3 Size of farm/holding (please tick relevant box)

SMALL � MEDIUM � LARGE �

(70–299 acres) (300–499 acres) (500–800 acres)

OTHER � (please specify) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

4 LENGTH OF TENANCY/OWNERSHIP (years or months) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

5 (Please complete following table)

FAMILY MEMBERS RELATIONSHIP TO 1 AGE OCCUPATION(Living or working at farm)

1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

OTHER (please specify) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

6 TYPE OF FARMING ACTIVITIES (please tick relevant boxes)

DAIRY � HORTICULTURE �

ROUGH PASTURE � CROPS �

(GRAZING)STUD FACILITIES �

LIVESTOCK � SHEEP � PIGS � COWS �

POULTRY � HORSES � GOATS �


7 (Please complete following table)

EMPLOYEES DUTIES FULL/PART TIMEor SEASONAL

1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .


8 HAVE YOUR FARMING ACTIVITIES CHANGED DURING YOUR TENANCY/OWNERSHIP OFTHE FARM? (please tick relevant box)

YES � NO �

(Continued)

3151-Ch-05.qxd 3/2/04 2:43 PM Page 139


140

b IF YES, HAS THIS INVOLVED DIVERSIFICATION? (please tick relevant box)

YES � NO �

9 IF YES TO QUESTION 8b, PLEASE BRIEFLY DESCRIBE THE FORM(S) OF DIVERSIFICATION(you may want to use the list below as a reference when answering this question)

ACCOMMODATION AND CATERING ANSWER: . . . . . . . . . . . . . . . . . . . . . . . . . . . .(e.g. bed and breakfast) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .AGRICULTURAL SERVICES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .(hay, cutting and baling, . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .machine rental, etc.) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .EQUINE (e.g. pony trekking, . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .stud, livery, etc.) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .CROP-BASED PRODUCTION AND . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .PROCESSING (e.g. processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .of foodstuffs, animal feed) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .TRADING (e.g. farm shop, . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .retailing, etc.) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .MANUFACTURING (e.g. craft . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .goods, etc.) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .FORESTRY (e.g. timber . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .processing, rental of land for . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .plantations, etc.) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .BREEDING (e.g. rare breeds, . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .livestock, etc.) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .MISCELLANEOUS (e.g. land- . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .building facilities, landscape . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .services, pet care, vehicle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .repair, etc.) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

b HOW LONG AGO DID YOU DIVERSIFY? (years or months) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

10 DO YOU RENT and/or LEASE OUT ANY PART OF YOUR LANDYES � NO � (please tick relevant boxes)

b IF YES (please specify) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

11 PLEASE COULD YOU DESCRIBE, IN YOUR OWN WORDS, WHY YOU DIVERSIFIED?(you may want to use the list below as a reference when answerinng this question)

To raise or maintain farm or ANSWER: . . . . . . . . . . . . . . . . . . . . . . . . . . . .family income . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .To utilize available human . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .resource, and/or hired labour, . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .land, buildings or machinery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Personal interest . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Utilization of spare capital . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Market opportunity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .To gain a competitive . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .advantage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Tax incentives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Introduction of government . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .legislation (e.g. quotas) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

(Continued)

3151-Ch-05.qxd 3/2/04 2:43 PM Page 140

TALKING TO PEOPLE

141

12 DID YOU RECEIVE ANY (please tick relevant boxes)

ADVICE ASSISTANCE GRANT RESISTANCE

FROM THE FOLLOWING:

NATIONAL PARK AUTHORITY � � � �

MINISTRY OF AGRICULTURE,FISHERIES AND FOOD � � � �

NATIONAL FARMERS UNION � � � �

DEVON COUNTY COUNCIL � � � �

YOUR DISTRICT COUNCIL � � � �

BANKS/FINANCIALINSTITUTES � � � �

COUNTRYSIDE COMMISSION; � � � �

FORESTRY COMMISSION � � � �

PUBLICATIONS � � � �


b WOULD YOU LIKE TO EXPAND ON YOUR ANSWER TO QUESTION 12?(e.g. the advice, assistance, grant received, the resistance met, etc.)

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

13 HOW DID YOU FIND OUT ABOUT THE DIFFERENT SCHEMES, OR INFORMATION,OFFERED BY VARIOUS ORGANIZATIONS? (please identify the organizations in your answer)

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

14 DID YOU APPROACH ANY OF THESE ORGANIZATIOINS (or the ones you have specified)?

YES � NO � (please tick relevant box)

15 IF YES TO QUESTION 14. WHY DID YOU FINALLY APPROACH THESE ORGANIZATIONS?(please identify the organization in your answer)

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

16 IN YOUR OWN WORDS, HOW WOULD YOU EXPRESS, OR DESCRIBE, YOUR LEVEL OFACHIEVEMENT IN DIVERSIFICATION?

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

17 WHAT FACTORS HAVE HAD AN AFFECT ON YOUR DIVERSIFICATION? (you may want to usethe list below as a reference when answering this question)

Changes in the market ANSWER: . . . . . . . . . . . . . . . . . . . . . . . . . . . .Lack of market growth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Labour difficulties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Lack of finance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

(Continued)

3151-Ch-05.qxd 3/2/04 2:43 PM Page 141

of attitude statements (probably somewherebetween sixty and 100) is assembled to forman item pool. These items will be analysedand submitted to a scaling procedure … Ifall goes well, the outcome will be a muchsmaller selection of items (say, between oneand two dozen) which together have theproperties of an attitude scale that can yielda numerical score for each respondent. Theanalysis will also have thrown light on thetopic or dimension which the scale isdesigned to reassure.

In its simplest multidimensional form, then,attitude scaling will consist of a group ofquestions for which the respondent willreceive a score for his or her answer. Figure 5.3illustrates a set of questions on the subject of

traditionalism. The score for each answerreflects its association with the particular atti-tude being measured, and the total score willtherefore be used to reflect the respondent’s‘position’ on the abstract continuum of opin-ion on this issue.Thus a high score in Figure 5.3is held to denote a very highly traditionalistattitude, and vice versa. Much more complexstatistical manipulations can be undertaken,often using factor analysis to produce a math-ematical value for how well each question‘loads on to’ the overall mathematical valuefor the attitude in question.3

As with many statistical analyses, thesemultidimensional scaling techniques take us


142

Operating difficulties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Inappropriate support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .from external sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

18 HAS THE NATURE OF YOUR PRIMARY FARMING ACTIVITY CHANGED DUE TODIVERSIFICATION?

YES � NO � (please tick relevant box)

b IF YES, WOULD YOU LIKE TO EXPAND ON YOUR ANSWER?

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

19 WOULD YOU LIKE TO ADD ANY FURTHER COMMENT, OR INFORMATION, ON ANYRELEVANT ISSUES, OR FACTORS, YOU BELIEVE WE HAVE, OR HAVE NOT, COVERED IN OURQUESTIONNAIRE?

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

THANK YOU FOR YOUR CO-OPERATION IN COMPLETING THIS QUESTIONNAIRE. IF YOUWOULD BE WILLING TO PARTICIPATE IN A SHORT (15–30 MIN.) INTERVIEW AT A LATER DATE,PLEASE LET US KOW BY TICKING THE RELEVANT BOX.

YES � NO � TO INTERVIEW

Figure 5.2 A postal questionnaire to farmers on diversification of agricultural enterprise

Source: Clare Bover, Mike Gadd, Pam Hovell, Karen Inston and Pauline MacKenzie

3151-Ch-05.qxd 3/2/04 2:43 PM Page 142

further and further away from the actualsource of the attitude – that is, the respon-dent. In practice it is extremely difficult toconnect a series of different questions withone specific attitude because, typically, theanswers given to questions represent a con-fused conflation of a number of different atti-tudes. By the time such an analysis reaches afactor analysis it is difficult to know what theover-riding mathematical factor actually rep-resents. Many researchers, then, have choseneither to stick with simple open-ended ques-tions and with uni-dimensional scales, or toseek out other ways of dealing with questions

about attitudes. These range from questionswhich invite respondents to complete a sen-tence in their own words to the use ofvignettes and projective techniques.Vignettestend to be short stories in which particularcircumstances are specified within a particularstandardized context (Box 5.3 illustrates atypical vignette). Respondents can then beasked simple questions such as ‘what shouldthey do?’ and ‘why?’, and attitudes towardsthese kinds of circumstances could then beinterpreted from their answers. Projectivetechniques rely on a spontaneous interpretationby the respondent of a written, verbal or

TALKING TO PEOPLE

143

Strongly Stronglyagree Agree Undecided Disagree disagree

a If you start trying tochange things very [ ] [ ] [ ] [ ] [ ]much you usually 4 3 2 1 0make them worse.

b If something grows upover a long period,there will always be [ ] [ ] [ ] [ ] [ ]much wisdom in it.

c It is better to stick bywhat you have than to [ ] [ ] [ ] [ ] [ ]try new things you donot really know about.

d We must respect thework of our forebears [ ] [ ] [ ] [ ] [ ]and not think that weknow better than theydid.

e A person does notreally have much [ ] [ ] [ ] [ ] [ ]wisdom until he is wellalong in years.

Person Item Scale score Interpretationa b c d e

1 4 + 4 + 4 + 4 + 4 = 20 Very highly traditionalist

2 4 2 1 0 3 = 10 Moderate3 2 1 2 1 1 = 7 Moderate4 0 0 0 0 0 = 0 Very non-traditional

Figure 5.3 Multidimensional attitude scaling: an illustration

Source: de Vaus (1991: 251)

3151-Ch-05.qxd 3/2/04 2:43 PM Page 143

Box 5.3 Using a vignette

John and Barbara Dixon are both in their mid-forties and their children now have jobsoverseas. Barbara is a research scientist in a medical laboratory and John is an accoun-tant. They own their own home and are comfortably off. [JOHN/BARBARA] (A) is offereda promotion that would mean [A LARGE RISE IN INCOME/ONLY A SLIGHT INCOME RISE BUT A MUCH MORE

ENJOYABLE JOB] (B) but would also mean that they would have to sell up and move sev-eral hundred kilometres. [BARBARA/JOHN] could also get a job in the new location [BUT IT

IS NOWHERE NEAR AS GOOD AS THEIR PREVIOUS/JOB WHICH IS VERY SIMILAR TO THEIR CURRENT JOB](C). [JOHN/BARBARA] would like to make the move but [JOHN/BARBARA] does not want togive up his/her current job. [The references to John and Barbara in the last sentencewould be varied so that the person who is offered the new job wants to move whiletheir partner does not want to move.]

Source: de Vaus (1991: 91)

visual stimulus. This can involve asking for aresponse to a calculated statement on a parti-cular issue or showing respondents a series ofvisual images, often on cards which representhuman figures in various situations thatrespondents are asked to describe and inter-pret in the form of a story. Saarinen (1966)used one such ‘thematic appreciation test’ inhis study of farmers’ attitudes to drought haz-ards in the Great Plains of the USA. Oncepopular, many of these techniques havebecome largely disused because of the diffi-culties experienced by researchers in inter-preting the responses given in such a way as tobe relevant to the research question they wereinvestigating, although they may well enjoya revival now that the inherent difficulties ofinterpreting any form of data are being morewidely acknowledged. As with all the pro-cedures discussed in this section, the decisionstaken about constructing a questionnaireshould be on the basis of logical and practicalinterconnections between research question,data construction and data interpretation.Employing an innovative technique which isonly loosely connected to your research ques-tion and which constructs uninterpretabledata will not be a fruitful way of practisinghuman geography!

Carrying out thequestionnaire

Unless the questionnaire is directed to allmembers of a particular target group, one earlydecision to be made in carrying out a ques-tionnaire will be the selection of a particularsample of people to interview. Again it will beimportant here to use a method of samplingwhich is in tune with the interpretation strat-egy for the research. For example, some statisti-cal analyses make clear assumptions about thecharacteristics of the sample from whichnumerical information has been drawn. Inthese cases probability samples – that is, thosewhere each person in the population has anequal, or at least measurable, chance of beingchosen as a potential respondent – will bepreferable. On the other hand non-probabilitysamples are those where some people have ahigher chance (although not a measurable one)of being chosen and may be appropriate inother circumstances.There are four main typesof probability sample (Box 5.4) and three maintypes of non-probability sample. Selection fromthese different sampling methods will verymuch depend on the necessity or otherwise fora statistically representative or otherwise indica-tive sample of interviewees. A ‘representative’


144

3151-Ch-05.qxd 3/2/04 2:43 PM Page 144

Box 5.4 Sampling methods

Probability samples

1 Random sample: Each potential interviewee is given a number and then therequired sample size is drawn from the list using random number tables. This sam-pling method is best where good information exists about the total populationconcerned, and where that population is concentrated geographically.

2 Systematic sample: Interviews are drawn at regular intervals from a population. Ifa 10% sample is required, every 10th person from the list of potential intervieweesis selected. This method shares the disadvantages of the random sample but, inaddition, exposes the survey to the risk that the list of potential interviewees maycontain a certain kind of person at regular intervals. For example, if a membershiplist of a pressure group consisted of male and female partners listed together, a50% systematic sample would pick out all male or all female.

3 Stratified sample: The list of potential interviewees is first grouped by any charac-teristic which the researcher wishes to see evenly represented in the sample – forexample, age, gender, location, etc. The sample can then be selected either ran-domly or systematically from within these groups.

4 Multi-stage cluster sample: The original list of potential interviewees is sampledrandomly in a number of stages, first to select a sample of areas (say, within a city),then to select a sample of smaller areas (perhaps neighbourhoods), then samplestreets and households within those neighbourhoods and, finally, sample individu-als within those households.

Non-probability samples

1 Purposive sample: Cases which are judged to be of particular interest to theresearcher.

2 Quota sample: where stipulated quotas of cases will be samples without randomselection procedures. Interviewers will simply select cases which conform to the cri-teria set (for example, males and females between the age of 50 and 60). By notusing random sampling the cases may ‘self-select’ because they will be the kinds ofpeople who will most readily agree to be interviewed.

3 Availability sample: Where anybody who agrees to be interviewed becomes part ofthe sample. This can happen, for example, where an advertisement is placed in thenewspaper to ask whether people are willing to answer questions of a certain type.Those who respond are thereby ‘self-selected’.

Source: After de Vaus (1991)

sample can be used to make assumptions aboutthe total population from which it is drawn.‘Indicative’ samples may give impressions aboutthe total population but are inappropriate forstatistical analyses which seek to interpret thattotal population. The size of the sample isalso important in achieving a representativegroup of interviewees. Broadly, the necessarysample size will depend on both the extent to

which the population varies and the preciserequirements of the survey (see Ebdon, 1985;Williams, 1986).

It will also invariably be necessary topilot the survey before starting on the maindata-gathering stage. Piloting serves many func-tions: it familiarizes the questioner with thequestionnaire design in practice; it provides anopportunity to find out which questions do not

TALKING TO PEOPLE

145

3151-Ch-05.qxd 3/2/04 2:43 PM Page 145

work well so that they can be improved orreplaced; very importantly, it allows for theremoval of extraneous questions which arefound not to construct data which are useful forthe overall research question(s); and it gives aclear idea about how long the survey takes tocomplete.Pilot surveys are often treated haphaz-ardly or perfunctorily by researchers of all levelsof experience but they are fundamental inallowing major improvements to be made priorto the main process of data construction.

The implementation of questionnairessubdivides sharply according to the form of sur-vey chosen. For a postal questionnaire itis important to remember that the paperworkwhich is sent is effectively a research encounterby proxy. Just as the researcher will be looking tointerpret aspects of the respondent through themedium of his or her replies, so the respondentwill be interpreting the researcher through themedium of the survey instrument. Even with-out meeting, therefore, a socially constructedrelationship will be established between theresearcher and the researched by using a postalquestionnaire. Most researchers will thereforewant to represent themselves as efficient, profes-sional, yet careful over important issues of ethics.They will also implicitly be seeking as high aresponse rate as possible. Therefore a carefullyworded letter establishing the position of theresearcher, the purpose of the survey, what willbe done with the results, the safeguards for con-fidentiality and the anonymity of the respon-dent, and the hope for a returned completedquestionnaire by a specified date will usuallyaccompany the questionnaire itself, along, ofcourse, with a stamped, self-addressed envelope.Figure 5.4 illustrates the letter from the group offirst-year Lampeter students which accompa-nied the postal questionnaire on farm diversifi-cation (Figure 5.2). It may be useful to send acarefully worded reminder to non-respondentsafter a reasonable length of time.

Much the same information has to begiven to potential respondents in telephone

and face-to-face questionnaires. A letter inadvance of the call or visit and perhaps eventhe booking of appointments for calls and vis-its are increasingly appropriate as a means ofrespecting people’s privacy and time manage-ment. Over the telephone, voice andlanguage will feed information in both direc-tions which will embellish the relationshipbetween researcher and respondent, and inface-to-face questionnaires matters of dresscode, body language and personality willadditionally contribute to this relationship.Our experience is that some researchers willtend to ‘hide behind’ their survey, giving verylittle of themselves in establishing a rapportwith their respondent, while others tend topresent the survey and themselves as distinctlydifferent entities allowing the researcher tocomment on the technicalities of the surveywhen the respondent finds it hard going.

The textbooks on survey methods presenta very rigid view of ‘doing a standardizedquestionnaire’. Fowler (1988), for example,although using the term ‘interviewer’, sets outa strong prospectus for standardization whenundertaking face-to-face questionnaire work.He suggests that

survey researchers would like to assume thatdifferences in answers can be attributed todifferences in the respondents themselves –that is, their views and their experiences –rather than to differences in the stimulus towhich they are exposed – that is, the ques-tion, the context in which it was asked, andthe way it was asked. The majority of inter-viewer training is aimed at teaching traineesto be standardised interviewers who do notaffect the answers they obtain. (1988: 109)

There is also an insistence on standardization ina number of different areas: presenting thestudy; asking the questions; probing for fulleranswers; recording the answers; and interper-sonal relations.On this latter point,Fowler says:

Inevitably, an interviewer brings someobvious demographic characteristics into an


146

3151-Ch-05.qxd 3/2/04 2:43 PM Page 146

interview such as gender, age, and education.However, by emphasising the professionalaspects of the interaction and focusing on thetask, the personal side of the relationship canbe minimised. Interviewers generally areinstructed not to tell stories about themselvesor express any views or opinions relatedto the subject matter of the interview.Interviewers are not to communicate anyjudgements or answers that respondentsgive. In short, behaviours that communicatethe personal, idiosyncratic characteristics ofthe interviewer are to be avoided. (1988: 110)

Such comments are clearly designed toprompt us about maintaining the reliability ofour surveys, and the goal of reliability contin-ues to be an important one in many circum-stances. However, in our view, these rules ofstandardization can place an unreal and anunjustifiable burden of expectation on humangeographers who choose to practise their sub-ject through questionnaires. It is well nighimpossible to exclude personal characteristics

TALKING TO PEOPLE

147

Saint David’s University CollegeUniversity of Wales

Professor Peter Beaumont: Department of Geography

24th February 1992

Dear Sir/Madam

We are a group of Human Geography students at St. David’s University College, who are currentlyinterested in researching the causes of upland hill farm diversification in the Dartmoor area, as acompulsory unit of our degree course.

We are contacting 40 farmers asking for 10 minutes of your time to fill in a short questionnaire aboutdiversification. We will also be conducting a few short follow-up interviews in April.

We would like to reassure you that your participation in this questionnaire will be treated with the utmostconfidentiality. Similarly, results from the study will be treated in the strictest confidence.

We would greatly appreciate a quick response as we have a tight deadline to meet. Therefore to facili-tate this, please find enclosed a stamped, self-addressed envelope.

Thank you for your co-operation

Karen Inston

On behalf of:

Pam HowellMike GaddClare BorerPauline MacKenzie

LAMPETER, DYFED, SA48 7ED, WALES, UK TELEPHONE: LAMPETER (0570) 422351 FAX: 0570 423423

Figure 5.4 Letter to accompany a postal questionnaire

Source: Karen Inston

3151-Ch-05.qxd 3/2/04 2:43 PM Page 147

from the social relationship formed in thiscontext.The researcher carrying out a question-naire is not a white-coated automaton whocan treat all respondents in exactly the sameprofessional manner. The very differences ofpersonality and experience which inspirevariations in the geographical imagination inthe first place will inevitably surface duringthe questionnaire.There are some respondentsyou just don’t relate to, or they don’t relate toyou, and encounters with them are just notthe same as those with people with whomyou establish a much more mutually enjoyableand supportive relationship during the meet-ing, regardless of your ‘professionalism’ inrecording answers and asking questions.Equally, the ‘researcher as robot’ model is alsoflawed in its own terms because the supposedobjective neutrality of the researcher will beresponded to differently by different people.Being scientifically neutral may appeal to somerespondents, giving an impression of profes-sionalism. To others, however, it may appearrude, uncaring and exploitative. Clearly, then,the idea that adopting a position of scientificneutrality to ensure that social situations ofresearch are either asocial or totally determinedby one of the participants is doomed to failure.‘Research as robot’ is itself a social role.

Equally it is our view that questionnaireresearchers should respect the respondent’srights to know about the survey and aboutwhat will happen to the results. This willsometimes involve researchers in tellingstories about themselves and commenting onsubject-matter relevant to the issues involvedin the survey. When you have sat down insomeone’s house for an hour or two, acceptedhis or her refreshments and been extraordi-narily nosy about his or her life and lifestyle,it is perhaps only fair to respond to his or herquestions about who you are and what youare doing with something other than ‘mytraining manual doesn’t allow …’. There isobviously a balance to be struck here. An

on-street questionnaire may be conducted asthe textbooks suggest, but it seems to us thatin general there are relational and ethical con-siderations which make it extremely difficultto accede to the demands of doing a stan-dardized questionnaire.

It may be that in coming to this conclusionwe are attempting to move the questionnairesurvey away from its position as a techniqueof conventional scientific method. Perhaps ifthe flow of the research issue (to data con-struction, to data interpretation) is compelledby a positivistic programme of hypothesis-testing and statistical analysis, the need toeradicate interviewer bias will be viewed ascrucial in order to set reliable standardsthroughout the survey. Our view, however, isthat any such form of data constructioninvolves a variable social relationship betweenthe researcher and the respondent.While it isentirely valid to attempt to construct compa-rable data about a potentially large number ofrespondents, thereby accepting the stricturesof a structured survey format, it is often alsovalid to listen to and record other relevantinformation which is offered by the respon-dent, even if the definition of ‘relevant’ in thiscase is dependent on the researcher as an indi-vidual rather than as a standardized researchinstrument.

Interviewing

The transition from questionnaires to inter-views is by no means a straightforward one.Many face-to-face questionnaires will containelements of interviewing, and some of ourdiscussion of questionnaires will be relevantto the carrying out of interviews. However,we want to argue that there are important dis-tinctions between questionnaires and inter-views and that these distinctions arise as muchfrom the intent and purpose of the inter-viewer as from the ways in which questions


148

3151-Ch-05.qxd 3/2/04 2:43 PM Page 148

are asked. Interviewing has been described as‘conversations with a purpose’ (Webb andWebb, 1932; see also Burgess, 1984) and,although the conversations vary across a rangeof structured, semi-structured and unstruc-tured formats, their purpose is to ‘give anauthentic insight into people’s experiences’(Silverman, 1993: 91). If the scientific meth-ods so often associated with the use of ques-tionnaires attempt (often unsuccessfully) tomirror the social and geographical worlds ofrespondents, interviews will employ know-ingly interactive research so as to gain accessto the meanings which subjects attribute totheir experiences of these worlds. Humangeographers throughout the 1980s and espe-cially the 1990s have increasingly found amatch between what interview research canoffer and the overall questions and issueswhich spark their geographical imaginations.At the same time, however, there has beenincreasing concern over the rigour of thisresearch. Baxter and Eyles (1997: 511), forexample, argue the need

to be explicit about the principles for mak-ing this work rigorous. Being forthcomingabout these criteria will better equip thosewho do not traditionally work within thequalitative paradigm to judge its approachand findings and, perhaps more impor-tantly, these criteria will be made public forconstructive scrutiny and debate.

The criteria to which they refer are framedvery much in the context of an accusativeculture in geography in which battle-lines arereadily drawn up between the qualitative andthe quantitative:

questioning how things are done – anessential component of self-reflection –allows qualitative research to demonstratethe relevance of the single case (credibility)and to move beyond it (transferability) witha degree of certainty (dependability or con-firmability). Context, contingency and thespecific positioning of subjects (includingresearcher-as-instrument) are central toqualitative inquiry and are not threatened

by the application of a general set of criteriafor evaluating rigour. These criteria providereasonable anchor points for a paradigmwhich is often inappropriately accused ofengaging with ‘anything goes’ science.(1997: 521; see also Bailey et al., 1999)

It is evident from this that the seeminglysimple goal of gaining access to meanings andexperiences via the use of interviews is by nomeans an unproblematic task. It is possible toapproach interviewing in a rather blasé fash-ion, somehow assuming that all you have todo is to ask the right questions and the realityof your respondent’s feelings, thoughts andexperiences will fall into your notebook ortape-recorder like picking ripe apples from atree. Even with the now extensive literatureon how to conduct interviews without con-taminating them with ‘bias’ (see, for example,Fowler and Mangione, 1990) researchers canseem to approach interviews as methods ofdetecting existing truths from willing respon-dents.As Holstein and Gubrium (1997) pointout, the supposed ‘trick’ is to formulate goodquestions, create an atmosphere conducive torapport and produce undistorted communi-cation between the interviewer and therespondent. They term this the ‘vessel-of-answers’ approach:

In the vessel-of-answers approach, the imageof the subject is epistemologically passive,not engaged in the production of knowl-edge. If the interviewing process goes ‘by thebook’ and is non-directional and unbiased,respondents will validly give out what sub-jects are presumed to merely retain withinthem – the unadulterated facts and details ofexperience. Contamination emanates fromthe interview setting, its participants andtheir interaction, not the subject, who, underideal conditions, serves up authentic reportswhen beckoned to do so. (1997: 117)

We believe that interviewing in this concep-tual context usually turns out to lack a criticalsensitivity to the interactions and exchangesbetween the interviewer and the interviewee.The interview thus becomes less rather than

TALKING TO PEOPLE

149

3151-Ch-05.qxd 3/2/04 2:43 PM Page 149

more rigorous because it assumes that aconstant set of ‘truths’ is available to be garneredfrom the interviewee regardless of the socialand cultural conditions in which the inter-view is constructed and carried out. Instead,we argue strongly that interviewers are them-selves implicated in the construction ofmeanings with their interviewees. Such inter-subjectivity is crucial and unavoidable, and thedata which result are essentially collaborative(see Alaasutari, 1995).

This explicit recognition of the intersub-jectivity of interviewing means that the‘respondent’ is acknowledged as an active sub-ject. He or she mediates and negotiates what istold to the interviewer and he or she engagesin a continuous process of assembling andmodifying answers to questions.The supposedvessel of objective knowledge is actually anactive maker of meaning. Answers on oneoccasion may well not be replicated onanother because the circumstances of produc-ing those answers will differ, and because con-veying situated experiential realities is by nomeans the same process as supposedly testinganswers against an immutable bank of ‘true’answers within the respondent.These preceptsthrow a rather different light on the key com-ponents of interviewing:

understanding how the meaning-makingprocess unfolds in the interview is as criticalas apprehending what is substantivelyasked and conveyed. The hows of interview-ing, of course, refer to the interactional,narrative procedures of knowledge produc-tion, not merely to interview techniques.The whats pertain to issues guiding theinterview, the content of questions, and thesubstantive information communicated bythe respondent. A dual interest in the howsand whats of meaning production goeshand in hand with an appreciation of theconstitutive awareness of the interviewprocess. (Holstein and Gubrium, 1997: 114,emphasis in original)

Interviews, then, are rather like an ‘interper-sonal drama with a developing plot’ (Pool,

1957: 193).The interview has a script, at leastin outline, roles to be played out and a stageon which conversation occurs. However, ithas a plot which develops rather than beingstatic. The waxing and waning of the inter-view will fashion script, roles and stage, some-times unpredictably. There is plenty of roomfor improvisation, both by the interviewerand by the subject; each is complicit in theproduction of a narrative. The result is farfrom being an objective, representative andtechnical exercise. Indeed, the strengths ofusing interviews lie in the very acknowledge-ment of intersubjectivity which permits adeeper understanding of the whos, hows,wheres and whats of many aspects of humangeography research.

When are interviewsappropriate?

Practising human geography is often a matterof choosing horses for courses, and the use ofinterviews is no exception to this. It will beclear from the above that broad surveys ofsurface patterns involving large numbers ofpeople will be more compatible with ques-tionnaire techniques than with in-depthinterviewing. However, human geographershave become increasingly interested in issueswhich fall between or below analyses of sur-face-level comparability. Instead there is anemphasis on explaining processes, changingconditions, organization, circumstances andthe construction, negotiation and reconstruc-tion of meanings and identities. Here therewill be a turn to intensive rather than exten-sive research strategies (see Table 5.1), and theuse of interviews is both legitimate and nec-essary in order to create the kind of informa-tion which lends itself to the types ofexplanation required (see Chapter 9).

As an example, in a recent study of home-lessness in rural areas it was found to beentirely appropriate to undertake a postal


150

3151-Ch-05.qxd 3/2/04 2:43 PM Page 150

questionnaire survey to homelessnessofficers in local authorities in order to garnera wide spread of comparative informationon the scale and perceived significance ofhomelessness in rural areas of the UK.However, when seeking to understandsomething of what homelessness is like inrural areas, the informal interview presenteda more appropriate way of approaching andtalking to homeless people about the feel-ings and meanings they attach to their expe-riences of being homeless (Cloke et al.,1999; 2000a; 2000b).

A second context for using interviews iswhen we as researchers wish to consider our-selves as active and reflexive in the process ofconstructing information.To some extent it isinevitable that the researcher is an activeinfluence throughout the research process,whether or not he or she is seeking to beneutral or objective. However, interviewsoffer a more deliberate route than this sug-gests for those who believe that ‘people’sknowledge, views, understandings, interpreta-tions, experiences and interactions are mean-ingful properties of the social reality’ (Mason,1996: 39) and who wish to practise thesebeliefs by interacting with people by talkingto them, listening to them and negotiatingwith them over narrative accounts. It used tobe the case that the word ‘I’ was persona( l pronoun) non gratis in human geographi-cal writing. However, the recognition of theactive subject and the reflexive self in humangeography has rendered entirely legitimate anapproach which makes explicit the intersub-jectivities inherent in interview practices.

A third reason for using interviews lies inthe desire of human geographers to ‘givevoice to’ others as an integral part of theresearch process. Questionnaires by their verynature will tend to aggregate individuals intothe categories which constitute extensive pat-terns. Not only are human geographers nowinterested in the inclusion of lay discourses in

academic discourses (see, for example,Halfacree, 1993) but many are also deter-mined to look through different ‘windows’ onto the geographical world by undertakingresearch in which a range of different voicescan be heard (see, for example, Philo, 1992).Such research might be aimed at understand-ing the position(s) adopted by one or moresocial groups on a particular place, practice orprocess, or the focus might be on particularindividuals in the form of oral histories, lifehistories or life stories. Interviews offer amethodological strategy by which thesevarious aims can be realized although, onceagain, care should be exercised in too blasé anassumption of what interviews can offer.Indeed, it is important to reiterate that the‘voices’ that are heard are very much subjectto the editorial and authorial concerns of theresearcher, as well as to the interpersonaldrama of the interview itself.

A final set of reasons for choosing toundertake interviews reflects rather morepragmatic circumstances. In some contexts(for example, an undergraduate dissertation)the time and resources available often do notpermit the undertaking of an adequately rep-resentative questionnaire survey. Here, thepragmatic view could dictate that a moremanageable programme of interviews beundertaken, and research aims might have tobe altered accordingly. From a position inwhich the questionnaire was the ‘orthodox’research tool, many human geographers havecome to realize that in circumstances oflimited resource or limited access to certaintypes of respondent it is more appropriate toundertake a series of interviews rather thana questionnaire survey. Even so, we wouldreiterate that interviews are neither a softoption – they need to be thoroughly prepared,carefully undertaken and then painstakinglytranscribed and interpreted – nor are they adirect replacement of other research toolssuch as the questionnaire.

TALKING TO PEOPLE

151

3151-Ch-05.qxd 3/2/04 2:43 PM Page 151

Preparing for an Interview

Having established that interviews can beregarded as conversations with a purpose,which represent interpersonal drama with adeveloping plot, we now turn to more practi-cal issues of how to prepare for and carry outan interview.As a very practical starting pointto our discussion of ‘doing’ an interview wewant to emphasize the need for considerableforethought and care about how the inter-view is to be recorded. If the interviewee iswilling for a tape-recorder to be used, tapingis by far the best way of recording the pro-ceedings. Researchers need to be aware, how-ever, that the presence of a tape-recorder cansometimes present a psychological barrier tothe interviewee, whose responses might bemore guarded and ‘formal’ than if the tape wasnot running.

It is also important to be well prepared fortape-recording. Prior familiarization with themachine and regular battery checks are invalu-able in avoiding the interviewer’s nightmare offinding out, after the event, that the interviewhas not yielded any text to quote from due tothe tape-recorder not working properly!Interviewers should also be aware that thedecision to tape an interview commits them tolong hours of transcription afterwards! Note-taking, on the other hand, will reduce theopportunity to use verbatim quotes from theinterview and can distract the interviewer’sconcentration from the task of listening to theinterviewee, thus reducing some of the poten-tial intersubjectivities of the exchange. If youare thinking of note-taking, various forms ofshorthand can be deployed to make the taskmore simple and speedy. These technical detailsmay appear trivial but, in our experience, theymatter because they fundamentally determinethe character of the data constructed, and shapethe purposes to which they can be put.

Aside from these technicalities it is alsoimportant from the outset to acknowledge a

constructive and critical tension here betweenmethodological rigour and dramaturgicalspontaneity. In our collective experience, theunexpected ‘chat’ with key individuals hasoften proved to be a most fruitful researchmoment. Equally, a well-prepared interviewcan launch off into unanticipated territorywith beneficial consequences for the under-standing of particular meanings and experi-ences. In contrast to many textbooks ongeographical methods, then, we would sug-gest that preparation, though important, is notthe be-all and end-all of interviewing. Indeed,one of the key skills of interviewing is a sen-sitivity of listening to what is being said,linked with an innate flexibility to permit andencourage encounters with the unexpected.

For these reasons, the idea of the ‘textbookinterview’ can be misleading and, at worst, canrender interviewing immune to the veryintersubjectivities which lie at the heart of theprocess. Nevertheless, methodological rigourdoes have a significant part to play in the cre-ative and critical tensions of interviewing, andpart of that rigour involves thorough prepara-tion where applicable. Box 5.5 outlinesJennifer Mason’s excellent guide to planningand preparing for qualitative interviewing. Inthe context of her research on how familiesdeal with issues of inheritance, she worksthrough the process with very helpful detailand contextualized expectation. From theidentification of the big research question, anumber of smaller mini questions are devel-oped, each of which can be thought throughin terms of ‘what you need to know’ and‘what’. Cross-referencing of these smallerresearch issues helps to keep the big researchquestion in mind.The next ‘stage’ is to thinkthrough how the interview might go – at leasthow the researcher might want it to go – andthereby to develop a loose sense of format orstructure for the interview. This, of course, hasto be flexible but it often leads to a series of‘topic’ or ‘issue’ prompts that can be used in


152

3151-Ch-05.qxd 3/2/04 2:43 PM Page 152

Box 5.5 Planning and preparing for qualitative interviewing

Step 1List or assemble the ‘big’ research questions which the study is designed to explore.

Example of one of the ‘big’ research questions in the ‘Inheritance’ project1 How do families handle issues of inheritance?

Step 2Break down or subdivide the big research questions into ‘mini’ research questions.Thelinks between the big questions and the subcategories of them – the mini questions –should be clearly expressed, for example by using corresponding numbers or codes, orby laying the two sets of questions out in a chart, or by using cross-referenced indexcards. It is possible to establish a perfectly workable manual system, or you can use acomputer graphics package and/or database to help you.

Example of mini research questions which are subcategories of the bigresearch question given above

1 (a) Are negotiations about inheritance treated as part of a wider set of negotiationsabout support in families? Or is inheritance treated as a totally separate matter?

(b) Do people in any way take into account the possibility of inheritance in for-mulating their own life plans?

(c) Is a clear distinction maintained between ‘blood relatives’ and ‘in-laws’ in theprocess of negotiating inheritance?

Step 3For each mini research question, start to develop ideas about how it might be possible toget at the relevant issues in an interview situation. This means converting your big andmini examples of ‘what you really want to know’ into possible interview topics, and think-ing of some possible questions – in terms of their substance, and the style you might use toask them. These will not form a rigid ‘script’ for you to use in the interview, but the processof developing possible topics and questions will get you thinking in ways appropriate to aninterview interaction. Again, make sure that the links between this set of questions and theother two (that is the big and mini research questions) are clearly expressed.

Examples of interview topics and questions related to mini research questions

1 (a) Family inheritance history, and history of other family support – what hap-pened in practice in relation to specific events and instances? How did peopledecide what was the most appropriate course of action?

(b) Knowledge of the inheritance plans, content of wills etc., of other familymembers. Have people thought about inheritance at all? Have they made wills?Do people have life plans, for example, do people have a sense of what theywill be doing, where living, and so on, in 5 or 10 or 20 years’ time? How werethese plans arrived at?

(c) Ascertain composition of family and kin group, and what kinds of relationships existwith specific others. Explore whom people count as ‘blood kin’, whom as ‘in-laws’ or‘step-relatives’ – establish this so that family inheritance history, and specific eventsand instances, can be contextualized in the sense that we will know the ‘kin status’(as conceptualized by the interviewee) of relevant parties. Explore the detail of dis-tributions of assets, and negotiations about them, in relation to kin of differentstatus. Who has legitimate interests? How do people decide whom to include andexclude? Possibly ask directly whether people think about their blood relatives andtheir in-laws in different ways in relation to inheritance, and other matters.

TALKING TO PEOPLE

153

3151-Ch-05.qxd 3/2/04 2:43 PM Page 153

Box 5.5 (Continued)

Step 4Cross-reference all the levels, if you have not done so already, so that you know thateach big research question has a set of corresponding mini research questions, andeach of these has a set of ideas about interview topics and questions. Make sure thecross-referencing works in reverse, so that your interview topics and questions reallyare going to help you to answer your big research questions.

Step 5Start to develop some ideas about a loose structure, or format, for interviews. You willwant this to be highly flexible and variable, but you should be able to produce somekind of guide to the key issues and types of questions you will want to discuss.

Examples of loose interview structure/format developed for the ‘Inheritance’projectIn this project we developed a loose interview format, based on key topics and types ofquestions we were likely to want to ask. With each interviewee we anticipated followingup lines of enquiry specific to their circumstances, which we would not be able to anticipatein advance. We therefore wanted maximum flexibility, but also some kind of guide orprompt for the interviewer about the key issues and questions with which the study wasconcerned. We did not produce a script of questions, but rather a set of index cards to takeinto each interview. One card contained a flow chart of a possible interview structure,which could be readily modified on the spot. The other cards contained shorthand notesabout specific topics and issues for the interviewer’s use at relevant points in the interview.These notes were non-sequential, so that they could be drawn upon at any time, in relationto the specific context of the interview in progress. Here are examples of each type of card:

‘Loose structure/format of interview’ card

Possible main structure Specific topics and issues – to be asked in relation to any of the main structure sections (there are cards for each of these sets of questions)

Introductory explanation↓

Brief social/personalcharacteristics

↓Composition of kin group Inheritance history, other responsibilities and and spouse’s kin group relationships, inheritance family and kin group

↓Family inheritance history

↓Specific questions (if Formal and external factors, including the lawnot covered elsewhere)

↓Questions about the Principles and processes of inheritance and familylaw check responsibility

↓Personal characteristics check Social and personal characteristics (current and over time)


154

3151-Ch-05.qxd 3/2/04 2:43 PM Page 154

‘Example of specific topics and issues’ cardInheritance history, other responsibilities and relationships, inheritancefamily and kin group

Experience of inheritance: personal/others – as testator, beneficiary, executor;patterns characteristic of own family; how many generations; experience of legalprocedures and services; expected and unexpected; experience of will making; when,why; professional advice; intestacy laws; lifetime transfers.

Inheritance and other aspects of kin relationships/wider patterns of responsi-bility: family relationships affected by inheritance? conscious of possible inheritance inrelationships with relatives?; conflicts – how resolved; life plans and inheritance, e.g.housing, geography, timing; death and how it is dealt with; making formal statementsabout relationships?; part of ongoing reciprocity and exchange – explicit/implicit?; ideaof final settlement?

The inheritance family or kin group: who has legitimate interests?; in-laws/ step-relatives/secondary kin?; excluions and principles of exclusion/inclusion; inheriting viasomeone else.

Step 6Work out whether you want to include any standardized questions or sections in yourinterviews. There may be certain questions which you want to ensure that you askeverybody. In the example above the introductory explanation was fairly standardized,as were some of the questions about personal and social characteristics (forexample, age, marital status). You might also want to think of some standardizedcomments and assurances which you will make about confidentiality of data to yourinterviewees.

Step 7Cross-check that your format, and any standardized questions or sections, do coveradequately and appropriately your possible topics and questions.

Source: Mason (1996: 44–51)

Box 5.5 (Continued)

the interview as a checklist to ensure coverageof the required ground. Such a checklist willoften involve important standardized state-ments at the beginning and at the end of theinterview to explain to the subject what willbe done with the information which is con-structed therein. Issues of confidentiality andfurther communication will be importanthere.

Helpful though these guidelines for prepa-ration are, they form only part of the picture.Intersubjective conversations with a purposeinvolve an active collusion between participants.

A mechanical trudging through of aninterview checklist or schedule by the researcherwill restrict the possibilities for interpersonaldrama, and therefore of plot development.Wedeal with issues relating to conducting inter-views below. However, the participation ofthe research subject will also vary enor-mously. Part of the preparation for interviews,then, lies in making decisions about whom tointerview. Much of the traditional methodo-logical training for human geographers pointsthem towards the idea of dealing with ‘repre-sentative samples’ in their research. Although

TALKING TO PEOPLE

155

3151-Ch-05.qxd 3/2/04 2:43 PM Page 155

principally associated with the constructionof data which can be analysed statistically, thesense of somehow wanting to talk to peoplewho represent a wider social group has alsospread across into qualitative interviewing.Indeed, where interviews are designed todocument a series of different properties ordiscursive principles, the researcher will wantto choose to interview across that range ofdifference, and will therefore use a loose formof ‘representative’ samples of interviewees.However, such choices are not representativein the statistical sense, and the idea of spread-ing interviews across axes of difference – bethey geographical, social, identity related,practice related or whatever – soon runs intoproblems if individuals are somehow beingasked to represent their ‘category’. For example,when interviewing an elderly, middle-classrural woman who is active in local politics ora young gay urban man working in financialservices, the multiple and shifting identitiesinvolved will usually defy easily categorizedexplanations relating to, say, gender, sexualityor place.

Choosing whom to interview, then,involves a targeting of people who are likelyto have the desired knowledge, experiences orpositionings, and who may be willing todivulge that knowledge to the interviewer. AsSusan Smith (1988: 22) has argued: ‘anyattempt on the part of an analyst [geographer]to enter the life-world of others is above all,strategic … it makes both moral and analyti-cal sense to expose the power relations inher-ent … at an early stage of the research.’Sometimes researchers will have difficulty ingaining access to an identity group or organi-zation. Here we often encounter ‘gatekeepers’who either are in an official position to handleinquiries about an organization or who arepowerful figures within a particular group(Hughes and Cormode, 1998). Gatekeeperswill often attempt to structure the access ofthe researcher to others, pointing us towards

‘helpful’ or ‘safe’ interviewees who in thejudgement of the gatekeeper are appropriateonward contacts. Naturally, there are pitfallsinherent in these decisions such that theresearcher

must retain the leeway to choose candidatesfor interview. Otherwise there is a gravedanger that the data collected will be mis-leading in important respects, and theresearcher will be unable to engage in thestrategic search for data that is essential to areflexive approach. However, gaining accessto informants can be quite complex, some-times as difficult as negotiating access to asetting. Indeed, it may even be necessary tonegotiate with gatekeepers before one cancontact the people one wants to interview.(Hammersley and Atkinson, 1995: 134)

Sometimes, however, researchers have to workwith the interviews they can get, and it is cru-cial that the context in which knowledge isbeing produced is fully acknowledged inthese circumstances.

‘Working with what you can get’ alsoapplies to groups of subjects who will not beamenable to direct approaches or who aredifficult to get hold of. Here ‘snowballing’can be used whereby a small initial group ofinformants (this might be one person) can beasked to provide interview contacts withfriends and acquaintances within a particularsocial or identity group, thus permitting achain of interviewees to emerge. Similarissues to those relating to the role of gate-keepers arise here, although they are oftenless obvious in snowball samples. Where theemphasis is to use interviews to hear differentstories, the propensity of people to pass theresearcher on to people whom they considerto be like-minded needs to be acknowledgedin the intertextualities which underpininterviews.

Finally, in terms of how to prepare forinterviews, it is worth noting that we shouldbe prepared for different levels and types ofresponse from research subjects. This may be


156

3151-Ch-05.qxd 3/2/04 2:43 PM Page 156

illustrated in terms of Erving Goffman’s ideasabout the ‘two-sidedness’ of people perform-ing occupational roles: ‘One side is linked tointernal matters held dearly and closely, suchas image of self and felt identity; the other sideconcerns official position, jural relations, andstyle of life, and is part of a publicly accessibleinstitutional complex’ (1968: 119).

Goffman’s work suggests that there will bea ‘front’ and a ‘back’ to the responses given byinterviewees. In the case of interviews withrepresentatives of agencies, the front responsesmay be the readily available public relationsnarratives designed for public consumption.The ‘back’ responses will be more hidden andyet may be the areas of most interest to theresearcher in that they may reflect oppositionto, or negotiation with, the public face of thefront responses. It can also be argued, however,that a much wider range of interviews willinvolve these public/private distinctions,which can be anticipated in interview prepa-ration. In the interpersonal drama of inter-viewing, however, such distinctions emerge ina much more messy and complex form andmay well defy the notion of clearly definedback or front regions of knowledge andresponse. Indeed, the active subject will be cre-atively making meanings during the interviewitself, and so the idea of fixed reference pointsof back and front knowledge should representinstead rather less fixed markers in the criticaltension in interviews between rigorous prepa-ration and dramaturgical spontaneity.

Conducting an interview

By using the metaphor of drama to emphasizeboth interpersonal performance and thedevelopment of an plot within interviews, welay ourselves open to an account of conductingan interview in which the script is tightlyadhered to, and the interviewer plays all thepowerful roles. For example, Berg (1989: 35)suggests that interviewers must play the role

of actor, director and choreographer (see Box5.6) during the ‘performance’ of an interview:

interviewers must be conscious and reflec-tive. They must carefully watch and inter-pret the performance of the subject. Theirinterpretations must be based on thevarious cues, clues, and encoded messagesoffered by the interviewee. Included inthe information these interactions supplymay be the communication of a variety ofmoods, sentiments, role portrayals, andstylised routines, which represent the inter-viewee’s script, line cues, blocking, andstage directions. From this information,interviewers must take their cues, adjusttheir own blocking, and effect new orresponsive stage directions.

This extended use of dramaturgical metaphorperhaps overemphasizes the interview processin which improvisation is at least as importantas delivering ‘classic’ lines. However, it doespoint us towards a discussion of the degree towhich an interviewer can influence the inter-subjectivity of the interview by providing asituated context, by employing interviewingskills and by asking the ‘right’ kinds of questions.

Sometimes the researcher will not be incontrol of where an interview takes place.Thesubject may choose where to be interviewed –in a place of paid employment, in the homeor perhaps in some neutral location – andgatekeeping individuals or agencies may welldictate the interview location. For example,many interviews in schools take place inbroom cupboards marked ‘office’. Equally, theresearcher will not always be able to choosewho else is present during the interview, andinevitably the presence of parent, child or co-worker will alter the interpersonal dynamicsof the interview. As an extreme example, it isnot unknown for the friendly attention of thepet dog or cat to induce such a strong allergicreaction that an interviewer cannot completethe interview concerned (see Popkin, 1998)!

The ‘textbook’ interview calls for the inter-viewer to establish a ‘rapport’ with the subject.The achievement or otherwise of rapport is

TALKING TO PEOPLE

157

3151-Ch-05.qxd 3/2/04 2:43 PM Page 157

Box 5.6 Roles played during interviews

Interviewers as Actors. Interviewers as actors must perform their lines, routines, andmovements appropriately. This means that in addition to reciting scripted lines (theinterview questions), they must be aware of what the other actor (the interviewee) isdoing throughout the interview. Interviewers must listen carefully to line cues in orderto avoid ‘stepping on the lines’ of the interviewee (interrupting before the subject hascompletely answered a question).

Interviewers as Directors. Simultaneous to their performance as actors, interviewersmust also serve as directors. In this capacity, interviewers must be conscious of how theyperform lines and move as well as of the interviewee’s performance. Interviewers mustreflect on each segment of the interview as if they were outside the performance asobservers. From this vantage point, interviewers must assess the adequacy of their per-formance (for example, whether line cues from the interviewee are responded to cor-rectly and whether avoidance messages are being appropriately handled).

Interviewers as Choreographers. The various assessments made in the role of direc-tor involve a process similar to what Reik … described as ‘listening with the third ear’.By using what they have heard (in the broadest sense of this term) in a self-aware andreflective manner, interviewers manage to control the interview process. As a result,interviewers, as choreographers, can effectively block their movements and gesturesand script their response lines.

Source: Berg (1989: 35)

certainly influenced by whether the locationof the interview appears to reinforce thepower either of the interviewer or the subject.Rapport is also thought to be influenced bythe appearance and demeanour of the inter-viewer (see, for example, Gorden, 1987).Characteristics of age, gender, race and the likecannot be ‘fine tuned’ unless a team of differ-ent interviewers is involved. It may sometimesbe crucial, however, to undertake interviews ina native language in bilingual or multilingualareas. Other characteristics such as style ofdress, manner of speech and more generaldemeanour can be adjusted to suit a particularoccasion. For example, casual clothes may beessential when interviewing young people onsensitive issues so that the interviewer does notcome across as an authority figure. Conversely,the interviewer will perhaps wish to comeacross as an authority figure when talking to

subjects who have that kind of expectationabout what their interviewer should look like(Walford, 1994). Thus a business suit is oftenworn when interviewing high-ranking man-agers in industry or commerce.

These attempts to ‘fit in’ are valid reflec-tions of the wish to establish rapport with theresearch subject, but they by no means guar-antee that barriers to in-depth intersubjectiv-ity will be removed. Many active subjects will‘see through’ obvious attempts to establishrapport and will rely much more on the sen-sitivity displayed by the interviewer duringthe interview.This in turn begs the questionof the degree to which the acquisition andpractice of interviewing skills will enable theinterviewer to achieve the desired depth andscope of conversation. On the surface, suchskills are both demanding and formidable. AsMason (1996: 45–6) points out:


158

3151-Ch-05.qxd 3/2/04 2:43 PM Page 158

At any one time you may be: listening towhat the interviewee(s) is or are currentlysaying and trying to interpret what theymean; trying to work out whether whatthey are saying has any bearing on ‘whatyou really want to know’; trying to think innew and creative ways about ‘what youreally want to know’; trying to pick up onany changes in your interviewees’ demeanourand interpret these …; formulating the nextquestion which might involve shifting theinterview onto new terrain; keeping an eyeon your watch and making decisions aboutdepth and breadth given your time limits.At the same time you will be observingwhat is going on around the interview; youmay be making notes or, if you are audio orvideo tape recording the interview, keepinghalf an eye on your equipment to ensurethat it is working; and you may be dealingwith ‘distractions’ like a wasp which youthink is about to sting you, a pet dog whichis scratching itself loudly directly in front ofyour tape recorder microphone, a tele-phone which keeps ringing; a child crying,and so on.

Key skills for interviewing will thereforeinclude: listening sensitively; rememberingwhat has already been said; achieving an effec-tive balance between listening and speakingout (with questions, prompts and responses toquestions from the interview subject); sensi-tivity to unspoken signals, particularly in thearea of body language and demeanour; andtechnical response in tape-recording or note-taking. For some, such skills are artistic andintrinsic; for others they can be taught andlearnt mechanistically (Roth, 1966).The pres-ence or absence of these skills will alter theintertextual construction of narrative in aninterview.

Another facet of the ‘textbook’ interview isto ask the right kinds of questions. For example,questions using affective words (that is, wordswhich carry some kind of emotional chargeor which are likely to elicit an emotionalresponse) are supposedly to be avoided. Forexample, according to Berg (1989: 24), ‘theword why in American culture tends to

produce in most people a negative response’and should therefore be avoided. Equally,overly complex or misleading questions arealso to be shunned. While all this in theorysounds like good advice, there is a danger hereof interviewers being so trained up indemeanour, skills and language that theirinterpersonal contribution to the interview isas a walking textbook rather than as a sensi-tive and honest human being. In our experi-ence, interviewers who present themselves assomething other than themselves will be rela-tively transparent to many research subjects.Indeed, it could be argued that the variousattempts to characterize the ‘textbook’ inter-view represent a conceptual clinging to thenotion of interviewing as extracting grains oftruth from the vessel-of-answers that is theinterviewee. Gaining access to the meaningswhich subjects attribute to their experiencesof their social and geographical worlds maynot be a technical issue at all. Critical aware-ness in interviews will come from acknowl-edgement of intersubjectivity, which in turndepends on the interviewer giving somethingof her or his self in the interview. The creativetension between methodological rigour anddramaturgical spontaneity will, therefore, beas much to do with intellectual and personalhonesty as to the demands of textbookinterviews.

Discussion groups

Introduction

One form of questioning and interviewingthat deserves special mention is the group dis-cussion, which has become increasingly popu-lar as a tool of qualitative inquiry in humangeography over recent years. Although theterm ‘focus group’ tends to be used generi-cally, Burgess (1996) makes the very important

TALKING TO PEOPLE

159

3151-Ch-05.qxd 3/2/04 2:43 PM Page 159

distinction between focus groups, which referto single-group interviews in the traditionand culture of market research, and in-depthdiscussion groups, referring to interactive groupinterviews in the tradition and culture ofpsychotherapy. Focus groups have been widelyused in contemporary society, particularly inareas of market research, ranging from thetesting of particular food products to the pre-liminary evaluation of policy initiatives fromthe New Labour government in Britain. Anexcellent example of the use of focus groupsin human geography is the research by PeterJackson and colleagues designed to investigatethe cultural significance of the new genera-tion of men’s lifestyle magazines (such asArena, GC, Loaded, FHM and Maxim). Some20 one-off focus groups were used to promptinteractive discussion in this research (Jacksonet al., 1999). For an example of best practicein the use of in-depth discussion groups inhuman geography we would point to thework of Jacqueline Burgess and colleagueswho have made excellent use of in-depth dis-cussion groups to explore the contestedmeanings associated with nature conservationand environmental values (see Burgess et al.,1988; 1988b; 1991).

To avoid confusion over terminology weprefer to use the generic term ‘discussiongroups’ and then to differentiate betweenfocus groups and in-depth discussion groupsas appropriate. Although there is a range ofspecialist literature which underpins themethodologies deployed in discussiongroups,4 we suspect that for some humangeographers the apparent ease of ‘doing a dis-cussion group’ leads to instances of dangerousmisuse.This risk should be reviewed alongsidethe very positive potential advantages fromthis technique when deciding whether dis-cussion groups are an appropriate method ofinquiry in a specific research context.

In simple terms, the generic idea of discus-sion groups may be defined as ‘a qualitative

method involving a group discussion, usuallywith six to twelve participants, focused aroundquestions raised by a moderator’ (Pratt, 2000a:272). However, as Goss (1996: 113) points out,such groups are more specifically

‘a confusion of the focused interview, inwhich an interviewer keeps a respondent ona topic without the use of a structured ques-tionnaire, and a group discussion, in whicha relatively homogeneous, but carefullyselected group of people discuss a series ofparticular questions raised by a moderator’.

It is important therefore to position discus-sion groups critically among the array ofother methods involved. As an interview tech-nique, group discussions allow researchers todraw out interaction between participants andmake direct comparisons between the experi-ences and opinions narrated in the group.However, the ability to go into depth withparticular individuals is commensurately lessthan in individual interviews. As a group obser-vation technique, discussion groups presentthe opportunity for researchers to observevery significant interactions between indivi-duals in confined space and time. Indeed, theprompting of interaction by the moderatorfeeds back into the interviewing process asparticipants begin to narrate material as aresponse or addition to what they have heardfrom other members of the group. However,for some, the discussion group may representan unnatural setting in which they are lesslikely to enter into sensitive or difficultdiscussion.

Just as qualitative methods literatures havetended to specify a ‘model interview’, so toohave ‘the common sense and the preferredpractices of a few researchers … been reifiedinto rules of thumb, or myths, … that specifythe ideal form of the group discussion’ (Goss,1996: 113). In an informative account of thetechniques of discussion group work, Morgan(1997) identifies particular phases of researchand poses key questions about each so as to


160

3151-Ch-05.qxd 3/2/04 2:43 PM Page 160

provide appropriate guidelines for undertakingdiscussion groups. Two phases are importanthere, as follows.

PLANNING GROUPS

This involves consideration of how big thegroups should be, how many are required,who will participate in the groups and howstructured the discussions will be and whatlevel of involvement is envisaged from themoderator. Morgan’s rule-of-thumb answersto these questions suggest that groups projects‘most often (a) use homogeneous strangers asparticipants, (b) rely on a relatively structuredinterview with high moderator involvement,(c) have 6 to 10 participants per group, and(d) have a total of three to five groups per pro-ject’ (1997: 34).

He insists, however, that such norms are apoint of departure rather than some kind ofindustry standard. Interestingly, geographershave used a wide variety of methods torecruit and conduct discussion groups (seeHolbrook and Jackson, 1996). The Jacksonet al. (1999) study on reading men’s magazines,for example, deliberately chose to use pre-established groups rather than groups ofstrangers and completed 20 such groups, eachlasting for around an hour.

CONDUCTING GROUPS

This involves questions about the site of theinterview, the content of the interview andthe role of the group moderator.The site canbe in the home or in a ‘neutral’ venue such asa room in a public building. It should permitthe participants to be comfortable but alsoallow the interviewer to record and moderateproceedings. In terms of content, Mertonet al. (1990) have suggested four basic criteria foran effective discussion group: ‘It should covera maximum range of relevant topics, providedata that are as specific as possible, foster inter-action that explores the participants’ feelings insome depth, and take into account the personal

context that participants use in generatingtheir responses to the topic’ (cited in Morgan,1997: 45).

The role of moderation should include anhonest introduction to the topic, a series ofground rules for the discussion, an ice-breakingdiscussion starter, the use of prompts to con-tinue the discussion and an effective conclu-sion which debriefs participants and informsthem about any further involvement with thetranscriptions and resultant narratives that willbe produced. In each of these aspects there isconsiderable room for experimentation andadaptation to particular circumstances whichcontextualize the research.

Discussion group strategy

It will be clear that in the above discussion wehave skated over a number of important con-ceptual as well as technical issues which needto be resolved critically and carefully in thedeployment of discussion groups. However,our main concern here is to suggest, first, thatfocus group methodologies permit consider-able flexibility in decisions about recruitment,group composition and discussion formatand, secondly, that many of these detaileddecisions tend to fall into place once the strat-egy of the focus group is clear.This is whereBurgess’s (1996) distinction between focusgroups and in-depth discussion groups is ofclear significance, as this distinction offershuman geographers a strategic choice in theuse of discussion group techniques in theirresearch – a choice which may be illustratedfrom Jacqueline Burgess’s own expert studies.She argues that focus groups are most advanta-geous in circumstances where available field-work time is short or when researchconclusions have to be delivered quickly, suchas when urgent policy recommendations arerequired. For example, she was commissionedin 1993 by the Community Forest Unit of theCountryside Commission to undertake a

TALKING TO PEOPLE

161

3151-Ch-05.qxd 3/2/04 2:43 PM Page 161

speedy one-year research study of the socialand cultural dimensions of fear in recreationalwoodlands in the urban fringe (Burgess,1996). With only five months available forthe empirical parts of the research, she com-bined participant observation of smallgroups of people who were taken on a guidedwalk through woodland with a follow-up90-minute focus group in which their directexperiences and the deeper fears which mayhave underpinned that experience were dis-cussed. Given her aim to discover how riskand fear associated with woods are sociallyconstructed, she chose single-gender whiteand black focus groups, recruited thoughexisting contacts and subsequent snowballing.Her experience was that the focus group dis-cussions enabled people to share their viewsand feelings about the issues, and that the roleof moderation was to ensure that the follow-ing mental checklist of issues was addressed inthe discussion:

1. The pleasures of being in woodlands.2. Perceptions of risk.3. Specific features of the wood and how

design and management might alleviatefear.

4. Crime and safety issues in woods com-pared with built environmental settings.

5. Tensions between ‘wild’ and ‘tame’woodlands.

Burgess’s report on this methodology stressedthat

These five topics were discussed by all thegroups, although the depth with whicheach was discussed and the sequence inwhich it was discussed depended on thesalience of the topic for the group. Havingdone the walk together, the groups weremuch more comfortable with one anotherthan would otherwise have been the case,while the shared experience provided acommon focus for discussion in early stagesof the group. (1996: 133)

The results of her research illustrated clearlythat not only do feelings of fear and anxietyvary between groups of different gender, ageand cultural background but also people’sfears are socially constructed rather thanprompted by the nature of woodlands per se.

In general, the focus group setting is bettersuited to investigating publicly available dis-courses than to understanding deep-seatedpersonal and group values. Jackson (2001:207) emphasizes the need to make sense ofthe world through ‘systems of intelligibility’that are shared by individuals and groups whoare part of the same ‘interpretative commu-nity’. He suggests that focus groups method-ology allows researchers to highlight thesediscursive repertoires which characterize theways in which different groups attempt tomake sense of particular texts or issues. Focusgroups equally permit analysis of how indi-viduals and groups adopt particular discursivedispositions to the identified repertoires.

In-depth discussion groups, on the otherhand, are much more suited to research wherean intimate and sustained grouping ofpeople allows participants to share their feel-ings with each other and to create a strongsense of mutual commitment. The group formsits own culture and decides what is to beincluded or omitted from its discourse.Burgess et al. (1988) argue that this philoso-phy is particularly appropriate in the study ofenvironmental values and meanings. In a pro-ject designated to explore local experiencesand local provision of open space in theGreenwich area of London, Burgess and hercolleagues recruited mixed-gender groupsfrom three contrasting localities within thearea, and an additional group of Asian womenwith which to explore the needs of minoritypopulations. Groups met in the local commu-nity centre for six consecutive weeks, for a90-minute session. The research team pro-vided both a ‘conductor’ (moderator) and anobserver.


162

3151-Ch-05.qxd 3/2/04 2:43 PM Page 162

The emphasis in group discussions was toanalyse the broad themes which emergedfrom group analytic processes. Thus Burgesset al. (1988b) identified three importantprocesses:

1. The developmental phases of the groupmatrix where people first establish andthen deepen their relationships by meansof increased trust and intimacy.

2. The manifest and latent content of com-munications, especially where latentmeanings reflect individual and collectiveunconscious feelings.

3. The role of the conductor, with clearboundaries between the work orientationand a therapeutic orientation for the groupdiscussions.

It was from these processes that more system-atic themes relating to open space emerged.Three are emphasized in the importance ofsensory experience, the rate and scale ofchange in the urban environment, and differ-ences in values for, and uses of, green spacesin the contrasting urban and rural areas.Thusapplication of the principles of group analyticpsychotherapy allowed researchers to addressa human geography problem.A strategy of in-depth discussion, therefore, resulted in verydifferent research findings from those that afocus group would have. As Burgess et al.(1988b: 475) reflect:

By means of sharing in the history of thegroup matrix, one witnesses a break downof rational intellectualisation to the discus-sion of emotions and feelings. Over time,too, as people become closer to oneanother and feel safe within the group,they are able to talk about their negativefeelings for environments and landscapes,as well as their preferences … In a once-only group, we could have gained anunderstanding of the consensual values ofa sensitive middle-class group which, aboveall else, valued solitude in open spaces.Only by means of the life of the group

were we all able to realise the extent towhich ‘being alone with nature’ is a verysmall part of highly complex, holistic valuesystems in which open space is chargedwith multiple individual and collectivemeanings.

This strategic distinction between focusgroups and in-depth discussion groups willnot answer specific questions about whathappens if only two people turn up, or howdirective we should be when people in thegroup talk freely about their interests but notabout ours. It does, however, provide a veryimportant framework within which to posekey epistemological questions about thenature of investigative research in group set-tings. Conceptual awareness of what we arelooking for – and here the distinctionbetween generally available discursive reper-toires and deep-seated personal/group valuesis crucial – is a necessary first step to deci-sions about what kinds of group discussionto deploy. Practical methodological queriescan also be assessed against this more con-ceptual framework.A one-off focus group oftwo people may not be as instructive as alarger group for investigating discursiverepertoires and dispositions. However, in-depth discussion with two people couldreveal interesting deep-seated values.A focusgroup that is ‘side-tracked’ may be problem-atic whereas an in-depth discussion uses theinterests of participants in the developmentof a group matrix from which a deeperunderstanding of systematic issues willemerge. Once again we suggest that intellec-tual honesty and criticality are as importantas textbook blueprints in the methodologicalrigour of this form of work. Technicalmethodological details matter not so muchbecause they lead to ‘successful’ research butbecause they determine the nature of thedata being constructed and influence thepurposes to which those data can bedirected.

TALKING TO PEOPLE

163

3151-Ch-05.qxd 3/2/04 2:43 PM Page 163

Ethics: animportant end note

Over recent years it has become increasinglyimportant to emphasize ethical considerationsin the conduct of research (see, for example,Penslar, 1995; Smith, D., 1995; Hay, 1998).Thistrend is to be applauded, not only because ofa previous neglect of ethics, particularly whenusing seemingly objective and remote quanti-tative methodologies, but also because clearinterconnections have needed to be madebetween ethical issues and the kinds of sub-jects to which qualitative inquiry has beendirected. For example, it has been stronglyargued by Mason (1996) that the use of qual-itative methods to construct thick descrip-tions of the lives of individual peopleinevitably raises more ethical difficulties thanthe use of other, more quantitative, researchstrategies in maintaining the confidentialityand privacy of those with a personal involve-ment in the research.

Janet Finch (1993) also reminds us thatthese qualitative methods are being applied tovery different kinds of people. Accordingly,she reminds us that it is now commonplace inresearch to distinguish between powerlesssocial groups – where the need to protect therights to privacy are extremely important –and powerful social groups who are well ableto provide their own protection. In this way,an ethical approach to research will need toembrace notions of power and powerlessness,as Finch herself maintains in the context ofresearch into gender: ‘because issues of powerare central to gender relations, one cannottreat moral questions about research onwomen as if they were sanitised “ethicalissues” divorced from the context whichmakes them essentially political questions’(1993: 178).

We want to stress, therefore, that the use ofinterviews and focus groups as discussed inthis chapter cannot be assumed to represent a

collaboration of equals.While it is possible toimagine that interviewers can in some waysbe overpowered by their subjects, it isabsolutely essential that human geographersare aware of the power which they potentiallyhold over their subjects. Thus, some veryimportant questions need to be raised aboutthe ethical conduct of researchers when inter-viewing and conducting focus groups. Forexample, the questions asked will sometimesdelve into personal issues which could bepainful, traumatic or confidential for the sub-ject. An appropriate response may be to avoidsuch questions or to be free to switch off thetape-recorder and terminate the interview ifthe subject is upset by the issues raised (seeEngland, 1994). The way in which questionsare asked may also be unethical, particularly ifthe questioning is aggressive, underhand ordiscomforting. Here there can be an effectiveimposition of power relations over the sub-ject, not only through the nature of the ques-tioning but also in a broader sense by the wayin which a rigid agenda is set for the inter-view and strong control is established over theresulting ‘data’. For example, there is strongevidence to suggest that everyday rules andhabits of conversation between men andwomen operate such that the effects of socialpower will produce bias in favour of the viewsof men over those of women (see Fishman,1990). Moreover, Ann Oakley’s (1979) workwith women about their experiences of tran-sition into motherhood rejected ‘textbook’interviewing, which pointed to a one-wayprocess of gaining information from subjectsbut not answering their queries. She arguedthat engagement should become a key principleof feminist research (see also Oakley, 1984;1990).

These social theories of power raise obvi-ous moral and political questions about prac-tising human geography. In terms of genderedpower relations, a praxis of ‘doing feministresearch’ (Roberts, 1990) has emerged.There


164

3151-Ch-05.qxd 3/2/04 2:43 PM Page 164

are, however, other axes of social and culturalpower which present equivalent challengesfor researchers working with issues of, say,race, sexuality, children, disability, poverty andhomelessness.Acknowledgement of intersub-jectivity in interviews must include referenceto the uneven distribution of social power insociety as well as to the particular relations ofdramatic collusion which occur in the inter-view itself.

Human geographers have sought to answerthese questions about ethics in two significantways. First, there have been attempts to gen-erate genuine ‘research alliances’ betweenresearcher and researched, and in particularbetween interviewer and interviewee. Pile(1991: 467), for example, has suggested that inthe past ‘geographers have acted as if theystand outside the specific historicity and geo-graphity of their subjects; this has enabledthem to comment on the reality of the sub-ject’s view of their own situation, while notallowing the subject’s valid versions of reality’.He draws on psychoanalytic notions of the‘therapeutic alliance’ to suggest the possibilityof fostering ‘research alliances’ between theresearcher-interviewer and the researched-interviewee. Such alliances entail a process bywhich the balance of power of the emotionalrelationship in an interview is transferredfrom the interviewer to the interviewee, thuscreating the interview as a safe space in whichdeep, mutual and trustworthy interactions andunderstandings emerge. This outcome permitsthe subject’s version of reality to be aired andvalued while the interviewer’s project benefitsfrom a more sensitive and critical grasp on theintersubjectivities involved.

It is not necessary to buy into psychoana-lytical ideas to applaud the basic sentiments ofthe ‘research alliance’. Indeed, Hester Parr(1998b) has warned against the implicationsof figuring the research interview throughthe psychoanalytic lens of doctor (therapist,counsellor)/patient (psychologically damaged

individual) relations. In part she does thisthrough reporting on her experience of inter-viewing people with mental health problems,many of whom have been through suppos-edly therapeutic interviews. From this contextshe argues for a more flexible approach tointerviewing which draws on the spirit butnot necessarily the letter of good psycho-analytic practice.

The second response to ethical questionshas been a discussion about how to constructa widely agreed set of ethical guidelineswhich can be used to establish appropriate‘research alliance’ relations between researchersand the people they work with. For example,four such guidelines are recognized byHammersley and Atkinson (1995):

1. Informed consent: The researcher shouldinform the researched about the researchin a comprehensive and accurate way, andthe researched should give their uncon-strained consent.

2. Privacy: The researcher should resist makingthings public which are said and done forprivate consumption.

3. Harm: The researcher should avoid nega-tive consequences both for the peoplestudied and for others.

4. Exploitation: The researcher should avoid‘using’ respondents to gain informationwhile giving little or nothing in return.

We would add a fifth concern to this list:

5. Sensitivity to cultural difference and gender : Theresearcher should be sensitive to the rights,beliefs and cultural context of the researched,as well as to their position within patriarchalor colonial power relations.

The ethnographic turn in human geographyhas brought with it two particular strands ofresearch which make particular demands onthis kind of ethnographic strategy.The first isan acknowledgement of the need for reflexivity

TALKING TO PEOPLE

165

3151-Ch-05.qxd 3/2/04 2:43 PM Page 165

Box 5.7 Challenges to ‘textbook’ ethical guidelines

Informed consent?

Gaining permission for an interview is not a cut and dried event at the start. Even withagreement to conduct an interview, it seems as though the interviewee doesn’t reallyrelease their permission until they begin to trust you. Equally, their responses are gen-erally more interesting when permissive trust has been established. I certainly encoun-tered instances of Punch’s mutual deceit, where I was on my best ethical behaviour,often asking questions which were peripheral to my main interest, until I was beingtrusted, and effectively given permission to be more personal and directive. There is nodoubt that I was more prone to exaggeration, concealment and other trust-seekingdevices during the early part of interviews. This is a really distressing, though not per-haps uncommon, conclusion to reach (see Punch, 1986).

Privacy?

I had many opportunities to take photographs of the homeless people I met, and I sensedthat many would have given me permission so to do. Nevertheless, there was somethingin my moral/ethical map which directed me away from this practice. Was it about thepotential transgression of privacy? Was it a judgement about exploiting the subjects ofthe photographs? Maybe. However, all of these factors rolled into an imagination of howin the future I might give talks on the research. Somehow I knew that I would not wantto flash up images of these people in order to impress or stimulate an audience. It was thissure knowledge that worked back into the research practice of not using photography.

Harm?

I had been quite keen to talk to James, and, in the course of a number of fairly super-ficial conversations, he did express a willingness (even desire) to tell me his story andto put it to tape. In the end, it took three or four different ‘attempts’ to actually sitdown and talk with James, because each time things had worked out such that thiswould have been rather inappropriate. On one occasion, anxiety over an impendingcourt appearance had just got too much; on another James was struggling over theanniversary of a friend’s untimely death; and on two others, I came to the conclusionthat had I gone ahead it could have jeopardised James’s efforts at the time to quit alco-hol. I would not want to pretend that these were necessarily easy decisions to take, northat the whole process was not extremely frustrating; for in the end it could havemeant ‘losing’ James’s story (and there were others that I did lose in this way).

Exploitation?

I had written recently about the dangers of exploitative research tourism, and theunethical nature of ‘flip’ ethnographies in which academics flip in and flip out of thelives of other people staying just long enough to collect juicy stories. Yet I found myselfdoing just that in a particular village. Staying for just a weekend I had a chanceencounter with two homeless men, and their stories have become integral to our find-ings. My commitment to them was so small, and the fact that I have a more regularcommitment to homeless people in Bristol seemed at the time to be a disconnectedexcuse for flip ethnography (see Cloke, 1997).

As I entered this area I become aware of ‘Tony’, who was begging under the mainarchway. Here, then, was the opportunity to produce an image relevant to my forth-coming presentation. However, I was also aware that the local press had been some-what hostile to people begging on the streets of this town over recent months. As such,


166

3151-Ch-05.qxd 3/2/04 2:43 PM Page 166

Box 5.7 (Continued)

I needed to give some thought to how I would approach Tony, introduce myself andthe project, and ask permission to take his photograph … I first placed some money inTony’s polystyrene cup… and crouched down next to him. I began making conversation …I asked Tony whether he had heard of our homelessness project … I provided somebrief details … how we had been trying to talk to homeless people in the local areaand record their stories. He didn’t seem that interested … I explained to Tony that I wasalso keen to collect some visual images to place alongside the stories, and as suchwished to take a photograph of him. ‘You’re not working for the papers are you?’, heasked. I stressed that I was conducting research. ‘OK, but I’m not giving you my name’.‘That’s OK, I don’t want your name, just your picture!’ We reached an agreement andI moved across the cobbled street to compose my image. It was a hurried shot …I returned across the street, thanked Tony for his help and wished him well.

Sensitivity to cultural difference?

The stronger part of my self-identity is that of a single parent, and my own strugglewith the negative representations of my ‘group’ in the media and wider society. WhileI have never been homeless, I have, on occasion, had difficulty (as many people do) withtrying to maintain my rent payments and provide for my family. These experiencesallowed for a certain degree of identification with my interviewees but, in certain sit-uations, this proved problematic – at times it felt as if I was at once the researcher andthe researched, perhaps too much of a ‘insider’ in some ways.

Source: Cloke et al. (2000a)

which has resulted in a conscious analyticalscrutiny of the researcher’s self.The second isa growing acceptance of Bakhtin’s (1986) dia-logics by which researchers seek to encounterotherness through the potential of dialogue.Together, reflexivity and dialogics suggestresearch to be a process rather than a product.As England (1994) has argued, research isthereby transformed by the inputs of thosewho previously have been regarded as the‘subjects’ of research. Furthermore, theresearcher is now recognized to be an integralpart of the research process. Hence the inter-view and the focus group become unfoldingperformances, permitting the telling of partic-ular narratives rather than quarries for thetruth.

These changes have direct ethical conse-quences and often serve to problematize any

straightforward application of the ethicalguidelines listed above.To illustrate these dif-ficulties we turn to an account by Cloke et al.(2000a) of the ethical issues in research withhomeless people in rural areas. Extracts fromthe research diaries of individuals in theresearch team (Box 5.7) clearly indicate thatin each case the five ‘textbook’ principles forethical interviewing were challenged by thepracticalities of interviewing homeless people.It is clear that standard ethical approaches tothe relationship between the ‘researcher’ andthe ‘researched’ are both turbulent and prob-lematic.The practice of human geography cannever be a neutral exercise. As Cloke et al.(2000a: 151) suggest:

For good or ill, the very act of entering theworlds of other people means that theresearch and the researcher become part

TALKING TO PEOPLE

167

3151-Ch-05.qxd 3/2/04 2:43 PM Page 167

co-constituents of those worlds. Therefore,we cannot but have impact on those withwhom we come into contact, and indeed onthose with whom we have not had directcontact, but who belong in the social worldsof those we have talked to … Ultimatelysuch matters are entwined with the need toavoid exploitation of research subjects, andto give something back to them throughthe research process. These are matters ofcomplex negotiation … to suggest that adegree of negotiation does not regularlytake place over differential ethical risk, inorder to garner material with which toachieve certain ends, is to hide behind ethi-cal standards so as to obscure the real-timedilemmas of research.

These issues of research ethics are crucial tothe undertaking of the kinds of methodolo-gies discussed in this chapter. In order to dealappropriately with the responsibilities wehave in the worlds that we so readily gatecrash

in the name of research, we require moredetailed and critical ethical maps by which wecan traverse this difficult terrain.

Notes

1 See, for example, Moser and Kalton (1971),Bynner and Stribley (1978), Dillman (1978),Dixon and Leach (1978), Sudman andBradburn (1982), Rossi et al. (1983), Fink andKosecoff (1985), Bell (1987), Fowler (1988),De Vaus (1991), Oppenheim (1992), Foddy(1993), Czaja and Blair (1996), Lindsay (1997)and Parfitt (1997).

2 See Downs (1970) as an example of the useof semantic differential scales in retailinggeography.

3 See Oppenheim (1992: chs 10 and 11) for afull account.

4 See, for example, Goldman and McDonald(1987), Greenbaum (1988), Kreuger (1988),Stewart and Shamdasani (1990) and Morgan(1993; 1997).


168

3151-Ch-05.qxd 3/2/04 2:43 PM Page 168

Introduction: what isethnography and howcan it be geographical?

Elsewhere in this book we have begun ouranswers to such questions by going to Greekroots. And this is quite useful. While geo-graphy literally means ‘earth-writing’, ethno-graphy means ‘people-writing’ (Hoggart et al.,2002), the kind of ‘writing’ most often associ-ated with early and/or traditional anthropolog-ical fieldwork. Here, Western researcherstypically spent a year or more living in far-off,non-Western, small-scale, isolated, rural com-munities (Jackson, 1983).They learnt local lan-guages, watched and participated in day-to-dayactivities and (sometimes with research assis-tants and translators) entered into conversationsabout these activities with local people. Theyasked what was going on and why, took noteson what they saw and heard, sketched or tookphotographs of particular people, places orevents, kept tallies, collected objects and any-thing else that could help to record, make bet-ter sense of and/or better represent to otherswhat they were doing and learning. Ideally,such extended, intensive and detailed workwould enable the researcher to understand that‘people’s’ ‘worldview’, ‘way of life’ or ‘culture’,like an inquisitive ‘insider’ (Herbert, 2000).Ethnography has its roots in European imperi-alism and in the roles that academic disciplinessuch as anthropology and geography played in

this (see Asad, 1973; Livingstone, 1992;Gregory, 1996). Subsequent adoptions, adapta-tions and transformations, however, mean that,today, its practitioners, politics, fieldsites, theo-retical content and purposes complicate thisstereotype considerably.

This doesn’t mean that our opening ques-tions can’t be answered.Wherever it is practised,‘ethnography’ has common characteristics.First, it treats people as knowledgeable, situ-ated agents from whom researchers can learna great deal about how the world is seen, livedand works in and through ‘real’ places, com-munities and people. Secondly, it is anextended, detailed, ‘immersive’, inductivemethodology intended to allow groundedsocial orders, worldviews and ways of lifegradually to become apparent.Thirdly, it caninvolve a ‘shamelessly eclectic’ and ‘method-ologically opportunist’ combination ofresearch methods but, at its core, there must bean extended period of ‘participant observa-tion’ research ( Jackson, 1985: 169). Fourthly,participant observation uniquely involvesstudying both what people say they do andwhy, and what they are seen to do and say toothers about this. Fifthly, ethnographyinevitably involves tricky negotiationsbetween researchers’ words and deeds, bothwithin standard field settings and betweenthese and other settings. Finally, therefore,ethnography involves a recognition that itsmain research tool is the researcher and the

Doing Ethnographies6

3151-Ch-06A.qxd 3/2/04 2:43 PM Page 169

ways in which he or she is used to acting inmore familiar circumstances and learns to actin the often strange and strained circum-stances of his or her research settings. Heredifferently ‘theorized’ (academically andotherwise) and/or taken-for-granted world-views, ways of life, self-understandings, rela-tionships, knowledge, politics, ethics, skills,etc., are accidentally and deliberately rubbedup against one another.

Ethnographic findings are not therefore‘realities extracted from the field’ but are‘intersubjective truths’ negotiated out of thewarmth and friction of an unfolding, iterativeprocess (Parr, 2001; Hoggart et al., 2002). Inmany ways, Steve Herbert argues ‘all humansare ethnographers whenever they enter a newsocial scene; one moves from outsider toinsider as one comprehends the world fromthe insider’s point of view’ (2000: 556). Thismay be true, but novice ethnographers mustrecognize, develop, complement and sometimesunlearn existing attitudes, habits, sentiments,emotions, senses, skills and preferences.A goodethnographer is someone willing and able tobecome a more reflexive and sociable versionof him or herself in order to learn somethingmeaningful about other people’s lives, and tocommunicate his or her specific findings,including their wider relevance, to academicand other audiences.

Ethnography’s methods other than partici-pant observation (and what is involved in under-standing the data generated through them andin representing that understanding) are dealtwith elsewhere in this book. Here, therefore,we focus on issues that geographers havebrought into, and got out of, ethnographicresearch and the participant observation at itsheart. Recent geographical reviews havepointed out that, although participant obser-vation is now a standard approach in the dis-cipline’s methodological literature,1 only atiny proportion of mainstream publicationsappear to be based on its use. For instance, of

the human geography articles published inthe Annals of the Association of AmericanGeographers and Society and Space between1994 and 1998, only 3.8% and 5%, respec-tively involved participant observationresearch (Herbert, 2000). Moreover, only ahandful of ‘proper’, book–length ethnogra-phies by geographers have been published inthe past 30 years, and these have usually beenrevised versions of PhD theses.

This poor showing, according to Herbert(2000), results from a general geographicalscepticism about ethnography’s merits. Crang(2002) has added more practical explanationsincluding problems of arranging and con-ducting extended periods of fieldworkaround family, work and other commitments,and of funding bodies’ requirements forresearch proposals specifying ‘outcomes’which are difficult to predict when using ‘amethod in which the units and boundaries ofinquiry are as little predetermined as possible’(Jackson, 1983: 45). Under these circum-stances, it is not surprising that qualitativeresearch in human geography is primarilybased on interviewing or, at best, ‘short-termethnographies’ (Thrift and Dewsbury, 2000:424; Crang, 2002).

Ethnography is often cast as entering geog-raphy as a response to the dehumanizingeffects of the quantitative revolution in the1960s (see Billinge et al., 1984). So, the storygoes: in the 1970s it was used by humanisticgeographers keen to put ‘real people’ backinto the discipline and their work subse-quently inspired the widespread adoption ofqualitative methods during the cultural turnthe early 1990s (see Philo, 1991; Cook et al.,2000b). Historians, however, of geographyhave argued that the sponsorship and publish-ing of ethnographic studies – i.e. ‘first handaccounts of travels and explorations’(Livingstone, 1992: 161) – go much furtherback. In a 1899 edition of the Journal of theAmerican Geographical Society of New York, for

PART 1 Constructing geographical data

170

3151-Ch-06A.qxd 3/2/04 2:43 PM Page 170

example, Titus Munson Coan set out in his‘Hawaiian ethnography’ the ‘physique’, ‘men-tal and moral traits’, ‘language’, ‘religion’,‘social usages’ and ‘arts’ of the ‘brownPolynesian’ who was ‘everywhere substantiallythe same – or was before the obliteratinghand of European civilisation swept over thefinely graven plate of Nature’s etching’ (1899:24). Between 1906 and 1910, 0%, 3% and 14%of the articles published in GeographischeZeitschrift, The Geographical Journal and theBulletin of the American Geographical Society wereclassified as ‘anthropology and Ethography’ and60% of articles in The Geographical Journal werecategorized as ‘exploration and mapping’(Close, 1911: 743; Genthe, 1912). This there-fore means that participation, observation and‘ethnography’ have, one way or another, com-prised a core methodological tradition ingeography since the beginning of the nine-teenth century (Livingstone, 1992; Anderson,1999). This tradition also includes numerouswomen’s accounts of ‘travel and exploration’which, historically, were not deemed ‘propergeography’ or ‘proper anthropology’ by thesedisciplines’ male-dominated professional bod-ies.2 The work of ‘geographers’ as diverse asMary Kingsley, Titus Munson Coan, CarlSauer and Bill Bunge could therefore be seenas part of geography’s long-running ethno-graphic tradition.3 The quantitative revolution‘may, therefore, be seen as an aberration ratherthan a revolutionary paradigm … that itwas claimed to be at the time’ (Winchester,2000: 14).

In this book we want to make a contribu-tion to the kinds of ethnographic researchthat geographers want to do now. And wewant to offer more than a potted method,which just says what ethnography is, what itinvolves and how to do it well. Instead, wewant to outline ethnography as a long-runningand distinctive geographical practice that haslearnt much from, and can contribute muchto, wider interdisciplinary debates.

For the past 30 years, we argue below, theexploration of relationships between people,place and space has been the hallmark of geo-graphical ethnography. However, the ways inwhich these relationships have been theo-rized, designed into research practice, ‘discov-ered’ (or not) in the ‘field’ and written up (ornot) as ‘fieldnotes’ have changed as the disci-pline has changed.All sorts of fascinating the-oretical ideas, impressive ethnographic work,missed opportunities and thoughts for thefuture have come and gone.

Currently, participant observation is enter-ing a new phase of popularity, with agenda-setting geographers making fresh calls for itsuse (e.g., Thrift, 2000b; Smith, S.J., 2001;Hoggart et al., 2002). But these latest ‘cuttingedge’ ethnographies could learn a great dealfrom older, pioneering studies. There areimportant connections to be made betweenhumanistic ethnographies from the 1970s and1980s and the ‘new’ ethnographies subse-quently advocated as geography took its cul-tural turn. Thus, below, we outline andexemplify key debates about theory and prac-tice within and between these ethnographies,use these to provide some top tips forprospective ethnographers and then finishwith a discussion of ‘field-noting’ – the pro-duction of data from participant observationresearch.

Geography’s humanisticethnographies

Humanism advocates a people-centredunderstanding of the world and its workings,and argues that humanity must be central tointellectual endeavour (Cloke et al., 1991).Geographers’ commitment to these idealsdates from the late nineteenth century andtheir influence connects the educationalhumanism of Peter Kropotkin and H.J. Fleure,Paul Vidal de la Blache’s school of la géographie

DOING ETHNOGRAPHIES

171

3151-Ch-06A.qxd 3/2/04 2:43 PM Page 171

humaine, Carl Sauer’s Berkeley School ofcultural geography, J.K. Wright’s (1947)‘geosophy’ and David Lowenthal’s (1961)writings on ‘geographical epistemology’.4 Allwere concerned with understanding people’sminds, hearts, imaginations and their com-bined, ‘properly human’, geographies. Theyquestioned relationships between ‘the worldoutside and the pictures in our heads’(Wright, 1947, cited in Jackson and Smith,1984: 21), and attempted to understandhuman agency in relation to experiential,societal and physical ‘structures’. Some ofthese ideas were absorbed into the quantita-tive revolution’s behavioural geography,but a self-consciously humanistic geographyemerged in the early 1970s as an ethical andphilosophical critique of the dominant spatialscience tradition in human geography whichwas ‘insensitive to the humanity of both geo-graphers and people under study’ (Clokeet al., 1991: 67).

Here, in response to spatial scientists’claimed ‘distance’ from what they studied,humanistic geographers emphasized thehumanity of all researchers and how thisunavoidably entangled them with theirobjects of study. In response to spatial scien-tists’ claimed neutrality, objectivity and abilityto work without values, they emphasized thecentrality of human values in everyday lifeand argued that claims to rise above themwere not only untenable but also unethical,irresponsible and ‘cloaked their complicity inglobal inequality and exploitation’ (Rose,1993b: 42). Humanistic geographers arguedthat ‘Quantifiers and spatial analysts havefailed to make a single important discoveryeither about mankind or about the humanenvironment’ (Prince, 1980: 294). ‘Human’geography therefore needed a ‘radical reori-entation of thought and vision’ (Buttimer,1976, cited in Rose, 1993b: 43). This reori-entation involved heated, often acrimonious,debates, with accusations of ‘irrelevance’ and

‘hypocrisy’ frequently made on the pages oflearned academic journals and books.However, what these debates contributed togeography and to the foci of its ethnographicresearch was a distinctive and important set ofarguments about the inescapable entangle-ment of people and place.

To those who know their history of Anglo-American human geographic thought, muchof this is a familiar story. Readers might there-fore expect us to start outlining humanisticgeography’s main philosophical critiques ofspatial science and its attempts to imagine anew, fully ‘human’ geography. They mightexpect us to discuss the importance of phe-nomenology and existentialism, the work ofHusserl, Heidigger, Merleau Ponty and Sartre,among others and some key principles pickedup from this work. These might include theblurring of positivism’s object/subject andfact/value distinctions to take a positionwhere ‘one’s goals, intentions, and purposescan never be totally isolated from one’s expe-rience and knowledge of the world’(Entrikin, 1976: 626) and the belief thatpeople’s goals, intentions, purposes, experi-ences, knowledge and bodily movementswere always ‘incarcerated’ in places as diverseas ‘movements, art-works, buildings … cities’and other entities that ‘organise space as cen-tres of meaning’ (Tuan, ested in Entrikin,1976: 626). We might then say somethingabout humanistic geographers’ concern tounderstand the ‘human experience of spaceand place’ and the, often taken for granted,‘geographies of the lifeworld’. Readers mightthen expect us to conclude by discussing aslew of ethnographic studies that groundedand extended these ideas in concrete placesand everyday lives.This empirical connectionwas very rarely made because humanisticgeography was, however, a largely reflexivephilosophical exercise.

Humanistic geography’s primary concernwas to understand the human experience of


172

3151-Ch-06A.qxd 3/2/04 2:43 PM Page 172

space and place, the ‘geographies of thelifeworld and so on – in other words, the essen-tial qualities of the humanity which spatialscience had brushed aside. For example, its phe-nomenological method involved, in principle,researchers being able to ‘bracket out’ all worldlypresuppositions in order that the raw,pure,direct‘essences of man, space, or experience’ could berevealed (Tuan, cited in Entrikin, 1976: 628).This was an intuitive, introspective, idealistic,conceptual process which did ‘not depend uponempirical verification’ (Entrikin, 1976: 629).5 Sowhile being philosophically rich, humanisticgeography’s intuitive, introspective, idealistic andessentialist approach – ironically – kept it distantfrom the people whose humanity it cared somuch about (Eyles, 1986).

It seems, therefore, that ethnographicresearch should, in principle, have been part ofthis new,humanistic geography but, in practice,was not (Jackson and Smith, 1984).Yet thingsare not so disappointing as they seem becausehumanistic geography

made its most significant contribution tohuman geography not in directing theattention of a few researchers to the deep-est phenomenological and existential con-nections people have with their places, butin sensitising numerous researchers … tothe everyday and yet often quite intimateattachments all sorts of people (and not justphilosophically inclined scholars) have tothe places that encircle them. (Cloke et al.,1991: 81).

Humanistic geography literature producedonly a handful of ethnographies detailingpeople’s attachments to place, but none drewheavily on the philosophies and phenomeno-logical methods outlined above. Only threewere discussed in any significant detail: DavidLey’s (1974) The Black Inner City as FrontierOutpost, Graham Rowles’ (1978a) Prisoners ofspace? and John Western’s (1981a) OutcastCape Town. All three were revised AmericanPhD theses in urban social geography and, aswell as being published in their own rights,

were also showcased in edited collections ofhumanistic geography at the time. Theseethnographic chapters often stood out fromthe philosophical crowd, and gave onereviewer ‘greater confidence in the ability ofthese humanists to open new realms of geo-graphical discovery’ (Prince, 1980: 295).

These three ethnographies are nowregarded as milestones in geographical inquiry.Two have been ‘revisited’ as ‘classics in humangeography’ in Progress in Human Geography(Ley’s in 1998 and Western’s in 1999), and allthree can be seen as important precursors ofgeographers’ current interests, first, in represen-tation–reality, structure–agency and space–bodyrelations and, secondly, in the practice of ethno-graphic research.They are discussed below bothto demonstrate their impacts and influences onthe discipline and to illustrate how ethno-graphic research proceeds and what the finishedproducts can look like.

The Black Inner City asFrontier Outpost

David Ley’s (1974) The Black Inner City asFrontier Outpost: Images and Behaviour of aPhiladelphia Neighbourhood was submitted as aPhD thesis at Penn State University in 1972and revised and published two years later as anAssociation of American Geographers’ mono-graph. At the time of its publication it wasseen as ‘a solid research achievement and amodel for further studies’ (Palm, 1975: 573).Addressing ‘radical concerns about racialisedinjustice’ at the time, Ley’s book has beenrecently described as geography’s ‘first and inmany ways most lasting example’ of ethno-graphic social research and as ‘signalling aturning point in the development of [urban]social geography’ (Jackson, 1998: 76, 75).

Ley lived in Philadelphia’s black inner-citycommunity of Monroe for a six-monthperiod in 1971, and then spent another two

DOING ETHNOGRAPHIES

173

3151-Ch-06A.qxd 3/2/04 2:43 PM Page 173

and a half years writing proposals for itscommunity association. While there, heparticipated in community activities, becameacquainted with residents, undertook a ques-tionnaire survey of 116 residents and exam-ined numerous newspaper articles and ‘cityfiles’ (Darden, 1976; Palm, 1998). Behind thescenes, he drew on an ‘eclectic range of intel-lectual sources including a humanistic readingof the French tradition of la géographiehumaine, an ethnographic interpretation of theChicago School of urban ecology, combinedwith an interest in various strands of pheno-menology together with a strong influencefrom behavioural science and environmentalpsychology’ ( Jackson, 1998: 75; see alsoJackson and Smith, 1984; Cloke et al., 1991).He also drew inspiration from more ‘off thewall’ sources, such as the travels of explorerJames Cook, cybernetics and accounts ofCaesar’s Gallic Wars – the source of his ‘frontieroutpost’ model (Ley, 1998; Palm, 1998).

What this meant, he has written recently,was that the book was ‘too conceptually busy’(1998: 79). It was also busy in a methodolog-ical sense. Along with the tapes and notesfrom his participant observation research, heanalysed the results of his questionnaire ‘usingsemantic differential tests to determine stressprofiles, and factor analysis to derive compo-nents of a “stress index”’ (Palm, 1998: 77). Inthe mid-1970s, this range of theoretical andmethodological approaches was seen as anexemplary combination of ‘the rigour of pos-itivist science with the insights of experientialexposition’ (Palm, 1975: 572). Twenty-threeyears later, Palm (1998: 77) described this as‘the highest possible compliment I could[pay] at the time’.

Ley’s findings were presented in two parts.In the first – ‘The shadows’ – he examineddominant ‘outsider’ representations of blackinner-city neighbourhoods in the USA asunited and oppositional to mainstream whitesociety. Here, he examined both mass media

and academic representations, most often puttogether ‘by white investigators who havenever lived in a black community but whonevertheless emerge as so-called authoritieson the life and behaviour of blacks’ (Darden,1976: 117), but also by ‘members of the city’sBlack Muslim population’ ( Jackson, 1998: 76).In this section, Darden comments:

He concludes that the popular white imageof black America is false, for it is groundedin history and American institutions, nur-tured by social and spatial segregation, andtraditionally reinforced by the white mediafrom which most Americans derive theirmost important and in many cases onlyinsight into black America. Ignorance andlack of firsthand information have helpedwhite Americans believe in the existence ofa monolithic, hostile, organised black com-munity that is preparing for revolution incollaboration with outside agitators …Occasional visits to the inner city will notchange these false images. (1976: 116)

The sustained and systematic explorations ofan urban ethnographer could, however,change those images. Thus, in the secondsection of Ley’s book – ‘The real’ – he out-lined the findings of his research within thecommunity (i.e. his ‘insider’ perspective).Here, in what remains the most ‘compelling’part of the book ( Jackson, 1998), he describedin rich ethnographic detail a communitydivided by ‘gang territorialization’, and howthe everyday lives of its people were full ofcomplexity, suspicions, mistrust, disunity,‘unpredictable alliances’, ‘uncertainty andeven chaos’ (Palm, 1975: 572).

According to Ley, what characterized thiscommunity was not a black oppositional unitybut ‘individuation’ or the ‘retreat into self andfamily as a means of coping with the environ-ment’ (Palm, 1975: 573). And this was wherethe frontier outpost model came in. AsDarden (1976: 117) explained, the

social group, the community association,the street gang – all may be frontier out-posts. Each builds up an internal defensive


174

3151-Ch-06A.qxd 3/2/04 2:43 PM Page 174

repertoire, each stockades itself and pursuesa policy of individuation, each seeksto aggrandize itself, sometimes at theexpense of members of its own kind, eachdevelops a group image, its own symbolsand stereotypes.

Representations and realities of the blackinner city were therefore poles apart, and thiswas something which a white outsider such asLey could only have discovered throughdeveloping an insider’s perspective. However,as he noted in the book’s Afterword, while heprised ‘representation’ and ‘reality’ apart in itstwo main sections, he could not be sure thathe had done so in practice.

In 1998 he recounted the ‘gropings’ of hisearly weeks of participant observation whenhe found that his ‘socially constituted stock ofknowledge … was challenged at every turnby everyday life in North Philadelphia’ (p. 79).The role of social scientists, he had written inthe book, was to ‘search for order in [such]chaos’ (cited in Palm, 1998: 78). But thismeant that, while his conclusions clearlyderived ‘from the reality of the area studied’,they also derived from:

• ‘the postulates of the scholar’s conceptualframework, postulates which are probablyethnocentric, or, at best, limiting’ (Palm,1975: 572);

• his presence in the field as ‘the neighbour-hood exotic’ whose British accent ‘localadolescents sought to mimic … when wepassed on the street’; and, in particular,

• his ‘two key informants, a middle-agedwoman and an adolescent male, bothactive in the neighbourhood’ (Ley andMountz, 2001: 239).

Despite these criticisms, however, when PeterJackson revisited Ley’s book in 1998, heargued that it was

justifiably regarded as a classic because ofthe way it embodied and advanced … trends

[in ethnography and radical anti-racistpolitics], prefiguring many subsequent devel-opments in the discipline: the concern forsocial theory, for qualitative research meth-ods, for the politics of representation and therise of cultural and media studies. (p. 75)

Outcast cape town

While Ley’s ethnography was based in a blackinner-city neighbourhood in the USA whichseemed to be fragmenting itself, Western’s(1981a) Outcast Cape Town was based in a‘coloured’ inner-city neighbourhood inSouth Africa which was being forcibly frag-mented as a result of apartheid’s 1950 GroupAreas Act. South Africa’s two and a quartermillion ‘coloured’ people were

the creation of miscegenation amongWhites, who established themselves inSouth Africa after 1652, moving towardsthe interior from Cape Town; their slaves,who were imported mainly fromMadagascar and the East Indies; and theautochthonous Khoisan peoples, otherwiseknown as Hottentots and Bushmen, nowalmost completely wiped out. Today themajority of Capetonians are Coloureds,neither Whites nor Black Africans … Thirtypercent of all the coloureds in the republiclive in Greater Cape Town, forming aboutsixty percent of its total population of overone million. (Western, 1978: 298)

Despite the fact that the people of Cape Town‘range[d] along a phenotypic continuum fromthose who appear to be Nordic Whites tothose who appear to be Black Africans’(Western, 1978: 298), the governmentimposed ‘a discrete hierarchy of White-Brown(Coloured)-Black (Bantu) categories throughapartheid legislation’ (1978: 298). Residentialsegregation of black South Africans had beenin place for decades. The Group Areas Acttherefore was aimed at those ‘in the middle’ –the ‘coloureds’ whose genetic make-up mademost nonsense of apartheid’s discrete racialcategorizations. Couched often as the clear-ance of slums adjacent to white districts,

DOING ETHNOGRAPHIES

175

3151-Ch-06A.qxd 3/2/04 2:43 PM Page 175

neighbourhoods were systematically razedand their coloured inhabitants were relocatedto ‘new ghettoes … typically situated severalmiles from the white towns, with a bufferzone in between’ (van den Berghe, cited inWestern, 1978: 300).

As reviewers of Western’s work noted thiswas a ‘chilling’ topic (Pirie, 1984: 259), and hisaccounts of its history, justifications and effectson the people of Mowbray (his case-studyCape Town neighbourhood) carefully andvividly captured this. Some saw him as argu-ing ‘from the position of an objective observerbasing his conclusions upon historical evi-dence, statistics, and personal interviews’(Entrikin, 1979: 256), while others praised hiswork as ‘no clinically-detached survey’ (Pirie,1984: 259), as ‘original, frank … modulatedrather than angry … earnest … sincere’ (Pirie,1984: 260) but ‘hard hitting’ ( Jones, 1980:115). For one, it combined ‘serious scholarshipwith passion in a way that I have never seenthem intertwined’ and, despite some unusualsource materials, should therefore ‘be read byevery geographer, by every graduate studentin geography, and by many nongeographers.The book is a superb example of human andhumanistic geography, of painstaking, exhaus-tive, and explorative research, of penetratinganalysis, and of civilised, literate reporting. Inshort, the book is geography at its relevantbest’ (de Blij, 1982: 493).

Western had not set out to work in SouthAfrica, let alone in Cape Town. He initiallywanted to work in Burundi but was unable tobecause of the ‘African holocaust’ of the early1970s in which 300,000 people were killed(de Blij, 1982). He first saw South Africa inJuly 1974, ‘a time that in retrospect seems thehigh water-mark of apartheid’ and throughhis experiences there he became ‘really impas-sioned about “politics”’ (Western, 1999: 425).He explained, it ‘was only when rummagingaround apartheid’s legacy of empty andgentrifying formerly “coloured” areas of

Mowbray that I sensed something wicked hadthis way come’ (Western, 1999: 425).

In 1999 Western mentioned two peoplewho had made a big difference in his work.The first was Robert Coles, whose passionatetalk at UCLA in 1973 had been inspirationfor Western in the audience. Coles’ tales ofethnographic research had fired Western upand made him ask ‘Couldn’t I, as an inquisi-tive urban geographer, do such hands-on,please-tell-me-in-your-own-words research?’(Western, 1999: 426). The second was VictorWessels, a South African ‘Unity Movementleader then banned under the Suppression ofCommunism Act’, and one of the first peoplehe met in Cape Town (Western, 1999: 426).As Western recounted,‘He befriended me andchallenged me to, damn it, do something’.‘Areyou indeed a so-called scholar?’Wessels askedhim. ‘Then write your best contribution toour struggle’ (cited in Western, 1999: 426).Through his extended ethnographic researchthat struggle became ‘real’ for Western, aswhat he saw, heard and read profoundly dis-tressed him.

When Outcast Cape Town was revisited as a‘classic in human geography’ in 1999, DavidSimon (a geographer who had grown up inCape Town and been radicalized as a child bythe processes described in Western’s book)recalled his ‘admiration’, as a postgraduate, ‘atthe quality of writing and vividness of theaccounts of gross inhumanity and sufferingconveyed’ (p. 423). For many South Africansocial scientists, this was their ‘first exposure tohumanistic geography’ (Dewar, 1999: 421), andit created ‘a sense of frustration, even indigna-tion, on the part of some … that a USA-basedBritish geographer had come to Cape Townand produced an acclaimed masterpiece intheir own back yard’ (Simon, 1999: 424).

Like Ley’s account of Monroe, Western’saccount of Mowbray drew on, and ‘showcased’,the sociologically oriented phenomenologyof Alfred Schutz (Entrikin, 1979). Like Ley, he


176

3151-Ch-06A.qxd 3/2/04 2:43 PM Page 176

juxtaposed outside representations of CapeTown with the real social geographies on theground. And, like Ley, his ethnography wasmethodologically eclectic and busy. OutcastCape Town combined ‘every skill at the geog-rapher’s disposal – cartographic, quantitative,literary, and visual’ and a wide range ofsources, ‘from Gamtaal poetry to governmen-tal statistics, from questionnaire-generateddata to real-estate advertisements in newspa-pers’ (de Blij, 1982: 493). Western even‘indulge[d] in some informal modelling,reconstructing the ideal apartheid cityaccording to the canons of enabling legisla-tion’ and recorded ‘an 82 per cent congruencebetween his deductive model for Cape Townand the actual case’ (Pirie, 1984: 258).

This multiple methodology allowed him tochronicle ‘the effects of separate developmentphilosophies and practices in a city that beforethe mid-twentieth century was “by far the leastracially segregated in Southern Africa, and per-haps in all of sub-Saharan Africa”’ (de Blij,1982: 493). He cast a critical eye over accountsof the pre-apartheid city, the designs thatapartheid planners produced for it, the ways inwhich these plans were put into practice andthe effects that this had on Mowbray’s removedand resettled coloured population ( Jones,1980; de Blij, 1982; Dewar, 1999). Western’sethnographic research ‘meticulously docu-ment[ed] their histories, experiences and per-ceptions’ (Pirie, 1984: 258) during and after‘removal’. It revealed how the:

racially motivated eviction of coloureds[had] induced collapse of community, aswell as apprehension, social and private dis-tress, material disadvantage … fear … civicunrest … [and the] humiliating damagewhich removals and destruction of homeshad on self-esteem and security, and theway in which people had to learn againtheir place on the edges of a callous urbansociety. (Pirie, 1984: 258)

Western’s ethnography therefore had importantimplications. For South African students, it

contained an important political lesson that ‘“people are living there”; that there arehuman consequences to the use (and abuse)of power and that impersonal, unseen struc-tural forces are not entirely deterministic.There is a primacy in human agency’ (Dewar,1999: 422). And for geographers more widely,it showed how space could be ‘manipulated topreserve and promote a divided society inwhich whites will be master.A distinctive ter-ritoriality cements a racial lifestyle and out-look, entrenching black powerlessness andservile deference’ (Pirie, 1984: 258).

Fifteen years later, reviewers could see howit had shown how ‘social relations are … bothspace-forming and space-contingent’ andhow ‘in remaking the city, apartheid alsoremade Cape Town’s citizens’ (Dewar, 1999:421). Thus, it seemed, Western’s ethnographyhad ‘foreshadow[ed] the insights of structura-tion theory and postmodernism’ (Dewar,1999: 423). Moreover, as Western argued in1999 he had been an ‘accidental humanist’.When he began his research in 1974, thenotion of humanistic geography hadn’t beenproperly crystallized. By the time he had fin-ished it, this had fortunately emerged as ‘anaccepting subdisciplinary niche into which todeliver my now-completed study’ (Western,1999: 426). Not surprisingly, then, he hadn’tdrawn heavily on its literature. However, pre-cisely because the Group Areas Act uprootedpeople from places where they had lived forgenerations and clearly caused so much suffer-ing, the situation (and Western’s book about it)could not help but illustrate a core theme withinit – i.e. ‘the complex and inextricable linkbetween people’s identities and places in theircommunity and society’ (de Blij, 1982: 493).

Prisoners of space?

Both Ley’s and Western’s ethnographies are,without doubt,‘classics in human geography’.Both were hugely influential in the subsequent

DOING ETHNOGRAPHIES

177

3151-Ch-06A.qxd 3/2/04 2:43 PM Page 177

development of urban social geography, withits combinations of political, theoretical andmethodological concerns ( Jackson and Smith,1984). However, we want to argue thatGraham Rowles’ (1978a) Prisoners of Space?Exploring the Geographic Experience of OlderPeople is an equally classic humanistic ethno-graphy. Rowles’ reviewers were as raving asLey’s and Western’s.They praised his painstak-ing, inductive, ‘experiential’ approach, his‘fluid’, ‘intelligent’ and ‘exciting’ accounts andthe way he gave ‘a sense of holism to individ-uals within their physical environment’ (Stutz,1979: 473). They praised his engaging, firstperson, ‘novelistic’ writing style which tookhis arguments ‘beyond explanation and gener-alisation’ and ‘stop[ped] one short, con-fronting one’s perceptions’ (Western, 1981:276, 275).They praised the social relevance ofhis detailed ethnographic work and believedthat it would inspire and allow ‘the eventualemergence of more realistic theory and moreappropriate planning’ that was sensitive topeople’s lived experiences (Ley, 1978: 356).And one reviewer, singled out Rowles’chapter in Ley and Samuels’ (1978) collectionas its ‘most humanistic and engaging’ contri-bution (Prince, 1978: 295).

Despite these plaudits, Rowles’ work doesnot appear to have had the same disciplinaryimpact as Ley’s or Western’s. This is perhapsbecause he has worked on the substance of hisPhD – i.e. the human aging process – in theinterdisciplinary field of ‘gerontology’.6

However, the recent resurgence of interest inhumanistic geography’s core concerns – e.g.geographical knowledges, taken-for-granted,‘performative’ everyday lifeworlds and body-space-identity relationships7 means that itmight now be time to revisit what is geogra-phy’s most direct and detailed humanisticethnography.

Rowles is interesting because he under-took his PhD in a hot-bed of humanisticgeography – Clark University – where he was

supervised by Anne Buttimer and studiedalongside David Seamon and others (seeButtimer and Seamon, 1980; and Buttimer,2001). His work grew out of and challengedthe principles, abstraction and introspectionof this new geography and he believed that its‘introspection and anecdotal or literary evi-dence in part reflects a reluctance to engage in[what he called] experiential field work’(Rowles, 1978a: 174). Like Ley and Western,he used ethnographic research methods togain an insider perspective of a group toooften represented in simple, abstract, stereo-typed terms. However, while they did ethno-graphies on a conventional neighbourhood orcommunity scale, he worked at a muchsmaller scale, developing an experientialapproach to ‘interpersonal knowing’ intended‘to eliminate the intellectual separation ofobserver and observed and thus mediate thedistinction between objective and subjectiveknowledge’ (Entrikin, 1979: 256; see alsoJackson, 1983). In practice this meant that:

Unlike conventional participant observa-tion, where the researcher aims to vary his[sic] social contacts and maintain a certaindetachment from them, interpersonalknowing is predicated upon a deepeningfriendship between participant andresearcher, a relationship which takes timeto establish and in which otherwise hiddenor suppressed aspects of personal under-standing become knowable. Ethically, theobjective is one of mutual discovery, asresearcher and researched are bothengaged in a mutually rewarding project.(Ley, 1978: 355)

Prisoners of Space? was, in a general sense, anexploration of the differences between‘Cartesian and lifeworld space’ (Western,1981b: 276). However, more specifically, it wasan ethnographic critique of behavioural sup-positions about the constriction of elderlypeople’s geographical lives as a result of theageing process. Rowles undertook hisresearch in Winchester Street, a working-classneighbourhood in an unnamed East Coast


178

3151-Ch-06A.qxd 3/2/04 2:43 PM Page 178

American city.This was the context for whatthe book was really about: the biographies,geographical experiences and everyday life-worlds of elderly people living there. He ini-tially intended to recruit large numbers ofelderly research participants but only fivestuck with him – Stan (aged 68), Marie (83),Raymond (69), Evelyn (76) and Edward (80).However, as he later wrote, this ‘anxiety pro-voking [recruitment] ‘failure’ turned out to bea blessing in disguise’ (Rowles, 1978b; 177).

Through developing close, long-term rela-tionships over a three-year period, he was ableto appreciate in vivid detail what it was like tobe, live and see the world as an elderly person.In his book, one chapter was devoted to his‘interpersonal knowledge’ of each person.Two passed away before he finished writingbut he discussed draft chapters with the othertwo before settling on his final versions. Somegeographers found this interpersonal, person-centred approach intriguing (Entrikin, 1979),but, for others, it was much more problematicthan ‘more orthodox participant observation’because, first, the intensity of its commitmentmeant its participants would be largely self-selecting; secondly, its ethical goals could onlybe ideals, given the inevitable imbalancesbetween the “co-explorers” concerns, actionsand ‘public disclosure’; and, finally, some ofthe details being sought were so intimate that‘the outsider should not, perhaps, have beengiven such privileged access’ (Ley, 1978: 356).

Like Western, Rowles examined people’sattachments to places. While these came tothe surface in Mowbray through forceddetachments, in Winchester Street theyemerged from the relationships Rowlesdeveloped with these five people. In the earlystages of his work this didn’t work at all toplan. Marie, for example, wasn’t interested intalking to him about the kinds of issues raisedin the academic literature on ageing, howevercarefully he worded his questions. So, herecounted,

Instead, as we sat in her parlor poring overtreasured scrapbooks in which she kept arecord of her life, she would animatedlydescribe trips she had taken to Florida manyyears previously. She would muse on thecurrent activities of her granddaughter inDetroit, a thousand miles distant. She woulddescribe incidents in the neighbourhoodduring the early years of her residence.Blinded by preconceptions, I could not com-prehend at first the richness of the taken-for-granted lifeworld she was unveiling.(1980: 55)

As a result of such ‘frustrating’ initial experi-ences, his research became less focused andmore time-consuming as he went with theflow of their lives and their ways of thinkingabout the experiences in which he was inter-ested. He ‘hung out’ with his elderly partici-pants, one by one, chatting in their homes,going for walks, meeting friends, going for abeer, attending community meetings, what-ever they liked to do. This allowed relation-ships to develop at their own pace. And herehe found that hours and hours of interaction,lots of casual chatting and long periods ofsilence occasionally unearthed striking flashesof insight. Marie’s account of her localchurch’s Christmas dinner party, for example,shocked him:

‘We had a big banquet. There was over twohundred people, two hundred and fortypeople. We all bought our tickets. We had acaterer.’ ‘What about the poorer olderpeople?’ I interjected. ‘The poor ones?’ shereplied incredulously. ‘There’s no more poor.Don’t talk about poor people. The oldpeople? There’s no more of that. The oldpeople now live better than the young onesbecause they have pensions. They have goodpensions, too. So they all live like million-aires now,’ she concluded emphatically. Themore I pondered this unsettling exchangethe more apparent it became that Marie’simage of poverty, framed by conditionsduring the depression, was completely dif-ferent from my own. Gradually I came tointernalise rather than merely to acknowl-edge that the world she inhabited was a dif-ferent ‘Winchester Street’, not the run-downenvironment I could view. (1978b: 178–9)

DOING ETHNOGRAPHIES

179

3151-Ch-06A.qxd 3/2/04 2:43 PM Page 179

As these relationships developed, Rowlesproduced reams of notes, photographs, sketchmaps and taped conversations with or abouteach participant. He slowly learnt and inter-nalized the richness, complexity and intimatedimensions of the worlds they inhabited. Hisconclusion was that these elderly people wereattached to Winchester Street via their:

• intimate bodily memories of the dimen-sions, routes through and physical featuresof its many interconnected spaces and thealmost effortless routine movements thisallowed;

• autobiographical memories which con-nected who they were to where they werethrough the continual sparking of memo-ries by the ‘incident places’, mementoesand other bits and bobs of their lives; and

• social networks, near and far, everyday andoccasional, experienced and imagined,through which their lives were inter-twined with others, alive and dead.

These attachments showed that the stereotypeof elderly people’s deteriorating physical andperceputal capabilities leading the world toclose in on them was a ‘demeaning oversim-plification’ (Rowles, 1978b: 183). In its place,Rowles’ work ‘emphasise[d] the vibrancy andcreativity of the human mind as well as thetyranny of biological, economic, and socialconstraints’ (Ley, 1978: 356). And theseagency/structure relations had importantpolicy implications. As Rowles later argued,family, medical and social service professionalsshould listen to elderly people’s requests to‘age in place’ because this could prolong theirlives. Removing them to nursing homes, forexample, severed intimate physical, psycho-logical and social attachments to particularplaces that could by no means be easily reestab-lished in a new setting. And, for those who hadto relocate, the design of their new homescould be made more sensitive to residents’

multidimensional geographical knowledges,experiences and bodily capabilities.

Geography’s ‘new’ethnographies

The above discussion has no conclusionbecause these ethnographies foreshadowedmuch of what was to come as Anglo-American geography took its cultural turn inthe early 1990s. ‘Cultural turn’ is a ‘shorthanddescription for a vast number of different andsometimes incommensurable trends withinhuman geography, united only by theirdiverse appeals to concepts of culture andto the intellectual field of cultural studies’(P. Crang, 2000: 142). These included thediverse insights of feminism, postmodernism,poststructuralism, ‘queer theory’, anti-racism,postcolonial theory, psychoanalysis, social psy-chology, cultural studies, science studies andsocial anthropology which, together, broughtinto geography concepts and concerns aboutdifference, identity, embodiment, language,knowledge, texts, discourses, images, commu-nication, value, ways of seeing, agency andstructure, power relations and resistance, toname a few (Philo, 2000; Shurmer-Smith,2002a). Many were not, however, entirely newto the discipline. There were for example,some striking similarities between old human-istic and new feminist geographies. Bothattempted to study and understand better:

• how different people understood theworld and their place(s) within it;

• the routinization of everyday lives in timeand space (including the ‘home’); and

• the ways in which embodiments, memo-ries, emotions and feelings tied togetherplaces and social/personal identities.

Moreover, in studying these relationships,both refused scientific rationality, demanded


180

3151-Ch-06A.qxd 3/2/04 2:43 PM Page 180

researchers’ reflexivity and dispensed withdualisms – e.g. ‘analysis and empathy, insiderand outsider, thought and pleasure, body andmind, individual and context’ – which ‘arealways bound into the power relations of asociety’ (Rose, 1993b; 48). While humanisticgeography left a lot to be desired (e.g. becauseof its inadequate theorization of power rela-tions and its neglect of gender relations; Clokeet al., 1991; McDowell, 1992a; Rose, 1993b),its biggest influence on a cultural turn which‘sent shockwaves throughout the length andbreadth of human geography’ (Philo, 2000:28) was arguably via the widespread adoptionof its qualitative research methods (textual,visual and intersubjective; Dwyer and Limb,2001; Ley and Mountz, 2001).These, enabled

economic geographers [to] ‘discover…’ [the]embeddedness of local economies in localsocial practices; political geographers [tobecome] aware of new nationalisms andnotions of identity in boundary formationand exclusion; urban geographers [to]turn … their attention to lifestyle and … [tobecome] enthusiastic about cultural regen-eration of cities; … [and] retail geographers[to become] enthusiastic about sites of con-sumption, as opposed to patterns of distrib-ution’. (Shurmer-Smith, 2002a: 2)

The now substantial body of work outliningthese methods and how to use them well is,therefore, an overdue but vital element of thiscultural turn literature.

Geography’s cultural turn was also its ethno-graphic turn because social and cultural geogra-phy were brought closer together, in part, underthe influence of the new and feminist ethnogra-phy emerging from anthropology and sociol-ogy.8 Here, theory, politics, poetics and theirrelationships with research design and practicewere discussed in substantial critical detail.Below we focus on three ways in which thesebecame central to geography’s new ethnogra-phies: in debates about, first, categorizing subjectcommunities; secondly, relating abstract theoryto everyday life; and, thirdly, being reflexive.

Categorizing subjectcommunities

Literal definitions of ethnography as ‘peoplewriting’ can be misleading because, originally,these people weren’t the kind you might meetin the library, on holiday or in your research.Early ethnographers set out to describepeople in the sense of a people, a culture, anation, a community – a group of people.Thus‘the assumption of a sociocultural unit, spa-tially and temporally isolated, is deeplyembedded in the conventional framing ofsubjects for ethnographic analysis’ (Marcusand Fischer, 1986: 86). Geographers takingthe cultural turn were therefore critical of theway in which ethnography had been used tobolster an understanding of the world asdivided into discrete peoples or cultures. Butthey were also critical of humanistic geogra-phy’s belief that, beneath the surface in therealm of human essences, all people were thesame (Ekinsmyth and Shurmer-Smith, 2002).So how could relationships between groupand individual identities and the places andspaces of everyday life be understood?Clearly, ethnographic research methods arewell suited to this kind of study. But what doresearchers mean when they say they’re usingthem to understand the ‘inner workings’ of apeople, a culture or a community? How aretheir boundaries defined? By whom? At whatstage(s) in the research process? What differ-ence can this make to research findings? Andwhat can answers to these questions meanboth for the design of, and data constructionwithin, geography’s new ethnographies?

As we have already noted, the ethnographictraditions in geography and anthropology arehistorically entangled. However, as geographybegan to take its cultural turn this entangle-ment took on a new significance. No accountof this turn is complete without discussingJames Duncan’s (1980) critique of the ‘super-organic in American cultural geography’.9

DOING ETHNOGRAPHIES

181

3151-Ch-06A.qxd 3/2/04 2:43 PM Page 181

Duncan’s paper was a critique of the theory ofculture that, through the work of Carl Sauerand his Berkeley School students, dominatedtwentieth-century American cultural geogra-phy. Sauer, as we said earlier, is an importantfigure in the history of geography’s ethno-graphic research. He was an advocate of field-work in ‘primitive rural areas’ (Duncan, 1980:194); he ‘supervised some 40 PhD theses, themajority on Latin American and Caribbeantopics’; and he conveyed ‘to all his students hisfirm belief in the need for first hand fieldexperience and for learning the language ofthe people being studied’ ( Jackson, 1989: 10).So why did Duncan have such a problemwith his ‘cultural’ geography? Why, for some,did Duncan’s 1980 paper have such an impacton the discipline, helping to ‘open … the wayfor a “new” cultural geography’ (Shurmer-Smith, 1998a: 567)? Why, for others, had hispaper made so little difference to this newcultural geography (Mitchell, 1995)? What dowe mean by culture? And what differencedoes this make to the way we, first, understandother related words such as ‘people’,‘commu-nity’,‘neighbourhood’ or ‘place’ and, secondly,undertake ethnographic research with orwithin them?

Duncan’s (1980) main critique of superor-ganicism was that it reified cultures as things.He argued that Sauer and many of his stu-dents saw culture as existing above, and beingindependent of, the everyday actions ofhuman beings. It acted on them. It had thepower to make things happen. It evolved‘according to its own internal logic and pre-sumed set of laws’ (Zelinsky cited in Duncan,1980: 187). Individuals were merely its‘agents’, ‘bearers’ or ‘messenger[s] carryinginformation across the generations and fromplace to place’ (Duncan, 1980: 184). Its ‘valuesgrip men’s [sic] minds and force them to con-form to its will’ (1980: 184). It was a ‘power-ful nearly sovereign primal force [which]

should share star billing in our research andpedagogy, along with geomorphologicalagents, climatic process, biological process,and the operation of economic laws’(Zelinsky cited in Duncan, 1980: 188). Sauer’swas another ‘human geography’ which seem-ingly ‘had nothing to do with individuals’(Jackson, 1989: 18). At best it implied thatpeople were ‘passive and impotent’ and theties that bound them were ‘external to them’(Duncan, 1980: 190). Like Coan’s (1899: 24)descriptions of ‘the brown Polynesian’ whowas ‘everywhere substantially the same’, thisgeography involved the analysis of distinct,bounded, essential, internally created, self-perpetuating and homogeneous cultures.Individuals, at best, were represented as ‘modalpersonality types’ (Matthewson, 1998: 570).

The superorganic theory of culture hadbeen brought into geography in the 1920s and1930s largely through Sauer’s association withthe Berkeley anthropologists Alfred Kroeberand Robert Lowie. As early as the 1940s,though, it had been more or less dismissed inan anthropology which had changed its focusto address ‘questions of how individuals, inter-acting with other individuals through institu-tions, create, maintain, and are in turnmodified by their environment … [and ques-tions of] how individuals exercise choice, howthey are strategists who manipulate the con-texts in which they find themselves’ (Duncan,1980: 196). While this ‘superorganic’ theorycame ‘under devastating attack and [had] longsince been rejected by the vast majority ofanthropologists’ (1980: 182), generations ofAmerican cultural geographers ‘ignored thevariety of alternative definitions’ (p. 182).Instead, they ‘continued to adhere uncritically’to this theory in their research ( Jackson, 1989:18–19) and ‘promoted it to generations of stu-dents through textbooks and undergraduatecourses outlining the kind of ‘world regionalgeography … [which] classified the globe’s


182

3151-Ch-06A.qxd 3/2/04 2:43 PM Page 182

population into culture “worlds”, as if culturesand regions were homogeneous and easilybounded entities’ (Anderson, 1999: 4). Only inthe mid-1970s did this understanding of culturebegin to wane as ‘a number of geographers …distanced themselves from [this] tradition’(Anderson, 1999: 4).

More recent critics of superorganicismpoint out that, despite being academicallydefunct, it continues to maintain its power inpopular understandings of culture, and canoften sneak back into academic research in acasual, taken-for-granted way (Mitchell,1995).We are used to hearing talk of Britishculture, Islamic culture, gay culture’, youthculture, black culture, or urban culture. Manypeople are fascinated by their ‘own’ and/or‘other cultures’, see themselves as moving‘between cultures’ and/or as having a ‘mixedculture’. Culture is often used to explain con-ditions and events. All this arguably reflects atendency towards culturalism: the ‘assumptionthat culture “independently” exists, that cul-tural distinctions are necessarily real androoted in the peoples being analysed, and thatculture can be used as explanation’ (Mitchell,1995: 108). However, while definitions ofculture may be taken for granted, they are notinnocent. Take Peter Jackson’s (1989: 27–8)critique of its use in a newspaper report:

On 12 May 1984, the Los Angeles Timesreported that a high ranking official of thefederal Department of Housing and UrbanDevelopment had sought to explain thepoor living conditions of the Hispanic popu-lation in terms of their cultural ‘preference’for overcrowded housing. Rather than citingpoverty as the explanation for their over-crowded condition, or their uncertain politi-cal status as ‘undocumented workers’, theofficial chose to emphasise their large familysize and ‘preference’ for living with theirextended families. He failed to recognisethat ‘preferences’ are subject to materialconstraint and that it is unreasonable toinfer ‘preferences’ directly from behaviour.Neither can the statement be rejected simply

as a matter of individual ignorance. It is areflection of institutionalised attitudes andstructured inequality. The implication thatHispanic ‘culture’ explains their poor livingconditions, regardless of their materialcircumstances, is … a culturalist explanation.

The main critique of ‘superorganism’ andrelated popular culturalist explanations is thatthey are ‘divisive and dangerous as [they]underpin … unnecessary oppositions andenmities’ (Shurmer-Smith, 2002a: 3; see alsoMassey, 1991b), and ‘can represent and reifydifference by [both] obfuscating connected-ness’ and ‘over-valorising localism’ (Mitchell,1995: 111).These obfuscated connections arenot only those between different, spuriouslyseparate cultures (peoples, communities,nations, etc.) but also between culture, eco-nomics, politics and other, spuriously separaterealms of existence (Latour, 1993: Thrift,1995). Categorizations are central to implicitand explicit theories of how the world isorganized, how it works and how it can orcannot be changed. But as the ‘cultural turn’convinced many geographers, all categoriesand theories are social constructions whichhelp to establish, maintain, justify, and trans-form unequal power relations. Western’s(1981a) study of the categorization and treat-ment of South Africa’s Cape coloureds illus-trated this better than anything.

So how can geography’s new ethnogra-phers study a people, a culture, a communityor any other social group when the politics ofcategorization are so dangerous? Let’s go backto Jackson’s (1989) newspaper article. Imagineyourself as an undergraduate in Los Angeles in1984.You’re fascinated by urban social geog-raphy and are planning a dissertation on this.You come across that article in the Los AngelesTimes. It’s topical. It’s on your doorstep. Itseems to reflect a number of issues you’vebeen reading about in the literature. You’reprobably not part of this Hispanic culture but

DOING ETHNOGRAPHIES

183

3151-Ch-06A.qxd 3/2/04 2:43 PM Page 183

you think you could study it using ethnographicmethods. But your literature review alsoinvolved reading Duncan’s (1980) paper. Soyou’re suspicious of the newspaper’s descrip-tion of this discrete, homogeneous, Hispanicculture and what that official said it wasresponsible for.This is a superorganic story soyou can’t uncritically accept what it says. Butyou still think it’s worthwhile doing researchon housing problems in this part of LA anddecide to base your study in one, predomi-nantly Hispanic, neighbourhood. To get thisgoing, you know you need to find a way intothat neighbourhood, that community, thatculture via one of its gatekeepers, ideally acommunity leader. Once he or she acceptsyou, you’re in.You will be able to move frombeing an insider in your community orculture to an outsider in theirs.

But doesn’t this thinking also reflect thatsuperorganicism you’re so worried about?Isn’t it reproducing, again, that divisive anddangerous, separate and homogeneous, ‘themand us’ mentality? The boundary between oneculture and another is policed by that gate-keeper, isn’t it? But, hold on, if the gatekeeperis a community leader, hasn’t everyone gonesuperorganic? This is confusing, especiallyafter you have spent some ethnographic timein that neighbourhood, being introduced tomore people, talking to them and to others,hanging out, getting involved in things, likeDavid Ley did in Philadelphia. Here itbecomes very clear that the people most keento view this community (or culture) as a sep-arate, bounded, homogeneous entity are thosewho wish to govern, police, represent, create,maintain and/or transform it. However, youremember Pamela Shurmer-Smith’s questionthat, in anyone’s day-to-day life, ‘Which of usshares every aspect of our life with even oneother person, let alone a group’ (2002a: 3)?She’s right.The people you speak to seem tohave all kinds of affiliations to do with their

age, their gender, their financial circumstances,their jobs, their religious beliefs and prac-tices, their politics, the languages they speak,their hobbies, their tastes, their diverse lifeexperiences, whom they choose to (and can’thelp but) spend their time with, and morebesides.

Moreover, these affiliations aren’t only withlocal people or only in the present day.Whatabout those first, second, third-generationMexican, Guatemalan and Nicaraguanmigrants you talked to? It’s like that area ofLondon where Doreen Massey lives because:

while [‘it’] may have a character of its own,it is absolutely not a seamless, coherentidentity, a single sense of place which every-one shares. It could hardly be less so.People’s routes through the place, theirfavourite haunts within it, the connectionsthey make (physically, or by phone or post,or in memory and imagination) betweenhere and the rest of the world vary enor-mously. (1991b: 28)

So how can you set out to study a subjectcommunity whose existence as a neatlybound entity is, at best, highly suspect? Howcan you identify a people who may have fewconcrete boundaries, and whose precisemembership may be difficult to gauge? Thismakes it difficult to research or talk about‘them’.

The answer is that, in an ontological sense,these people-grouping terms (like manyothers, most notably ‘race’) are meaningless,but the ways in which they are defined,deployed and contested in real-world situa-tions have profound effects (Mitchell, 1995;Shurmer-Smith, 2002a). So, however suspect,they can’t simply be ignored. They must betackled, head on. Take culture, for instance.According to Don Mitchell (1995: 108),this cannot and should not be studied as athing which people possess but as a ‘power-fully determined idea’ bound into specific,grounded circumstances, and specific, contested


184

3151-Ch-06A.qxd 3/2/04 2:43 PM Page 184

power relations. Like Duncan, Mitchell drawsupon the work of Clifford Geertz. His ‘inter-pretive anthropology’ has been held up asan ‘attractive alternative to “superorganic”definitions’ ( Jackson and Smith, 1984: 39)because its:

approach is essentially semiotic: regardingculture as a series of signs and symbolswhich convey meaning. Alluding to Weber’sidea that man [sic] is an animal suspended inwebs of significance which he has himselfspun, Geertz argues that culture comprisesthese webs and that their analysis is ‘not anexperimental science in search of law but aninterpretive one in search of meaning’(Geertz 1973, 5) … The diverse and contra-dictory evidence of human behaviourobserved in the field is thus recognised forwhat it is, rather than as aspects of somepostulated cultural entity: ‘what we call ourdata are really our own constructions ofother people’s constructions of what theyand their compatriots are up to’ (Geertz1973, 9). (Jackson and Smith, 1984: 38)

In terms of research design and practice thismeans a number of things. First, researcherscould choose to undertake research to chal-lenge stereotypical, popular, culturalist expla-nations of the living conditions of onecategorized and located group (like Rowles’elderly people). Here ethnographic researchcould be undertaken to understand better thepower of categorization in specific circum-stances, how this may also be part of specificpeople’s self-categorization and the addi-tional and alternative categorizations in use‘on the ground’.10 Here, researchers may startwith culturalist explanations but not neces-sarily end up with them if the partiality, com-plexity, contradiction and complications ofbelonging to any group are both expectedand followed through as part of the researchdesign.

Secondly, instead of framing ethnographicresearch as a process of moving betweencultures it could better be framed as a process

of moving through, and perhaps extending,existing social networks which, themselves,cross boundaries (sometimes with difficulty)between discursively and physically separatedcultures, communities and/or places. Thirdly,it might not only be worth detailing in field-notes the networking that is the researchprocess but also to design research whichallows the tracing of network relations thathelp to create the conditions experienced bythe people under study. Here researcherscould plan to undertake ethnographic studiesin more than one place, with people who maythink they lead separate lives until, perhaps,that research is done and those boundary-crossing connections can be more clearly seen(Marcus and Fischer, 1986). In addition(and/or contrast), they could study the livesof people who routinely make these connec-tions themselves: on foot, in cars, on planes,on the Internet.11 More about this later.

From beginning to end ethnographicresearch is all about networking so this shouldideally be done from the beginning of a project.Ethnographers, have to learn how to workthrough networks, to make appropriate con-nections and to ‘go with the flow’ when pre-conceptions come to light and alternativeinterpretations begin to make more sense. Soit is not a good idea to follow the ‘standardthree-stage … read-then-do-then-write’model of research design (Cook and Crang,1995: 19) to prepare for such an unpredicat-ble process. Nipping culturalist thoughts inthe bud by networking as early as possible is,therefore, a wise damage limitation exercise.

Relating abstract theoryand everyday life

During the 1980s it was not only the retheo-rization of culture that made a difference togeographical inquiry. Theory, in general,became much more important. Central to this

DOING ETHNOGRAPHIES

185

3151-Ch-06A.qxd 3/2/04 2:43 PM Page 185

were attempts by geographers like Nigel Thrift(1983b) to bring together two bodies of workthat developed largely in parallel during the1970s.On the one hand, there was the human-istic geography ‘which [hoped] to capture gen-eral features of place through the specifics ofhuman interaction’ and, on the other, was theMarxist geography ‘which [hoped] to read offthe specifics of places through the general lawsof tendencies of capitalism’ (Thrift, 1983b: 23).Somehow, they argued, the former’s ‘volun-tarist understanding of agency had to berelated to the latter’s ‘deterministic’ under-standings of structure (Agnew,1995).But a fineline had to to trodden because:

the ‘nitty gritty’ of everyday life [could not]be presented as raw, unmediated data – theempiricist fallacy, data speaking for itself –nor [could] it be presented through abstracttheoretical categories – the theoreticist andidealist trap, the lack of interest in empiricalfindings … [What’s] best for the relation –data/theory – is the ‘surprise’ ‘ … that eachcan bring to the other … [through a] con-tinuous process of shifting back and forth …between ‘induction’ and ‘deduction’. (Willisand Trondman, 2000: 12; see aslo Ley andMountz, 2001)

Ethnographic research into everyday lifedesigned to test or illustrate abstract theory orto generate grounded theory by ‘bracketingout’ that abstract theory therefore keepstheory and everyday life separate (Dwyer andLimb, 2001). Structuration theory was devel-oped in the 1980s better to understand theirentanglements, and ethnographic researchplayed an important part in this.

For structuration theory’s main advocate,sociologist Anthony Giddens (1984: 16),structure was ‘often naively conceived of interms of visual imagery, akin to the skeletonor morphology of an organism or the girderof a building’ and was therefore treated as‘“external” to human action, as a source ofconstraint to the free initiative of theindependently constituted subject’. He

argued, instead, that there was a duality ofstructure which placed human agency in aweb of cause and effect which was bothburied in the routinized, often taken forgranted, ‘ “how to do” and therefore “how tobe” messages’ in/of everyday life (Thrift,1983b: 43) and stretched out of awareness andcontrol through the ‘systemness’ of society.Here, structure and agency were effectively thesame thing; the structuration process wasunfolding and unfinished; people’s everydayactivities – consciously, intentionally or other-wise – were central to this process; and, there-fore, their consequences were countless, oftenunintended and inevitably folded back intothe contexts for future activities.

For ethnographers, structuration theory wasinteresting because it emphasized:

• how, where, when and with whom peoplelearnt those ‘how to do’ and ‘how to be’messages;

• how these messages were continually rein-forced components and results of rou-tinized, compartmentalized and channelled‘activity experiences’;

• how much of the knowledge in thesemessages comprised a taken-for-granted‘practical consciousness’, which was onlyrationalized into ‘discursive consciousness’when routines were disrupted and/or actiityexperiences jarred when moving betweenlocales (Giddens, 1984;Thrift, 1985);

• how much of this knowledge was geo-graphical and bodily, as Graham Rowleshad discovered; and

• how participant observation’s uniquelylong-term, situated and embodied approach(watching, talking and doing) could there-fore play a vital role in translating practicalconsciousness into discursive consciousness.

Ethnographic studies reconnecting structureand agency could therefore address important


186

3151-Ch-06A.qxd 3/2/04 2:43 PM Page 186

questions about the ‘potency of humanagency in an unfinished world’ and drawattention ‘to the prospects for effective actionrather than determinacy in social life’ (Smith,S.J., 1984: 353). And, through unearthing,putting into words and questioning the takenfor granted, they could potentially ‘restorehuman beings to their worlds in such a waythat they can take part in the collective trans-formation of their own human geographies’(Gregory, 1981b: 5).

Despite this potential structuration theoryhas more or less disappeared from geographi-cal view. One explanation is that it couldn’tturn its theoretical principles into new formsof research practice because ‘Giddens … con-sistently advocated a methodological bracketingthat allows for either the analysis of strategicconduct or the analysis of institutions …[thereby] transpos[ing] the dualism between‘“agency”and “structure” from a theoretical toa methodological level’ (Gregory, 2000a: 800).Nowhere was this better illustrated than inPaul Willis’ (1977) Learning to Labour: HowWorking Class Kids Get Working Class Jobs.This was a landmark ethnography due to itsdetailed combination of political economictheory and ethnographic research.Nigel Thrift(1983b) used it to illustrate structuration’sarguments about the intended and unintendedconsequences of social action. And, subse-quently, it was a springboard for GeorgeMarcus’ sustained methodological rethinkingof structure/agency relations (1986; 1998;Marcus and Fischer, 1986; 1999). For Thrift(1983b: 43), Learning to labour

shows how some working-class schoolchild-ren build up a resistance to mental work outof their resistance to the authority of theschool. Manual work is invested with seem-ingly opposite qualities to mental work –aggressiveness, solidarity, sharpness of wit,masculinity – and becomes a positive affir-mation of freedom. Here resistance has theeffect that working-class schoolchildren

accept working-class jobs through ‘free’choice. They willingly embrace their ownrepression.

Thrift did not criticize Willis’ book, despite itsresemblance to the kind of functionalist workhe was arguing against. Learning to labour wasan attempt to trace ‘the passage of a group ofyoung men from school to work … [to]show… how their unequivocal rejection ofmiddle-class values represents a symbolicopposition to their structural subordination’( Jackson, 1989: 6, emphasis added). Willisundertook ethnographic research with agroup of ‘lads’ and contextualised their livesand choices within abstract arguments about‘class structured society built into the tradi-tion of Marxist theory’ (Marcus, 1986: 186).Half the book was detailed ethnographicdescription and half was dense theoreticalargumentation. Agency was studied empiri-cally and structure was theorized. That wasthe problem.

Marcus’ reading of Willis’ approach wasmore critical than Thrift’s. For him it simplyadded a ‘recognition of the messiness of reallife … to conventional notions of functionalistorder’ (1986: 183). It supported that ‘unin-tended consequences’ argument only in thesense that ‘agents do what they do in theirvarious contexts, and hey presto!, there isneat order in the system’ (1986: 183). Willismade ‘the lads real, but … [reified] the largersocial system in which they [lived] … [byignoring the fact that] what is “the system” forthe lads is the other’s (the middle class) culturalform’ (p. 186).Thus, Marcus concluded:

[If] the system thoroughly operates byunwitting interdependencies and unin-tended, but coordinated consequences, … itwould be difficult to sustain the conven-tional way of representing the dominantclass or classes in Marxist theory ... If [Willis]had instead constructed his text on cross-class ethnographic juxtapositions, he wouldhave had to face the full implications of his

DOING ETHNOGRAPHIES

187

3151-Ch-06A.qxd 3/2/04 2:43 PM Page 187

unintended consequences more squarely.(pp. 186–7)

This is a powerful and important structure/agency, theory/data argument.But what wouldan ethnography that took it into accountactually look like? Marcus and Fisher (1986: 91)answered:

What we have in mind [is the kind of work]that takes as its subject not a concentratedgroup of people in a community, affected inone way or another by political-economicforces, but ‘the system’ itself – the politicaland economic processes spanning differentlocales, or even different continents.Ethnographically, these processes are regis-tered in the activities of dispersed groups orindividuals whose actions have mutual,often unintended, consequences for eachother, as they are connected by markets andother major institutions that make theworld a system.

Marcus (1986: 171) argued that ‘the most obvi-ous views of systems as objects for experimen-tation with multi-locale ethnographies’ were‘markets (Adam Smith’s invisible hand) andcapitalist modes of production, distribution, andconsumption (Marx’s version of the invisiblehand – commodity fetishism)’. However, afterundertaking a review of multi-locale ethnogra-phies some years later, he (1995; 1998) foundthat Marxism was not the only body of theorythat, in practice, could enjoy such a following.Poststructural theory worked just as well.

One of poststructuralism’s core beliefs isthat dominant categorizations and binaryoppositions must be unsettled and resisted inorder to create a ‘non-oppressive politics’(Pratt, 2002: 626). As Donna Haraway (1991:177) has argued:

self/other, mind/body, culture/nature [male/female,] civilised/primitive, reality/ appear-ance, whole/part, agent/resource, maker/made, active/passive, right/wrong, truth/illu-sion, total/partial, God/man [and, we wouldadd, global/local] … have all been systematicto the logics of domination of women,

people of colour, nature, workers, animals – inshort domination of all constituted asothers, whose task is to mirror the self.

Working against this, therefore, requires ‘newconnective modes of theorising’ (Pratt, 2000b:626) and this is where geographers’ contem-porary fascination with non-representationaltheory comes in (Thrift 1995, 2000b).This iswhere:

• the ‘good bits’ of structuration theory havearguably resurfaced in new, improved forms;

• calls have been made for new types ofresearch ‘where “theory” is not “applied”in a particular “context” but where thetheoretical and empirical fold into oneanother in a glutinous mix of text andcontext, and facticity and fictionality’(Thrift, 1995: 530); and

• ‘observant participation’ is now beingtouted as a key means to accomplish this(Thrift, 2000b; Smith, S.J., 2001: 35–6;Hoggart et al., 2002: 258).

We can only give a flavour of the argumentshere, in part because they aren’t yet ade-quately developed in the literature. None ofhuman geography’s qualitative research textsdeals with this in sufficient, substantial detail,and its finished products have tended to be‘unspecific or unclear’ about research practices(Davies, 1998: 76).

Like much of the literature we have dis-cussed so far, non-representational theory isprimarily concerned with the ‘mundaneeveryday practices, that shape the conduct ofhuman beings towards others and themselvesat particular sites’ (Thrift, 1997, cited in Nash,2000: 65). But its advocates have criticizedqualitative research in human geography, first,for registering only ‘narrow realms of sensatelife’ (Thrift, 2000a: 3) and, secondly, for rely-ing on a model of consciousness where ‘allacts are secondary to the processing of (or


188

3151-Ch-06A.qxd 3/2/04 2:43 PM Page 188

even deliberation over) knowledge asrepresentation’ (Hinchliffe, 2000: 575). Newresearch practices are therefore necessary toreach beyond what can be rationalized intowords (i.e. from, or translated into, ‘discursiveconsciousness’) into ‘realms of sensate life’that, for most people, cannot: for example, theworkings of the ‘unconscious’, the body’s pre-cognitive practices (e.g. kinaesthesia)12 and the‘rounded sensory experience’ of everyday life(Smith, S.J., 2001: 31; see also Gregory, 2000a;2000b; Nash, 2000).

However, as shown in geographical studiesof embodiment, experience and perfor-mance (e.g. Rowles, 1978b; Crang, 1994;Malbon, 1999), it is surprising what non-specialists can and do turn into words whenethnographies are conducted appropriately,in sufficient detail, with members of care-fully chosen ‘subject communities’, and inthe right times and places. Thus non-representational theory could usefully sensi-tize prospective ethnographers to push theboundaries of research design (by employ-ing, for example, new media technologieswhich can help to shift attention beyondwords) and to consider the experiential sideof its methods in a more holistic manner.Going beyond this could, however, be prob-lematic because:

How can … precognitive body practices beknown, or is the effort to understand andcommunicate abandoned in favour ofabstract theorising of the non-repre-sentable? Are ethnographic research meth-ods as redundant as textual or visualsources, since they invite people to speakand therefore cannot access the preverbal?What happens to the project of ‘giving avoice’ to the marginalised, if the concern iswith what cannot be expressed rather thanwhat can? … How does the focus onnoncognitive everyday practices positionthe academic? The energy spent in findingways to express the inexpressible, … seemsto imply a new (or maybe old) division oflabour separating academics who think

(especially about not thinking or thenoncognitive) and those ‘ordinary people’out there who just act. (Nash, 2000: 662)

To avoid turning into another ethnographicdead-end – calling for much but deliveringlittle – geographers need to undertake andpublish substantial studies which show hownon-representational theory and its observantparticipation can surprise one another.

To an extent, this has already begun in onearea of non-representational theory – actornetwork theory or ANT – whose mostnotable full-blown empirical study – BrunoLatour’s (1996) Aramis, or the Love ofTechnology – is packed with ‘surprises’ (seeLaurier and Philo, 1999). ANT is, arguably,what structuration theory could have becomeif Giddens had more than a ‘limitedencounter with post-structuralism’, if he’dbeen able to think of ‘action as [more than]individual and [had more] fully consideredthe ghost of networked others that continu-ally informs that action’ (Thrift cited inGregory, 2000a: 800). In ANT, agency is farmore than the activity of individual humanbeings. Rather, it is an attribution of ‘the abil-ity to act’ to ‘functional collectives’ which can‘include non-human actors and new cou-plings of human beings and machines’(Gregory, 2000b: 350). It is ‘a relational effectgenerated by a network of heterogeneous,interacting components whose activity isconstituted in the networks of which theyform a part’ (Whatmore cited in Gregory,2000b: 350). And its structures cannot, there-fore, be seen as ‘lying somewhere outside orabove the fray’ but are, like everything else,‘intricate weavings of situated people, artefacts,codes, and living things and the maintenanceof particular tapestries of connection acrossthe world’ (Whatmore and Thorne, 1997:288, emphasis in original).

These are perhaps the barest of ANT’sbare bones. Fleshed out, it can be a

DOING ETHNOGRAPHIES

189

3151-Ch-06A.qxd 3/2/04 2:43 PM Page 189

mind-bogglingly abstract labyrinth of ideas.However, in Aramis and in other ANT-inspired empirical studies, it is often a moreobvious and thought-provoking ‘way of see-ing’. Here the dividing lines between agency/structure, theory/data, global/local and plentyof other binary oppositions disappear fromview in the mundane, situated complexities ofeveryday life.13

In this kind of study, the research processtakes shape after researchers (1) latch on tospecific people, things, metaphors, plots/stories/allegories, lives/biographies and/orconflicts; and (2) follow them, see what ‘sticks’to them, gets wrapped up in them, unravelsthem and where that takes a researcher who is(un)able to follow myriad possible leads andto make myriad possible connections (Marcus,1995; 1998). Here, questions are asked of‘emergent object[s] of study whose contours,sites, and relationships are not known before-hand’ but come into being and take ‘unex-pected trajectories’ through following up theanswers to those questions, and the questionsthat they then raise, and so on, where possible(Marcus, 1998: 86, 80).As Jonathan Murdochhas put it, ‘We let them [the people, thethings, etc.] show us where to look, whatmaterial they use in the course of networkconstruction and how they come to berelated to others’ (cited in Davies, 1998: 77).Theory is therefore ‘“held lightly” and ismade accountable to fieldwork’ (Dwyer andLimb, 2001: 11) rather than ‘imprison[ing]field data within preconceived positions’ (Leyand Mountz, 2001: 236).The big explanationsin terms of politics, economics, organisation,and technology always turn up, without fail:“It’s politically unacceptable”. “It isn’t prof-itable”.“Society isn’t ready for it”.“It’s ineffi-cient”’ (Latour, 1996, cited in Laurier andPhilo, 1999: 1054).

What is interesting in these studies is howthe connections being made or followed are

‘ordered’ by their human and non-humanparticipants (including researchers), and howthe research process is thereby at once com-plexly empirical and ‘theoretical’ (in both aca-demic and everyday senses of ‘sense-making’).Ian Cook et al.’s (2002) papaya paper, forexample, brings together theoretical debatesabout:

culture and economy, consumption andproduction, postcolonialism and globalisa-tion, commodity chains, political economy,actor network theory, multi-local ethnogra-phy, plant physiology, retailing, gender,‘race’ and class, labour relations. Morebesides. [These are] Academic debateswhich could never be synthesised into abig, complex theoretical position. By me atleast. But they do work together in anempirical account tracing the biographiesand geographies of what might initiallyappear to be a discrete thing. In all of itscomplexities and contradictions.

Similarly, Mike Michael’s (2000: 13) study ofdog leads ‘simultaneously explor[es] the inter-relations between dogs and humans, culturalrepresentations of animals, community, localand central government, the environment andthe body’.What this approach does is ‘remove[the idea of context] as a subtle safety net formany theorists who will often invoke it as aterm akin to social order or social structure’(Laurier and Philo, 2003: 1055). It can, there-fore, be unpopular because it decentres estab-lished radical standpoints like the ‘resistanceand accommodation framework that hasorganised a considerable body of valuableresearch’ (Marcus, 1998: 85). What it offersinstead, however, is an approach which, first, is‘modest’, ‘pragmatic’, ‘recursive’ and ‘commit-ted to an ordering inquiry into ordering, ratherthan to an ordered inquiry which uncovers otherroot orders’ (Law, 1994: 97), and, secondly,enables the ‘mega concepts with which con-temporary social science is afflicted … [to] begiven the sort of actuality that makes it possible


190

3151-Ch-06A.qxd 3/2/04 2:43 PM Page 190

to think not only realistically and concretelyabout them, but, … creatively and imagina-tively with them’ (Geertz, 1973: 23, emphasisin original).

Few of these studies exist and few ofthem – including Aramis – are, strictlyspeaking, ethnographic in the sense thatthey have been based on a substantial periodof participant observation research. Aramis,for example, is participant observation instyle but is, in fact,‘a “novelistic” report’ basedon ‘a rich archive of documentary sources(planning and policy papers, publicity mate-rials, consultancy reports, memos, scribblednotes), iconographic sources (photographs,maps, diagrams) and also in-depth inter-views with many of the relevant humanactors’ (Laurier and Philo, 1999: 1050). Suchstudies are therefore of little use if we wantanswers to basic methodological questions‘such as how to construct the multi-sitedspace through which the ethnographer tra-verses’ (Marcus, 1998: 89). Saying that ‘they’will ‘show us where to look’ doesn’t helpmuch because:

How do we know where they are telling usto look without knowing their vocabulary?How do we understand the significanceof their materials without using existingcategories? How do we understand theirrelationships without drawing on priorexperiences? And, perhaps most reasonably,how do you get an immensely busy commu-nity of people to do your work for you?(Davies, 1998: 77)

These questions are rarely addressed ingeography’s ANT (and after) literature(although, see Davies, forthcoming). But, for-tunately, they have been addressed in geogra-phy’s growing, process-oriented qualitativemethods literature. Here, numerous ‘warts andall’ stories show that research practice alwaysinvolves working through, extending andthen (often) hiding network relations(Marcus, 1998; see also Keith, 1992; Cook and

Crang, 1995; Davies, forthcoming). A keyfeature of anthropology’s cultural turn washow it took behind the scenes, ‘corridor talk’about what shaped fieldwork experiences andobjects of study, put them on the pages of aca-demic books and journals and turned theminto ‘up front’ debates about the theory andpractice of ethnographic inquiry. This wascalled ‘being reflexive’, and geographersjoined in.

Being reflexive

Geography’s cultural turn led many to ‘ques-tion the constitution of the discipline – whatwe know, how we know it and what differ-ence this makes both to the type of researchwe do and who participates in it with us, aseither colleagues or research subjects’(McDowell, 1992a: 399–400).The ‘social con-struction of knowledges and discourses’within the discipline ‘and the relations ofpower embedded within them’ were hotlydebated (McDowell, 1992a: 399–400). Ingeography’s ‘house journals’, awkward andprofound questions were asked:

What is the relationship between our ownbackground, current position and values,and our own research agenda? How do weknow what we know? Through what sort oflenses is our knowledge filtered? Who isincluded and who is excluded by the socialpractices and academic subject matter ofacademic geography? For whom are wewriting? And what is, what should be, therelationship between our theories and ourpolitics, between our thinking and action?(McDowell, 1992b: 56)

And this has continued. Why are mostresearchers ‘commonly white, male and mid-dle class, with disproportionately low ratiosbetween researcher and researched for non-whites, the poor, women and so on’ (Hoggartet al., 2002: 261)? Why does research so ‘ofteninvolve those with more economic and

DOING ETHNOGRAPHIES

191

3151-Ch-06A.qxd 3/2/04 2:43 PM Page 191

cultural capital studying those with less’(Hoggartet al., 2002: 261)? Why do researchers mainlyreport their findings to people like themselves(Ley and Mountz, 2001)? How can you justifyusing methods designed to develop sufficienttrust for participants to yield sensitive infor-mation and then risk betraying this by pub-licly writing what may ‘upset (or perhaps evendisadvantage)’ them (Ekinsmyth, 2002: 183;see also Ley and Mountz, 2001)? Given itsimperial roots, how can ethnographic research(in the ‘Third World’ or elsewhere) avoidbeing the ‘handmaiden to broader colonialistprojects that invent oppressed groups as ameans of controlling them’(Herbert,2000:562)?So, ‘who has the right to write about whom’(Hoggart et al., 2002: 263)? Are there circum-stances where researching ‘the Other … [is]ethically defensible’ (Hoggart et al., 2002:263)? And, how can ‘the Other’ be repre-sented anyway, ‘when researchers are so thor-oughly saturated with the ideological baggageof their own culture’ (Ley and Mountz,2001: 235)?

These questions are awkward andbewildering, and geographical research couldeasily have ground to a halt under theirweight. However, if anything, the opposite hashappened. Qualitative studies of ‘non-dominant, … socially excluded or margin-alised groups’ (Smith, S.J., 2001: 25–6) haveflourished (Dwyer and Limb, 2001).This mayreflect, first, the increasing involvementand/or visibility of members of non-dominant groups in academic human geogra-phy and, secondly, critiques of oversimplified,essentialist distinctions between ‘the Self ’ and‘the Other’ in debates about culture and thepolitics of difference (Ekinsmyth, 2002). Here,identity has not been seen in such starkeither/or categories but as ‘partial in all itsguises, never finished, whole, simple there andoriginal; it is always constructed and stitchedtogether imperfectly, and therefore able to join

with another, to see together withoutclaiming to be another’ (Haraway, 1996: 119,emphasis in original).

These new studies seem to have resultedfrom the political choices by a, perhaps, morediverse, body of geographers committed tofeminist, anti-racist, postcolonial and like-minded projects. They have seen their rightscomplicated by their responsibilities to docertain kinds of research. Many geographershave wanted to recover and centralize …‘“marginalised” voices and stories’ (Parr,2001: 181) with the aim of ‘unsettling thestatus quo, or redefining what is relevant, use-ful and legitimate [geographical] knowledge’(Smith, S.J., 2001: 25–6). Chris Philo’s (2000:29) verdict, at least, is that this has had thedesired effect, creating ‘a new ambience forhuman geography that has blown away manycobwebs of convention, conservatism andeven downright prejudice’. And qualitative,‘intersubjective’ methods have played theirpart in this because they make it ‘easier toequalise power relations’ by ‘allow[ing] theresearched more opportunity to “frame”(i.e. identify salient issues etc.) the research’(Ekinsmyth, 2002: 181) ‘without beinginvasive, colonising and violent’ ( Johnston,1997b: 291).

These principles have been a core contri-bution of feminist ethnography (Madge et al.,1997; Nagar, 1997a; Ekinsmyth, 2002) andwere, arguably, central to our three humanis-tic ethnographies. Rowles’ work, for instance,not only provided vivid insights into the geo-graphical lives and imaginations of elderlypeople but also challenged stereotypes aboutthe ageing process both in American societyin general and in the academic disciplines thatclaimed expertise in this field.This happenedbecause he organized his research around thebelief that elderly people knew an awful lotmore about the ageing process than he did,and that he therefore had much to learn from


192

3151-Ch-06A.qxd 3/2/04 2:43 PM Page 192

them. He used ethnographic methods to doresearch with elderly people, and his datadetailed not only what he learnt but also howand where it was (re)framed: in conversation, insitu, over time and with difficulty.

This ethnographic tradition has continuedin geography. However, it has been supple-mented by new work that pushes its politicsto their logical extreme and/or studies peoplewho wouldn’t normally be considered to bemarginal. First, more explicitly interventionistprojects have drawn upon participatoryaction, activist and/or emancipatory researchmethods. Here, researchers do ‘not simply[aim] to empathise with research subjects’ but,through contributing professional skills andcontacts to the struggles of marginalized peo-ples, aim ‘to help to produce changes in theirlives’ (Dwyer and Limb, 2001: 7).14 Secondly,it is important to recognize that shining somuch light on the relatively poor and power-less leaves in the shade huge parts of theprocesses that human geographers aim tounderstand. Thus there is a growing body ofethnographic research with more wealthy andpowerful groups.15 While such ethnographiesare often framed as geographical inflections ofparticular ethnographic traditions, they arealso the result of much more practical choicesabout when, where and with whom researchcan be done.

To help think through the problems andpracticalities brought up in these debates,many geographers have followed other socialscientists by becoming much more reflexive.Reflexivity can be seen to take three formswhich attempt to ‘link the micro-level activi-ties of the lone ethnographer with macro-level processes and imperatives’ (Herbert,2000: 562). First, the power relations structur-ing the interpersonal relationships that ethno-graphies are based upon have been scrutinized.Questions have been asked about whom andwhat researchers have power over, whom and

what has power over them, how their powerand status can change as they move throughtheir ‘expanded field’, and what relative effectsresearch has on the lives of researchers andresearched (See Robson, 1994; Cook, 2001).16

Secondly, geographical institutions, traditions,ways of seeing and the ways they can affectthe ‘conditions and modes of producingknowledge about other cultures’ (Callaway,1992: 32) have also been put under the micro-scope. It has been noted, for instance, thatgeographical traditions and institutionalarrangements attract only a narrow range ofstudents, who work through established socialand/or disciplinary networks which guidetheir research in certain directions (seeMcDowell, 1992b; Sidaway, 1997;Twyman etal., 1999). Finally, wider questions have beenasked about how these issues are related to ‘aworld historical situation in which, as oneface of imperial domination, differentiallypermeable barriers to the movement ofpeople, and to the flow of information, havebeen erected and have asymmetrically orga-nized contact between the world’s people’(Dwyer, 1977: 148).

Calls to be more reflexive in these waysmight, again, seem quite daunting. Someargue that too much reflexivity might pro-duce research that is more self- than other-fascinated and thereby discourages the studyof issues that really matter, ‘out there’ in thewider world (Pile and Thrift, 1995).That said,this self-reflection has, however, resulted innew ways of thinking about the planning andconduct of ethnographic research.

First, writing about the prehistories ofresearch projects has shown how the personalis the political, that concerns, motivations andpassions rooted in people’s life experiencesplay an important role in energizing anddirecting their research (see Bristow, 1991;Cloke, 1995; Saltmarsh, 2001). Secondly,reflexive writing has shown that, however ‘out

DOING ETHNOGRAPHIES

193

3151-Ch-06A.qxd 3/2/04 2:43 PM Page 193

there’ and ‘other oriented’ a project appears tobe, it will typically have taken shape throughexploiting and extending networks of whichresearchers and/or their supervisors arealready, or could easily become, part. Researchis often done in places where:

• trusted contacts have already been estab-lished in local universities, NGOs, and soon (see Robson and Wills, 1994a);

• spaces and/or social groups can be enteredthrough paying a fee (see Cook, 1997;Desforges, 2001);

• a job, work placement, or voluntary posi-tion is only an application away (see Crang,1994)

• no-fee events, campaigns or other activi-ties, provide relatively free access (seeJackson, 1988; Parr, 2000; Anderson, 2002);

• researchers work out that they knowpeople who know people who can helpthem to develop contacts in seeminglyimpregnable or distant organisations andplaces (see Keith, 1992; Davies, 2003;Cook and Crang, 1995);

• researchers choose to work within andthrough their own family, friendship andleisure networks (see Lakhani in Hoggartet al., 2002; Saltmarsh, 2001); or.

All this reflection has led to a rethinking ofwhat counts as ‘the field’: where and when ‘it’begins and ends, and where and whenresearchers should consider themselves to be‘on duty’ (or not). First encounters in tradi-tionally ascribed fields – neatly bounded and outthere – are by no means the only ones thatmake a difference to research findings. Ethno-graphic and other research always takes placein an expanded field,where researchers inhabitand move between a number of differentlocales, people and frames of meaning, andtheir work inevitably involves complex trans-lations of meaning between all these settings.17

Here, the variety of research participants is

much wider than is usually acknowledged.They include not only those people in aresearcher’s proper subject community butalso those who inspire, fund, direct and act asthe official audience for his or her work, andthe friends and family members whose sup-port and/or understanding (or lack of it) canplay an important role in getting things done(see Cook, 2001). In addition to the residentsof Mowbray, for example, Robert Coles andVictor Wessels were key participants inWestern’s (1981) expanded fieldwork.

Secondly, moving between locales and rela-tions where norms of speech, appearance,status, ethics, manners and so on may be verydifferent often involves a complex, confusing,anxiety-provoking, and sometimes bizarreprocess of ‘identity management.’18 Some ofthis is specific to participant observation asdecisions have to be made about how covert orovert this should be (see Hoggart et al., 2002:256–7). But, research more generally,‘can be a tricky, fascinating, awkward, tedious,annoying, hilarious, confusing, disturbing,mechanical, sociable, isolating, surprising,sweaty, messy, systematic, costly, draining, iter-ative, contradictory, open-ended process’(Cook et al., in press). It is invariably an intel-lectually and emotionally challenging experi-ence, and prospective researchers need to beprepared for this. Ideally such research reachesa point where the researcher has the confi-dence to ‘follow impulses, change directions,and co-ordinate with other people …[because he or she understands that local]unpredictability has its distinctive tempo, andit permits people to develop timing, coordi-nation, and a knack for responding to contin-gencies’ (Rosaldo, 1993: 112). But this maynot always happen, especially if the researchisn’t given sufficient time.

So, to conclude, reflexive writing about theprocess of doing ethnographies can helpresearchers to explain – to themselves and toothers – what they did and why under the


194

3151-Ch-06A.qxd 3/2/04 2:43 PM Page 194

circumstances. This is where questions ofpositionality, politics and, indeed, ethics getworked out, often in quite messy ways(McDowell, 1992a). Expanded field work,therefore, requires expanded field notes, andprospective ethnographers can learn much fromthe growing body of geographers’ reflexive,‘warts and all’ accounts of their experiences.

Top tips for prospectiveresearchers

Before we conclude this chapter by discussingthe construction of ethnography’s prime ‘data’ –its fieldnotes – we want to take stock of thearguments presented so far and to condensethem into some top tips for anyone thinking ofdoing participant observation research. If youhave got this far and still want to do this,we rec-ommend that you do the following.

Read some ‘wartsand all’ accounts

Read such accounts no matter how irrelevantthey may seem to what you plan to do andeven if you don’t plan to use participantobservation as a formal research method.19

Try to read them before undertaking yourresearch. They provide tales from the(expanded) field that can be full of ideas,inspiration, confidence, mistakes, warnings,realizations, etc., from which you could learnin planning and conducting your research. Ifor when your research doesn’t quite go toplan, it isn’t necessarily because it’s going badlybut because that’s what happens in participantobservation research. Like those authors, youwould have to cope under those circum-stances, think on your feet, change the wayyou do things, try to maintain some focus andlive with the fact that you can’t be sureexactly how things will turn out.

Read them alongsidesome ‘methodologicalcookbooks’

Read those texts that don’t provide such richtales from the field but outline the practical-ities involved in using a variety of researchmethods, including participant observation.20

Despite the inevitable twists and turns ofresearch practice, it’s important to set out todo research properly, systematically, rigor-ously, to have a plan and to know how to dodifferent types of research well. But, if thingsdon’t go to plan and/or if opportunitiesunexpectedly arise in the field to undertake,for instance, participant observation orarchive work, you need to know how to dothis as well.Take one or more of these books(or their equivalent) into the field, just incase.

Familiarize yourselfwith the history ofethnographic researchin geography

Ethnographic research has been undertakenacross a number of disciplines, all over theworld and for over a century. So it is reason-able to assume that someone, somewhere, hasdone research which somehow relates to whatyou’re thinking of doing. As we have shown,calls for geography’s latest, cutting-edge,‘poststructural’ ethnographies could usefullydraw upon some neglected ‘humanisticethnographies’ undertaken in the 1970s. Evenif their ‘conceptual bookends’ may seem inap-propriate or out of date, the ‘thick descrip-tion[s] of everyday life’ that they contain maystill be informative and thought provoking(Ley, 1998: 79). It is important to recognizethat there is a long-running and distinctiveethnographic tradition within the discipline,so geographers don’t only have to look else-where for inspiration.

DOING ETHNOGRAPHIES

195

3151-Ch-06A.qxd 3/2/04 2:43 PM Page 195

Read some results ofparticipant observationresearch

Have in mind the kind of final product youmay want, or be expected, to produce fromyour research. ‘Warts and all’ and ‘cookbook’reading can help you to picture what yourmethodological writing might look like. Butwhat about the findings of your research? Howcould they be presented? How have others pre-sented their findings? What theoretical ideas,methodologies, field settings and findings havethey put together? Have they made themappear to be somewhat discrete? Or have theypresented them as jumbled up, more or less asthey fitted together in the expanded field? Havethey limited themselves to the usual combina-tions of words, pictures, maps, diagrams, etc., onpaper? Or have they expressed things differentlyon and/or off the page? What academic prece-dence is there for doing this? And, having readthis work and decided what works best for you,what kinds of data do you therefore want toproduce, in what sorts of detail and in whatstate of separation for your write-up?21

Mix up your reading,doing and writing

Open a research diary as soon as you startthinking about your dissertation or your nextresearch project especially, but not only, if you’reconsidering using participant observation. It’sessential to undertake research that is interest-ing (to you and to others), relevant (in an aca-demic, popular, policy, etc., sense) and ‘doable’(in a practical sense). Interesting, relevantresearch that can’t be done, or doable researchthat is irrelevant and/or uninteresting shouldbe avoided if possible! Interesting, relevant,doable research has to be pieced together overtime – through trial and error, false starts, dead-ends and breakthroughs, flights of imagination

and reality bites and, ideally, tentative foraysinto the field to make contacts and to trythings out. Good combinations of theory,methodology and practice should grow andwork together.

So why not record in a research diary anyinteresting and relevant ideas from academicreading, newspapers and other sources andhow these have gone down and made sense asyou’ve discussed them with your peers, super-visors or other people? A diary can also besomewhere to have a go at writing fieldnotesbefore your really need to.These notes couldconcern the expanding field of your project.They could outline your early contact-making and suggest how your ideas coulddevelop. They could also be experiments infield-noting where you go somewhere and/ordo something new and try to finds ways ofvividly conveying that experience to an acad-emic audience (fellow students, qualitativemethods lecturers or your supervisor; seeWarren et al., 2000). So what guidance can wegive on this?

Conclusion: field-noting

Until recently, geography’s prospective ethno-graphers had only a limited methodology lit-erature to draw upon when planning theirresearch. In development geography, forexample, students were often expected toprepare for fieldwork through ‘a good grasp oftheory and a well written literature review’but, rarely, through a good grasp of the actual‘rigours of fieldwork’ (Robson and Willis,1994b: 1).This is what they were expected tolearn through a ‘“baptism of fire” in the field –probably mirroring their supervisors’ ownexperiences’ (Robson and Willis, 1994b: 1). Indisciplines where ethnography is a coremethodology – most notably in anthropology –the ways in which ‘fieldwork dilemmas arise


196

3151-Ch-06A.qxd 3/2/04 2:43 PM Page 196

and are resolved’ have been an important partof research training (McDowell, 1992a: 408).In geography, it has not. Here, far more atten-tion has been paid to the products than to ‘theprocesses of actually undertaking the research’(Dwyer and Limb, 2001: 2, emphasis in orig-inal). Yet field noting is necessary to recordand to make sense of key events as they hap-pen, to think about how the research is takingshape, to try to maintain and/or regain itsfocus and to plan its next stages.These notesmay be systematically written in a properresearch diary but they may also take otherforms: as researchers scribble notes on straypieces of paper or around the edges of ques-tionnaires; as they write emails and lettershome to friends, family and supervisors; or asthey continue to fill the pages of a personaldiary or journal. Field-noting is not the solepreserve of ethnographers, then, because allresearch involves a combination of participa-tion and observation in projects that changeas they proceed, and such changes are invari-ably noted in some way or other. Proper field-notes are, however, the prime data ofparticipant observation research. These arewhat should enable researchers to give thosewho read their work that unique sense of‘being there’.

Katy Bennett has recently recalled theproblems she had doing participant observa-tion in a British farming household. Shewrote about ‘the angst of not knowing whatto write down, what ought to be written,what’s the most important part of a conversation,an observation, an event …[and] the feelingsof drowning in voices and “data” and notknowing where to start.’ (Bennett andShurmer-Smith, 2001: 255). Her supervisorsimply said: ‘Just keep writing’ (2001: 255).What sort of writing are we talking about,though? How spontaneous and/or systematicshould it be? It’s not just a simple matter ofcollecting or recording experiential data from

the field. Fieldnotes can never be anythingother than ‘fictions in the sense that they are“something made”, “something fashioned”’(Geertz, 1973: 15 see also Willis andTrondman, 2000; Bennett and Shurmer-Smith, 2001; Hoggart et al., 2002). And theseare ‘fictions’ which shape as much as theyreflect researchers’ experiences of the researchprocess (Wolfinger, 2002). As soon pen is putto paper, or fingers to keys, the recording andinterpretation of data blur into one another(Burgess, 1986).

Field-noting is an ongoing sense-makingprocess. It is a process of creative writingbased on first-hand experience. It involvesattempts to tie together minutiae of theoreti-cal and empirical detail gleaned in andbetween the different locales of a project’sexpanded field (Batterbury, 1994). It usuallyinvolves more than one kind of writing formore than one audience. Some of what getswritten should be suitable for publication inraw or refined forms. Yet significant parts,however ‘juicily’ relevant to the research, mayhave to remain private to protect theresearched.22 Moreover, for a researcher oftenisolated in the field, a notebook may be aplace to write things purely for him or her-self, to write out the frustrations of beingthere, struggling to do that research, to let offsteam, let rip, write all kinds of intemperate,intolerant, lusty, clumsy, emotional things(Parr, 2001) that could never, ever, beincluded in a final product but might be eas-ier to cope with once they have been put intowords. For many researchers, this self-reflexive,soul-searching, worry-writing may comprisethe majority of their fieldnotes.

Whatever the contents, field-noting isoften done in stages, starting with hastilywritten scratch notes jotted down then andthere (‘inscriptions’), which are then elabo-rated upon at regular intervals (ideally eachday) in more flowing narratives (‘descriptions’),

DOING ETHNOGRAPHIES

197

3151-Ch-06A.qxd 3/2/04 2:43 PM Page 197

parts of which may later be processed into themore polished tales from the field that canappear in finished products.23 After the initial,highly detailed scenes, characters and roleshave been written, field-noting may narrowits foci as researchers mention only changes tothese established scenes, etc., and focus in on,latch on to and follow up parts of these(Wolfinger, 2002). Here it is common forresearchers to begin to identify patterns andregularities, rhythms and routines, dominantdiscourses and ways of seeing/doing, to checkfacts and verify claims made in onetime/place/setting in others, and so on(Jackson and Smith, 1984). Identifiable indi-viduals, places and so on often have to beanonymized, usually in the last stages of writ-ing a finished product, but sometimes as thenotes are being written if there is a dangerthey could fall into the wrong hands (seeWarren et al., 2000; Dowler, 2001).The sense-making in these notes may be the work notonly of the researcher but also of co-workerssuch as research assistants and translators, andthe ways these perspectives come togethermay also need to be described (Twyman et al.,1999). Particularly where researchers work infield areas radically different from their homecontexts, field-noting may involve writingabout taken-for-granted aspects of identitiesand experiences that may become morenoticeable but may also be difficult todescribe (Punch, 2001).

Field-noting takes up a considerableamount of time, with every hour of partici-pant observation requiring, according to oneestimate, two hours of note-taking (Junkercited in Burgess, 1986).This means that a care-ful balance has to be struck, in the field,between the what, when and where of doingand writing. Field-noting (along with otheradministrative elements of data constructionsuch as filling and sending mail) is not some-thing that should simply be done in the

researcher’s spare time (Crang, 1994).Moreover,wherever and whenever it is done, it provokesemotions (especially feelings of guilt that thisis a betrayal of participants’ trust) as much as itattempts to describe their role in and betweenresearch settings throughout the researchprocess (Stacey, 1988; Cook, 2001; Bennett,2002). Great care has to be taken to ensurethat these notes and other bits of data don’t getlost or confiscated. Access to a photocopierand/or a laptop computer is therefore essentialto make no-extra-time duplicates which canbe kept in or sent to a number of safe places(Dowler, 2001). Finally, only the tiniest pro-portion of these notes end up on the pages ofcompleted works. So, how can good notes bewritten and what do they look like?

Participant observation research is supposedto provide its readers with a vivid impressionof ‘being there’. So how might we conjure upa sense of place of the locale(s) in which theresearch is based and how this might changeover the minutes, hours, days, weeks andmonths of the research? How might wedescribe the rhythms and routines of thoseactors who inhabit and/or comprise that set-ting? How might we move from these obser-vations to describe our interactions in thatsetting with those actors, initially and as theydevelop over time? How might we describenot only what we are learning but also theoften strange and strained circumstancesunder which that learning is taking place?How might we describe the difference thismakes to the way the research seems to shap-ing up, where it can and should go, and whatpower we have over this? And how would wedescribe how this develops in and betweenthe locales of our expanding fields?

To answer these questions,we echo the guid-ance given to researchers planning semi-structured interviews:draw up a checklist whichreminds you of the issues you should probablycover, in whatever order they come up, but be


198

3151-Ch-06A.qxd 3/2/04 2:43 PM Page 198

prepared to ‘go with the flow’ when theunexpected happens (Evans, 1988). For moreobservational or substantive field notes, attentioncould be paid to the following list of prompts:

1. Space: the physical place or places; 2. Actor:the people involved; 3. Activity: a set ofrelated acts people do; 4. Object: the physicalthings that are present; 5. Act: single actionsthat people do; 6. Event: a set of related activ-ities that people carry out; 7. Time: thesequencing that takes place over time; 8. Goal:the things that people are trying to accom-plish; [and] 9. Feeling: the emotions felt andexpressed. (Wolfinger, 2002: 91)

These could be supplemented by more spec-ulative prompts which can help the researcherto think through what might have been goingon there, why that might have been the caseand how this might be further investigated.For example:

Who is he? What does he do? What do youthink she meant by that? What are theysupposed to do? Why did she do that? Whyis that done? What happens after ________?What would happen if ________? What doyou think about ________? [and] Who isresponsible if ________? (Lofland andLofland cited in Wolfinger, 2002: 90)

This is where observational writing begins toblur with participatory writing. So in additionto and/or, wrapped up in, these substantivenotes, it is essential to discuss the more reflex-ive, participatory aspects of the researchprocess. These are the ‘methodological field-notes’: the ‘detailed records of the researchprocess, research relations, and research roles’,which provide a means both to ‘allow theresearcher to talk things through; a dialoguewith oneself ’ and to provide the raw materi-als to ‘construct an autobiographical accountof the major phases in the research process’ inthe final report (Burgess, 1986: 58–9).

Methodological footnotes often comprisedescriptions of the ways in which researchunfolds or takes shape in or between the settings

and interactions described in the substantivefieldnotes. And their prompts should relatemore to the research process and theexpanded field as they take shape in practice.These could include descriptions of:

• The selection of research topics.• The networking that helps them to take

place and shape.• The ways in which the researcher man-

ages first and subsequent impressions inand between the various social settingsand interactions which comprise the field.

• How people seem to be reacting to himor her as a result of his or her age, gender,skin colour, language skills, nationality,politics, dress, etc., and how he or sheattempts to work with and/or against thisto get the research done.

• The ways in which he or she identifiesand (un)successfully tries to build relation-ships of trust with key informants,what heor she learns from shadowing and talkingto them and how others appear to seethese relationships.

• How circumstances change and break-throughs are made.

• How research methods and theoreticalstandpoints are tailored to match what’spossible, what’s happening, and what hasunexpectedly turned up through theresearch process.

• The way that field-noting, photographyand other forms of data construction havebeen able to take place and what effectstheir inscription and description have onthe ongoing research process.

• How the researcher becomes part of thescenes he or she is describing, the effectsthis has on the research process, and on hisor her sense of self and his or her findings.

• How all these issues are brought back to,and used in, the institutional contextwhere the research is formally based.

DOING ETHNOGRAPHIES

199

3151-Ch-06A.qxd 3/2/04 2:43 PM Page 199

To conclude, in Boxes 6.1–6 we provide arange of raw and polished examples of thesenotes from published and unpublished geo-graphical ethnographies. These boxes movefrom the most observational, substantive andpublic to the most participatory, method-ological and private field-noting. They sug-gest six layers of description, movingprogressively from outside to inside perspec-tives: from locating an ethnographic setting,to describing the physical space of that set-ting, to describing others’ interactions withinthat setting, to describing researchers’ partic-ipation in those interactions, to their reflec-tions on the research process, and endingwith self-reflections provoked by theresearch.We ask to what extent they begin to‘take you there’, to that drop-in centre,courtroom, street, restaurant, farm, village orbus journey, and to what extent they canplace you in the shoes of the researchers who

produced them? What level of detail seems towork best as you read these notes? Given thekind of final product you are workingtowards, do you think it would be better todisentangle your substantive and methodolog-ical notes when you produce that final prod-uct or when you produce its fieldnotes (seeHoggart et al., 2002)? What could be addedto these depictions by using alternativemedia? All we can suggest is that you have ago at this yourself, before embarking on yourproper research. Go to an unfamiliar placeand think about how you would try to sharethat experience with someone who hadnever been there. Go back a few times with acamera, sketch pad, whatever. Get somethingdown, in detail. Hand it over to your likelyaudience. See what works for you and forthem. Go back if necessary to elaborate onthe thinner parts of your description, etc.,until it seems to click.


200

Box 6.1: The First Layer:Locating an Ethnographic Setting

Parr’s ‘St Peter’s’ drop-in centre.

‘The drop-in (which I shall call ‘St Peter’s’) is located in Nottingham’s inner city, in adistrict well known for residential hostels, care homes, sheltered housing and dayschemes for people with mental health problems. … St Peter’s (see Figure 1 [above]) isa drop in located within a church hall and had been established for about 10 yearswhen the research took place’.

Source: (Parr, 2000: 226)

3151-Ch-06A.qxd 3/2/04 2:43 PM Page 200

DOING ETHNOGRAPHIES

201

Box 6.2: Layer 2: Describing the Physical Space of that Setting

a) Sketching a setting: Parr’s drop-in centre (the re-drawn version).

Source: Parr (2000: 230).

b) Describing a setting: Spradley’s courtroom:

‘There were rows of spectator benches, all made of heavy dark wood, oak or walnut,to match the panelled walls. The rows of benches went for more than twenty-five feetuntil they met a railing that seemed to neatly mark off a large area for ‘official busi-ness’ … At the right of the area behind the railing were twelve high-back leather chairsbehind another railing. A large oak table with massive chairs all faced toward a highlectern which I took to be the judge’s bench.

Source: Spradley (1980 cited in Wolfinger, 2002: 88).

CAR PARK

OTHER ACTIVITIES

DRIVEWAYTO STREET

CARPARK

OFFICESTAFFONLYMAIN DOOR

LOBBY

LONERS’ SEATING

POOL TABLE

HATCHFOR

FOOD

GROUP TABLES/SEATING

KITCHENSTAFF ONLY H

IGH

WIN

DO

WS

Box 6.3: Layer 3: Describing Others’ Interactions within that Setting

Smith’s observations of street policing:

‘… one quiet afternoon along the Lozells Road, the orientation towards ‘fun’ of anebullient group of black youths eating the fares of ‘Dolly’s Takeaway’ encountered the‘leisured’ frame of two police officers approaching from the opposite direction. A chainof subtle gestures was observed as the two parties became aware of each others’approach. An almost imperceptible sequence of exchanges provided enough impetusto transform the episode from one frame (fun and leisure) to another (danger and sus-picion). The sequence was as follows. The police officers seemed to stiffen their walkand move closer together. One twitched his hand towards what might have been aradio, and both fixed their eyes on the oncoming group with the consequence that

3151-Ch-06A.qxd 3/2/04 2:43 PM Page 201


202

Box 6.5: Layer 5: Reflections on the Research Process

Putting plans into action? Cook’s experience on a Jamaican papaya farm

‘I tell Jim [the farm manager] what I would like to do … which has now become sim-plified to the impact on the farm and surrounding villages of changes in consumertastes in the UK. I tell Jim as we leave to mount our bikes again, that I have decided …to make all the contributors to my work anonymous because I don’t want to provokea consumer boycott based on a highly selective reading of the negative aspects ofwhatever I might find. I emphasised that in my work, I would try to be even handed. I told

their subsequent conversation spilled melodramatically from the side of their mouths.The effect was to reframe a leisurely stroll into a display of professionalism. The youths,in turn, exaggerated their usual bounce into what could have been construed as aswagger. They talked more loudly and gesticulated more often. A frame orientedtowards fun slipped through into one characterised by ritual demonstration. At thesame time, the reserved window–shopping of two elderly white pensioners became anexpression of passive solidarity as they linked arms and drew their bags to their bodies.This became, too, an expression of deference, as they crossed the road to remove them-selves from the path of the encounter. However, as the groups moved closer and oneofficer’s request for a chip was laughably answered with the flash of an empty packet,a friendly shrug of the shoulders was sufficient to dispel the atmosphere of confronta-tion and replace it with one of amiable sparring’.

Source: Smith, S.J. (1988: 32–3).

Box 6.3 (Continued)

Crang’s experience of waiting in a restaurant

‘An initial panic and despair at not being able to cope with the pace of work, coupledwith a demonstrative rushing around and impatience with any delay is eventuallyreplaced by a fragile order, the reduction of customers to ticks on a list. All movementnow has a direction, there is none of the earlier aimless congregating; I map out mypriorities, I am heading somewhere all the time, and have the next four or five destinationsplanned out too. Yet this order is vulnerable. Firstly, any interruptions by customers arelikely to throw it all into disarray. For that reason I try to control all contacts with tables,for example moving around all my tables to see if there are any queries when I have aspare minute rather than waiting to be called. Secondly, for me that sense of directionand the sheer amount of energy I’m expending mean I need a constant boost; and thatis given by tips. If they don’t come my motivation collapses. Luckily last night they werepretty good so that feeling of occasional despair as it all gets too much seems worth it’.

Source: Crang, (1994: 683).

Box 6.4: Layer 4: Your Participation in Interactions in that Setting

3151-Ch-06A.qxd 3/2/04 2:43 PM Page 202

him I’d change his name and a few key identifying features. He said that this had beenone thing he had been kind of worried about, and seemed reasonably happy with thearrangement. I think he trusts me, now …? I said to him at the gate that I thought thatmy nosing into everyone’s business was what each wanted to know and that, hopefullywhat each gave would be at least equally compensated by what they learned. Ofcourse, there’s no way I can guarantee this and it might be dangerous, but we can workwith it … ? We cycled back to the banana field … We met Philipps, Jim’s right hand man …in his white pickup. Jim told him what I was doing as I had told him and asked whetherhe would introduce me to the packing house people and tell them what I was doing. Iput my bike in the back and we drove there. Walked in (the house had been rearrangedfor reconcreting the floor), and Philipps called the women one by one, then all at once,and introduced me as being from the Univ of Bristol, interested in the relationshipbetween this and that and would be asking questions about things like their wages andchanges in the area. They looked. I thanked Philipps and thought that was that, I’dcome back tomorrow and start. He then took me over to the wrapping bench andshowed me how they were wrapped. I thought, OK, that’s what I’ll do tomorrow. Butwe stayed there for a few minutes like wet lemons, so I started doing it. I was in for theday’.

Source: (Unpublished notes, Wednesday April 1 1992)

Data reflecting/affecting the scene? Robson taking photos in aNigerian village

‘… taking photographs caused a lot of problems and was a very uncomfortable part ofthe fieldwork. People became demanding that I should come and take their pictureand if I did then every time they saw me after that would insistently want to knowwhere their picture was. This led me to only taking my camera out on special occasionsand carrying it concealed in a bag so people would not see it and demand to have theirpicture taken. It was also frustrating because I wanted to be able to take natural pic-tures. However, the villagers had very set ideas about photography and unless theywere unaware that they were being photographed, they would insist on putting ontheir best clothes and pose rigidly for the camera with very serious expressions, womenalways with downcast eyes’.

Source: Robson, (1994: 45)

DOING ETHNOGRAPHIES

203

Encountering Tanzanian racism on the bus: Nagar’s letter ‘home’

‘At times it pains me so much to have brown skin here [in Dar es Salaam]. I take the busevery day to town and most of the times I am the only Asian riding on the bus.Sometimes I am greeted sarcastically, ‘Kem Chho?’ (Gujurati for ‘how are you?’). I feelangry … because I can hear and feel the resentment against the brown skin in the voicethat is greeting me. And I feel like screaming: ‘I am not an exploitative, racist …Muhindi (Asian) from here. Don’t look at me like that. I have nothing in common with

Box 6.6: Layer 6: Self-reflections

Box 6.5 (Continued)

3151-Ch-06A.qxd 3/2/04 2:43 PM Page 203

Notes

1 See, for example, Cook and Crang (1995),Flowerdrew and Martin (1997), Hay (2001),Kitchin and Tate (2000), Limb and Dwyer(2001) and Hoggart et al. (2002).

2 See Domosh (1991a), Blunt (1994b), Behar(1995) and McEwan (2000).

3 See Merrifield (1995), Gregory (1996), Fuller(1999) and Hoggart et al. (2002).

4 See Buttimer (1978), Ley and Samuels (1978),Jackson and Smith (1984), Livingstone (1992),Johnston (1997b) and Philo (2000).

5 See also Rowles (1978a), Jackson and Smith(1984) and Eyles, (1986).

6 Since 1985 Rowles has held a joint position atthe University of Kentucky as a professor inthe Department of Geography and in theSanders–Brown Center on Aging. Findingsfrom his Winchester Street study, and the fol-low-up study in a rural Appalachian towncalled Colton, were largely published outsidegeography in, for instance, the Journal ofHuman Development (1981a), TheGerontologist (1981b), the Journal ofEnvironmental Psychology (1983) and TheHandbook of Clinical Gerontology (1987).

7 For example, Nash (2000), Parr (2000), Thriftand Dewsbury (2000), and Smith, S.J. (2001).

8 For example, Clifford and Marcus (1986),Marcus and Fischer (1986), Stacey (1988) andBehar and Gordon (1995); see Cosgrove andJackson (1987), Gregory and Ley (1988),McDowell (1992a), Gregory (1996) andShurmer-Smith (1998a).

9 Duncan’s (1980) paper was ‘revisited’ as a‘classic in human geography’ in 1998 (see

Duncan, 1998; Mathewson, 1998; Shurmer-Smith, 1998a).

10 As well as the humanistic ethnographiesdiscussed above, the culture industry ethno-graphies of Dydia DeLyser (1999) and MikeCrang (2000), and the small-scale culturewar ethnographies of Garth Myers (1996)and Richa Nagar (1997b), are worth readingin this respect (see Herbert, 2000; Crang,2002).

11 See Clifford (1992), Marcus (1998), Hoggartet al. (2002) and Laurier and Philo (2003).

12 A standard dictionary definition of kinaes-thesia is ‘the sensation by which bodily posi-tion, weight, muscle tension, and movementare perceived’ (Anon, 2001: 809).

13 For example, in making a cup of coffee(Angus et al., 2001), using a remote controlor dog lead (Michael, 2000), windsurfing(Dant, 1999), working in an office(Hetherington and Law, 2000) or car (Laurierand Philo, 2003), or making natural historyfilms (Davies, 1999; 2000; forthcoming).

14 Sara Kindon’s (2003) participatory videowork with members of a Maori tribe inAotearoa, New Zealand, and DuncanFuller’s (1999) work with/for a credit unionin Hull in the north of England are notable‘full blown’ examples of such work.However, Batterbury’s (1994: 81) admissionthat the NGO lobbying that was not part ofhis formal research strategy gave him the‘most personal satisfaction’ shows how thisaction research can work on a much smaller,informal scale.

15 This includes work by Michael Keith (1991;1992) and Steve Herbert (2000) with policeofficers, by Steve Hinchliffe (2000) with


204

them!’ Being resented like that gives me some sense of what an antiracist white personmust feel in a black Chicago neighbourhood. But then my position here as an Indianstudent from India is not quite the same as a radical Muhindi from Tanzania either. Myproblem is that I am a foreigner here, and at times I want to be recognised as such. Butanyone who does not know me thinks that I am just an arrogant and aloof Muhindiwho does not know enough Kiswahili in spite of having lived in Tanzania all her life. Ithurts me so much to be resented here … And what really frustrates me is that thenature of my research keeps me away from the Waswahili and I just don’t have enoughtime to get absorbed in their culture because I am too busy trying to get absorbed intothe Asian communities’.

Source: Nagar, (1997: 215)

Box 6.6 (Continued)

3151-Ch-06A.qxd 3/2/04 2:43 PM Page 204

managers on training courses, by PamShurmer-Smith (1998b) with Indian adminis-trative service workers, by Eric Laurier andChris Philo (2003) with mobile sales teamsand by Gail Davies (1999; 2000; forthcom-ing) with natural history film-makers.

16 Elspeth Robson (1994: 57–8), for example,wrote how ‘in my “normal” life as a PhDstudent [in the UK] I play the role of a stu-dent, a learner and “subordinate”, but in thefield [in the Nigerian village where I worked]roles were reversed and I found it difficult tobe the teacher, the one calling the tune,deciding what we should do each day, direct-ing my assistants to do work for me’.

17 See Jackson and Smith (1984), Keith (1992),Katz (1992; 1994), Madge (1994), Clifford(1997) and Cook et al. (in press).

18 See Crapanzano (1977), Madge (1994),Robson (1994) and Nagar (1997a).

19 We recommend starting with Robson andWillis (1994a), Hughes et al. (2000), Limb andDwyer (2001) or Ogborn et al. (forthcoming).

20 Here we recommend starting with Cookand Crang (1995), Flowerdew and Martin(1997), Hay (2000), Kitchin and Tate (2000)or Hoggart et al. (2002).

21 We are referring here to the possibilitiesoffered by CD or DVD technologies or byonline publishing such as that in the radicalgeography journal ACME (www.acme-journal.org).

22 See Burgess (1986), Stacey (1988), Bennett andShurmer-Smith (2001) and Dowler (2001).

23 See Clifford (1990), Crang (1994), Cook andCrang (1995) and Capps (2001).

DOING ETHNOGRAPHIES

205

3151-Ch-06A.qxd 3/2/04 2:43 PM Page 205

3151-Ch-06A.qxd 3/2/04 2:43 PM Page 206

Geographical interpretations

If human geographers do use data in their research (see the introductionto Part I), it is obvious that we have do something with these data in thecourse of bending them into intelligible findings amenable to the drawingof conclusions, the offering of speculations and even, on occasion, thetendering of predictions. A few human geographers might suppose that‘the data speak for themselves’, implying that once you have gathered orgenerated the relevant data it is possible simply to ‘write up’ or ‘write out’these data as findings which unproblematically mirror what is happeningin the world beyond the pages of the academic text. The majority ofhuman geographers would dispute such a supposition, though, not leastbecause any sensitivity to the complexities of how data are constructed(whether pre- or self-constructed) demonstrates that data can neverstraightforwardly mirror the world. All the countless complications integralto what data sources entail, the myriad factors integral to how they arefirst ‘made’ and then brought together, whether by the researcher or byother people, give the lie to any belief in the innocence, naturalness,transparency and so on of data.

The discussions throughout Part I of our book serve to challenge anysimplistic assumptions about both what data entail and what they canreveal. This is not the end of the matter, however, since there is morethat also needs to be said about the very definite ‘labour’ which the humangeographical researcher cannot avoid expending in the process of turningdata into plausible findings, conclusions, speculations or predictions. Inshort, attention needs to be drawn to the very substantial amount of workthat has to be done on data before they can, as it were, be ‘made to speak’.Part of this work does involve being fully cognizant of how data sourceshave been constructed, as we have emphasized, but there is a sizeablefurther dimension to this work that comprises a diversity of ‘strategies’through which these data sources are interpreted by the researcher. Theinterpretations that thereby result must themselves be seen asaccomplishments, as ‘made’ by researchers through their labours on theirdata, and this is why we use the slightly odd-sounding phrase ‘constructinggeographical interpretations’ to capture what we see occurring here.

PART II CONSTRUCTING GEOGRAPHICAL INTERPRETATIONS

3151-part-02.qxd 3/2/04 2:52 PM Page 207

A few existing textbooks on matters of methodology touch upon the workof interpretation, but none of them lays out, explores and critiques as wedo below the overall portfolio of different interpretational strategiesavailable to human geographers. We might immediately add that suchtexts tend to take ‘interpretation’ as referring to a specific strategy akin towhat we term ‘understanding’ in Chapter 10, whereas we are expandingwhat we mean by ‘interpretation’ to the status of an omnibus termencompassing all of the different ways of working on data covered inPart II.

Interpretational strategies

Such interpretational strategies have not always been acknowledged byhuman geographers, some of whom have been interpreting data withoutrealizing they have been doing so, and this is lamentably true of thoseresearchers who conceive of what they are doing as just ‘common sense’ oras merely following conventional scientific procedures. In our view, suchhuman geographers are most definitely adopting interpretationalstrategies, ones whose inner logics can actually be specified quite closelyand ones carrying with them implications about what sorts of data countas credible and what kinds of findings, conclusions and so on can be arrivedat. These geographers are interpreting their data, they are working ontheir data according to frameworks of interpretation which we can readilyidentify, even if they fail or refuse to recognize that this is the case. Hencein the chapters that follow we will, to some extent, be addressing a rangeof unacknowledged interpretational strategies, showing how they stillperform the task of allowing researchers to ‘make sense’ of their data.

For many human geographers today, there clearly is an awareness ofbringing interpretational strategies to bear on their data – even if theymight not use this terminology – and such geographers are more or lessself-reflexive (see Chapter 1) about the work they are doing inendeavouring to move from data into findings, conclusions and beyond.This being said, the common format in which such issues are debatedrevolves around the different philosophies and social theories beingmobilized to inform research projects: the philosophies which fashion aresearcher’s prior assumptions about the nature of the knowledge beingproduced (assumptions about ontology and epistemology), and the socialtheories which frame a researcher’s root vision of how human society isconstituted (the inter-relationships between the everyday existence ofindividuals or groups and the more enduring properties of institutions,systems and structures). This is the terrain of conceptual orientations inhuman geography, as dealt with in a text such as Paul Cloke et al. (1991) orthat by Richard Peet (1998), and there is extensive and often heated debate

PART II Constructing geographical interpretations

208

3151-part-02.qxd 3/2/04 2:52 PM Page 208

about the philosophies (e.g. from positivism to humanism topoststructuralism) and the social theories (e.g. from social Darwinism toMarxism to feminism) that human geographers could or should entertain.And yet related to these conceptual orientations – cross-cutting them inways which are not always as obvious as might be supposed – arenumerous different interpretational strategies standing as bridgesbetween abstract conceptual concerns and the data on which researcherswork when striving for findings, conclusions and the like. It is upon theseinterpretational strategies that we now focus, teasing out identifiablestrategies for working on data that can be adopted, and urging a greaterself-reflection on such strategies that is independent of – or at least notdetermined by – more overt, we might say more abstract or high-level,philosophizing and social theorizing. In the chapters that follow referencewill be made to philosophies and social theories so that concepts ofpositivism, Marxism, psychoanalysis and such will be mentioned, but theinterpretational strategies under the microscope here cannot be ‘reduced’to any one of these a priori conceptual orientations. To our thinking, theseinterpretational strategies do entail something different again: certainstances before data – modes of probing into and manipulating data – thaturgently need to be unpacked by human geographical researchers. In thisconnection, then, we are unquestionably operating on a methodologicalplane as this is described in Box 1.4.

In the chapters that follow we identify and discuss five interpretationalstrategies, or perhaps it would be better to speak of three such strategiescross-cut by two interpretational processes or moments (those of both‘sifting and sorting’ and ‘representing’) that cannot but be bound up in theother three. For simplicity, though, we will refer to all five as strategieswhile admitting that it is only our middle three (‘enumerating’, ‘explaining’and ‘understanding’) that offer researchers the most obviously coherent –and, as such, often differentiated and perhaps counterposed – modes ofstrategic intervention in the tackling of data. We have debated long andhard how to characterize these interpretational strategies, what names togive them and what considerations are germane to each, and we realizethat the organization and the contents of these five chapters will seemunfamiliar and even strange to many readers. There are also doubtlessthings that are missing here, stances and modes of working on data aboutwhich we do not say enough. Even so, we remain convinced that thestrategies identified here do embrace the principal tasks – some beingessential and some being contrasting possibilities – that humangeographers can and do bring to bear in turning data into findings,conclusions, speculations and predictions. Some of these strategies or atleast aspects of them are complementary, open to being combined, whileothers are likely to be pursued sequentially in a given research project. Asalready stated, no project can avoid having a ‘sifting and sorting’ moment,

CONSTRUCTING GEOGRAPHICAL INTERPRETATIONS

209

3151-part-02.qxd 3/2/04 2:52 PM Page 209

for instance, and no project that is written up in some form can avoidhaving a ‘representing’ moment (when the findings, conclusions and thelike are written through for an audience of readers or listeners). Similarly,aspects of ‘enumerating’ and aspects of ‘explaining’ clearly link into oneanother on occasion, while the gulf between ‘explaining’ and ‘interpreting’is rarely as wide as certain conventional accounts of methodology imply.This being said, some elements of these different interpretationalstrategies may be quite hostile to each another, and a simple point is thatdifferent, often very different, human interests – ones often wrapped upin very different visions of what human geographical research is for in thewider world – fuel the advancing of claims about the vastly superior meritsof differing interpretational strategies over others. Nowhere is this moreapparent than when those who purport to explain the world (especially ifcrunching through quantitative data) come into conflict with those whopurport to understand the world (especially if immersing themselves inqualitative data). A modest ambition of what follows is thus to questionthe oppositions and hostilities that sometimes arise in this respect.

From initial siftings to final writings

Once researchers start to encounter data, choosing to examine certain datasets rather than others – as soon as they are selecting certain columns ofnumbers from a table for attention, choosing to concentrate on certainpassages, minutes or entries from a written document, being captivated bycertain responses in a questionnaire return or remarks in a taped interview,electing to record certain observations and impressions in a field diary andso on – they are taking the first steps in the interpretational strategy weterm ‘sifting and sorting’. This strategy is, therefore, intimately bound upin the most basic of dealings with data, and when data are being self-constructed it is probably true to say that the interpretational work ofsifting and sorting has already begun. It is a primitive of all humangeographical research, then, but in its universality and ever-presence it isalmost never subjected to self-reflexive scrutiny: it is almost never broughtinto the light for special examination and critique. This being said, whilethe effort to impose order on data cannot but be there from the veryoutset of a project, we suppose there to be dimensions of the sifting andsorting strategy that become more prominent once the researcher directlyconfronts the masses of data obtained from the field. It is what happenswhen the researcher begins to order the data by identifying ‘entities’within them that are reckoned to be obvious, significant or meaningful; itis what happens when he or she begins to categorize things, to detectbinaries in the data and to draw up hierarchies of more or less importantitems in the data. No project can proceed without doing work along these


210

3151-part-02.qxd 3/2/04 2:52 PM Page 210

lines, even if it is rarely acknowledged that this work is taking place andmaking a difference to what the data are finally allowed to tell us. Chapter 7duly tackles sifting and sorting, reflecting quite abstractly, even philoso-phically, on the issues involved.

The next three chapters cover ground which will probably be morefamiliar to readers, albeit doing so in a fashion that might seem unfamiliar,disconcerting even. Chapter 8 tackles ‘enumerating’, arguing that a primeinterpretational strategy for many human geographers is one that convertsdata into numerical forms – one that demands quantification of the worldby counting or measuring its contents – and then proceeds to all mannerof numerical analyses of these quantitative data ranging from the simplestof descriptive statistics (e.g. calculating a mean) to the most complicated ofspatial modelling and geocomputational visualization. Some humangeographers might suppose that enumeration is simply working with thefacts of the world, believing that the numbers are somehow ‘out there’ inthe world, but most of us would now acknowledge that to pin a numberon to something is a decisively human activity which translates thatsomething into a human ‘language’ wherein it can be technically manipulatedstatistically, mathematically and computationally. However simple orcomplex the manipulations, they cannot but insert a distance – we wouldcall this an interpretational distance – between the world (and dataextracted from it) and the researcher’s findings, conclusions and the like(when efforts are made to translate back from numerical languages intowords more obviously referring to tangible phenomena, events andprocesses around us).

Chapter 9 tackles ‘explaining’, arguing that many human geographersoperate within explanatory frameworks which, albeit in a host of differentways, purport to provide the tools allowing us to answer ‘why’ questions:Why did that settlement get located where it did? Why did that part of thecity end up derelict and riddled with poverty? Why did people in thatregion develop a strong sense of their own ethno-regional identity andhence wish to secede from their host nation-state? Explanatoryframeworks are approaches to knowledge production which think interms of cause-and-effect relations or, more subtly, of underlying causalmechanisms which exert an influence through differing media – people’spsychologies, cultural expectations, economic tendencies – to effectoutcomes which might include the location of a settlement, theimpoverishing of a city district or the upsurge of dissent in a given ethno-region. The job of ‘science’ has usually been regarded as that ofexplanation, particularly if explanation can feed into prediction, and this isas true of conventional science (the hypothesis-testing science that most ofus know from the school science room) as it is of more recent post-Newtonianscience (which now accepts ‘uncertainty’ as a feature of explanation) and anynumber of other intellectual positions which have claimed the mantle of


211

3151-part-02.qxd 3/2/04 2:52 PM Page 211

science (from Freudian psychoanalysis to Marxist historical materialism toHusserlian phenomenology). There are, therefore, numerous differentversions of what explanation entails and of how it can be achieved and, toa limited extent, these can be contrasted as ‘surface’ or ‘depth’ models ofexplanation, but for us they are all interpretational strategies whichresearchers operationalize in their efforts to make data yield up definitiveanswers to the sorts of ‘why’ questions listed above.

Chapter 10 tackles ‘understanding’, arguing that many humangeographers – notably ones who reject the prevailing models ofexplanation – prefer approaches where the chief objective is to ascertainthe meanings which are held in the hearts and minds of people central tothe human situations under study. In contrast to Dilthey and Weber’sconstruct of Eklären (i.e. ‘explanation’), many human geographers favourtheir construct of Verstehen (i.e. ‘understanding’), which is all about theexcavation of meanings, values, beliefs, hopes and fears which coursethrough what are occasionally called the ‘thought-worlds’ of people whoare the researched (whether they be princes or paupers, company excutivesor the socially excluded) (e.g. Gregory, D., 1978a: 133). For these geographers,the research task is basically complete once the researcher has obtainedsome understanding of the meanings, values and the like figuring in thethought-worlds of the researched. Drawing inspiration chiefly from thehumanities and cultural studies, the tendency is to depend upon highlyqualitative methods, like in-depth interviewing or participant observation.And so the anticipation is that, through varying forms of encounter,empathy and dialogue, the researcher can come to understand keydimensions of how the researched think (and feel) about their lives and howthey then act – what they do, whether routinely or as one-off occurrences –as decision-makers in their everyday lives at home, in the street, at work orwhatever. This being said, in more sophisticated versions of understanding,consideration is given to the broader discourses (ideas, beliefs, opinions) thatframe what research subjects think (and feel) about their worlds, in whichcase the researcher’s understanding does not lodge solely in theidiosyncrasies of individual people encountered in the research process. Justas there are different models of explanation, so there are different modelsof understanding, and there are also different levels of work involved thatinclude addressing both highly conceptual problems (as in rarefied debatesabout hermeneutics) and mundanely technical ones (to do with how to ‘codeup’ qualitative data sources in the hunt for clear patterns in the realm ofmeanings, values and beliefs). Once again, though, to our minds these are allfacets of an overall interpretative strategy which is understanding: they areall directed to how the data can be shaken to reveal the fundamentallymeaningful qualities of those portions of humanity that we purposefullymeet or accidentally stumble across in our inquiries.

Integral to some of our endeavours when we are gathering or generat-ing data is a process which can rather grandly be called ‘representing’.


212

3151-part-02.qxd 3/2/04 2:52 PM Page 212

Whenever we write notes on a document found in an archive, scribble lineson the context and proceedings of an interview or more formally write upour observations and impressions in a field diary, we are starting torepresent our data in textual form. Indeed, in many cases the data are thistextual form because we have nothing else to carry back with us to ouroffices and computer suites. What we write and how we write it at thisstage is, therefore, already part of the interpretative process that we call‘representing’ geographical texts, and it should be evident that what weare doing here is also very much embedded within the sifting and sortingcovered in Chapter 7 (at this very preliminary stage these twointerpretative processes go hand in glove). Yet what principally concerns ushere is less these initial acts of writing through data, important thoughthey are, and more what happens when human geographers – in tandemwith subjecting their data to some combination of sifting, sorting,enumerating and understanding – seek to produce in their written outputfindings, conclusions, speculations and perhaps predictions to be presentedto an audience. This audience could be other academics reading a journalpaper; it could be students reading a textbook; it could be policy-makersreading a report; it could be people who have been researched reading areport back; or it could even be an audience hearing about the research atone-step remove, perhaps the general public reading a summary preparedby a journalist for a newspaper article. A host of considerations weavethrough these acts of representation, these re-presentations of peoples,lives, space and places which have been the focus of study, and one way ofquickly grasping what is involved is to think about how the people whohave been researched might react to representations of them. But morethan this, because we are in no doubt that acts of representation (i.e.constructing geographical texts) comprise a maze of different possibleinterpretative possibilities bound up in the diverse conventions, styles andrhetorics of writing available to the researcher. Chapter 11 duly tackles thisamorphous task of representing, reflecting quite widely, often critically, onthe issues involved.


213

3151-part-02.qxd 3/2/04 2:52 PM Page 213

3151-part-02.qxd 3/2/04 2:52 PM Page 214

Sitting downwith your data

So far in this book we have dealt at lengthwith the many ways in which geographicaldata are constructed, whether by the researcheror by others whose labours he or she thendraws upon.Whatever the route by which thedata have been constructed, a point arises –although it may well not preclude furthergenerating or collecting of data – when theresearcher ‘sits down’ with all or some of hisor her data and starts to puzzle out exactlywhat to do with them. This can mean quiteliterally ‘sitting there’, perhaps in something ofa daze, surrounded by piles upon piles ofmaterials: by box-files of photocopied officialsources (minutes, policy statements, structureplans); by results from questionnaire surveys;by photographs, mental maps and snippets ofpoems; by tapes and notebooks carryingthousands of interview words; by field diariesbrimful of participant observations; and thelike. (And it might help at this point if youwere to look at the discussion of the conceptsof ‘induction’ and ‘deduction’ as outlined inBox 7.1.) While some attempt might havebeen made to impose a measure of order onthis chaos of materials as it was constructed,perhaps by keeping together seemingly similarsorts of data (say, all the official sources) orperhaps according to where the data weregathered and from whom (say, from the

planners in a particular local authorityplanning department), to all intents and pur-poses this is still a chaos which needs to bebrought into some kind of order before it canbe tackled further.

There are all manner of decisions to bemade in this respect: deciding that several fieldobservations should be categorized together asexemplifying (say) a particular type of ‘settle-ment layout’ or ‘settlement social life’; decidingthat several passages from several official plan-ning documents should be categorizedtogether as exemplifying (say) a particularvariety of interventionist policy or ideologicalcommitment; or deciding that certain state-ments from a transcribed interview should becategorized together as exemplifying (say) cer-tain themes such as ‘belonging’, ‘nostalgia’ and‘alienation’. In each case decisions have to betaken about what comprises a clearly delim-ited field observation, documentary passage orinterview statement, and then additional deci-sions have to be taken about what are basically‘like’ and ‘unlike’ observations, passages andstatements worthy of being grouped togetherinto categories. It is usually these categoriesthat the researcher will utilize in subsequentinterpretations, but he or she will often wishto refer back to the original observations, pas-sages and statements (and the sensitiveresearcher will try hard to retain some alert-ness to the nuances within such individualpieces of data).

Sifting and Sorting7

3151-Ch-07.qxd 3/3/2004 4:08 PM Page 215

While the philosophical ground here is contested, a standard distinction is drawnbetween these two basic approaches to the conduct of academic research, each ofwhich implies a fundamentally different stance before the data constructed (whethercollected or generated) in a project. One geographer (Lindsay, 1997: 7; see also the con-cise statements in Walford, 1995: 4) suggests that inductive reasoning ‘works from theparticular to the general’, while deductive logic ‘works from the general to the partic-ular’. Induction depends upon careful examination of a substantial amount of ‘rawdata’, seeking to discern in them patterns and regularities which can be interpreted asbeing of some generality, importance or meaning. It is commonly supposed that induc-tion depends on statistical procedures to establish the presence of such patterns andregularities, although our argument throughout Part II of this book is that there arevarious strategies (quantitative and qualitative) which can be deployed to do this.Deduction entails prior specification of theories, models and laws which can be used toaccount for the details found in ‘raw data’ as derived from studies of particular phe-nomena. Again, it is commonly supposed that deduction depends upon statistical pro-cedures to compare expected outcomes (from the theories) with observed eventualities(what actually takes place), but our argument here is that there are various strategies(quantitative and qualitative) which can be deployed to do this. The usual thinking,moreover, is that only induction requires close attention to data – to ‘letting the dataspeak for themselves’ (Gould, 1981) – whereas deduction does not need to bother overlywith the niceties of data. In geography, regional geographers have been regarded asinductivists obsessed by data, while spatial scientists have been regarded as deductivistslittle troubled by them (a claim with a measure of justification; see Chapter 1). A similardistinction might also be detected between humanistic geographers with an inductivefeel for data and Marxist geographers with a deductive line on theory, but there ishardly a clear-cut divide here, and the picture becomes even more muddled in relationto feminist, postcolonial, poststructuralist and postmodernist geographies (see Chapter 1and later in this chapter). Our own view is that such distinctions are invariably over-drawn and that most research tends to contain both inductive and deductive momentsin a constant, if uneven and sometimes unappreciated, dialogue. Moreover, we believethat virtually all human geographical research projects, wherever we might wish to posi-tion them on the inductive–deductive continuum, display a sustained engagement withdata. And this means that they then have no choice but to ‘sit down’ with their data andto begin sifting and sorting them in the manner which is the concern of this chapter.

Box 7.1: Induction and deduction


216

It is worth adding that perhaps the mostclear-cut example of this kind of activity inhuman geography today arrives whenresearchers are undertaking what is called the‘coding up’ of textual evidence (notably tran-scripts of tape-recorded interviews), since itbecomes so obvious here that the researcher isseeking to discern an order in – and in effectto impose an order upon – the original chaosof data.We return to the activity of coding inChapter 10 when discussing what we describe

as the ‘artisanal’ approach to understandingqualitative data (see also Crang, 2001; Jackson,2001). The basic point, however, is that theresearcher simply cannot avoid having to do alot of preliminary work on the chaos of his orher data as the first stage of moving into theinterpretative part of his or her project. It isthis work that we refer to here as ‘sifting andsorting’, and which Mike Crang (2001) imagi-natively refers to as ‘filed work’ (as opposed to‘field work’), because it cannot but involve

3151-Ch-07.qxd 3/3/2004 4:08 PM Page 216

SIFTING AND SORTING

217

a process analagous to that of sorting outpaperwork ready to be filed in the appropriatedrawers and dropfiles of a filing cabinet. If aresearcher’s attempt at ‘constructing geographi-cal interpretations’ is going to be at all success-ful, there needs to be a lot of this careful siftingand sorting. It simply has to happen in aresearch project, it always has to happen and itshould therefore be reflected upon quite con-sciously.The fact is that little attention is usu-ally given to this activity, partly because it doesseem like – and probably often is – such athoroughly dull aspect of the research process.As Crang (2001: 215) puts it, the ‘relativesilence’ about what is entailed in sifting andsorting is unsurprising given that it ‘feels likean unglamorous almost clerical process’(which returns us to the hardly very excitingimage of office filing). The purpose of thischapter is, none the less, to examine thisabsolutely crucial yet rarely acknowledged ordiscussed part of the research process. In sodoing, we aim to show that it is precisely not asimple aspect of the process, even if some ofthe practical actions involved appear straight-forward and not warranting much comment.Indeed, we acknowledge that it is actuallyriddled with taxing conceptual issues whichcannot be assumed away by a reflexive humangeography.

Entitation

This sifting and sorting of data – combingthrough them to ascertain exactly what isthere; laying them out, rearranging them,classifying them, putting individual items ofdata into boxes and lists; perhaps going on todetect patterns in the lists and boxes – is not apart of the research process which receives theattention it should. It would be wrong toimply that this deficiency has never beennoted, though, and here we reproduce a use-ful observation taken from a book on ‘systemsanalysis’ in geography. Given that we are wary

of systems analysis as an approach to humangeography, for reasons relevant to this chapter(see Box 7.2), we have altered this quotationto remove references to ‘systems’. Even so, thesense remains apt for our purposes:

The [initial] phase of [data interpretation]involves two steps: the identification of thebasic components or entities of which a [sit-uation or problem under study] is com-posed; and the definition or description ofthese components in terms of measurableproperties or state variables. This phase is ofenormous importance but it has not alwaysbeen given the attention that it would war-rant. The process of identifying entities hasbeen termed entitation; the process of mea-suring entities has been styled quantifica-tion. … Entitation is vastly more importanta stage in [data interpretation] thanquantification: Gerard (1969) demonstratedthis point with the example of primitiveastronomy where stars (basic entities) weregrouped into constellations which wenow know were, in reality, meaningless.(Huggett, 1980: 29)

The term entitation (see also Langton, 1972;Chapman, 1977; Gregory, 1986) usefully cap-tures something of what we think is entailedin the preliminary ordering of data, a task thatresults in the basic entities relevant to aresearch project being detected, delimited anddescribed, and perhaps also seen in basic pat-terns relative to one another. Another termfor this chapter could hence have been ‘geo-graphical entitation’. In the quotation it issuggested that entitation precedes quantifica-tion, the numerical measurement of the enti-ties so revealed, and in Chapter 8 (on‘enumerating’) we pursue something of whatthis arguably more familiar part of a researchproject (namely, quantification) can involve.But in this chapter we agree with the quota-tion that many human geographers have con-flated entitation and quantification (orenumeration), and have thereby regarded theattribution of numerical values to items ofdata as the primary stage in the interpretation –or, as they would be more likely to say, the

3151-Ch-07.qxd 3/3/2004 4:08 PM Page 217

Considerable confusion surrounds the introduction of thinking about ‘systems’ in thegeographical literature. One branch of this thinking, derived from the ‘general systemstheory’ of Ludwig von Bertalanffy (1968), has seen geographers seeking for fundamen-tal properties (usually spatial ones) which feature in many different but seeminglyanalogous worldly systems such as the hydrological system, the ecosystem, the regionalsystem and the settlement system (e.g. Chorley, 1964; Coffey, 1981). Less grandly, a vari-ety of ‘systems theory’ has sprung up in the discipline which conceptualizes the world interms of interlocking ‘systems’, wherein a system is taken simply as an identifiable por-tion of reality which ‘consists of a set of entities with specifications of the relationshipsbetween them and between them and their environment’ (Wilson, 1986: 476). Put in thismanner, a great many geographers could be cast as systems thinkers (and one argumentis that we are all in effect systems thinkers), but few explicitly utilize the specializednotions (to do with ‘inputs’, ‘outputs’, ‘feedbacks’ and the like) which have been devel-oped to describe the supposed workings of systems by the likes of Chapman (1977),Bennett and Chorley (1978) and Wilson (1981b). These notions are sometimes referredto as the tools of ‘systems analysis’ (Huggett, 1980), although in recent years the latterterm has been increasingly been restricted to a ‘mathematical approach to the model-ling of systems’ (Hay, 1994: 615). Numerous objections might be raised to systems think-ing in its many guises (e.g. Kennedy, 1979; Gregory, 1980), not least because there is atendency here to conflate the ‘real’ and the ‘system’: to suppose that the world really isa bundle of interacting systems, thereby forgetting that this is only a thoroughly humanpicturing of this world, which can lead to simplistic attempts at trying to ‘control’ theseimagined real systems. Moreover, by the lights of the current chapter, there are gravedangers associated with the basic impetus of systems thinking to ‘put things into boxes’:to suppose that it is a quite straightforward exercise to identify distinct entities whichcan be named, specified, measured and inserted into the ‘boxes and arrows’ depictionof a typical systems diagram (see the example illustrated in this box).

Box 7.2: Systems analysis


218

LABOURINPUTS

CLIMATICINPUTS

(Constant)

PRODUCTIONSUB-SYSTEM

(Closely linked tosocial organisation)

OUTPUT

SOIL−LANDINPUTS

(Variable)

DEMAND(Directly related topopulation numbers)

influences including regulations and constraints

ThroughputFlows

InteractionLinks

One wayMultiple

A simple subsistence farming system

Source: Carr (1987: 121, Figure 13.1)

3151-Ch-07.qxd 3/3/2004 4:08 PM Page 218

SIFTING AND SORTING

219

analysis – of geographical data.This is obviousfrom what are arguably the standard remarksmade by geographers about the sifting andsorting of data, as we briefly review in amoment.The problem here is that attributingnumerical values presupposes that theresearcher has decided in advance exactlywhat are the basic entities or perhaps patternsbetween entities warranting enumeration. It isa view which overlooks the very real concep-tual labour that goes into entitation, effectivelyregarding it as automatic or unimportant, oras somehow folded into the more dramaticmoment of fixing the numbers to the bits of‘stuff ’ under scrutiny.

These points about entitation do not onlyhold in relation to enumeration. Indeed, wewould insist that in any human geographicalresearch project, however quantitative or qual-itative in its practical execution (see Box 1.7),there must be a phase of entitation which pro-vides a preliminary ordering of the data uponwhich subsequent interpretative strategiesoperate. By the latter we could be talkingabout the utilization of a conventional statisti-cal test or a fancy GIS procedure on numeri-cal entities, but we could also be talking abouttextual or iconographic interrogations of non-numerical entities entailing words and images(a policy objective in a planning document,a statement in an interview transcript or aphotograph in an archival collection). Theprinciple is the same: the entities must havebeen extracted from the chaos, the mix, ofmaterials comprising what is sometimes calledthe ‘raw data’ (see Box 1.3).What we thereforedo find strange is the absence of general claimsabout entitation, and we suspect that this isbecause many researchers take it as too self-evident a stage in what is being attempted towarrant comment. Such a belief emerges fromconventional assumptions about the ‘nature’ ofthe data,of the real-world situations behind thedata,which are being subjected to a preliminaryordering. The data, and hence the realities,

are not seen as problematic; or, rather, theassumptions underlying how these ‘data of theworld’ are to be sifted and sorted – badgeredinto boxes and lists – are rarely examined at allcritically because they lie so fundamentally atthe heart of much taken-for-granted knowl-edge production and consumption in Westernacademia.

If this sounds like a big claim, it is meant tobe, for it genuinely strikes us that what isinvolved in entitation runs up againstextremely thorny philosophical dilemmasabout the positing of identities, binaries andhierarchies.We are not proposing that humangeographers can somehow resolve thesedilemmas – far from it – but we are recom-mending that we should become aware ofthem so as to be more circumspect and evenhumble about the grounds on which entita-tion occurs in any (and every) research pro-ject. In the course of developing this line ofargument, we will take a look at various mat-ters which are usually viewed in a rather dif-ferent way in the more philosophicalliteratures of human geography.

Standard remarks

In so far as geographers have thought muchabout the preliminary ordering of data, ratherthan taking them for granted, they havetended to do so in terms of what JamesLindsay (1997: 22–9) calls both the provenanceand the types of data available to them. Interms of provenance, he refers to the differingsources from which data have been derived,offering a distinction between ‘primary’ data(generated by the researcher), ‘secondary’(generated but not interpreted by somebodyelse) and ‘tertiary’ data (already interpretedand available for consultation in existing textsand journals). We have some problems withthe starkness of these distinctions, as should beapparent from our detailed examination ofprecisely these matters throughout this part of

3151-Ch-07.qxd 3/3/2004 4:08 PM Page 219


220

the book (see also Box 1.3), but we agree withLindsay that the preliminary ordering of datamust always be highly sensitive to the prove-nance of these data: to the when, the whereand the why of their construction. In a simplesense, this means that the researcher mustnever lose sight of which items of data havebeen gleaned from the field through his orher own labours, which have been trawledfrom the surveys and reports of others, andwhich have already come to him or herpreinterpreted from the library. What opera-tions can legitimately be performed on thesedata, what kinds of inferences and conclusionscan plausibly be drawn from them, dependintimately on such matters of provenance andconstruction. But these matters, havingalready been addressed previously in thebook, are not our main focus here.

When turning to types of data, Lindsaybegins by considering the oft-cited distinc-tion between qualitative and quantitative data,a distinction that we are seeking to blur invarious ways even if it continues to havemeaning for us (see Box 1.7 and Chapter 8).But Lindsay introduces this distinction in thecourse of turning to a more sustained exami-nation of data in terms of their potential to bequantified (or enumerated), and he therebyechoes what have perhaps been the most typicalobservations ever made by geographers aboutthe preliminary ordering of data (e.g. Silk,1979: 6–7; Gregory, S. 1992: 136). This exer-cise merely divides up data under the fourstandard headings of nominal, ordinal, intervaland ratio (see Box 7.3), and explains theextent to which it is possible to apply numer-ical values to, and then to use statistical testingprocedures on, the different types of dataidentified.While there are varieties of statisticswhich have been developed for use on datawhich are merely nominal or ordinal (andgeographers have contributed usefully to thedevelopment of ‘categorical data analysis’appropriate for such data; see Wrigley, 1985;

Robinson, 1998: ch. 6), the common ambitionis still to find ways of converting such datainto forms where they can be treated numer-ically and thereby subjected to more familiarforms of measurement and manipulation.Unsurprisingly, most credence has been givento data capable of measurement on intervaland ratio scales, which is thus data which canbe most easily made suitable for the numeri-cal work favoured by many geographers(everything from the most basic of descriptivestatistical testing, much of which depends onbeing able to calculate arithmetic means,through to the complexities of multivariatestatistical analyses and beyond). It was not solong ago that a methods course in geography,even one in human geography, would quicklydispense with any data that could not be sub-jected to this toolbag of statistical-mathematicaltechniques. As indicated, in Chapter 8 weconsider at greater length what is involved inthe various faces of ‘enumerating’.

At this point our concerns are rather dif-ferent, however, in that we wish to digbeneath what is really being presupposed insuch treatments of the sifting and sorting ofgeographical data. Returning to Lindsay for amoment longer, it is significant to note thathe couches his discussion of data types verymuch in terms of ‘measurement’. He claimsthat ‘all data are obtained by some sort ofmeasurement, although at its simplest themeasurement might amount to no more thanallocation to a class’ (1997: 25–6). By his lattercomment he means something like choosingto allocate a data item (e.g. a listing of some-one’s occupation in a historical document as‘carpenter’) to a broader class or categoryready for further interpretation or analysis(e.g. ‘woodworkers’). None the less, it isrevealing that he talks of measurement in thisregard, given Richard Huggett’s argumentabove that measurement equates with quan-tification (an impulse to enumerate phenom-ena) and that quantification is itself an act

3151-Ch-07.qxd 3/3/2004 4:08 PM Page 220

1 Nominal This implies data that fall into categories or groups, such as landforms,land-use types, occupational categories, diseases or nationalities. These cannot bemeasured, but their frequencies in given areas can be counted.

2 Ordinal In these cases the relative magnitude of the phenomena can be known,so that they can be ordered from largest to smallest, even though actual values arenot available. This is often because of difficulty in obtaining exact measurements,either due to instrumentation problems or because of limited access to moredetailed data.

3 Interval With interval data the actual magnitudes of the phenomena areavailable, such as slope angle, rainfall, production figures or population.

4 Ratio These include proportion or percentage measurements, such as votingpercentages, relative humidities or chemical concentrations, as well as a variety ofindices, such pH values. These ratio data may have finite upper or lower limits,which obviously affect the way they can be analysed statistically.

Source: Gregory, S. (1992: 136)

Box 7.3: The standard view of the basic data types

SIFTING AND SORTING

221

which necessarily comes after the prior act ofentitation. In other words, we would arguethat Lindsay’s examination of types of data(the main part of his chapter on ‘data handling’in human geography), basically misses out thestage of entitation which cannot but occur –even if taken for granted, even if done com-monsensically – before arriving at the stagewhere the researcher can sort out his or herdata into piles distinguished as nominal, ordi-nal, interval and ratio. And this is notwith-standing Lindsay’s own admission that ‘if dataare not naturally given objects, then we haveto recognise that they are chosen and shapedin a way which might have unconscious aswell as conscious elements’ (1997: 22).

Our contention is that, even for a piece ofdata about some phenomenon, event oraspect of the world to be designated as one ofthe seemingly simplest (nominal) type, a deci-sion must already have been taken to namethat phenomenon, event or aspect as onething rather than another (as a ‘carpenter’whocan also be classed as a ‘woodworker’ but notas a ‘textile worker’). In Huggett’s example,for instance, a decision has been taken toidentify a small pinprick of light in the nightsky as a star which might be given a name

(e.g. the North Star) and also classed as part ofone given constellation rather than another(e.g. the Big Dipper). All this has occurredprior to any assessment of what type of dataare involved, let alone of how they might bealtered to render them amenable to numericalmeasurement and manipulation. A furtherimplication of Huggett’s example is thatmedieval astronomers viewed their classifica-tion of a star to a constellation as a meaning-ful scientific act, whereas modern astronomerswould suggest that this is less meaningful thanidentifying a given star in terms of the preciseportion of deep space where it is located(which may be nowhere adjacent to any ofthe other stars in its constellation as it appearsto us on earth).The basic lesson is that differ-ent decisions may be taken in the naming,classifying and, in effect, entitation of a givenstar, and in the process all manner of differingassumptions may be made about both its‘nature’ and, interestingly, its time-space rela-tions.All this occurs in the preliminary order-ing of data pertaining to that star, beforequestions kick in about the type of datainvolved, and yet we probably take too littlenotice of what is happening in the process: wefail to register a whole series of moments in

3151-Ch-07.qxd 3/3/2004 4:08 PM Page 221


222

the sifting, sorting and entitation which thendoes have such an influence on any subsequentinterpretations of the star as provided byacademics.

Spatial referencing

The astronomy example also hints at anotherissue which we need to note before exploringthe deeper problems bound up in sifting, sort-ing and entitation – namely, that of the spatialreferencing of data.The medieval astronomersidentified stars according to their position inthe geography of the night sky and hencegave priority to constellations as orderingdevices, while modern astronomers think interms of the precise locations of stars relativeto those nearby in the particular sectors ofdeep space which they occupy. For geogra-phers the space in question tends to be earthlyrather than celestial, but the issue of spatialreferencing is central to the concerns of thediscipline, both in its historical developmentand still today (see Chapter 1). So much of thegeographer’s data directly describes spatialunits of the earth’s surface (e.g. specifying thedimensions of length, breadth and area pos-sessed by given units) or more indirectlydescribes aspects of particular spatial units(e.g. prevailing religious beliefs or stated vot-ing intentions of human populations living inthat unit). Much work by regional geogra-phers was precisely concerned with the taskof identifying meaningful areal units and ofdeciding which data sets most appropriatelyserve to represent the character of theregional entities so delimited (e.g. Hartshorne,1939), while spatial scientists expresslyequated the wider scientific task of ‘classifica-tion’ with the narrower geographical task of‘regionalization’ (of allocating individualpieces of data to given areal units while alsodeciding upon realistic typologies of regions).1

Even in the more unusual of more recentprojects in human geography, virtually all the

data consulted retain some element of spatialreferencing. Thus a humanistic geographermight be talking about what creates feelingsof security, well-being and belonging forpeople living in a given place (e.g. Ley, 1974;Relph, 1976), or a feminist geographer mightbe talking about differences in the ‘genderdivision of labour’ between one locality andthe next (e.g. McDowell and Massey, 1984).The latter geographers may be less concernedthan either regional geographers or spatialscientists about the exact configuration of theregions and spatial patterns involved, beingmore concerned with the human meaningsand political-economic dynamics of the situ-ations under study, but the spatial referencing(the embeddedness in place or the variabilitybetween localities) remains absolutely crucialto the inquiry.

For virtually all human geographers, there-fore, sifting and sorting their data continues toinclude some reference to space, place, loca-tion and environment. We realize that byraising these issues we are entering into aminefield of debate about what exactly geog-raphers understand, and should understand, bysuch terms and notions. Right at the heart ofany geographical study assumptions will bemade which, even if not reflected upon at allexplicitly, cannot avoid carrying with them anenormous freight of philosophical reasoningabout such terms and notions. Even the seem-ingly simplest of sifting and sorting operations –choosing to sort out data on wage rates bynamed standard planning regions or choosingto sift out data on people’s feelings by types ofenvironment (e.g. run-down inner city oridyllic lowland village) – presupposes a wholearray of understandings about what are rele-vant spatial references, about the likely inter-twining of spatial and social processes acrossgeographical scales, about the ‘absolute’ or the‘relative’ properties of space itself and so on.This is not the occasion to begin debatingsuch issues2, but we do wish to flag them as

3151-Ch-07.qxd 3/3/2004 4:08 PM Page 222

SIFTING AND SORTING

223

ones which press upon the practising ofhuman geography.

What we should also underline is the extentto which there is a long tradition within thediscipline of thinking quite technically aboutwhat one writer terms ‘geocoding’, which‘can be defined as the process by which anentity on the earth’s surface, a household, forexample, is given a label identifying its loca-tion with respect to some common point orframe of reference’ (Goodchild, 1984: 33). Atone level geocoding widens into the broaderissue of spatial co-ordinate systems, the speci-fication of (x, y) co-ordinates for a givenobject (the attribution of standard grid refer-ences or latitude and longitude, for example),which means that the discussion of geocodingpotentially covers the efforts of cartographers,surveyors, remote sensors and various other‘topographic scientists’. Indeed, it could beargued that cartography is a primary orderingdevice for geographers which should be dis-cussed further here (for an outline, seeBlakemore, 1994; Box 7.4), particularly sincethere was a time when many geographersdeclared that ‘if it cannot be mapped, then itis not geography’.We would not echo such adeclaration, although we do support thosewho are seeking for new, different and moreexperimental means of mapping phenomena(e.g. emotions and power relations) whichhave hitherto been regarded as too‘ephemeral’ (see Dorling, 1998). We wouldalso note criticisms of how a cartographicobsession dovetails with a prevalent ‘occular-centricism’ in the discipline’s history (seeChapter 1), and at the same time suggest thatmost procedures integral to standard map-making exercises inevitably run into thetricky problems of fixing (id)entities whichwe will examine shortly.

Much the same is true of the most recentinheritor of this technical tradition, geograph-ical information systems (GIS),wherein highlysophisticated computer-assisted techniques are

deployed to take ‘geographic information,[which] in its simplest form, is informationwhich relates to specific locations’ (Martin,1996: 1), and then to store this information ina database which both preserves the spatial ref-erencing and renders it amenable to furthermanipulation (in seeking to detect inter-relationships within the data). Many see GIS asthe future of the discipline (e.g.Thrall, 1985),and we would agree that GIS offers an impor-tant new tool for human geographers to uti-lize in the stages of collecting, storing andexploring large spatially referenced data sets.There are thus compelling reasons for why weshould treat GIS further (for an outline, seeGoodchild, 1994; Box 7.5), particularly giventhe key role of GIS in sifting and sorting data,but we are reluctant to accord it greaterprominence both here and elsewhere in thebook because serious criticisms can be voicedabout its conceptual limitations and practicalapplications.3 And, once again, GIS arguablyintensifies the tricky problems integral to thefixing of (id)entities.

What happens whenwe put things intoboxes and make lists

In effect, we have already been discussing theimplications of listing and boxing thingsduring the previous chapters. Indeed, we havealready been taking seriously what we are call-ing entitation, as when noting the differingways in which the raw data of UK unemploy-ment can be configured and reconfigured totell different ‘stories’ about the UK economy(see Chapter 2). Decisions taken in the entita-tion of unemployment, in giving it shape andpossible meaning, have duly been unpacked,and attention has been paid to the ‘particularunderstandings of reality’ possessed by differ-ent agents and bodies involved in the processesof economic data collection, manipulation

3151-Ch-07.qxd 3/3/2004 4:08 PM Page 223

Cartography can be defined as a ‘body of theoretical and practical knowledge thatmapmakers employ to construct maps as a distinct mode of visual representation’(Blakemore, 1994: 44). It is commonly assumed that one of the prime tasks of the geo-grapher is map-making, and at least one well-known history of the discipline(Livingstone, 1992) has a book cover depicting ‘the geographer as map-maker’. Manytextbook introductions to the discipline concentrate heavily on map-making and thuson the diverse and often quite complex knowledges and skills involved in what isdefined above as cartography. In an old textbook such as Bygott (1934), ‘mapwork’ and‘practical geography’ are taken as virtually synonymous, and the focus is basically onhow to translate ‘real earth’ (terrains, settlements, etc.) into ‘paper landscapes’ (repre-sentations of that earth, together with often less visible human constructions likepolitical and administrative boundaries). Attention is thereby given to everything fromsurveying to projections to map-reading to understanding different kinds of maps(at different scales, with different projections, made by different agencies). A morerecent textbook such as Lindsay (1997) still includes a whole chapter on ‘maps and map-ping techniques’ although here it is acknowledged that much map-making – particu-larly of high-quality maps – is done by specialists (by cartographers who would notnecessarily call themselves geographers). It is also recognized that the maps which geo-graphers themselves produce today tend to be representations of particular bodies ofspatially variable data – e.g. population profiles, industrial location, income level, qual-ity of life, etc. (Dorling and Fairbairn, 1997; Dorling, 1995b; 1998) – rather than ‘basemaps’ attempting to convey the ‘sense’ or ‘ground truth’ of overall physical and humanenvironments. These geographers’ maps of today are produced in the context of par-ticular research projects and they become tools in helping to answer given researchquestions and perhaps (when used more inductively) to frame new ones. In this regard,the technical quality of the cartography becomes of less importance than what themaps can demonstrate, how they can be interpreted and what new possible avenuesfor inquiry they may open up. Numerous critiques are now directed at standardassumptions about the value neutrality of cartography, the root argument being thatall maps are made for given purposes which determine exactly what is shown and howit is shown (Harley, 1988; 1989) and, in this connection, maps must be seen as official(or occasionally non-official) sources to be examined as part of given research projects.Thus, how maps are constructed in given situations, for instance by a Nazi geopoliticianor by a local planning department, becomes something that must itself be subjected toresearch (see Chapters 2 and 3). What is perhaps less reflected upon is how construct-ing maps, choosing what is or is not an object (entity) fit for (some kind of) spatial rep-resentation, is central to the sifting and sorting of ‘raw data’; and much cartography isthereby paradigmatic of the thorny issues to do with the attribution and fixing of iden-tities that are central to this chapter (see also Kitchin and Tate, 2000: 156–64).

Box 7.4: Cartography


224

and (re)presentation. These are the sorts ofunderstandings which we are urging humangeography researchers to critique, although itis important to distinguish between thoseassumptions which are, as it were, taken forgranted (because they are reckoned to be socommonsensical) and those which emergemore wittingly from the professional-political

interests of the agents and bodies involved inproducing, formatting and making availablesuch data. And such a distinction applies asmuch to researchers themselves when con-structing their own data as it does toresearchers interrogating the data constructedby others, since from the outset of a researchproject researchers will be making myriad

3151-Ch-07.qxd 3/3/2004 4:08 PM Page 224

SIFTING AND SORTING

225

decisions about data which, more or lessself-consciously, cannot but feed into the enti-tation of the phenomena, events and aspectsunder study (a key message, if not quitecouched in these terms, throughout Chapters 5and 6).

It is absolutely vital to explore the moreconscious decisions involved here, whethertaking into account the political orientationsof governments, the aesthetic preferences ofphotographers, the personal motivations ofresearchers themselves or whatever: all thesedimensions will play upon the sifting, sorting

and entitation of data. But the chief concernfor the rest of this chapter is with the com-monly much more hidden, indeed taken-for-granted, dimensions of these acts. As statedabove, therefore, this does lead us to tackle somequite deep-seated, ultimately philosophicalquestions to do with identities, binaries andhierarchies. Before tackling such questionsdirectly, though, let us begin with two briefcameos from the geographical literature – onediscussing an example of entitation in actionand the other entailing a particular line ofthought about such acts of entitation – from

Geographical information systems (GIS) can be defined in various ways, some accentingmore the technological elements, some the data transformations and some the appli-cations (Kitchin and Tate, 2000). At bottom, though, GIS entails a combination of com-puter hardware, software and ‘liveware’ (the researchers) which creates a ‘system forcapturing, storing, checking and displaying data which is spatially referenced to theearth’ (Chorley Committee, 1987, cited in Kitchin and Tate, 2000: 164). For many, theuse of GIS – sometimes talked about in terms of a broader sweep of geographical infor-mation technologies (GIT) – offers the way forward for the discipline as a whole, pro-viding an arsenal of technologies and data-handling skills, ones specifically applicableto spatially referenced data, which can then be the discipline’s distinctive contributionto a broad constituency of social and environmental disciplines. For others, as indicatedin the main text, there are many problems inherent in the kind of ‘geography’ that GISentails and in the uses to which GIS tends to be put (cf. Clark, 1998; Flowerdew, 1998).This is not the place to discuss the workings of GIS nor of specific GISes (different sys-tems) since such specialist texts are already numerous (e.g. Martin, 1996; Longley et al.,1998a; Kitchin and Tate, 2000: 164–87), but it is appropriate in the context of thischapter to note the following:

Two major traditions have developed in GIS for representing geographical distrib-utions. The raster approach divides the study area into an array of rectangularcells, and describes the content of each cell; while the vector approach describes acollection of discrete objects (points, lines or areas), and describes the location ofeach. In essence, the raster approach ‘tells what is at every place’ and the vectorapproach ‘tells where everything is’. (Goodchild, 1994: 219, emphasis in original)

As the latter comments make clear, GIS is indeed fundamentally about scrutinizing spa-tial data sets in such a manner as to ascertain exactly what ‘things’ are present withinthe graticules of a preset spatial grid or exactly where are specific ‘things’ (identifiedand identifiable objects or, perhaps better, shapes) relative to one another. Such aneffort of establishing identities and fixing them spatially, in many ways merely an elab-oration of the time-honoured map-maker’s craft, comprises the foundation of all GISwork. Again, then, and like cartography, the fashion in which GIS operates on raw datais indeed highly germane to the deeper, even philosophical, concerns of this chapter.

Box 7.5: GIS

3151-Ch-07.qxd 3/3/2004 4:08 PM Page 225


226

which we will draw out points to inform themore abstract reflections that follow.

Occupations in historical-geographical perspective

An important topic for research concerns thecharacter of occupations (the jobs of workpeople do), and human geographers haveoften asked about the differing occupationalprofiles associated with different places andregions in given parts of the world.The his-torically minded of such geographers havealso asked about how these local and regionalcompositions have changed through timeand, in particular, interest has centred on theemergence in eighteenth- and nineteenth-century Europe of specialist ‘industrial’ dis-tricts wherein large numbers of people shiftedinto occupations ranging from basically craft-based pursuits (e.g. furniture-makers) to beingfactory hands on early production lines (e.g.textile workers in the ‘dark satanic mills’).Thedata available today on occupations fromsomething like the British Census are cer-tainly not as straightforward as many mightsuppose, and there are difficulties withdeploying the 1990 Standard OccupationalClassification (Dorling, 1995b), but the datafrom periods prior to the 1841 census – whendetailed occupational recording began – aremore inconsistent again, patchy both sec-torally and spatially, and hence resistant to easyinterpretation. In a monograph explicitlyaddressing this matter, Paul Glennie (1990)discusses at length the problems involved withsuch data and, in so doing, explores the sortsof qualifications already voiced in Chapter 2about using official sources (in his casepre-census English population listings, taxreturns, parish registers, probate inventoriesand the like).

More specifically, Glennie stresses theproblem that the entities which we mightwish to extract from the data (the occupations

themselves) do not have the simple, coherentand stable qualities that the researcher wouldideally require.While it may be true that there‘are few occupational terms whose generalmeaning remains [entirely] elusive’ (Glennie,1990: 14), such that the basic designation of‘bricklayer’ can probably be understood byus today in a manner akin to that of aneighteenth-century citizen, it is also the casethat ‘even quite familiar designations encom-passed a range of labour relations and forms ofwork’ (1990: 14). In other words, a bricklayerin the eighteenth century might go about his(or her?) business rather differently from a brick-layer today, but so too might an eighteenth-century bricklayer in Lancashire be doingsomewhat different things set within a some-what different set of contractual relationsfrom a fellow bricklayer in Kent. Further-more, significant differences may existbetween how individuals themselves mightclassify their own occupation and how thismight be defined by ‘social superiors’ such asthe gentry, the clergy, session clerks or over-seers of the poor; and another considerationis that the occupations of women, even formalpaid occupations, remained largely invisible tosuch authorities.The occupations that ‘mate-rialize’ out of the historical data, therefore,reflect a host of assumptions, decisions andexclusions, ones made by both those in thepast who recorded occupations andresearchers in the present who adopt or adaptthese occupational classifications.

There is nothing overly surprising aboutthese claims, but what we wish to emphasizeis Glennie’s further line of argument aboutwhat is involved in talking about occupations,and in being prepared to identify particularnamed occupations with both clarity andcertainty:

[To] have an ‘occupation’ in its modernsense is a comparatively recent idea. ‘JohnSmith is a tailor’ in the twentieth centurycarries a rather narrower meaning than in,

3151-Ch-07.qxd 3/3/2004 4:08 PM Page 226

SIFTING AND SORTING

227

say, 1600 or 1800. It is not so much that theterm covers several possible relations of pro-duction … , rather what differs is the exclu-sivity of activity embodied in moderndefinitions … The counting of occupationsis important, but mere tabulation of occu-pations amounts to tacit acceptance that‘an occupation is an occupation is an occu-pation’, regardless of context. (Glennie,1990: 11, emphasis in original)

What is evidently in Glennie’s mind is the factthat few people in premodern England pos-sessed anything like one single occupationand, instead, tended to be involved in a rangeof different occupations, such that for some ofthe time they might be farming (either ontheir own small-holding or as a waged agri-cultural labourer) whereas at other times theymight be cotton-spinning in their attics orfurniture-making in their yards. These wouldbe the intersecting activities of the ‘proto-industrial household’ (Medick, 1976; Gregory,1982), ones involving all household members,including women and children whose work inthis respect rarely figures in the archives, andones which often slipped into the realms of aninformal economy which was again unlikelyto be fully recorded.The key issue is that thedetermination to force the data into a mouldwhich speaks of definite occupations, convey-ing the impression of clearly delimited andexclusive entities akin to how we tend tothink of such things today, is itself somethingthat must always be acknowledged.This is notto say that the researcher should not speak ofsuch occupations – indeed,Glennie is preparedto do so in constructing a general typologyof pre-census occupational classifications – butit is to insist that we can never completelyforget about how we are carving up the data,and in a sense the reality behind the data, priorto any effort of naming, enumeration andinterpretation. And his recognition of howwe in effect envisage singular occupationalidentities (person A is in occupation X; personB is in occupation Y), identities posited as

devoid of overlap or ambiguity, allows us toforeshadow the more philosophical questionsat issue here.

A geographical detourinto set theory

In a little-known book on ‘the mathematicsof geography’, Keith Selkirk (1982: 5) consid-ers the principles underlying much geographi-cal thinking and begins by suggesting that westart our studies – or should start our studies –through a careful sorting out of what mightbe termed raw data into sets:

A set is a collection of distinguishableobjects, for example the set of active volca-noes in South America or the set of inhab-ited islands in Indonesia. It is necessary tohave a criterion by which we know whetherany object is included in such a set or not. Inthe above examples we need a clear defini-tion of the word ‘active’, and to distinguishwhether ‘inhabited’ includes seasonal habi-tation or not.

There is a formal notation which attaches to‘set theory’, talking about ‘elements’ or‘members’, ‘universal sets’, ‘subsets’ and so on(and also developing a related mathematicalnotation; see Selkirk, 1982: 5), but the keypoint to underline is the assumption thatitems of data can be relatively straightfor-wardly allocated to sets in a process whicheffectively entitizes them as certain varietiesof phenomena rather than others. In terms ofoccupations, for instance, the aim is to allocateparticular individuals to particular occupa-tions in as clear a fashion as possible or toallocate particular specialist occupationalclassifications (e.g. bell-making) to more over-arching occupational sets (e.g. engineering) ina similarly clear fashion. While we may notconceive of what we are doing in this vocab-ulary, there can be no doubt that we are ineffect striving to specify readily identifiablesets (particular sets such as bell-makers or

3151-Ch-07.qxd 3/3/2004 4:08 PM Page 227


228

more general sets such as engineers), and thento allocate individual pieces of data to theappropriate sets. The properties of such setsare often visualized in terms of Venn dia-grams, themselves an intriguing translation ofcomplex conceptual processes into theabstract spaces of diagrams made up of linesand shapes comprising borders and bound-aries.Two examples of Venn diagrams convey-ing simple geographical information areprovided in Figure 7.1, and it might be arguedthat such diagrams offer a useful way ofachieving a preliminary ordering of geo-graphical data prior to any further work beingconducted on these data (see also Abler et al.,1971: 152–5).

It should be obvious, as can be seen fromthe Venn diagrams, that thinking in terms ofsets does not necessarily commit human geo-graphers to supposing that data can be allo-cated to only one set or another set and neversimultaneously to two or more sets. In fact, itis quite possible to conceive of the intersectionof two or more sets, with the intersectioncontaining those elements reckoned to beproperly members of more than one set (as inFigure 7.1). This means that in set terms wecan identify an individual person who issimultaneously a farmer and a spinner or toidentify an occupation (e.g. a railway buffetattendant) that could be allocated simultane-ously to different occupational classifications(e.g. a worker in both services and transport).None the less, at root the logic of set theorydoes conceptualize the situation very much interms of phenomena either being included ina set or excluded from it, and the further sug-gestion is that the principles of set theorybecome rather meaningless if the researcherstarts wanting to allocate certain phenomenato four, five or more sets (the implicationbeing that the analytical simplicity inherent inthinking in set terms then disappears, leavingus back with the chaos of reality). Selkirk alsoargues that conventional set theory does not

handle well the situation where all the elementsof a set can be assigned to one or other of twodisjoint subsets, the example which he cites(and which is perhaps not as obvious as hesuggests; see below) being that of the set of allhuman beings and its two disjoint subsets offemales and males. Selkirk refers to such asituation as dichotomous and notes that dichoto-mous divisions are the basis of many classifi-cation systems (a consideration to which wewill return).

Interestingly, he then makes the followingobservations with respect to Venn diagrams, aswell as introducing another possibility forhow we might be more flexible in our envis-aging of sets:

The Venn diagram has a number of disad-vantages … Firstly, Venn diagrams are noteasily drawn for more than three sets …,and secondly they emphasise the inclusion–exclusion aspect at the expense of the ideaof dichotomising a set into two parts. Analternative which seems little known togeographers is the Karnaugh map. (1982: 7)

The latter alternative, the Karnaugh map, con-fronts the problem that many phenomena can(and should) be allocated simultaneously tomore than one set, deploying a system whichpositions individual items (pieces of data) indifferent sets relating to segmentations of boththe horizontal and vertical axes of a grid (foran example, see Figure 7.2). What we havehere, then, is another way of translating con-ceptualizations into the abstract spaces of adiagram – one that, its rather ungainly rectan-gular quality notwithstanding, is actually moreflexible than is the conventional Venn diagram.The possibility of alternative visualizations –of ‘other’ ways to arrive at conceptual maps –is something we will revisit shortly, albeit in asomewhat different fashion.At this point let usmerely underline the extent to which set the-ory does tend to proceed in terms of an ‘inclu-sion–exclusion aspect’, referred to in otherliteratures as an ‘either/or logic’, and also the

3151-Ch-07.qxd 3/3/2004 4:08 PM Page 228

SIFTING AND SORTING

229

extent to which the conventional diagram-matic representations seemingly reinforce thisapproach to entitation. This is ultimately aphilosophical issue, and it is now to suchphilosophical reflections that we (have nochoice but to) turn.

Identities

The basis of what we are calling entitation (inany project) lies in the attributing of identity.It entails being open to the raw data of theworld as they come to the researcher, whetherthrough his or her own efforts or in a formmediated by others, and it highlights theprocess whereby the researcher decides moreor less consciously to identify the componentphenomena, events or aspects seeminglyexisting in (and constituting) that portion ofreality relevant to the study. It involves thecrucial link between the verb ‘to identify’ and

the noun ‘identity’, two similar words possessingsimilar etymological roots, with the act ofidentification establishing the identities of thethings reckoned to be present and worth dis-cussing.The usual process of identification, ofestablishing the identities of what are to bediscussed, is one which proceeds by delimit-ing a phenomenon as a singular, coherent andstable entity (and thus we also speak of ‘enti-tation’), perhaps giving it a proper name (e.g.Merrybank Farm) or depicting it as a partic-ular type of phenomenon (e.g. a sheep farm).Even when we are working with less tangiblephenomena such as statements in an inter-view transcript, our ambitions are still con-ventionally to delimit them in such a singular,coherent and stable manner. Thus we wouldwant to identify a given statement as a clearinstance of, for example,, somebody’s expres-sion of ‘topophilia’ (love of place),‘topophobia’(hatred of place),‘neighbourliness’,‘alienation’

Austria

Austria

etc.

etc.

Spain

Spain

Switzerland Bulgaria

Iceland

Iceland

Norway

Norway

Sweden

Sweden

Finland

Finland

Denmark

Denmark

UK

UK

Eire

Eire

Greece

NetherlandsBelgium

Belgium

W Germany

France

France

Italy

Italy

Luxembourg

Poland

Poland

BA

A

Figure 7.1 Two Venn diagrams used by Selkirk: one showing the set of Scandinaviannations as a subset of the set of European nations, and the other showing the overlap-ping sets of Scandinavian and (the then) EEC nations

Source: Selkirk (1982: 6, Figures 1.1 and 1.2)

3151-Ch-07.qxd 3/3/2004 4:08 PM Page 229


230

or whatever.4 When working with more tan-gible phenomena, the goal is normally toattribute not just a name but also a number,and thereby to give numerical values to anamed phenomenon (e.g. Merrybank Farmhas a land-holding of 75 acres) or to adepicted type of phenomenon (e.g. sheepfarms are ones with over 70% of their incomederived from sheep; sheep farms in the districtcomprise 60% of all farms; or sheep farms inthe district employ 50% of local maleemployees). In order for enumeration (orquantification) to occur, it is normally essen-tial for singular, coherent and stable identitiesto have been etablished, since numbers can

only really be fixed at all meaningfully toidentities boasting these properties.

Thinking generally, though, we would sug-gest that in any given research project manyentities are being ascribed with such well-defined identities to facilitate further inter-pretation of them and of how they apparentlyrelate to one another.The upshot is a concep-tual ‘mapping’ of the phenomena concerned,creating what might be termed an ‘entitymap’ which effectively positions these pheno-mena in a stabilized ‘space of identities’, and –while this may sound like an odd way ofputting things – such a map is being gener-ated (it cannot help but be generated) in

Netherlands

R

S

P

Q

EireFrance

Belgium

W Germany

Luxembourg

Finland Iceland

Denmark

Norway

Sweden

Switzerland

E Germany

Hungary

Czechoslovakia

Spain

PortugalItaly

Austria

Poland, etc.

UK

Figure 7.2 A Karnaugh map showing European nations divided on four categories:P = per capita income exceeds $1000 per annum; Q = over 300 daily newspapers per1000 population; R = over 68 years life expectancy at birth for males; S = over 15 milesof railway per 100 square miles

Source: Selkirk (1982: 8, Figure 1.8)

3151-Ch-07.qxd 3/3/2004 4:08 PM Page 230

SIFTING AND SORTING

231

every conceivable research project. It is theoutcome of the sifting and sorting which weare talking about here and its production isthe unspoken objective of this crucial butunacknowledged activity.

To begin thinking about the notion ofidentity, and to begin criticizing how it is nor-mally treated by geographers and other acad-emics, we propose to take as our guide thehuman geographer Gunnar Olsson. This issomeone who is often regarded as a rathermaverick figure within the discipline, some-one who has forsaken the conventionalexpectations of how to proceed (indeed, ofhow to practise) as a human geographer, butwe would counter that in text such as Birds inEgg/Eggs in Bird (1980; see also Philo, 1984)he quite brilliantly confronts the issue ofidentity and its problems. Here is one set ofclaims which he makes about identity:

The problem of identity is to recognisesomething as the same again. In conven-tional reasoning the equality sign is in thedefinition that a and b are identical if a andb have exactly the same properties. Identicalwords refer to identical things. The samelabelling word must always refer to exactlythe same phenomenon. What is called onething is never called another. The advantageis that others will know what I hold as iden-tical because they can see what I point to.The disadvantage is that, while categoriesare frozen, the world is not. (1980: 65b,emphasis in original)

This passage lays out the common supposi-tion that the identity of a phenomenonresides in, or at least depends upon, its singu-larity (each thing being clearly delimited as‘one’, and then as different from any otherthing), its coherence (each thing having givenproperties) and its stability (each thingremaining the same from one moment to thenext). Another coupling of key words there-fore surfaces: that of ‘identical’ and ‘identity’.The form of ‘entity map’ implied is one pos-sessing entities which are clearly identified intheir own pristine individuality, perhaps

named as A or B, and they are pictured aswholly separate with stark boundaries run-ning between them which fix certain ‘stuff ’ inone part of the map (e.g. in A) and other‘stuff ’ in another part (B). The connectionback to the Venn diagrams mentioned aboveshould be obvious, and we can reference liter-atures which carry a wider examination ofsuch boundary-based ‘topologics’ and theirimplications (e.g. Reichert, 1992). Mentionmight also be made here of what MarcusDoel (1999) critiques as the ‘pointillism’ ofconventional geography, particularly quantita-tive geography, residing in the discipline’sobsession with ‘One’: with the wish always tobe able to specify, delimit and count on thesimplicity of the one identifiable, set-off andenduring entity, event, process or whatever asclearly distinct from any and every other onein the world (Doel, 2001).

Despite this critical tone, it must beacknowledged that Olsson appreciates thebenefits of such a straightforward approach toidentity in that it is the basis for people tocommunicate about phenomena, to shareknowledge about what they agree – if tacitly– to regard as the same phenomenon with adefinite identity. He also suggests that the ulti-mate grounding for agreeing on identity con-sists of people physically pointing at the sameentity, in which case he hints at a groundingwhich is given less by pure logic and more bythe contingent, embodied act of pointing(a suggestion with resonances is explored furtherin Olsson,1988;1991a).‘Trust’ between peopleregarding the attribution of identity is therebytaken to be more about pointing, showingand demonstrating (see also Olsson, 1995)than about the ethereal logical deliberationsof philosophy, an observation that hints atthe variability of the ‘real world’ whichrequires more the gestures of everyday lifethan the abstractions of philosophy to createthe preconditions for successful humancommunication.

3151-Ch-07.qxd 3/3/2004 4:08 PM Page 231


232

This is probably to run into deeper watersthan we need for our book, but it does flowfrom the critical claim that Olsson makesabout the problem with the conventionalapproach to identity: that arising because,‘while categories are frozen, the world is not’.Thus, to return to the earlier example, anyattempt at imposing a single occupationalclassification on a citizen of eighteenth-century Britain,whether by an authority at thetime or by the researcher in the present, is torisk steamrollering over the fact that he or shemay really have done a number of differentjobs. Moreover, it risks obscuring the possibil-ity that what the citizen did as (say) an ‘inn-keeper’ could vary considerably from whatsomeone else given the same title might do inanother part of the country. But even some-thing as simple as identifying Merrybank as a‘sheep farm’ might not be so straightforward:in an earlier decade the farm might have con-centrated more on arable farming; in the pre-sent it might devote half its attention to sheepand half to a diversified activity such as run-ning a farm zoo; over half its employees mightwork full time in the farm zoo but over halfits income might still derive from the sale ofsheep and wool; and many other complica-tions can easily be envisaged. The identitiespostulated by the categorizations of ‘inn-keeper’ and ‘sheep farm’ are indeed frozen,implying a simple, obvious and unvaryingcharacter (in both time and space) which thethings of the world almost never possess. Inother words, while the world beyond acade-mia is a thoroughly messy place, one whichmost of us would acknowledge from everydayexperience as far from being marked by anysingularity, coherence and stability, academicsspend much of their time striving to imposejust such a definitive quality on it with theirneat namings, depictions and ‘mappings’ ofthe entities which they purport to find there.The sifting and sorting thus end up generat-ing a picture of the world which is too tidy,

an artificially ordered approximation whichcannot fully ‘mirror’ or ‘correspond to’ thephenomena, events and aspects which arethere, present and happening, in everydayreality. The complaint that our models andtheories are too neat and tidy, and hence toosimplified, compared to the messiness of theworld is actually quite commonplace (they arethe life-blood of ‘postmodernism’; see Box 7.6;see also Cloke et al., 1991: ch. 6), but Olsson’sphilosophical point is that such a neateningand tidying-up is integral to our routineassumptions about the identities of phenom-ena: it is indeed endemic to the very founda-tions of Western thought, for ‘[b]anned asoutcasts are … fuzziness and indeterminacy’(Olsson, 1980: 62b).

Another ‘outsider’ geographer, DavidSibley (1981b; 1998), has long complainedabout the obsession with order, or at leastwith a certain orthodoxly geometric concep-tion of order, which has marked the efforts ofmany human geographers over the years(most notably spatial scientists). Drawing onan ‘anarchist’ critique of conventional scientificreasoning, he objects to approaches whichinsist on trying to fit awkwardly shapedchunks of reality into neat boxes, a processrequiring the edges of the real ‘stuff ’ to beplaned off and the remains shoehorned intothe simple (often symmetrical, sometimes ele-gant) ‘grids and networks’ of standard intellec-tual imaginings. Arguing along very similartracks, a number of human geographersinspired by ‘poststructuralism’ (see Box 7.7)have also begun to question the usual pro-cedures and outcomes of how the disciplinehas gone about its sifting and sorting of data.In the course of an innovative paper record-ing a staged conversation between a spatialscientist (here called a spatial analyst) and apoststructuralist, Deborah Dixon and JohnPaul Jones (1998a; see also Natter and Jones,1993; 1997) outline what they refer to as the‘epistemology of the grid’ which underpins

3151-Ch-07.qxd 3/3/2004 4:08 PM Page 232

‘Postmodernism’ has quickly become a much maligned and misunderstood umbrellaterm for what should probably be best regarded as a stance taken by many academics,geographers included, to fundamental questions about what the human world is(ontological questions) and how ‘we’ can acquire knowledge about it (espistemologi-cal questions). It cannot be regarded as a coherent theory, model or approach, despiteoccasionally being presented as such. While many writers explore the character of whatthey describe as the ‘postmodern’ epoch in which we presently live, one marked by thecomplexity, hyperactivity, hybridity and mutability of economies, polities, societies andcultures (e.g. Harvey, 1989a), the stance of ‘postmodernist’ thought is one that effec-tively mirrors – and argues that our thought should mirror – this real ‘chaosmos’ byrefusing conceptual certainty: by refusing any simple ‘grand theories’ or ‘meta-narratives’(big stories) about the human world and its workings, whether these be supposedlyscientific laws, humanistic emphases on human agency and spirit, Marxist accounts ofcapitalist determinations and socioeconomic structure, psychoanalytic excavations ofrepressed desires, or whatever.5 The watchword of postmodernism as a way of under-standing is hence difference: an alertness to every line or scrap of difference which pre-vents things, peoples, situations, etc., ever being truly the same as one another, andwhich thereby insists on taking seriously the ‘jagged edges’ of particularity that mostother intellectual stances before the human world wish to smooth away. This is only togive the briefest sketch of postmodernist thought and, in practice, there are nowcountless debates about its genealogies, claims and (de)merits (e.g. Benko andStrohmayer, 1997), but it is worth providing such as sketch here because in manyrespects the very basis for this chapter – with its querying of conventional approachesto the problem of identity obsessed with finding sameness (between things) ratherthan celebrating differences – lies squarely in the postmodernist critique.

Box 7.6: Postmodernism

SIFTING AND SORTING

233

the research of the spatial analyst (and whichthey suggest is most formally propounded ina paper such as Berry, 1964; see Figure 7.3):

By our estimation, your [approach] isgrounded in a particular way of knowing,one that depends upon, and propounds, the‘epistemology of the grid’. The grid is at oncea procedure for locating and segmenting acomplex, relational and dynamic social real-ity. It works to create a systematical horizon-tality that stabilises both objects and theconcepts associated with them – i, j – suchthat both can be rigorously investigatedusing linear cause-and-effect systems of logic[see also Chapter 9]. The grid segments sociallife so that it may be captured, measured andinterrogated. (Dixon and Jones, 1998a: 251)

Moreover, Dixon and Jones stress that ‘thegrid[’s] … powers of segmentation fashionborders and supervise inter-relations amongobjects and events in space and time’ (1998a:251).

Writing chiefly against the assumptionsbuilt into spatial science but also against thestabilities of identity postulated by ‘structural-ism’ (see Box 7.7), Dixon and Jones go on tohighlight the importance of paying sustainedattention to context: to the specific circum-stances framing a particular phenomenon,event or aspect, as anchored in the peculiari-ties of a given time and space. Because the‘stuff ’ of the world varies so much from onecontext to the next, from one portion of time–space to the next, researchers need constantlyto be revising the names, categories and con-cepts which they deploy to isolate, identifyand then interpret this ‘stuff ’: their orderingsneed to shift accordingly; their sifting, sortingand entitation cannot remain frozen. Theymust be ever vigilant to the possibilitythat what was meant by ‘costermonger’ ineighteenth-century London was not identical

3151-Ch-07.qxd 3/3/2004 4:08 PM Page 233


234

to what it meant in eighteenth-centuryLiverpool, for instance, or that what wasmeant in the eighteenth century by the verynotion of ‘occupation’ is not the same as whatwe tend to mean by it today.

Dixon and Jones thus conclude that‘context stabilises meaning, but never perma-nently’ (1998a: 256), and they advocate a‘deconstructive’ stance (see Box 7.8) alert tohow the identities of phenomena (along withassociated terms, concepts and disciplines)which have been established with reference toone particular context constantly need to bereassessed with reference to other contexts. Itis a stance whose imperatives are at once ‘tolineate and de-lineate, to fix and unfix’(Dixon and Jones, 1998a: 255); or, in a voca-bulary more closely bound up with feminist

geography, the aim is never to let anything‘settle’ (Rose, 1991; 1993b).8 Allowing suchunfixity into research, permitting identities tobe unsettled, introduces an openness, a fluid-ity, into our thinking which is very much hos-tile to the ‘epistemology of the grid’. It runsagainst the normal assumptions and proceduresof practising human geography, suggestingthat the entitation part of a research projectmust always look to the context for guidance,to the temporal and spatial specificity of thesubject-matter under study, rather than to anypreset grids of expectation.

Binaries and hierarchies

In the course of his reflections on identity,Olsson makes a number of further points

CHARACTERISTICS Column j

Row i

Cell ij

Box or submatrix

Places

Figure 7.3 Brian Berry’s depiction of ‘the geographical matrix’. ‘A row of this matrixpresents the place-to-place variation of some characteristic or a spatial pattern of somevariable which can be mapped. Each column contains a “geographic fact”: the valueassumed by some characteristic at some place’ (1964). Berry also introduced time intothe picture by imagining an array of geographical matrices in temporal sequence.

Source: Berry (1964: 6, Figure 1)

3151-Ch-07.qxd 3/3/2004 4:08 PM Page 234

‘Structuralism’ is an intellectual movement emerging from de Saussure’s linguistics,Piaget’s psychology and Lévi-Strauss’s anthropology, and at root its insistence is thateverything occurring in the human world – including the seemingly most ‘human’ ofphenomena such as ‘our’ language and communication, cognition and comprehension,myth and taboo – is ultimately determined in its form and function not by individualhuman beings making conscious decisions which reflect their own (or any special fea-tures of) humanity but by anonymous ‘structures’ (of language, cognition, myth, etc.)which effectively lie beyond our individual knowing and control. These structures, con-ceived as relations between different entities identifiable within language, cognition,myth, etc., are then said to ‘structure’, to impose limits and possibilities, on what isactually said, understood, done, etc., by particular people in particular instances. Suchactual happenings are hence regarded as surface manifestations of deeper structures,and the researcher’s task as a structuralist is therefore to ascertain the contours of theseunderlying and not immediately visible structures. While the linguistic, psychologicaland anthropological versions of structuralism have occasionally influenced geographers(e.g. Tuan, 1972), the main route whereby structuralism entered the discipline wasthrough an engagement with Marxism during the 1970s. At this time various notionsto do with the determining influence of capitalism, a ‘structural realm’ comprised bythe dynamics of capitalist accumulation, crisis and contradiction, were imported in amanner which many duly termed ‘structuralist’.6

‘Poststructuralism’, closely related to postmodernism (see Box 7.6), continues thestructuralist insight that the human world is ‘made’ by structures, where these are con-ceptualized as relations of all sorts in which human beings are caught up (and throughwhich ‘we’ are all constituted as particular sorts of human beings with particular attrib-utes, aptitudes, attitudes and so on). Yet whereas structuralism tended to suppose thatthere are ultimately definite entities which can be identified and between which pos-sible relations can be established, poststructuralists are much more sceptical aboutthere being any such ‘pre-existing’ entities which are not in turn themselves merelyconfigurations of different relations (economic, political, social, cultural, discursive,psychical, spiritual, etc., all as cross-cut by axes of human difference such as class, gender,race, ethnicity, sexuality, (dis)ability, etc.). Any stabilities with which academics have tra-ditionally dealt, even those named in the rigorous structuralist texts of de Saussure,Piaget, Lévi-Strauss and others, are here thrown into question. This is not necessarily toreach the point where there is said to be nothing but chaos in the human world (as inextreme postmodernist positions), but it is to accept that everything under study has tobe subjected to the most uncompromising of relational analysis (see also Massey,1999b; Thrift, 1999), in the course of which attention must always be directed at thespecific time–space contexts in which the relations of relevance – whatever preciselythese may entail – become ‘grounded’, ‘earthed’, ‘real’ in the whirls and twists of every-day human geographies. This, then, is the tricky conceptual terrain from which Dixonand Jones (1998a) develop their powerful attack on the ‘epistemology of the grid’which is a central to the arguments of this chapter about identities and their (un)fixing(as integral to the sifting and sorting of raw data).7

Box 7.7: Structuralism and poststructuralism

SIFTING AND SORTING

235

which are germane to our discussion of siftingand sorting. In the course of talking about‘laws of identity’ as linked to categorization,he writes as follows:

These are themselves embedded in thelaw of the excluded middle. Everything isidentical to itself and nothing is identicalto anything else. Nothing is itself and notitself at the same time. And yet inherent

3151-Ch-07.qxd 3/3/2004 4:08 PM Page 235

Sometimes regarded as the methodology of postmodernism (e.g. Dear, 1986; 1988),‘deconstruction’ is first and foremost a ‘technique’ for analysing texts – primarily philo-sophical and literary works – and as such it was introduced by Derrida in his 1967 textOf Grammatology. As Poole (1988: 206) explains:

Derrida showed that, by taking the unspoken or unformulated propositions of atext literally, by showing the gaps and the supplements, the subtle internal self-contradictions, the text can be shown to be saying something quite other thanwhat it appears to be saying. In fact, in a certain sense, the text can be shown notto be ‘saying anything’ at all, but many different things, some of which subtlysubvert the conscious intentions of the writer.

This technique for teasing out the incoherencies, limits and unintentioned effects of atext – an exercise that scuppers any simple notion of being able to discern the ‘true’meaning of a text (a notion that privileges the author over the reader) – has had agreat influence, notably upon Anglo-American literary theorists. In this guise decon-struction commonly works by ‘tracking the traces of oppositional elements’ within atext, and there is a ‘two-step process’ through which ‘one first reverses and then dis-places the oppositions (eg. male/female, culture/nature or subject/object)’ (Pratt, 1994:468). This reference to isolating and then disrupting binaries (oppositions) is very per-tinent to claims further on in this chapter. What has also taken place is that the principleof deconstruction has been extended to the analysis of the contradictions lying in thehuman world beyond written texts: in all manner of cultural products, social settingsand power struggles. As a result, many researchers, human geographers included, nowtalk about ‘deconstructing this, that or the other’ when they are not so much pursuinga line of questioning directly informed by Derrida as simply highlighting aspects of agiven theory or real-world situation that do not appear to ‘add up’ or ‘fit together’ allthat well. See also notes on ‘the deconstructionist’ in Chapter 10.

Box 7.8: Deconstruction


236

in this stringent definition is neverthelessan element of dialectics [see below], forit is only by learning what somethingis not that we know what it is. Builtinto the positive is its own negation.(1980: 62b)

In a related statement he declares that in ‘con-ventional thought, all categories stem fromthe principle of the excluded middle, whichsays that nothing can be one thing and itsopposite at the same time’ (1980: 36b). Herehe is underlining the assumed singularity ofidentity, the straightforward oneness and non-contradictoriness of the phenomena beingidentified, but in so doing he raises a second

feature of conventional Western reasoning.This is one which tends to see things in termsof binary oppositions between supposedlydichotomous or mutually exclusive ‘pureessences’,not as condensations of many differentphenomena, events and aspects fusing togetherin an overlapping fashion. This is not just toreiterate the finding that conventional reason-ing prefers to hold things apart, assigningthem to different and separate positions in the‘entity map’ (see above); it is also to indicateits tendency to fix identities by settingup opposites which in a sense define oneanother. Olsson is perfectly aware of this ten-dency, elsewhere termed the role of the

3151-Ch-07.qxd 3/3/2004 4:08 PM Page 236

SIFTING AND SORTING

237

‘constitutive outside’ (Rose, 1993b; 1995),which arises when what are identified as thecharacteristics ‘inside’ a given phenomenon(e.g. health, sanity, normal behaviour) aredefined by those deemed to be ‘outside’ it(e.g. sickness, insanity, deviant behaviour).Although it is not necessarily the case thatthese binary oppositions will prioritize one ofthe two terms or identities over the other, it isoften the case in practice that they do so.‘[B]uilt into the positive [of a term or iden-tity] is its own negation’, states Olsson (1980:62b), which means that written into the heartof what is taken as good about a phenome-non, what is valued about it, are suppositionsabout the bad, devalued, unwanted and evendespised elements reckoned to be typical of itsopposite.

Arriving by a different route from Olsson,binaries have become a central focus of recentdebates in feminist geography, which echoOlsson’s earlier work while extending it insignificant new directions. Indeed, a rich veinof substantive research by feminist geogra-phers from Linda McDowell (1983) onwardshas examined a variety of pervasive worldlypractices, notably spatial practices, whichreflect a simple binary logic separating offpublic and formal ‘realms’ from private andinformal ‘realms’. As the WGSG authors statenear the start of a chapter addressed to thisissue:

Binary categories often suggest the exis-tence of discrete spaces, that is spaces asso-ciated with different types of activities …[They] often depend on the drawing ofsharp lines between the two halves withinthe binary category. These lines are likeboundaries, or fences; they are put upbetween the two sides which comprise thebinary category. So, for example, thepublic/private distinction works by estab-lishing a boundary between the publicand private sphere, and by drawing aline between public and private space.(1997: 112)

In effect echoing Olsson’s claims, or those ofSibley or Dixon and Jones, about how theconceptual fixing of identity goes hand inglove with ‘conservative’ practices defensive ofthe existing social order, these feminist geog-raphers show how a deeply worrying transla-tion occurs from the conceptual ‘binarizing’of women and men into the creation ofeveryday gendered geographies. As is wellknown, women tend to be portrayed as ‘natu-rally’ one kind of entity (caring, emotional,intuitive) and men as another (calculating,rational, cerebral), which relates to womenbeing expected to perform certain kinds ofrole (nurturing mother, carer, counsellor) andmen others (providing father, wage-earner,disciplinarian). Notwithstanding the countlessscramblings of these gendered distinctions inordinary life, there is still a clear link into pre-vailing assumptions about it being men whopopulate the public and formal spaces ofsociety (conducting the important affairs ofeconomy and state) while women remain inthe private and informal spaces (managing theroutine tasks of child-minding, home-makingand house-keeping). And such assumptionsstill have an influence on what actually hap-pens in the grounded geographies of who(out of women and men) is found where anddoing what. As a result, ‘[f]eminist geogra-phers continue to critique the gendering ofdichotomous categories used to create spaceand places, the boundaries which these set up,and the ways in which boundaries work todefine some people (and practices) as being“in bounds” while others are located “outof bounds’’’ (WGSG, 1997: 114; see alsoCresswell, 1996). In so doing, these authorsinsist that researchers should endeavour toensure that the basic ingredients stirred in atthe outset of a project, those through whichsense begins to be made of the raw data, arenot ones which unwittingly reinforce stereo-typical gendered binaries. More positively,

3151-Ch-07.qxd 3/3/2004 4:08 PM Page 237


238

they propose that researchers should bepurposefully ‘illustrating that the boundariesdrawn between binary categories are fre-quently more blurred than the dichotomyconstructs them as, very often becausewomen and men struggle to resist them’(1997: 115, emphasis in original).

Yet there is more to say here, since there isa line of feminist critique which emphasizesthe centrality of binary opposites to conven-tional Western thought and insists that agreat many such opposites, even if it is notimmediately obvious, are heavily freightedwith gendered meanings and implications.The WGSG authors provide a clear summaryof what they refer to as the problem of‘dualisms’:

A dualism is a particular structure of meaningin which one element is defined only in rela-tion to another or others. Dualisms thus usu-ally involve pairs, binaries and dichotomies,but not all pairs, binaries and dichotomiesare dualisms. What makes dualisms distinc-tive is that one of the terms provides a ‘core’,and it is in contrast to the core that the otherterm or terms are defined. Thus dualismsstructure meaning as a relation between acore term A and subordinate term(s) not-A …The reason some feminist geographers havepaid attention to dualistic ways of construct-ing knowledge is that dualisms are very oftengendered and hierarchised, so that the coreterm A is masculinised and prioritised, andthe subordinate term(s) not-A are feminised.(1997: 84–5)

Intriguingly, in his account of set theory ingeography, Selkirk commented on the preva-lence and importance of dichotomous thinkingand proposed that the most obvious exampleof such a dichotomy is that between womenand men (see above). Feminists agree that thisparticular dichotomy has been prominentwithin Western thought, and they argue thatit has functioned as something of a ‘master’dualism whose effects have rippled out toprovide an underlay of expectations and even

prejudices present in the dualistic foundationsof both academic and common-sense knowl-edges. In the process they have exposed themanner in which these knowledges becomestructured by oppositions between attributescoded as masculine (and in the processdeemed as important, good and valued) andattributes coded as feminine (and in the processdowngraded as trivial, relatively bad and notworthy of being valued). Gillian Rose (1993b:esp. 66–77) traces such oppositions in somedetail, noting their installation within theengine-room of Western thought by malescholars effectively casting intellectual activityin their own self-satisifed and self-congratulatoryself-image, and she offers a vehement critiquewhich aims to dismantle these oppositions inthe course of laying out new possibilities for anon-dualistic human geography whose tactic‘is suggested by the notion of other fields ofknowledge beyond A/not-A’ (Rose, 1993b:85; see Box 7.9 for materials covered by Rosein her exploration of gendered dualisms).

Probably the most controversial proposal toemerge in this respect would involve question-ing what Selkirk (and indeed virtually all therest of us) do take as so self-evident: namely, thedichotomy between women and men. Instead,the WGSG authors speculate that

The notion that we all fit into either a maleor female body is just that, notional. It can-not be sustained, either over time or space.Simply put, there is nothing ‘natural’ about‘male’ and ‘female’ bodies. There is nothingnatural about everyone being forced intoone sex or the other. (1997: 195)

However disturbing this claim may be tomany readers, there is mounting evidence tosuggest that the obviousness of biological sexcan no longer be taken for granted, and thatboth sex and gender are much more entan-gled, much less dichotomous, than is usuallyassumed.There are many versions of woman,

3151-Ch-07.qxd 3/3/2004 4:08 PM Page 238

The following are Gillian Rose’s specification of gendered dualisms, including part of apoem which she borrows from Helene Cixous (where the first term in each line is codedas ‘male’ and the second term as ‘female’) and four other listings (with the same cod-ings of first and second terms). Rose’s argument is that in each coupling the ‘masculine’term is conventionally regarded as a ‘superior’ personal attribute, academic quality,social condition, cultural space or whatever:

Box 7.9: Gendered dualisms

SIFTING AND SORTING

239

female and feminine, and there are many ver-sions of man, male and masculine: all of us arealmost certainly complex hybrids of bodilyfeatures, social behaviours and cultural attrib-utes which are conventionally allotted to thenon-overlapping boxes marked ‘woman’ and‘man’, ‘male’ and ‘female’ or ‘masculine’ and‘feminine’, but which in truth are always mix-ing together in unpredictable ways withinindividual bodies, minds and personalities. Forfeminists and feminist geographers,destabilizingthese binary opposites – or even the thoughtexperiment which contemplates destabilizingthem – is a vital manoeuvre in challenging thetyranny of dualistic thought and in throwing

up new possibilities for thinking about thesifting and sorting of identity. The implica-tions for the more mundane practices whichare ostensibly the focus of this chapter maynot be all that apparent, but if such a challengewere to strike a decisive blow against the stan-dard binary ‘inclusion–exclusion’ or ‘either–or’ logics of how to impose order on raw data,the impacts for all sifters and sorters would beimmense. And there are one or two immedi-ate implications as well, which will be coveredshortly when we draw this chapter to a close.

Before doing this, though, let us concludethis section by briefly mentioning how thespecifying of binary opposites in which certain

Activity/passivity,Sun/Moon,Culture/Nature,Day/Night,

Father/Mother,Head/heart,Intelligible/sensitive,Logos/Pathos

culture – naturecity – countrysidespace – natureculture – primitive

theory – empiricsgeneral – specificabstract – concretenomological – contextualizing

modern – postmoderndeep – shallowseminal – playfulgreat – fecundthrusting – titillatingpenetrating – veiled

time- – humanisticgeography geography

public – privatetransparent – opaquesocial – bodyknowledge – maternityrational – emotionalspace – place

Source: Rose (1993b: 67, 74, 75)

3151-Ch-07.qxd 3/3/2004 4:08 PM Page 239


240

things are valued over others often becomeselaborated in the further ordering device ofthe hierarchy.The WGSG authors discuss ‘theimplicit hierarchies’ (1997: 114) which are setup in the process of valuing some phenomenawhile devaluing others, and in so doing theyhint at the underlying principles of hierar-chies wherein entities are more or less explic-itly listed or ranked from those which arereckoned to be most important or fundamen-tal through to those which are reckoned to beunimportant and trivial. (And there is a clearconnection here to the so-called ‘depthmodels’ of explanation which we review inChapter 9.) The assumptions which feed intosuch ranked listings should be reflected uponat some length since what might seem like aninnocent act of getting raw data into a neatorder prior to the real work of interpretationwill actually have a decisive impact uponwhat interpretations are finally provided(constructed). To give a stark example, aMarxist geographer coding up an interviewtranscript may look for statements which canbe named as to do with ‘class’, and will moreor less deliberately put all these togetherunder this heading at the top of his or her listof things to take into account. Other state-ments which might be coded under otherheadings may also be noted, for instance ‘gen-der’, but will then be placed rather lowerdown the list. A feminist geographer lookingat the same transcript might arrive at similarcodings but will almost certainly reverse theorder so that statements under the heading of‘gender’ are listed higher in his or her inter-pretative hierarchy than are those under theheading of ‘class’.9

Geographical scale is often used as a basisfor hierarchical order in the sifting and sort-ing stage of a project. Factors reckoned to beof relevance from an inspection of the rawdata are allotted to levels identified as ‘global’,‘national’, ‘regional’ and ‘local’. There is then

often some prior supposition on the part ofthe researcher that one of these scales is moresignificant in explaining a given situation (e.g.an international war, a separatist movement, afire in a factory) than are any of the others(Smith, N., 1993). This supposition appearsquite logical, although it must be acknowl-edged that human geographers of differentpersuasions do tend to give priority to onescale over another whatever the subject-matter under study (e.g. some Marxist geog-raphers always prioritize the workings ofglobal capitalism; some humanistic geogra-phers always prioritize the hopes and fears ofpeople living in their local neighbourhoods).In line with the broader trajectory of our rea-soning, though, it should be added that manyhuman geographers now urge us to avoid pri-oritizing one geographical scale over anotherand propose instead that we should alwaysstrive to see the intersections of influencesemanating from different scales (particularlythe articulations of the global and the local,giving what is sometimes referred to as the‘glocal’; see esp. Massey, 1993; 1994: chs 6and 7). More generally, there is a powerfulargument which insists that human geogra-phers, and indeed other academics, shouldendeavour to avoid the deceptive seductionsof hierarchical thinking, or at least to becomemuch more self-aware than hitherto aboutthe urge to list and to rank things accordingto their supposed worldly importance andhence explanatory relevance (again, seeChapter 9). And once again there are someimmediate implications which will be men-tioned presently.

Alternatives andrecommendations

Much more could be said in this chapterabout what happens once researchers start to

3151-Ch-07.qxd 3/3/2004 4:08 PM Page 240

SIFTING AND SORTING

241

sift and to sort their data, and about how whatinitially may seem quite self-evident tasks,such as the naming of entities and the allocat-ing of them to boxes and lists, are really quiteproblematic and rooted in a host of question-able assumptions.To be sure, these are activitiesand assumptions deeply entrenched in theconventions of Western thought and action,and as such they are mostly taken for grantedand rarely exposed for critical attention. Yetwith the rise of new approaches both withinand beyond human geography, notably femi-nism (see esp. McDowell and Sharp, 1997),poststructuralism (see esp. Doel, 1999) andpostmodernism (see esp. Strohmayer andHannah, 1992; Benko and Strohmayer, 1997),such taken-for-grantedness is now beingextracted and subjected to intense question-ing.The foundations of knowledge, both aca-demic and common-sense, are thereby beingsubjected to a searching critique which to manymay seem esoteric and irrelevant but which,none the less, has far-reaching implications forsome of the apparently most mundane of ourpractices as human geographers. Indeed, asshould be clear from the discussion through-out this chapter, what might usually beregarded as the simplest, most uncontroversialpart of research – the basic sifting and sortingof data, the preliminary ordering in whichentities are identified,named,boxed and listed –is really one of the most problematic. Whatoccurs here in terms of the attributing ofidentities, the specifying of binaries and thepositing of hierarchies can no longer betreated as self-evident but must become a fur-ther element of the research endeavour aboutwhich we are self-critically aware. But what atbottom can this mean?

For many of the human geographers men-tioned in this chapter, there is a need to keepon worrying away at these foundations in thehope of being able to formulate alternative waysof thinking about how we order knowledge

(about what we take knowledge to ‘look like’ inthe first place).This has certainly been Olsson’sambition for two decades, commencing withhis earliest attempts to work with alternativesystems of logic (termed ‘many-valued’ and‘fuzzy logics’), which transcend simple‘either–or’ assumptions in preference for analternative ‘grammar of change and action’able to grasp certainty and ambiguity simulta-neously (Olsson, 1980: Portfolio 2). Few geo-graphers have followed this lead althoughthose closest to having done so are the likes ofStephen Gale with his ‘fuzzy set’ theory (1972;and note that Olsson describes Gale as some-one ‘who was certain about ambiguity’ –1980: n.p.) and also Stan Openshaw with hisconcern to deploy artifcial intelligence (AI) intandem with ‘fuzzy logic, fuzzy systems andsoft computing’ (Openshaw and Openshaw,1997: esp. ch. 10; Openshaw, 1996; 1998).Olsson himself quickly moved on from thesetechnical solutions, however, and even in his1980 text he framed them with a more‘dialectical’ vision derived from Marx (seeBox 7.10) wherein ‘every category includesboth itself and its opposite’ (Olsson, 1980:36b). Matters of identity and binaries weretherefore to be recast in terms of ‘dialecticalchange’, and it is worth quoting Olsson furtherin this connection:

It was exactly these forces of dialecticalchange that Marx tried to capture by usingbatlike words in which he could see bothbirds and mice at the same time. Thus, heallowed his categorial frameworks to be inconstant flux because he realised that thethings he was talking about were in con-stant flux. Put differently, he conceived ofhis seemingly inconsistent words as reallybeing consistent, because the worldhe saw was consistently inconsistent.(1980: 12b)

With dialectics he still sought to find someway of disrupting the usual obsession withfixing things in sealed boxes, and thereby to

3151-Ch-07.qxd 3/3/2004 4:08 PM Page 241

Boasting a heritage stretching back into Antiquity, ‘dialectics’ is a ‘logic of reasoning’which operates differently from standard logical reasoning that is rooted in dataderived from sensations, and which is thereby positioned as subservient to the suppos-edly identifiable, stable and coherent objects in the world. Whereas ‘[s]tandard or ana-lytic logic … is rigid and abstract, a matter of fixed connections and exclusiveoppositions’, it is suggested that ‘[d]ialectical logic sees contradictions as fruitful colli-sions of ideas from which a higher truth may be reached by way of synthesis’ (Quinton,1988: 225). Hence, rather than being anchored in an apparently constant world wherefixed ideas attach to clearly delimited phenomena, dialectics embraces the possibilityof a more dynamic, changing and fluid world at the same time as ‘freeing’ the ideasinspired by this world to enter into higher-level alliances, affiliations and syntheses pro-ductive of wholly new insights about human life on earth.

Marx took on board such dialectical logic, arriving at a deep perception of how phe-nomena are constituted through their relations (and see various remarks throughoutthis chapter about relational thinking), and developed a broader understanding ofhuman history – his ‘dialectical materialism’ – in which every society contains withinitself the ‘seeds’ of another society that it might ultimately become (or, more precisely,it encompasses within itself contradictions whose intensification and eventual resolu-tion will tip that society over into being something quite different and new). Nobodyhas done more than Harvey (esp. 1995) to bring Marxist dialectics into geographicaltheory, emphasizing that ‘[d]ialectical thinking prioritises the understanding ofprocesses, flows, fluxes and relations over the analysis of elements, things, structuresand organised systems’ (1995: 4). Moreover, he states that ‘[e]lements or “things” … areconstituted out of flows, processes and relations’ (1995: 5), and that it is vital to appre-ciate the sheer ‘heterogeneity’ of things in that they are always themselves constitutedthrough ‘internal relations’ holding together contradictory forces or tendencies within‘smaller’ component things. The implications for thinking about issues of identity,binaries and hierarchies are legion.

Box 7.10: Dialectics


242

see A and not-A as at once inter-related andmutually constitutive. This manoeuvre con-nected him to Marxist geographers such asDavid Harvey, for whom Marx’s dialectics wasthe preferred starting-point in sifting andsorting the data of the world,10 but he stillfound himself feeling that such an orientationwas too prone to collapsing the dialecticalmoment into rigid binaries.The upshot is thathe began to turn increasingly to the explodedlogics and free-flowing creativities of surreal-ism – ‘every time I returned, I saw more sur-realism’ (Olsson, 1980: n.p.) – and this hasmeant that, while striving to produce a new

‘invisible cartography’ freed from the irongeometries of conventional thought,11 he hasincreasingly turned for inspiration to thepoetic and ‘dreamlike’ writing of surrealistslike Joyce wherein ‘truth emerges when iden-tities are violated and opposites unified’(1980: 47e).

We realize that it may seem odd toencounter such notions in a book about prac-tising human geography, and we are fullyaware of the criticisms which may be levelledat Olsson (e.g. Billinge, 1983; Sparke, 1994),but we are still prepared to suggest thatresearchers could do a lot worse than allowing

3151-Ch-07.qxd 3/3/2004 4:08 PM Page 242

SIFTING AND SORTING

243

such ‘crazy’ notions to swirl around theirheads at the stage when they begin seeking toimpose order on their materials. Just sittingthere with your data, knowing that arespected human geographer such as Olssonhas contemplated the possibilities of very‘other’ ways of ordering knowledge, evenways which in a conventional sense seementirely disordering and remote from the nor-mal expectations of ‘science’, will potentiallybe a very liberating (if unnerving) experience.But it is not just Olsson’s example whichmight be pondered, for many other respectedhuman geographers have been equally chal-lenging of standard orderings, none more soperhaps than the feminist geographers whohave queried the common dependence ondualistic constructions of knowledge anchoredin an equation of what is deemed ‘superior’with what is deemed ‘masculine’. Even sothere is a difficulty about leaving readers withthe impression that in their own individualresearch projects they should emulate theexamples of Olsson and certain feminist geo-graphers, and hence devote their energies tochallenging the very pillars of Westernthought. While we do wish to encourage acritical edge, one that does recognize thesepillars to be somewhat more unsteady than iscommonly supposed, we appreciate that formuch of the time researchers will have littlechoice but to accept the normal proceduresof sifting, sorting, boxing and listing, as predi-cated on standard assumptions about identity,binaries and hierarchies. But does this there-fore imply that most of our discussion in thischapter is irrelevant to the everyday practisinghuman geographer? To some extent we haveto be honest and say ‘yes’, but this is notentirely the case because we can offer someconcrete proposals about a critically self-reflexive sifting and sorting which can be ofimmediate relevance.

It is with a listing of such proposals that weshall now conclude the chapter. First of all, itis crucial to be self-critical about the identities thatyou identify. In most pieces of research, we willwant to sort out our raw data in such a man-ner that we can identify definite things withinthem, whether as nameable entities (e.g.Merrybank Farm) or specifiable categories(e.g. sheep farms). More specifically we willwant to take particular statements (whetherfrom documentary sources, interview tran-scripts, field diaries or whatever) and codethem up under clearly distinguishable head-ings (e.g. ‘class’, ‘gender’, ‘community’, ‘dis-ability’). We must be aware of what we aredoing in this respect, though, and give con-siderable thought to the act of entitationinvolved. We must remember that we couldalways be identifying entities and categoriesdifferently from how we are doing, and that itis not somehow ‘natural’ or automatically cor-rect to be identifying certain ones rather thanothers. The secret is to be self-critical aboutour identified entities and categories andalways regard them as ‘open questions’, asthings that might need to be revisited,renamed and reclassified, even abandoned, asthe project progresses. We should endeavournever to let them achieve the solidity, thefrozen quality, which is all too typical of muchsocial-scientific research.

Secondly, it is crucial to be alert to the messi-ness of identity. Following from the above, wemust acknowledge that even seemingly self-evident entities and categories, such as anamed place, an occupational grouping orsupposed fundamentals like ‘woman’ and‘man’, may not be quite as straightforward asmight be supposed.There may be tricky the-oretical debates which suggest this to us, asindicated earlier, but it may be that within thedetails of our raw data this also becomesapparent. A simple spatial designation like

3151-Ch-07.qxd 3/3/2004 4:08 PM Page 243


244

‘Glasgow’ can turn out to be quite complex,for instance, in that a researcher may find thatwhat is being called ‘Glasgow’ in differentsources and by different people refers to terri-tories of differing sizes with differing centralpoints and differently drawn boundaries.Thepotential variability of an occupationalgrouping like ‘carpenter’ (or even of the verynotion of ‘occupation’) has already been dis-cussed and could easily be detected fromempirical data, as Glennie shows us. Thepotential variability of the entities ‘woman’and ‘man’ has also been discussed and, while itis unlikely that much of our data will call forus to question the continued use of these cat-egories, it is not impossible that it might (asLewis and Pile, 1996, found when studyinggender performances at the Rio Carnival,Brazil).The implication is that we should beprepared to accept that identified entities andcategories are almost always going to be lesssingular (more overlapping), less coherent(more fragmented) and less stable (morechangeable) than is commonly expected of theidentities upon which research is based. Thisdoes not mean that we should stop workingwith them as the basis for further interpreta-tions, but it does mean that we must consis-tently add qualifications about their messiness(if not sometimes foregrounding the fact so asto question the orthodoxies of others).

Thirdly, it is crucial to be alert to the contextsof identity. Following from the last two points,we must ensure that anything which we doname or classify is meaningful to the particu-lar context of our research: in other words,that it is appropriate and relevant to the people,situation, time and space under study. Ideallythis means that the entities and categories atwhich we arrive should be ones recognizableand understandable by the people in the givensituation, time and space. This cannot andneed not always be the case since there will be

occasions when the researcher has to operatewith names and classes imposed very muchfrom outside the situation (perhaps whenstudying prehistoric human geography forwhich virtually no ‘written’ records remain, orperhaps when purposefully attempting ahighly conceptual account dependent on termsderived from prior theories and models).None the less, our usual preference would be,as far as possible, for striving to use the words,phrases and discourses of situated peoplethemselves – whether derived from a planningdocument, an interview, a casual conversation,even an overheard remark – in order to guideus in identifying entities and categories perti-nent to the subject-matter of the research. Asan aside, this is a strong claim now cominginto geography from scholars influenced byethnomethodology (e.g. Laurier, 2001;Laurier and Philo, 2003), who give priority toclosing the gap between the terminologies ofacademics and the ‘lay’ vocabularies throughwhich situated peoples cannot but thinkand, in effect, perform their own worlds.Moreover, from the previous chapters of thebook, we can note that there are many waysof constructing data which preserve the subtletexture of what is meaningful to people in thehistorically and geographically specific con-texts of their everyday lives, work, play, lovesand hates. It should be added, of course, thatthis stance should not preclude us from beingcritical of the entities and categories favouredby the people under study: they are certainlynot beyond criticism, even if this must bedone with a contextual sensitivity, and theirviews can be commented upon judgemen-tally from the standpoint of the researcher (seeChapters 1 and 12).

Fourthly, it is crucial on occasion to considerscrambling binaries and hierarchies. In followingthe previous recommendations, we shouldalso aim to avoid falling unthinkingly into

3151-Ch-07.qxd 3/3/2004 4:08 PM Page 244

SIFTING AND SORTING

245

specifying dualisms or adopting hierarchies. Inthe spirit of what we have just said aboutidentities, this is not to suggest that we needto purge all traces of binaries and hierarchiesfrom our research since this is probablyimpossible anyway (given just how pivotal arebinaries and hierarchies to Western thought).But it is to suggest that we are self-critical inthis respect as well, and that we do our best tospot binaries and hierarchies emerging fromthe preliminary ordering of our data. If webegin to figure that contrasts between the cityand the country might be germane to theresearch, or those between highlands and low-lands, Gemeinschaft and Gesellschaft, Fordismand post-Fordism and so on, then we are def-initely bringing binaries into the heart of ourresearch. It is therefore incumbent upon usto ascertain whether or not such contrastsare meaningful to the particular study beingconducted, perhaps asking if they are onesregistered as significant by the peopleinvolved; and we must beware of unwittinglyprioritizing one element of a binary over theother in a fashion according it with valuewhile dismissing its opposite as valueless(which would be to fall into the dualistic trapwith its likely gendered underpinnings cri-tiqued by feminist geographers). This beingsaid, it may be that the people under studythink in terms of binaries, and quite possiblyin terms of dualistic and gendered binaries, inwhich case our task is to report on this findingwhile retaining a critical distance alert to thesociospatial inequalities perhaps resultingfrom grounded actions informed by suchbinaries. Finally, the same principles apply tohierarchies. From the moment we start to listthings which we reckon to be relevant to ourresearch, the very ordering of the list has thepotential to become an unexamined rankingof what is considered more or less important,determining or of explanatory value.We must

therefore explore the relevance of our list-ings to the people under study, noting theirsense of what is or is not important, andkeeping their rankings (as far as these can bediscerned) as a check upon our own.That weshould do this self-critically, but also with acritical eye on the largely taken-for-grantedhierarchies harboured by our research sub-jects, can presumably now go pretty muchwithout saying.

Notes

1 See, for example, Philbrick (1957), Grigg(1965; 1967), Abler et al. (1971: ch. 6) andSemple and Green (1984).

2 See Blaut (1961), Harvey (1969), Sack (1980),Gregory and Urry (1985), Gregory (1985;1994), WGSG (1997: esp. ch. 1), Massey(1992; 1994; 1999a; 1999b) and Thrift (1996;1999).

3 See Goss (1995b), Pickles (1995), Curry(1998); but see Clark (1998) and Flowerdew(1998).

4 This is precisely what we are attempting todo in most versions of ‘coding up’ transcripts andother textual sources, as explained further inChapter 10; see also Crang (1997) who usesthe phrase ‘sifting and sorting’ precisely inthis connection.

5 See Gregory (1989b), Cloke et al. (1991: ch. 6),Strohmayer and Hannah (1992); see alsoChapters 9 and 10.

6 For good surveys of structuralism in geogra-phy, noting its links to Marxism, see Gregory(1978a: ch. 3; 1994) and Peet (1998: ch. 4).

7 For further guidance on poststructuralism ingeography, see Dixon and Jones (1996a),Natter and Jones (1993; 1997), Peet (1998: ch. 6)and Doel (1999).

8 Interestingly, elsewhere Jones utilizes anumerical modelling technique called the‘expansion method’ as a formal spatial-analytic vehicle for coping with the variabil-ity of context; see Jones and Casetti (1992).

9 Of course, attention should be given to thetotal numbers of statements under eachheading, in which case a simple numerical measure can be deployed to ensure that

3151-Ch-07.qxd 3/3/2004 4:08 PM Page 245

the priorities in the understandings of theinterviewee are not completely over-ridden;see what we say about numerical ‘contentsanalysis’ in Chapter 8.

10 See Harvey (1973: esp. ch. 7; 1995; 1996); seealso Castree (1996).

11 See Olsson (1991a; 1991b; 1994); see alsoPhilo (1994).


246

3151-Ch-07.qxd 3/3/2004 4:08 PM Page 246

Enumerating8

Enumeration andhuman geography

In this chapter we turn to the use of numbersin the practising of human geography.Numbers are a fundamental facet of oureveryday lives. They convey information tous; they underwrite key categorizations ofsize, standard, price, importance, progress andso on; they are used to present significant evi-dence to us in favour of, or in opposition to,some opinion, policy, product, way of life orplace; and sometimes they are offered as proofof some underlying generalization. Thus webuy clothes of a stipulated size and we buyalbums from numbered charts. Our schools,universities and hospitals are measured interms of their performance and our housingvalues and our incomes are taken as a reflec-tion of the standard of the area and the taxa-bility of the people who live there. It isinevitable, therefore, that human geographerswill be informed by, and in turn will seek toinform, this numerical world using processesand practices of enumeration.

For some human geographers, quantitativeanalysis can be represented as a modusoperandi for addressing the nature of majorsocial and political problems in their ownterms. A recent critical review by Johnstonet al. (2003) charts national-scale problemssuch as failing schools, unemployment, ill-health and hospital waiting-lists alongside

international-scale problems associated withliving standards, famine, AIDS and environ-mental degradation, and contends that suchproblems are difficult to assess, let alonerespond to, without qualitative analysis.Theyargue that

The identification of these and many otherproblems – their what, where and inten-sity, all of which have to be addressedbefore one can turn to why and then howto remove them – involves measurement.Although many of the problems involvealleviating (at least) the situations of indi-viduals, nevertheless they can only beattacked by focussing on populationaggregates, such as famine sufferers, orthe unemployed, or the long-term sick inparticular areas – without committing theecological fallacy. Knowing where suchpeople are involves first measuring thesymptom … and then identifying othercharacteristics of either the people them-selves or the places they live in, so as to iso-late potential causative factors for theirconditions and then analyse the impact ofattempted cures (aggregate social researchinvolving the ability to handle large datasets). (2003: 158)

The use of enumeration to interpret humangeography data reflects in some ways anuncomplicated story by which ever moresophisticated quantitative techniques havebeen developed to analyse numerical spatialdata and to construct and test mathematicalmodels of spatial processes. There is anothersense, however, in which the development of

3151-Ch-08.qxd 3/2/04 2:45 PM Page 247

quantitative geographies has been implicatedin a rather tense and often dichotomous rela-tionship between quantitative and qualitativegeographers undertaking their quantitativeand qualitative analyses, sometimes in opposi-tion to each other.Thus alongside the evolv-ing sophistication of quantitative analysis,there has come a proselytizing zeal to presentenumeration as the way of doing things,matched only by the ‘reverse swing’ of thosewho would dismissively regard qualitativeanalyses as the way. It would be very easy tocharacterize, or even satirize, the relationshipbetween quantitative and qualitative geogra-phies as a debate between two entrenched,mutually antagonistic camps. Such a debatewill, however, be variously constructed andperceived. Our team of authors includes indi-viduals who have encountered trenchantopposition to the introduction of qualitativemethods courses into the curriculum of par-ticular universities. On the other hand, whenreflecting on the ‘two entrenched, mutuallyantagonistic camps’, Ron Johnston (pers.comm.), a quantitative researcher of highrepute, reflects:

the more I think about it, the more I feelthat this is not the case, and that there isan asymmetry: few ‘quantifiers’ disputethe validity of the general questionsbeing posed by the ‘qualifiers’ … onthe other hand, many ‘qualifiers’ deny thevalidity of the questions the ‘quantifiers’ask, and brand all of the latter as logicalpositivists.

In a recent theme issue of Environment andPlanning A on ‘reconsidering quantitativegeography’, Philo et al. (1998) recall how theemergence of the quantitative ‘revolution’ inthe early 1960s was itself flavoured by anantagonistic divide between the older qualita-tive and new quantitative approaches. Theyquote Floyd, writing in 1963, as an illustrationof how this new geography was received atthe time:

Faced with the intellectual immensity of thegeographer’s goal, a number of youngworkers … would have us eschew the inte-grative, global philosophy of the subject,time-honoured as it is, and turn our atten-tion to sterner, more ‘masculine’ things: arigorous wrestling with restricted data andareas utilising the ‘new’ scientific methods,quantitative ways of analyses, statistical andmathematical tools and models, and soforth. The old ways often lead to a never-ending trail of ‘trivia, boredom and ennui’,the new presents challenging, brain-crackingproblems in combinational topology, scal-ing, theories of central place location, near-est neighbours, the journey to work andlinear (or at least quasi-linear) program-ming. (p. 15)

As Philo et al. point out, Floyd’s recognitionof the excitement and potential associatedwith emergent quantitative geographies wastempered by a strong concern that theseinnovations would signal the death knell ofthe existing mainstream position of qualitativemethodologies, to the extent that ‘the non-quantitative geographers are politely if firmlyrelegated to the lower order of workerswithin the geographical hierarchy’ (Floyd,1963: 15).

The speedy dominance achieved by quan-titative approaches in geography has oftenbeen accompanied by self-representations ofconsiderable machismo:

Quantitative geography is here to stay. Therevolution is over. It is not a ‘new’ geogra-phy in the sense that it alters the nature ofthe discipline itself; but only in that it offersnew techniques – for the solution of oldproblems as well as new ones – techniqueswhich have brought a new rigour into geo-graphical thinking, and have made possiblethe exploration of new fields. (Hammondand McCullagh, 1974: xi)

On the other hand, this ‘new rigour’ has beenheld by others to be at the expense of under-standing meanings and explaining phenomenadue to the uncritical conceptualization ofgeographical objects and subjects. As Dey(1993: 3) suggests:‘The growing sophistication


248

3151-Ch-08.qxd 3/2/04 2:45 PM Page 248

of social science in terms of statistical andmathematical manipulation has not beenmatched by comparable growth in the clarityand consistency of its conceptualisations.’

Gould’s (1979) reflections of geography inthe 1960s and 1970s suggest that many earlyborrowings of statistics by geographers wereinappropriate and that the quantitative ‘revolu-tion’ which at first seemed to liberate humanthought was in danger of being later trans-formed into dogma and myth. Subsequently,the quantitative project in geography hasbecome subject to increasing criticism. In aninfluential paper, Barnes (1994) points to twostrands of criticism. First, and perhaps mostobvious, are criticisms from Marxist andhumanistic geographers about the applicabilityof mathematical knowledge to the representa-tion both of social and political conflicts, andof the significant characteristics of humanaction. Secondly, Barnes focuses on the internalblind spots and dead-ends of mathematics as astructured system of knowledge. He seesmathematics as a prime example of logocentrism– a term from Derrida (see, for example, 1991)which describes a belief in a fundamentallyordered world that can lead to a further beliefthat there is a fundamental level of knowledgebeyond which we need not go.What is beingsuggested here is that quantification in humangeography tends to present the world in anordered way which can be accessed usingnumeric data and understood by using thesedata both to describe and infer characteristicsof the human world, and to provide evidencefor or against hypotheses and laws by whichthat world can be understood. It is importanthere to discriminate between different under-lying approaches to quantification. For some,the privileging of observations represents anempiricist approach while others (whose test-ing of repeatable observations leads to theoryconstruction) may take a more positivistapproach (see Box 8.1).Yet others use quanti-tative techniques to test non-positivist theories

(see, for example, Plummer and Sheppard,2001). Albeit underpinned philosophically inthese different ways, the quantitative mind-set,once adopted (and sometimes it becomesdeeply ingrained in the intellectual psyche),makes it difficult to see beyond the numericand problematic to accept that there are other(non-numeric) bases of knowledge which canbe equally relevant to the broad project ofunderstanding the human world. In theseterms, quantification can be viewed as a(rather partial) attempt to impose a rather par-ticular form of order to the world.

Qualitative movements have emerged inchallenge to these quantitative orthodoxies,suggesting that some matters of concern tohuman geographers cannot simply becounted or measured, and therefore deployinga range of methodologies which focus onintensive research of a non-quantitative kind.Perhaps because of the strength of numbers ofquantitative practitioners in human geogra-phy, and the consequent difficulties thatothers have sometimes encountered in gain-ing legitimacy (in postgraduate theses, teach-ing, research funding, etc.) for their qualitativemethodologies, there has been rather a defen-sive posturing from qualitative geographerswho have sometimes seemed unwilling tosanction any numbers-based approaches totheir subject. Hence quantifiers have beenoften dismissive of qualifiers, and vice versa,and, in this context, enumeration is thereforepositioned as a game played by quantitativegeographers.

Some commentators now argue that thequantitative take-over from the qualitative, ascommented on by Floyd, has been reversed.By 1973, Harvey had pronounced that ‘thequantitative revolution has run its course, anddiminishing marginal returns are apparentlysetting in’ (p. 28) and, as the millenniumturned, Johnston (2000a: 131) was moved todeclare that ‘Spatial analysis isn’t dead (or evendying), therefore, so why is it being committed

ENUMERATING

249

3151-Ch-08.qxd 3/2/04 2:45 PM Page 249

Empiricism

Gregory (2000e) describes empiricism as a philosophy of science which privileges empir-ical observations over theoretical statements. Empiricism assumes that observed state-ments are the only way of making direct reference to the real world and that suchstatements can be judged as true or false without reference to theory.

Positivism

Positivism builds on these empiricist foundations but, additionally, insists that scientificobservations have to be repeatable, that the generality of observations leads to theformal construction of theories which when verified empirically can be regarded asscientific laws. In turn, these scientific laws can be unified progressively and integratedinto a single system of knowledge and truth (see Gregory, 2000f).

Box 8.1: Empiricism and positivism

to the discipline’s past by many of thoseconcerned with human geography’s future?’

The worry now in some quarters is that theincreasing dominance of qualitative method-ologies may signal the death knell of the pre-viously mainstream position of quantification,representing an increasingly difficult ‘contestfor resources, power and influence within thediscipline’ ( Johnston, 2000a: 131) and cer-tainly marking a ‘downturn’ in the fortunes ofquantitative geography during the 1980s and1990s (Graham, 1997; Johnston, 1997a).Fotheringham et al. (2000) suggest five issueswhich set such a downturn in context(Box 8.2).They argue that many of the criti-cisms traditionally levelled at quantitativegeography no longer apply. For example, notall quantifiers can be assumed to be positivists:

just as some quantitative geographers believein a ‘geography is physics’ approach (natural-ism) which involves a search for global ‘laws’and global relationships, others recognisethat there are possibly no such entities … theemphasis of quantitative analysis in humangeography is to accrue sufficient evidencewhich makes the adoption of a particular lineof thought compelling. (2000: 5)

Neither, it is argued, should they be presentedas being driven by any deep-rooted politicalor philosophical agenda: ‘For most of its

practitioners, the use of quantitative techniquesstems from a simple belief that in many situa-tions, numerical data analysis or quantitativetheoretical reasoning provides an efficient andgenerally reliable means of obtaining knowl-edge about spatial processes’ (2000: 4). Such arecognition of diversity among quantitativepractitioners is both significant and, in ourview, very helpful. We want to recognize amyriad of entry points into human geography –those places, events, readings or interests whichspark an individual’s imagination to go deeperinto the subject. It is clear to us that quantita-tive methods can be, for some people, just suchan entry point. For some researchers andstudents it is the compelling association withmathematical modelling, statistical analysis andcomputing which is the unambiguous pointof departure from which the journey throughthe terrain of human geography begins.Indeed, the selective nature of that terrain willoften be mapped out by decisions about whatkinds of subject-matter provide ‘good fit’ withthe desired methods. It seems likely in thesecircumstances that quantitative geographies, orperhaps a search for order which underpinssome of them, will be the dominant factorinforming judgements about what geographyis and what it should be.


250

3151-Ch-08.qxd 3/2/04 2:45 PM Page 250

1. A disillusionment with the positivist philosophical underpinnings of much of theoriginal research in quantitative geography and the concomitant growth of manynew paradigms in human geography, such as Marxism, postmodernism, structural-ism and humanism, which have attracted adherents united often in their anti-quantitative sentiments.

2. The seemingly never-ending desire for some new paradigm or, in less polite terms,‘bandwagon’ to act as a cornerstone of geographical research. The methodology ofquantitative geography had, for some, run its course by 1980 and it was time to trysomething new.

3. A line of research that appears to be better accepted in human geography than insome related disciplines is one that is critical of existing paradigms. As quantitativegeography was a well-established paradigm it became, inevitably, a focal point forcriticism.

4. As part of the broader ‘information revolution’ which has taken place in society,the growth of geographical information systems (GIS), or what is becoming knownas geographical information science (GISc), from the mid-1980s onwards, has hadsome negative impacts on quantitative studies within geography. Interestingly,these negative impacts appear to have resulted from two quite different percep-tions of GISc. To some, GISc is seen either as the equivalent of quantitative geogra-phy, which it most certainly is not, or as the academic equivalent of a Trojan horsewith which quantitative geographers are attempting to reimpose their ideas intothe geography curriculum (Taylor and Johnston, 1995; Johnston, 1997a). To others,particularly in the USA where geography has long been under threat as an acade-mic discipline, GISc has tended to displace quantitative geography as the para-mount area in which students are provided with all-important job-related skills(Miyares and McGlade, 1994; Gober et al., 1995).

5. Quantitative geography is relatively ‘difficult’ or, perhaps more importantly, is per-ceived to be relatively difficult both by many academic geographers, who typicallyhave limited quantitative and scientific backgrounds, and by many students. Thisaffects the popularity of quantitative geography in several ways. It is perceived bymany students to be easier to study other types of geography, and their exposureto quantitative methodology often extends little beyond a mandatory introductorycourse. It deters established non-quantitative researchers from understanding thenature of the debates that have emerged and which will continue to emergewithin quantitative geography. It also makes it tempting to dismiss the whole fieldof quantitative geography summarily through criticisms that have limited validityrather than trying to understand it.

Source: Fotheringham et al. (2000: 1–3)

Box 8.2: Reasons for the ‘downturn’ in quantitative geography

However, it is also important to recognizethat a single methodological impulse – whetherquantitative or qualitative – need not occupy anall-embracing position in the personality ofhuman geography. Practices of enumeration,statistical analysis and spatial modelling can alsoform part of a wider portfolio of techniquesopen to human geographers. In this case, a

particular topic, issue, subject or question maybe the entry point which sparks the geograph-ical imagination and the selection of suitabletechnique(s) with which to practise humangeography in this area can be dictated by thenature of that subject-matter.

Indeed, we would strongly suggest thatquantitative and qualitative analytical techniques

ENUMERATING

251

3151-Ch-08.qxd 3/2/04 2:45 PM Page 251

need not be viewed as discrete choices, andthere are now signs of a growing interest inanalyses which span the quantitative–qualitativedivide. Although there are geographers whocontinue to use mathematics in a fundamen-tally logocentric way, there are many otherswho have sought in different ways to explorethe potential interconnections between enu-meration and other interpretative methodolo-gies (see, for example, Fielding and Fielding,1986; Fielding and Lee, 1991). For example,essentially quantitative geographers (seeWilson, 1974) have suggested that multilevelmodelling (see below) supposedly represents away of using quantitative approaches toaddress some of the contextual and conceptualissues being raised in highly politicized and‘subjective’ arenas. A recent move in thisdirection has come from Openshaw (1996;Openshaw and Openshaw, 1997). Startingfrom an unpromising re-articulation of thequantitative (‘hard’)–qualitative (‘soft’) dividein which soft methodologies are recognized ashaving ‘good’ and ‘bad’ traits (Box 8.3),Openshaw opens out some interesting possi-bilities for translating ‘soft’material into a formwhich is compatible with computationalmethods derived from artificial intelligence.His advocacy for using ‘fuzzy logic’ not neces-sarily in a numeric way represents a small butpotentially significant step towards intercon-necting the quantitative and the qualitative.

From a very different starting point, someusers of qualitative methodologies (see Dey,1993) have suggested that the growing impor-tance of computer-based qualitative analysis,dealing for instance with enumerating theuses of different metaphors and the links indiscourse, will bring qualitative and quantita-tive analysts closer together.Yet others, oftenusing a realist philosophy of science, foreseethe linking of causal mechanisms and thevariety of concrete events determined bythose mechanisms in terms of quantitativeenumeration (see Foot et al., 1989). One

further area of interconnection which hasconsiderable potential is where mathematics isused critically, recognizing that it is a languagejust like any other. Researchers such as Sibley(n.d.), Pratt (1989) and Sheppard and Barnes(1990) have rejected the idea of mathematicsas a logocentric project yet have used numericdata in an exploratory way as ‘a technique fororganising one type of information, informa-tion that needs necessarily to be comple-mented by a more relational, contextualunderstanding, as well as more abstract theo-retical development’ (Pratt, 1989: 114).

Many of the arguments rehearsed so far inthis chapter illustrate the potential for defensivedivisiveness in human geography’s methods. Ifyou are quantitative then you are presumed tobe anti-qualitative, and vice versa. While suchstereotypes are no doubt given expression inlecture theatres, tutorial rooms and coffeehouses throughout the international communityof human geography (and certainly the widercommunity of geography, in which physicalgeographers often have a strong contribution tomake on these issues), we want to argue thatthis need not be so. We view enumeration asone of a range of available interpretative strate-gies for human geographers. Moreover, weagree with Philip (1998) that there is consider-able value to be gained from scrambling thesupposed quantitative–qualitative divide andfocusing more often on multiple methodolo-gies which deploy different techniques and dif-ferent stages of a research process, and whichrevisit the supposed binary relationshipbetween objectivity and subjectivity.

In the remainder of this chapter we presenta brief account of some of the methods ofanalytical enumeration available to humangeographers.We do not seek here to present adetailed manual of individual techniques.There are many other texts which performthis task. Neither do we shrink from intro-ducing critical debates around the intellectualcontext of particular forms of enumeration


252

3151-Ch-08.qxd 3/2/04 2:45 PM Page 252

Possibly good aspects

• They are a means of coping with the task of understanding complex human systems.• They are concept rich and theory rich.• They strengthen links with other social sciences.• They focus on developing a social and political dimension.• They are extremely flexible and non-constraining on the topic of study.• They emphasize the importance of social context.

Possibly bad aspects

• They are ignorant of the vast amounts of computer information that exists.• They neglect information technology.• They engender severe computer skill deficiencies.• They are intolerant of non-conformist research.• They exaggerate the problems of geographical information systems and quantita-

tive approaches.• They involve no understanding of science.• They involve a denial of science.• They focus both on the uniqueness of the world and on interpretations of the

world.• No consensus-based research agenda is possible.• They are more strongly linked to fringes of other social sciences than to geography.

Source: Openshaw (1998: 325)

Box 8.3: The goods and bads of ‘soft’methodologies in human geography

since the deployment of this ‘critical edge’ isalso part of the portfolio of skills for practisinghuman geography. We do, however, regardthese forms of analysis as potentially powerfuland useful tools with which to make sense ofmany kinds of human geography data.

Describing, exploring,inferring

Following the quantitative revolution of the1960s, the analysis of numeric data using sta-tistical techniques became the orthodox meansof practising human geography and, in theUSA, for example, this orthodoxy has contin-ued to be dominant in many educational cur-ricula and research programmes. In the UK,as in parts of the US system, the growth of

qualitative geographies has presented a challengeto any notion that quantification is the ortho-dox methodology, but numeric analysisremains a significant part of the human geo-grapher’s toolkit. In the remainder of thischapter we explore some of the manoeuvresavailable to researchers wishing to analyse andinterpret numeric data.We begin with perhapsthe most straightforward practices: summariz-ing data through the descriptive use of statis-tics; exploring data through visualization; andinferring aspects of the population fromsamples – and it is worth noting at this pointthat this statistical pathway should not beequated merely with solely quantitative researchstrategies.Although by definition the interpre-tative techniques offered by enumerationrequire the data to be in the form of numbers,it is certainly the case that such interpretation

ENUMERATING

253

3151-Ch-08.qxd 3/2/04 2:45 PM Page 253

can be implemented on data (for example,words, visual images, sounds) which have beencategorized or otherwise transformed intonumbers, as well as data constructed in numer-ical form. As such, enumeration is a validoption for interpreting many of the differenttypes of data construction discussed earlier inthis book, and not just for the more obviousconstructions of official data, social researchdata and questionnaire interview data. Indeed,several aspects of more qualitative researchstrategies are open to interpretation by enumer-ation. For example, Krippendorff ’s (1980: 7)account of content analysis suggests that itshistory has been firmly rooted in a ‘journalis-tic fascination with numbers’, and basic tech-niques such as word counts and textualcontent analyses will certainly transform dataconstructed in non-numeric form into num-bers which can then be interpreted by enu-merative methods. Moreover, although it isoften the case that human geographers will befocusing on the singularities of particular textsrather than looking for patterns within them,such singularities are likely to be embedded inimages or language which will probably con-tain a number of implicit classifications andcomparisons. As Dey (1993) argues, classifica-tion should not be viewed merely as the stuffof structured questionnaire interviews; rather itmay be the aim of various strategies of qualita-tive research to recognize and specify the com-parisons and classifications used by the subjectof that research. Thus enumeration is oftenbound up with classification (see Chapter 7):

Enumeration is implicit in the idea of mea-surement as recognition of a limit or bound-ary. Once we recognise the boundaries tosome phenomenon, we can recognise andtherefore enumerate examples of that phe-nomenon. Once we know what a ‘school’ is,we can count schools. Indeed it may be hardto describe and compare qualities entirelywithout enumerating them. (Dey, 1993: 27)

Having established that statistical description,exploration and inference can be applied to

many different types of data, we now broachone of the most thorny issues in enumerativehuman geographies – that of how differenttypes of statistical manoeuvres should them-selves be categorized. The approach of manytextbooks is to divide groups of statistics intotheir usual analytical function. In this, way, weare presented with ‘descriptive statistics’,‘inferential statistics’ and so on.This approachcan lead to considerable confusion overwhether there are separate descriptive andinferential statistics. In our view there are not.All statistics can be used to describe, and allcan be used to make inferences.Thus a simplemeasured mean can be used to describe thecentral tendency of the data set but can alsobe used to infer the characteristics of the meanof the wider population from which the dataset has been drawn. Similarly, a regressionequation can be used to describe the relation-ships in a set of values, as well as to infer rela-tionships more widely if the data are from asample. Bearing these potential confusions inmind, it is important to reiterate the impor-tance of appropriate sampling strategies (seeChapter 5). Some of the most importantstatistical devices in human geography arethose which focus on sampling error and yet,at the undergraduate level, it is often easy tounder-emphasize their role (for example, indissertation research).

In our account of enumerative techniqueswe have chosen to discuss processes of usingstatistics to ‘describe’ and to ‘infer’ rather thanseeking to identify ‘descriptive’ or ‘inferential’statistics per se.

Using statisticsto ‘describe’

There has been a regular flow of textbooksover the years which have sought to introduceand expand the subject of statistical methodsin geography.1 Although there have beensome interesting innovations (see Wrigley and


254

3151-Ch-08.qxd 3/2/04 2:45 PM Page 254

Bennett, 1981; Bailey and Gatrell, 1995; Roseand Sullivan, 1996, for example), this body oftextual material suggests a well-trodden corepathway of statistical interpretation. It is usualto point out first that in most instances geo-graphers do not have access to all the cases forwhich data are, in theory, available. Althoughmost official data such as population censusesand employment statistics are supposed to becomprehensive, in practice it is often neces-sary to study only part of the data set, select-ing samples of data with which to work (seeChapter 5). Having acknowledged the partialnature of the data being used, the next stagein statistical methodologies is to describe andanalyse the characteristics of these samples ofdata. Here we are introduced to statisticswhich provide measures of central tendency,the variability to data and the nature of theoverall data set. These statistics can be pre-sented numerically or graphically and theyperform two main tasks.

SUMMARIZING DATA

The statistical toolkit provides measures ofcentral tendency (mean, median, mode) andmeasures of dispersion (standard deviation,coefficient of variation) which help to sum-marize large data sets. Establishing central ten-dency is a useful but potentially problematicprocess.As an illustration,Table 8.1. draws ona major survey of ‘Lifestyles in rural England’(Cloke et al., 1994), part of which attemptedto measure poverty in the rural areas ofEngland. Surveys of 250 households were car-ried out in each of 12 study areas, and data onhousehold incomes were analysed to producemean and median values for each area.Thesedifferent ‘average’ values were then used tosuggest levels of affluence and poverty in therural areas concerned.The table highlights theresults of three attempts to provide an indica-tor of those households whose income levelsmight be described as being in or on the mar-gins of poverty. Such indicators are by nature

fraught with potential definitional problems(see Cloke et al., 1995a; 1995b; Woodward,1996) but, leaving these aside in this presentcontext, the first two indicators make use ofaverages to represent households in this way.

The first identifies those households whoseincome falls below 80% of the mean householdincome in all 3000 households in each studyarea surveyed, and the second substitutes themedian value for the mean.As can be seen fromthe table, not only is the range of percentages ofhouseholds with incomes below 80% of aver-age incomes markedly different according towhether the mean (higher values) of themedian (low values) is used but the rank order-ing of different case studies also varies accord-ing to whether the mean or median is used. Forexample, Shropshire has the fifth highest num-ber of households in or on the margins ofpoverty according to the mean indicator butthe eleventh highest according to the medianindicator,whereas West Sussex is tenth using themean indictor but second using the medianindicator. Such variations in the use of ‘average’statistics render the simple interpretation ofsuch data extremely problematic.The distribu-tion of incomes in particular study areas, andespecially the ways in which the distributionsdeviate around the mean, would require closeattention before any real sense could be madeof what the statistics are describing.

EXPLORING DATA

Statistical procedures can also be used as partof a more inductive framework of analysis bypresenting data in an appropriate form forsuggesting hypotheses about underlyingprocesses, and thereby suggesting fullermodelling practices which could be employedto investigate these processes further. Suchexploration often uses visual devices – scatter-plots, boxplots, histograms, pie charts and soon – to represent frequencies and to suggestpatterns, trends and processes which can befollowed up using other techniques.

ENUMERATING

255

3151-Ch-08.qxd 3/2/04 2:45 PM Page 255

As with the task of summarizing data, suchrepresentation and exploration are both fun-damental to some research tasks and yet opento variable interpretation, either due to lack ofknowledge or by design.To illustrate, we pre-sent in Box 8.4, a dated but still interesting useof the exploration of frequency statistics,drawn from Brown et al. (1992). Taking dataat a world scale on the electrical generatingcapacity of nuclear power plants, the analysisperforms several simple interpretative devicesin order to offer different facets of descrip-tion. A figure of 326 000 megawatts capacityfor the year 1991 has some interest in its ownright but it is enhanced by being placed in thecontext of the capacities recorded in previousyears. Not only does this allow a comparisonwith the early years of nuclear power genera-tion to show the dramatic increase that hasoccurred over the 40 years concerned but italso affords a year-on-year comparison whichinforms us that, between 1990 and 1991,nuclear generating capacity fell for the firsttime since commercial nuclear power wasintroduced.

Graphic display of these data (often by barcharts as well as lines) helps to interpret the

pattern concerned, although in this case therecent downward turn is hardly discernible intheir Figure 1. In their Figure 2 we are invitedto interpret the trend in nuclear generatingcapacity in terms of the concomitant trend inthe number of construction starts for nuclearreactors. It may have been assumed that theseemingly inexorable rise in generating capa-city was associated with a continuing flow ofnew reactors. However, this frequency distrib-ution shows a context of peak construction inthe 1970s since when the number of starts hasbeen subject to a jagged downward trend.Such processes of providing numerical contextfor the original frequency distribution couldhave gone further, of course, perhaps illustrat-ing the number of reactors shut down over theperiod, the relative performances of coal-firedor gas-fired electricity generation, the geo-graphical distribution of generators and so on.The authors themselves introduce these andother factors in their summary interpretationwhich provides an excellent example of well-crafted description. Here we are presentedwith an informative and convincing accountbased on enumeration of ‘facts’. It remainsdescription, however, because there remain so


256

Table 8.1 Indicators of households in or on the margins of poverty in 12 rural studyareas

Mean* Median** Income support***

North Yorkshire 61.9 Devon 47.1 Nottinghamshire 39.2Northumberland 61.1 West Sussex 45.8 Devon 34.4Notttinghamshire 58.8 Essex 44.4 Essex 29.5Essex 53.3 Cheshire 43.6 Northumberland 36.4Shropshire 51.4 Suffolk 42.6 Suffolk 25.5Devon 50.0 Nottinghamshire 41.2 Wiltshire 25.4Wiltshire 49.3 North Yorkshire 40.5 Warwickshire 22.6Warwickshire 48.4 Warwickshire 38.7 North Yorkshire 22.0Cheshire 46.2 Northumberland 37.5 Shropshire 21.6West Sussex 45.8 Wiltshire 36.6 Northamptonshire 14.8Suffolk 44.7 Shropshire 35.1 Cheshire 12.8Northamptonshire 43.6 Northamptonshire 34.5 West Sussex 6.4

Notes:*in or around margins of poverty: < 80% mean**in or on the margins of poverty: < 80% median***in or on the margins of poverty: < 140% income supplement entitlement

Source: Cloke et al. (1994: 94)

3151-Ch-08.qxd 3/2/04 2:45 PM Page 256

ENUMERATING

257

Box 8.4: Describing frequencies: the example of the electricity-generating capacity of nuclear power plants, 1950–91

Between 1990 and 1991, total installed nuclear generating capacity declined for the firsttime since commercial nuclear power began in the fifties. There were 421 nuclear plantsin commercial operation in January 1992, fewer than at the peak in 1989. These plants sup-plied 17 percent of the world’s electricity and had a total generating capacity of 326,000megawatts – only 5 percent above the figure reported 3 years earlier. (See Figure 1.)

19500

5

10

15

20

25

30

35

1960 1970 1980 1990 2000

Gigawatts

Source: IAEA

19500

50

100

150

200

250

300

350

1960 1970 1980 1990 2000

Gigawatts

Source: IAEA, UN

World net installed electricalgenerating capacity of nuclearpower plants, 1950–91Year Capacity [megawatts]

1950 01951 01952 01953 01954 51955 51956 501957 1001958 1901959 3801960 8301961 8501962 1,8001963 2,1001964 3,1001965 4,8001966 6,2001967 8,3001968 9,2001969 13,0001970 16,0001971 24,0001972 32,0001973 45,0001974 61,0001975 71,0001976 85,0001977 99,0001978 114,0001979 121,0001980 135,0001981 155,0001982 170,0001983 189,0001984 219,0001985 250,0001986 275,0001987 298,0001988 311,0001989 321,0001990 329,0001991 326,000

Figure 1 World electrical generating capacity ofnuclear power plants, 1950–91

Figure 2 World nuclear reactor construction starts1950–91

Sources: R. Spiegelberg, Division of Nuclear Power, International Atomic Energy Agency, Vienna, Austria, unpublished printout,March 18, 1992; United Nations, Department of International Economic and Social Affairs, Energy Statistics Yearbook, 1950–1974(New York: 1976); Greenpeace International, WISE-PARIS, and Worldwatch Institute, The World Nuclear Industry Status Report:1992 (London: 1992).

Source: Brown et al. (1992: 48–9)

3151-Ch-08.qxd 3/2/04 2:45 PM Page 257

many other questions about context, inten-tions, meanings and processes in which thesefacts are embedded.To use one brief exampleof this, some understanding of the authorship of

this piece on nuclear energy provides acontext for the facts presented. Brown et al.(1992) are writing on behalf of the WorldwatchInstitute, an environmental pressure group. In


258

Forty-nine nuclear plants are under active construction worldwide, a quarter as many asa decade ago (see Figure 2), with a combined capacity of 39,000 megawatts. Many of theseare nearing completion, so in the next few years worldwide nuclear expansion will nearlystop. It now appears that by decade’s end the world will have at most 360,000 megawattsof nuclear capacity – only 10 percent above the current figure. This is less than one tenththe figure predicted for the year 2000 by the International Atomic Energy Agency in 1974.

Some 76 reactors, with a total generating capacity of 17,000 megawatts, have alreadybeen retired, after an average service life of less than 17 years. As technical problems con-tinue to crop up at many older plants, retirements seem likely to accelerate. Dozens morecould be closed in the next few years, nearly cancelling out reactors coming on-line.

In Western Europe, nuclear expansion plans have been stoppped everywhere but inFrance. And even the French program is now in jeopardy due to rising public opposi-tion and the $38-billion debt of the state utility. Only six plants are under construction,and just two have been started since 1987. The United Kingdom has one final plantbeing completed. Canada and the United States together have just three under con-struction. No new plants have been ordered in the United States in 14 years, and it hasbeen nearly two decades since a plant was ordered that was not later cancelled.

Japan still has an active nuclear construction program, but there too, public oppositionis rising and utility executives are considering alternatives. In the Third World, operatingnuclear plants generate 18,000 megawatts – 6 percent of the world total. Many are wellover budget, behind schedule, and plagued by technical problems. As a consequence,only a handful of orders have been placed in the Third World in the past decade.

Recently, nuclear programs in Eastern Europe and the former Soviet Union have alsobegun to come unglued. With the arrival of democracy and with some 300,000 peoplenow being treated for radiation sickness, a torrent of public criticism has been unleashed,focusing on the failure of nuclear plants to meet western safety standards. Scores ofnuclear plants have been cancelled in Bulgaria, Czechoslovakia, Hungary, Poland, Russia,and the Ukraine. As concern about deteriorating equipment and discipline grows, pressuremounts to close those that remain.

This international trend away from nuclear power is propelled by the two serious acci-dents at Three Mile Island and Chernobyl, rapid cost escalations, and rising concern abouta healthy environment. Many people are worried about the danger of accidents and acontinuing failure to develop safe means of disposing of nuclear wastes. Opinion polls inmost countries indicate lopsided majorities against the construction of more reactors.

In addition, nuclear costs have risen to the point where nuclear power is no longercompetitive with many other energy sources. Not only coal plants, but also new, highlyefficient natural gas plants and new technologies such as wind turbines and geother-mal energy are less expensive than new nuclear plants.

The old market niche that nuclear power once held is nearly gone. Since 1988, nuclearadvocates have tried to use concern about global warming as a reason for reviving theindustry, but they have had no impact so far, as reactor orders have continued to dwindle.Nuclear power is an expensive way to offset fossil-fuel-fired power, and several hundredplants would have to be built in order to reduce carbon emissions significantly. Given thecurrent economic and political state of the industry, a major revival seems unlikely.

Box 8.4 (Continued)

3151-Ch-08.qxd 3/2/04 2:45 PM Page 258

the Foreword to their book they express aclear desire for their descriptive analysis to beinterpreted in particular ways:

We hope that this volume will meet manyneeds: that national governments will use itto compare their efforts to reduce carbonemissions or soil erosion with those for theworld as a whole; the industrial firms willuse it to compare the share of recycledmaterials in their output with that of therest of the world; and that individuals whoare helping to establish bicycle-friendlytransportation systems will be interested inthe global trends in the use of bicycles andautomobiles. More generally, we hope thatnot only environmental activists but con-cerned citizens everywhere will want to findout whether population growth is slowing,how rapidly CFC production is being phasedout, or whether the world’s nuclear arsenalis shrinking, as disarmament agreementspromised. (1992: 12)

Thus the presentation of descriptive enumer-ation of nuclear power plants may be inter-preted as being part of this environmentalistframework, and the statement in Box 8.4 that‘several hundred plants would have to be builtin order to reduce carbon emissions signifi-cantly’ suggests that the commonly used linkbetween nuclear power and environmentalpolicy is being rejected, and therefore that theauthors have an underlying opposition tonuclear power. In fact the display of these databegs many other hypotheses (for example,relating to the costs of nuclear energy and thereduced strategic need for military nuclearmaterials) which would require further inves-tigation and analysis.

Using statistics to ‘infer’

Using statistics, with their associatedexploratory manoeuvres, to perform descrip-tive functions is often regarded as the founda-tion level of analysis, to be built on bystrategies of hypothesis-testing so as to infereither difference or association. Statistics areused for inferential purposes in two main

ways. First, statements can be made about thepopulation from which a sample has beendrawn and analysed. Secondly, statements canbe made regarding the likelihood of some-thing being observed in the data set beinganalysed with no reference to a wider popu-lation. Most usage adopts the formerapproach, which has led to the innovative ideaof inference as an exploratory data analysis aswell as a confirmatory analysis.

Inference is commonly viewed as a valuableprocess of enumerative analysis in its own rightand as a necessary progression from descriptiveparameters.As Gregory (1992: 137) argues:

Useful as this [descriptive parameters] maybe in a limited sense, it often fails to ensuremaximum scientific benefit from the datacollection that has been carried out. Ideally,the items forming the sample, and the sam-ple itself, should be as representative as pos-sible of the total data set from which thesample has been selected. If this representa-tiveness can be ensured, then – within cer-tain limits – the characteristics of the samplecan be used to infer the characteristics ofthe total data set, i.e. the population fromwhich the sample has been drawn.

It is worth noting here that there is a cleardiscursive bias in favour of going further in enu-merative interpretation than simple description.Description is ‘limited’ and ‘fails’. Inference givesmore ‘scientific benefit’ and it makes the most ofthe process of data collection. Gregory is cer-tainly not alone in using such a discourse.Thelogical step in almost all accounts of statisticalmethods in geography is to progress smoothlyfrom descriptive parameters to inferred estima-tions and then to inferential tests.

Use of statistics for inference varies accord-ing to differing types of data (nominal, ordi-nal, interval, ratio) and in addition may be‘parametric’ – that is, where the sample dataare assumed to belong to an overall popula-tion of data which accord to a normalfrequency distribution – or ‘non-parametric’where no such assumption is necessary,although the results of all inferential tests

ENUMERATING

259

3151-Ch-08.qxd 3/2/04 2:45 PM Page 259

(including non-parametric ones) are held tobe ‘reliable’ only if they are applied to statisti-cally ‘unbiased’ samples (see Ebdon, 1985: 17).This does seem to be rather a big ‘if ’ in thecase of much human geography data. Indeed,spatially distributed data will often violatemany of the assumptions which need to bemet if parametric statistics are to be used forinferential purposes. This is fully realized bythe leading proponents of quantification andhas often inspired a quest for thoughtfulnessand sensitivity in the whole trajectory ofspatial statistics (see Haggett et al., 1977).

The nub of inferential tests is the establish-ment of a null hypothesis which suggests thatthe difference occurring in the sample data isthere by chance and not because of any ‘real’difference or relationship existing in the overallpopulation of data. Interest particularly ariseswhen the null hypothesis can be rejected witha degree of statistical confidence (using signifi-cance levels such as 95%, 99%, 99.9%, etc.) infavour of the alternative hypothesis. Thechi-squared,Kolmogorov–Smirnov and Mann–Whitney U-tests are common non-parametrictests used for inference, while the F-test orStudent’s t-test are equivalent parametric tests.There are a series of manoeuvres which need tobe understood in this process. Clearly, thenotion of hypothesis implies assumed reason-ing: x is different from y; x is the same as y; x isrelated to y; x is explained by y. In this way,inferred relationships or differences can easilybecome explanatory in the minds of researchers.Huck and Sandler (1979), who themselvesfavour the use of statistics for inference, com-parison and relationship, have produced a bookfull of cases where explanations for events basedon hypotheses can be challenged because of theexistence of ‘rival hypotheses’ which belie theoriginal explanations given (an illustration ispresented in Box 8.5), and Huff (1973) providesa series of examples where statistically ‘proven’hypotheses hide spurious associations betweendifferent data.

One further issue which is crucial to theuse of statistics for inference is that of thespatial characteristics of data and analysis. Ashuman geographers will usually be interestedin understanding spatial processes, the spatiallocation of the data being used becomes animportant component in the analysis. Somestatistical procedures effectively discard thespatiality of data by regarding input informa-tion as strings of numbers which can bereshuffled regardless of where they are locatedspatially. Other procedures clearly identify thespatiality of the data. For example, the pres-ence of spatial autocorrelation – ‘the presenceof spatial pattern in a mapped variable due togeographical proximity’ (Hepple, 2000: 775) –is widespread in data which are mapped outgeographically. Spatial autocorrelation can beanalysed to suggest similarities and differencesbetween adjoining areas, although care needsto be exercised in the definition of an appro-priate scale of neighbourhood unit in suchanalysis.

Considering the spatial characteristics ofdata, then, adds both to the potential utility ofstatistical inference and to the potential prob-lems encountered in their use (see Box 8.6).As with more qualitative strategies of analysis,rigorous technical expertise is the key touseful output. This is particularly so whenextending inferred differences to situationswhere statistics are used to describe therelationship between two or more variablesmeasured from a particular sample of data.The range of bivariate or multivariate tech-niques is wide but is often held to be thekey area of interpretation in geography. AsGregory (1992: 138) again puts it:

In most geographical problems, factors andcauses do not operate singly, but rather inintricate interaction with one another. It isthe unravelling of such interactions and theevaluation of the relative importance ofeach contributory element, that often pro-vides the focus and the very raison d’être ofthe geographical enquiry.


260

3151-Ch-08.qxd 3/2/04 2:45 PM Page 260

Box 8.5: Huck and Sandler’s illustration of a rivalhypothesis: newspaper advertising

Newspaper advertising I: the hypothesis

Each year, millions of dollars are spent on newspaper advertising. Obviously, the peoplewho pay for the ads feel their financial outlay is worthwhile. In other words, there is apresumed cause-and-effect relationship involved here, with the newspaper advertisingbeing the cause while subsequent increased sales constitute the effect. But do the adsreally bring about, in a causal sense, more consumer purchases? Recently, two researchersconducted a little study designed to answer this simple yet important question.

The setting for this study was a small town in northern Illinois (population about3000). The subjects were 142 female customers who regularly purchased items at a gro-cery store. A list of 28 items appeared in the local newspaper for four consecutive daysprior to the day of the study. Each of these 28 products was advertised at a reducedprice, and the purpose of the study was to determine whether this advertising made adifference in sales of these items.

Following the four days of advertising, the data of the experiment were collected.The procedural aspects of the study on this fifth day were as follows. As each subjectcame through the checkout counter, the clerk examined her purchases to see if any ofthe advertised sale items were included. If one or more of the 28 items were about tobe bought, the clerk was instructed to ask whether or not the consumer had readabout the sale items in the newspaper ad. Although several of the subjects came backto the store a second (or third) time on the data-collection day of the study, responsesfrom each consumer were recorded only for her first time through the checkoutcounter.

Results of the study indicated that all 142 subjects purchased one or more of the 28advertised items. Ninety-nine of the subjects stated that they had read about theseitems in the newspaper while 43 subjects admitted that they had not. Based on thesefigures, the researchers concluded that ‘reading the newspaper advertising seemed toincrease purchase of advertised items more than not reading the paper’ (Peretti andLucas, p. 693). To probe the data further, each subject’s socioeconomic status (SES) wasassessed by the store clerks (who knew the customers quite well) through an instru-ment called the Index of Status Characteristics. Using this SES information, a group oflower-class consumers was compared with a group of middle-class consumers. Resultsindicated that a significant difference existed between the two groups, and in thewords of the researchers, ‘advertising tended to affect lower-class consumers’ buyingmore’ (p. 693).

Based upon the data that were collected in this study, can we conclude that a cause-and-effect relationship has been established? Does newspaper advertising lead toincreased purchasing behavior? And does the alleged effect really exist more for lower-class consumers than for those in the middle class?

Newspaper advertising II: rival hypotheses

There are, in our judgment, three problems associated with this study that potentiallyinvalidate the conclusions. The first problem has to do with the question that was askedof subjects as they came through the checkout counter. The second has to do with thepersonnel used to collect these answers. And the third problem has to do with thehonesty of responses.

ENUMERATING

261

3151-Ch-08.qxd 3/2/04 2:45 PM Page 261

The main problem with this investigation relates to our inability to tell how manyof the subjects would have bought a sale item even without the newspaper ads.Surely some of the subjects made plans to visit the grocery store on what was actu-ally the fifth day of the study and read the local newspaper during one or more ofthe four preceding days. Although these subjects responsed affirmatively to thequestion about having seen the newspaper ad, this does not necessarily indicatethat the ad caused them to go to the store or to buy any of the 28 sale items. Verypossibly, they were planning to visit the store to purchase (among other things) asubset of the 28 ‘critical’ products before they saw the advertisement. Did they seethe ad before shopping? Yes. Did the ad cause them to buy advertised products?Possibly not.

To get around this problem, a different sort of question could have been asked.Instead of inquiring whether or not the consumer had read about the sale items inthe newspaper, the clerks could have asked, ‘Would you be buying these sale itemsif they had not been advertised in the paper?’ Or, ‘Did our newspaper ads aboutthese sale items cause you to purchase the items you have selected?’ However, wefeel that an alternative strategy would have produced more valid data and alsomade the questioning unnecessary. Without too much difficulty, the printed adver-tisement could have taken the form of an insert and been included in a random halfof the home-delivered newspapers. A record could have been kept concerning whogot the ads and who did not, and then the data could have been collected regard-ing the presence or absence of the 28 sale items in the market basket on day five ofthe study.

The second problem of this study relates to the personnel used to collect the data.These were the store clerks at the checkout counter. They examined each subject’s pur-chases to see if any of the 28 advertised items were present, and if any were they askedtheir question about having read the newspaper ads. However, since the store clerkswere also the ones who assessed each subject’s socioeconomic status, we wonderwhether the clerks’ knowledge of customer SES might have biased the way in whichthey interacted with the subjects as the subjects came through the checkout counter.Possibly more (or less) time was spent talking with the lower SES subjects because theclerks knew they were in this category. We feel that the personnel used to classify sub-jects into SES categories should not have been the ones to collect the data regardingthe effects of the newspaper ads.

Finally, we wonder about the honesty of responses to the question: ‘Did you readabout these sale items in the newspaper ads?’ A significantly greater number of lower-class subjects than middle-class subjects answered affirmatively. This almost gives theimpression that lower-class consumers read the paper (or at least the ads about sales)more than middle-class consumers. Maybe this is true. However, a rival hypothesis isthat middle-class consumers are embarrassed to admit that they are buying itemssimply because they are on sale. A partial (but not complete) way to get around thisproblem would have involved a cut-out coupon in the paper that had to be presentedat the checkout counter before the consumer could get the sale price.

[Note: Problems of establishing causes and effects are by no means limited to survey-based methodologies or to techniques of enumeration. Honesty is equally an issue forqualitative geographies.]

Source: Huck and Sandler (1979: 66, 184)


262

Box 8.5 (Continued)

3151-Ch-08.qxd 3/2/04 2:45 PM Page 262

1. The modifiable areal unit problem. In the analysis of aggregate spatial data, theconclusions reached might be dependent upon the definition of the spatial unitsfor which data are reported.

2. Spatial non-stationarity. Relationships might vary over space making ‘global’ modelsless accurate. Global models will also hide interesting geographical variations.

3. Spatial dependence. To what extent is the value of a variable in one zone a functionof the values of the variable in neighbouring zones? Strong spatial dependence canaffect statistical inference.

4. Non-standard distributions. Examine the nature of the data being analysed. Not alldata are normally distributed! Consider experimental methods of statistical inferenceif data have an unusual distribution.

5. Spurious relationships. Make sure that any relationships identified are meaningfuland are not caused by a variable or variables omitted from the analysis.

Source: Fotheringham (1997: 167)

Box 8.6: Problems encountered in the use of inferentialstatistical analyses in human geography

Measuring the correlation of two variables canhave a descriptive use but, more usually, tests ofsignificance are carried out to see whethervariations in one variable ‘cause’ those in another(but note here that these are tests of co-variationand not of cause – see also Chapter 9).The product–moment correlation coefficientis the usual parametric test here, and theSpearman rank correlation is the non-parametric equivalent.Correlations stem directlyfrom regression analyses where the depen-dence of one variable on another is measured.Regression statistics are used to produce trendlines but also to encompass non-linear rela-tionships, and they allow for the prediction ofrelationship between variables. Other, morecomplex multivariate techniques such asprincipal components analyses, factor analysisand cluster analyses use multiple relationalstatistics to sort out particular components orfactors in the data set and indicate howstrongly particular variables relate to thesecomponents (see Johnston, 1978, and, for anexample, Pattie and Johnston, 1993). There isoften, however, some ambiguity as to whatthese components or factors actually represent,particularly in factor analysis where data are

rotated in different ways to find the optimalconfiguration of variables or where differenttime periods of analysis are being dealt with.

The use of these multivariate techniquesoffers human geographers explanatory powerin their interpretation of data. However, thesetests can be highly manipulative and open tosignificant user error (in relation to whetherdata are really fit to meet the requiredassumptions) and to considerable user dis-cretion (for example, in relation to which ofthe many different ways to construct a regres-sion line or decide which factor-analysingtechnique is chosen in a particular case).Moreover, there is a strong sense in which theresearcher is never quite sure what is causingwhat he or she is measuring, even thoughallegedly causal explanations are being offeredby the technique concerned.As Marsh (1988:225), an exponent of correlation and regressionanalyses, points out:

It is one thing to declare confidently thatcausal chains exist in the world out there. Itis quite another thing, however, to find outwhat they are. Causal processes are notobvious. They hide in situations of complex-ity, in which effects may have been pro-duced by several different causes acting

ENUMERATING

263

3151-Ch-08.qxd 3/2/04 2:45 PM Page 263

together. When investigated, they willreluctantly shed one layer of explanation ata time, but only to reveal another deeperlevel of complexity beneath.

Some human geographers will now rejectentirely the notion of cause-and-effect chains,and others will want to stipulate that manycauses of events, circumstances and practices inthe contemporary world are not easily measur-able in numeric form, and can therefore beinvisible to these statistical techniques. Otherswill merely argue for a more careful and rigor-ous approach in pursuit of statistical explanatorypower.These issues of the authority of enumer-ation will be revisited later in this chapter.

Modelling spatialprocesses

Much of the pioneering quantitative research inhuman geography in the 1960s and 1970sstemmed from a desire to test hypotheses aboutthe spatial ordering of points, lines, areas andflows, the patterns for which were suggested byeconomic themes (for example, relating tocentral places and industrial location).The sem-inal work of Peter Haggett (1965) set the tonefor much which was to follow and, as such,emphasized points and flows and paid far lessattention to areal patterns. In a recentreview, Johnston (2000a: 127) has suggested thatthis rich heritage has led to the energy of thequantitative revolution being used in particularways, and he identifies other applications whichhave not been so widely exploited:

Geographers have made much more use ofstatistical methods for the spatial analysis ofdata than of mathematical procedures formodelling spatial processes, though this isless true of physical than human geogra-phers. The books by Bennett and Chorley(1978) and by Wilson and Bennett (1985)introduced a wide range of procedures forthe spatial analysis of spatial systems, com-plex and otherwise, but changes in thephilosophical orientation of much humangeography have seen them little used.

Johnston’s analysis of the relative under-development of spatial modelling in humangeography is an interesting one. It might beargued (cf. Fotheringham, 1998) that there aretwo main constraints on making progresswith quantitative research in human geogra-phy – the intellectual power which articulateshow spatial processes work and how they canbe modelled, and the computational power totest, refine and run such models. Initially itwas the computational power which laggedbehind. It is worth remembering that as late asthe early 1980s many geography studentswere learning about quantitative methods byusing hand-held calculators, and that even thesubsequent computerization was for a consid-erable time restricted by the boundaries ofclassic software packages such as SPSS andMinitab. Yet as computational power hasincreased, the philosophical swings towards,first, political economic theory and, secondly,the postmodern and poststructural underpin-nings of the ‘cultural turn’ have directed thegaze of increasing numbers of human geogra-phers away from an adherence to techniquesof mathematical modelling. Those who havebeen turned on by the power and thrill ofcomputing have typically forged a geocom-putational trail (see below), with the propen-sity to model being directed by the task ofdeveloping existing procedures of spatialmodelling. Thus the intellectual powerdirected at modelling spatial processes hastypically been vested in a relatively smallgroup of researchers, often with links outsidegeography (for example, with mathematicsand econometrics) as well as inside (focusing,for example, on ‘modellable’ topics such ashealth, voting, urban processes and so on).

In these circumstances it is unsurprisingthat the cornerstone of progressive attemptsto model spatial processes has been the gen-eral linear model, which effectively extendsthe relational use of statistics discussed earlierin this chapter.What we explore here, then, is


264

3151-Ch-08.qxd 3/2/04 2:45 PM Page 264

how the general linear model has been devel-oped and how interesting new directions,especially those relating to modelling thelocal, the complex and the chaotic, haveopened up yet greater opportunities forhuman geographers to practise using thisgenre of approach and technique.

Developing from thegeneral linear model

O’Brien (1992: 3) has defined generalized lin-ear models as ‘a class of linear statistical mod-els which share some important mathematicalcharacteristics’ of which three are particularlysignificant:

1. They can be reduced to a common mathe-matical form: where y = a ‘response’ com-ponent and µ = a systematic componentto be estimated from the data.

2. They investigate the behaviour of theresponse by linking it to ‘explanatoryinformation’ drawn from survey data usinga form of linear, or additive, structure.

3. They usually draw the error componentfrom exponential probability distributions.

Techniques contributing to the general linearstable of modelling include multiple correla-tion and regression, multivariate analysis ofvariance, principal components analysis, factoranalysis and discriminant analysis.These havebeen used in a range of contexts, includinggravity models, optimization models and net-work models. As Johnston (1978) admits,researchers have increasingly realized that, to aconsiderable degree, human geography appli-cations tend to violate one of the basicassumptions of the general linear model – thatconcerning independence among the obser-vations. During and since the 1970s, there-fore, there have been considerable efforts toproduce more sophisticated versions of thegeneral linear model which address these

violations of key assumptions. For example,Openshaw’s (1984) seminal work on the mod-ifiable areal unit problem (MAUP) demon-strated that aggregation and scale effects wereimportant, separate but often interdependentinfluences on observed relationships. In otherwords it is possible to get whatever correlationyou want when ecological data sets are createdfrom individual data sets or when, as forexample in the census which aggregates indi-viduals into areas, the ecological data set has tobe taken as given.

An important development was alsoachieved by Wrigley (1976), whose realizationthat standard regression procedures should notbe used on closed number sets – such as per-centages or proportions – (because such usagecan produce nonsensical results such as per-centages below 0 and above 100!) stimulatedmore sophisticated analyses using, for example,logistic regression. Other developmentsfocused on the issue of autocorrelation (Cliffand Ord, 1973), much of which is assumed tobe spatial in nature, but the fact that adjacentareas have similar values (for example, residualvalues from a regression) does not automati-cally mean that a spatial process is causal oreven involved. It is clear that some humangeographers continue to be concerned withthe shortcomings of techniques such as linearand spatial regression models, and thereforewith the shortcomings of the general linearmodel itself (Haining, 1990).

An excellent illustration of these concernsis provided by Fotheringham et al. (2000)(Figure 8.1). They use 1991 census data tomap owner-occupation levels of housing inTyne and Wear (a), noting that the highestrates are at the periphery of the county whilethe lowest rates are in the central area. Thenthey map male unemployment rates for thesame area (b), noting that the highest rates arein the centre.These problems suggest a link-age between owner-occupation and unem-ployment and a linear relationship is indicated

ENUMERATING

265

3151-Ch-08.qxd 3/2/04 2:45 PM Page 265


266

(a)

(b)

Owner-occupation (%)< 94:67

< 6.32

Male unemployment (%)< 45.77

< 5.61

3151-Ch-08.qxd 3/2/04 2:45 PM Page 266

ENUMERATING

267

Figure 8.1 Problems with linear and spatial regression models

Source: Fotheringham et al. (2000: 163–6)

(Male unemployment)2

0

10

20

30

40

50

60

70

80

90

100

2.5 3 3.5 4 4.5 5 5.5 6 6.5 7

Ow

ner-

occu

patio

n (%

)++

+++

+

+++ +

++

+++ +

++++

++++++ +

+

++ + +++

++

++

+++++ +

++

++

+

++

++ +

++

+ ++

++

+ +

+ ++

+++

++

++

+ ++

++ +++ +

+

++

+

++ +

++

++

+

++ +

+

++++++

+++ ++ +++

+++

++

+ ++

(c)

(d)

Residual value< 25.40

< −20.31

3151-Ch-08.qxd 3/2/04 2:45 PM Page 267


268

by a scatterplot of the square root of maleunemployment against owner-occupation(c). However, a map of the residuals of a regres-sion model (d) clearly suggests that theseresiduals do not vary randomly over space butseem to occur in localized clusters. Moreoverthere are sudden changes between low andhigh values.The model therefore transgressesthe assumption that each observation is inde-pendent of the others. Any assumption thatmale unemployment is a good predictor ofowner-occupation is (by this evidence)unsafe, and additional modelling develop-ments are required.While some human geog-raphers have sought these developmentswithin the general linear model, others havepreferred to develop alternative models.

One example of this search for alternatives isthe development of dynamic linear modelling,as in the field of space–time forecasting models,which Burrough (1998: 166) defines as

a mathematical representation of a real-world process in which the state of a locationon the Earth’s surface changes in response tovariations in the driving forces. Any systemfor modelling space-time processes mustinclude procedures for discretising space-time, and for the computation of new attrib-utes for the spatial and temporal units inresponse to the driving forces.

This time-series, autoregressive movingaverage has been used extensively in physicalgeography, addressing issues such as catchmenthydrology, slope stability and so on, usingPCRaster programming. However, the space–time forecasting model has been less success-fully employed in human geography, perhapsbecause its ‘black-box’ structure deals less wellwith the complex spatialities and socio-economic behaviours typically reflected insocial science issues.

A second illustration of the search for alter-native models is the move into dynamic non-linear modelling, as in the example of spectralanalysis (see Bennett, 1979). Spectral analysis isa technique which models the oscillations in a

time-series using calculations of the relativeimportance of different frequency bands. It canbe applied to regional economic cycles and todiffusion models (for example, in epidemio-logy).Here, cross-spectral analyses between dif-ferent time-series permit a multidimensionalappreciation of how, say,measles diffuses duringdifferent epidemics in different places, allowingthe development of more general predictivemodels.As with the example of dynamic linearmodelling, however, this ‘alternative’ has notdisplaced the continuing dominance of previ-ous general linear models.There are, however,two further sets of initiatives in modellingspatial processes which are demanding of veryserious attention in terms of the future ofquantitative human geographies.

Modelling ‘the local’

The first of the new directions in spatial mod-elling stems from the recent focus in spatialanalysis on identifying and accounting for dif-ferences rather than samenesses across space.An emphasis on difference has pointedresearchers towards statistical techniqueswhich are aimed at the local scale. In a reviewof these moves, Fotheringham (1997) notes anincreasing tendency to question the assump-tions inherent in regression analyses about thestationary nature of relationships over space.He discusses four types of research which aremaking a particular contribution to model-ling the local:

1. Local point pattern analysis (see, for example,Boots and Getis, 1988), which involves thedevelopment of exploratory methods fordetecting spatial point clusters.

2. Local measurement of univariate spatial rela-tionships (see, for example, Ord and Getis,1995), where the measurement of spatialassociation inherent in data permits themodelling of how the values of anattribute are spatially clustered.

3151-Ch-08.qxd 3/2/04 2:45 PM Page 268

3. Local measurement of multivariate spatialrelationships – the measurement of morecomplex relationships in their local form,using either ‘expansion’ ( Jones and Casetti,1992) or ‘multilevel’ ( Jones, 1991) model-ling methods.

4. Mathematical modelling of flows (see, forexample, Fotheringham and O’Kelly,1989), which involves the building ofspatial interaction models which can beapplied to migration, residential choice,retail and recreational behaviour and so on.

Some of the implications of these concernswith ‘the local’ can be seen in the example ofmultilevel modelling. For instance, Jones andDuncan (1996) compile a list of the necessaryrequirements for modelling approaches whichtake the notion of place seriously and whichallow for an examination of difference(Box 8.7).They argue that the use of multilevelmodelling can detect ‘effects’ or differenceswhich operate at different scales. For examplein a study of immunization uptake amongchildren in Britain, an important policy-relatedfactor is that general practitioners are rewardeddifferentially by reaching targets based onimmunizing certain percentages of children.Jones and Duncan (1996: 94) argue that ‘multi-level models, by modelling at the individual andclinic levels simultaneously, are able to get somepurchase on which practices are performingwell given their client groups, and thereby per-mit the development of a contextualised mea-sure of performance’. This type of modelling(which has a range of applications in health,education and political geographies) doesseem to offer considerable scope for a morespatially sensitive form of spatial analysis infuture quantitative geographies.

Modelling ‘the chaotic’

The second set of initiatives draws on theoriesof chaos and complexity. Theoretically, there

has been an upsurge of interest from humangeographers in ideas of chaos (see Prigogine,1984).Wilson’s (1981a;1981b) research showedthe nature and importance of the constraints tosolutions of flow patterns at both the originand the destination of those flows. He arguedthat an entropy-maximizing approach was a bet-ter mechanism for modelling constrained flowsthan the unconstrained regressions that had beenused previously.With the acknowledgement ofconstraints has come an awareness of the poten-tial for chaos.Mathematically,chaos suggests rela-tionships between variables which demonstrateno sense of order in either magnitude, space ortime. Such attributes have a direct application inphysical geography (for example, in the model-ling of some aspects of climate change) but, asyet, are largely untapped in human geographymodelling contexts. Even so, ideas of the chaoticappeal strongly to those human geographerswho are interested in concepts of deconstructionand difference, and so there are prospects forquantitative initiative in this area.

One example of the complexity of the sub-chaotic in human geography comes fromBatty and Longley’s (1994) modelling of frac-tal cities. Fractal characteristics are reflected inobjects with an irregular spatial form, theirregularity of which is repeated geometricallyacross different scales. As Batty and Longley(1994: 4) suggest:

Cities have quite distinct fractal structures inthat their functions are self-similar acrossmany orders or scales. The idea of neigh-bourhoods, districts and sectors inside cities,the concept of different orders of transportnet, and the ordering of cities in the centralplace hierarchy which mirrors the economicdependence of the local on the global andvice versa, all provide examples of fractalstructure which form the cornerstones ofurban geography and spatial economics.

This acknowledgement of alternative geo-metries, or even in the case of chaos theorydisordered non-geometries, is opening upinteresting new lines of inquiry for spatial

ENUMERATING

269

3151-Ch-08.qxd 3/2/04 2:45 PM Page 269


270

Contextual differencesModelling must recognize that ‘people make a difference and places make a difference’so that people are not reduced to statistical aggregates and places to generalizations.In a contextual model of health, the setting in which the behaviour literally takes placeis seen as crucial for understanding. For example, in studying tobacco consumption weneed to consider simultaneously the micro-scale of individual characteristics and themacro-context of local cultures in which people live.

Place heterogeneityThe effects of context can potentially be very complex, with relationships varying in dif-ferent ways. For example, there may be places where people of all ages have a highconsumption of tobacco, whereas in others it is the young who have a relatively highrate, but in others again it is the elderly. This age differential may also be accompaniedby complex differences for other individual characteristics such as gender and class. Ifthis is the case, it is not possible to have a simple ‘map’ of smoking for consumption willvary according to who you are in relation to where you are.

People/place interactionsModelling must also take into account the possibility that individual differences may’interact’ with context. For example, a person of low social class may act quite differentlyaccording to the social class composition of the area in which he or she lives. To put it inanother way, between-place differences need to be examined in relation to the socialcharacteristics of individuals in combination with social characteristics of places.

Individual heterogeneityIt is not only the higher-level contexts that must be allowed to vary in their effect, forindividual differences may be too complex to reduce to an overall ‘average’. Thuspeople of low social class may not only smoke more on average but they may also bemore (or less) variable in their consumption. Heterogeneity between people as well asbetween places must be expected and modelled.

Changing people, changing placesChanging contexts and changing behaviour are important too, so that the approachmust be able to handle longitudinal data and time (as well as place) as context.

Inter-related behavioursIt is important that differing but not unrelated behaviours are not considered sepa-rately but modelled together. Any meaningful model of health-related behaviourwould need to consider smoking alongside alcohol consumption, for example.

Quality and quantityThe behaviour and actions of individuals have both a qualitative aspect (occurrence –does it occur?) and a quantitative element (amount – how much, how often, howmany?). Both these need to be considered simultaneously. For example, the variableswhich are closely related to whether a person smokes or not may reveal nothing aboutthe number of cigarettes smoked. There may be places where there are a few peoplewho smoke, but those who do so smoke heavily; an average figure would be verymisleading.

Box 8.7: Requirements for spatial modelling whichtakes ‘place’ and ‘difference’ seriously

3151-Ch-08.qxd 3/2/04 2:45 PM Page 270

Box 8.7 (Continued)

modelling. However, as is evident from thereflexive acknowledgement of ‘nets’,‘ordering’,‘hierarchy’ and so on above, the driving forcefor many existing applications remains thedesire to model the order of disorder and thecapacity of GIS to manipulate vast quantities ofdata into these models. As is discussed in thenext section, the acknowledgement of differ-ence in more recent spatial models tends to bedriven by the ability to compute large data setsrather than by any rapprochement between thephilosophical underpinnings of spatial analysisand the qualitative cultural turn.

Geocomputation

From the early 1980s onwards, the availabilityand affordability of increasingly sophisticatedand powerful computing technology havecreated a whole new world for enumerativeanalysis in human geography. Not only didthese technological changes aid the continu-ing use of spatial modelling but they alsoprompted the creation of geographic infor-mation systems (GIS) which set up integratedcomputerized processes for handling andanalysing geographic data (see Chapter 7).Muchhas been written about the origins and develop-ment of GIS2 but, in essence, what began asthe computerization of mapped informationvery quickly became a toolbox of proceduresfor the input, storage, manipulation and out-put of geographical data.

As Martin (1997: 216) suggests, the earlydevelopment of GIS was quickly wrapped upin the needs of commercial users:

In commercial use, one of the most importantroles of GIS is as an organising structure forextremely large quantities of operationalinformation. Such inventory applicationstend to require compact data storage andrapid retrieval, but do not place particularlyheavy analytical or statistical demands onthe system. For this reason, much of thedevelopment effort in GIS (and the corre-sponding literature) has been directedtowards ways of holding and queryinggeographical data more efficiently.

Two main spatial organizing systems emerged(Figure 8.2). Raster systems work on a frame-work of rectangular cells and input informa-tion on the basis of each cell. This relativelycontinuous modelling of space has been

ENUMERATING

271

Multiple contextsIt is likely that there will not be a single context but many. For example, smokers may beinfluenced not only by the milieu in which they live but also by the work environment.

Source: Adapted from Jones and Duncan (1996: 81–2)

Continuous object:population densitysurface

Area object:parliamentaryconstituency

Line object:road location

Point object:postcode location

Example ofspatial object

Vectordata

x

Rasterdata

Figure 8.2 Vector and rasterdata structures

Source: Martin (1997: 217)

3151-Ch-08.qxd 3/2/04 2:45 PM Page 271

favoured in physical geography applications.Vector systems use points, lines and areas toinput the location of discrete objects usingmap co-ordinates. Social science applicationsof GIS have generally found vector systems tobe more useful to their needs.

From these simple beginnings, the techno-logy of GIS has matured, providing the basisfor sophisticated modelling of spatial distri-butions, flows, networks and hierarchies.However, as Longley (1998: 3) has stressed,GIS provides the environment in which spa-tial processes can be explored, but not theguiding principles for that exploration: ‘Initself, GIS provides little guidance as to howbest spatial problems might be solved. It isclear that big technical and epistemologicalquestions remain about the ways in whichselected facets of geographical reality are dis-carded or emphasised in the process ofmodel-building.’

Thus, the computer has been transformedfrom being an adjunct to the research envi-ronment to being the research environment(Goodchild and Longley, 1998). It is thecoupling of the input, throughput and outputcapacities of GIS with analogous models foranalysing, visualizing and modelling dataassociated with spatial patterns and processesthat has brought about the most significantapplication of GIS in human geography.Theterm ‘geocomputation’ has been used toconvey this coupling of technology and ana-lytical purpose (see Longley, et al., 1998b).Here the intensive computational environ-ment of GIS is used for research-led problem-solving purposes and thus creates intellectualspace for continuity with the longstandingconcerns of quantitative geography. SoMacmillan (1998), for example, regards geo-computation as being founded on the ana-lytical philosophies and scientific methodswhich have marked the established tradi-tions of quantitative human geography.Thisview is not universal, however. For example,

Couclelis (1998) finds geocomputation tobe lacking in systematic connection withspatial theory or, indeed, with computa-tional theory – she regrets that the GISenvironment has produced a grab-bag oftools rather than a more ‘revolutionary’unifying epistemology.

The reasons for these different viewpointsare clear. On the one hand, most users of GISenvironments will, implicitly or explicitly,adopt a ‘scientific’ research design involvinghypothesis-testing using data analysis, andemploying rational, ordered reasoning.However, the GIS world has become popu-lated by ‘an uncoordinated proliferation of dig-ital datasets (many of which are collected byunorthodox or even profoundly unscientificmeans)’ (Longley, 1998: 7).The rich diversity ofdigital representation in GIS can thus lead toepistemological conflict with the ‘answers’,‘truths’ and ‘spatial processes’ which they helpto generate.Thus the fruits of geocomputationcan prompt debate about whether quantitativetraditions are being enhanced, or diluted, inGIS environments.We illustrate these claims inthree areas of applied geocomputation.

GIS and spatial analysis

In another of his reviews of quantitativehuman geographies, Fotheringham (1998)sees the role of geocomputation as a crucialguiding force for innovation. In his view,computation is driving the form of analysisundertaken rather than just providing aconvenient environment for applying inde-pendent techniques. One benefit of geocom-putation is therefore an emphasis on variationand differences rather than similarities. Theprocesses and relationships under analysis arenot regarded as homogeneous, and the geo-computational ability to analyse small-scalespatial variations has thus enhanced the abilityto study differences in broad spatial and social


272

3151-Ch-08.qxd 3/2/04 2:45 PM Page 272

processes. As a result, geocomputation hasboosted exploratory (rather than confirmatory)techniques and provides a mainly inductiveinstrument by which the representation ofdata can lead to new hypotheses about theprocesses which produce the data.

Fotheringham (1998) outlines a number ofimportant geocomputational applications inhuman geography, of which two are particu-larly relevant here.

SPATIAL INTERACTIONS3

Geocomputation has permitted unidirec-tional and bidirectional interactions betweenthe GIS and spatial analytical routines. Thus,for example, data derived from a GIS can beused to undertake a statistical analysis such asregression, and the resulting statistics (forexample, residual values) can be returned tothe GIS to create fuller maps (see, for example,Figure 8.1). Such integration has becomeincreasingly dynamic, creating a seeminglyseamless movement between mapped dataand analysis.The use of Windows software, forexample, permits the review (or ‘brushing’) ofscatterplotted data in one window anda simultaneous map-referencing of thosepoints in another.These improved exploratorytechniques do not, of course, guarantee anygreater insights into the spatial processes con-cerned, but the ability to move between ana-lytical and mapping software in Fotheringham’s(1998: 286) view produces ‘a reasonably highprobability of producing insights that wouldotherwise be missed if spatial data were notanalysed with a GIS’.

Another geocomputational advance in spa-tial analysis has been in addressing the modi-fiable areal unit problem.The use of areal dataat different scales (for example, census data atenumeration district, ward and district levels)is often problematic because the results ofanalyses can be sensitive to the particular spa-tial units in which the data are presented.Theflexibility provided by raster- and vector-based

systems has presented one route (but not theonly route) by which those problems can beaddressed and overcome.

ARTIFICIAL INTELLIGENCE4

Geocomputation has prompted human geo-graphers to begin to work with artificial intel-ligence systems, the ultimate aim of which is‘to produce computer programs which cansolve problems for which they have notbeen programmed’ (Fotheringham, 1998: 289).These developments have involved applica-tions of heuristic searches, using point patternanalysis to identify particular clusters of pointsin particular locations, and neurocomputing,where neural nets are trained to represent apriori connections in simulation processes.There has also been development of evolu-tionary computing techniques in order torepresent the growth of objects (for example,cities) (Batty and Longley, 1994).

GIS and geodemographics

Geocomputation has also been at the core ofan important phase of applied modelling oflocational analysis in human geography overthe last two decades or so. Using longstandingideas about locational modelling alongsideaspects of operational research and computa-tional theory, a group of human geographershas set about using GIS environments to ful-fil the commercial needs of a range of businessand service providers (Longley and Clarke,1995). The success of this approach is illus-trated by GMAP – a consultancy companyestablished by the University of Leeds in 1987and subsequently taken over into majorityownership by the American marketing com-pany, Palk, in 1997. Members of GMAP andothers re-pioneered geographical expertise inmodelling the best locations from which tomaximize sales to particular target groups. Inso doing they built up a more general exper-tise and interest in geodemographics.

ENUMERATING

273

3151-Ch-08.qxd 3/2/04 2:45 PM Page 273

Category A: ‘thriving’

1 Wealthy achievers, suburban areas2 Affluent ‘greys’, rural communities3 Prosperous pensioners, retirement areas

Category B: ‘expanding’

4 Affluent executives, family areas5 Well-off workers, family areas

Category C: ‘rising’

6 Affluent urbanites, town and city areas7 Prosperous professionals, metropolitan areas8 Better-off executives, inner-city areas

Category D: ‘settling’

9 Comfortable middle-agers, home-owning areas10 Skilled workers, home-owning areas

Category E: ‘aspiring’

11 New home-owners, mature communities12 White-collar workers, multi-ethnic areas

Category G: ‘striving’

13 Older people, less prosperous areas14 Council house residents, better-off homes15 Council house residents, high unemployment16 Council house residents, greatest hardship17 People in multi-ethnic, low-income areas

Source: Clarke (1999: 580)

Box 8.8: ACORN geodemographic classification

Clarke (1999) suggests a number of method-ological advances relating to geodemographics.First, the location of customer types hasrequired a profiling of geographical areas interms of the key customer segment typeswithin them, using classifications such asACORN (Box 8.8).During the 1990s, increas-ingly sophisticated systems linked populationdemographics with commercially obtainedlifestyle information. Secondly, GIS environ-ments have also been used to assess retail sitelocation through the geocoding of populationswithin catchment areas and the calculation of

demarcated (‘buffered’) travel times around thestore location. Population catchment estimatescan then be translated into sales estimates, tak-ing account of other competing stores in thearea. Such calculations are not without prob-lems, however, with both the definition ofcatchment areas and the modelling of competi-tion being subject to considerable debatewhich, in turn, has sponsored new develop-ments in interaction and flow modelling.

In some ways geodemographics raises inpractice the theoretical tensions (discussedearlier in this chapter) between the scientific


274

3151-Ch-08.qxd 3/2/04 2:45 PM Page 274

philosophy of research and the sometimesuncoordinated proliferation of digital datasets. Many of the commercially derived dataimported into geodemographic modelling arepatchy, incomplete and therefore ‘unscientific’in pure terms. It also rests on the (somewhatarguable) theoretical assumption that ‘you arewhat you consume’. Nevertheless, there islittle doubt that geodemographics have assumedan important and applied role in contempo-rary analysis by enumeration.

Visualization

Geocomputation has also used GIS to explorenew ways of displaying, viewing and ‘reading’quantitative data that represent more complexspatial processes and systems. To begin with,

human geographers have begun to developnew forms of cartography (Dorling andFairburn, 1997). For example, customizedpopulation cartograms have been used to cre-ate more realistic and reliable understandingsof demographic distributions (Dorling,1995a), and new diagrammatic ways of repre-senting the complex relationships betweenvariables have enhanced the graphical displayof data. Figure 8.3 illustrates a mapping of theoutcome of the 1997 General Election inEngland. Instead of a conventional spatial pat-terning by constituency area, constituenciesare represented by circles, the size of which isproportional to the area of the constituency.The visualization of electoral data in this casepermits a holistic view of the spatial distribu-tion of party control and the types of

ENUMERATING

275

Labour

Liberal

Conservative

Independent

Figure 8.3 The outcome of the 1997 General Election in England

Source: Upton (1999: 401)

3151-Ch-08.qxd 3/2/04 2:45 PM Page 275

constituencies represented. Figure 8.4 uses theouter circle to display 18 values, three each forsix variables relating to voting in that election.These values are then linked using a cobwebdiagram, with the thicker lines correspondingto stronger relationships. The room forcomputer-driven experimentation here islimitless, with multivariate data being repre-sented by anything from triangularization ofspace to the variable features of a frowning orsmiling face situated at each data point.

The cognitive processes inherent in thesenew cartographies are clearly underpinned byparticular statistical modelling of data.Visualization becomes much more difficultwhen dealing with hypervariate data than

with univariate or bivariate data and, asFotheringham (1999: 604–5) stresses: ‘therewould appear to be a trade-off between pre-senting a clear image…and understandingwhat that image means in terms of the originaldata.The problem is not new in multivariatedata analysis, and the same issue arises inprincipal components analysis, factor analysisand discriminant analysis.’ Visualization,then, merely reflects the wider epistemo-logical and interpretative issues of the tech-niques which underlie it.

Two other initiatives in visualization arealso significant in this context. First, the prin-ciples of the new cartography have focusedresearch attention on the possibilities


276

Low

Low

Low Hi

Hi

Hi

Low

Mid

Mid

Mid Mid

Mid

Area

Non-white %

Con %

Council %

Mid

Hi

Hi

Hi Low

Low

Manual %

Retired %

Figure 8.4 A cobweb diagram of sociological variables influencing voting

Source: Upton (1999: 412)

3151-Ch-08.qxd 3/2/04 2:45 PM Page 276

of developing realistic virtual spaces in whichthe data-rich GIS can be encountered indifferent ways. Batty et al. (1998), for example,have shown how immersive and remote envi-ronments can be built up using GIS such thatsingle and multiple users can interact in avariety of ways. They illustrate these ideasusing four of their different developmentalprojects:

a multi-user Internet GIS for London withextensive links 3-D, video, text and relatedmedia; an exploration of optimal retaillocation using a semi-immersive visualisationin which experts can explore suchproblems; a virtual urban world in whichremote users as aviators can manipulateurban designs; and an approach to simulat-ing such virtual worlds through morpho-logical mapping based on the digitalrecord of the entire decision-makingprocess through which such worlds arebuilt. (1998: 139)

There are obvious commercial and educationadvantages to these kinds of virtual visualiza-tions. The modelling aspect is again anenhancement which is only as effective as theunderlying models and techniques permit itto be.

The second initiative which displays thepotential of geocomputational visualizationis the development of ‘wearable’ computing.Clarke (1998) has shown that the availabilityof fully independent mobile computing anddevelopments in human–computer inter-action are leading to important new terri-tories of visualization. Hyperinteractivedevices will permit the conversion of physi-cal movements, sounds and perhaps eventhoughts into digital input streams. Returndisplays might be envisaged which simulatevision and sound. Once again we see thepotential for technology to be the drivingforce of science. Wearable computing couldopen up a whole new suite of questions tobe answered through geocomputation.Equally, the potential for the loss of privacy

or even, worse, invasions of the humanconsciousness is also apparent.As Clarke (1998:136) speculates:

I dread the day when I am woken from asound sleep by a noisy, flashing advertisementprojected onto my retina urging me todownload a new free Web-browser, onethat I cannot turn off without focussing ona dark grey ‘Decline’ button hovering at thefar range of my peripheral vision. Never-theless, as a Professor, I look forward toteaching students directly while walking inthe mountains or on the beach, and tositting again at a desk devoid of a massive,noisily fan-cooled, cathode ray tube.

The authority of numbers?

The questioning of science

Throughout this chapter we have attemptedto convey a sense of the growing sophisticationof quantitative analytical techniques in humangeography. The use of such techniques nowranges from relatively straightforwardexploratory analysis using descriptive enu-merations and statistics, recently aided by newand interesting forms of visualization in GISenvironments, to highly complex modelling,including sustained attempts to discover andexploit the spatiality of data. Earlier in thechapter we emphasized the potential forvarying interpretation, and even misrepre-sentation, when using the simplest ofdescriptive techniques. Such potential diffi-culties in no way undermine the legitimacyof descriptive analyses of enumeration, butthey do point to the likelihood that thedeployment of ever more sophisticated sta-tistical modelling will be accompanied by anincreasing risk that the underlying enumera-tion can be lost sight of during iterativemathematical manipulation.

The principal challenge to any presumedsupremacy for quantitative methods in humangeography has, however, taken the form of a

ENUMERATING

277

3151-Ch-08.qxd 3/2/04 2:45 PM Page 277

critique of the positivist philosophical roots ofquantification. That is, it is not the use ofnumbers per se which has been objected to;rather the framework of thinking which dictatesthe assumptions and interpretative directionswhich so often characterize numerical analy-ses. Putting this argument simply, the princi-pal language of science is mathematics;science in human geography has beenunderpinned by the flawed philosophy ofpositivism; therefore, mathematics carrieswith it the flaws of positivism and should berejected accordingly (see Barnes, 1996). Sucha critique is inexorably linked with theadoption of other philosophical frameworksin human geography. Thus Cosgrove (inGregory, 2000d) regards the modelling ofspatial science, with its privileging of abstrac-tion and generalization, as an intrinsic expres-sion of the modernist impulse. It followsthat the practice of quantification is seen asincapable of responding to the challenges of‘difference’ posed by poststructuralist philoso-phies. Dixon and Jones (1998a) follow up thisontological challenge with an epistemologicalone. Poststructuralism, they note, calls intoquestion any assumption that the ‘real’ can beseparated from the ‘represented’ – an assump-tion on which the epistemology of modelbuilding relies.

For Sibley (1998), the nub of the questionis about the search for order. He notes thatpractitioners of spatial science and GISexhibit considerable anxiety about thenotion of disorder and base their researchapproach on searching out order. Even themore recent turn towards enumerating dif-ference, as noted earlier in this chapter, hastended to emphasize the order of disorder.Sibley argues that the turn towards post-structuralist philosophies has posed signifi-cant questions about this emphasis on order.He especially highlights that practiceswhereby human geographers claim to adoptvalue-neutral approaches and ‘just look’ for

order in data are not innocent practices.Indeed, the radically non-innocent nature ofthe search for order raises unsettling implica-tions for the process of spatial modelling.Theupturn of these critiques has been a signifi-cant shift by many human geographers awayfrom the quantification of spatial science,thereby reinforcing a quantitative–qualitativedivide.

It is worth noting at this point that muchof this critique of the philosophy underlyingquantitative practices has failed to establishwhether geographers using quantitativemethods were ever actually wedded to posi-tivism (with its attendant philosophical andsociocultural baggage) as a framing philoso-phy, and even if they were whether posi-tivism was the main justification for their useof mathematical analysis. It may well be,for example, that quantitative practicesrely more on broad philosophies of empiri-cism than on the more specifically focusedphilosophy of positivism.This very significantpoint opens up a wider debate about thesocial and cultural contexts in which quanti-tative techniques are used – a debate whichhas been particularly drawn out by TrevorBarnes (1998) and Les Hepple (1998).Barnes argues that correlation and regressionanalyses – kingpins of the general linearmodel – emerged from very particular socialand cultural contexts, and that these con-texts were embued in the very nature of thetechniques:

Correlation and regression were the first ofthe inferential statistics, but they too weredesigned to advance a particular social andideological agenda, eugenics. As a resultthey were not a set of value free, neutraltechniques, but carried with them theassumptions, conditions, and wider prob-lematics of the intellectual movement inwhich they arise. As correlation and regres-sion analyses became black boxes, theymoved out of biometrics and into otherfields, including eventually human geogra-phy in the mid 1950s… Because statistics is a


278

3151-Ch-08.qxd 3/2/04 2:45 PM Page 278

social and cultural product, its frameworkcannot necessarily be transported to a dif-ferent context without inconsistenciesoccurring. (1998: 220–1)

We should note here that Barnes is not claimingthat regression analysis is somehow an inher-ently racist technique. However, he is empha-sizing that regression originated in a particularintellectual context (eugenics) and thereforecarries with it certain background assump-tions (the need for independent sampling,large numbers of samples, etc.) which ‘facili-tate control and power over certain bits of theworld’ (1998: 220).

Hepple debates with Barnes the extent towhich the emergence of regression and statis-tical inference has been constrained by thesespecific sociocultural assumptions and con-texts. He argues that the technical form ofthese analytical procedures was not signifi-cantly constrained by specific biometricassumptions and contexts. More widely, heargues that ‘quantitative and cultural approachesare both necessary, shedding light in differentways on social life’ (1998: 232).

Ground Truth

These questions about the connectionsbetween philosophy and practice in quantifica-tion, and about the way in which quantitativeanalytical techniques carry with them impor-tant sociocultural baggage, prompted intensi-fied debate in the era when GIS came to thefore.The early exchanges tended to be funda-mentalist in tone. For example, Taylor (1990:211–12) chastens quantifiers in human geogra-phy who have ignored and by-passed legitimatecritiques of their positivist underpinnings andtheir empiricist superstructure and sought tocapture the rhetorical high ground of the sub-ject by swapping ‘knowledge’ for ‘information’:

Knowledge is about ideas, about puttingideas together into integrated systems of

thought we call disciplines. Information isabout facts, about separating out a particu-lar feature of a situation and recording it asan autonomous observation … The positivist’srevenge has been to retreat to information,and leave their knowledge problems – andtheir opponents – stranded on a foreignshore. But the result has been a return ofthe very worst sort of positivism, a mostnaïve empiricism.

In reply, Openshaw (1991: 626) offers asomewhat imperialist vision of GIS:

GIS can provide an information systemdomain within which virtually all of geogra-phy can be performed. GIS would emphasisean holistic view of geography that is broadenough to encompass nearly all geogra-phers and all of geography. At the sametime it would offer a means of creatinga new scientific look at geography, andconfer upon the subject a degree of cur-rency and relevancy that has, arguably,been missing.

An important clarification of a more sophisti-cated critique of GIS came with the publica-tion of Ground Truth edited by John Pickles(1995). Ground Truth set out to tackle threesets of issues:

1. The impact of geography on the transfor-mation of data-handling and mappingcapabilities provided by GIS.

2. The ideas, social practices and ideolo-gies which have emerged with thesenew forms of data-handling and spatialrepresentation.

3. The situating of GIS as a tool of, and anapproach to, geographical informationin the context of late twentieth-centurytransformations of capitalism.

GIS, then, is viewed in Ground Truth as

a tool to protect disciplinary power andaccess to funding; as a way of organisingmore efficient systems of production; and asa reworking (and rewriting) of culturalcodes – the creation of new visual imaginar-ies, new conceptions of earth, new modalities

ENUMERATING

279

3151-Ch-08.qxd 3/2/04 2:45 PM Page 279

of commodity and consumer, and newvisions of what constitutes market, territoryand empire. (1995: viii)

Interestingly, some of those seeking to debateGround Truth immediately complain aboutthe non-comprehensibility of its language.For instance, Flowerdew (1998: 299), withreference to the above extract among others,reflects that ‘With considerable effort, I canwork out approximately what these sentencesmean (I think) but it is hard to imagine GISboosters, who perhaps need to hear theunderlying message, struggling to understandthem, let alone be convinced’.

This focus on language can be compared,somewhat ironically, with the list of reasonsfor the ‘downturn’ in quantitative geographyin Box 8.2, and specifically with the claimstherein that the ‘difficulty’ of quantitativegeography ‘deters established non-quantitativeresearchers from understanding the nature ofdebates’. The coding of quantitative geogra-phy as ‘difficult’ and of Ground Truth as ‘over-complex’ seems here to reflect ideologicalrepresentations and could easily be reversed.More importantly, the issue of language seemsto provide an excuse for some quantitativegeographers (Flowerdew clearly excepted) toeschew any meaningful engagement with theGround Truth project.

In Ground Truth there is a series of deepconcerns over the impacts of unmediatedtechnical practices in GIS. Three examplesillustrate these concerns. First, Curry (1995)argues that human geographers need to takeserious account of conflicting stances towardsethical issues in GIS.There are ethical concernsrelating to the expertise involved. GIS usersoften know how to operate the system butnot how it was conceived or how it worksin detail. User expertise might therefore bethought of as meeting ethical requirements,despite the fact that the systems in use maywell have been developed in the entirely dif-ferent ethical contexts of military and intelligence

use (see Smith, N., 1992). There are alsoethical concerns about how the analysis oflarge data sets tends to regard people-places asan appropriate category for interpretation,thus contributing to the dissolution of realindividuals. Clearly such concerns need alsoto be linked with an opposite risk in whichGIS can be used (explicitly or implicitly) forsurveillance purposes, thus perpetrating inva-sions of privacy.

Secondly, Goss (1995a: 161) reflects on theability of GIS to manipulate behaviour.As henotes, GIS

reduced the complex multidimensionalcharacter of human identity to a set of cor-related variables which are then aggre-gated into abstract constructs or clustersbased on place of residence; these clustersare then reified into ideal types of ‘con-sumer neighbourhoods’ used to predict con-sumer behaviour, which is in turn reducedto a simple scale representing the likelihoodof purchasing a particular commodity.

As Goss points out, GIS can thus become partof a self-fulfilling prophecy as social identity isdefined in terms of particular clusters of con-sumer goods consistent with idealizedlifestyles, and commercial interests will thentry to sell these identities back to consumersin the form of complementary consumergoods. Not only do such processes go a longway towards constituting social subjects butthey can also direct the gaze towards particu-lar (wealthy, high-status) subjects who are ofgreater commercial interest, thereby divertingthe gaze from other (impoverished, sociallyexcluded) subjects.

Thirdly, Harris et al. (1995) question therelationship between GIS and democracy.They argue that the use of GIS for research,planning and project assessment, particularlyin developing nations, has generally turnedout to be a ‘technicist legitimisation of thehistorical power relations associated with tra-ditional developmentalism’ (1995: 196). Amore radically democratic alternative is to


280

3151-Ch-08.qxd 3/2/04 2:45 PM Page 280

establish a participatory application of GIS inwhich an appreciation of local knowledge,community needs and specific social historiescan be incorporated into the developmentalprocess. Harris et al. (1995) report on theirexperience of attempting to introduce justsuch a participatory application in the contextof post-apartheid rural reconstruction inSouth Africa.Their conclusion is stark:‘In tra-ditional South Africa, it is likely that GIS willreinforce traditional market-based and techni-cist approaches to policy formation to thedetriment of alternative restructuring strate-gies: this process is inherently undemocratic’(p. 217).

The control over data collection, analysisand interpretation at the scale now offered byGIS environments remains heavily top-downand elitist.There are ways of imagining a moredemocratic and localized set of inputs but, inthe pragmatics of GIS use, in both the publicand private sectors, such democratic aims arelikely to be superseded by technocratic, com-mercial and bureaucratic concerns.

Quantitative responses

The responses to issues raised in Ground Truthand analogous debates suggest three strategiesfor taking forward practices of enumeration inhuman geography. The first – to ignore theemergent critique altogether – is reflected byClark’s (1998: 305) extraordinary use of mas-culinist war discourses to describe the GroundTruth attack on a productive profession ofquantifiers:

The attack on the complacency and flawedthinking of GIS launched by Pickles (1995) inGround Truth is an academic Blitzkrieg ofunusual power and ferocity. It strikes at thecore of a productive profession by usingweapons (arguments) from another world,in much the same way that the Panzer divi-sions rolled East to impose a new order onEurope’s agricultural and industrial heart-land. Both attacks were pre-emptive and

were fuelled by a mission rather than ameasured and reflective desire to contemplatealternative models. However, while Blitzkriegappeared initially to be immune to resis-tance, it is amazing that this most radicalcritique of GIS social implications has failedto raise even a whimper, let alone an enragedoutcry, from most of the practitioners ofGIS. Indeed, in a very real sense the attackmounted on Ground Truth has failed com-pletely to engage the enemy! The tirade hasgone unheard not because the fundamentalmessage was irrelevant or wrong butbecause the messenger spoke a foreign lan-guage and spoke in an empty room.

This kind of viewpoint suggests that anadherence to a positivist philosophy of geo-graphy will valorize objectivity as a goal andtake a benign view of technology as the meansof reaching that goal.

A second response might be to adhere toquantitative methodologies on grounds otherthan a philosophical attachment to positivism.Many quantitative human geographers wouldnow agree with Flowerdew (1998: 292) whenhe suggests that while ‘quantitative geographyhas lost and largely abandoned the philoso-phical debate, quantitative work is still impor-tant’. Earlier in this chapter we discussedFotheringham et al.’s (2000) view that thereare important differences in philosophy acrossthe protagonists of quantitative geography.For example, the philosophical tenets of prag-matism – what counts as knowledge is deter-mined by its usefulness – were developed byRorty (1979), who proposed a conversationbetween the mind and the world with nofixed end-points, strict rules or necessarylogic. Pragmatism can and does lead to a useof quantitative methods to provide theanswers required by particular end-users.Spatial analysis and modelling thereforeprovide ways of meeting the demands of thecommercial marketplace, the governmentdecision-maker and the research council anx-ious to be seen to be relevant to the needs ofthe nation. In these contexts numeric analyses

ENUMERATING

281

3151-Ch-08.qxd 3/2/04 2:45 PM Page 281

are perceived as useful (and sometimes evenvital), so applied human geographers supplytheir information in this form.

It is difficult to know whether quantitativehuman geographers have really given up onpositivistic philosophies, despite claims tohave set aside the philosophical debate. Forexample, Flowerdew himself is reluctant toveer from ‘value-free’ research (1998: 293) andsuspects that most GIS scholars and practi-tioners are reluctant to embrace alternativesocial theories which might, of course, pointout the links between ‘value-free’ research andpositivist philosophy. The suspicion remainsthat although some quantifiers recognize theplace for other, more qualitative, forms ofresearch (see, for example, Jones and Duncan,1996), most will not have rejected the princi-pal facets of scientific method in their ownwork, even if they feel the need to explainthese in terms of ‘the best way of doingthings’ rather than of positivist philosophies.

A third response has been to incorporateparts of the critique into new forms of quan-titative research. An important illustration ofthis comes from the Radical Statistics Groupwhich was formed in 1975 as part of the rad-ical science movement associated with theestablishment of the British Society for SocialResponsibility in Sciences.The group reflectsconcern about the mystifying use of technicallanguage to disguise social problems as tech-nical ones; the lack of control by the commu-nity over the aims of statistical investigations,the way these are conducted and the use ofthe information produced; the power struc-tures within which statistical and researchworkers are employed and which control thework and how it is used; and the fragmenta-tion of social problems into specialist fieldsobscuring connectedness (http://www.radstats.org.uk). Such concerns have prompted cam-paigns for better ‘official’ statistics and a betterunderstanding of statistics. The group isparticularly interested in demystifying the

selective use of social statistics to support thepolicies and status of those in power.They alsoaim to introduce the kinds of statistics needed‘to understand and improve life’ (Dorling andSimpson, 1999b; see Box 8.9). The approachreflected by the Radical Statistics Groupacknowledges that statistics are a product ofsomething and therefore are never neutral.Such statistics

cannot be ignored nor can they be substan-tially changed unless society itself ischanged. People who wish to make statisticsmore open … more democratically linked tothe priorities of local and other communi-ties, less dominated by government andtrade, are allies of anyone who wants tochange society in these ways. (Simpson andDorling, 1999: 415)

Here, then, is a very significant group ofquantitative social scientists who advocatethe explicit introduction of political valuesinto their work.This permits them to addressissues relating to ethics and democracy aspart of their research design. It may alsomake them more amenable to more qualita-tive practices which offer ‘evidence’ of thepolitically directed ‘targets’ of their research.

Individual responses

We began this chapter by emphasizing thattechniques of analytical enumeration providepotential powerful and useful ways of practis-ing human geography.We also suggested thatquantitative and qualitative techniques neednot be viewed as discrete choices. It will beevident from the account presented in thischapter that a deep and sometimes uncom-municative divide has often separated practi-tioners of the quantitative and the qualitative,with each questioning the philosophy, lan-guage, assumptions and usefulness of theother. Many will wish to unsettle some of themore dichotomous views of quantification


282

3151-Ch-08.qxd 3/2/04 2:45 PM Page 282

• There are few statistics on gender, and children are generally treated as propertiesof their households. This is a relic of an era when women and children were ‘goodsand chattels’.

• Most of the work done by people at home, including looking after children, shop-ping and cooking, is not measured by either economic or social statistics. Only paidwork is measured in the System of National Accounts and, until recently, we had noofficial time budget statistics.

• Unemployment statistics do not measure the number of people who want jobs.• Health statistics do not measure health or care needs.• Housing statistics do not measure homelessness or housing need.• Poverty statistics do not measure the number of poor people.• Economic and transport statistics do not measure the environmental and social costs

of new roads or industrial activity.• Until the 1990s one of the few sources of statistics on ‘ethnic minorities’ were the

‘mugging’ reports of the Metropolitan Police, but there were almost no statistics onracially motivated assault.

Source: Dorling and Simpson (1999b: 4)

Box 8.9: Examples of statistical deficiencies in Britain

and qualification as being, at least sometimes,a defensive and occasionally arrogant postur-ing which unthinkingly rejects the potentialfor using together both numeric and non-numeric data construction and data interpre-tation. This position can be founded onconcepts developed by Geertz (1973) andDenzin (1978) who differentiate between‘thick description’ and ‘thin description’.According to Denzin (1978: 33), thick descrip-tion includes ‘information about the contextof an act, the intentions and meanings thatorganise action and its subsequent evolu-tion’. Interpretation here, then, focuses onthe context, the motivation and meaning,and the process in which action is sedi-mented. Thin description focuses more onfacts about, and characteristics of, theresearch subject.We do not rule out the pos-sibility of using forms of enumeration asthick description – witness Prigogine’s(1984) use of quantitative catastrophe theorywhich has become the foundation of signif-icant strands of modern social theory.However, we do take a collective view on

the place of enumeration in thin and thickdescription – a view which is based on ourexperience as users of enumerative methodsin our research rather than exponents claim-ing to be experts in the field.

Our approach is that enumeration inhuman geography is best seen as a form ofthin description, capable of identifying cer-tain characteristics and patterns of data, butincapable of describing or explicating themeaningful nature of social life (see Chapters 9and 10).There are two key points to empha-size here. First, ‘thin description’ should notbe viewed as a pejorative or derogatory term.It describes a perfectly legitimate researchfunction, usually involving large numbers ofitems of information. Indeed, thin descrip-tion can often achieve a high level of impactwhen publicized in numeric form. For exam-ple, the data in Table 8.1 were used to madevery public statements about rural poverty inBritain. When, however, similar attemptswere made to use qualitative narratives toinfluence public agencies about the plightof the rural poor, there was much more

ENUMERATING

283

3151-Ch-08.qxd 3/2/04 2:45 PM Page 283

resistance to the ideas being advanced(Cloke, 1996). So there is no reason why‘thin’ and ‘thick’ descriptions cannot be usedsympathetically together. Indeed, in practicethey often are, without any explicit acknowl-edgement of that fact. Secondly, thin descrip-tion, although normally applied to numericdata, can in principle equally well be appliedto non-numeric data and is therefore avail-able as part of an interpretative strategy forwhat would generally be described as quali-tative research, as well as for the more obvi-ous quantitative work.

Clearly, then, there are individual decisionsto be made about whether the role of enu-merative analysis can or should go beyondthin description. Tests of inference, compari-son and relationship (and the models whichare founded on them) are clearly embeddedwithin a particular scientific logic of ‘explana-tion’. Some readers will want to pursue thisform of analysis and, in so doing, will need tobe aware of the philosophical and technolo-gical assumptions which they are endorsing inthese practices. Others will view theseanalytical techniques as attempting the taskof thicker description without sufficientrecourse to the context, meaning and processwhich can be delivered using more qualitativemethods. At a descriptive and exploratory

level, we see no necessary incompatibilitybetween the outcomes of quantitative andqualitative data analysis. Indeed, fruitful col-laboration may be envisaged between theseoutcomes. Collaboration beyond this point,however, will be more restricted, partlybecause technical choices will often dictatedifferent subjects for research, and partlybecause collaboration beyond this point willbe weighed down by the considerable philo-sophical and intellectual baggage whichcurrently accompanies quantitative and quali-tative approaches.

Notes

1 See, for example, Cole and King (1968),Hammond and McCullagh (1974), Smith(1975), Norcliffe (1977), Gregory (1978),Johnston (1978), Silk (1979), Matthews(1981), Unwin (1981), Clarke and Cooke(1983), Ebdon (1985), Williams (1986),Haining (1990), O’Brien (1992), Walford(1995), Robinson (1998) and Fotheringhamet al. (2000).

2 See, for example, Martin (1996), Chrisman(1997), DeMers (1997) and Longley et al.(1998a).

3 See, for example, Bailey and Gatrell (1995).4 See, for example, Openshaw and Openshaw

(1997).


284

3151-Ch-08.qxd 3/2/04 2:45 PM Page 284

The complexity ofexplanation

We will begin this chapter by returning to oneof the two geographers who featured inChapter 1.According to Linda McDowell, theraison d’être of geography ‘is the explanation ofdifference and diversity’ (1995: 280). Not manywould wish to argue with either part of thisstatement. Geography is concerned with dif-ference and diversity, its own distinctiveness asa discipline resting on its enduring interest inthe uneven development of both the social andthe natural worlds. As an academic pursuit, italso seems indisputable that geography shouldbe concerned with explaining these differ-ences, as well as with charting and describingthem.To this end it seems entirely reasonablethat human geographers are involved in pro-ducing explanations for the geography of socialevents and processes – accounting, in otherwords, for where, why and how things happenrather than just pointing out when and wherethey take place. This is what we would allexpect of academic research – it is undertakenin order to explain events and processes. Atfirst sight, then, the statement that the rationaleof geography is to explain difference and diver-sity seems fairly unproblematic.

On closer inspection, however, a number ofproblems do arise.These are largely due to theextremely complex nature and meaning ofthe term ‘explanation’.A hint of the difficulties

can be conveyed by considering the way weuse the term in day-to-day life.We ‘explain’ topassers-by how to reach a particular destina-tion: a largely descriptive form of explanationwhich simply involves the transfer of rela-tively accurate information. We ‘explain’ tofriends and colleagues why we acted in a cer-tain manner: a developmental form of expla-nation in which we attempt to account forthe development of particular sets of socialactions and activities.We also ‘explain’why wedidn’t take a particular course of action: ofteninvoking a comparative form of explanationby asserting what would have happened if wehad undertaken that course of action.We even‘explain’ why we are going to undertake aparticular course of action in the future: a pre-dictive form of explanation which is tied upwith predicting the outcome of a set of socialevents that have not yet happened.We usuallypay no heed to these different types of expla-nation, and there is usually no need to – socialconventions and accepted forms of commu-nication will normally ensure that each differ-ent meaning of the term ‘explanation’ isclearly conveyed and understood.

In human geographical research the mean-ing of the term is more restricted and is usu-ally confined to a developmental form ofexplanation – that which attempts to traceand account for the ‘development’ ofprocesses and events (see Mason, 1996: 127) –although comparative and predictive forms of

Explaining9

3151-Ch-09.qxd 3/2/04 2:46 PM Page 285

explanation are also found to a lesser extent.There are also well-established proceduresand principles that academics use when theyattempt to explain something. Unfortunatelythis does not lessen the complexity surround-ing the term because different researchers willwork with different understandings of whatconstitutes an adequate form of explanation,and they will use different procedures in orderto provide one. In his book on social sciencemethods, Sayer (1992: 232) states that thequestion ‘What is a good explanation?’ is oneof the most common yet most exasperatingfor a methodologist to be asked.As he pointsout, the answer first depends on what kind ofsocial activity or process is being investigatedand, secondly, on what the researcher wants toexplain about it.What the researcher wants toexplain about it is in turn linked to how he orshe conceives of the relationship betweenmore abstract conceptual and theoreticalissues and empirical research.Thus there is nosingle route to a good explanation, and thepath taken will depend on what needsexplaining as well as on what the researcherunderstands to be the components of a goodexplanation.This chapter will take us throughthe different forms of explanation commonlyused by geographers and will also highlighthow the researcher can help to select thosewhich are more suitable for his or her ownparticular piece of research.

Explanation throughlaws: geography asspatial science

In his book examining the development ofhuman geography, Chisholm opened thechapter on ‘Theories of spatial structure andprocess’ with the claim that ‘Explanatorytheories may be regarded as the apex of thescientific pyramid, the goal to which westrive’ (1975: 122).This statement is instructive

for three reasons. First, it assumes that humangeographers will seek to follow a particularversion of the ‘scientific’ form of explanation.Secondly, it states that we should deriveexplanation from theoretical statements.Thirdly, it implies that we should all followthis one prescribed route. These three claimsare, of course, linked, and the ‘scientific’ formof explanation, drawing on universal theoreti-cal statements or laws, came to have a hugeinfluence on the discipline of human geogra-phy from the mid-1960s onwards.

Prior to this, explanation in geography waslargely concerned with applying intuitiveunderstanding, based mainly on description,to a large number of individual and seeminglyunique cases.The notion held throughout thefirst half of the twentieth century that geo-graphy was an idiographic science, concernedwith studying the unique and particular, mil-itated against any search for generalized formsof explanation. It was felt that, because geo-graphers were largely concerned with studyingparticular places, they could search only forparticular understandings.This view was chal-lenged in the 1950s, most forcefully bySchaeffer (1953), who argued that geographyshould be nomothetic and seek to providegeneral laws about spatial organization. Thiscall gradually gained ascendancy and, formuch of the 1960s, geographers looked forways of formulating and applying such laws.

They did so by working within a positivistframework of inquiry. For the positivist, pre-dictive and explanatory knowledge comesfrom the construction and testing of theories,which themselves consist of highly generalstatements expressing the regular relationshipsthat are found in the world. In this process,scientific theories consist of universal state-ments whose truth or falsity can be assessedthrough observation and experiment (Keatand Urry, 1975). Those theories which wereempirically verified would then assume thestatus of scientific laws. The full transfer of


286

3151-Ch-09.qxd 3/2/04 2:46 PM Page 286

scientific method into human geographycame in 1969 with the publication of Harvey’sbook Explanation in Geography where heargued for the primacy of deductive theoret-ical forms of explanation. Under this form ofexplanation, the event or process to beexplained is deduced from a set of initial ordetermining conditions and a set of generallaws – given Law L, if initial conditions C andD are present, then event E will always occur.The event E is ‘explained’ by reference to thelaws and the initial conditions. There is nological distinction between the explanationand the prediction of any given event. Anexplanation simply refers to an event whichhas happened, and prediction to one in thefuture. In his book Harvey outlined a numberof explanatory forms in geography whichused this deductive form of reasoning.Whichis applied depends on the kinds of questionsbeing asked and the kinds of objects underinvestigation, but they all share the methods ofhypothetico-deductive science wherebyexplanation comes from the formulation oftheories which yield testable hypotheses. Bytesting these hypotheses we can discover theregularities of the universal laws which gov-ern patterns of events. The events are then‘explained’ by reference to the universal laws.As long as a correct set of initial conditions isspecified, these laws can be drawn on to pro-vide explanations. This kind of reasoningdominated human geography in the 1960sand most of the 1970s and generated thesearch for law-like statements of order andregularity which could be applied to spatialpatterns and processes. Hence the successionof models which appeared in geography overthis period – for instance, Christaller’s modelof settlement hierarchy, Alonso’s land-usemodel, Zipf ’s rank-size rule of urban popula-tions and Weber’s model of industrial loca-tion. All were an attempt to use law-likestatements in order to explain and predictspatial outcomes.

However, a fundamental problem with thistype of explanation is that what causes some-thing to happen has nothing to do with thenumber of times or the regularity with whichit happens: as Sayer points out, ‘unique eventsare caused no less than repeated ones, irregu-larities no less than regularities’ (1997: 115).And as he goes on to point out, even showingthat an individual event is an instance of auniversal regularity or law fails to explainwhat produces it.The inadequacy of positivistargument arises from confusing the provisionof grounds for expecting an event to occurwith giving a causal explanation of why thatevent occurred (Keat and Urry, 1975: 27).This problem was compounded in geographyby the fact that much of the work carried outwithin ‘spatial science’ was actually statisticalrather than deductive.And as Guelke (1978: 44)has pointed out:

The use of the word ‘explain’ in a statisticalsense is most misleading, because in such asense it carries no connotation of cause. Yetin everyday language it often has suchcausal implications. In consequence, and inspite of disclaimers, geographers have seenstatistical analysis as an end in itself, whensuch ‘analysis’ is basically a mathematicaldescription.

Thus, geographers working within a positivistscientific framework tended to confusedescription with explanation and looked toexplain events via statistical regularity ratherthan causal process (see also Chapter 8).Despite this, this method of explanation waspredominant for most of the 1960s and 1970sand has left a huge legacy within the disci-pline. A more recent example of its applica-tion can be found in an article by Bennett andGraham, published in 1998 and interestinglyentitled ‘Explaining size differentiation ofbusiness service centres’. In this case, the‘explanation’ is given in purely statisticalterms. As the authors themselves put it, ‘ourpurpose is to demonstrate statistically thegeneral pattern of differentiation of groups of

EXPLAINING

287

3151-Ch-09.qxd 3/2/04 2:46 PM Page 287

centres of differing sizes’ (1998: 1477). Butdemonstrating a statistical pattern, in this casea hierarchical one of five stages, does notexplain that pattern. The factors which pro-duced it are not interrogated at all, and analysisstops once the pattern has been identified.What takes over is intuition; the authors statethat ‘the five levels can be interpreted in anintuitive way’ (1998: 1478), but even this intu-ition results only in the five levels of businesscentre being labelled as national (and inter-national) foci, regional centres, secondaryregional centres, shadows and local centres.Again, there is no explanation of these differ-ences, merely a descriptive labelling process.What is instructive is that the authors felt thatsuch description amounted to explanation –once their statistical tests had yielded regularresults, they felt the term explanation couldbe applied – when what actually had beenprovided was a very detailed statisticaldescription.

Explanation as causation

In contrast to this emphasis on mathematicalorder and statistical description, Sayer (1992;1997) advocates a causal approach to explana-tion. In this view, an explanation amountsto statements about what actually causes anevent to happen, and an adequate causalexplanation requires the discovery of relationsbetween phenomena and of some kind ofmechanism which links them.To produce anexplanation we therefore need a knowledgeof the underlying structures and mechanismsthat are present and of the manner in whichthey generate or produce the phenomena weare trying to explain.

Sayer uses the example of manufacturingshift to explore the difference between thetwo types of explanation, pointing outthe difference between the view that empiri-cal observations themselves can generate

explanations and the opinion that they merelyact as a guide to identifying some of theunderlying processes at work (1997: 117–19).In order to make the distinction clearer, hedraws on research which examined the ten-dency for metropolitan areas to lose manufac-turing employment in the 1960s and 1970s,while small towns and rural areas enjoyedgains. One explanation that was sought forthis phenomenon explored the influence ofgovernment grants on business location bylooking at whether the areas of employmentdecline had assisted or non-assisted area status.The hypothesis that there was a link betweenbusiness success and grant aid was rejected, fornon-assisted areas such as East Anglia experi-enced employment growth as well as assistedareas such as Northumberland, and the largerurban areas within assisted areas suffereddecline in the same way as their non-assistedcounterparts.Thus, because the regularity wasnot found and the universal statement couldnot be made, the presence of governmentgrants was rejected as a potential source ofexplanation.

But as Sayer points out, there is anotherway of positing such a link which does notrely on universality but, instead, looks atmechanisms and processes. This wouldexplore the effect of regional aid on employ-ment growth and decline by thinking throughwhich kinds of firms might make use of suchaid and in what circumstances.Thus instead ofpositing, and looking for, a simple regularitybetween incentives and employment change,we would now be looking at the different cir-cumstances of individual firms and askingwhat kinds of firms would find regionalgrants attractive and why. This recasting of thequestion requires a different type of researchin order to provide a different type of expla-nation. Research would be less concernedwith gathering large-scale sets of data whichcould be statistically tested and more con-cerned with investigating the strategies and


288

3151-Ch-09.qxd 3/2/04 2:46 PM Page 288

responses of individual firms through interviewsand questionnaires. We would, of course,expect more complex answers to emerge – insome cases we might find the influence ofincentives on location and in others wewouldn’t – depending on the type and natureof the firm under investigation. Moreover,even if there was a positive influence this neednot necessarily translate into employmentgain: some firms may well use regional develop-ment grants in order to buy new machinerywhich might make workers redundant, result-ing in overall employment loss. This moremessy conclusion is a reflection of the com-plex nature of the problem under investiga-tion. It also emerges from seeking a differenttype of explanation – one based not on theidentification of a regularity of occurrencebut on searching for the processes and mech-anisms which cause an event to happen.

Intensive andextensive research

This in turn has implications for the type ofresearch methodology adopted and the typeof research questions asked. Remaining withthe above example (from Sayer, 1997: 118), ifwe ask what determines the location of manu-facturing investment, the answer will almostcertainly be that several different processes do,depending on the particular firm in question.In our research we might seek to find whichprocesses are most important in a limitednumber of cases. This is known as ‘intensive’research and it, focuses on a small number ofcase studies in order to find the causalprocesses and mechanisms behind a particularevent. But as Sayer (1997: 118) notes, ‘discov-ering how a mechanism works does not tellus how widespread that mechanism is: forexample in how many instances were regionalincentives influential?’ In contrast, the type ofresearch which may lead to an answer for this

type of question is termed ‘extensive’ – insteadof seeking causal mechanisms it asks about the‘extent’ of certain phenomena.The confusionbetween the two types of explanation wehave considered stems from this difference.The two types of research actually provideanswers to two very different kinds of inquiry.The problem is that ‘this second kind of ques-tion regarding distribution and extent isexpected to answer the first regarding causa-tion’ (1997: 118).This is not accidental or theresult of sloppiness on the part of theresearcher but, as we have seen, it is built intothe very foundations of scientific methodwhere evidence of regularity (or extent) andstatistical description are put forward as‘explanations’ of particular events. However,when looked at from a perspective whichviews explanation in terms of causation, whilethe search for regularities and generalizationsmay contribute answers to questions regard-ing the quantities and distributions of phe-nomena (extensive questions), it will throwlittle light on what produces them (intensivequestions).

Exploring this distinction between inten-sive and extensive research reminds us that thetype of explanation which can adequately begiven is highly dependent on research designand methodology. In other words, it is impor-tant to appreciate in advance the type ofexplanation being sought or the data whichare generated from the research will not havebeen constructed in an appropriate form. Asset out in Table 5.1 in Chapter 5, the differ-ences between intensive and extensiveresearch, which at first sight might appear tobe rather superficial ones centred around thescale of research being undertaken, are in factquite fundamental differences of researchdesign, technique, method and explanation.

As we noted in Chapter 5, each is actuallyaddressing a different type of researchquestion, and from this follows a differentcontent and form to the analysis. Because

EXPLAINING

289

3151-Ch-09.qxd 3/2/04 2:46 PM Page 289

intensive research is concerned to identifycausal relations and connections, the groups itstudies will be those whose members arethought to relate to each other, structurally orcausally. Specific individuals or ‘actors’ will bestudied because of the properties they possess(such as a being an owner-occupier or beinghomeless, or being unemployed or being amanaging director) and the relations theyhave to others (such as being in a position ofauthority or being a receiver of a certainservice). In contrast, extensive research will beconcerned to find patterns of regularity anddistribution and, as a result, will focus largelyon taxonomic groups whose members sharesimilar formal attributes but who need notnecessarily be connected to one another at allbeyond this. The research methods will alsotend to be different. Extensive research willfavour relatively large-scale and standardizedmethods such as formal questionnaires, whileintensive research will focus more around lessformal and more interactive methods, such asin-depth interviews or participant observa-tion. However, as we noted in the previouschapter, the roles of the two types of researchcan be seen as complementary rather thancompeting, and in assessing which to use weshould be mindful that both may well have aplace in the same research project. Prelimi-nary work of an extensive nature may well benecesary in order to set a broad context (thelevel of home ownership in two differentareas of a town, for instance) and intensivework can then be used to identify causalprocesses and relations within this (whyparticular people do or do not buy houses inthese two places). The important point tostress in this chapter is that, while extensiveresearch may be used to identify broad trendsand patterns, it will be less helpful in terms ofexplanation. This is because the relationshipsit identifies are those based on similarity andformal correlation rather than those based oncausal connections.

The search for a revolutionin geographicalexplanation

On one level, the discussions initiated bySayer around descriptive versus causal expla-nation were of great assistance in clarifyingthe position over the form and nature ofexplanation in geography, and they helped toformalize the growth of causally based expla-nation during the 1980s. On another level,however, they deepened a complex set ofdebates over precisely which elements ofsocial activity should be incorporated intocausal explanation. These debates had begunin the early 1970s as doubts began to emergeover the adequacy of supposedly ‘scientific’explanations rooted in positivism.

These doubts were at once conceptual andmoral. They were perhaps expressed mostforcefully by David Harvey in his 1973 bookSocial Justice and the City, where he charted hisown personal journey away from positivismand set out an alternative frame of referencefor understanding social and geographicalchange. The initial reasons given for distan-cing himself from the positivism which hisown book Explanation in Geography (1969)had done so much to establish were methodo-logical and conceptual. Harvey explained thathe turned away from positivism and towards aMarxist-inspired analysis because he could‘find no other way of accomplishing what Iset out to do, or of understanding what has tobe understood’ (1973: 17).Among the reasonsgiven for this shift were that the scientificform of explanation used within spatialscience has a ‘tendency to regard facts as sep-arate from values, objects as independent ofsubjects, “things” as possessing an identityindependent of human perception and action,and the “private” process of discovery as sep-arate from the “public” process of communi-cating the results’ (1973: 11–12).According toHarvey, each of these tendencies carried


290

3151-Ch-09.qxd 3/2/04 2:46 PM Page 290

implications for the adequacy of the explanationthat could be given by a positivist humangeography.

To these academic concerns underpinningthe shift towards a different type of explana-tion Harvey also added moral and ethicalimperatives. As he wrote in Social Justice andthe City:

There is an ecological problem, an urbanproblem, an international trade problem,and yet we seem incapable of saying any-thing of depth or profundity about any ofthem. When we do say something it appearstrite and rather ludicrous. In short, our par-adigm is not coping well. It is ripe for over-throw. The objective social conditionsdemand that we say something sensible orcoherent, or else forever…remain silent. It isthe emerging objective social conditionsand our patent inability to cope with themwhich essentially explains the necessity for arevolution in geographic thought. [Ourtask] does not entail…mapping even moreevidence of man’s patent inhumanity toman…the immediate task is nothing morenor less than the self-conscious and awareconstruction of a new paradigm for socialgeographic thought. (1973: 129, 144–5)

The search, then, was on for the conceptsand categories which could help accomplishsuch a revolution in geographic thought andexplanation.

Explanation throughabstraction

Having set out the critique, Harvey immedi-ately began to develop his own preferred‘explanatory schema’ (1989b: 3). In order to dothis he turned to the writings of Karl Marx.As he explained in Social Justice and the City,‘the most fruitful strategy at this juncture is toexplore that area of understanding in whichcertain aspects of positivism, materialism andphenomenology overlap to provide adequateinterpretations of the social reality in whichwe find ourselves.This overlap is most clearly

explored in Marxist thought’ (1973: 129). In aseries of extremely influential writingsstretching over the next three decades, Harveycontinued to explore Marxist thought in aneffort to develop this ‘area of understanding’.One reason for the influence of Harvey’swork, aside from its intellectual richness, is theclarity which he brings to the task of expla-nation and understanding. In all his work heis at pains to chart a precise route linkingtheory and concept with empirical findings.

The key to this journey lies in the notionof abstraction. This refers to the practice ofisolating in thought particular aspects of theobject under study. This is done in order tosimplify some of the complexities of theworld we are researching, at the same time asidentifying key or critical aspects of thatworld. The Marxist-inspired methodologyHarvey draws on utilizes the process ofabstraction in distinct stages. In order to showhow these stages proceed, Harvey (1989b:8–10) examines the various elements whichmake up a typical daily meal.As he points out,we can happily consume these withoutknowing anything of the conditions whichled to their appearance on our table – indeed,this is usually the case. If we think hard wemight remember where we purchased themand how much they cost. But this is about asfar as we would get in terms of understandingthe geographies behind the production of thismeal. Yet in terms of understanding, andindeed explaining, the exchange of money fora packet of cornflakes or a loaf of bread isalmost the very end of the process rather thanthe beginning. Hidden behind this simpleexchange is a whole host of social relationsand technical conditions which have helpedto govern how the commodities on the tablewere produced, traded and consumed.

To expose these relations and conditionswe use the process of abstraction: isolating inthought particular aspects of the object understudy. In this case we can begin to view the

EXPLAINING

291

3151-Ch-09.qxd 3/2/04 2:46 PM Page 291

food not simply as part of a meal but as acommodity, bought by us but produced else-where by manufacturers using machinery,labour power, raw materials, technical knowl-edge and money. The concepts used todescribe this process – buying, selling, pro-duction, commodities – are abstractions inthat they are describing partial aspects of thefood on the table. Issues to do with taste andsmell, and enjoyment, and the gender rela-tions behind who cooked the meal, and whowashes up, are being put to one side as weconcentrate on isolating those conceptswhich can help us explain the food’s produc-tion. (There is, of course, a strong case to bemade, even at this point in the analysis, for notsetting these things aside at all and for arguingthat these issues are actually bound up withthe production of the food, just as much asthose processes which we have traditionallydefined as ‘economic’.) In this way Harveyargues that we can reduce the ‘complexity ofeveryday life to a simple set of…representa-tions of the way material life is reproduced’(1989b: 9).

At this stage in the argument the abstrac-tions used are fairly straightforward and referto everyday processes expressed in everydaylanguage. But following the process of reason-ing used by Marx, Harvey now takes thingsfurther and introduces a further round ofabstraction, this time using non-observableand somewhat specialized concepts to relatethe production of our particular meal to theproduction of food within a capitalist socialsystem as a whole. He uses concepts such assurplus value, class relations and means of pro-duction to situate and integrate the individualcase within a much larger whole. In a similarmanner, the American geographer MichaelWatts has recently written of how he some-times brings an oven-ready chicken into hisundergraduate class and asks students to iden-tify ‘this cold and clammy creature which I’vetossed upon the lectern’ (1999a: 305). After

some five minutes of the students crying out‘it’s a chicken’, ‘it’s a dead bird’, ‘it’s a fryer’,Watts reveals how he then solemnly declaresthat it is in fact none of these but, instead, is abundle of social relations encompassing,among others, farmers, immigrant labourers,transnational feed companies, processingplants, fast-food companies and the biotech-nology industry (Figure 9.1 depicts the keyactors involved in the production and con-sumption of the chicken). Here we see therole of abstraction in moving beyond theimmediate object – the chicken – to isolateand identify particular aspects of its existence.Watts goes on to chart how the chickenindustry in the USA moved in the 1940s and1950s from the mid-Atlantic states to developlarge-scale integrated production complexesacross the US south as part of a global indus-try now facing competition from Brazil,China and Thailand. He concludes that ‘youstart with a trivial thing – the chicken as acommodity for sale – and you end up with ahistory of post-war American capitalism’(1999a: 308).

Interestingly,when Marx first elaborated thismethod of reasoning in his book Grundrisse, heuses a geographical example concerning thedistribution of population. In order to under-stand how population is distributed betweenplaces, and between different sectors of indus-try, Marx points out that it would seem correctto begin with an analysis of the populationitself: with what we might call the surfaceappearance. However, he goes on to argue thaton closer examination this route proves falsebecause our view of the population will beincomplete unless we include the classes ofwhich it is composed. In turn, these classes willbe ‘an empty phrase’ unless we are familiar‘with the elements on which they rest. E.g.wage labour, capital etc.’ (1973: 100) Theselatter in turn presuppose exchange, division oflabour, prices, value, money. Marx paints this asa movement towards ever thinner abstractions


292

3151-Ch-09.qxd 3/2/04 2:46 PM Page 292

until we arrive at what he terms ‘the simplestdeterminations’ (p. 100). From here, however,the journey has to be retraced until we arriveat the population again. In other words, oncewe have taken our object of study apart andisolated key relationships and componentsthrough abstraction, we have to put it backtogether again. This time, however, it appearsnot as an incomplete surface picture but ‘as arich totality of many determinations andrelations’ (p. 100).

To return to the example drawn fromWatts, we started with the simple and some-what trivial oven-ready chicken and, via theprocess of abstraction, we finished with therather complex commodity circuit set out inFigure 9.1. The ready-to-cook chicken at itscentre now appears as the totality of manydeterminations and relations. We have thusmade a double movement – from the originalempirical object of study to the abstract con-cept, and from the abstract back to the concrete,

EXPLAINING

293

FlockService

FlockService

FurtherProcessing

FurtherProcessing

ExportMarket7.9%

Pet Food& Renderers

9.7%

FurtherProcessors

7.5%

DomesticFood

Market82.4%

RetailGroceryStores42.7%

Foodservice(restaurants &

institutions)Governmentand others

39.7%

Distributors25.7%

EquipmentManufacturers

Broiler GrowoutCOMPANY

Broiler GrowoutCONTRACT GROWERS

ResearchCompanies

Hatching-egg FarmsCONTRACT GROWERS

Hatching-egg FarmsCOMPANY

GrowerOrganizations

Finance &Credit

AncillaryIndustries

Consulting andancillary industries

National BroilerCouncil

Biotechnologycompanies

University and stateextension and R & D

ProcessingPlants

Rendering

Feed Mill

Processors

PROCESSED CHICKEN

INT

EG

RAT

OR

PR

OD

UC

TIO

N

Hatchery

Feed Broiler Chicks

Breeders

Breeders/Eggs

Uniform Genetic Stock

Breeder feed Breeder Eggs

Waste

Live Broilers

Feed

Ready to Cook

FIL

IER

E

CO

NS

UM

PT

ION

& D

IST

RIB

UT

ION

Bro

kers

& o

ther

s* 3

.7%

Ret

ail 2

6.9%

6.0%

9.5%

25.7

%

7.7%

3.7%

Foo

d se

rvic

e 16

.4%

Gov

ernm

ent 1

%B

roke

rs &

oth

ers

3.6%14.9%

.7%

.2%

3.8%8.4%

1.5%

Figure 9.1 The production and consumption of chicken

Source: Watts (1999: 310)

3151-Ch-09.qxd 3/2/04 2:46 PM Page 293

with concrete here being understood as theunity of many diverse aspects. Harvey arguesthat we can use such movement to ‘help usunderstand all kinds of surface occurrencesthat would otherwise remain incomprehensi-ble’ (1989b: 10). He goes on to argue that theproof of this line of reasoning is found in itsusage – ‘explanatory power becomes thecentral criterion of acceptability’ (p. 10).

It should perhaps be stressed at this stagethat the procedure of abstraction is by nomeans limited to Marxist-informed research.Research on gender relations might draw onabstractions developed in feminist researchconcerning patriarchy. That on the servicesector might use abstractions about bureau-cracy originally devised by Weber. Work onpolitical geography and the state might drawon abstractions connected to notions ofpower and governmentality derived fromFoucault.Whatever the source of the abstrac-tion, the way in which such explanation pro-ceeds by relating the surface appearance of anempirical event to the actual processes takingplace behind it is summarized in Figure 9.2.The figure shows how Sayer envisages therelationship between the abstract and theconcrete, and between causal mechanisms andstructures and the events they produce. Itrepresents a complex social system in which theactivation of particular mechanisms produceseffects (or events) which may be unique to aparticular time and place. But with differentcircumstances and conditions the same mech-anisms may produce different events and,conversely, the same kind of event may havedifferent causes. The use of abstraction per-mits the analysis of objects in terms of their‘constitutive structures’ as parts of widerstructures and in terms of their causal powers.In order to explain any particular event weneed to undertake concrete research to exam-ine what happens when these combine.

Figure 9.3 develops the information pre-sented in Figure 9.2 in order to clarify the

relationship between different kinds ofresearch and different kinds of explanation. Itshows how the researcher can examine mech-anisms, structures and events in different waysto produce different types of account. Thusabstract theoretical research deals with structuresand mechanisms in an abstract sense, andevents are only considered as possible out-comes. Concrete research of an intensive naturedoes study actual events, treating them as phe-nomena that have been caused by specificmechanisms and structures (themselvesuncoverable through abstract research). Bycontrast, the method of generalization is under-taken through research of a more extensivenature, where the researcher seeks to establishregularities and patterns at the level of events.Most positivist law-like explanations actuallyconsist of generalizations made at this level ofthe research process. It is also possible to adda fourth type of synthesis research, whichattempts to combine abstract and concreteresearch findings with generalizations cover-ing a wide range of events. Interestingly, Sayernotes that ‘research of this kind is especiallycommon in…geography’ (1992: 236). This isbecause as a discipline geography has a tradi-tion of seeking to understand variation acrossa range of events as well as a history of exam-ining the causes of specific events. It is con-cerned with both the general and theparticular, and, often, with the interconnectionbetween them.

Some of the major problems in explanationarise when the capacity and purpose of thesedifferent types of research are misunderstood.In particular, as Sayer points out, ‘researchersoften over-extend them, by expecting one typeto do the job of the others’ (1992: 238,emphasis in original). This can lead to over-ambitious claims being made on behalf of theexplanatory power of specific pieces ofresearch. In particular, abstract research can beover-extended when it is used to explainevents directly, without relevant empirical


294

3151-Ch-09.qxd 3/2/04 2:46 PM Page 294

research. On the other hand, concreteresearch can also be over-extended when thefindings related to a particular event arethe subject of overgeneralization or when theexplanation of that event proceeds withoutreference to the broader sets of structures andprocesses that have brought it about (Clokeet al., 1991: 154). Sayer concludes that the keyrequirement with this type of explanation isto be able to conduct theoretically informedconcrete research so that events on theground can be properly linked via abstractionto the mechanisms and processes which pro-duced them. In this section we have set outthe central role that abstraction might play inthis process.

Explanationand subjectivity

If we return to Figure 9.2 we can see thatSayer essentially stops the research process atthe level of events. He does, however, contendthat there may be a case for adding a fourth

level to the vertical dimension ‘to covermeanings, experiences, beliefs and so forth’(1992: 116). After consideration he concludesthat, since these can form structures, functionas causes or be considered as events, they havealready been taken into account in thisexplanatory schema.There are others, though,who would dispute this, and the recent historyof geography would bear witness to the factthat many believe that events remain unex-plained unless we are able to account for whatpeople understand, mean or intend by theiractions. In these cases explanation proceedsthrough an engagement with subjectivity.

Human geography’s encounter with the sub-jective realm of meanings and experiencesgathered pace in the 1970s. Like the Marxist-inspired work which underpinned explanationthrough abstraction, it was also fuelled by a dis-satisfaction with positivist forms of explanation.In this case the critique focused on positivism’sneglect of human agency and its tendency toreduce the complexity of human emotions andfeelings to ‘little more than dots on a map,statistics on a graph or numbers in an equation’

EXPLAINING

295

mechanisms

structures

eventsCONCRETE

ABSTRACT

M1,

E1, E2, E3, E4, E5, E6, E7, Ek

S1, S2, S3, S4, Sk,

M2, M3, M4, M5,......................

............................

Mk

Figure 9.2 Structures, mechanisms and events


3151-Ch-09.qxd 3/2/04 2:46 PM Page 295

(Cloke et al., 1991: 69). According to Entrikin(1976: 616), spatial science presented a view ofbehaviour which was ‘overly objective, narrow,mechanistic and deterministic’. In contrast, andas a reaction, those geographers involved in thisparticular form of critique studied ‘the aspectsof [people] which are most distinctively“human”: meaning, value, goals and purposes’.This concern with human awareness, con-sciousness and creativity was soon labelled as‘humanistic geography’ (Tuan, 1976a).

In terms of explanation, those geographersworking under the humanistic banner sub-scribe to Schutz’s view that ‘all scientificexplanations of the social world can and, forcertain purposes, must refer to the subjectivemeaning of the actions of human beings fromwhich social reality originates’ (cited inGregory, D., 1978a: 125, emphasis in original).Thus the ‘scientific’ form of explanationderived from the natural sciences and followedby spatial science is seen as inadequate, not

only because human geographers examine aworld which is meaningful to its inhabitantsbut, crucially, also because those meaningsenter into the very constitution of that world(Barnes and Gregory, 1997: 356). In otherwords, the key to our understanding ‘is to befound somewhere within ourselves’ (Gregory,D.,1978a: 124). The task of humanistic geogra-phy is not to impose its own external frame ofreference on to the world but to inquire intoand understand the frames of reference drawnon by the social actors themselves. In thissense explanation becomes contextual as wellas causal through a search for the meanings ofparticular actions (see Daniels, 1985). AsSchutz puts it, the basic question becomes‘what does this social world mean for theobserved actor within this world and whatdid he [sic] mean by his acting within it’?(cited in Ley, 1977: 505).

Under humanistic geography, then, theissue of explanation becomes heavily tied to


296

Synthesis

Generalization (extensive)

Abstract research

Concrete research (intensive)

CONCRETE

ABSTRACT

mechanisms

structures

events

M1,

E1, E2, E3, E4, E5, E6, Ek

S1, S2, S3, Sk

M2, M3, M4, Mk

Figure 9.3 Types of research


3151-Ch-09.qxd 3/2/04 2:46 PM Page 296

an understanding of the meanings, perceptionsand motives of social actors. As Schutzimplies, however, there is actually a dual senseto the term explanation as it is used inhumanistic geography – ‘what does this meanfor the actor’ and ‘what did he mean by acting’in this manner. In some cases, then, research isconcerned with uncovering the antecedentsto an action: with uncovering what wasmeant by an action or with ascertaining whatmeanings and motives lay behind an action.This is somewhat akin to explanation as cau-sation, although the cause is not uncoveredthrough abstraction but through exploringthe meaning of the event for those involved.In other cases, however, the humanistic geog-rapher will seek to explain the meaning ofsomething for those involved – and in thissense the ‘something’ is not confined to asocial action but could equally refer to a placeperhaps, or a landscape, or a cultural event.

We will use some examples to show howparticular research within the field of human-istic geography can be placed at differentpoints along this continuum of explanation.At one end of the continuum, closer to thenotion of explanation as description, we cansituate the work of Edward Relph. In his pio-neering book Place and Placelessness, Relphexplicitly rejects a form of explanation basedon models or on abstraction. As he states(1976: 6–7):‘My purpose…is to explore placeas a phenomenon of the geography of thelived-world of our everyday experiences…Ido not seek to…develop theories or modelsor abstractions…My concern is with thevarious ways in which places manifest them-selves in our experiences or consciousness.’Hence Relph was concerned to demonstrateand describe a range of different reactions to,and relations with, place. He uses conceptssuch as ‘rootedness’, ‘rootlessness’, ‘insideness’and ‘outsideness’ to chart these reactions,which he views as intensely subjective andpersonal. This is no easy task and, as Relph

points out, a human geography then based ona positivist form of explanation was ill-equipped to tackle it. In his words,‘any explo-ration of place as a phenomenon of directexperience cannot be undertaken in terms offormal geography…It must instead be con-cerned with the entire range of experiencesthrough which we all know and make places’(1976: 6).As well as shifting geography’s viewof place – from ‘a set of externally observablefeatures found at a particular site’ (Barnes andGregory, 1997: 293) to subjective centres ofexperience and consciousness – Relph wasalso helping to redefine its attitude to theexplanation of place.This shifted from ‘scien-tifically’ collecting and analysing a set of ‘facts’about a particular place towards charting anddescribing people’s experiences of differentplaces and examining the meaning of thoseplaces for those people.

Although Relph’s work concentrated ondescribing the meaning and experience ofplace, and on recovering the character ofeveryday geographical encounters, othershave used experiential description as the firststage in more complex accounts of socialaction and behaviour. Such work tendstowards a causal view of explanation, and oneof the foremost practitioners of this type ofhumanistic geography is David Ley (see alsoChapter 6). In his path-breaking monographon life in Philadelphia’s black inner-city, Leyset out to ‘order and understand the world’around him (1974: 4). He, too, was concernedwith experiences and meaning but only as astep on the road to understanding social activity.In his words, the researcher

has to discover the salient environmentwhich prompts decision making and behavi-our, the environment as perceived andexperienced, and its forces which providerules for human action…Only by coming togrips with the experiential environment,…only by exploration of the topographies ofmeaning, may we uncover regularities inbehaviour. (1974: 4)

EXPLAINING

297

3151-Ch-09.qxd 3/2/04 2:46 PM Page 297

Ley explored these meanings, perceptions andexperiences through fieldwork in inner-cityPhiladelphia. From the outset he also arguedthat such research would involve a differentform of explanation from the positivismwhich was then dominant within spatialscience. As he put it:‘To understand a decision-making world, a more holistic paradigm isrequired, one which…proceeds from theindividual to the aggregate, and which recog-nises inductive understanding as no lessappropriate than deductive prediction…Tounderstand the environment as perceived… isan invitation to more subjective methodo-logies’ (1974: 9). In this case, these subjectivemethodologies were built around a six-monthperiod of participant observation (seeChapter 6) in which Ley ‘walked the streets,attended meetings, belonged to a church’ andwas involved in joint activities with the localcommunity which ‘ranged from helping withhomework, and a bake sale, to social eveningsat the bowling alley, to informal “raps”, andtwo out-of-town summer excursions’ (1974:17–18).Through these methods Ley hoped tobuild up a picture of the complex system ofblack inner-city life and of the role of theexternal environment in that system.

His findings were rich and detailed (seealso Ley and Cybriwsky, 1974). They led tothe proposition that space and environmentboth have a range of ascribed meanings andthat, crucially, these meanings contain cueswhich prescribe appropriate forms of behav-iour. In order to examine this, Ley looked atthe behaviour of inner-city street gangs and atwhere residents felt safe and unsafe. Figures 9.4and 9.5 show the gang territories, as definedby the location of graffiti and the pedestrianroutes taken by residents. Ley found that resi-dents differentiated their own neighbourhoodfrom others primarily through behaviouralrather than physical properties. Descriptiveindicators such as housing quality, traffic den-sity and land use were used to differentiate

one area from another less often thanbehavioural indicators such as gang activity,violence and quietness. He concluded thatwhen people ascribed properties to particularspaces, prior to decision-making or to move-ment, the perceived meaning of an environ-ment was more important than its purelyphysical characteristics.

For Ley, then, ‘actions are intentional andpurposive, they have meaning, but access tothis meaning requires knowledge of themotives and perception of the actor, his [sic]definition of the situation’ (1977: 505).Thus afull explanation of such action hinges aroundan understanding of the motives and percep-tions of those involved – around an apprecia-tion of their ‘definition of the situation’.Taylor(1971) has drawn attention to a problem withsuch a procedure. He agrees that, ultimately, ‘agood explanation is one which makes sense ofthe behaviour’.However, he points out that ‘toappreciate a good explanation one has to agreeon what makes good sense; what makes goodsense is a function of one’s readings; and thesein turn are based on the kind of sense oneunderstands’ (cited in Gregory,D., 1978a: 145).This in turn means that it is difficult to draw aclear distinction between subjective under-standing and subjective explanation, and thatwhat we judge to be an adequate interpreta-tion of social action will always be dependenton the constructs which are embedded in ourown frame of reference: in how we make thetranslation between the first-order intentionsof the actors involved and what we regard asadequate second-order representations ofthose actions (Gregory, D., 1978a: 145). Thisconcern has always surrounded the more sub-jective forms of explanation. As Gregorypoints out, overcoming it would mean thatgeography would ‘have to dismantle the oppo-sitions between subject and object, actor andobserver, and emphasise the mediationsbetween different frames of reference’ (1978a:146). We next turn to a form of explanation


298

3151-Ch-09.qxd 3/2/04 2:46 PM Page 298

which has attempted to do just this, by explicitlyacknowledging the enormous difficultiesinvolved in adequately representing, or perhapsmore accurately re-presenting, the actions andmotives of others.

Explanation and practice

In contrast to those types of explanation whichseek to uncover the causes of events, whetherthrough abstraction or through uncovering themeanings and perceptions of those involved,

some geographers have turned to a mode ofthought which explicitly rejects any search foran underlying causal process.This is known as‘non-representational’ thinking for it is basedon the premise that we can never hope ‘tore-present some naturally present reality’ (Thrift,1996: 7).This in turn is because it is our prac-tices, or performances, which constitute oursense of the real. Thus, reality is not sittingsomewhere as a pre-given, waiting to beuncovered, but is constantly constructedthrough social practice and social activity.According to such a view events ‘have an

EXPLAINING

299

OBSCENITIES PER BLOCK

Southampton Avenue

Home corner of gang turf

Approximate edge of gang turf

0

1

2

3

4.6

0 100 200 300

YARDS

Sutton

Acton

Reading

Lavant

Guildford

Henfield

Richmond

Rochester Avenue

Lynn

Tubm

an

Dub

ois

Eve

rs

36th 37th 38th 39th 40th 41st 42nd 43rd 44th 45th 46th

Kin

g

Gar

vey

Figure 9.4 Gang territories defined by the location of graffiti

Source: Ley (1974: 218)

3151-Ch-09.qxd 3/2/04 2:46 PM Page 299

imminent quality that means they are shaped asthey are made’ (Hetherington and Law, 2000:127). This understanding of events inevitablyleads to a ‘rather different notion of explana-tion’, one which prioritizes social practice andis more ‘concerned with thought-in-action,with presentation rather than representation’(Thrift, 1996: 7). As such, explanation and

understanding are not ‘about unearthingsomething of which we might previously havebeen ignorant,delving for deep principles or dig-ging for rock-bottom, ultimate causes as [theyare] about discovering the options people haveas to how to live’ (Thrift, 1996: 8). In summary,‘non-representational theory is an approach tounderstanding the world in terms of effectivity


300

ORIGIN

DESTINATION

ONE PEDESTRIAN ROUTE

Hastings Avenue

Reb

ecca

Ave

nue

Ham

pton

Jackson

46th

36th 37th 38th 39th 40th 41st 42nd 43rd 44th 45th 46th

47th 48th 49th

Murray R

oad

Col

lies

Allan

Clagton

George

Southampton Avenue

Ormond

Tubm

an

Youn

g

Dub

ois

Kin

g

Gar

vey

Eve

rs

Mal

colm

Sutton

Reading

Lavant

Guildford

Henfield

A

B

Richmond

Lynn

Rochester Avenue

Kentwood

Acton

Albeny Road

Milton

Rebecca Avenue

Figure 9.5 Pedestrian routes

Source: Ley (1974: 224)

3151-Ch-09.qxd 3/2/04 2:46 PM Page 300

rather than representation; not the what but thehow’ (Thrift, 2000d: 216, emphasis in original).

One strand of non-representational thoughtwhich considers how social practices areformed, maintained and changed is actor-network theory. Sometimes known as the‘sociology of translation’, actor-networktheory was initially developed within the socio-logy of science (see Law, 1986; 1991; 1994)and then taken up within human geographyduring the 1990s.1 A closer consideration ofthe theory, and the way it has been used ingeography, will help to clarify the distinctiveapproach taken by non-representational theo-ries towards the issue of explanation. Actor-network theory gains its name from thecentral tenet of its approach, which uses thenotion of a network to consider how socialagency is constructed and maintained. Unlikecausal explanation, where entities such as themode of production or classes are seen as thecauses of events, actor-network theory viewssuch structures, and the interests they serve, asa set of effects arising from a complex order-ing of network relations.And unlike a subjec-tive form of explanation (which prioritizesthe intentional activity of human subjects),actor-network theory views agency as aneffect distributed through a heterogeneousarrangement of actors and intermediaries,both human and non-human – such as texts,technical artefacts, money and machines.Thusactor-network theory understands events andeffects as processes which are generated by theinteraction of human and non-human agentsoperating through a collective network.

In order for the network to produce, ororder, social effects it must be stable. Such sta-bility is in turn achieved through ‘translation’:defined as ‘The methods by which an actorenrols others’ (Callon et al., cited in Woods,1998: 323). According to Woods (1998: 323),these include ‘defining and distributing roles,devising a strategy through which actors arerendered indispensable to others, and displacing

other entities into a forced itinerary’.Throughtranslation, then, an inter-related set of entities(both human and non-human) is enrolledinto the network in a manner which stan-dardizes and channels their behaviour in thedirections desired by the enrolling actor(Murdoch, 1995: 747–8). Effectivity, or thatwhich produces effects, springs from theinteractions of the various entities involved inthe network. Power is therefore associativeand invested in relations and connectionsrather than in the entities themselves. Whatmatters in producing change and promotingactivity is the manner in which the entitiesare brought together.As a result the researchershould turn away from ‘laws and regularities’and instead search for ‘exchanges and interfer-ences, connections and disconnections…Inother words we need to look for differentkinds of topologies, based on communicationand connectedness’ (Thrift, 2000d: 221–2).

Woods (1998) has used such an approach toresearch rural conflicts operating around landuse and the right to hunt deer with hounds.He takes the example of the attempt made bySomerset County Council to ban such hunt-ing, which itself must be viewed ‘as part of abroader longer-running political contest inthe county: between a conservative, largelyagricultural, elite which has traditionallydominated the local power structure andwhich has close connections with hunting,and an emerging liberal “elite” drawn largelyfrom the professional middle classes’ (1998:324). Woods uses concepts and methodsdrawn from the actor-network approach toexamine the five key sets of entities enrolledby the anti-hunting campaigners in theiractor network (see Figure 9.6).The continua-tion of stag-hunting is positioned as an obsta-cle to the realization of each of their goals,and thus the banning of hunting is identifiedas an ‘obligatory passage point’ which all theentities must support. In opposition, a pro-hunting network developed (see Figure 9.7)

EXPLAINING

301

3151-Ch-09.qxd 3/2/04 2:46 PM Page 301

which consisted of a further set of human andnon-human entities (some of which wereredefined through their enrolment into thesecond network, such as the deer and thecounty councillors).

Woods goes on to chart the differentelements involved in the enrolment andstabilization of each network, through four‘moments’:

1 Problematization, where the goals andproblems of each entity are defined by apotential translator in a manner whichpresents the translator’s objective as thesolution to the other entities’ problems.

2 Intressement, whereby entities are attractedinto the network by blocking other possiblealignments.

3 Enrolment, where the roles of the actors inthe network are clarified and distributed.

4 Mobilization, when the interests are enacted(1998: 323).

He shows how, despite an initial success inpassing an anti-hunting motion, the anti-hunting network was eventually defeatedwhen a second pro-hunting network emergedincluding two new entities – the judiciaryand lawyers.This proved a stronger and morestable network than the first pro-huntingone, and the inclusion of legal entities provedcrucial in effectively ‘invalidating the defini-tion of the county council as an entitywith the power to ban hunting on its land’(1998: 334).

Woods’s account is meticulously researchedand presented in telling a set of stories whichtogether trace the political and social conflictsoperating in Somerset around the issue ofhunting and land use between 1993 and


302

Banhunting

Banhunting

FollowLiberalideology

Representpublicopinion

Enjoyruralidyll

Livenaturallife

Goals ofEntities

LeagueAgainstCruel Sports

LocalCampaigners

CouncilLeadership

CountyCouncillors

LocalPeople

DeerEntities

Need toshowpublicsupport

Opp:CouncilMotion

Huntingcruel =illiberal

Publicopposehunting

Crueltyof huntingconflictswith ruralidyll

Huntingunnatural& cruel

OPP/Obstacles

Figure 9.6 The anti-hunting network (OPP = obligatory passage point)

Source: Woods (1998: 329)

3151-Ch-09.qxd 3/2/04 2:46 PM Page 302

1995.That he chooses to set these stories andnarratives in an actor-network framework issignificant because in such an account thetelling of the story and the presentation of thenarrative become the explanation. AsMurdoch puts it: ‘We must explain by usingthe descriptions of network construction, andnot by recourse to some underlying historicallogic’ (1995: 731).Thus, for the actor-networktheorist there is an elision between descrip-tion and explanation: the one literallybecomes the other, and it is only by thor-oughly describing the inter-relations operat-ing within the network that the researchercan hope to explain their effects (Murdoch,1995: 752).As Latour puts it:

The description of socio-technical networksis often opposed to their explanation, whichis supposed to come afterwards…Yet noth-ing proves that this kind of distinction is

necessary. If we display a socio-technicalnetwork we have no need to look for addi-tional causes. The explanation emerges oncethe description is saturated. (1991: 129)

This image of explanation through saturateddescription is certainly beguiling. If we followit, says Latour,

we will never find ourselves forced to aban-don the task of description to take up thatof explanation…There is no need to gosearching for mysterious…causes outsidenetworks. If something is missing it isbecause the description is not complete.Period. Conversely, if one is capable ofexplaining effects of causes, it is because astabilised network is already in place. (1991:129–30)

According to Latour, then, when we havefully described our network we have at oneand the same time written our explanation.But as Woods points out there are problems

EXPLAINING

303

Representconstituents

Continuehunting

Live Protectsport

Continueestablishedsocial life

Goals ofEntities

CountyCouncillors

Hunt Horses &Hounds

Other FieldSportsEnthusiasts

Localcommunity

Entities

Publicsupporthunting

Opp:Defeat/OverturnBan

Could beput down

Ban setsprecedent?

Social liferevolvesaround hunt

OPP/Obstacles

Maintainbalanceof herd

Deer

Banthreatensmanagement

Figure 9.7 The pro-hunting network (OPP = obligatory passage point)

Source: Woods (1998: 332)

3151-Ch-09.qxd 3/2/04 2:46 PM Page 303

with this, and certainly with transferring suchan approach from the sociology of scienceinto human geography (1998: 338). This isbecause a key difficulty remains in knowingwhich entities should and should not beincluded as actors in the network – withdrawing, in other words, the boundaries ofthe network to be described. As Woods putsit, ‘there are thousands of entities that mightbe considered to have an influence upon thenetwork.Where would we draw the limits ordo we accept the network as infinite?’ (1998:337).To a certain extent, then,we are left witha similar problem faced by those forms ofexplanation which stress subjectivity: how dowe faithfully translate and describe thethoughts, perceptions and meanings of others?This concern grows when we also considerthat any description of an actor network,however complete, can tell only the agent’snarrative, excluding alternative stories fromthose who are not enrolled in the network orthose of other entities less able to communi-cate their own narratives.

Thrift (2000d: 215) has written thatthrough the use of non-representationaltheory we can move away from positions ‘whichassume that a developmental account tells all,towards a “history of the present” (Foucault,1986)’. As we have just noted, the problemstill remains of which history of the present towrite and from which (or whose) perspective?In an effort to overcome these twin concerns,those involved in non-representational workhave retreated somewhat from the networkmetaphor and from the ‘panopticism of actor-network theory…where the analyst looksdown and counts the poor and huddledmasses of scallops and machines’ (Callon andLaw, cited in Woods, 1998: 337). Instead,non-representational theory has prioritizeda focus on ‘practice’ and on ‘performance’.Recognizing that in some senses the notionof an actor network was itself representa-tional, researchers have come to concentrate

on the event itself rather than on the networkbehind it.As Thrift points out this has signifi-cant implications for academic research, for‘what counts as knowledge must take on aradically different sense. It becomes some-thing tentative, something which…is a prac-tice and is a part of practice’ (2000d: 222).Drawing on the work of Gibbons et al.(1994), he goes on to note that this impliesthat we are entering a third type of knowl-edge production (1997: 143; 2000d: 223).Type 1 systems of knowledge productionwere disciplinary and homogeneous and wererooted in the ‘modern’ search for certaintyand fixity. Those ‘geographical laws’ producedby spatial science can be seen as the result ofsuch systems of knowledge.Type 2 systems ofknowledge production, by contrast, are trans-disciplinary and heterogeneous. They havebeen produced by a wider range of workwhich draws on less certain metaphors to dowith ‘complexity’, ‘transience’, ‘flexibility’,‘adaptability’ and ‘turbulence’. Actor-networktheory with its origins in the sociology ofscience and its stress on relations and connec-tions can be said to represent this type ofknowledge.

According to Thrift we are now entering athird type of knowledge production system –one which stresses the ways in which theworld is conducted through activity. It is atype of knowledge which privileges ‘theembodied non-cognitive activity in whichpeople engage’ (Thrift, 1997: 138), and onewhich shifts the interest towards issues ofemotions, desires and imaginations. It is arealm which examines the host of complex‘real-time encounters through which wemake the world and are made in turn’ (1997:138). Such a view would see the geographiesaround us as for ever imminent, as constantlyin use and essentially incomplete, and asalways in the process of becoming. If investi-gating this world requires a new type ofknowledge, it certainly implies a new form of


304

3151-Ch-09.qxd 3/2/04 2:46 PM Page 304

explanation. Or perhaps it implies a breakwith the notion of explanation altogether. Itsstress on imminence and becoming ratherthan on certainty and completeness wouldlead us to question whether we can ever fully‘explain’ anything. Perhaps those who wouldfollow such inquiries are better placed tosearch for an understanding of events ratherthan for an explanation of them. It is to theseconsiderations of geographical understandingthat we turn in the next chapter. Before wedo so, however, we want to provide a briefconclusion to our thoughts on explanation.

Concluding comments:from the explanationof geography to thegeographies withinexplanation

We have covered much ground since webegan this chapter, initially reviewing humangeography’s attempts to formulate and uselaws as part of positivist forms of ‘scientific’explanation, before moving on to examinevarious types of postpositivist explanation –whether through abstraction, subjectivity ornon-representational theory. By way of a con-clusion to this chapter on explanation inhuman geography we want to draw out oneof the key implications the discipline itself hashad to face as it has made this journey.

As we have noted throughout the chapter,as part of its move beyond spatial sciencehuman geography has necessarily engagedwith a variety of social and cultural theoriesin an effort to establish new forms of expla-nation.This engagement has left its traces onboth parties: as Dodgshon points out, ‘It iscrucial to any understanding of how socialtheory has colonised human geography torealise that the discovery, like all true discov-eries of one culture by another, was mutual’(1998: 5). Gregory characterizes the process of

these mutual discoveries as a double movementin which

a form of social theory was projected intohuman geography as a way of meeting aseries of supposedly ‘internal’ concerns, andthen, as a direct consequence of this con-ceptual mapping, the a-spatiality of theoriginal formulation was itself called intoquestion. It is those return movements thathave widened the discourse of geographystill further. (1994: 111)

Critically for our purposes in this chapter, thisbroadening of the geographical discoursethrough mutual discovery has had implica-tions for the content, as well as the nature, ofexplanation. We have moved from a positionwhere the prime object of human geograph-ical research was to explain the geographieswhich were uncovered around us, to onewhere those geographies become part andparcel of the explanation of a range of otherprocesses.

This shift was explored in an influential setof essays published in 1985 entitled SocialRelations and Spatial Structures, jointly editedby a geographer (Derek Gregory) and a socio-logist ( John Urry). In their introductory essay,the two editors conclude that ‘spatial structureis now seen not merely as an arena in whichsocial life unfolds, but rather as a mediumthrough which social relations are producedand reproduced’ (Gregory and Urry, 1985: 3).It is this view of space as constitutive of socialrelations which has underpinned the dialoguebetween human geographers and those work-ing elsewhere in the humanities and socialsciences.As this dialogue has proceeded it haschanged the notion of what geographersshould explain, as well as the idea of how theyshould explain it. At times it is an encounterwhich has shifted our thoughts away fromexplanation and more towards understanding.Indeed, as we have seen, for some scholars thedivision between understanding and explana-tion is difficult to sustain and the two meld

EXPLAINING

305

3151-Ch-09.qxd 3/2/04 2:46 PM Page 305

into each other. It is to a continuation of thisparticular aspect of the dialogue betweengeography and social theory that we nowturn, in inquiring further into the nature ofgeographical understanding.

Note

1 See, for instance, Murdoch and Marsden(1995), Thrift (1997), Murdoch (1995; 1997;1998) and Woods (1998).


306

3151-Ch-09.qxd 3/2/04 2:46 PM Page 306

Introduction

So far in this part of the book we have traceda path from the initial stages of interpretation(sifting and sorting), through the so-called‘thin description’ of enumeration, to the ana-lytical mode of interpretation known asexplanation. To oversimplify somewhat, eachof these procedures can be seen as steps in abroadly scientific enterprise.Although the par-allel is not exact, ‘sifting and sorting’ brings tomind the categorizing activities of the natu-ralists of the eighteenth and nineteenth cen-turies – the exercises in classification of aLinnaeus or a Darwin. Enumeration, with itsemphasis on quantification, generalizabledescriptions and statistical testing, sits squarelyin the ‘geography as science’ corner. Similarly,explanation (whether of a positivist, criticalrationalist, structuralist, Marxian or criticalrealist kind) usually involves a commitment tosome idea of science – understood (at mini-mum) as a process of logical reasoning sup-ported by empirical evidence. That said,throughout our earlier chapters we have triedto question some of the conventionalscientific assumptions about what can beachieved when we sift and sort, enumerateand explain.

None the less, the scientific aspirations ofmuch human geography are well known andreflect the two views of the subject that weredominant in the Anglophone literature until

at least the mid-1970s. One view held thatgeography was a unified discipline and that itshould be considered one of the naturalsciences.As Derek Gregory (1978a: 19) pointsout, this (at least implicit) belief in the naturalscientific status of the subject unites the sup-posedly ‘new’ geography of the 1960s with‘the ideals of its Victorian forebears’. Anotherview is that physical and human geographyare, in certain important respects, epistemo-logically distinct but that human geographycan still be seen as a science, albeit this time asa social rather than a natural science. Withoutwishing to gloss over the fundamental differ-ences between these two views, it is interest-ing that they both share a commitment to thepractice of science (even if they disagreeabout what that means).

In this chapter we want to move away fromthis scientific terrain to focus on some ratherdifferent approaches to geographical inquirythat have grown in popularity in Anglophonehuman geography over the last 30 years to thepoint where they have become the primarymodes of interpretation in many parts of thediscipline. These approaches, which we aregrouping together under the heading of‘understanding’, derive from the arts andhumanities as much as from the socialsciences.What they share is an interest in andconcern for the meanings that attach to thehuman geographies of the world, and a beliefthat the production and communication of

Understanding10

3151-Ch-10.qxd 3/2/04 2:46 PM Page 307

meaning are two of the defining attributes ofthe human condition and by extension of ourhuman geographies. These approaches tointerpretation are based in part on the con-viction that enumeration and explanation cantake us only so far; that there is a central ele-ment of what it means to be human that is leftuntouched and unexplored by scientificmethods, whether of a natural-scientific or asocial-scientific kind. Arguably, therefore, inour practices as human geographers we reacha moment in our attempts to make sense ofthe data before us where we have to do some-thing that is not really central to, even if per-haps on occasion still conformable with, thepractices of enumeration and explanation.

The point – so obvious in some ways andyet so lost on many researchers – is that, put atouch cryptically, people are not rocks. As far aswe know, rocks on a slope or any naturalobjects studied by, say, physical geographersdo not possess an ‘interior’ realm whereinthey are aware, conscious of and makingjudgements about their surroundings.As far aswe know, these rocks do not have the capacityto monitor their local environment nor tomake decisions about how to act within thatenvironment, choosing to do one thing ratherthan an other or to do nothing. People clearlydo possess such an interior realm, however,and our suggestion is that this realm cannotbut be meaningful, full of basic meaningscharged with emotions regarding the imme-diate conditions in which people find them-selves (ones of comfort, discomfort, love, hate,fear, disgust, anger, passions of all kinds), butprobably also containing more sophisticatedmeanings carried in their more or less devel-oped notions, opinions, ideas, ideologies andworldviews. Some of these meanings may bevery much taken for granted, indeed barelyeven acknowledged, by the people involved;others of these meanings, conversely, may beconstantly reflected upon and available to bediscussed in detail and even recorded in written,

sung, acted or otherwise represented forms(see Chapters 3 and 4).

We realize, therefore, that we are sandwich-ing together many different aspects of themental states possessed by human beings, notall of which many scholars would acceptunder the blanket term of ‘meanings’. Nonethe less, all that we wish to underline is thatunderstanding this ‘stuff ’, these interior realmsshot through with meanings of one sort oranother, almost unavoidably becomes a con-cern for human geographers wishing to takepeople seriously as people complete withthese messy interior realms replete withmeanings (see Cloke et al., 1991: ch 3).Thus,figuring out what spaces, places, environmentsand landscapes mean to people, or trying tounderstand the meanings that people in givensituations acquire, elaborate, share and perhapscontest in relation to activities of geographi-cal consequence (whether locating a factory,opposing the siting of a landfill site, choosingwhere to go on holiday, whatever): these arefundamentally different research practices ofinterpretation from those documented so farin our book.

Before turning to a discussion of what isinvolved in the practices associated withunderstanding meanings, it may be helpful tocontrast the concerns of this chapter withthose of the two previous ones in a little moredetail. First, part of our focus here is on whatthe anthropologist Clifford Geertz (followingGilbert Ryle) calls ‘thick description’ (Geertz,1973: 3–30). Indeed, in an early plan for ourbook this chapter was entitled ‘Geographicalthick description’. We will discuss Geertz’sapplication of this idea in more detail below,but, briefly, the concept of thick descriptionemphasizes the complex layers of meaningthat can attach to what are often apparentlysimple social behaviours.This richness is typ-ically only apparent to ‘insiders’ in a givensocial situation who are able to pick up on thevariegated subtleties of meaning because of


308

3151-Ch-10.qxd 3/2/04 2:46 PM Page 308

their intimate shared knowledge of the socialconventions, language and informal codesbeing used for communication within aspecific social group. In this approach thework of the anthropologist (or by extensionthe social or cultural geographer) involvesbecoming an insider (at least on a temporarybasis) in order to learn the conventionsthrough which the deeper meanings are con-veyed. Understanding in the sense of ‘thickdescription’ can thus be contrasted with the‘thin description’ that, we suggested inChapter 8, was the realm of enumeration.Thisis not to decry enumeration but it is to rec-ognize that the principal strength of enumer-ative methods is their ability to providecoverage in breadth rather than to develop therichly descriptive depth of the Geertzianapproach. We should perhaps note in passingthat the idea of social worlds having thismeaningful depth is itself not without its critics,a point to which we will return later.

Secondly, we can contrast the idea of under-standing with that of explanation elaborated inChapter 9. The distinction between verstehen(understanding) and erklären (explanation) ismost closely associated with two Germanthinkers: philosopher Wilhelm Dilthey(1833–1911) and sociologist Max Weber(1864–1920) (see Weber, 1962). If erklärenequates with the practices noted in Chapter 9,verstehen is used ‘to denote understanding fromwithin by means of empathy, intuition, orimagination, as opposed to knowledge fromwithout, by means of observation or calcula-tion’ (Burke, 1988: 894, emphasis added).Thedistinction can be loosely mapped on to anumber of others, including the distinctionbetween subject(ive) and object(ive), or thatbetween ‘lifeworld’ and ‘system’. JürgenHabermas (1972) usefully distinguishes betweenempirical-analytic and historical-hermeneutic formsof knowledge, and he argues that the produc-tion of knowledge is never neutral but ratheralways linked to particular social interests. In

broad terms, the empirical-analytic form ofknowledge is concerned with an object world(‘system’) and is associated with the goal ofattaining technical control over that world. Bycontrast, the historical-hermeneutic form ofknowledge is concerned with a subject world(‘lifeworld’) and is associated with the goal ofaiding mutual understanding between humansubjects occupying this social world. For thepurposes of our discussion of different modesof geographical inquiry, we can also see expla-nation as concerned with discovering whatproduces (leads to/causes/generates) a parti-cular phenomenon, whereas understandingtreats the phenomenon under investigation asitself a source of meanings (ideas/concepts/relationships) that need to be interpreted.

In this chapter we start our discussion ofthe practices of understanding in the specificsense of the interpretation of meanings, a tra-dition of inquiry also known in some circlesas ‘hermeneutics’. The word hermeneuticscomes from the Greek work hermeneus (‘aninterpreter’).The term was originally used torefer to the interpretation of biblical texts butwas brought into philosophy in the nineteenthcentury by Dilthey (who was also the first touse the ideas of verstehen and erklären in thesenses discussed above).The original emphasisof hermeneutics as the interpretation of (par-ticular kinds of ) written texts is one that hasbeen retained in geography, as in other disci-plines which are concerned with written lan-guage, and as a result at various points in whatfollows we will couch our discussion in dis-tinctly textual terms, thinking about ‘authors’,‘texts’ and the ‘reading of texts’. By this tex-tual model, authors wish to convey meaningsof various kinds through their texts, employ-ing all manner of ‘means’ to do so spanningfrom the very words themselves to countlessstylistic devices, and the meanings writteninto their texts are potentially available forreaders to discern, more or less faithfullyin line what the authors had intended, and

UNDERSTANDING

309

3151-Ch-10.qxd 3/2/04 2:46 PM Page 309

perhaps then to inform the lives and actionsof these readers. (Many now critique thesimplistic assumptions about how meaning isproduced and circulated that inheres in such amodel, as we tackle below when talking about‘the deconstructionist’.)

Dilthey broadened the task of hermeneu-tics, however, so as to include the interpreta-tion of the meaningful character of not justtexts per se but of all manner of humanactions, intentions and institutions.This activity,striving for a verstehen or understanding ofthe overall human world, was held to be theappropriate method of the Geisteswissen-schaften (‘sciences of the [human] spirit’),while the Naturwissenschaften (‘sciences ofnature’) were left to tackle the more familiargoal of explanation.1 More broadly con-ceived, therefore, hermeneutics has to do withthe recovery of the meanings that are pre-sumed to be present in all kinds of humansituations,whether expressed in verbal languageor not. This partly reflects the widespread (ifnot always unproblematic) extension of theconcept of text to include a range of expres-sive human products and activities, such that(for example) paintings, films, landscapes andeven social life in general come to be definedas texts for methodological purposes.Again, inour discussion below, such problems regardingthis extension of the textual metaphor willnever be far below the surface, although wefully support the claim that understandingmust be about far more than just careful read-ing of actual texts and must entail anencounter with meanings proliferating any-where and everywhere in the social world.

Seven modes ofunderstanding

So far, then, we have established that ‘under-standing’ involves the recovery of the mean-ings present (or presumed to be present) in

written texts, human utterances and in otherkinds of human artefacts and activities.Human geographers have adopted a range ofdifferent kinds of practices that relate to theidea of understanding. In this section of thischapter we duly consider seven different‘modes of understanding’ by looking at thepractices of seven different kinds of researchersor figures: ‘the critic’, ‘the artisan’, ‘the ethno-grapher’,‘the iconographer’,‘the conversation-alist’,‘the therapist’ and ‘the deconstructionist’.Each of these figures treats the problem ofunderstanding somewhat differently, adoptingrather different techniques and approaches,and giving rise to rather different kinds ofhuman geographical research and writing as aconsequence. Taken together, though, theyprovide a repertoire of practices throughwhich human geographers can engage withthe meaningful qualities of human life inthe social world. This being said, the modesof understanding indexed by these sevenfigures do overlap, are not entirely exclusiveof one another and can be to some extentcomplementary.

What we might also say by way of prelimi-nary clarification is that, broadly speaking, thefirst four figures all adopt a similar definitionof meaning as something that inheres in thetext/situation under study, and that the mean-ings involved can be extracted or recoveredthrough the application of appropriate skillsor expertise.The next two figures extend thisapproach to meaning, emphasizing the extentto which the specific meanings upon whichthe researcher works cannot but be generatedin the context of interactions between theresearcher and his or her research subjects (thepeople at the heart of the human geographi-cal situation under study). The last figureadopts a more questioning attitude to the verypossibility of meaning itself, however, and inso doing reflects some of the most challengingfrontiers of contemporary human geographywherein much that has been routinely taken


310

3151-Ch-10.qxd 3/2/04 2:46 PM Page 310

for granted by those attempting to understandthe human realm is pulled out for criticalreassessment.

The critic

Our first figure is ‘the critic’ or, as we are alsotempted to term it,‘the man of letters’.We usethe gendered term deliberately to emphasizethe specifically masculine subject positioninvolved.The archetype that we have in mindis that of a particular kind of literary critic,one perhaps personified above all by SirArthur Quiller-Couch (1863–1944). Quiller-Couch was a novelist, poet, essayist and liter-ary critic and Professor of English Literatureat the University of Cambridge. In his inau-gural lecture at the university on 29 January1913, Quiller-Couch set out the principlesthat he thought should guide his work as acritic. Foremost among these is the view thatthe purpose of studying any great work ofliterature is to discover the meaning as intendedby the author:

I put to you that in studying any work ofgenius we should begin by taking itabsolutely; that is to say, with minds intenton discovering just what the author’s mindintended; this being at once the obviousapproach to its meaning …, and the merestduty of politeness we owe to the great manaddressing us. We should lay our mindsopen to what he wishes to tell, and if whathe has to tell be noble and high and beauti-ful, we should surrender and let soak ourminds in it. (Quiller-Couch, 1916, emphasisin original)

This emphasis on ‘authorial intention’ involvesa number of assumptions. First, it assumes thatthe author has a clear, coherent and consciousintention. Secondly, it assumes that he or she isable to express it in words.Thirdly, it assumesthat the meaning of the text is the same as themeaning intended by the author, and that thetext is thus a neutral medium for the transmis-sion of the author’s message.This in turn takes

it for granted that the meaning of the text isnot influenced by the social and political con-text in which it is written, and nor is it influ-enced by the nature and limits of languageitself (except in so far as these lie under thecontrol of the author).As we shall see, all theseassumptions have been challenged by morerecent literary theorists, especially by thosewriting since the 1960s.

The other aspects of Quiller-Couch’sapproach revealed in the quotation above arethe emphases on the ‘greatness’ or ‘genius’ ofthe author, an apparently ineffable and inex-plicable quality in his insight, and on the sup-posed nobility and beauty of literary art.Thisfaith in the essential goodness of literature iseven clearer in a later passage in the samelecture:

[B]y consent of all, Literature is a nurse ofnoble natures, and right reading makes afull man in a sense even better than[philosopher Francis] Bacon’s; not replete,but complete rather, to the pattern forwhich Heaven designed him. In this convic-tion, in this hope, public spirited menendow Chairs in our Universities, sure thatLiterature is a good thing if only we canbring it to operate on young minds.(Quiller-Couch, 1916)

One of the difficulties with Quiller-Couch’sapproach is that, while the discovery of theauthor’s intended meaning is seen as the pri-mary goal of reading, it is very unclear how, inpractical terms, such discovery is supposed totake place. It took another Cambridge critic,I.A. Richards (1893–1979), to set out anapproach to ‘practical criticism’ that for thefirst time provided a clear methodology forthe recovery of the meanings that wereassumed to lie embedded in literary texts.Richards’s method is set out in his bookPractical Criticism which, since its publicationin 1929, has been very widely used inEnglish-speaking universities to teach thetechniques of literary criticism. Richardsadvocated interpretation based on ‘close reading’

UNDERSTANDING

311

3151-Ch-10.qxd 3/2/04 2:46 PM Page 311

of texts. He argued that the reader shoulddevelop judgements about the literary meritsof a work without being influenced byknowledge of the status (or even the identity)of the author, nor by concerns with the socialand political context in which a text has beenwritten.

A further development of close readingcame in the work of another Cambridgeliterary critic, F.R. Leavis (1895–1978). Leavis(and the journal Scrutiny, which he founded)took a strongly moral approach to literature,combining Richards’ practical criticism witha judgemental style that won him many ene-mies in the literary and academic worlds.According to Leavis the role of the literarycritic was not only to interpret the meaningof a work but also to come to a critical eval-uation of its artistic merit and moral worth.He was thus concerned with what he saw asthe social and cultural benefits of readinggreat literature, and he had strong (andunorthodox) views about which literaryworks really merited the label ‘great’. Leavisstressed values such as spontaneity, ‘authentic-ity’ in interpersonal relationships and a ‘rever-ent openness before life’, and he then judgedliterary texts according to whether or notthey promoted such values. Moreover, Leavisthought that making such judgements in aneffective way required the critic him- or her-self to have these personal attributes, andhence to have a kind of cultivated sensibilitythrough which great literature could beappreciated. His approach was thus unasham-edly elitist, a far cry from later development incultural studies stressing the importance ofboth popular culture and everyday ways ofunderstanding.2 It also focused on supposedlyuniversal human values and, in that sense, canbe seen as seeking to transcend the individualauthor and his or her intentions and torecover the more universal meanings revealedin particular ways in different texts. Ingeography the work of some early humanistic

geographers on geography and literaturecomes close to the approach of the critic orman of letters (see, for example, the contri-butions to Pocock, 1981a). Indeed, the inter-pretation of imaginative literature wascentral to the development of humanisticgeography more generally. For Pocock(1981b: 10–11):

It is the deliberately cultivated subjectivityof the writer which makes literature litera-ture and not, say, reporting. It is the work ofthe heart as well as the head; it is emotion,often recollected later, perhaps ‘in tranquil-lity’, when an earlier stimulus is reworkedand given expression, in the manner per-haps that the painter in his studio maydevelop his quickly pencilled field sketch …The truth of fiction is truth beyond merefacts. Fictive reality may transcend or con-tain more truth than the physical or every-day reality.

In Pocock’s respect for the ‘cultivated subjec-tivity of the writer’ and his claim that fictionreveals a ‘truth beyond mere facts’ it is not dif-ficult to see an echo of the approach of aLeavis or a Quiller-Couch.

The artisan

Like the man of letters the ‘artisan’, oursecond figure, works through a detailed andclose reading of texts. In this case, however,the goal is to understand the world as it isrevealed in the everyday experiences, encoun-ters and utterances recorded in written textssuch as interview transcripts, letters, diariesand notebooks (see Chapters 5 and 6).Whereas for the man of letters the purpose ofclose reading is to uncover what are presumedto be universal truths about the human con-dition inhering in the text, for the artisanclose reading involves a detailed understand-ing of the points of view, subjective experi-ences and ‘local knowledges’ (see below)of specific individuals and groups located in


312

3151-Ch-10.qxd 3/2/04 2:46 PM Page 312

particular contexts in time and space. For theartisan, therefore, the goal is not so much totranscend the author by recovering universalhuman meanings but to stand in the place ofthe author and to see the world through hisor her eyes

We have chosen to use the notion of arti-sanal practice to highlight those techniques,skills and procedures that can, in principle, beacquired by any interested and conscientiousresearcher. In other words, they do not requirethe researcher to have some rather mysteriousand especially refined sensibility as seems to bethe case with the man of letters. For the arti-san, while understanding undoubtedly requireshard work, it is a craft that can be learnt ratherthan a talent with which one is born (or not,as the case may be). Of course some practi-tioners will be better at it than others and, aswith any craft, skills improve with practice andexperience. That said, the basics are relativelystraightforward and it is an approach that canreasonably be adopted in, for example, anundergraduate student’s research project.This isnot in any sense to disparage the artisan – onthe contrary, we absolutely want to stress thatthis is a highly skilled activity but also thatthere is no reason why even the most inexpe-rienced researcher should not begin to acquireand practise the skills involved. There are anumber of different kinds of techniques thatcan be employed to gain the kind of empa-thetic understanding that the artisan seeks.Perhaps the best known and most widelyapplied (not least in human geography) is theapproach to qualitative analysis developed bythe American sociologist, Anselm Strauss(1916–96), and his erstwhile collaborator,Barney Glaser. Strauss called his approach‘grounded theory’ to emphasize that a socialresearcher’s concepts about the world(‘theory’) should always develop from and havetheir roots – and thus be ‘grounded’ – in theconcepts voiced by the people whose livesand activities are being studied. Analysing

qualitative data, such as a set of interviewtranscripts, can seem like a very daunting task.Grounded theory provides a step-by-step pro-cedure that both makes the task more feasibleand aims to ensure that the conclusions drawnby the researcher are validly supported byempirical research evidence.

There are a number of different versions ofgrounded theory (for example, the approachesof Glaser and Strauss diverged after their earlycollaboration). However, at the heart of theprocess lies the activity of ‘coding’. Codingbegins with an initial careful reading of the textbeing analysed, which could be a newspaperarticle, an interview transcript or a focus grouptranscript.The next step is a version of ‘siftingand sorting’ the raw material of this text (seealso Chapter 7), working through the textquite systematically, identifying sections of it(typically phrases, sentences or paragraphs) thatare relevant to the research and tagging thesewith shorthand labels or ‘codes’. For example,in response to a question about attitudes to theenvironment, an interviewee might say ‘envi-ronmental conservation is very important tome’. This phrase might be labelled with thecode ‘conservation’. The same code can thenbe applied to other text segments that touchon the same issue. Another phrase about therisks associated with climate change might begiven the code ‘risk’ or ‘climate change’ orboth.When the whole text has been coded theresearcher will have generated a list of codesand a set of associated text segments or quota-tions (Box 10.1 gives an example of a codedsegment of an interview from the work of thegeographer Meghan Cope). At this stage theartisan is already in a position to use the text innew ways. For example, it would now be pos-sible to pull out all the quotations that share aparticular code so that the key themes of thetext can be identified. Alternatively, one couldfind all the quotations that have been codedwith both ‘risk’ and ‘climate change’ to isolatean interviewee’s comments about the risks of

UNDERSTANDING

313

3151-Ch-10.qxd 3/2/04 2:46 PM Page 313

Box 10.1: Coding an interview

In vivo codes/Text description Analytic codes

Q: What do you expect tosee happen as people hit[welfare] time limits or arecut off from programs?

A: Yeah, the safety net safety net issue ofthing. I see them coming ‘creaming’ –into that. I suspect, and the early success ofbig picture here if you look those ready toat the [welfare] population workapproximately a third of theindividuals can probably getwork on their own with a who can worklittle bit of help, anotherthird need some fairly who needs services role of jobintensive services, basically training andall the services that the support employment and training servicesand other human servicescommunity can providethem, and then I believe –and this is my own opinion –that there is probablyanother third or so thatreally are going to have a ‘hard to reach’very very hard time of who is unable to work populationthings after the five yearlimit. They may be in asituation where they’rereally not going to be ableto work. There are somepeople who have seriousproblems if you want to callthem problems, if you want respondent isto label them as such, that sensitive toare going to prevent them ‘labeling’from working.

Source: Interview conducted in July 1997 by Meghan Cope with a staff member of asocial service organization in Buffalo, NY

Source: Cope (2003: 453)

climate change from those concerned withother kinds of environmental risk.

Codes can be derived from terms used inthe text (in the case of an interview transcript

that will mean the actual words spoken by theinterviewee). Such codes are called in vivo or‘emic’ codes. In the example above ‘conserva-tion’ is an emic code because it is a word


314

3151-Ch-10.qxd 3/2/04 2:46 PM Page 314

taken from the text itself. By contrast, ‘etic’codes are words or phrases that do not appearin the text itself but are chosen by theresearcher because they refer to his or herresearch questions or conceptual framework.Thus, if our hypothetical interviewee hadspoken of being ‘worried about global warm-ing’without using the words ‘risk’ and ‘climatechange’, ‘risk’ and ‘climate change’ would beetic, not emic, codes. It might be argued thatthe truly grounded grounded theorist shoulduse only emic codes as a means of keeping hisor her coding (and thus the subsequent develop-ment of his or her theories and concepts) asclose as possible to the sense and spirit of theinterviewees’ worldviews. In practice, how-ever, many commentators question whetherthe distinction between ‘emic’ and ‘etic’ codescan really be sustained. Although emic codesmay adopt the words of the research subjects,it is still the analyst who selects which specificwords are to be used out of the hundreds spo-ken in the course of a typical interview, andsuch a selection cannot but be influenced bythe worldview, interests and existing knowl-edge and concepts of the researcher. Thus it isprobably best to see coding as a constantmovement between the concepts of theresearcher and the interviewee, and it is onlya short step from here to the ‘dialogical’model of understanding embodied in thefigure of the ‘conversationalist’ (see below).Moreover, coding usually entails an iterativeapproach in which an initial analysis providesthe basis for a return to the field (see Chapter 6)for further research. This might involvefurther interviews with the same person orothers, which then provide more material forinterpretation.

Having undertaken this initial phase ofcoding, which is sometimes referred to as‘open coding’, the next stage consists of iden-tifying the links between the codes and cate-gories and combining them through so-called‘axial codes’ that express different kinds of

relationships (such as X ‘is an example of ’ Y,or X ‘is caused by’ Y,or X ‘contradicts’ Y).Oneway to present these relationships is in a dia-gram or ‘code map’. Cook and Crang (1995)provide an example of a code map taken fromMike Crang’s research, which is shown inFigure 10.1. The unpolished nature of themap (with hand-written labels and variousdeletions and additions) reveals its role as atool of research in progress.The final stage ofcoding is ‘selective coding’ in which one ormore core codes are identified to form thefocus of the researcher’s own written study,and it is when all other codes and subcodesare systematically related to these selectivecodes that one is said to have generated‘grounded theory’.

What grounded theory offers is a detailedand empirically rich interpretation of thesocial situation under investigation deeplyanchored, so it is claimed, in the understand-ings of the participants as revealed in inter-views or other accounts. It is artisanal in thesense that it provides a set of procedures thatoffer clear guidelines for practice, withoutbeing overly rigid or prescriptive. Under-standing human geography through thestrategies of textual coding is thus notmechanical and it does require the creativeinput of the researcher, but it is not mysteri-ous. On the other hand, it can be somewhatlaborious, which is why for many practition-ers an obvious development of these tech-niques has been the use of computer softwareto automate some of the more routine aspectsof the work (van Hoven, 2003). Softwarepackages typically allow the uploading oftexts into what is effectively a specializeddatabase, and then provide a graphical inter-face to allow on-screen coding. Thus textsegments can be selected using a mouseand codes assigned from a menu. One of themost useful features of such programs is theircapacity for rapid searching, allowing quickretrieval of phrases or longer quotations based

UNDERSTANDING

315

3151-Ch-10.qxd 3/2/04 2:46 PM Page 315

on searches for individual codes or codecombinations (such as the ‘risk’ and ‘climatechange’ example mentioned above). Otherfunctions might allow various forms of con-tent analysis (working out how often certainterms or codes appear, for example), moreeffective memoing or appending of notes

about a specific portion of interview text, andthe organization of collaborative research byenabling remote access by several researchersto the same store of textual materials and thesharing of code lists and results.

Although now a little dated, the only com-prehensive comparative survey of the various


316

Figure 10.1 A 'code map'

Source: Cook and Crang (1995: 89)

3151-Ch-10.qxd 3/2/04 2:46 PM Page 316

packages available (as well as advice abouttheir use) is that provided by Weitzman andMiles (1995). More recently,Weitzman (2000)has outlined the main advantages and disad-vantages of using computers in this way andhas summarized the five main types of pack-ages available:

1 Text retrievers that identify all the instancesof a particular word or phrase in a text ortexts.

2 Textbase managers that organize and storetexts in a database and allow them to besorted according to various criteria.

3 Code-and-retrieve programs that allow theresearcher to apply codes to sections oftext in a similar way to the manual processdescribed above and then to search thetext by code.

4 Code-based theory builders that have addi-tional functions to support the develop-ment of theoretical propositions andhypotheses based on searching largeamounts of coded text.

5 Conceptual network builders which facilitatethe construction and display of networkedvisualizations of codes, rather like the codemap in Figure 10.1.

While the use of computer software canundoubtedly assist with the frequentlylabour-intensive work of coding (and some ofthe additional functions, such as network dis-plays of code relationships, are well worthexploring), the development of computer-aided qualitative analysis has been seen bysome researchers as a rather mixed blessing.Crang et al. (1997), for example, while gener-ally supportive of the use of these packages,counsel against fetishizing the technical capa-bilities of software tools over the substance ofthe analysis.

Inevitably the primary focus of this kind ofanalysis has been textualized information ofvarious kinds and it works best in relation to

fairly specific research questions in relation toclearly defined social settings.However, not allqualitative research is of this form. Researchersseeking a more holistic understanding of thecomplex flows of social practices that makeup the ‘way of life’ of a whole community aremore likely to adopt the participant observa-tional methods discussed in Chapter 6. Therecord of such observational research is typi-cally a field diary or field notebook. Ofcourse, in principle a researcher could ‘codeup’ his or her own notes or diary entriesfollowing the process outlined above.We suspectthat in practice most field notebooks are notanalysed in that way, though, but rather aretreated as the first drafts of the ‘thick descrip-tions’ espoused by the anthropologist CliffordGeertz. It is to Geertz’s approach that we nowturn through a consideration of our thirdfigure: the ‘ethnographer’.

The ethnographer

The third figure we wish to consider is thatof ‘the ethnographer’, although we shouldimmediately acknowledge that the artisan andthe ethnographer are very closely allied interms of their approach to the domain ofmeaning. In practice, the two figures can endup appearing much the same in their closeattention to the meanings circulating in every-day human situations, and a number of humangeographers today would probably claim to bemoving between the two possibilities (indeploying artisanal tools to achieve ethno-graphic effects). There are important differ-ences, however, and tracing these helps toclarify different steps that can be taken on theway to the goal of understanding. While theartisan is arguably more inspired by the inter-pretative garb of modern sociological practice,taking on board the notions of groundedtheory and the tools of coding,the ethnographeris more likely to trade on anthropologicalsenses of the researcher’s intuitive ‘sensitivity’

UNDERSTANDING

317

3151-Ch-10.qxd 3/2/04 2:46 PM Page 317

to what it is that events and phenomena endup meaning in the lives of particular people.While the artisan is more concerned with themeanings held by people regarding specificissues, such as their feelings about the closureof a school, the persecution of asylum-seekersor the problems of water supply, the ethnogra-pher will usually be more concerned to recon-struct, as far as possible, the overall worldview – thebroader sets of meanings about love, life, death,history, politics and so on – possessed by anidentifiable human grouping occupying adefinite place in the world.

Thus, whereas the artisan will examine datafrom documents, interviews and other sourcesfor clues about a relatively restricted set ofmeanings, albeit often recognizing how suchmeanings are framed by more wide-rangingbeliefs, the ethnographer will want to pulltogether as many threads of data as possible –often derived from conversational and practi-cal interactions over long periods with particularpeople in a particular place – so as to discernsomething of the entangled patterns of mean-ing constitutive of shared and relativelyall-encompassing worldviews. In the textuallexicon, it might be reiterated that the artisanis aiming to stand in the place of the author,hoping to triangulate the meanings held by avariety of individual authors relative to anevent or phenomenon such as the openingof a superstore or the marketing of a city.Particular meanings here attach quite clearlyto particular authors, and the approach is quiteatomistic,with the artisan peppering his or herown account with the words of different indi-viduals apparently voicing distinct and some-times contradictory meanings (and the artisanis always on the look out for cleavages withinthe domain of meaning). Instead, whilerespecting the individual and sometimesrepeating the words of named individuals, theethnographer is happier with the goal ofweaving together the stories told by many dif-ferent authors – by many different folks in the

setting under investigation – with the goal ofencompassing these many authors and manystories in one (relatively) seamless narrativeaccount of their shared worldview (their col-lective framework for making sense of theworld around them, human and natural).

In this latter approach it is almost as if thewhole span of everyday life within a situatedtime and place becomes regarded as a ‘text’, arich layering of stories pregnant with mean-ings one on top of the other, and that the taskof the ethnographer is to ‘read’ this text, todecipher its bundle of constituent meanings.This is then done less by identifying tensionsbetween meanings than by registering deeperresolutions whereby they do ultimatelycohere (if imperfectly) in some fashion. It ishere, therefore, that we encounter one ofthe most influential routes into thinkingabout social life as text: one that has clearlyshaped a host of developments within con-temporary human geography (arguably moreso than other routes into the social life as textmetaphor).

Such an approach does still throw emphasison the ethnographer as the privileged personable to perform the skilful act of decipheringthe mesh of meanings, and in this regard thereare overlaps with the role adopted by theman of letters.This being said, the ethnogra-pher allows him- or herself to be thoroughlyimmersed in the everyday hurly-burly ofpeoples and places, something that the man ofletters would abhor, and the ethnographer isin no doubt that it is only through being soimmersed – by being a character in the text ofsocial life, not divorced from it or transcen-dent of it – that he or she will be able toappreciate the ‘local’ constellation of meaningsdictating why some events and phenomenaare reckoned meaningful and others mean-ingless. At this point, then, we should recallthat in Chapter 6 we have already discussed atlength the ethnographer, exploring the tech-niques of immersion, of participant observation


318

3151-Ch-10.qxd 3/2/04 2:46 PM Page 318

and field note-taking, through which he orshe constructs a wealth of data available forsubsequent interpretation. Putting aside thenot insignificant point that the ethnographercannot but commence the task of interpreta-tion long before returning from the field – heor she will constantly be trying to make senseof the situation in which he or she is partici-pating simply to get by, not to make mistakes,have one’s ‘cover blown’ and the like – whatwe need to spell out now is greater detailabout what the ethnographer’s mode of inter-pretation entails.

As an example of a prime exponent of themode of understanding we are tackling here,wewill mention the well-known social anthropol-ogist Clifford Geertz. It is actually Geertz whopopularized the term ‘thick description’, bor-rowed from Ryle, which we have already uti-lized in this chapter and elsewhere in the book.In many respects this notion – accenting theneed to achieve a real ‘thickness’, a richness orsubstance, in the ‘description’ (note, not expla-nation) provided – goes to the heart of how theethnographer approaches the meaningfulness ofthe human realm. In a chapter arguing for ‘aninterpret[at]ive theory of culture’,Geertz writesas follows:

The concept of culture I espouse … is essen-tially a semiotic one. Believing, with MaxWeber [see above], that man [sic] is ananimal suspended in webs of significance hehimself has spun, I take culture to be thosewebs, and the analysis of it to be thereforenot an experimental science in search of law[a standard explanatory approach: seeabove] but an interpretative one in searchof meaning. (1973: 5)

The ‘search for meaning’ is hence predicatedupon semiotic (see Box 10.2) assumptionsabout culture residing in ‘webs of signifi-cance’. Such webs are spun from meaningsattached to the occupants and goings-on inthe surrounding world, the attributions ofsignificance and insignificance, of meaning-fulness or meaninglessness: and all these are

the ingredients of what Geertz wishes tounearth in his fieldwork, usually immersed inthe everyday worlds – we sometimes say life-worlds – of particular peoples in particularplaces (of ‘others, guarding other sheep inother valleys’ – Geertz, 1973: 30). He presentsthese said ingredients as the food of theethnographer.

Geertz stresses that ethnography includesthe tools of ‘selecting informants, transcrib-ing texts, taking geneaologies, mappingfields, keeping a diary, and so on’ (1973: 6),but that it should never be reduced to suchtools nor to the techniques of, for instance,coding up interview transcripts. This meansthat he distinguishes the ethnographer fromthe artisan in a manner akin to that rehearsedabove:

[Understanding], then, is sorting out thestructures of signification – what Ryle callsestablished codes, a somewhat misleadingexpression, for it makes the enterprisesound too much like that of the cipher clerkwhen it is much more like that of the liter-ary critic – and determining their socialground and import. (1973: 9)

The reference to the ‘literary critic’ repeatsthat alignment with the man of letters alsomentioned above, and in this vein Geertz char-acterizes the apex of the ethnographer’s art asbeing the accomplishing of ‘thick descrip-tions’ which provide a ‘reading’ of everydaylifeworlds as ‘texts’ full of meanings. As heexpands:

ethnography is thick description. What theethnographer is in fact faced with … is amultiplicity of complex conceptual struc-tures, many of which are at once strange,irregular, and inexplicit, and which he [sic]must contrive somehow first to grasp andthen to render … Doing ethnography is liketrying to read (in the full sense of ‘constructa reading of’) a manuscript – foreign, faded,full of ellipses, incoherencies, suspiciousemendations, and tendentious commen-taries, but written not in conventionalisedgraphs of sound but in transient examplesof shaped behaviour. (1973: 9–10)

UNDERSTANDING

319

3151-Ch-10.qxd 3/2/04 2:46 PM Page 319

Box 10.2: Semiotics

Semiotics is the study or interpretation of signs and symbols. The founder of semioticsis usually regarded as Ferdinand de Saussure (1857–1913). Saussure argued that a signconsists of two elements: the ‘signified’ and the ‘signifier’. The signified is a concept oran idea and the signifier is the form taken by the sign (for example, the word ‘tree’ isa signifier and the concept of a tree that the signifier points to is the signified; togetherthey make up the sign for ‘tree’). Note that the signified (the concept ‘tree’) is not thesame as the physical specimen of a tree growing in the ground. Semioticians use theterm ‘referent’ to indicate the physical object the sign relates to.

One of Saussure’s key insights was that signs are arbitrary. That is, there is no auto-matic or inevitable connection between a particular signifier and a signified. Thus twopeople might use different signifiers (e.g. ‘tree’ and ‘bush’ to indicate the same ideaand to refer to the same physical specimen). For people over the age of 45 the sign‘wicked’ usually refers to things that are evil. By contrast when today’s teenagers say‘wicked’ they usually mean ‘excellent’. This ‘arbitrariness of the sign’ means that themeanings of words and images are not transparent and self-evident: this is one reasonwhy they require interpretation. Signs can convey meanings in two ways, through‘denotation’ and ‘connotation’. Thus the word ‘green’ denotes the colour that comesbetween yellow and blue in the rainbow spectrum, but it connotes a whole set of otherideas such as ‘envy’, ‘nausea’ (as in green face), ‘springtime’ (green shoots), ‘environ-mentalism’ (as in Green Party), ‘naivety’, ‘Ireland’ and so on. Here the interpretation ofmeaning is more like using a thesaurus than a dictionary. Authors have a lot of free-dom to choose between many different possible words to describe or explain the samephenomena. The denoted meaning may be the same in each case but the connotationsmay be quite different and even contradictory. Moreover, no author can think throughall the possible connotations of every word and phrase before making a selection, sotexts are always open to multiple interpretations – many of which may not have beenintended by the author.

Semiotics also stresses the relational character of meaning. The meaning of any signis constituted through its differences from all the other signs in the system. We learnthe meaning of ‘green’ by coming to understand how it differs from other colours.People with certain kinds of colour blindness cannot understand ‘green’ in the sameway because they cannot see all the differences that give ‘green’ its meaning.

According to some forms of semiotics, the production of a text involves the ‘encoding’of meaning in the form of a pattern of signs, while the interpretation of a text involvesthe ‘decoding’ of those signs. However, because of the arbitrariness of the sign and theinfluence of social context on processes of encoding and decoding, no interpretationcorresponds perfectly to the meanings encoded during textual production.Communication thus involves the transformation of meaning rather than its simpletransmission. When an interpretation is written down, meaning undergoes a furthertransformation, or ‘recoding’.3

As it happens, Geertz clearly does attend to‘graphs of sound’, to the words and storiesuttered by his field informants and acquain-tances. As a result, the metaphor of readingthat which is not straightforwardly written yetwhich still clearly means so much – a text ofconduct, movement, gestures, struggles, anger,

laughter and tears, together with the spokenproducts of an oral culture – is one that hassince come to pervade much of contempo-rary human geography. We provide somemore flavour of Geertz’s mode of understand-ing in Box 10.3, where we follow hisencounter with the Balinese cockfight, but let


320

3151-Ch-10.qxd 3/2/04 2:46 PM Page 320

Early in April 1958 Geertz and his wife arrived in a Balinese village and were soongreeted to the unnerving spectacle of a large and illegal cockfight being held in thepublic square to raise money for a new school:

Attempting to meld into the background [Geertz and his wife] watched theBalinese villagers watching the cocks fighting in the ring. All of a sudden, thelocal police arrived, armed with machine guns and looking menacing, and imme-diately the gathering dispersed: People raced down the road, disappeared head-first over walls, scrambled under platforms, folded themselves behind wickerscreens, scuttled up coconut trees … Everything was dust and panic. (1973: 414–15)

The Geertzs ran too, ending up tumbling into the courtyard of another fugitive,whereupon ‘his wife, who had apparently been through this sort of thing before,whipped out a table, a tablecloth, three chairs and three cups of tea, and we all, with-out any explicit communication whatsoever, sat down, commenced to sip tea andsought to compose ourselves’ (1973: 415).

One of the policemen marched into the yard, to meet fierce rebuttals from theGeertzs’ unexpected host, who declared that all three of them had been there all thetime drinking tea and discussing matters of culture. The next day, news of this eventhad travelled throughout the village, rendering the Geertzs an interesting novelty forteasing – locals mimicked their awkward running style – but also leading them to beaccepted and then allowed into the everyday rituals of village life in a manner thatmight not otherwise have been the case. Moreover, as Geertz elaborates:

It [this event] gave me the kind of immediate, inside-view grasp of an aspect of‘peasant mentality’ that anthropologists not fortunate enough to flee headlongwith their subjects from armed authorities normally do not get. And, perhapsmost important of all, … it put me very quickly on to a combination of emotionalexplosion, status war, and philosophical drama of central significance to thesociety whose inner nature I desired to understand. By the time I left [Bali] I hadspent about as much time looking into cockfights as into witchcraft, irrigation,caste or marriage. (1973: 417)

As a consequence, therefore, Geertz devoted much energy to reconstructing the mean-ings that the Balinese villagers derived from and possessed about the cockfights, partlyby hearing their own countless stories about such cockfights in the past, but also sim-ply by attending and carefully observing similar events in the present. As well as not-ing the double entendre of ‘cocks and men’ which is obvious in English but also presentin the Balinese village worldview, he arrived at the following generalizations about theplace of the cockfight in what might be termed ‘local knowledge’ (Geertz, 1983) oreven the local ‘economy’ of meaning:

What sets the cockfight apart from the ordinary course of life, lifts it from therealm of everyday practical affairs, and surrounds it with an aura of enlargedimportance is not … that it reinforces status discriminations (such reinforcementis hardly necessary in a society where every act proclaims them), but that it pro-vides a metasocial commentary upon the whole matter of assorted human beingsinto fixed hierarchical ranks and then organising the major part of collective exis-tence around that assortment. Its function, if you want to call it that, is inter-pret[at]ive: it is a Balinese reading of Balinese experience, a story they tellthemselves about themselves. (1973: 448)

Box 10.3: Clifford Geertz and the Balinese cockfight

UNDERSTANDING

321

3151-Ch-10.qxd 3/2/04 2:46 PM Page 321

The reference to the cockfight as a story is, of course, telling in that it plays upon theidea of social life as text, although it is interesting that Geertz suggests that it is notjust him – the clever anthropologist – reading this story, it is in effect a story that theBalinese villagers are telling themselves to affirm their own identity. We are certainlyin a sophisticated hermeneutical realm here, one where meanings are being producednot only by words but also by actions, and one where meanings are circulating invarious directions (between the villagers themselves and between them and the visit-ing professor of anthropology). By examining in microscopic detail the cockfights andby seeking answers to simple questions – who owns a fighting cock? in what order doowners fight their cocks? who is on-looking and who is gambling? – Geertz is indeedable to advance a ‘reading’ of the cockfight that can stand as a prime exemplar of theethnographer’s mode of understanding. We might also add that attention to the geo-graphies of the cockfight, to its staging in public space and to the bodily spaces of whois positioned where during the event within the square, can further enrich the accountthat is being given.

Box 10.3 (Continued)

us now briefly trace the movement of hisideas into the practising of (certain strains of )human geography.

Following James Duncan’s (1980) criticismof an older way of thinking about culture inhuman geography, one associated with CarlSauer (see Chapter 1) wherein Cultureappears as this undifferentiated ‘super-organic’entity somehow existing apart from the trans-actions of everyday life, the search for alterna-tive conceptualizations of culture almostinevitably took some geographers towardsGeertz’s take on culture as a web of meaningsin constant affirmation and change. As PeterJackson and Susan Smith (1984: 39) put it,‘[f]ollowing Duncan’s (1980) critique …, aninteresting debate arose in which Geertz’sinterpretative anthropology was alluded to asan attractive alternative to “superorganic” def-initions’. Miles Richardson (1981: 285) dulyfollows Geertz in supposing that cultureresides in how people talk to each other,thereby sharing with one another ‘in someincomplete way’ their ‘feelings about the past’,their ‘thinking about the present’ and their‘plans for the future’. Talk is not narrowlyconceived by Richardson, however, for hesupposes it to include not only spoken words

but also written words, bodily gestures,physical activities, mime, dance, ritual and themanipulation of material objects. Throughsuch expanded talk, people build up sharedmeanings about the world around them andsometimes beyond, and Richardson underlinesthat talk is indeed intersubjective – sharedbetween separate human subjects – andthereby social, as well as acquiring a measureof ‘tradition and permanence’ sedimentinginto a recognizable if slowly changeable ‘holis-tic pattern’. He also speculates that mucheveryday talk tends to be about materialobjects (artefacts from hoes to houses ortrainers to televisions) which assume a ‘sym-bolic’ quality in the sense of having pinned tothem a cluster of meanings with a ‘capacity’for impelling reactions from people whoencounter the said objects. In addition, as ageographer, he proposes that a given group ofpeople in a given part of the world, interact-ing with each other and the material objectswithin their environs, cannot but formulate ashared ‘image’ of their place:

Th[is] image is not the underlying schemetalked about by students of mental maps[people’s ability to represent, say, the layoutof a local neighbourhood] as being


322

3151-Ch-10.qxd 3/2/04 2:46 PM Page 322

antecedent to acts, but rather it is anexplanation, an interpretation arising fromour acts. The image may be explicit anddetailed or barely formed in our conscious,but in either case we ‘extended it out’ or‘hang it on’ the material setting so as to givea sense of place to that place. (1981: 286)

Our claim would be that, while not specifi-cally following Richardson’s lead, much thathas since appeared in the guise of a ‘new’ cul-tural geography has effectively taken onboard such a Geertzian semiotic line onculture. It has become assumed that theeveryday production, circulation and con-sumption of meanings can, in principle atleast, be ‘read’ by the careful human geogra-phy researcher.

Such a sense of ‘reading’ meanings isdoubtless present in the work of the artisans,as discussed earlier, even if the ambitionof reconstructing overall worldviews or‘local knowledges’ has featured less than therecovery of meanings swirling around specificissues. None the less, some geographers –notably ones working on the boundariesbetween geography and anthropology such asDonald Moore (1997) or Tad Mutersbaugh(1998), and most likely researching peoplesand places beyond the West – have sought tounderstand the human realm in a fashionakin, albeit not always explicitly referencing,Geertz’s writing of ‘thick descriptions’ thatembrace the rich totality of local culturalmeanings. Moore’s patient inquiries into thestate-administered Kaerezi ResettlementScheme in Zimbabwe’s Eastern Highlands,fringing the border with Mozambique, mightbe an example. When he reflects upon the‘livelihood struggles’ of Angela, a woman inher late 50s, and her neighbours, running upagainst the demands of patriarchal, chiefly andstate power, a host of meanings come intoplay that evidence the extent to which ‘placematters; it has a politics and is produced throughmyriad and symbolic struggles’ (Moore, 1997:103, emphasis in original).

Another example might be Mutersbaugh’ssimilarly patient inquiries into the Oaxacan(Mexican) indigenous community of SantaCruz,where he lived for a year with his spouseundertaking intensive ethnographic work fullof interviews, conversations and participantobservation to establish the gendered dimen-sions of labour within the overall village order.‘Labour–organisation discourse encompassesboth techniques of power (sanctions, surveil-lance) qua material practices and negotiationof meanings of labour’, writes Mutersbaugh(1998: 440), all of which depend upon ‘sociallyconstructed sites – twilit front porches, smokykitchens, the village assembly hall – wherevillagers converse and map out present andfuture projects’. It is true that both Moore andMutersbaugh are interested in wider consider-ations to do with power and resistance or gen-der, labour and technology, and are therebyasking particular questions of their study areas,but in the process they are clearly ‘reading’ the‘texts’ of local meanings. We would suggest,moreover, that in order to answer their ques-tions they are indeed striving to ‘read’ thewhole book of locally interwoven culturalmeanings, not just a few scattered pages. Assuch, they are very much, in our terms here,Geertzian ethnographers in their mode ofunderstanding human geographies.

The iconographer

The fourth figure we wish to consider is thatof ‘the iconographer’, and here we move toconsidering a mode of understanding dealingnot with words per se, whether written orspoken, but instead with images. As westressed in Chapters 3 and 4, visual images inphotographs, paintings, cartoons, films, televi-sion programmes and advertisements, to namebut a few possibilities, can all be sources ofgeographical data. Once a researcher hasdecided that he or she will draw upon images,it becomes important to establish just how

UNDERSTANDING

323

3151-Ch-10.qxd 3/2/04 2:46 PM Page 323

these images are to be interpreted so as toreveal matters of interest about the particularhuman situations under study.The researchermight elect to treat images in a formal man-ner, perhaps classifying and counting up par-ticular sorts of images appearing in, say, a runof newspapers as a prelude to reaching con-clusions about what meanings the images aresupposed to convey. A very specific examplemight be finding lots of images in a news-paper showing supposedly homeless vendorsof The Big Issue magazine apparently wearingexpensive trainers, leading the researcher tosuppose that the newspaper in questionwishes to convey a message about such ven-dors not necessarily being homeless and poorbut really ‘scamming’ it. As this indicates,however, it is hard to avoid the conclusionthat most (if not all) attempts to understandimages are ultimately about teasing out themeanings present within images, whether aspurposefully ‘coded’ into them by their pro-ducers or as ‘derived’ from them by their con-sumers (and there may be many differentconstituencies of consumer for the sameimages, each of whom may derive differentmeanings from the images).

It would, therefore, not be inappropriate tothink in terms of meanings being both ‘writ-ten into’ and ‘read from’ images, therebyextending the textual metaphor to the workthe human geographer might do in relationto images. It is easy to find instances of geog-raphers explicitly talking about ‘reading’images as ‘texts’, one instance being StuartAitken’s (1997) reference to ‘media texts’ and,more specifically, ‘film texts’ where there isnormally a confluence of the image (movingvisuals) and the written (the script) to gener-ate an overall text. Tim Cresswell offers anexample when he ‘reads’ the film FallingDown, extracting the possible meanings aboutpower, domination and resistance wrapped upin the story of how the mainstream whitedefence industry worker, D-Fens, turns into

the hero/anti-hero who,gun in hand,‘questionsurban constructions of space, time and moneyin turn as he wanders across the city’ (Cresswell,2000: 256).The meanings considered are boththose of the film-makers, who wished toemphasize that alienation in the USA fromthe forces of both capital and the city is notsolely a province of African-Americans, anddifferent audiences, such as the liberal criticswho saw it as a worrying expression of whitemale angst directed violently against womenand people of colour. Cresswell does not layout any explicit methodology guiding hisaccount of Falling Down – it might actually beregarded as tending towards being decon-structionist (see below) in its focus uponfaultlines of meaning in the film text – and itwould have to be acknowledged that it is onlyrecently that ‘visual methodologies’ (Rose,2001) have shifted on to the agenda of humangeography in a major way.

There is a significant exception to this rule,though, in that a handful of geographers con-cerned with landscape have for some timebeen influenced by an approach known as‘iconography’; and hence our naming of thefigure of the iconographer to cover this modeof understanding meanings in images.We willreturn to landscape presently but, for themoment, let us stick with iconography, whichbegan life (and here the link with the historyof hermeneutics is instructive) as the ‘inter-pretation of symbolic imagery’ (Daniels andCosgrove, 1988: 1) derived from the reper-toire of the Classical world and adaptedwithin the medieval Christian tradition. Theproject of iconography was revived in theearly twentieth century by the school of arthistory under Aby Warburg, and a summary ofwhat it came to entail runs as follows:

In opposition to the purely formalistic tradi-tion of art interpretation associated withHeinrich Wölfflin (which analysed picturespurely in terms of the surface patterns ofcolour, chiaroscuro, line and volume, relating


324

3151-Ch-10.qxd 3/2/04 2:46 PM Page 324

them principally to other works of art),iconographic study sought to probe mean-ing in a work of art by setting it in its his-torical context and, in particular, to analysethe ideas implicated in its imagery. While,by definition, all art history translates thevisual into the verbal, the iconographicapproach consciously sought to conceptu-alise pictures as encoded texts to be deci-phered by those cognisant of the culture asa whole in which they were produced. Theapproach was systematically formulated byWarburg’s pupil, Erwin Panofsky. (Danielsand Cosgrove, 1988: 2)

There are hints here of a distinction akin tothat between the practices of sifting, sortingand enumerating, possibly together with themore formal ‘coding’ exercises of the artisan(see above), and the practice of understanding –and most specifically that as pursued by theGeertzian ethnographer – which calls for tak-ing seriously the broad sweep of meaningsconstitutive of a given ‘culture as a whole’.For Panofsky, iconography (he actuallyfavoured the term ‘iconology’ in this respect)aimed to reveal the ‘intrinsic meaning’ of art-works as linked into the ‘attitudes’ of a givenperiod, national, religious or even class con-sciousness. To some extent, the successfulachievement of this task would depend uponthe skills of the iconographer since, as heacknowledged:

There were no established conventions orspecific methods that would ascertain theseprinciples; they were to be reconstructed bya kind of detective synthesis, searching outanalogies between disparate forms likepoetry, philosophy, social institutions andpolitical life: ‘To grasp these principles’,wrote Panofsky, ‘we need a mental facultycomparable to that of a diagnostician’.(Daniels and Cosgrove, 1988: 2)

The iconographer as the one who diagnosesthe complex cross-cuttings of meaningswithin images, tracing from them to the playof poetic, philosophical, social and politicalforces in the time and place of their production:

these are hence the traits requisite of thismode of understanding. Once again, and aswith both the man of letters and the ethno-grapher, it is possible to go only so far in item-izing the particular practices constitutive ofsuccessful work in this mode; it is possible togo only so far in learning the skills apparentlyneeded by the iconographer since, ultimately,a measure of individual creative response can-not be denied.

There are other traditions of image analy-sis, including the Marxist-inspired approachof someone like John Berger (1972), whichrecognizes that different ‘ways of seeing’ (ofcomposing and responding to images) arealways bound up with different forms ofeconomic, social and political organization. Inparticular, the likes of Berger suppose thatmodern capitalism has an ‘interest’ in deploy-ing images both to ‘mystify’ the masses (toobscure from them the real conditions oftheir existence) and to ‘sell’ things to them (toprompt them into the consumption of com-modities and experiences). This is not theplace to debate such traditions, however, andinstead we wish to underline, with StephenDaniels and Denis Cosgrove (1988: 3), that‘Panofsky applied this approach of “readingwhat we see” to built as well as to paintedforms’. In other words, he opened up thepossibility for an iconography of buildings,townscapes and, by extension, landscapes of allkinds. A landscape can be the actual visualscene before the researcher’s eye – the imagehere being what the eye latches on to whenbeholding the scene – or the representation ofa given landscape in a photograph, drawing orpainting.While entirely natural landscapes inthe world – if any remain – cannot be takenas deliberately encoded with meanings, exceptperhaps by God, most worldly landscapesbearing human imprints and all representedlandscapes are clearly open to an interpreta-tion of their meaningful contents as bothinscribed within and extractable from them.

UNDERSTANDING

325

3151-Ch-10.qxd 3/2/04 2:46 PM Page 325

In short, such landscapes are amenable tothe interpretative practices of a Panofsky, aBerger, a feminist critic (Rose, 2001) orequivalent, leaving the possibility of a humangeographical treatment of landscapes embed-ded within (and influencing) a more generalhuman geographical treatment of images. InChapter 4 we provided a familiar example ofcritiquing the meanings held within andprompted by a landscape image: that ofThomas Gainborough’s late seventeenth-century painting of Mr and Mrs Andrewsseated before the scene of their countryestate.

Cosgrove and Daniels’s The Iconography ofLandscape (1988) collection paved the way inthis regard, building upon past achievementsby Cosgrove (1984) when relating ‘symboliclandscapes’ to ‘social formations’ (or, puttingthings more cryptically, connecting spaces ofmeaning to places of material life).This bookalso anticipated future accomplishments byboth Cosgrove (1993), who ‘reads’ real andimagined ‘Palladian’ landscapes, and Daniels(1993), who ‘reads’ the cultivated, painted,cartooned and photographed landscapes inte-gral to English national identity. A morerecent collection edited by Cosgrove (1999)extends these principles to the ‘reading’ ofmaps from a host of different periods andplaces, showing how such maps, as very par-ticular kinds of images, have often been asmuch vectors of cultural meaning as expres-sions of factual scientific (geographical)knowledge.One chapter here (by Paul Carter)focuses on both maps and other images,including an engraving by W. Hamilton called‘The View of the Island of Ouby fromFreshwater Bay on Batchian’ taken fromThomas Forrest’s A Voyage to New Guineaand the Moluccas (1779) (see Figure 10.2). AsCarter discusses:

Th[is] remarkable engraving … may pose asan ethnographically inflected picturesqueview – in reality it assimilates the coast to

the exhibition diorama. Type specimens of amultiplicity of exotic objects present them-selves to the scientific gaze, washed-up,exposed, already detached from their envi-ronmental matrix, lost it seems unless theycan be classified and correctly arranged. Thenative collectors, handling physically whatthe English savant examines mentally, pleas-antly flatter the latter’s sense of intellectualsuperiority. (1999: 133)

This image is hence subjected to a criticalinterrogation – again, it teeters on being adeconstruction (see below) – exposing themeanings that the white European (English)educated viewer might bring to bear, onesinformed by an embryonic ‘scientific racism’and as bound into a philosophical apprecia-tion of how the products (or ‘trophies’) ofexploration get lifted from distant shoresready for cataloguing at leisure in the scholar’smuseum. Carter as iconographer thereby sig-nals the complexity of meanings that canindeed be ‘read’ from such an image, one thatat first might seem quite straightforward inappearance. It might therefore be argued thatwhat he provides (along with Cosgrove,Daniels and others) is akin to a Geertzian‘thick description’, here of an image (or aselection of images) rather than of a particularpeople in a particular place.

The conversationalist

We now move to two rather different figuresfor whom the act of understanding drawsaway from the assumption that the humanworld comprises texts and situations some-how ‘full of ’ meanings that are more or lessstraightforwardly available for interpretation bythe researcher. Actually, in practice, all theprevious four figures, perhaps excepting thecritic, have run up against problems with thisassumption. Indeed, whenever we have men-tioned researchers being aware of their ownrole in deciding what are the meanings pre-sent in a text or situation, as when the artisan


326

3151-Ch-10.qxd 3/2/04 2:46 PM Page 326

debates ‘etic’ coding, we have anticipated thefigures that we will now call ‘the conversa-tionalist’ and ‘the therapist’.

For reasons to be explained shortly, thefigure of the ‘conversationalist’ is meant tocharacterize researchers who are fully awarethat the meanings of which they speak – themeanings that become the basis for their ownstudies in ‘interpretative human geography’(Smith, D.M., 1988b) – cannot but be gener-ated in the context of a two-way ‘relationship’between researcher and research subjects.Revisiting the textual metaphor, the keyrecognition is that the reader cannot simply‘read’ out the meanings intended by the‘author’ since, in practice, what really occurshere is that meanings are produced preciselythrough the encounter between reader andauthor. Readers cannot avoid bringing to theact of reading what they find meaningful fromtheir own vantage point, their own ‘place’ inthe world, and, as such, there is no possibilityof them ever being that neutral conduit of an

author’s original meanings hoped for by theman of letters. Readers respond to an author’sintentions, picking up on some but probablynot all the meanings the latter has striven toimprint on the pages, and responses from read-ers may span the whole gamut of empathy,sympathy, indifference, hostility or even flyingoff on an wildly imaginative curve spurred by(yet in no sense contained by) the text beforethem. Seen in this light, meaning becomes amessy, fragile and multiple thing, devoid of thestability that a more scientific, explanatory takeon research would ideally require.Yet we thinkit vital and appropriate to identify a mode ofunderstanding that proceeds from a startingpoint wherein the relationship between theresearcher and his or her research subjects –whether ‘authors’ of documents, utterers ininterviews, actors in real events and places orwhatever – is directly akin to that of the readerresponding to an author.

We can now return to the ‘conversationalist’,a figure partially inspired by the participant in

UNDERSTANDING

327

Figure 10.2 ‘The View of the Island of Ouby from Freshwater Bay on Batchian’ by W. Hamilton

Source: Carter (1999)

3151-Ch-10.qxd 3/2/04 2:46 PM Page 327

something like the staged conversaziones of‘polite’ society, someone happily exchangingwith others not just chit-chat but more devel-oped opinions and thoughts on a variety ofweighty matters.An important feature of suchconversations is that, in principle at least, allparticipants are supposed to bring to the‘table’ their own informed viewpoints –respect and ‘space’ duly being accorded to allthe viewpoints present and expressed. Thismodel has a number of significant theoreticaloff-shoots, whether in Richard Rorty’s (e.g.1979) pragmatist call for intellectuals to ‘keepthe conversation going’ (not to cut off possi-ble positions and arguments) or in JürgenHabermas’s (e.g. 1984) ‘theory of commu-nicative action’ influenced by the image of aconversation-rich eighteenth-century ‘publicsphere’. For us, though, it simply introducesthe sense of the researcher who ‘converses’with his or her research subjects, the people atthe heart of a study being conducted, whetherliterally (as an interviewer or a participantobserver) or imaginatively (when reading awritten document or perhaps a transcript).More than this, though, for what such aresearcher also appreciates is the extent towhich it is during ‘conversations’ that the verymeanings integral to his or her research areactively produced.These meanings are henceneither solely the ‘property’ of the researcher,as it were, nor of the research subjects. Rather,both put their own viewpoints into the con-versation so that the meanings which emergereflect both parties and in effect are hybridpositions – summations of the ‘discussion’,resolutions of any ‘debate’ – irreducible toeither or any of those initially taken by par-ticipants in the conversation.

The figure of the conversationalist iscentral to thinking within the realm ofhermeneutics, as introduced above, and mostparticularly in the writing of Hans-GeorgGadamer, a German philosopher. One com-mentator summarizes Gadamer’s insistence

that ‘we’ as scholars cannot avoid ‘makingsense’ of whatever we study through themeaning-endowing lenses of who we are, whenand where we are living and what we have wit-nessed or learned about:

We never come upon situations, issues orfacts without already placing them withinsome context, connecting them to someother situations, issues or facts and, in short,interpreting them in one way or another. Theparameters of these interpretations, more-over, derive from our own circumstances andexperience, and these circumstances andexperience are always already informed bythe history of the society and culture towhich we belong. (Warnke, 1987: 169)

Gadamer usefully underlines that the mean-ings with which we make sense of the worldare never entirely personal but rather haveroots in, if not being determined by, manydimensions of the historically and geographi-cally specific ‘society and culture’ where welive. He nevertheless believes that this shouldnot render us closed to what others differentfrom ourselves, perhaps from other times andplaces, can tell us about themselves and theirworlds; and it is revealing that here he looksto ‘the structure of dialogue’ (Warnke, 1987:169) for guidance on what might be possible.‘As he depicts genuine conversation,’ contin-ues Georgia Warnke in her exegesis, ‘it israther characteristic of it that all participantsare led beyond their initial positions towards aconsensus that is more differentiated andarticulated than the separate views withwhich the conversation-partners began’(1987: 169).Warnke admits that the notion of‘consensus’ is troublesome in Gadamer’s workbut concludes that his most fruitful accountruns as follows:

[C]onsensus refers simply to a ‘fusion ofhorizons’, an integration of differing per-spectives in a deeper understanding of thematters in question. According to thissecond idea of dialogic consensus, one isrequired to take account of the positions ofothers in discussing an issue or subject-matter


328

3151-Ch-10.qxd 3/2/04 2:46 PM Page 328

with them. Here, even if one holds to one’sinitial point of view, one has nonetheless todeal with the objections, considerations andcounter-examples that others introduce. Inthe end, whether one changes one’s posi-tion or maintains it, the view that results ismore developed than the one with whichone began, and the same holds for theviews of all participants. Whether they con-clude by agreeing or disagreeing in a sub-stantive sense, their positions are nowinformed by all the other positions. They areable to see the worth of different consider-ations, incorporate different examples anddefend themselves against different criti-cisms. In this way their views acquire agreater warrant; they are less … one-sided.(1987: 169–70)

It is worth quoting this lengthy passage, partlybecause it so neatly builds upon the figure ofthe conversationalist but also because it strikesus as a highly apt description of a mode ofunderstanding that is now pivotal – if notquite expressed like this – to how a greatmany human geographers proceed wheninterpreting data from their encounter withresearch subjects.

For Gadamer, ‘hermeneutics is a form ofjustification involving the dialogic adjudica-tion of both beliefs and standards of rational-ity’ (Warnke, 1987: 170). Such an interest inestablishing the parameters of rationality isnot that relevant to human geographers, butthe model of conversation, dialogue andmediating between ‘horizons’ of meaning hasclearly excited several.Anne Buttimer talks of‘dialogue’ in her famous 1976 paper, forinstance, proposing that each ‘participant in adialogue needs to become aware of his [sic]own stance, and the stance assumed by theother, so that language for dialogue couldemerge, ie. be jointly created, or at least jointlyaccepted, by both participants’ (p. 278). Onone level she is asking for a dialogue betweengeography and phenomenology (a formalphilosophy) but, on another, she wishes toeffect a dialogue between the ‘world horizons’(1976: 281) – recall the reference above

to achieving a ‘fusion of horizons’ – of theacademic ‘stranger’ exploring the senses of‘social space’ integral to a ‘foreign culture’:

The knowledge acquired in one’s ownsociety is inadequate; one has to questionthe former ‘givens’ of social life and searchfor common denominators for dialoguewith the other. To gain a foothold, or basisfor dialogue, one needs to grasp the innersubjective meanings common to that othergroup, its sociocultural heritage, and its‘stream of consciousness’. (1976: 285–6)

Much the same construction, this timeexplicitly borrowing from the hermeneuticsof Gadamer, shapes Derek Gregory’s remarkthat ‘interpretation is not a process of some-how overcoming the distance between oneframe of reference and another’ (1978a: 145,emphasis in original). He contemplates thisissue chiefly in terms of the temporal distancethat confronts historical geographers (see alsoGregory, 1978b) but he accepts that the ‘dis-tance’ between researcher and research sub-jects can arise for other reasons too, perhapsspatially (as ‘us, here’ research ‘them, there’)but often due to social, cultural, bodily andother differences between the ‘positionalities’of researcher and researched (see also Chapter 1).Recognizing that there is merit in researchers‘immersing’ themselves in the worlds of theirresearch subjects, he adds that ‘this does notmean and cannot involve the substitution ofone frame of reference for another’ (1978a:145), quoting Gadamer in the process tothe effect that the researcher’s ‘placing’ ofhim- or herself in a situation does notrequire self-effacement but rather a fullrecognition of ‘ourselves’ – and our frameof reference – standing in a dynamic, eventense, encounter with the people whom weare studying.

We would argue that just such a sense ofwhat has variously been termed the‘hermeneutic circle’ or the ‘double hermeneu-tic’ has now become an article of faith for

UNDERSTANDING

329

3151-Ch-10.qxd 3/2/04 2:46 PM Page 329

many human geographers. The result is awidespread recognition that human geo-graphical research cannot but grow out of anexchange between two different bodies ofmeaning: that of the researcher, on the onehand, and that of the researched on the other.In-depth interviews where questioning andanswering becomes co-equal discussion; par-ticipant observation full of shared practicesand associated everyday talk; and researchersgetting research subjects to comment on tran-scripts, field diaries or summary reports – allthese common aspects of current humangeography accept and actively promote ashared ‘meaning production’ in the construc-tion of geographical data (see Chapters 5 and 6).Moreover, this sharing has almost become anunspoken assumption, the bridge betweenacknowledging one’s ‘positionality’ (Jackson,2000) in a research project – being self-awareand even self-critical of one’s own attributes,attitudes, biography and social contexts (seeChapter 1) – and seeking to unearth allmanner of meanings circulating ‘out there’ ina world of texts, images and situations availableto us for interpretation.Very few human geo-graphers would now take the stance adoptedby Leonard Guelke (e.g. 1974; 1982) in his‘idealist human geography’, wherein he statedthat the researcher is an empty, neutral vesselsimply finding and recording the ideas (thetheories or meanings) present in the heads ofresearch subjects (which for Guelke were usu-ally people from the past, such as the Dutchcolonizers of South Africa). Most humangeographers today would hence be in nodoubt that they are bringing their own mean-ingful baggage to a project – Harrison andLivingstone (1979; 1980) speak in theirresponse to Guelke of the geographer’s ‘pre-suppositions’ – and that somewhere in theheart of the research there must be an inter-mingling of meanings (‘ours’ and ‘theirs’). AsTrevor Barnes (2000c: 335–6, emphasisadded) writes:

the explicit working out of the hermeneuticalapproach has become less significant …Nonetheless, the spirit of hermeneuticalinquiry, that is, the recognition of the impor-tance of interpretation, open-mindedness,and a critical, reflexive sensibility, is as greatas it has ever been, and certainly evidentin the discipline’s recent cultural turnand interest in the poetics and politics ofrepresentation.

Barnes admits that not all that much has beendone recently by human geographers in termsof explicitly theorizing the hermeneutic bal-ancing act between the ‘frames of reference’held by both researcher and researched. Yetthe balancing act is now a familiar one, so hecontinues, such that all manner of work nowconducted by geographers within the ‘culturalturn’ (Cook et al., 2000b; Philo, 2000) orwhen discussing the ‘crisis of representation’(see Chapter 11) is indeed alert to the reflex-ive moment in interpretation. In short, thesegeographers worry deeply about how theresearcher’s interpretations of a research sub-ject’s interpretations of the world are inextri-cably locked within a play of meanings acrosslifeworlds, a conversation to be heard and itspoints of (dis)connection registered, but morewhen debating the substance of a project’sconclusions than as part of any grander epis-temological hand-wringing.

The therapist

In order to introduce our sixth figure (that of‘the therapist’), it is worth pausing to under-line – and to return to a theme of Chapter 5 –just how significant to contemporary humangeography is the practice of conversationbetween researchers and those people underresearch. More than just a metaphor allowingus to consider the hermeneutics of meaning,as above, much human geography has cometo rest on real conversations, real strips of dia-logue, in grounded field settings. DavidDemeritt and Sarah Dyer (2002: 229–30) coin


330

3151-Ch-10.qxd 3/2/04 2:46 PM Page 330

the phrase ‘dialogical geography’ to capturewhat is going on here, suggesting that such aterm ‘refers to qualitative research based ondialogue, or conversation, between theresearcher and other human informants’.Moreover, and as a valuable cue for our argu-ments, they elaborate that

Some forms of qualitative research involveliteral dialogues with other people. Theprevalence – indeed almost dominance – ofsuch qualitative research methods in humangeography today has given rise to a set ofdebates about the proper practice and thecredibility or truth status of research basedupon methods of literal dialogue. (2002: 229)

Something of these debates has already beenrehearsed above when discussing the conver-sationalist, but what we want to stress now isone very recent strand of inquiry where thegrounded conversation between researcherand researched has itself very much been fore-grounded. This is the growing adoption ofpsychoanalytic methods within human geog-raphy (Rose, 2000), as buttressed by a moredeep-seated turn to taking seriously theclaims of psychoanalysis that we will mentionshortly.

Building on contributions from JacquelineBurgess (Burgess et al., 1988; Burgess andPile, 1991), Steve Pile (1991) has introducedgeographers to what he regards as an ‘inter-pretative’ approach that goes beyond the stan-dard forms of ‘qualitative’ research to embracea still more intensely ‘intersubjective relation-ship’ between researcher and researched. Sucha relationship is seen as full of interpersonaldynamics, the recognition, enhancing andproblematization of which are central to howpsychoanalysis operates in the therapeuticencounter between therapist (or analyst) andclient (or analysand):

The patient works with the analyst becauseof the desire to relieve her or his symptoms[of mental distress], and because of the sup-port provided by the analyst; this unspokenand provisional agreement of work

together is known as the ‘therapeuticalliance’ … I suggest that there is an analo-gous, though different, ‘research alliance’,and we can use concepts developed in thecourse of therapy to excavate the multiple-layered interactions between researcherand researched. (Pile, 1991: 461)

Pile explores quite technical psychoana-lytic notions of ‘transference’ and ‘counter-transference’, as played out in an exercisepredicated on bringing to consciousness the‘repressed’ experiences of the patient to becritically worked through (and hopefullyresolved).This is an exercise where ‘the ther-apist’s unconscious and conscious reactions tothe patient’s transferences’ (Parr, 1998b: 344)cannot be overlooked as part of the overallprocess, the basis of which is the ‘therapeuticalliance’ that Pile wishes to take as a possiblemodel for interpersonal ‘research alliances’.Objections have been raised to Pile’s sugges-tions, with Linda McDowell (1992a) wonder-ing if more needs to be known aboutpower/social relations between therapist(researcher) and patient (research subject).Hester Parr (1998b: esp. 351) then worriesthat the therapeutic model commits geogra-phers to an approach that is ultimately toorigid, too hung up on control and insuffi-ciently open to ‘messy methodologies whichseek to “tune in” to research participants’ dif-ferent ways of telling’. A handful of humangeographers have since considered furtherwhat methodological prompts can be derivedfrom psychoanalysis, perhaps in terms ofDonald Winnicot’s therapeutic session asfacilitative ‘holding environment’ or even theuse of a Jungian ‘sand play’ for probing people’sdeepest interpretations of landscape.4

What we must now clarify, however, is thatwithin the broader ‘psychoanalytic turn’(Philo and Parr, 2003) affecting some cornersof human geography there is also a trajectoryof criticizing standard modes of understandingin the discipline. Such a critical stance arisesbecause the psychoanalytic emphasis on the

UNDERSTANDING

331

3151-Ch-10.qxd 3/2/04 2:46 PM Page 331

unconscious, positing that so much of ahuman being’s mental life or ‘libidinal econ-omy’ is occurring at a subterranean levelunavailable to conscious inspection, chal-lenges the conventional assumption thatpeople can know, reflect upon and articulatewhat is meaningful to them. Even more sothan for the humanistic geographers drawingupon phenomenological notions of thepreconscious (Relph, 1970; Seamon, 1979),psychoanalytic geographers propose that verysignificant domains of the human condition –so many things, events, actions and processeswithin the human world, from both everydayincidents such as shunning a person unlikeourselves on the street to mass projects of‘ethnic cleansing’ – are ultimately energized byunconscious drives about which we are dimlyif at all aware. It is argued that such drives stemfrom buried passions, desires and fears whosecontents remain ‘repressed’ from consciousnessbut which are, none the less, the foundation ofan individual and even collective ‘psychiceconomy’ impelling all manner of goings-oncentral to worldly human geographies.5

Such drives could arguably be termed‘meanings’, and it is possible to envisage psy-choanalytic procedures (even the use ofhypnotherapy) that might enable these sub-terranean strata of meaning to be exposed forcritical inspection. Without a doubt, though,we are now working with a very differentsense of what meaning entails from thatfavoured by the likes of the critic, the artisan,the ethnographer, the iconographer and eventhe conversationalist.The figure of the thera-pist (as someone who strives to retrievethreads of intelligibility from the debris of aperson’s unconscious) thus acquires a rolevery different from that of just another parti-cipant in a two-way conversation or dialogue.Whether geographers could or shouldattempt to deploy the skills of the trainedtherapist or counsellor is a vexed matter, but itis interesting that a small number are now

seeking to gain just such an expertise as aninput to their future research practice.

It also needs to be stated that in the litera-ture of psychoanalytic geography there isuncertainty about whether psychic materialscan ever be truly ‘knowable’, given the suspi-cion that any conscious articulation of themwill always amount to a mis-translation. Oneof Gillian Rose’s (1997a; 1997b) key claims –rooted partially in psychoanalytic theory – isthat much of the human world is ‘inoperable’in the sense that it precisely does not entailhuman subjects communicating preformedmeanings to one another through languagebut rather a flow of conducts, includingspeech acts, that in effect comprise a resistance‘to all the forms and all the violences of sub-jectivity’ (Nancy, 1991: 35, cited in Rose,1997b: 188). Much of the content of this flowis hence unknowable and uncertain, simplyunavailable to the conscious researcher, andsuch a claim undercuts much that is assumedabout (how to access) meanings by at least thefirst five of the figures described above. AsRose (1997b: 184) ponders, are there notsome ‘things beyond discourse, representation,interpretation, translation’; and she then asks‘what [are] the implications for my inter-pret[at]ive project, with its transcripts, codes,categories, records and papers?’ The answerappears to be that such a project hits big prob-lems, now being unable to interpret meaningsas had been hoped, and Rose is left supposingthat we cannot do much more than attendimaginatively to the ‘silences’ and ‘absences’revealing that there is often nothing beneath, nocoherent meanings, just ‘emptiness’. Perhapsthe truth about one’s unconscious is not itscapacity to generate meaning but, rather, itslocus as the ‘death’ of meaning where thingsfall apart, become no-things, rather thanbecoming assembled – parcelled up as recog-nizable frames of reference – in the guise sup-posed by all who cleave to the hope ofunderstanding human situations. Whatever


332

3151-Ch-10.qxd 3/2/04 2:46 PM Page 332

the unconscious impels people to do thuscannot be taken as to do with meaning, atleast not in anything like a conventional senseof what meanings entail, and so any equationof drives with meanings should be made onlywith the greatest of theoretical care.

The deconstructionist

Although they vary substantially in theirapproach to the interpretation of meaning,the figures we have discussed so far share acommon starting point: namely, that theinterpretation of meaning is just about possi-ble. They all work from the assumption thattexts, images, conversations and so on aresources of more or less stable and coherentmeanings and that these meanings are in prin-ciple accessible to researchers with the appro-priate skills and experience. In other words,they take the view that the role of theresearcher is either to recover meanings thatinhere in texts and other sources or, in thecase of the ‘conversationalist’, to constructmeaning in dialogue with the source. Ourfinal figure takes a different view of meaningand the possibility of understanding it. Thedeconstructionist starts from the assumptionthat texts do not contain stable and coherentmeanings. No matter how skilled theresearcher, the deconstructionist argues, it isnot possible to reconstruct the meaning of asource because the very idea of meaning onwhich such reconstructions are based isflawed. According to the deconstructionist,the meanings of any text are always unstable,partial and contradictory. Deconstructionistsare particularly scathing about the idea thatmeanings are put into texts by their authors.Under the slogan ‘death of the author’ (pro-posed originally by the semiotician, RolandBarthes), they argue that authors never con-trol their texts and that through the processof writing the text takes on an independentexistence beyond authorial intention.

Deconstruction (see also Box 7.8) is mostclosely associated with the work of the Frenchphilosopher Jacques Derrida. According toDerrida, the truth claims of any text areundermined by the unspoken (or perhaps weshould say unwritten) assumptions on whichthe explicit arguments of the text depend.Derrida argues that the Western philosophicaltradition, and by extension Western thoughtin general, is logocentric.That is, it assumes thatthere is some ultimate foundation for knowl-edge (for example, ‘logic’ or ‘rationality’) andthat this reflects the fundamentally orderedcharacter of the universe. In other words, weusually assume that behind the words, con-cepts and theories through which we under-stand the world an underlying principle(‘logic’, ‘rationality’, etc.) is present that guar-antees the validity of our accounts. Thisassumption is what Derrida calls a ‘metaphysicsof presence’. He uses the term ‘metaphysics’here because we can have no independentverification of the presence of the underlyingprinciple; we understand it only through thevery words, concepts and theories that it sup-posedly validates. Because we cannot accessthe underlying principle directly (it is, in theend, just an assumption), all logocentric sys-tems are without the secure external founda-tions they claim. What purports to be anultimate foundation is actually part of thesystem itself, or may in fact contradict it.Deconstruction examines this disjuncturebetween the claims of a text and the basis forthose claims. More specifically, Derrida arguesthat claims to truth typically rest on a series ofbinary oppositions. Examples of such binarypairs include inside/outside, self/other,male/female and rational/irrational. Moreover,these binary concepts are also hierarchical sothat the self is more important than the other,male dominates female, rationality is view pos-itively and irrationality negatively and so on.Derrida seeks to show is that it is never possi-ble to exclude all traces of the subordinate

UNDERSTANDING

333

3151-Ch-10.qxd 3/2/04 2:46 PM Page 333

element from its dominant pair so that anyargument based on a sharp distinctionbetween the binary elements can be shown tobe internally incoherent. This process ofuncovering the necessary instability of appar-ently logical and coherent texts is an exampleof deconstruction.

References to Derrida’s work and theprocess of deconstruction have becomewidespread in contemporary human geogra-phy, although there are still relatively fewexamples of geographers working with thetechniques of deconstruction in any sustainedway.Trevor Barnes provides one of the clear-est attempts to follow the spirit of Derrida’sarguments in a study of geographical thought(1994), though it is perhaps worth noting inpassing that valuing conventional ‘clarity’ inwriting is an example of the very logocen-trism that Derrida seeks to unsettle, as Barneshimself recognizes (1994: 1023). Barnes looksat the role of mathematics in geography andparticularly at the value placed on mathemat-ical procedures by proponents of the so-called quantitative revolution in geography inthe 1960s and early 1970s. Supporters ofquantification argued that mathematicswould provide a solid foundation for geogra-phers’ accounts of the world that was univer-sal, logical, objective, simplifying and precise.Through a deconstructive approach, Barnesshows that each of these claims about mathe-matics is untenable because each of themdepends on its binary opposite. For example,drawing on philosophers such as Wittgensteinand Russell, Barnes argues that attempts todemonstrate the universality of mathematics(the argument that mathematics is true in alltimes and all places) always founder becauseany justification for universality must dependon conventions that are socially and tempo-rally specific (1994: 1029–30).There are sim-ilar problems with the idea that mathematicsoffers greater precision than accounts writtenin ordinary language. This is because the

concept of ‘precision’ is not itself part of themathematical system but is actually an ordi-nary language interpretation of the relation-ship of mathematical models to the thingsthey are supposed to represent. One canonly claim that mathematics is more precisethan other languages by using those other,supposedly imprecise and thus fallible,languages to express the claim (Barnes,1994: 1033–4). In his conclusion Barnes re-emphasizes the value of the deconstructiveapproach:

I argued in this paper that mathematics is aprime example of logocentrism. For it isabove all an attempt to impose order on theworld. But … that order neither inheres inmathematics nor the world as such, but onlyin the local institutional authorities thatcontrol mathematical practices. Of course,there have been attempts to justify thosepractices as something more than practices:for example, the claims that mathematics isuniversal, or that statistics is a foolproofmethod for making inferences. But such jus-tifications always unravel … They unravelbecause at bottom they assert that there isa realm of certainty that lies beyond lan-guage. It is here that the arguments ofdeconstructionist are so valuable. For theyshow that there is nothing outside the text;our truths are only those that we write toourselves. (1994: 1037)

In addition to the analytic approach taken byBarnes, a number of other geographers havefollowed Derrida in a somewhat differentway, but seeking to introduce something ofhis experimental approach to the use of lan-guage and writing into geography. Althoughnot explicitly a deconstructionist, GunnarOlsson was an early geographical proponentof playing with the style and form of writtentexts to try to destabilize their dominantmeanings (1980). More recently the geogra-pher Marcus Doel, among others, has adopteda self-consciously non-traditional writingstyle as a way of subverting the foundationalknowledge claims of conventional geographi-cal texts (1994; 1999).


334

3151-Ch-10.qxd 3/2/04 2:46 PM Page 334

Conclusion: betweenunderstanding andexplanation

In this chapter we have sought to unpack thecomplex notions of ‘meaning’ and ‘under-standing’ and suggested that geographers haveapproached both of these in strikingly differ-ent ways. Sometimes these approaches arecomplementary. For example, the ‘artisan’ andthe ‘ethnographer’ are really close cousins, dif-fering in their use of particular techniquesperhaps but not fundamentally at odds overthe role of meaning in social life and the tasksof the researcher. In other cases there are whatappear to be irreconcilable differences. Mostdramatically the ‘deconstructionist’ doubts thevery existence of the authorial intentions thatsome ‘critics’ seek to reveal.

Whatever their differences, however, itseems to us that, taken together, our sevenfigures do speak to a more general issue inas-much as that they confirm Derek Gregory’sview that ‘there can be no clearly defined dis-tinction between understanding and explana-tion’ (1978a: 145). At the start of this chapterwe referred to the ideas of Dilthey and Weber.Gregory argues that there is an important dif-ference between their approaches: ‘whereasDilthey had regarded verstehen (understand-ing) and erklären (explanation) as two quiteseparate procedures, Weber always connectedthem closely together … the ideal type wassupposed to provide a means of movingbetween verstehen and erklären’ (1978a: 133).

There are thus important connections tobe drawn between the arguments of thischapter and those of Chapter 9. Explanationscannot afford to ignore meanings, particularlywhere the meanings with which social actionsare imbued constitute an important part ofthe motivation of the actors involved. In thissense, meanings can be part of the explanationfor the processes and patterns studied byhuman geographers.At the same time, peoplereflect on their circumstances and devise theirown explanations for their own activities andthose of others (which may or may not accordwith the explanations advanced by geogra-phers). In this way explanations can becomepart and parcel of the meaningful character ofsocial life.

Notes

1 For a detailed discussion of Dilthey’s ideasand their application in geography, see Rose(1981).

2 A development with many echoes in contem-porary human geography – e.g. Burgess andGold (1985), Aitken and Zonn (1994a) andLeyshon et al. (1998).

3 For more information on semiotics, go towww.aber.ac.uk/media/Documents/S4B/semi-otic.html. This website has now been pub-lished as a book – see Chandler (2001).

4 See Bingley (2003) and Kingsbury (2003); seealso Bondi (1999; 2003a; 2003b) for veryrecent ideas about the spaces of counselling.

5 Sibley (1995), Pile (1991; 1996), Wilton (1998),Nast (2000) and Callard (2003).

UNDERSTANDING

335

3151-Ch-10.qxd 3/2/04 2:46 PM Page 335

Introduction

The literal meaning of ‘geography’ is ‘earthwriting’ – taken from the Greek geo (‘theearth’) and graphien (‘to write’) – and writingis indeed central to the practice of humangeography. For while in principle geogra-phers could be involved in the production ofall sorts of texts – films, paintings, soundtracksand, of course, maps – it is writing that everysort of human geography and human geogra-pher has in common. Academic libraries arefull of books and journals containing thewritten output of various types of geographi-cal practice, and you will present your ownresearch results to your tutors largely in theform of written essays, projects and disserta-tions. It is not only the form of communica-tion that is common. Whether you arewriting a book, a journal article, a dissertationor an essay, many of the issues involved in theprocess of writing are similar. The ‘terror’ ofthe blank page and the ensuing trials andtribulations of authorship are faced by every-one sitting down to write: the authors of thisbook can all vouch for the fact that they arenot restricted to students.This chapter, how-ever, seeks to move beyond a consideration ofthe mechanics of writing to explore issues ofrepresentation. In other words, it concen-trates on writing as a critical part of those‘practices by which meanings are constitutedand communicated’ (Duncan, 2000: 703).

Thus the central motif running through thechapter is that writing is a form of represen-tation (or indeed re-presentation) whichhelps to create, rather than simply reflect, ourgeographical experiences. As part of thiscreative process, the author is inevitably inter-preting his or her data through the veryprocess of writing.

This in turn means that writing, and tex-tual construction more generally, are notprecursors (‘writing things down’) nor lateradd-ons (‘writing up’) to the practices ofdata construction and interpretation; theyare a critical part of those constructions andinterpretations.Writing and the constructionof academic texts, whether whole books, sin-gle chapters, articles, essays, dissertations orreports, are not something separate thatcome at the end of the research process aslater add-ons. Writing needs to be seen assomething integral to research – betterexpressed perhaps as ‘writing through’ – andindeed it is only in the process of writingthat our own ways of seeing and interpretingthe world are conveyed to others. All practi-tioners of human geography, whether under-graduate dissertation students or seniorprofessors, have to find productive ways ofwriting in order to tell others about thegeographies they have found. In this senseacademic writing concerns the constructionof a narrative in order to communicate with,engage, and, ultimately, convince your

Representing Human Geographies11

3151-Ch-11.qxd 3/3/2004 3:27 PM Page 336

particular audience. This audience could beother academics reading a journal paper; itcould be students reading a textbook; itcould be government policy-makers readinga report; or it could be a lecturer reading andmarking undergraduate essays. But becauselanguage is never simply mimetic, and can-not mimic or imitate the world it seeks toconvey, writing for any audience is alwaysabout the translation and interpretation ofthat world to others.

In this chapter we explore this role in termsof a move from issues of presentation (of writ-ing as a craft, a skill, a tool) to issues of repre-sentation – writing as about the production ofknowledge and ways of seeing/conceiving theworld. Thus we look at the work of writingand at ways of presenting research before wemove on to consider various moments of tex-tual representation. Although the chapter islargely concerned with writing, it also pre-sents a brief consideration of other ways ofrepresenting geographical practice, throughfilm and notation.We should say at the outsetthat some geographers are very suspiciousabout undue attention being given to ques-tions of writing and representation, fearingthey deflect attention away from more press-ing matters of substance and the politics ofresearch to questions of poetics. Nigel Thrift,for instance, cautioned against ‘over-wordyworlds’ well over a decade ago, in 1991. Weshare many of his concerns over the worryinghegemony of culture within social research,for instance, and the move towards philosoph-ical rather than empirical means of argumen-tation, and recognize the pitfalls of stressingpoetics above politics (Barnes and Gregory,1997) – not least because of the danger thatthis all becomes part of a cultural reposition-ing of geography as (second-rate) art orfiction, and geographers as nothing other thancultural intermediaries. Given this, we shouldperhaps stress that, in the argument presented

here, a concern with textual production is notseparate from questions of empirical sub-stance, matters concerning the relations of theresearcher to the researched or issues aroundthe point and politics of research; indeed, ourkey point is that it is heavily implicated in allthese.

The work of writing

This section looks at the practice of textualconstruction as work. It argues that all practi-tioners of human geography research, fromundergraduates sweating at their dissertationsto lecturers labouring over their latest book,ultimately have to find fruitful and rewardingways to ‘write through’ their data and ideas.While each individual will vary in how easyor difficult, how fun or painful he or she findswriting, all human geographers have to workat it. Box 11.1 builds on this idea of writing aswork by setting out Northedge’s ideas on thecraft of writing, where he develops the notionthat writing must be seen as taking placethrough a number of distinct, but interlinked,stages.

This section explores some of these stagesand offers some advice on how to organizethe craft of writing. In doing so it seeks toopen up for debate some of the feelingspeople have about writing they often keepprivate. Although it acknowledges that ‘writ-ing through’ one’s research is far from easy, itdoes attempt to combine this with groundsfor optimism and ways forward. It starts fromthe position that there is no neat and simpletransition, or straightforward translation, fromconstructed data and interpretations into finaltexts.We have already seen in earlier chaptershow the data that form the ‘results’ of researchare subject to many manoeuvrings, fromsifting and sorting onwards. At each stage theform these data take in any account of the

REPRESENTING HUMAN GEOGRAPHIES

337

3151-Ch-11.qxd 3/3/2004 3:27 PM Page 337

According to Northedge in his book The Good Study Guide (1994: 157–8, emphasis inoriginal), writing can be conceived of as a craft. He first makes a case for this view bycounterposing it to writing as inspiration:

How do you picture really experienced writers setting about the job of writing?Do you imagine them sitting down with a blank sheet of paper, or a blank wordprocessor screen and just spilling out words? Do you picture them being visitedby inspiration and immediately starting to pour out beautiful and compellingsentences, which are ready to be sent off for publication? It is almost never likethat.

He then makes the point that one reason why this never happens is because of thenature of the writing which we all do as geographers. As in most other academic disci-plines, geographers are largely concerned with ‘expository’ writing – writing whichseeks to analyse, understand, explain and argue – in contrast to ‘expressive’ writingwhich attempts to express feelings, and ‘narrative’ writing which tells stories. As weshall see later in the chapter, these three types of writing are not necessarily totally dis-tinct, and the human geographer may well draw on all three. But for now we can stickwith Northedge when he explains that the process of trying to develop a carefully con-structed argument, and maintaining it over a number of pages drawing on empiricalevidence and the ideas of others, is actually a very complex one. As he notes: ‘Puttingwell-formed sentences on to paper ready for sending to your tutor is only the last of aseries of stages in the process of putting together an essay. Before that a great deal ofthinking and preparatory work has to be undertaken’ (1994: 157–8).

Finally, he develops the notion of writing as a craft by drawing an analogy withfurniture-making – in this case making a table. Like writing, this does not simply ‘happen’:

First [the furniture-maker] has to conceive of a design,… then choose the wood,prepare it, measure it, mark it, cut it, shape it, make the joints and finally put ittogether. And even then it still has to be smoothed, waxed and polished. Writingessays may not be quite as elaborate … but it does have some of that quality –requiring you to work methodically through a whole series of closely linked activ-ities. If you simply sit down when you have finished reading the course texts andtry to write a whole essay in a single sweep you will get nowhere. The job is toobig. You have to break it down into stages. Then you can take it stage by stageand work your way to a finished product. (pp. 157–8, emphasis in original)

When we write elsewhere in the chapter of ‘writing as a craft’, it is in this sense thatwe do so.

Box 11.1: The craft of writing

research is actively constructed through theprocess of writing; and in turn, this processwill inevitably help to construct data andinterpretations rather than simply reportingon them. Research writing, and the represen-tation of geographical practice, is thus alwaysabout the work of writing through, and notjust writing up.

Why writing throughresearch can be hard

Two main reasons are usually advanced as towhy writing through research can be difficult.First, it involves organizing and rationalizingfrom vast masses of data and ideas. As alreadynoted, this process begins at the very earliest


338

3151-Ch-11.qxd 3/3/2004 3:27 PM Page 338

stages of sifting and sorting.This is intrinsicallydifficult for most people. It can also be quitescary. There can be worries about what onemight lose as one organizes the data, accom-panied by concerns about doing justice to theresearch you have done and to the data youand your research subjects have constructed.There will also be worries about how topackage the interesting, but still somewhatvague, ideas you have had about how to inter-pret those data, and about the sheer complex-ity of the phenomena and lives you’ve foundout about. Indeed, perhaps at the root of allthis there may well be worries about whetherany sort of order will ever emerge.As Becker(1986: 133, 134) puts it:

That is the deepest cause of the anxiety thatstrikes writers when they begin. What if wecannot, just cannot, make order out of thatchaos? I don’t know about other people,but beginning a new paper gives me anxi-ety’s classical physical symptoms: dizziness, asinking feeling in the pit of the stomach, achill, maybe even a cold sweat … The …world may be ordered, but not in any simpleway that dictates which topics should betaken up first. That’s why people stare atblank sheets of paper and rewrite first sen-tences a hundred times. They want thosemystical exercises to flush out the One RightWay of organizing all that stuff.

The second reason writing can be difficult isbecause it involves presenting one’s ideas, andindeed oneself, for scrutiny. Scrutiny by others –both real (actual readers you give materials to)and imagined (those whom, as you write, youimagine reading your text, often with theresult you immediately delete whatever youhave written as visions of their disappointedfrowns pass through your head: your supervi-sors, examiners, peers, role models). Scrutinytoo from your own critical and often unreal-istically demanding eye. Such scrutiny meansthat writing can be considered a risky process –it can become a very emotional form oflabour through which you are offering your-self for external judgement. Box 11.2 sets out

some thoughts from the sociologist PamelaRichards who developed her own notion ofwriting as a ‘risk’.

As Richards concludes, the process ofwriting may be difficult but it can also beimmensely rewarding and give you a greatsense of achievement. However, there are cer-tain strategies that can be followed in order tomaximize the pleasurable aspects of writingand minimize the costs of self-doubt. Manybooks are now available which help studentswith different aspects of the writing processand, together, they give a huge array ofadvice.1 As our emphasis in this chapter is onwriting as representation and not writing aspresentation, we will not seek to duplicate orsummarize these but will pick out four relativelysimple steps as a reminder that writing can beorganized in productive ways, before we moveon to look at some of the more detailedmechanics behind ‘good’ writing.

Organizing the workof writing

The first suggestion is to view writing as anongoing constructive process rather than as aone-off act of research presentation. Practi-cally, this means being prepared to write mul-tiple drafts (even beginning with what deSouza, 1988, calls a ‘spew draft’), starting writ-ing as early as possible (including in memosand letters), and not expecting early writingto meet the same criteria one would apply tolater drafts.As Becker (1986: 14, 17) puts it:

writing need not be a one-shot, all-or-nothing venture. It [can] have stages, eachwith its own criteria of excellence … Aninsistence on clarity and polish appropriateto a late version was entirely inappropriateto earlier ones meant to get the ideas onpaper … a mixed-up draft is no cause forshame … Knowing that you will write manymore drafts, you know that you need notworry about this one’s crudeness and lack ofcoherence. This one is for discovery, not forpresentation.


339

3151-Ch-11.qxd 3/3/2004 3:27 PM Page 339

Richards explains how she came to view the process of writing as a risk, after havingvivid dreams about people commenting on her work and being exposed as an acade-mic fraud. In trying to understand why she had these dreams of self-doubt she came tothe following conclusion:

For me, sitting down to write is risky because it means that I open myself up toscrutiny [from colleagues and myself]. To do that requires that I trust myself, andit also means that I have to trust my colleagues. By far the more critical of theseis the latter, because it is colleagues’ responses that make it possible for me totrust myself. So I have dreams of self-doubt and personal attacks by one of myclosest and most trusted friends … If you give someone a working draft to read …you’re asking them to decide whether you are smart or not … If there are noflashes of insight, no riveting ideas, what will the reader conclude? That you’restupid … Hence the fear of letting anyone see working drafts. I cannot face thepossibility of people thinking I’m stupid. (Richards, 1986: 113–15)

She then goes on to explain the classic Catch-22 situation that this gives rise to – shecan receive the affirmation she seeks from others only through the process of writing,but it is only this affirmation that frees her to write:

I trust myself (and can therefore risk writing down my ideas – things that I havemade up) primarily because others I trust have told me that I am OK. But no onecan tell me that until I actually do something, until I actually write somethingdown. So there I am, faced with a blank page, confronting the risk of discoveringthat I cannot do what I set out to do, and therefore am not the person I pretendto be. I haven’t yet written anything, so no one can help me affirm my commit-ment and underscore my sense of who I am. (1986: 117)

However, there is considerable light at the end of this particular tunnel. As Richardsgoes on to explain, this fear of risk does not paralyse her writing abilities completely.Instead, through the process of writing she comes to the conclusion that it is not an all-or-nothing proposition. What gets written is usually neither a literary gem nor purerubbish. Most writing is somewhere in between and, by recognizing this, Richards wasable to come to terms with writing as a process – and one which, like most things, ishelped considerably by practice:

In some ways, writing gets easier the more you do it, because the more you do it,the more you learn that it’s not really as risky as you fear. You have a history onwhich to draw for self confidence, you have a believable reputation among awider number of people whom you can call on the phone, and best of all, youhave demonstrated to yourself that taking the risk can be worth it. You took therisk, you produced something, and voila! Proof that you are who you claim to be.Though I must also admit that it’s not as easy as I’m making it sound. My writinghistory gives me some confidence, but I look at my past work with mixed emo-tions. It looks awkward and full of errors … This means that every time I sit downto write I find myself wondering whether I can really do this stuff at all. So writ-ing is still a risky activity. But what I seem to be learning as I spend more time writ-ing is that the risks are worth taking. Yes, I produce an appalling amount of crap,but most of the time I can tell it’s crap before anyone else gets a chance to lookat it. And occasionally I produce something that fits …, something that capturesexactly what I want to say. Usually, it’s just a sentence or two, but the number ofthose sentences grows if I just keep plugging away. This small hoard of good stuffalso helps me take risks … It reminds me there are two sides to risk. You can lose,but you can also win. (1986: 118–19)

Box 11.2: Writing as emotional risk


340

3151-Ch-11.qxd 3/3/2004 3:27 PM Page 340

Although the notion of seeing writing as amultistage process can be helpful, you shouldbe aware that experience of writing for onepurpose may not necessarily suit another.Thepreparation of writing many short essays, forinstance, can actually be unhelpful when itcomes to writing through one’s own researchin the form of a much longer dissertation orfieldwork report. Indeed Becker claims that:

The student’s situation rewards quick, com-petent preparation of short, passablepapers, not the skills of rewriting andredoing … Students find the skill of writingshort papers quickly less useful as theyadvance … eventually they have to writelonger papers, making more complicatedarguments based on more complicateddata. Few people can write such papers intheir head and get it right on the first try,though students may naively think that goodwriters routinely do … So students flounder,‘fear getting things wrong’, and don’t get itdone on time. (1986: 19)

Although Becker is writing from the stand-point of American sociology, the step up todoing a 10 or 12 000-word geography disser-tation can be similar. The moral here is toadapt your writing strategy to the task inhand. What suits the production of a 1500-word essay linked to a tutorial or seminar dis-cussion may not work for a dissertation tentimes as long based on extensive fieldwork.But in each case multiple drafts and revisionswill be far more beneficial than one-off burstsof activity as the deadline approaches.

Secondly, approaching writing as anongoing part of the practice of human geogra-phy also means recognizing it needs to beorganized: in the same way that any othercraftsperson organizes his or her work. Itshould be viewed as an extended processrather than as something done in short, last-minute bursts, and as something one purpose-fully arranges in one’s own diary, finding thetimes when and the places where one writesmost productively. It can always be arranged assomething one tries to make a pleasure,

through rewards and incentives (a few hoursthen a snack, short walk or some music).There are even tips on how to start and con-clude each writing session carefully (so beginwhen fresh and not jaded; finish so that youhave something easy to pick up on uponreturn).

Thirdly, this organization of the writingcraft is not just individual and private. Indeed,one potential problem researchers may havewith writing is the fact that by its very natureit tends to be organized and carried out inhighly individualized and isolating ways. Onemeans of overcoming such isolation com-monly used by professional writers is todevelop a circle of ‘reading friends’ who willlook through early drafts and offer construc-tive advice – to be given, and taken, in theright spirit.According to Becker, these friendswill always treat ‘as preliminary what is pre-liminary’ (1986: 21) and help the author sortout matters of style and presentation. It is sur-prising how often an external eye can pick uppassages of writing which are unfocused orimprecise and point to sections where theideas are unclear and muddled.This tends tobe because the author gets too close to whathe or she has written and finds it difficult tostand back and take a detached look. This isprecisely what a writing friend can offer.These relationships work best if they are reci-procal – form a group and take it in turns toread through each other’s work.

Fourthly, and finally, one needs to developstrategies for dealing with different stages ofthe writing process.Wolcott (1990) lists theseas ‘getting going’, ‘keeping going’, ‘linkingup’, ‘tightening up’ and ‘finishing up’, anddevotes a chapter to each stage. Getting goingcovers the sorting, organizing and filing ofmaterials, as well as deciding where, when andhow to begin. We have already covered thepreliminary stages of this in our discussion ofsifting and sorting in Chapter 7, but this alsoinvolves the initial processes of editing and


341

3151-Ch-11.qxd 3/3/2004 3:27 PM Page 341

selecting appropriate data. As Wolcott puts it:‘The critical task … is not to accumulate allthe data you can, but to “can” (i.e., get rid of )most of the data you accumulate’ (1990: 35).So even in these very preliminary stages of theprocess we encounter the notion of writingbeing heavily implicated in the constructionand interpretation of data rather than simplyreflecting them. Once you have made a start,the trick is to keep going. Keeping going willinvolve more than writing in its literal sense –putting pen to paper or fingers to keyboard –and encompasses reading, organizing, refiningand revising.Wolcott (1990: 51) even includesthe process of ‘just staring into space’ as part ofthis phase. The key point here is to keepgoing; as Wolcott says,‘you have to have some-thing written before you can begin to improveit’ (1990: 55, emphasis in original).Try not toget stuck, either because of material you donot have or cannot understand. Make a noteto yourself of what needs inserting or revisingand go on to the next section. Once you havethat crucial first draft, even with some spacesand unfinished sections, it can become abetter second draft and a good third draft.

Linking up reminds us that all writing isembedded in broader academic and socialcontexts. It is done for particular reasons, tomeet specific goals. These may necessitatelinking your work to that of others – whetherin terms of literature reviews or broader con-siderations of concepts, theories and methods.Tightening up takes place once these linkshave been drawn and everything is in place. Itrefers to another round of reviewing and edit-ing but takes place once the material is moreor less complete.This is the final stage in theconstruction of data, as the entire script isevaluated and appraised as a complete entity:

• Does it hang together?• Are its points well made?• Are the arguments illustrated with refer-

ence to other work or your own research?

• Does the empirical material relate to themore abstract sections?

• Is any material superfluous to the argument?• Is each section introduced appropriately?• Can the reader follow where the argu-

ment is going?

Once all these issues have been attended to, thepiece of writing needs to be ‘finished’ andmade fit for its purpose. Issues such as the title,references, word length, headings and subhead-ings, footnotes and endnotes, figures andappendices all have to be addressed. Once thisstage is complete, you are ready for ‘getting itout the door’ (Becker, 1986: 121) where thetension between making it better and getting itdone is finally resolved in favour of the latter.Usually for students this resolution is forced bydeadlines, but if these do not come into play asense of the productive and rewarding possibil-ities of participating in the collective intellec-tual endeavour that is the practice of humangeography should help you to let go.

Presenting research

In this section we will keep the same instru-mental, craft quality of writing to the fore butwill do so in terms of accounts that advise notonly on how to write but also how to writewell. It deals with the presentation of researchdata, as part of its writing through, in threeparts. In the first we will look at the ways inwhich some academics have sought to offerguidance on the textual presentation of research,and at how this guidance makes public what areusually more private commentaries fromexaminers, journal editors and referees.Thereare some recurrent themes in this which placean emphasis on clarity and direct expression(including the avoidance of passive sentenceconstructions) and highlight how to engageaudiences through the use of particular acad-emic conventions and the adoption of a certain


342

3151-Ch-11.qxd 3/3/2004 3:27 PM Page 342

‘author-itative’ persona. The second sectionexamines some of the positive aspects thatstem from these kinds of prescriptions,emphasizing in particular the way they makeus think about the process of writing as a skillthat has to be, and indeed can be, worked atand developed.They also help us to focus onissues of audience engagement and remind usthat ultimately writing should be viewed asrelational – its purpose is to forge a relation-ship with the reader, however solitary aprocess it feels. Lastly, however, we go on topoint out that, while there is much of value inthese kinds of calls for ‘good writing’, theyalso tend to oversimplify and close downmany of the issues they seek to confront.These issues will be picked up more fully aswe move on to discuss representation andrhetoric later in this chapter.

Good writing: ‘overcomingthe academic pose’

There are now many books and guides avail-able which set out tips and prescriptions for‘good writing’ and most frame these tips interms of clarity and brevity.Again, because ouremphasis is on representation rather thanpresentation, we will simply outline the mainpoints of these prescriptions. De Souza sum-marizes the advice as follows: ‘Change passiveto active constructions. Take long sentencesapart. Eradicate redundancies. Eliminatepompous phrases … Get rid of qualifications …clarity is a virtue’ (1988: 2–3).Time and timeagain one comes across the same injunctions toavoid unnecessary words or phrases, to elimi-nate the passive voice (especially any form ofthe verb ‘to be’), and to make a virtue out ofshort and simple sentences whenever you can.2

One reason why most books on academicwriting emphasize these points about brevityand clarity is that there tends to be a misplacedfeeling that ‘scholarly’ writing is of necessitycomplex and convoluted. Indeed for some

people it seems that the more difficult a passageis to follow, the more intelligent it must be.Thisfeeling is utterly misplaced and confuses thestyle of an argument with its substance.The complexity of the latter is not indicated bythe complexity of the former – indeed,quite theopposite.As Hanson (1988: 6) puts it:

Also necessary is the author’s ability to com-municate that argument in lucid, jargon-free prose. Ideas that are heavily drapedwith jargon are likely to set an editor mus-ing about the moral of ‘The Emperor’s NewClothes’. Let us not convey to our students –however unintentionally – the idea thatobscure, convoluted or jargon-laden writingis somehow ‘scholarly’.

Many academics, however, seem unwilling tolet go of this idea.According to de Souza:

Articles published in scholarly geographicaljournals should clearly communicate ideasand results to those who read those jour-nals. Yet many of us ignore the rules thatgovern good writing and turn out stiltedand overblown prose. We use langauge thatobscures what we want to say, drowningmeaning in murkiness. (1988: 1)

According to Mills, this use of excessivejargon and obscure language actually has verylittle to do with the complexity of thesubject-matter being written about. Insteadhe links the reason to the psychological stateof the writer and to academic insecurity:

lack of ready intelligibility, I believe, usuallyhas … nothing at all to do with the profun-dity of thought. It has to do almost entirelywith certain confusions of the academicwriter about his [sic] own status … Desire forstatus is one reason why academic men [sic]slip so easily into unintelligibility … To over-come the academic prose you have to over-come the academic pose. (1959: 239–40)

For the student, overcoming the academicpose with good prose can help to raise marksconsiderably.As Steinberg says:

better prose produces better examinationresults. When examiners say vaguely that so


343

3151-Ch-11.qxd 3/3/2004 3:27 PM Page 343

and so wrote an ‘economical’ answer oruse adjectives like ‘neat’, ‘precise’ or ‘to thepoint’, they have, without always noticingit, reacted to vigorous, energetic writing. Ahealthy verb cannot transform a ploddingwaffler into Maitland but it can make thewaffle sound ‘2.lish’ and not ‘2.2ish’…Strong verbs give the writer a sense ofpower. (1985: 90)

Identifying the positivesin these kinds ofprescriptions

The moral in these prescriptions is to avoidthe ‘academic pose’ through active, clear andprecise writing. There are four main benefitsto this kind of advice. First, it does force youto think about writing as a craft skill, as some-thing that you have consciously to developabilities in rather than something you simplypick up through an imitation of what seemsscholarly.Thus these prescriptions do remindus that writing does not simply spring on tothe page ready formed through mental inspi-ration – it is something that, quite literally, hasto be worked at and worked on. In this senseit can be understood as a kind of carving orsculpture where the rough edges are constantlyhoned and refined before we reach the fin-ished product. These kinds of prescriptionstell you how to undertake this refinement.Secondly, this advice reminds you that theimpact of your style of writing is neverneutral – it is something that has a huge influ-ence on what any reader makes of what youare saying.Thus the meaning of what you aresaying is intimately connected to how you aresaying it. As we noted at the beginning ofthe chapter, this is because writing bothconstructs and communicates meaning ratherthan simply reflecting or presenting an exter-nal reality.

Thirdly, most of these texts on how towrite well do highlight the particular problemwhich exists in a temptation to use certainsorts of academic, exclusive vocabularies – both

as a means to make oneself sound scholarlyand as a way of associating oneself with thosewhose intelligence and scholarliness oneadmires. Consider, for instance, Box 11.3,which cites part of a letter to Howard Beckerfrom one of his students talking through howshe related to academic language in this way.As we shall go on to examine later in thechapter, the final point that Rosanna makesabout scholarly writing and being a scholar isnot actually the case – as she herself realized.Often, the very best scholars are those whoare able to explain complex ideas in the mostaccessible writing style.

This brings us on to the fourth positiveaspect of these prescriptions. They centreissues about audience engagement and somake sure that writing is seen as a practicewhich is all about building a relationship toreaders rather than a solitary, solipsisticactivity. This helps to remind us that thepoint of ‘writing through’ research is toengage readers, not to convince yourself ofyour own worth. The audiences that readthe results of geographical research are wideand varied: students, other academics,policy-makers, specialist researchers or eventhe public at large.What all these audienceshave in common, though, is that they are ata distance and removed from the writer. Inthis sense, writing is a ‘very special form ofconversation’ (Northedge, 1994: 146). Theparticular concerns this raises are set out inBox 11.4.

Problems with thesesorts of prescriptions

However, these kinds of prescriptions havetwo sorts of problems. First, they oversimplifythe criteria upon which the quality of writingcan and should be judged. For instance, theydefine clarity in very instrumental ways –about the conveyance of a clear and strongmessage or argument – and ignore other roles


344

3151-Ch-11.qxd 3/3/2004 3:27 PM Page 344

Box 11.3: On the temptation of academic jargon

In his book, Writing for Social Scientists, Howard Becker (1986: 26–8) recounts how heonce spent a considerable time editing and shortening a chapter written by one of hisPhD students – Rosanna Hertz. Becker polished the grammar, shortened overlengthsentences, removed superfluous words and cut repetition. Instead of being pleasedwith his efforts, Rosanna was a bit put out. Although she accepted that his version wastighter and clearer, she maintained that her original version was ‘classier’ (1986: 28).

Becker, in turn, was somewhat taken aback by this and asked her to explain, in writ-ing, what she meant by ‘classier’. This is what she wrote by way of explanation:

Somewhere along the line, probably in college, I picked up on the fact that articu-late people used big words, which impressed me. I remember taking two classesfrom a philosophy professor simply because I figured he must be really smart sinceI didn’t know the meanings of the words he used in class. My notes from theseclasses are almost non-existent. I spent class time writing down the words he usedthat I didn’t know, going home and looking them up. He sounded so smart to mesimply because I didn’t understand him … The way someone writes – the moredifficult the writing style – the more intellectual they sound. (1986: 28)

Becker goes on to explain that these types of feelings partly arise from the very hier-archical nature of academic life, where professors and lecturers are seen just to ‘knowmore’ and are viewed as figures of intellectual authority – who should be imitatedwhether or not they make sense. As Rosanna explains:

When I read something and I don’t know immediately what it means, I alwaysgive the author the benefit of the doubt. I assume this is a smart person and theproblem with my not understanding the ideas is that I’m not as smart. I don’tassume … that the author is not clear because of their own confusion about whatthey have to say. I always assume that it is my inability to understand or that thereis something more going on than I’m capable of understanding. (1986: 29)

Rosanna is pointing to the classic feeling of inferiority which many students (andindeed some lecturers) feel when first confronted with academic writing which is com-plex and even unintelligible. She goes on in her letter to highlight the processes whichhelp to maintain this situation:

academic elitism is part of every graduate student’s socialization. I mean that aca-demic writing is not English, but written in a shorthand that only members of theprofession can decipher … Ideas are supposed to be written in such a fashion thatthey are difficult for untrained people to understand. This is scholarly writing. Andif you want to be a scholar you need to reproduce this way of writing. (1986: 30)

that (differently clear) writing can play – todescribe, evoke, give a feel for, to expressambiguities and a lack of certainty.This tendsto presume particular sorts of audiences(a general geography readership, for example,or policy-makers) and may not be most effectivewith some others (for example, those whoselives are being written about).

Miller (1998), for instance, describes howone of her research subjects, having read herresearch report, comes back at her on the lackof humanity, feeling and relationship in thetext. The forceful, assertive, active and vigor-ous language used by Miller in the reportmade her feel that this was the author’s textand her own involvement had only been as a


345

3151-Ch-11.qxd 3/3/2004 3:27 PM Page 345

Box 11.4: Writing as a conversation: speaking to your reader

Writing is a very particular kind of conversation in the sense that you cannot see theperson you are talking to, and he or she never joins in the conversation. Neverthelessyou are charged with the responsibility of convincing him or her of your argument viathe written word. To make matters worse, you know he or she will be ‘listening’ hard,harder sometimes than in a real conversation, since he or she will usually be readingevery word, in contrast to a conversation where the listener may get distracted orwords are drowned out by background noise. Northedge (1994: 146–7, 166–7) sets outsome pointers for taking part in such a conversation:

• You have to develop a sense of your audience because the style used will vary fromaudience to audience. When you read research reports written for policy makers,they will be in a different ‘voice’ from those written for an academic argument.

• The standard formula for students is to ‘write for the intelligent person in thestreet’: assume that your reader has not read the books and articles you have usedbut that he or she is interested in the topic and can pick up your arguments quickly,provided you spell them out clearly.

• You have to take the reader carefully through all the points of the conversation,illustrating any points made, constructing a logical route through the points andmaking sure they follow on from one another.

• The talking you do in this conversation takes a lot of effort. This is why the contentshould have been prepared and sorted out in advance. If you try to think up the con-tent at the same time as addressing yourself to your reader, you will usually do oneor the other badly.

• Since writing is a very formal activity compared to speaking, it sometimes takes a lotof practice to find your ‘voice’. Don’t be worried if you do not initially convey whatyou intended. This is what drafts are for.

resource: that she was just a life to be slottedinto the researcher’s argument (here we linkinto the question of research ethics discussedin the next chapter).

One might also challenge the (once and forall) equation of good writing and clarity withaccessibility. In certain contexts, for certainpurposes, there is nothing wrong with work-ing through a specialized vocabulary (forexample, of particular theoretical schools) ifone’s aim is to engage with other initiates orinitiate others into that conceptual frame-work. Technical language is also necessary tothink through aspects of human geographicalrealities that are not articulated in everydaylife and language. It can also be used as adevice to think otherwise, beyond currentorthodoxies. Moreover, the reverse also holds –maths is a clear language but it isn’t highly

accessible; and we don’t always expect it to beso (even if sometimes we do expect mathe-maticians to be able to explain something oftheir work in lay terms).The same is true ofmedicine. Sometimes writing in complex,technical and sophisticated language is thebest option.The problem with prescriptions isthat they can be too prescriptive.

Secondly, these kinds of prescriptions for‘good writing’ not only oversimplify evaluativecriteria but they also fail to follow through onthe implications of many of the importantissues they raise. They close them down intodisciplining templates and standard voicesrather than using them to open up the possibil-ities (and responsibilities) of writing as textualproduction. They also don’t situate the issuesthey raise in a wider consideration of the char-acter of academic knowledge, or do so only


346

3151-Ch-11.qxd 3/3/2004 3:27 PM Page 346

implictly. For example, the question of authorialpersona and voice is reduced into a binaryopposition between imitative, jargon-riddenpretentiousness and plain speaking, confident‘neutrality’. This ignores broader issues aboutthe social evaluations of certain kinds of voices.Rose (1993b), for instance, describes how agreat deal of geographical writing conveys amale perspective on both nature and culture;indeed, she argues that the separation betweenthe two is artificial, constructed through ‘man-made language [which] divides what should beunited’ (1993b: 71). As a consequence, thesubjectivity expressed through geographicalthought and writing is one which is ‘dependenton separation and domination, especially on theseparation between masculine Culture andfeminine Nature’ (1993b: 71).

Moreover, the languages that such prescrip-tions applaud are bound up with certain fram-ings of the world (for example, a separation ofthe known and knower, the latter untroubledand undisturbed by the former). Rose againpicks up on this theme in her discussion of theway geographers have traditionally looked atlandscape. As she notes: ‘the pleasure of themasculine gaze at beautiful Nature is temperedby geography’s scientism … The gaze of thescientist has been described … as part of mas-culinist rationality, and to admit an emotionalresponse to Nature would destroy theanonymity on which that kind of scientificobjectivity depends’ (1993b: 88). Hence theinjunction to clear, simple and precise writingcan mask, and indeed stem from, a false sepa-ration and artificial distanciation between theobserver and the observed. If one takes theview, presented in Chapter 1 and developedthroughout Part I, of research as a social activ-ity, where data are constructed through arelationship between the researcher and theresearched, it becomes difficult, if not impossi-ble, to maintain this separation.

This in turn means that concerns withgood writing style can take us only so far in

thinking about the nature and significance oftextual production as a human geographicalresearch practice. As Richardson (1988: 202)puts it, the ‘solution to the writing problem isnot the extermination of jargon, redundan-cies, passive voice, circumlocution and (alas)multisyllabic conceptualization and referentialindicators’. Instead we have to involve our-selves with issues of representation and rhetoricand consider the numerous ways in whichwriting as a process extends far beyond thewords and sentences on the page.

Representationand rhetoric

This section begins by setting out the notionsof representation and rhetoric as ways tothink about writing as a process that is impli-cated in the very creation of ‘interpretations’,‘data’ and ‘reality’. It then develops the argu-ment that textual construction and writingshould be viewed not so much as regrettablynecessary processes of communication – andthus something to be undertaken in ways thatdraw as little attention to themselves as pos-sible – but as stylistic processes of persuasionthat are fundamental to research practice.These points are then illustrated across twocontrasting fields that seem to sit very differ-ently in relation to textual construction andits importance.The first is regional and land-scape description, an area of geography wherecreative writing has been explicitly encour-aged in order to convey to the reader a senseof how a particular landscape looks or evenfeels.The second draws on examples of sup-posedly more ‘objective’ types of representa-tion in order to establish the point that evenscientific objectivity does not lie outsiderhetoric but, rather, is in part constructedthrough it. Thus, despite the differencesbetween the textual styles and strategiesadopted across these two fields, a number of


347

3151-Ch-11.qxd 3/3/2004 3:27 PM Page 347

common issues concerning the nature ofwriting materialize across both. These issueswill then be explored in more detail later inthis chapter when we examine a number ofspecific facets, or moments, of textual con-struction which all human geographers haveto negotiate.

The device of rhetoric

For much of its history, human geography hasrather unproblematically accepted the con-ventional view of language which sees wordsas mirrors of the worlds they represent. ForEagleton, the centrepiece of such a view isthat words are ‘felt to link up with theirthoughts or objects in essentially right andincontrovertible ways’ (cited in Barnes, 2000a:588).Thus the trick of writing, of represent-ing the world to others, becomes reduced toa mechanical one of lining up the right wordsin the right order (Barnes, 2000a: 588).Recently, however, this view has been chal-lenged to emphasize that the relationshipbetween words and the worlds they purportto represent is less straightforward and morecomplex.As Richardson puts it,‘writing is notsimply a true representation of an objectivereality, out there, waiting to be seen. Instead,through literary and rhetorical structures,writing creates a particular view of reality’(1990: 9).This sensibility towards the inherentcomplexity of representation foregrounds thenotion of poetics within geography. By thiswe do not mean the appreciation or practiceof poetry. Instead we are referring to poeticsas a ‘critical practice that involves taking intoaccount the force, exactness and power ofwords’ (Barnes and Gregory, 1997: 4). In otherwords poetics is about ‘treating words withrespect, recognizing their power, passion andpotential’ (Barnes, 2000a: 588). Said (1977;1993), for instance, has shown how the powerof words was used to construct very particularideas of ‘orientalism’.This notion in turn was

used to structure representations of Asia (andAsians) into a worldview that was seen from,and compatible with, projects of Westernimperialism in the nineteenth and earlytwentieth centuries, where the Western cultu-ral self was increasingly defined in contrast tonon-Western ‘Others’.This definition was thenimportant in helping to underpin and validatethe Western exploitation of these ‘other’ lands.

Once we accept that words are not neutralmirrors of the world they represent, we areled to examine just how they do gain their‘power, passion and potential’. Richardson hasnoted how ‘all language has grammatical, nar-rative, and rhetorical structures that constructthe subjects and objects of our research,bestow meaning, and create value’ (1990: 12).Hence, part of the ability of words to bepowerful and ‘bestow meaning’ comes fromthe ‘rhetorical’ devices used by their writer toestablish authority and persuade readers.Rhetoric is defined here as the art of argu-ment, and not used in its more pejorativeeveryday sense of pretence and display. As partof such an ‘art’, metaphors, imagery, citations,appeals to authority, anecdotes and even jokesare all used as ‘instrumental devices’ in the‘persuasive discourse of science’ (Richardson,1990: 15). In other words all texts seek to ‘per-suade’ readers of the truth of their claims, andmost will make use of rhetorical devices to dothis. What is interesting about this is thatdiferent elements of science, different disci-plines and subdisciplines all have their own setof literary devices which help this process ofpersuasion. We will explore some of thedevices used in geography later in this chapter.For now we simply want to concentrate onrhetoric as a fundamental, and unavoidable,part of research practice: as a device which isconstructive in the making of data and inter-pretations, as well as in the making of authorsand readers.

Making these broad claims about theimportance of rhetoric is not about privileging


348

3151-Ch-11.qxd 3/3/2004 3:27 PM Page 348

style (how something is written or represented)over substance (what is being said) but recog-nizing that style is a part of that very substance:

The distinction between style and substancehas burrowed like a worm deep into ourculture … yet it has few merits. It is all styleand no substance. Consider. What is the dis-tinction of style and substance in ice-skatingor still-life painting or economic analysis?…By style we mean properly the details ofsubstance … Style is not a frosting added toa substantial cake … The substance of acake is not the list of basic ingredients. It isthe style in which they are combined.(McCloskey, 1988: 286)

If one accepts that the style and the substanceof an argument are bound up together, andthat the device of rhetoric plays a key role inthe binding, one can then think about theproductive uses of a concern with rhetoric inboth the writing and reading of human geo-graphical research.These seem threefold. First,an appreciation of rhetoric allows you toexplore how different understandings of ‘whatresearch is for’ are put into practice. Secondly,they help you gain a critical sense of why orwhy not something ‘grabs you’ and holds yourinterest – of why a particular piece of writingmight appeal to you more than another.Thirdly, this concern with rhetoric can beused in analysing how the stylistic substanceof research texts is not only a matter of thefree choices of authors but also stems fromthe social contexts and institutional structuresof academic research practice, which privi-lege certain genres over others. We willexplore these issues in more depth now,beginning by looking at the way geographershave tackled the ‘art’ of regional and land-scape description.

Regional and landscapedescription

We discussed in Chapter 1 geography’s longengagement with regional and landscape

description. We return to it here because itis a field of geographical research withinwhich issues of representation and rhetorichave been debated explicitly for some time.Indeed, it is an area of the discipline where ithas long been acknowledged that geogra-phers should write ‘creatively’ in order torepresent their subject-matter to the reader.For many geographers, the representation ofregions and landscapes placed geographyfirmly within the arts and humanities, andthey sought to introduce large elements ofcreativity into their own work. One of themost influential of all geographers, H.C.Darby, for instance, drew a firm distinctionbetween geography as a science and geogra-phy as an art:

Geography is a science in the sense thatwhat facts we perceive must be examined,and perhaps measured, with care and accu-racy. It is an art in that any presentation (letalone perception) of those facts must beselective and so involve choice, and tasteand judgment … It would be interesting toorganise … a competition for geographers,and to invite an account of the regionalgeography of south Lancashire in the man-ner of say, Estyn Evans or Dudley Stamp orS.W. Wooldridge. I venture to say that thevarious accounts would not be identical, anymore than a portrait of a man by GrahamSutherland would be identical with one ofthe same man by Henry Lamb. Each picturewould have its own validity. (1962: 6)

Here Darby is applauding creativity throughdescription and is well aware that differentauthors will inevitably produce differentaccounts of the same region – but, as he says,‘each picture would have its own validity’.Hart (1982:2) drew a similar distiction betweengeography as a science and geography as anart, and argued that ‘the highest form of thegeographer’s art is producing good regionalgeography’ – which he defined as ‘evocativedescriptions that facilitate an understand-ing and an appreciation of places, areas andregions’.


349

3151-Ch-11.qxd 3/3/2004 3:27 PM Page 349

Darby, though, also recognizes that suchevocation and creativity are not particularstrengths of geographical writing. As he putsit: ‘good description of landscape or town-scape is not an outstanding feature of thewriting of professional geographers … Welook in vain for – to use an old-fashionedword – a “likeness” of, say, chalk downlands orclay vales or mountain uplands’ (1962: 1).Another famous regional geographer,E.W. Gilbert, also drew a contrast between theability of geographers to portray the ‘person-ality’ of a region and the skills of novelists. Hewrote that

English regional novelists display many mer-its that geographers can recognise and envy …The regional novelists have been able toproduce a synthesis, ‘a living picture of theunity of place and people’, which so ofteneludes geographical writing. The geogra-pher often speaks of the ‘personality’ of theregion and this is exactly what somenovelists have brought out so strongly.(1972: 124–5)

By making explicit this contrast Gilbert ishighlighting a particular geographical problemwith how we might represent landscapes andregions textually. In many ways novelists havemore licence to weave images of people andplaces together across the length of a piece offiction. For geographers the difficulty remainsof translating the geographies that we per-ceive into words. This difficulty partly arisesbecause ‘what one sees when one looks atgeographies is stubbornly simultaneous’ (Soja,1989: 2) – picture an urban street scene ofpeople driving to work, young children walk-ing to school, local residents going shopping,old men chatting, all set against a backdrop ofvaried architectures and styles. Even the mostphysical of landscapes will have several, if notmany, different elements: topography, vegeta-tion, aspect, climate. According to Soja, the‘problem’ with trying textually to representsuch simultaneity is that ‘what we write downis successive, because language is successive’

(1989: 247). In other words, the sequential,linear flow of textual narrative is not bestsuited to describing the complexity of anygiven space at any given moment in time.Ultimately Soja is left to conclude that ‘thetask of comprehensive, holistic regionaldescription may therefore be impossible’(1989: 247).

However, he has sought his own way out ofthis dilemma by using a particular rhetoricaldevice which runs through his powerful writ-ings on Los Angeles. Soja aims to describe thatcity in a ‘free-wheeling’ style (1989: 2), whichmakes its argument through the compositionof that description. In this case Soja aims to‘creatively juxtapose … assertions and inser-tions of the spatial’ (1989: 2). It is through theforce of this juxtaposition that the descriptioncomes alive.Box 11.5 illustrates the manner inwhich he ‘asserts and inserts’ the spatial.

Soja, then, uses the rhetorical devices ofinsertion and juxtaposition to try to capturethe multiple meanings of Los Angeles withinhis ‘fulsome geographical text’. But this, ofcourse, is not the only route to creativeregional description. Figures 11.1 and 11.2show Quoniam’s annotated paintings ofArizona. The rhetorical device used here isone which is centred on the shape andappearance of the text rather than on itsstructure and content as with Soja. Thefigures represent Quoniam’s notes from a tripthrough Arizona. As Cosgrove and Domoshpoint out:

The notes are in fact illegible, except for theoccasional stimulus word like ‘Coronado’, sothat the text for the most part is meaning-less as language. Its meaning derives fromits formal qualities as blocks of inscribedsigns which we recognize as written notesand thus assume to have recorded andproduced meaning. (1993: 32)

In other words, we assume that these noteshold geographical knowledge, even if thenature of such knowledge is not immediately


350

3151-Ch-11.qxd 3/3/2004 3:27 PM Page 350

All the following quotations come from Chapter 9 of Soja’s (1989) PostmodernGeographies, where the reader is taken on a textual tour of Los Angeles.

What is this place? Even knowing where to focus, to find a starting point, is noteasy … (p. 222). What follows … is a succession of fragmentary glimpses, a freedassociation of reflective and interpretive field notes (p. 223). However … LosAngeles … must be reduced to a more familiar and localized geometry to be seen.Appropriately enough, just such a reductionist mapping has popularly presenteditself. It is defined by an embracing circle drawn sixty miles out from a central pointlocated in the downtown core of the City of Los Angeles (p. 224). An imaginativecruise directly above the contemporary circumference of the Sixty-Mile Circle can beunusually revealing … The circle cuts the south coast … near one of the key check-points regularly set up to intercept the northward flow of undocumented migrants …The first rampart to watch, however, is Camp Pendleton Marine Corps base, thelargest military base in California in terms of personnel … After cruising over themoors of Camp Pendleton … we can land directly in Rampart no 2, March Air ForceBase (p. 225). What in the world lies behind this Herculean Wall? What appears toneed such formidable protection? In essence we return to the same question withwhich we began: What is this place? (p. 229). The Sixty-Mile Circle is ringed with aseries of … outer circles at varying stages of development, each a laboratory forexploring the contemporaneity of capitalist urbanization. At least two are combinedin Orange County, seamlessly webbed together into the largest and probably fastestgrowing outer city complex in the country (world?) … The Orange County complexhas also been the focus for detailed research into the high technology industrialagglomerations that have been recentralizing the urban fabric of the Los Angelesregion (p. 231). These new territorial complexes seem to be turning the city inside-out, recentering the urban to transform the metropolitan periphery into the coreregion of advanced industrial production (p. 233). The downtown core of the City ofLos Angeles … is the agglomerative and symbolic nucleus of the Sixty-Mile Circle(pp. 235–6). Looking down and out from City Hall the site is especially impressive …Immediately below and around is the largest concentration of government officesand bureacracy in the country outside the federal capital district … Included withinthis carceral wedge are the largest women’s prison in the country and the seventhlargest men’s (p. 236). Looking westward now, toward the Pacific and the smog-huedsunsets which brilliantly paint the nightfalls of Los Angeles, is first the CriminalCourts Building, then the Hall of Records and Law Library … Along the northernflank is the Hall of Justice, the US Federal Courthouse and the Federal Building, com-pleting the ring of local, city, state and federal government authority which com-prises this potent civic center (p. 237). From the center to the periphery … theSixty-Mile Circle today encloses a shattered metro-sea of fragmented yet homoge-nized communities, cultures and economies, confusingly arranged into a contin-gently ordered spatial division of labor and power (p. 244). I have been looking atLos Angeles from many different points of view and each way of seeing assists insorting out the interjacent medley of the subject landscape. The perspectivesexplored are purposeful, eclectic, fragmentary, incomplete and frequently contra-dictory, but so too is Los Angeles and, indeed, the historical geography of everyurban landscape. Totalizing visions … can never capture all the meanings and signi-fications of the urban when the landscape is critically read and envisioned as a ful-some geographical text (p. 247).

Box 11.5: Taking Los Angeles apart: the art ofdescription through juxtaposition


351

3151-Ch-11.qxd 3/3/2004 3:27 PM Page 351

recognizable. The form of the notes is alsoimportant.Text is not always aligned horizon-tally but is sometimes inserted at anangle, resembling more the usual method of

representing geological strata. Cosgrove andDomosh conclude that the ‘structure of writ-ing thus becomes the structure of language, itsmeanings layered like those of the earth itself ’


352

Figure 11.1 Quoniam’s annotated painting of Arizona: Tucson

Source: Cosgrove and Domosh (1993: 33)

3151-Ch-11.qxd 3/3/2004 3:27 PM Page 352

(1993: 32). Here also the rhetorical deviceextends beyond words to encompass drawings,maps and sketches. However, in Quoniam’s artit is these pictures which becomes the text.Ashe says,‘painting becomes the text of my ideasand the text becomes the descriptive picture

of the landscape’ (cited in Cosgrove andDomosh, 1993: 35). Cosgrove and Domoshconclude that ‘by revealing the interchange-ability of word and picture, he forces us torecognize words as simply another surface. Butthat surface, like a painting or an illustrated


353

Figure 11.2 Quoniam’s annotated painting of Arizona: Coronado

Source: Cosgrove and Domosh (1993: 34)

3151-Ch-11.qxd 3/3/2004 3:27 PM Page 353

geological section, is not superficial: it containsand creates meanings’ (1993: 35).

This section has sought to convey that themeanings created by our text are partly struc-tured by the rhetorical devices we use. As wenoted, the use of different rhetorical devices isexpected, and even encouraged, in the ‘art’ ofregional description. However, we do not seerhetoric and representation as optional extrasin human geographical research that only somehuman geographers – the more arty types,those leaning to the humanities – engage in.The degree to which human geographers areexplicit about rhetoric and representaionwithin their texts varies, but the degree towhich they engage in them does not.Rhetoricis as implicated in constructions of scientificobjectivity as it is in constructions of academicsubjectivity. As Richardson (1990: 15) puts it‘science does not stand in opposition torhetoric; it uses it’.We will now illustrate howscience uses rhetoric by looking at some of thedevices contained within the most ‘objective’pieces of writing.

Rhetorics of objectivity

This section develops the notion thatscientific objectivity does not stand outsiderhetoric but, rather, is in part constructedthrough it. In so doing it helps to confirm themessage that, ultimately, all research is textu-ally constructed. It is, as we noted at thebeginning of the chapter, written throughrather than simply written up. Before wecome on to the way research is presented inwritten accounts we will spend a short timeexamining what might be claimed as some ofthe most objective sort of geographical infor-mation we can find: railway timetables.We dothis to illustrate how even these are highlyrhetorical and implicated in wider connota-tions.As Bonsiepe puts it:‘Information withoutrhetoric is a pipe-dream … “Pure” informa-tion exists for the designer only in arid

abstraction. As soon as he [sic] begins to giveit concrete shape, to bring it within the rangeof experience, the process of rhetorical infil-tration begins’ (cited in Kinross, 1985: 30).

Figures 11.3 and 11.4 show two LondonNorth Eastern Railway (LNER) and time-tables, the first from 1926 and the secondfrom a redesign shortly after this date. Themajor difference between them is that oftypeface, although dashes have been substi-tuted for dots in alternate rows and the dotsare further apart. The redesigned versionshown in Figure 11.4 uses Gill Sans typeface.It was chosen to give all LNER printedmatter a common identity. More than this,however, it was felt to possess some intrinsicqualities which made it especially suited tothis purpose.As explained at the time, ‘it is so“stripped for action” that as far as glance read-ing goes, it is the most efficient conveyor ofthought’ (Kinross, 1985: 23, emphasis inoriginal). Such efficiency was necessary to helpthe ‘passenger being jostled on a crowdedplatform on a winter evening, and trying withone eye on the station clock to verify theconnections of a given train’ (Kinross, 1985:23). Kinross goes on to contrast these exam-ples with the design of later timetables fromboth British Rail (Figure 11.5) and the Dutchrailways (Figure 11.6). Each of these uses dif-ferent rhetorical combinations of typeface,type style, symbols, spaces, dots and lines toimpart their ‘information’. When takentogether these carefully designed combina-tions of textual features tell us something‘about the nature of the organization thatpublishes them’ (Kinross, 1985: 21).The ‘senseone has of … their respective contexts … is aconsequence of the rhetorical devices theyemploy’; in short, these examples remind usthat ‘nothing is free of rhetoric’ (1985: 29),and that even the most straightforwardgeographical information (the time and placeof the train) is conveyed to the reader throughdeliberate rhetorical devices.


354

3151-Ch-11.qxd 3/3/2004 3:27 PM Page 354


355

Figure 11.3 LNER Railway timetable, 1926

Source: Kinross (1985: 19)

Figure 11.4 Redesigned LNER Railway timetable


3151-Ch-11.qxd 3/3/2004 3:28 PM Page 355


356

Figure 11.6 Dutch national railway timetable, 1970–1


Figure 11.5 British Rail timetable, 1974–5


3151-Ch-11.qxd 3/3/2004 3:28 PM Page 356

Our second example of the deliberatecrafting of particular forms of expression inthe context of varying social fields is con-cerned with the process of research and, moreparticularly, with the way this process is pre-sented in objective scientific papers. It isdrawn from the work of Gilbert and Mulkay(1984), which explores the contextuallyspecific written and spoken discourses of bio-chemists in the field of ‘oxidative phosphory-lation’. Although not directly taken fromgeography this is a powerful example of theway that the same piece of research isaccounted for differently at different timesand to different audiences – including injournal papers, coffee-room chat and gossip,and formal explanations to sociologists in aninterview.Their central message is not to pulldown the façade of objectivity (which is pre-sented through written journal articles) but torecognize it as a constructed voice, and as onlyone of the many voices constructed duringresearch. When seen in this light ‘scientific’writing loses its position as the voice ofuniversal remit and is seen instead as a partic-ular rhetorical construction set up to serve aspecific purpose of persuasion.

Gilbert and Mulkay are particularly con-cerned to compare accounts of the same pieceof experimental science as represented in apublished paper and in conversations aboutthe research. The opening paragraph of theintroduction to the scientific paper reads asfollows:

A long held assumption concerning oxida-tive phosphorylation has been that theenergy available from oxidation-reductionreactions is used to drive the formation ofthe terminal covalent anhydride bond inATP. Contrary to this view, recent resultsfrom several laboratories suggest thatenergy is used primarily to promote thebinding of ADP and phosphate in a catalyti-cally competent mode [ref. 1], and to facili-tate the release of bound ATP [refs 2 & 3]. Inthis model, bound ATP forms at the catalyticsite from bound ADP and phosphate with

little change in free energy. (Gilbert andMulkay, 1984: 41)

The formal account of the research presentedhere contains several rhetorical devices to per-suade the reader of the validity of the paper’sargument. The opening sentence portraysother scientists’ work, but not as reasonablyheld beliefs associated with experimental evi-dence. Instead these existing accounts arelabelled as mere ‘assumptions’. This preparesthe reader for the alternative view, which isduly delivered in the second sentence. Thistime, however, the reader is led to believe thatthis alternative opinion is based on hard dataobtained from several laboratories – clearly acase of ‘one group’s results undermining theother group’s assumptions’ (Gilbert andMulkay, 1984: 43). This point is then drivenhome in the third sentence which refers to amodel. Gilbert and Mulkay conclude that ‘inthe course of three or four sentences, the texthas conveyed a strong impression … that therest of the paper is based upon a well estab-lished analytical position which constitutes amajor advance on prior work’ (1984: 43).However, no biochemical findings have beenpresented to support this impression: it hasbeen achieved through rhetorical deviceswhich convey particular views.

That these are also partial and highly selec-tive views, one voice among many, becomesapparent when conversations with the paper’sauthor are explored. In these he describes hisreaction to the ideas embodied in the ‘model’referred to in the paper when these werefirst suggested to him by the head of hislaboratory:

He came running into the seminar, pulledme out along with one of his other post-docs and took us to the back of the roomand explained this idea that he had … Hewas very excited. He was really high … Ittook him about 30 seconds … And so we satdown and designed some experiments toprove, test this…. It took him about 30 secondsto sell it to me. It really was like a bolt. I felt,


357

3151-Ch-11.qxd 3/3/2004 3:28 PM Page 357

‘Oh my God, this must be right! Look at allthe things it explains’. (1984: 47)

Thus contrary to the formal paper, where weare told that experimental results suggested amodel which seemed an improvement onearlier ‘assumptions’, the interview accountsuggests a dramatic revelation which wasimmediately seen to be right. Gilbert andMulkay point out that ‘Whereas in the [for-mal account] the model is presented as if itfollowed impersonally from experimentalfindings, in the [informal account] thesequence is reversed and the importance ofintuitive insights is emphasised’ (1984: 48).

Several other differences are revealedbetween the two accounts, concerning thestatus of experiments, the degree of commit-ment to the model and the fact that only twolaboratories seem to have been involved ratherthan ‘several’. Gilbert and Mulkay concludethat the participants’ actions and beliefs weresystematically different in the two settings,leading them to identify ‘two major interpreta-tive repertoires, or linguistic registers, whichoccur repeatedly in scientific discourse’ (1984:39). Our point here is that these competingrepertoires are built up through different kindsof rhetoric.The formal scientific register, usedin the paper, constructs the research processand those who take part in it in a particularway – as rational, logical and impartial. A cer-tain style of writing is used which constructsthe scientists’ actions and beliefs as followingunproblematically from the results of imper-sonal empirical findings.However, in conversa-tion, the scientists constructed accounts of theiractions which were much more contingentand dependent on personal and social circum-stance.They also revealed their thoughts to beinspired and institutional, even quasi-religious.The differences in these two ‘linguistic regis-ters’ reveal that scientific objectivity is alsopartly constructed through sets of rhetoricaldevices which are deliberately employed aspart of a particular writing strategy.

Representation in practice

The previous section has shown how rhetoricis used to structure all written accounts ofresearch, whether humanistic and cultural orscientific and objective. This general pointabout the importance of rhetoric can now bebroken down into more detail by looking atsome specific aspects of textual representationwithin human geography. All geographershave to negotiate these different aspects whenthey sit down to write, although they will doso to varying degrees depending on theirobject and method of study. However, theyusually deal with them implicitly and, inmany cases, almost automatically.We will con-sider them here in order to raise a set ofexplicit questions around key facets ormoments of textual construction.

The textualconstruction of data:inscription, transcriptionand description

The interpretation of data through the processof writing begins with the notes, commentsand observations which are made during theactual research process.Part I of the book drewattention to these from the perspective of dataconstruction. In this section we will shift theemphasis to look at the ways in which whatwe write when we undertake research alsobegins to shape our interpretation of the datawe produce. As we pointed out at the begin-ning of this chapter, writing is not somethingwhich is done only at the end of the researchwhen we are ‘writing up’. Writing is some-thing which is carried out all the way throughthe research process – and which is integral toit – from the jotting down of the very firstideas, to the notes made in the field, to thefinished article or dissertation.

The way in which elements of interpreta-tion are present in such writing, long before


358

3151-Ch-11.qxd 3/3/2004 3:28 PM Page 358

any formal analysis is undertaken, can beillustrated through the example of ethno-graphic research which we explored inChapter 6.There we referred to the many lay-ers of description which we undertake in thefield and through which ethnograhic data aretextually constructed. Clifford (1990) notesthat running across all these different layers ofdata collection are the three elements ofinscription, transcription and description (seealso Sanjek, 1990). For the ethnographer,inscription refers to the notes and jottings(sometimes just a mnemonic word or singlephrase) made by the researcher so that he orshe can fix an observation or later recall whatsomeone has said. These are attempts torecord social life ‘in motion’ as it happensaround the researcher and are thus usuallyswift and somewhat partial. They are madecontemporaneously with, or soon after, theevents which are observed or words which areheard. Description is the name given to theelement of writing which takes place at aslight distance. The account made is fullerthan in inscription and may well be done inthe evening when reflecting on the fieldworkundertaken during the day – often using theinscribed notes made in the field.The idea isto put together, to construct, a more or lesscoherent representation of what the ethno-graphic researcher has seen and heard duringhis or her research.Transcription refers to thecareful recording of an informant’s precisewords as they are spoken during the fieldworkencounter.

The point we wish to make here is that allthree elements are inevitably part of a creativeprocess, one in which data are always as muchtextually and intertextually constructed associally constructed. This is because therecording of data in the field is far frommechanical.As Clifford (1990: 57) puts it, ‘the “facts” are selected, focused, initially inter-preted, cleared up’. In many ways they have tobe because what we are creating through

fieldwork is a ‘recontextualised, portableaccount’ (Clifford, 1990: 64). The ‘presentmoment’ of the actual fieldwork encounter is‘held at bay’ and a ‘reordering goes on’.Thus‘fieldnotes are written in a form that willmake sense elsewhere, later on’ (Clifford,1990: 64). Hence the experiences that arerecorded in the field become ‘enmeshed inwriting … that extends before, after andoutside the experience of empirical research’(1990: 64). In this process of reordering, sothat events can be made portable and revealedelsewhere (in a paper, report or dissertation),an initial round of interpretation has alreadytaken place. Certain events and particularinsights have been privileged and others givenlesser prominence. In particular, in our desireto present a rounded coherent narrative,much of the complexity and messiness of thereal world gets written out and smoothedover. Hence Clifford remarks that ‘we seldomencounter in published work any cacophonyor discursive contradiction of the sort foundin actual cultural life and often reflected infieldnotes. A dominant language has over-ridden, translated and orchestrated thesecomplexities’ (1990: 59).

However, the manner in which we textu-ally fashion those whose behaviour andactions we are researching does not rely solelyon these ‘technical’ issues of recording, select-ing and reordering. There is always a moraldimension to how we describe and representothers. As Richard Rorty puts it, ‘whateverterms are used to describe human beingsbecome “evaluative” terms’ (1987: 245). Evenif the individual researcher uses particular ter-minology in a neutral ‘descriptive’ manner,‘there is no way to prevent anybody using anyterm “evaluatively”’ (Rorty, 1987: 245).Thus,while many social scientists, including socialgeographers, have used the term ‘underclass’to describe that proportion of the populationwho are denied access to work and as a resultend up having to cope with poverty and


359

3151-Ch-11.qxd 3/3/2004 3:28 PM Page 359

deprivation, others have used the term in amore evaluative sense to depict a shiftless, lazyand dangerous substratum of society. Weshould be wary of the ways in which our initialcategorizations and descriptions of those weare studying can carry a moral interpretationas well as an analytical interpretation. However,whether our interpretation is moral or analyt-ical, the key point to make here is that eventhe most preliminary notes and observationsmade in the field are already casting a partic-ular interpretive light on the eventual ‘find-ings’ of the research.

The textual construction ofinterpretations: metaphors

A second major way in which geographershave begun to cast an interpretive light ontheir data through the style of their writing isvia the use of metaphors. Metaphors arepowerful rhetorical devices which function tomake something intelligible to an audiencethrough equating it to something else. In thismanner the characteristics of one object (usu-ally familiar) are transferred to those ofanother object (usually less familiar) (Barnesand Gregory, 1997: 510). However, in thistransference metaphors become central to theconstruction of interpretations and ideas.Barnes and Duncan (1992) distinguishbetween big and small metaphor use.3 Smallmetaphors are those we use in an almosteveryday sense, which ‘pepper’ (Barnes andDuncan, 1992: 11) our writings. An exampleis, of course, the peppering of our prose withmetaphors.Many of these are unconscious andwe use them all the time – we accuse others ofglaring errors, we praise someone for tying upthe loose ends of his or her argument, we ripthe work of our competitors to shreds. Someof these everyday metaphors are geographicalin their content and are built around spatialelements. Pratt (1992), for instance, writes ofthe way social and cultural theorists use

metaphors of mobility, travel, marginality andborderlands to help construct their arguments.Geographers, for instance, have been especiallyfond of the metaphor of the frontier to helpconvey all kinds of images: of diffusion, spread,development, economic change, social contes-tation (see Smith, N., 1996). Small metaphorssuch as these are now part of the language andcannot be avoided – we simply need to recog-nize they are there and look to mobilize(another metaphor) the most effective ones tohelp make our arguments as convincing as wecan.

Big or large, metaphors are also importantin academic writing. They help to structurewhole schools of thought and ‘shape the veryexplanatory framework’ (Barnes, 2000b: 500)which these schools offer.They can also helpto determine research agendas. Barnes (1992;1996) uses the example of economic geogra-phy to discuss the manner in which theselarge metaphors have been employed. Heshows how spatial scientists drew onmetaphors of the gravity model and rationaleconomic man to construct their models ofspatial economic behaviour in the 1960s.Places became masses which interact accord-ing to a function of their distance apart, andurban land use is determined by the rationaluse of that space by individual actors. Ofcourse neither situation corresponds to thesocial and economic world we study, but themetaphors drawn from physics were used toconstruct models which were in turn used to‘understand’ urban and regional change (seealso Chapter 9 on the type of explanationfavoured by spatial science). However, Barnesshows that the use of metaphor has continuedin economic geography and was not restrictedto spatial science. Warde (1985), for instance,has criticized Massey’s use of a geologicalmetaphor where she conceives of distinctlocalities in terms of successive layers ofeconomic activity, laid down on top of oneanother. Similarly, Amin and Robins (1990)


360

3151-Ch-11.qxd 3/3/2004 3:28 PM Page 360

note that Scott and Storper’s work oneconomic agglomeration relies on themetaphor of flexibility. Once accepted, thesebig metaphors all helped to structure the wayhuman geographers thought about andexplained the uneven development of econo-mic activity. Again, such metaphors are diffi-cult to avoid. One simply has to be aware ofthem and employ them to maximum effect inone’s own writing.

Questions of genre

This section will concentrate on how textualconstruction operates in social and institu-tional contexts and will argue that one of themain expressions of this is through concep-tions of genre and generic conventions.Thesestructure what authors (can) write, and seek toconnect authors and readers into commonframes of reference. Sometimes the genre usedis defined by specific conventions. For example,most universities have rules and regulationswhich determine the length and style ofundergraduate dissertations. Different acade-mic journals also have different specificationsin terms of word length and stylistic features,such as footnotes. Publishers will have parti-cular expectations of textbooks and researchmonographs in terms of how these are laid outand constructed. However, we need to movebeyond these basic stylistic features of differentgenres to appreciate that modes of interpreta-tion are also predicated upon different textualconventions. Sayer (1989: 262), for instance,argues that one of the problems with muchphilosophical literature is that it assumes that‘the content of knowledge is indifferent to itsform’. In contrast, Sayer highlights how thecontent of knowledge is intimately connectedto the form.Thus enumeration, understandingand explanation are all predicated on differenttextual conventions. Using the example ofregional geography, Sayer shows how differentauthors structure their texts in different ways

depending on whether they are presentingatheoretical descriptive narratives, positivistmodels, structural analysis or realist concreteaccounts. This means that ‘writing … isinfluenced by its location within ongoingdebates and against rival academic and politi-cal groupings. So we not only write aboutthings but for, with and against others. In thissense, the researcher’s motives are not individ-ual but relational’ (Sayer, 1989: 269). Becausewriting is relational it is always partly definedand understood within, and in opposition to,particular genres. Broadly, it flows from thisthat genres and their institutional enforcementneed to be understood and recognized byreaders (see Agger, 1989a; 1989b), understoodand used by authors for strategic purposes, andpushed and bent by authors when appropriate.

Composition and style

This final section on individual elements ofrhetoric examines how the overall construc-tion of the text should be considered in termsof interpretation and is an important part ofwhat one actually says. Put simply, differentsorts of research, based on different materialsand with different audiences in mind,may sug-gest different compositional strategies.We canillustrate this with references to two pieces ofgeographical research on consumption – AlanPred’s book Recognizing European Modernities(1995) and Ben Malbon’s book Clubbing:Dancing, Ecstacy, Vitality (1999). Pred’s book issubtitled ‘A montage of the present’ and is verycarefully composed in a montage style. At theoutset Pred (1995:11) quotes approvingly fromWalter Benjamin: ‘Method of this work:literary montage. I have nothing to say, only toshow.’ Pred uses visits to three spectacularspaces – the Stockholm Exhibition of 1897,the Stockholm Exhibition of 1930 and acontemporary arena, the Globe – to examinehow new forms of consumption are entangledin changing discourses of power.He deliberately


361

3151-Ch-11.qxd 3/3/2004 3:28 PM Page 361

uses a montage style, producing a newcomposite whole from fragments of words andpictures. In so doing Pred (1995: 28) notesthat the avant-garde Soviet film-maker, LevKuleshov, referred to his own ‘rapid intersplicingof differently situated people and urban land-scapes as “creative geography”’. His book repre-sents his own attempt at creative geography, themontage of text, photographs, poems, illustra-tions and citations being used to explore differ-ent elements of European modernity.

The idea is to use these different textualdevices in a montage to ‘make visible the mul-tiple pasts layered within the present moment’(Pred, 1995: 264). Pred is attempting to usemontage to reveal connections and inter-relationships between elements of modernityusually held separate.The style and composi-tion of the book are very much a part of thestory being told. A new style of narrative isdemanded in order to convey the complexintermeshings of political economy and cul-tural politics, at different – yet related – histori-cal moments. Pred’s ultimate message is thatthe multiple modernities that have existedwithin any one European country, and thatcontinue to exist in the face of hyper- andpostmodernity, can be revealed only through atextual strategy which invokes multiple mean-ings and representations. The book ends(1995: 264) with an echo of the Benjaminquotation which began it (along with a nod toone of Lenin’s most important works):

What are the tension-riddled images to beevoked?What is to be shown?What is to be done?

In Pred’s case the very deliberate compositionof the book through a montage of fragmentsserves to highlight the complexity and inter-relation of different elements of consumptionwithin European modernity. Malbon (1999)uses a different compositional strategy toinform his understanding of consumption. In

his case the textual motive is not montage buta night out. Malbon’s object of study are thecultures and spaces of clubbing. He uses theseto explore a key element of the geography ofcontemporary youth culture in the UK. Asthe book is about the experience of goingclubbing, Malbon divides it up into threeparts which mirror a night out.The openingpart is labelled ‘Beginnings’ and is designed tocontextualize the night out yet to come. Inthis opening part Malbon considers the keyacademic starting points for analysing club-bing – young people at play, consumption andidentity formation. The second part com-prises the main empirical section of the book,which investigates the various spaces andpractices of clubbing. The reader is takenthrough four sections in this part, beginningwith club entry and finishing at the end of thenight. The final part mirrors those post-clubbing times of reflection in seeking tounderstand the events and experiences of theprevious night. Malbon deliberately mirrorsthe chronology of a night out in the narrativestructure of the book. His self-avowed aim isto examine the practices, imaginations andemotions of the clubbers themselves, and hehas carefully designed the composition of thebook to reflect this. Perhaps Pred and Malbonare extreme examples of the way that compo-sition can mirror and enhance content, butthey both remind us that the structure of ourwriting can help inform the reader’s interpre-tation of what we write.

Beyond the book

As we move into an age of multimedia pre-sentation, digital photography, video stream-ing, text messaging and digital storage, excitingnew possibilities are opening up for the pre-sentation of geographical research. In short, thepotential is now enormous for textualizingresearch through means other than books and


362

3151-Ch-11.qxd 3/3/2004 3:28 PM Page 362

journals. Geographers have only just begun tograsp this potential, so examples are limited –partly because of the disciplinary effect ofaccepted genres referred to earlier in thischapter. One example which we can referreaders to is the recent work of Michael Prykeon the redevelopment of Berlin followingreunification. Potsdamer Square, at the eco-nomic and cultural centre of the former EastBerlin, is being rebuilt with Western capital toform one of the largest redevelopment sites inEurope. Pryke has sought to capture theemerging economic, visual and audio rhythmsof this new urban landscape with a visual andaudio montage of Potsdamer Square. Theresulting ‘paper’ (Pryke, 2000) can be found atwww.open.ac.uk/socialsciences/geography/research/berlin/.This contains both audio andvideo clips of the area being redeveloped,and the ‘reader’ can click on different imagesand sounds which convey an impression of thenew Berlin.These visual and audio montagesare amenable to manipulation by the ‘reader’:you can zoom in or out of the video clips andstop and mix the audio sounds. The soundsavailable through the mini-disk recordingscomplement the visuality of the video clips.Asa result, the ‘reader’ is introduced to a range ofsensory experiences and can gain an apprecia-tion of the visual and audio rhythms emergingin the new Berlin which would be impossiblethrough written text alone.The different rela-tions between ‘author’ (or photographer andsound recordist) and ‘reader’ (or watcher andlistener) open up a realm of new interpreta-tions and perceptions.The result is a fascinat-ing glimpse of the possibilities which the newdigital media are opening up for geographicalcommunication (see also Pryke, 2002).

Conclusion

There are two main points to draw out of thischapter and to reiterate in this conclusion.

The first is that all human geography isfiction in the sense that it is textually madeand fashioned. But unlike most novels it islargely about real people and real lives andreal situations.The moral is that human geo-graphers have to find ways to craft accountsthat best suit the character of those lives andpeople and situations, while also attending tothe trick of engaging their audience. This isnot an easy process and partly explains whywriting is such a difficult part of the researchprocess.We have been at pains to stress howwriting is indeed part of the whole researchprocess and not just something which istagged on at the end when the ‘writing up’occurs. Secondly, this chapter has sought topresent some helpful tips on what to con-sider when writing through research – butwithout undue prescription. It does not seekto use writing as a disciplining device,though it acknowledges it is that. Rather, ithas tried to offer some routes through theprocess of textual construction that will letauthors develop their own voices, their ownstyles, their (and their research subjects’) ownmaterials in ways which are appropriate fortheir research. It has also sought to providesome criteria upon which to judge anddefend the choices they have made. But, ulti-mately, these choices will be made within ahost of competing political, moral and ethi-cal contexts. It is to these that we now turnin the final chapter.

Notes

1 See, for example, Becker (1986) and Wolcott(1990).

2 For those who wish to follow up such advicein detail, Northedge (1994) contains twoexcellent chapters on the craft and mechanicsof writing.

3 See also Barnes and Curry (1992) who writeof large and small metaphors.


363

3151-Ch-11.qxd 3/3/2004 3:28 PM Page 363

We want to end the book with some reflectionson the politics of geographical practice andon the power relations involved in both con-structing and interpreting geographical data.One of the central messages running throughthis book has been the idea that all geographi-cal research is a social activity. Far from oper-ating in some kind of sealed environment orsocial vacuum (let alone an ivory tower), indi-vidual researchers are always working in par-ticular social contexts. Indeed, the practice ofresearch should not be viewed as a neutral, ortechnical, procedure but should instead beseen as a social process which is heavily ladenwith competing, and often conflicting, sets ofdemands and values. May (1997: 46) pointsout that the values of those involved enter theresearch process at all stages – in the interestsleading to research (of both researcher and,where appropriate, funding agency); in theaims, objectives and design of individual pro-jects; in data construction and interpretation;and in the use made of research findings.Through an investigation of each of thesestages this chapter looks to situate the contextwithin which research practice takes place, aswell as to provide an appreciation of thedemands and pressures which are placed onthe researcher from a number of differentsources. It is these demands and pressures, andthe values which inform a response to them,which underpin what we term the ‘politics’of research. The critical point is that all

research takes place within these multiple andoverlapping sets of social relations and willinevitably be influenced by them.

We will indicate some of the key ways inwhich these social relations might affect theresearch process by focusing on different ele-ments within the overall practice of research.We are not seeking here to be exhaustive, andthere are complete books devoted to thepolitics of research.1 Rather, we will pick outa range of indicative issues which will help usto show that human geographical research cannever be viewed as value-free.We will beginby looking at those ‘personal’ politics whichare implicated in the nature of the research wechoose to do and in the ways we do it. Wethen move on to look at some of the broaderpolitics which press in on the individualresearcher and help to structure the contextwithin which academic research takes place –both those which operate within the researchprocess itself and those which influenceresearch practice externally.We will finish thechapter with a look at how the twin issues ofethics and morality underlie each of theseaspects of research.

The ‘personal’ politics ofgeographical practice

The first stage at which these various influ-ences are felt occurs at the very outset of the

The Politics of Practising HumanGeography12

3151-Ch-12.qxd 3/2/04 2:50 PM Page 364

research process when the research topic isdevised and initial research questions are for-mulated. We might find the inspiration forthese ideas in the books and articles thatothers have written or in discussions we havehad with colleagues or fellow students. Fieldtrips can be another source of inspiration, andencounters with new places and new peoplescan trigger all kinds of research ideas.Ultimately, though, we all carry with us ourown unique geographical imaginations whichare fired in different ways through differentmeans.As one of us has written elsewhere:

An understanding of why we are interestedin particular subjects/objects/modes ofhuman geographical study will not solely befound in the quality and persuasiveness ofthe canon of study which already exists,important though that is. Such an under-standing will also demand some scrutiny ofthe subjectivities, identities, positionalitiesand situated knowledges that we as individ-uals bring to the collectivity of that canon.(Cloke, 2004a: 1)

In this section we will talk through some ofthe ways in which these subjectivities, identi-ties, positionalities and situated knowledgesprovide the prompts for our individual prac-tice of human geography. We can combinethese influences under the label of ‘personal’politics – deliberately written with a small ‘p’because they are not reducible to the moreformal political world of elections, parties andpressure groups.This is not to deny the influ-ence of formal politics (Harvey, 1996; 2000;Watts, 2001) but to acknowledge that formany of us the impulses which push our geo-graphical sensibilities in a particular directionare much more diverse.Tracey Skelton (2001:89), for instance, has written of how the field-work for her PhD research ‘was initially aboutmy wishes, dreams and desires’.These in turnwere originally prompted by a school visitfrom Miss Peacock (a missionary who hadbeen in India) which left a childhood dreamof spending time abroad and learning about

other ways of life. When combined with adesire to conduct research from a feministstandpoint, which emerged through herdegree, this led to a decision by Tracey topursue research into gender relations in theCaribbean. Here we see an example of theway that the ‘persuasiveness of the canon ofstudy’ – in this case feminist geography –combined with Tracey’s own subjectivity andpositionality to forge her particular researchpractice.

In the opening chapter of the book weintroduced you to Carl and Linda and toldyou something of the way they practised theirparticular human geographies.We also intro-duced you to some of the sensibilities andvalues which influenced such practice. Wenow want to introduce you to a third geogra-pher, whom we will call Rachel. Like Carland Linda, Rachel is a real individual. Shecould well be you, for her story begins whenshe was an undergraduate student. RecountingRachel’s story allows us to illustrate themanner in which personal identities, emotionsand values help to shape the nature of theresearch we do, and also how the very prac-tice of that research may well encourage us toreflect back on, and even change, our ownvalues and identities.We have chosen Rachelto illustrate these points because she has writ-ten so powerfully, eloquently and honestlyabout the manner in which her own geogra-phy has been practised. Her route into geo-graphical research, and the influences on hersubsequent journey through it, are outlined inBox 12.1.

Rachel’s journey is instructive for severalreasons. It confirms how the impetus to doresearch often comes from deep within us, outof our personal engagement and desires. Italso shows how the influence of others – inthis case, academics and fellow students – canhelp to shape our geographical imagination(see Cloke, 1994, for a similar account of suchinfluences).And it shows how our positionality

POLITICS OF PRACTISING HUMAN GEOGRAPHY

365

3151-Ch-12.qxd 3/2/04 2:50 PM Page 365

In 1992 Rachel left a small Yorkshire mining town to begin a geography degree. Herinitial experiences at university were far from positive and the social and culturalenvironment she found jarred very strongly with the close-knit working-class commu-nity she had just left. As she has written, ‘I arrived at university and suffered massiveculture shock. I felt like an alien from another planet…my first year at university was anightmare, and all I wanted was to wake up’ (Saltmarsh, 2001: 142). But out of thisnightmare came anger – at academia and academics – and this began to inform herwriting and research. Gradually she replaced the opinion of others with her own, andher work became semi-autobiographical. As she explained:

I was being forced to think about the social and cultural implications of my being[at university] due to the very nature of my degree subject. Lectures about class,urban and social geography, cultural geography and the way in which we wereintroduced early on to the politics of knowledge construction and research tech-niques, really forced me to think about how people and places were beingdescribed and analysed by mainly middle-class geographers and how this was tiedup with my own experiences of being caught between two cultures. Crucially itwas also encouraging me to think about who was speaking for whom and why.It was about power and I felt at the time that I had none. (2002: 4)

In an effort give voice to those who were powerless, Rachel decided to undertake herundergraduate dissertation on pit closures and the subsequent loss of mining cultures.That it was written from an autobiographical standpoint helped her to deal with herpast and her present; as she put it, ‘It was at [university] that I discovered who I was,who I wasn’t and who I would never be’ (2002: 2).

Rachel’s dissertation topic and style had come directly from her need to explore herown roots and her desire to let working-class people speak for themselves. Her geo-graphical imagination had been fired by her experience of university and her loss ofcommunity. She explained how writing this dissertation in turn ‘kick-started my jour-ney into further PhD research based around my class-culture, situated knowledge andautobiographical writing’ (2002: 1).

But when she returned to her community to undertake her PhD research and con-tinue with her particular form of geographical practice, she found that something hadchanged:

The most startling issue I found during my PhD was the fact that my own class cul-tural identity was being re-created…In contrast to the experiences of my disser-tation research, I often felt out of place in the town where I was born…The ironyis that my class identity has kind of reversed itself on my journey through my edu-cation. I was held up by some of my interviewees as being living proof that it waspossible to transform yourself from being working class into someone who ismiddle class through the education system. (2002: 7)

But it wasn’t just Rachel’s interviewees who saw her in a new light and were placing anew identity on her. In the course of her PhD research she also experienced what shehas described as a ‘loss of faith’. As she puts it:

I can’t explain what this loss of faith is exactly, but it has something to do withmy changing identity. Going back to the places and people where I came from hasmade me realize that I have been gradually changing over time, without even

Box 12.1: Rachel’s journey


366

3151-Ch-12.qxd 3/2/04 2:50 PM Page 366

Box 12.1 (Continued)

realizing it…I am a different person to the one who left the working class, councilhousing estate to go to university. I’m a different person to the one who came tobegin my PhD…I am re-negotiating my identity constantly. (2001: 146–7).

Rachel’s account of her journey is evocative and moving. In the course of it she opensup all kinds of issues about class, gender and community, and the painful process (forsome) of ‘gaining’ an education. We can only deal here with the elements which concernresearch practice, but her writing reveals much much more about the wider culturaland social politics of higher education.

within the research process, discussed inChapter 1, is constantly changing and alwayshaving to be renegotiated. In the course oftwo or three years, between undertakingfieldwork for her undergraduate dissertationand fieldwork for her PhD – in the sameplace – Rachel had changed in the eyes ofsome of those she was interviewing.They nolonger saw her as one of them but, instead,positioned her as part of an educated middleclass. This repositioning by the research sub-jects will have inevitably affected the practiceof research itself – for what people tell you,what they reveal and what they let you knoware conditioned by how they see you and howthey respond to you.As Claire Madge puts it,the practice of fieldwork involves

playing out a multiplicity of changing rolesduring the course of the research. Theseroles, which are sometimes complementary,sometimes clashing, and which are contin-gent on our positionality, will affect thedata given/gained and our subsequentinterpretations. In other words they willinfluence what we produce as knowledge.Personal relationships with people willinfluence the ethical decisions we makeregarding what we create as knowledge.Power, ethics and knowledge are inter-connected. (cited in WGSG, 1997: 94–5)

We will now explore this interconnection alittle further by talking about the politics ofthe research process itself. In doing so we willshift attention from the subjectivities of theresearcher, which help to shape what and how

he or she chooses to research, to look at thewider politics which impinge upon the practiceof research once it is under way.

The politics of researchpractice

Whatever the motivations for the shape andcontent of the research questions, there is anincreasing recognition that the practices ofresearch are themselves highly politicized.This issue can be explored on a number oflevels but all involve viewing the researchprocess itself as an exercise of power –whether power over the type of knowledgeproduced, power over those who are being‘researched’ or power within the researchteam itself. We have been at pains throughoutthe book to stress that research is not simply aneutral activity for unveiling information but,instead, is intimately involved in the actualconstruction of that information. What isunder consideration here is the manner inwhich this is done.

One of the reasons why it could be main-tained for so long that the purpose of researchwas rather unproblematically to ‘collect facts’is the distance that has traditionally beendrawn, and indeed encouraged, between theresearcher and the researched. It is a long-heldnotion that the more ‘objective’ the researcheris, the better and more reliable are the research


367

3151-Ch-12.qxd 3/2/04 2:50 PM Page 367

results. But such ‘objectivity’ is impossible insocial research and, as Acker et al. put it, ‘nei-ther the subjectivity of the researcher nor thesubjectivity of the researched can be elimi-nated in the [research] process’ (cited in May,1997: 52). Indeed, as we have already noted inChapter 5, the practice of research can neverbe a neutral exercise since ‘the very act ofentering the worlds of other people meansthat the research and the researcher becomepart co-constituents of those worlds.Therefore we cannot but have impact onthose with whom we come into contact’(Cloke et al., 2000a: 151, emphasis in origi-nal). Thus, rather than being viewed as twoseparate parts of a purely technical researchprocess, the researcher and the researched areco-constituted during the practice ofresearch. Each influences the other and isaffected by the other.

This situation is made more complex, andeven more difficult to negotiate, by the factthat neither researcher nor researched are sta-ble and singular categories. Each will shift asdiffering identities come to the fore duringthe process of research. The identity of bothresearcher and researched will be multiple –embracing and crossing, for instance, categ-ories of gender, class, race, sexuality, age anddisability. This helps to problematize thesomewhat straightforward categories ofresearcher and researched. We have alreadygiven an example of how this might happenin Chapter 5 where we quoted a researcherreflecting on her position vis-à-vis those shewas interviewing for a project on rural home-lessness (see also Box 5.7):

The stronger part of my self-identity is thatof a single parent, and …I have, on occa-sion, had difficulty (as many people do) withtrying to maintain my rent payments andprovide for my family. These experiencesallowed for a certain degree of identifica-tion with my interviewees but, in certain sit-uations, this proved problematic – at times itfelt as if I was at once the researcher and the

researched, perhaps too much of an ‘insider’in some ways. (Cloke et al., 2000a: 143)

Such experiences are worth repeating here notonly because they alert us to the sensitivity ofthe relation between researcher and researchedbut also because they indicate the micro-politics involved in the very practice of research,as the researcher strives to balance distance andcloseness. Just how these issues are balanced, andthe position taken on the continuum betweeninsider and outsider, will affect the constructionof research data. We can explore these micro-politics of research a little further by introducingyou to a fourth geographer, called Robina. Heraccount of her research with Muslim women isset out in Box 12.2.

The complexity of the identities negotiatedby Robina and her respondents during the actof research, and the impact these had on thedata constructed, remind us that research is anembodied process and that people’s position-alities can enable as well as constrain. Hencethe outcome of research is always dependenton the micro-politics of the research process.As Robina puts it,‘the research “findings” canbe seen to be “produced” out of a set of inter-actions and there will be differences in whateven different researchers of the same colour,gender and/or class may find or give signifi-cance to’ (Mohammad, 2001: 108).

Thus, in addition to considering themultiple roles and positions of any individualresearcher, it is also the case that differentresearchers will approach the same researchsituation differently, and thereby constructdifferent data from it. Cloke et al. draw ourattention to these differences, and to the poli-tics involved in negotiating them, when theyrecount the experience of a particularresearch situation:

Although I was equipped with a mobilephone, there were some situations in whichI felt I was putting myself in danger and thisrestricted my ability to do some of theresearch. There were some places I just wasn’t


368

3151-Ch-12.qxd 3/2/04 2:50 PM Page 368

Robina Mohammad was involved in a series of research projects with Muslim womenbetween 1994 and 1996, using participant observation, questionnaires, focus groupsand semi-structured interviews. The way that she has written about these experiencesreveals important insights into the positionality of the researcher and the politics of theresearch process (Mohammad, 2001). Robina is a Pakistani Muslim by birth but wasbrought up in Britain. Despite cultural and social differences forged through herupbringing and education, the Pakistani women Robina was researching saw her asone of ‘them’. In order to play out this role she was expected to attend Friday prayersat the Muslim Women’s Association (2001: 107). However, this status as insider was bothbeneficial and detrimental to the research. It was beneficial because it allowed accessto material that other researchers would have been denied. As Robina puts it:

I would often hear remarks such as ‘we shouldn’t really say this since it is disloyal[to the group] but since you are one of us it is all right’, or ‘I can tell you becauseyou’ll know what I’m talking about’. These comments were followed by the shar-ing of gossip about the group or personal experience that otherwise, it was sug-gested, would have been withheld or presented differently. (2001: 108)

As well as opening up closed knowledge, her status as ‘insider’ also hampered the con-struction of data. In these cases, Robina almost appeared ‘too close for comfort, mak-ing people wary of sharing information’ (2001: 109). Ultimately, Robina had tonegotiate her positionality with each of her respondents, illustrating that the positionof researcher is not fixed and immutable but highly fluid and somewhat vulnerable.But her position also affected the conduct of the research practice. Robina tended todress in a Westernized style, which for some interviewees placed her on the‘“insider”/”outsider” border [and] mediated both mine and the researched’s perfor-mances in the situation of the interview’ (2001: 110). She was seen variously as ‘same’,‘other’ and simultaneously ‘same’ and ‘other’ – some women spoke to her only becauseshe wore Western dress, while for others this created a barrier. The implications for theconstruction of data were manifold. As Robina concludes: ‘I had a rough researchagenda to follow, but what was created out of the process of fieldwork has never beenentirely what could have been predicted…the knowledges produced are alwaysversions. They represent one out of other possible truths’ (2001: 113).

Box 12.2: The researcher as insider/outsider

able to visit [and] missed opportunitieswhich resulted from my position as a lonefemale researcher who lacked experience inhandling certain situations. (2000a: 145)

The research process is always a complexsocial situation and the data constructed stemfrom an interactive set of activities which areconstantly negotiated and renegotiatedbetween the researcher and the researched.Thepolitics surrounding these negotiations – interms of positionality, identity and power –will of necessity impact upon the data con-structed. As Robina suggested (see Box 12.2),

the knowledges which are constructedthrough these negotiations and renegotiationsare always ‘versions’. On a different day, withother participants, other possible (and com-peting) narratives would emerge.

Another dimension to the identity of theresearcher is added by their potential positionin a research team. Both undergraduate andpostgraduate research tends to be carried outon an individual basis, but academic staff willoften work in research teams. Large-scalefunded projects are often undertaken by morethan one person and may involve a team of


369

3151-Ch-12.qxd 3/2/04 2:50 PM Page 369

several staff. The politics surrounding theoperation of such teams are at once externaland internal. Internally, the research team willaim to function as a unit, like all teams.Project directors, research staff, administrativeand secretarial staff will all attempt to worktogether and will have various tasks withinthe overall framework of the research project.Inevitably different sets of social relations willbe involved between the different members ofthe team, and the quality of the research datawill be partly dependent on the characteristicsof these relations. Externally, the politicsinvolved will operate around the fact that dif-ferent staff members perform different tasksand are involved in different elements of theresearch process. Cloke et al. (2000a) notehow hierarchical positioning within the pro-ject team may lead to particular responsibili-ties being discharged within the researchprocess.As one of the research staff put it:

I can’t help but wonder whether myincreased sense of ethical dilemma was notalso partly due to the fact that, as anemployee and part of a team, I felt some-how more beholden to my employers andcolleagues to do a particularly good job ofwork and come up with the stories I was col-lecting…the ‘status’ and ‘role’ of anyresearcher in a team are bound to affect thedifferent ethical dilemmas and political‘conflicts of loyalty’ they are likely to experi-ence. (2000a: 144)

The politics of theacademy

The task of research teams is to begin to shiftour attention towards those external pressureswhich push down on the researcher in certainsituations. In many ways the intellectualhistory of the discipline of geography can beread through the lens of these external pres-sures. In his book tracing this history, DavidLivingstone wrote that ‘The heart of my argu-ment is simply that geography changes as

society changes, and that the best way tounderstand the tradition to which geogra-phers belong is to get a handle on the differ-ent social and intellectual environmentswithin which geography has been practised’(1992: 347). Hence geography’s formation asa university subject in the late nineteenthcentury was linked to its utility as a science ofexploration and mapping to be used in theservice of colonial expansion, and the famousexplorer H.M. Stanley noted how geography‘has been and is intimately connected withthe growth of the British Empire’ (cited inLivingstone, 1992: 219).The shift to quantifi-cation in the 1950s and 1960s has been inter-preted by Harvey (1984) as ‘a strategicmanoeuvre to escape the political suspicionfalling on social science’ (cited in Livingstone,1992: 324) during the McCarthy era ofwitch-hunts and repression in the USA, andeven if one accepts the alternative argumentthat quantification stemmed from a search forscientific credibility, there is still a broaderpolitics here in terms of what counts asscience and what is seen as credible. Likewise,the backlash against quantitative spatial sciencethat swept through the discipline from the late1960s onwards was fuelled in part by a polit-ical desire to provide a critique of social andeconomic inequality. The opening editorial ofthe newly founded journal Antipode in 1969(Vol. 1, no. 1) advocated the ‘replacement ofinstitutions and institutional arrangements inour society that can no longer respond tochanging societal needs, that stifle attempts toprovide us with a more viable pattern forliving, and often serve no other purpose thanperpetuating themselves’. And, of course, themore recent rise of new concerns withingeography (such as those operating aroundissues of gender, race and sustainability) hasmirrored the growth of these as politicalmovements within society more generally.

Those seeking a handle on the social andintellectual environments which are helping


370

3151-Ch-12.qxd 3/2/04 2:50 PM Page 370

to shape the current practice of geography areforced to confront significant change. Overthe past two or three decades the rather con-genial relationship between well-funded uni-versities, government and society has begunto breakdown. Under this system, public trustand government funds combined to allowresearch to be largely driven by individualintellectual curiosity. This is no longer the caseand, as Demerrit notes, the change has hadimplications for research as ‘older narrativesabout scientific knowledge as a public goodor as an end in itself no longer hold sway’(2000: 309). Instead, university funding isincreasingly subject to tight budgetary controland external accountability.This is not neces-sarily a bad thing but, as Demerrit notes, ‘theproblem with the new social contract [sus-taining academic research] is not so muchwith the idea of making academic researchpublicly relevant and accountable, but withthe narrowness with which public relevance,accountability and value have been defined’(2000: 324).

The narrowness of the definition of rel-evance has had a significant impact on thecontent of geographical research, especiallyin the USA. Here, the US National ResearchCouncil published a report entitled Redis-covering Geography: New Relevance for Scienceand Society (1997). In it, a case was made forgeography’s utility as ‘good science and socie-tally relevant science’. But such utility wasinterpreted in a very particular and partialmanner. As Johnston points out in his reviewof the report:

it has identified a market for geographerswith GIS and related analytical skills and hasdecided that this is the market that the dis-cipline should be orientated towards…Radical geography, despite its major contri-butions – theoretical, empirical and practical –to the discipline’s development over the lastthree decades of the 20th century, wasimplicitly characterised as ‘nonscientific’ andexcluded from…the priorities for…teachingand research. (2000c: 988)

There are thus all kinds of politics surroundingthe practice of geographical research which,together, help to define and delimit particularspheres of research activity as ‘relevant’ andtherefore more legitimate than others. Someof these are internal: the US NationalResearch Council report, for instance, wasdrafted by a committee of 16 practising geo-graphers from within the discipline.2 In otherinstances the pressure to define legitimate andacceptable fields of research inquiry comesfrom outside the discipline – especially fromthose who actually provide funds for research.In the UK those public sector agencies andcharities that fund large-scale social science –such as the Economic and Social ResearchCouncil (ESRC), the Joseph RowntreeFoundation, the Leverhulme Trust and theNuffield Foundation – all distinguish a set ofparticular themes as priorities for researchfunding (their current priorities can be foundon their respective websites). It remains easierto obtain funding for research which is insidethese themes rather than outside them.Individual researchers, or research teams, areable to shape their research proposals to someextent but, in order to win research monies,they must remain within the contours set bythe funders. In this sense most research isreactive rather than proactive in that itresponds to calls from funding agencies whowill all have a particular politics to theirresearch themes and funding streams. In otherwords, the model of the lone scholar produc-ing research on items of his or her choosing isnow fading and most large-scale geographyresearch projects now take place within asystem which is shaped by the demands of theresearch funders.

The private sector also has a role in deter-mining which aspects of research are deemedas ‘relevant’ or ‘legitimate’ and therefore wor-thy of funding. These pressures are perhapsmost clearly felt in those areas of academiawhich have a directly applicable economic


371

3151-Ch-12.qxd 3/2/04 2:50 PM Page 371

benefit (such as biotechnology,genetics, softwareengineering or pharmaceuticals) but they doimpinge upon particular areas of geography.Curry (1998), for instance, shows how indus-try sponsorship of GIS research and teachinglaboratories has tended to undermine acade-mic autonomy (see Demeritt, 2000: 324).Another example of the encroachment of pri-vate sector concerns into academic geographyis provided by what became known as the‘Shell debate’ within the Royal GeographicalSociety–Institute of British Geographers(RGS-IBG). In 1995 the RGS-IBG becamethe professional body representing academicgeographers in the UK following a merger ofits two constituent organizations – althoughsome have argued that, given the unequalpower relations which existed between theparties involved, the word merger is perhapsnot entirely appropriate (Tickell, 1999). Theoil company Shell was a corporate sponsor ofthe RGS, giving around £45 000 a year to itsExpedition Advisory Centre for young people(Gilbert, 1999: 224;Woods, 1999: 229).Therewere anxieties about such sponsorship in dis-cussions over the merger, with many acade-mic geographers expressing concern aboutthe environmental impacts of Shell’s miningand drilling activities.

These concerns reached a peak inNovember 1995 when the Nigerian writerand political activist, Ken Saro-Wiwa, andeight other activists from the Ogoni region ofthe Niger Delta, were murdered by theNigerian military regime (Gilbert, 1999).Saro-Wiwa was a critic of Shell’s oil opera-tions in the Niger Delta, claiming they devas-tated the environment and did little to benefitthe Ogoni people. Following the executions,a campaign was mounted by some academicgeographers to end the RGS-IBG’s relation-ship with Shell. This was unsuccessful andShell is still a sponsor of the merged society.Many of the geographers active in the cam-paign against Shell’s sponsorship have resigned

from the RGS-IBG. Others have remained inthe merged organization.This episode raises awhole host of issues concerning the relation-ship between academic research and privatesponsorship (Kitajima, 1999;Watts, 1999b).AsGilbert notes:

the debates over the future of the Shell con-nection have forced British geographers tothink about environmental and politicalethics, about the public role of their profes-sional body, about the relationship betweenexternal funding and academic indepen-dence, and about the tensions between theacademy and activism…It is hard to imaginean issue more calculated to test understand-ings of the nature of the discipline, and thepolitical and ethical consequences of ‘beinga geographer’. (1999: 220)

What is interesting is that it was the actionsof a private company on another continentwhich stimulated such a fundamental ques-tioning of the consequences of ‘being ageographer’.

For undergraduates, the consequences of‘being a geographer’ may well be a temporaryconcern, one which is over in three years. Butfor many academics the nature, if not the con-sequences, of ‘being a geographer’ is some-thing which has to be confronted andnegotiated on a fairly frequent basis. A largepart of being a geographer in a contemporaryacademic setting is bound up with research,and there are innumerable choices that haveto be faced over both the content and direc-tion of this research activity. Some of thesechoices are closely linked to the kinds ofexternal pressures we have just sketched out.Most academic research, for instance, requiresfunding of some kind or another – to pay forresearch staff; to fund visits to specialistlibraries and archives; to finance fieldwork; tobuy computers and software; to pay for secre-tarial support; to purchase research reportswritten by others; and so on.Yet as we notedabove, those who fund research are becomingincreasingly prescriptive. Particular types of


372

3151-Ch-12.qxd 3/2/04 2:50 PM Page 372

research on particular subjects stand morechance of finance than do others.

In addition to determining the nature ofthe research agenda, the funding agencies willalso play a key role in shaping research output.Hence what is written within the discipline isalso shaped by external pressures. Just as stu-dents have to respond to a set of rules andregulations concerning what they produce fortheir degrees, academics have to meet exter-nal requirements which shape what theywrite. Each research project will almost cer-tainly have its own output requirements.TheESRC has traditionally expected research dis-semination to concentrate on academic out-puts in the form of books, chapters or journalarticles but, in its desire to make itself more‘relevant’, it now expects the projects it fundsto produce outputs which are also aimed atthe ‘users’ of research, ranging from the gen-eral public to the policy community.3 TheJoseph Rowntree Foundation requires a finalreport to be presented to it rather than anyacademic output, but it also expects usergroups to be informed of the research results.The Nuffield Foundation also requires aresearch report. These requirements willinevitably shape the interpretation and dis-semination of the research and help to deter-mine how, where and in what format theresearch is presented. Again, this is not amatter purely for the researcher to determine,and we need to move away from the notionof the individual academic huddled over adesk writing up his or her research accordingto his or her own wishes.While the desires ofthe individual researcher to make a particularintervention in a certain set of academicdebates will play a part in determining whatgets written, a larger factor will be the needsand requirements of the funding agencies.

In the UK, research output is also driven bythe Research Assessment Exercise (RAE).Those who fund higher education now usethis exercise to determine the level of baseline

research funding to each academic department.In total, around £5 billion of research fundswere distributed in response to the results ofthe 2001 exercise.These exercises, which havebeen running since 1986, have considerablyheightened the culture of performance andevaluation within UK universities (seeDemeritt, 2000; Smith, D., 1995b; 2000). Eachacademic is expected to produce a certainnumber of publications over the lifetime ofeach research cycle, which are then rated forquality by a panel of expert reviewers.This initself has been enough to alter the kind ofresearch undertaken and the nature of theoutput produced. David Demeritt has writtenof his own experiences of these demands, as ageographer at the beginnings of an academiccareer.As he puts it:

I feel such pressures impinging on my careerpath. While academic employment dependson a generalized ‘excellence’ in research,some research clearly pays better than doesother research. Having trained as a histori-cal geographer, I have found success in theacademic labour market only by emphasiz-ing my environmental expertise, whichpromises the kind of relevance valued byfunding agencies. (2000: 321)

In addition to impinging upon the careers ofindividual researchers and academics, thepressure to deliver certain types of research inparticular formats to satisfy performancereview criteria has also had an impact onresearch dissemination. Publishers and journaleditors have reported a large rise in the typesof submissions most favoured by the RAE –journal articles and research monographs –and a fall in the type of work not consideredRAE relevant, most noticeably textbooks andunpublished research reports. This may wellbe a major reason for the fact that, despite agrowing sensitivity within human geographyto exploring the world through a ‘polyphonyof voices’, this has not yet been matched by afull recognition of multiple audiences, otherthan those which are demanded by research


373

3151-Ch-12.qxd 3/2/04 2:50 PM Page 373

funders (Milbourne, 2000). As Cloke et al.conclude, ‘New words and new worlds havenot really travelled to new ears and eyes’(2000a: 147). In this sense research is still pri-marily carried out ‘on’ others rather thanbeing done ‘for’ others.

To raise these issues is to begin to appreci-ate the various ‘politics’ which are involved ingeographical research. As we noted earlier, inthe words of Claire Madge, power, ethics andknowledge are indeed interconnected.And aswe have seen, these politics and ethical con-siderations operate at every level throughoutthe research process – from the macro-scalepolitics involved in national-level researchfunding, to the micro-scale politics involvedin the relation between researcher andresearched.When wrapped up and looked atas a whole, what they tell us is that the designand implementation of research should not beseen as a narrowly technical issue, somehowuncontaminated by political and ethical ques-tions. Instead, all stages of the research processare affected by these kinds of issues. Questionsof how to negotiate and resolve them arecomplex and multidimensional, involvingresearch funders and users as well as theresearchers and the researched themselves.One way to plot a route through such com-plexity is to seek guidance from broader setsof moral and ethical considerations. This iswhat we will turn to next in the final sectionof the chapter.

Ethics, morality andgeographical research

By embracing the larger concerns of ethicsand morality we can start to open a windowon a number of other questions which areimplicated in the process of geographicalresearch. The questions involve whom andwhat we choose to research; how we conductthat research and how we relate to our

research ‘subjects’; how we use interpretativeand authorial power in constituting ourresearch findings; and what use we make, andallow others to make, of these findings(Cloke, 2002; 2004b). To a certain extentthese questions have been implicit in much ofwhat we have written throughout this book.Engaging with issues of ethics and moralitymakes them explicit.

Philo (1991b) identified three main ways inwhich geographers might engage with theconcerns of moral philosophy. He referred tothe first as the geography of everyday morali-ties which are established through the‘different moral assumptions and supportingarguments that particular peoples in particularplaces make about “good” and “bad”/“right”and “wrong”/“just” and “unjust”/“worthy”and “unworthy”’ (1991b: 16).As Philo pointsout, there can be little argument that theseassumptions do vary considerably – from onenation to another, one community to anotherand one street to another. A second form ofengagement might come from the way in wewe investigate the geography in everydaymoralities, which refers to the ways in which‘moral assumptions and arguments often havebuilt into their very heart thinking aboutspace, place, environment, landscape, and (inshort) geography’ (Philo, 1991b: 16). Again,there is little doubt about the ways in whichparticular behaviours and practices are viewedas suitable for particular spaces and places.

It is, however, the third form of engagementbetween geography and morality which wewant to concentrate on in the remainder ofthe chapter, for this refers to the morality ofgeographical practices.This engagement arisesthrough geographers reflecting ‘upon theirown morality, their own sense of “good” and“bad”/“right” and “wrong” and so on thatpermeates and maybe informs their own practiceas academic geographers’ (Philo, 1991b: 17) –including the practice of research. Such reflec-tion may often be uncomfortable but, in many


374

3151-Ch-12.qxd 3/2/04 2:50 PM Page 374

senses, geographers are ideally placed toundertake it for our whole discipline isfounded on investigations of social and spatialinequality. If people and places were equal,geography would lose its rationale: it is builtaround uncovering and exploring division,difference and uneven development.

Yet the notion of a ‘moral human geogra-phy’ would surely have to do more thanuncover and explore these differences. AsHarvey remarked some 30 years ago, our taskshould not entail ‘mapping even more evi-dence of man’s patent inhumanity to man’(1973: 144). It would instead have to involvea commitment to helping to secure a moreequal and just society. In the words of DavidSmith, ‘it is in seeking the new institutionsand social arrangements required to equalisehuman life prospects, whoever and whereverpeople are, that the creation of a moral geog-raphy can be envisaged’ (2004: 207).There aremany places where geographers can con-tribute; their work, after all, embraces a widerange of relevant issues, including the distrib-ution of employment and wealth; the locationand provision of welfare services; theexploitation and conservation of naturalresources; the emergence of new cultural andpolitical identities; and the development ofinequality by gender, race, sexuality, disabilityor age.We could go on to produce a score ofrelevant examples where geographical work iscentral to issues of equality and justice.All willoffer the human geographer the chance ofexplicitly engaging with the ‘articulation ofthe moral and the spatial’ (Philo, 1991b: 26).

However, as David Smith (2004) goes on topoint out, a project with such huge implica-tions is conceivable only if driven by the mostpowerful moral argument. As individuals,geographers will bring their own moral

understandings to such a task – and these willbe variously informed by a range of personalcommitments. One possible way to unitethese various understandings through thepractice of research is to draw on the work ofthe French anthropologist Marc Augé. Augéargues that two strands of sensitivity arerequired in relation to others (to those we doresearch ‘on’): a sense of the other and a sensefor the other (Cloke, 2002; 2004b).We wouldargue that in recent years geographers haveincreasing shown a sensitivity towards othervoices as their research has broadened to gain asense of what has meaning for others.However, Augé argues that our researchshould show a sense for the other, and forotherness, in the same way that we develop asense of direction, of rhythm, of belonging.Wewould argue that this has been much less evi-dent in geography, making it much more diffi-cult to engage with others in an emotional,committed and connected manner. If we canunite a sense of the other with a sense for theother it becomes possible to do research withpeople and for people rather than just onpeople.And if we can make these connectionsit becomes more likely that we will be able toredirect our future research in order to engagewith a truly moral human geography whichseeks to contribute to a more just and equalsociety.

Notes

1 Bell and Roberts (1984), Punch (1986) andHammersley (1993; 1995).

2 See Johnston (2000c: 982) for a list ofmembers.

3 On the link between geographical researchand public policy, see Peck (1999), Massey(2000) and Martin (2001).


375

3151-Ch-12.qxd 3/2/04 2:50 PM Page 375

Abler, R., Adams, J.S. and Gould, P. (1971) Spatial Organization: The Geographer’s View of the World.Englewood Cliffs, NJ: Prentice Hall.

Ackerman, E.A. (1965) The Science of Geography. Washington, DC: National Academy of Sciences, NationalResearch Council.

Agger, B. (1989a) ‘Do books write authors? A study of disciplinary hegemony’, Teaching Sociology,17: 365–9.

Agger, B. (1989b) Reading Science: A Literary, Political and Sociological Analysis. Dix Hills, NY: General Hall.Agnew, J. (1995) ‘Classics in human geography revisited: Thrift, N.“On the determination of social action

in space and time”: commentary 1’, Progress in Human Geography, 19: 525–6.Aitken, S.C. (1994) ‘I’d rather watch the movie than read the book’, Journal of Geography in Higher Education,

18: 291–307.Aitken, S.C. (1997) ‘Analysis of texts: armchair theory and couch-potato geography’, in R. Flowderdew and

D. Martin (eds) Methods in Human Geography: A Guide for Students Doing a Research Project. Harlow:Longman, 197–212.

Aitken, S.C. and Zonn,L.E. (1993) ‘Weir(d) sex: representation of gender–environment relations in Peter Weir’s“Picnic at Hanging Rock” and “Gallipoli”’, Environment and Planning D: Society and Space, 11: 191–212.

Aitken, S.C. and Zonn, L.E. (eds) (1994a) Place, Power, Situation and Spectacle: A Geography of Film. Lanham,MD: Rowman & Littlefield.

Aitken, S.C. and Zonn, L.E. (1994b) ‘Re-presenting the place pastiche’, in S.C.Aitken and L.E. Zonn (eds)Place, Power, Situation and Spectacle: A Geography of Film. Lanham, MD: Rowman & Littlefield, 3–25.

Alaasutari, P. (1995) Researching Culture: Qualitative Methods and Cultural Studies. London: Sage.Allen, J., Massey, D. and Cochrane,A. (1998) Rethinking the Region. London: Routledge.Amaral, D.J. and Wisner, W.B. (1970) ‘Participant observation, phenomenology and the rules for judging

sciences: a comment’, Antipode, 2: 42–51.Amedeo, D. and Golledge, R. (1975) An Introduction to Scientific Reasoning in Geography. Chichester: Wiley.Amin, A. and Robins, K. (1990) ‘The re-emergence of regional economies? The mythical geography of

flexible accumulation’, Environment and Planning D: Society and Space, 8: 7–34.Anderson, J. (2002) ‘Researching environmental resistance: working through secondspace and thirdspace

approaches’, Qualitative Research, 2: 301–21.Anderson, J.R. (1961) ‘Towards more effective methods of obtaining land use data in geographic research’,

The Professional Geographer, 13: 15–18.Anderson, K. (1999) ‘Introduction’, in K. Anderson and F. Gale (eds) Cultural Geographies. Harlow:

Longman, 1–21.Anderson, K. and Gale, F. (eds) (1992) Inventing Places: Studies in Cultural Geography. Melbourne: Longman

Cheshire.Anderson, K. and Smith, S. (2001) ‘Editorial: emotional geographies’, Transactions, Institute of British

Geographers, 26: 7–10.Angus, T., Cook, I., Evans, J. et al. (2001) ‘A manifesto for cyborg pedagogy?’, International Research in

Geographical and Environmental Education, 10: 195–201.

References

3151-References.qxd 3/2/04 2:52 PM Page 376

Anon (2001) ‘Kinaesthesia’, in Collins Concise Dictionary. Glasgow: HarperCollins, 809.Archer, J.E. and Dalton, T.H. (1968) Fieldwork in Geography. London: Batsford.Asad, T. (ed.) (1973) Anthropology and the Colonial Encounter. London: Ithaca Press.

Bailey, C.,White, C. and Pain, R. (1999) ‘Evaluating qualitative research: dealing with the tension between“science” and “creativity”’, Area, 31: 169–83.

Bailey, T. and Gatrell,A. (1995) Interactive Spatial Data Analysis. Harlow: Longman.Baker, A.R.H., Hamshere, J. and Langton, J. (eds) (1970) Geographical Interpretations of Historical Sources.

Newton Abbot: David & Charles.Bakhtin, M. (1986) The Dialogical Imagination.Austin,TX: University of Texas Press.Barnes,T. (1992) ‘Reading the texts of theoretical economic geography’, in T. Barnes and J. Duncan (eds)

Writing Worlds. London: Routledge, 118–35.Barnes, T. (1994) ‘Probable writing: Derrida, deconstruction and the quantitative revolution in human

geography’, Environment and Planning A, 26: 1021–40.Barnes,T. (1996) Logics of Dislocation: Metaphors, Models and Meanings of Economic Space.New York,NY:Guilford.Barnes, T. (1998) ‘A history of regression: actors, networks, machines and numbers’, Environment and

Planning A, 30: 225–34.Barnes,T. (2000a) ‘Poetics of geography’, in R.J. Johnston et al. (eds) The Dictionary of Human Geography

(4th edn). Oxford: Blackwell, 588–9.Barnes,T. (2000b) ‘Metaphor’, in R.J. Johnston et al. (eds) The Dictionary of Human Geography (4th edn).

Oxford: Blackwell, 499–501.Barnes,T. (2000c) ‘Hermeneutics’, in R.J. Johnston et al. (eds) The Dictionary of Human Geography (4th edn).

Oxford: Blackwell, 334–6.Barnes,T. (2001a) ‘Lives lived and lives told: biographies of geography’s quantitative revolution’, Environment

and Planning D: Society and Space, 19: 409–30.Barnes, T. (2001b) ‘Retheorising economic geography: from the quantitative revolution to the “cultural

turn”’, Annals of the Association of American Geographers, 91: 546–65.Barnes,T. and Curry, M. (1992) ‘Postmodernism in economic geography: metaphor and the construction

of alterity’, Environment and Planning D: Society and Space, 10: 57–68.Barnes,T. and Duncan, J. (eds) (1992) Writing Worlds. London: Routledge.Barnes,T. and Gregory, D. (eds) (1997) Reading Human Geography. London:Arnold.Barnett, C. (1997) ‘“Sing along with the common people”: politics, postcolonialism and other figures’,

Environment and Planning D: Society and Space, 15: 137–54.Barrell, J. (1982) ‘Geographies of Hardy’s Wessex’, Journal of Historical Geography, 8: 347–61.Batterbury, S. (1994) ‘Alternative affiliations and the personal politics of overseas research: some reflections’,

in E. Robson and K. Willis (eds) Postgraduate Fieldwork in Developing Areas: A Rough Guide. IBGDeveloping Areas Research Group Monograph 8. London: IBG, 60–89.

Batty, M., Dodge, M., Doyle, S. and Smith,A. (1998) ‘Modelling virtual environments’, in P. Longley et al.(eds) Geocomputation: A Primer. Chichester:Wiley, 139–61.

Batty, M. and Longley, P. (1994) Fractal Cities. London:Academic Press.Batty, M., Longley, P. and Fotheringham, S. (1989) ‘Urban growth and form: scaling fractal geometry and

diffusion-limited aggregation’, Environment and Planning A, 21: 1447–72.Baxter, J. and Eyles, J. (1997) ‘Evaluating qualitative research in social geography: establishing rigour in

interview analysis’, Transactions, Institute of British Geographers, 22: 505–25.Bayliss-Smith,T. and Owens, S. (1990) ‘Introduction: aerial photography, the hidden agenda’, in T. Bayliss-

Smith and S. Owens (eds) Britain’s Changing Environment from the Air. Cambridge: Cambridge UniversityPress, 7–14.

Beauregard, R. (1993) Voices of Decline. Blackwell: Oxford.Beaver, S.H. (1962) ‘The Le Play Society and field work’, Geography, XLVII: 225–40.Becker, H. (1986) Writing for Social Scientists. Chicago, IL: University of Chicago Press.Behar, R. (1995) ‘Introduction: out of exile’, in R. Behar and D. Gordon (eds) Women Writing Culture.

Berkeley, CA: University of California Press, 1–29.

REFERENCES

377


Behar, R. and Gordon, D. (eds) (1995) Women Writing Culture. Berkeley, CA: University of California Press.Bell,C. and Roberts,H. (1984) Social Researching: Politics, Problems, Practice. London:Routledge & Kegan Paul.Bell, J. (1987) Doing your Research Project. Milton Keynes: Open University Press.Bell, M. (1993) ‘“The pestilence that walketh in darkness.” Imperial health, gender and images of South

Africa c. 1880–1910’, Transactions, Institute of British Geographers, 18: 327–41.Benedikt, M. (ed.) (1991) Cyberspace: First Steps. Cambridge, MA: MIT Press.Benko, G. and Strohmayer, U. (eds) (1997) Space and Social Theory: Interpreting Modernity and Postmodernity.

Oxford: Blackwell.Bennett, K. (2002) ‘Participant observation’, in P. Shurmer-Smith (ed.) Doing Cultural Geography. London:

Sage, 139–49.Bennett, K., Ekinsmyth, C. and Shurmer-Smith, P. (2002) ‘Selecting topics for study’, in P. Shurmer-Smith

(ed.) Doing Cultural Geography. London: Sage, 81–93.Bennett, K. and Shurmer-Smith, P. (2001) ‘Writing conversation’, in M. Limb and C. Dwyer (eds)

Qualitative Methodologies for Geographers. London:Arnold, 251–63.Bennett, K. and Shurmer-Smith, P. (2002) ‘Handling case studies’, in P. Shurmer-Smith (ed.) Doing Cultural

Geography. London: Sage, 199–209.Bennett, R. (1979) Spatial Time Series. London: Pion.Bennett, R. and Graham, D.J. (1998) ‘Explaining size differentiation of business service centres’, Urban

Studies, 35: 1457–80.Bennett,R.J. and Chorley,R.J. (1978) Environmental Systems: Philosophy, Analysis and Control. London:Methuen.Benton, L. (1995) ‘Will the real/reel Los Angeles please stand up?’, Urban Geography, 16: 144–64.Berg, B. (1989) Qualitative Research Methods for the Social Sciences (3rd edn). London:Allyn & Bacon.Berger, J. (1972) Ways of Seeing. London: Penguin Books.Berleant-Schiller, R. (1991) ‘Hidden places and Creole forms: naming the Barbundan landscape’, The

Professional Geographer, 43: 92–101.Berry, B.J.L. (1964) ‘Approaches to regional analysis: a synthesis’, Annals of the Association of American

Geographers, 54: 2–11.Bertalanffy, L. van (1968) General Systems Theory: Foundation, Development, Applications. London:Allen Lane.Billinge, M.D. (1983) ‘The Mandarin dialect: an essay on style in contemporary geographical writing’,

Transactions, Institute of British Geographers, 8: 400–20.Billinge, M.D., Gregory, D. and Martin, R. (1984) Recollections of a Revolution: Geography as Spatial Science.

New York, NY: St Martin’s Press.Bingham, N. and Thrift, N. (1999) ‘Some new instructions for travellers: the geography of Bruno Latour

and Michel Serres’, in M. Crang and N.Thrift (eds) Thinking Space. London: Routledge, 281–301.Bingley, A. (2003) ‘In here and out there: sensations between self and landscape’, Social and Cultural

Geography, 4: 329–46.Blake, L.A. (1999) ‘Pastoral power, governmentality and cultures of order in nineteenth-century British

Columbia’, Transactions, Institute of British Geographers, 24: 79–93.Blakemore, M.J. (1994) ‘Cartography’, in R.J. Johnston et al. (eds) The Dictionary of Human Geography

(3rd edn). Oxford: Blackwell, 44–8.Blaut, J. (1961) ‘Space and process’, The Professional Geographer, 13: 1–7.Blomley, N. (1994) ‘Activism and the academy’, Environment and Planning D: Society and Space, 12: 383–5.Blunt, A. (1994a) ‘Mapping authorship and authority: reading Mary Kingsley’s landscape descriptions’, in

A. Blunt and G. Rose (eds) Writing Women and Space: Colonial and Post-colonial Geographies. New York,NY: Guilford, 51–72.

Blunt,A. (1994b) Travel, Gender and Imperialism: Mary Kingsley and West Africa. New York, NY: Guilford.Blunt, A., Gruffudd, P., May, J., Ogborn, M. and Pinder, D. (eds) (2003) Cultural Geography in Practice.

London:Arnold.Bondi, L. (1999) ‘Stages on a journey: some remarks about human geography and psychotherapeutic prac-

tice’, The Professional Geographer, 51: 11–24.Bondi, L. (2003a) ‘A situated practice for (re)situating selves: trainee counsellors and the promise of coun-

selling’, Environment and Planning A, 35: 853–70.


378


Bondi, L. (2003b) ‘“Unlocking the cage door”: counselling, care and space’, Journal of Social and CulturalGeography, forthcoming.

Bonnett,A. (1992) ‘Art, ideology, and everyday space: subversive tendencies from Dada to postmodernism’,Environment and Planning D: Society and Space, 10: 69–86.

Boots, B. and Getis,A. (1988) Point Pattern Analysis. Harlow: Sage.Bowen, M. (1981) Empiricism and Geographical Thought: From Francis Bacon to Alexander von Humboldt.

Cambridge: Cambridge University Press.Brandth, B. (1995) ‘Rural masculinity in transition: gender image in tractor advertisements’, Journal of Rural

Studies, 11: 123–34.Brewer, J. (1989) The Sinews of Power: War, Money and the English State, 1688–1783. London: Unwin

Hyman.Brice, A.M.H. and Bain, J.A. (n.d., c. 1918) Two Great Explorers: Sir Henry M. Stanley and Fridtjof Nansen.

London: Partridge & Co.Bridge, G. (1992) ‘Questionnaire surveys’, in A. Rogers et al. (eds) The Student’s Companion to Geography.

Blackwell: Oxford, 196–206.Bristow, J. (1991) ‘Life stories: Carolyn Steedman’s history writing’, New Formations, 13: 113–31.Brown, L., Flavin, C. and Kane, H. (1992) Vital Signs 1992: The Trends that are Shaping our Future. London:

Earthscan.Bunge,W. (1962) Theoretical Geography. Lund: Gleerup.Bunge,W. (1971) Fitzgerald: The Geography of a Revolution. Cambridge, MA: Schenkman.Bunge,W. (1975) ‘The point of production: a second front’, Antipode, 9: 60–76.Bunge,W. (1988) Nuclear War Atlas. Blackwell: Oxford.Burgess, J. (1985) ‘News from nowhere: the press, the riots and the myth of the inner city’, in J. Burgess and

J. Gold (eds) Geography, the Media and Popular Culture. London: Croom Helm, 192–228.Burgess, J. (1987) ‘Landscapes in the living-room: television and landscape research’, Landscape Research, 12: 1–7.Burgess, J. (1990) ‘The production and consumption of environmental meanings in the mass media: a

research agenda for the 1990s’, Transactions, Institute of British Geographers, 15: 139–61.Burgess, J. (1996) ‘Focusing on fear: the use of focus groups in a project for the Community Forest Unit,

Countryside Commission’, Area, 28: 130–55.Burgess, J. and Gold, J. (eds) (1986) Geography, the Media and Popular Culture. London: Croom Helm.Burgess, J., Harrison, C. and Maiteny, P. (1991) ‘Contested meanings: the consumption of news about nature

conservation’, Media, Culture and Society, 13: 499–520.Burgess, J., Limb,M. and Harrison, C. (1988) ‘Exploring environmental values through the medium of small

groups. 2. Illustrations of a group at work’, Environment and Planning A, 20: 457–76.Burgess, J. and Pile, S. (1991) ‘Constructions of the self: some thoughts on the need for a psychoanalytic

perspective’, in C. Philo (compiler) New Words, New Worlds: Reconceptualising Social and CulturalGeography. Lampeter: Social and Cultural Geography Study Group of the Institute of BritishGeographers, 31–2.

Burgess, R. (1984) In the Field: An Introduction to Field Research. London:Allen & Unwin.Burgess, R. (1986) ‘Being a participant observer’, in J. Eyles (ed.) Qualitative Approaches in Social and

Geographical Research. Occasional Paper 26. London: Department of Geography and Earth Science, QueenMary College, University of London, 47–66.

Burke, P. (1988) ‘Verstehen’, in A. Bullock et al. (eds) The Fontana Dictionary of Modern Thought. London:Fontana, 894.

Burrough, P. (1998) ‘Dynamic modelling and geocomputation’, in P. Longley et al. (eds) Geocomputation:A Primer. Chichester:Wiley, 165–92.

Burton, F. and Carlen, P. (1979) Official Discourse. London: Routledge.Burton, I. (1963) ‘The quantitative revolution and theoretical geography’, The Canadian Geographer,

7: 151–62.Butler, R. (2001) ‘From where I write: the place of positionality in qualitative writing’, in M. Limb and

C. Dwyer (eds) Qualitative Methodologies for Geographers: Issues and Debates. London: Edward Arnold,264–76.

REFERENCES

379


Buttimer, A. (1974) Values in Geography. Commission on College Geography Resource Paper 24. Washington,DC:Association of American Geographers.

Buttimer,A. (1976) ‘Grasping the dynamism of lifeworld’, Annals of the Association of American Geographers,66: 277–92.

Buttimer,A. (1978) ‘Charisma and context: the challenge of la géographie humaine’, in D. Ley and M. Samuels(eds) Humanistic Geography: Prospects and Problems. London: Croom Helm, 58–76.

Buttimer, A. (2001) ‘Home-reach-journey’, in P. Moss (ed.) Placing Autobiography in Geography. Syracuse,NY: Syracuse University Press, 22–40.

Buttimer,A. and Seamon, D. (eds) (1980) The Human Experience of Space and Place. London: Croom Helm.Bygott, J. (1934) Mapwork and Practical Geography. London: University Tutorial Press.Bynner, J. and Stribley, K. (1978) Social Research: Principles, Procedures. London: Longman.Byron, M. (1993) ‘Using audiovisual aids in geography research: questions of access and responsibility’, Area,

25: 379–85.

Callard, F. (2003) ‘The taming of psychoanalysis in geography’, Social and Cultural Geography, 4: 295–312.Callaway, H. (1992) ‘Ethnography and experience: gender implications in fieldwork and texts’, in J. Okely

and H. Callaway (eds) Anthropology and Autobiography. London: Routledge, 29–49.Campbell, D. (1974) ‘Role relationships in advocacy geography’, Antipode, 6: 102–5.Capel, H. (1981) ‘Institutionalisation of geography and strategies of change’, in D.R. Stoddart (ed.)

Geography, Ideology and Social Concern. Oxford: Blackwell, 37–69.Capps, P. (2001) ‘The retail workplace of gender, space and power dynamics: a case study of a camping busi-

ness in south-east England’, in M. Limb and C. Dwyer (eds) Qualitative Methodologies for Geographers:Issues and Debates. London:Arnold, 291–3.

Carr, M. (1987) Patterns: Process and Change in Human Geography. London: Macmillan.Carter, H. (1990) ‘Towns and urban systems 1730–1914’, in R. Dodgshon and R. Butlin (eds) An Historical

Geography of England and Wales. London:Academic Press, 401–51.Carter, P. (1999) ‘Dark with excess of bright: charting the coastlines of knowledge’, in D. Cosgrove (ed.)

Mappings. London: Reaktion Books, 125–47.Cassidy, F.G. (1947) The Place-names of Dane County, Wisconsin. Publications of the American Dialect Society 7.

Greensboro, NC:American Dialect Society.Castells, M. (1989) The Informational City. Oxford: Oxford University Press.Castree, N. (1996) ‘Birds, mice and geography: Marxisms and dialectics’, Transactions, Institute of British

Geographers, 21: 342–62.Chamberlain, P.G. (1995) ‘Metaphorical vision in the literary landscape of William Shakespeare’, The

Canadian Geographer/Le Géographie canadien, 39: 306–22.Chandler, D. (2001) Semiotics: The Basics. London: Routledge.Chapman, G.P. (1977) Human and Environmental Systems. London:Academic Press.Chisholm, M. (1975) Human Geography: Evolution or Revolution. Harmondsworth: Penguin Books.Chorley, R.J. (1964) ‘Geography and analogue theory’, Annals of the Association of American Geographers,

54: 127–37.Chouinard,V. (1994) ‘Radical geography: is all that’s Left right?’, Environment and Planning D: Society and

Space, 12: 2–5.Chrisman, N. (1997) Exploring Geographic Information Systems. New York, NY: Wiley.Clanchy, M.T. (1979) From Memory to Written Record: England 1066–1307. London: Edward Arnold.Clark, M.J. (1998) ‘GIS – democracy or delusion?’, Environment and Planning A, 30: 303–16.Clarke, D. (1984) ‘Film/space’, Environment and Planning A, 26: 1821–3.Clarke, D. (ed.) (1997) The Cinematic City. London: Routledge.Clarke, G. (1999) ‘Geodemographics, marketing and retail location’, in M. Pacione (ed.) Applied Geography:

Principles and Practice. London: Routledge, 577–92.Clarke, G. and Cooke, D. (1983) A Basic Course in Statistics (2nd edn). London: Edward Arnold.Clarke, K. (1998) ‘Visualising different geofutures’, in P. Longley et al. (eds) Geocomputation: A Primer.

Chichester:Wiley, 119–37.


380


Cliff,A. and Ord, J. (1973) Spatial Autocorrelation. London: Pion.Clifford, J. (1986) ‘Introduction: partial truths’, in J. Clifford et al. (eds) Writing Culture: The Poetics and

Politics of Ethnography. Berkeley, CA: University of California Press, 1–26.Clifford, J. (1990) ‘Notes on (field)notes’, in R. Sanjek (ed.) Fieldnotes. Ithaca, NY: Cornell University Press,

47–70.Clifford, J. (1992) ‘Travelling cultures’, in L.Grossberg et al. (eds) Cultural Studies. London:Routledge, 96–112.Clifford, J. (1997) Routes: Travel and Translation in the Late Twentieth Century. London: Harvard University

Press.Clifford, J. and Marcus, G. (eds) (1986) Writing Culture: The Poetics and Politics of Ethnography. Berkeley, CA:

University of California Press.Cloke, P. (1994) ‘(En)culturing political economy: a day in the life of a “rural geographer”’, in P. Cloke

et al. (eds) Writing the Rural: Five Cultural Geographies. London: Paul Chapman, 149–90.Cloke, P. (1996) ‘Rural lifestyles: material opportunity, cultural experience, and how theory can undermine

policy’, Economic Geography, 72: 433–49.Cloke, P. (1997) ‘Country backwater to virtual village? Rural studies and “the cultural turn”’, Journal of

Rural Studies, 13: 367–75.Cloke, P. (2002) ‘Deliver us from evil: prospects for living ethically and acting politically in human geo-

graphy’, Progress in Human Geography, 26: 587–604.Cloke, P. (2004a) ‘On vision and envisioning…’, in P. Cloke et al. (eds) Envisioning Human Geographies.

London:Arnold, 1–10.Cloke, P. (2004b) ‘Spiritual geographies’, in P. Cloke et al. (eds) Envisioning Human Geographies. London:

Arnold, 228.Cloke, P., Cooke, P., Cursons, J., Milbourne, P. and Widdowfield, R. (2000a) ‘Ethics, reflexivity and research:

encounters with homeless people’, Ethics, Place and Environment, 3: 133–54.Cloke, P., Crang, P. and Goodwin, M. (1999a) Introducing Human Geography. London:Arnold.Cloke, P., Goodwin, M., Milbourne, P. and Thomas, C. (1995b) ‘Deprivation, poverty and marginalisation

in rural lifestyles in England and Wales’, Journal of Rural Studies, 11: 351–65.Cloke, P., Milbourne, P. and Thomas, C. (1994) Lifestyles in Rural England. London: Rural Development

Commission.Cloke, P., Milbourne, P. and Thomas, C. (1995a) ‘Poverty in the countryside: out of sight and out of mind’,

in C. Philo (ed.) Off the Map: A Social Geography of Poverty. London: Child Poverty Action Group,83–102.

Cloke, P., Milbourne, P. and Widdowfield, R. (1999b) ‘Homelessness in rural areas: an invisible issue’, inP. Kennett and A. Marsh (eds) Homelessness: Exploring the New Terrain. Bristol: Policy Press, 61–80.

Cloke, P., Milbourne, P. and Widdowfield, R. (2000b) ‘Homelessness and rurality: out of place in purifiedspace’, Environment and Planning D: Society and Space, 18: 715–35.

Cloke, P. and Perkins, H. (1998) ‘“Cracking the Canyon with the Awesome Foursome”: representations ofadventure tourism in New Zealand’, Environment and Planning D: Society and Space, 16: 185–218.

Cloke, P., Philo, C. and Sadler, D. (1991) Approaching Human Geography: An Introduction to ContemporaryTheoretical Debates. London: Paul Chapman.

Close, C.F. (1911) ‘The purpose and position of geography’, Bulletin of the American Geographical Society,43: 740–53.

Coan, T.M. (1899) ‘Hawaiian ethnography’, Journal of the American Geographical Society of New York,31: 24–30.

Coffey,W.J. (1981) Geography: Towards a General Spatial Systems Approach. London: Methuen.Cohen, S. (1995) ‘Sounding out the city: music and the sensuous reproduction of place’, Transactions,

Institute of British Geographers, 20: 434–46.Cole, J. and King, C. (1968) Quantitative Geography. London:Wiley.Coleman,A. (1954) ‘Fieldwork in schools: a sample traverse in east Kent’, Geography, XXXIV: 262–71.Coleman, A. and Lukehurst, C.T. (eds) (1974) Field Excursions in South East England. Field Studies for

Schools 11. London: Rivingtons.Colenutt, R. (1971) ‘Postscript on the Detroit Geographical Expedition’, Antipode, 6: 85.

REFERENCES

381


Cook, I. (1997) ‘Participant observation’, in R. Flowerdew and D. Martin (eds) Methods in HumanGeography: A Guide for Students Doing Research Projects. Harlow: Longman, 127–49.

Cook, I. (2001) ‘“You want to be careful you don’t end up like Ian, he’s all over the place”: autobiographyin/of an expanded field’, in P. Moss (ed.) Placing Autobiography in Geography. Syracuse, NY: SyracuseUniversity Press, 99–120.

Cook, I. et al. (2000a) ‘Social sculpture and connective aesthetics: Shelley Sacks’ “exchange values”’,Ecumene, 7: 338–44.

Cook, I. et al. (2002) Commodities: The DNA of Capitalism (http://apm.brookes.ac.uk/exchange/texts/IanCook2.pdf) accessed 23 June 2003.

Cook, I. et al. (forthcoming) ‘Follow the thing: papaya’, Antipode.Cook, I. et al. (in press) ‘Positionality/situated knowledge’, in D. Atkinson et al. (eds) Cultural Geography:

A Critical Dictionary of Key Ideas. London: I.B.Tauris.Cook, I. and Crang, M. (1995) Doing Ethnographies. Norwich: Environmental Publications.Cook, I. and Crang, P. (1996) ‘Commodity systems, documentary filmmaking and new geographies of food:

Amos Gitai’s Ananas.’ Paper presented at the 1996 annual conference of the Institute of BritishGeographers/Royal Geographical Society conference, Glasgow, January (http://www.ges.bham.ac.uk/draftpapers/iancook-pineapple.htm) accessed 23 June 2003.

Cook, I., Crang, P. and Thorpe, M. (1999) ‘Eating into Britishness: multicultural imaginaries and the iden-tity politics of food’, in S. Roseneil and J. Seymour (eds) Practising Identities: Power and Resistance. London:Macmillan, 223–48.

Cook, I., Crouch, D., Naylor, S. and Ryan, J. (eds) (2000b) Cultural Turns/Geographical Turns: Perspectives onCultural Geography. Harlow: Longman.

Cook, I. and Harrison, M. (2003) ‘Cross over food: re-materialising postcolonial geographies’, Transactions,Institute of British Geographers, 28: 296–317.

Cope, M. (2003) ‘Coding transcripts and diaries’, in N.J. Clifford and G.Valentine (eds) Key Methods inGeography. London: Sage, 445–59.

Corrigan, P. and Sayer, D. (1985) The Great Arch: English State Formation as Cultural Revolution. Oxford:Blackwell.

Cosgrove, D. (1984) Social Formation and Symbolic Landscape. London: Croom Helm.Cosgrove, D. (1993) The Palladian Landscape. Leicester and London: Leicester University Press.Cosgrove, D. (ed.) (1999) Mappings. London: Reaktion Books.Cosgrove, D. and Daniels, S. (eds) (1988) The Iconography of Landscape: Essays on the Symbolic Representation,

Design and Use of Past Environments. Cambridge: Cambridge University Press.Cosgrove, D. and Domosh, M. (1993) ‘Author and authority: writing the new cultural geography’, in

J. Duncan and D. Ley (eds) Place/Culture/Representation. London: Routledge, 25–38.Cosgrove, D. and Jackson, P. (1987) ‘New directions in cultural geography’, Area, 19: 95–101.Couclelis, H. (1998) ‘Geocomputation in context’, in P. Longley et al. (eds) Geocomputation: A Primer.

Chichester:Wiley, 17–29.Crang, M. (1995) Doing Ethnographies (CATMOG 58). Norwich: Environmental Publications.Crang, M. (1996) ‘Watching the city: video, surveillance, and resistance’, Environment and Planning A,

28: 2099–104.Crang, M. (1997) ‘Analysing qualitative materials’, in R. Flowerdew and D. Martin (eds) Methods in Human

Geography: A Guide for Students Doing a Research Project. Harlow: Longman, 183–96.Crang, M. (2000) ‘Playing nymphs and swains in pastoral idyll’, in A. Hughes et al. (eds) Ethnography and

Rural Research. Cheltenham: Countryside and Community Press, 158–78.Crang, M. (2001) ‘Filed work: making sense of group interviews’, in M. Limb and C. Dwyer (eds)

Qualitative Methodologies for Geographers: Issues and Debates. London: Edward Arnold, 215–33.Crang, M. (2002) ‘Qualitative methods: the new orthodoxy?’, Progress in Human Geography, 26: 647–55.Crang, M., Crang, P. and May, J. (eds) (1999) Virtual Geographies. London: Routledge.Crang, M., Hudson,A., Reimer, S. and Hinchliffe, S. (1997) ‘Software for qualitative research. 1. Prospectus

and overview’, Environment and Planning A, 29: 771–87.Crang, P. (1992) ‘The politics of polyphony: reconfigurations of geographical authority’, Environment and

Planning D: Society and Space, 10: 527–49.


382


Crang, P. (1994) ‘It’s showtime: on the workplace geographies of display in a restaurant in southeastEngland’, Environment and Planning D: Society and Space, 12: 675–704.

Crang, P. (2000) ‘Cultural turn’, in R.J. Johnston et al. (eds) The Dictionary of Human Geography (4th edn).Oxford: Blackwell, 141–3.

Crapanzano,V. (1977) ‘On the writing of ethnography’, Dialectical Anthropology, 2: 69–73.Crapanzano, V. (1986) ‘Hermes’ dilemma: the masking of subversion in ethnographic description’, in

J. Clifford and G. Marcus (eds) Writing Culture: The Poetics and Politics of Ethnography. Berkeley, CA:University of California Press, 51–76.

Cresswell,T. (1993) ‘Mobility as resistance: a geographical reading of Kerouac’s On the Road ’, Transactions,Institute of British Geographers, 18: 249–62.

Cresswell,T. (1996) In Place/Out of Place: Geography, Ideology and Transgression. Minneapolis, MN: Universityof Minnesota Press.

Cresswell,T. (2000) ‘Falling down: resistance as diagnostic’, in J.P. Sharp et al. (eds) Entanglements of Power:Geographies of Resistance/Domination. London: Routledge, 256–68.

Cresswell,T. (2001) The Tramp in America. London: Reaktion Books.Crewe, L. and Goodrum, A. (2000) ‘Fashioning new forms of consumption: the case of Paul Smith’, in

S. Bruzzi and P. Church Gibson (eds) Fashion Cultures: Theories, Explorations and Analysis. London:Routledge, 25–48.

Crush, J. (1994) ‘Post-colonialism, de-colonisation and geography’, in A. Godlewska and N. Smith (eds)Geography and Empire. Oxford: Blackwell, 333–50.

Cullen, M.J. (1975) The Statistical Movement in Early Victorian Britain. Hassocks: Harvester.Curry, M. (1995) ‘GIS and the inevitability of ethical inconsistency’, in J. Pickles (ed.) Ground Truth: The

Social Implications of Geographic Information Systems. New York, NY: Guilford, 68–87.Curry, M.R. (1998) Digital Places: Living with Geographic Information Technologies. London: Routledge.Czaja, R. and Blair, J. (1996) Designing Surveys: A Guide to Decisions and Procedures.Thousand Oaks, CA:

Pine Forge Press.

Dandeker, C. (1990) Surveillance, Power and Modernity: Bureaucracy and Discipline from 1700 to the Present Day.Cambridge: Polity Press.

Daniels, S. (1985) ‘Arguments for a humanistic geography’, in R.J. Johnston (ed.) The Future of Geography.London: Methuen, 143–58.

Daniels, S. (1989) ‘Marxism, culture and the duplicity of landscape’, in R. Peet and N. Thrift (eds) NewModels in Geography. London: Unwin Hyman, 196–220.

Daniels, S. (1993) Fields of Vision. Cambridge: Cambridge University Press.Daniels, S. and Cosgrove, D. (1988) ‘Introduction: iconography and landscape’, in D. Cosgrove and

S. Daniels (eds) The Iconography of Landscape: Essays on the Symbolic Representation, Design and Use of PastEnvironments. Cambridge: Cambridge University Press, 1–10.

Daniels, S. and Rycroft, S. (1993) ‘Mapping the modern city: Alan Sillitoe’s Nottingham novels’,Transactions, Institute of British Geographers, 18: 460–80.

Dant, T. (1999) Material Culture in the Social World: Values, Activities, Lifestyles. Philadelphia, PA: OpenUniversity Press.

Darby, H.C. (1950–1) ‘Domesday woodland’, Economic History Review, 3: 21–43.Darby, H.C. (1951) ‘The economic geography of England A.D. 1000–1250’, in H. Darby (ed.) An Historical

Geography of England before A.D. 1800. Cambridge: Cambridge University Press, 165–229.Darby, H.C. (1962) ‘The problem of geographical description’, Transactions, Institute of British Geographers,

30: 1–14.Darby, H.C. (1977) Domesday England. Cambridge: Cambridge University Press.Darden, J. (1976) ‘Review of Ley, D. (1974) The Black Inner City as Frontier Outpost’, The Geographical Review,

66: 116–17.Davies, G. (1998) ‘Networks of nature: stories of natural history film-making from the BBC.’ Unpublished

PhD thesis, University College London.Davies, G. (1999) ‘Exploiting the archive: and the animals came in two by two, 16 mm, CD-ROM and

BetaSp’, Area, 31: 49–58.

REFERENCES

383


Davies, G. (2000) ‘Narrating the Natural History Unit: institutional orderings and spatial practices’,Geoforum, 31: 539–51.

Davies, G. (forthcoming) ‘Researching the networks of natural history television’, in M. Ogborn et al. (eds)Cultural Geography in Practice. London:Arnold.

Davis, C.M. (1954) ‘Field techniques’, in P.E. James and C.F. Jones (eds) American Geography: Inventory andProspect. Syracuse, NY: Syracuse University Press for the AAG, 496–529.

Davis, D.H. (1926) ‘Objectives in a geographic field study of a community’, Annals of the Association ofAmerican Geographers, 16: 102–9.

Davis, M. (1990) City of Quartz: Excavating the Future of Los Angeles. London:Verso.Dean, M. (1999) Governmentality: Power and Rule in Modern Society. London: Sage.Dear, M. (1986) ‘Postmodernism and planning’, Environment and Planning D: Society and Space, 4: 367–84.Dear, M. (1988) ‘The postmodern challenge: reconstructing human geography’, Transactions, Institute of

British Geographers, 13: 262–74.de Blij, H. (1982) ‘Review of Outcast Cape Town’, The Geographical Review, 72: 493–4.Deleuze, G. and Guattari, F. (1984) Anti-Oedipus: Capitalism and Schizophrenia. Volume 1. London:Athlone.Deleuze,G.and Guattari,F. (1988) A Thousand Plateaus: Capitalism and Schizophrenia. Volume 2.London:Athlone.DeLyser, D. (1999) ‘Authenticity on the ground: engaging the past in a California ghost town’, Annals of

the Association of American Geographers, 89: 602–32.DeLyser, D. and Starrs, P. (eds) (2001) Special issue on ‘Doing Fieldwork’, The Geographical Review, 91:

ix–viii, 1–508.Demeritt, D. (1996) ‘Social theory and the reconstruction of science and geography’, Transactions, Institute

of British Geographers, 14: 635–58.Demeritt, D. (2000) ‘The new social contract for science: accountability, relevance and value in US and UK

science and research policy’, Antipode, 32: 308–29.Demeritt, D. and Dyer, S. (2002) ‘Dialogue, metaphors of dialogue and understandings of geography’, Area,

34: 229–41.DeMers, M. (1997) Fundamentals of Geographic Information Systems. Oxford: Oxford University Press.Denzin, N. (1978) The Research Act (2nd edn). New York, NY: McGraw-Hill.Derrida, J. (1967) De la Grammatologie. Paris: Les Editions de Minuit. Translated as Of Grammatology.

Baltimore, MD: Johns Hopkins University Press.Derrida, J. (1991) The Derrida Reader: Between the Blinds. Hemel Hempstead: Harvester.Desforges, L. (2001) ‘Tourism consumption and the imagination of money’, Transactions, Institute of British

Geographers, 26: 353–64.de Souza,A. (1988) ‘Writing matters’, The Professional Geographer, 40: 1–3.Deutsche, R. (1991) ‘Boy’s town’, Environment and Planning D: Society and Space, 9: 5–30.De Vaus, D. (1991) Surveys in Social Research (3rd edn). London: UCL Press.Dewar, N. (1999) ‘Classics in human geography revisited:Western, J. (1981) Outcast Cape Town: commen-

tary 1’, Progress in Human Geography, 23: 421–2.Dewsbury, J.D. and Naylor, S. (2002) ‘Practising geographical knowledge: fields, bodies and dissemination’,

Area, 34: 253–60.Dey, I. (1993) Qualitative Data Analysis. London: Routledge.Dickens, P., Duncan, S., Goodwin, M. and Gray, F. (1985) Housing, States and Localities. London: Methuen.Dickinson, G.C. (1963) Statistical Mapping and the Presentation of Statistics. London: Edward Arnold.Dilke,M.S. (ed.) (1965) Field Excursions in North West England. Field Studies for Schools 2.London:Rivingtons.Dillman, D. (1978) Mail and Telephone Surveys: The Total Design Method. New York, NY:Wiley.Dixon, C. and Leach, B. (1978) Questionnaires and Interviews in Geographic Research. Concepts and Techniques

in Modern Geography 18. Norwich: Geobooks.Dixon, D.P. and Jones, J.-P. III (1998a) ‘My dinner with Derrida, or spatial analysis and poststructuralism do

lunch’, Environment and Planning A, 30: 247–60.Dixon, D.P. and Jones, J.-P. III (1998b) ‘For a supercalifragilisticexpialidocious scientific geography’, Annals of

the Association of American Geographers, 86: 767–79.Dodds, K. (1993) ‘War stories: British elite narratives of the 1992 Falklands/Malvinas war’, Environment and

Planning D: Society and Space, 12: 619–40.


384


Dodgshon, R. (1998) Society in Time and Space. Cambridge: Cambridge University Press.Doel, M.A. (1994) ‘Deconstruction on the move: from libidinal economy to liminal materialism’,

Environment and Planning A, 26: 1041–59.Doel, M.A. (1999) Poststructuralist Geographies: The Diabolical Art of Spatial Science. Edinburgh: Edinburgh

University Press.Doel, M.A. (2001) ‘1a. Qualified quantitative geography’, Environment and Planning D: Society and Space,

19: 555–72.Domosh, M. (1991a) ‘Toward a feminist historiography of geography’, Transactions, Institute of British

Geographers, 16: 95–104.Domosh, M. (1991b) ‘Beyond the frontiers of geographical knowledge’, Transactions, Institute of British

Geographers, 16: 488–90.Dorling, D. (1995a) ‘The visualisation of local urban change across Britain’, Environment and Planning

B: Planning and Design, 22: 269–90.Dorling, D. (1995b) A New Social Atlas of Britain. Chichester:Wiley.Dorling, D. (1998) ‘Human geography: when it is good to map’, Environment and Planning A, 30: 277–88.Dorling, D. and Fairbairn, D. (1997) Mapping: Ways of Representing the World. London: Longman.Dorling, D. and Simpson, S. (eds) (1999a) Statistics in Society. London:Arnold.Dorling, D. and Simpson, S. (1999b) ‘Introduction to statistics in society’, in D. Dorling and S. Simpson (eds)

Statistics in Society. London:Arnold, 1–6.Dowler, L. (2001) ‘Fieldwork in the trenches: participant observation in a conflict area’, in M. Limb and

C. Dwyer (eds) Qualitative Methodologies for Geographers: Issues and Debates. London:Arnold, 153–64.Downs, R. (1970) ‘Geographic space perception: past approaches and future prospects’, Progress in

Geography, 2: 65–108.Driver, F. (1988) ‘Moral geographies: social science and the urban environment in mid-nineteenth century

England’, Transactions, Institute of British Geographers, 13: 275–87.Driver, F. (1989) ‘The historical geography of the workhouse system, 1834–1883’, Journal of Historical

Geography, 15: 269–86.Driver, F. (1991) ‘Henry Morton Stanley and his critics: geography, exploration and empire’, Past and

Present, 133: 134–66.Driver, F. (1992) ‘Geography’s empire: histories of geographical knowledge’, Environment and Planning

D: Society and Space, 10: 23–40.Driver, F. (1993) Power and Pauperism: The Workhouse System, 1834–1884. Cambridge: Cambridge

University Press.Driver, F. (1998) ‘Hints to travellers: the Royal Geographical Society and the culture of exploration.’Typescript.Driver, F. (2000a) ‘Field-work in geography’, Transactions, Institute of British Geographers, 25: 267–8.Driver, F. (2000b) Geography Militant: Cultures of Exploration and Empire. Oxford: Blackwell.du Gay, P. (ed.) (1997) Production of Culture/Cultures of Production. London: Sage.Duncan, J.S. (1980) ‘The superorganic in American cultural geography’, Annals of the Association of American

Geographers, 70: 181–98.Duncan, J.S. (1998) ‘Classics in human geography revisited:“The superorganic in American cultural geog-

raphy”: author’s response’, Progress in Human Geography, 22: 571–3.Duncan, J.S. (2000) ‘Representation’, in R.J. Johnston et al. (eds) The Dictionary of Human Geography

(4th edn). Oxford: Blackwell, 703–5.Dwyer, C. and Limb, M. (2001) ‘Introduction: doing qualitative research in geography’, in M. Limb and

C. Dwyer (eds) Qualitative Methodologies for Geographers: Issues and Debates. London:Arnold, 1–20.Dwyer, K. (1977) ‘The dialogic of ethnology’, Dialectical Anthropology, 4: 205–24.

Ebdon, D. (1985) Statistics in Geography (2nd edn). Oxford: Blackwell.Edwards, K.C. (1970) ‘Organisation for fieldwork’, The Geographical Magazine, 42: 314.Ekinsmyth, C. (2002) ‘Feminist methodology’, P. Shurmer-Smith (ed.) Doing Cultural Geography. London:

Sage, 177–85.Ekinsmyth, C. and Shurmer-Smith, P. (2002) ‘Humanistic and behavioural geography’, in P. Shurmer-Smith

(ed.) Doing Cultural Geography. London: Sage, 19–27.

REFERENCES

385


Elton, G.R. (1972) Policy and Police: The Enforcement of the Reformation in the Age of Thomas Cromwell.Cambridge: Cambridge University Press.

E.M.Y. (1967) ‘Review of Dilke’s The Purpose and Organisation of Field Studies and Field Excursions in North-west England’, Geography, LII: 228–9.

England, K. (1994) ‘Getting personal: reflexivity, positionality and feminist research’, The ProfessionalGeographer, 46: 80–9.

Ennion, E.A.R. (ed.) (1949–52) Field Study Books (a series of books published under the auspices of theCouncil for the Promotion of Field Studies). London: Methuen.

Entrikin, J.N. (1981) ‘Philosophical issues in the scientific study of regions’, in D. Herbert and R.J. Johnston(eds) Geography and the Urban Environment. Volume 4. Chichester:Wiley, 1–27.

Entrikin, N. (1976) ‘Contemporary humanism in geography’, Annals of the Association of AmericanGeographers, 66: 615–32.

Entrikin, N. (1979) ‘Review of Humanistic Geography: Prospects and Problems, ed. Ley and Samuels’, EconomicGeography, 55: 253–7.

Evans, M. (1988) ‘Participant observation: the researcher as research tool’, in J. Eyles and D. Smith (eds)Qualitative Methods in Human Geography. Cambridge: Polity Press, 197–218.

Eyles, J. (1985) Senses of Place.Warrington: Silberbrook Press.Eyles, J. (1986) ‘Qualitative methods: a new revolution?’, in J. Eyles (ed.) Qualitative Approaches in Social and

Geographical Research. Occasional Paper 26. London: Department of Geography and Earth Science, QueenMary College, University of London, 1–22.

Eyles, J. (ed.) (1988a) Research in Human Geography: Introductions and Investigations. Oxford: Blackwell.Eyles, J. (1988b) ‘Interpreting the geographical world: qualitative approaches in geographical research’, in

J. Eyles and D.M. Smith (eds) Qualitative Methods in Human Geography. Cambridge: Polity Press, 1–16.Eyles, J. and Smith, D.M. (eds) (1988) Qualitative Methods in Human Geography. Cambridge: Polity Press.

Farrow, H., Moss, P. and Shaw, B. (eds) (1997) ‘Symposium on feminist participatory research’ (a series ofpapers), Antipode, 27: 71–101.

Fielding, N. and Fielding, J. (1986) Linking Data. London: Sage.Fielding, N. and Lee, R. (eds) (1991) Using Computers in Qualitative Research. London: Sage.Finch, J. (1993) ‘It’s great to have someone to talk to: ethics and the politics of interviewing women’, in

M. Hammersley (ed.) Social Research: Philosophy, Politics and Practice. London: Sage, 166–80 (originallypublished in Bell, C. and Roberts, H. (eds) (1984) Social Researching: Politics, Problems, Practice. London:Routledge, 70–87).

Fink,A. and Kosecoff, J. (1985) How to Conduct Surveys: A Step by Step Guide. London: Sage.Fishman, P. (1990) ‘Interaction: the work women do’, in J. McCarl Nielson (ed.) Feminist Research Methods:

Exploring Readings in the Social Sciences. London:Westview Press, 224–37.Flowerdew, R. (1998) ‘Reacting to Ground Truth’, Environment and Planning A, 30: 289–302.Flowerdew, R. and Martin, D. (eds) (1997) Methods in Human Geography: A Guide for Students Doing Research

Projects. Harlow: Longman.Floyd, B. (1963) ‘Quantification: a geographic deviation?’, The Professional Geographer, 15: 15–17.Foddy, W.(1993) Constructing Questions for Interviews and Questionnaires.Cambridge:Cambridge University Press.Foord, J. and Gregson, N. (1986) ‘Patriarchy: towards a reconceptualisation’, Antipode, 18: 186–211.Foot, S., Rigby, D. and Webber, M. (1989) ‘Theory and measurement in historical materialism’, in

A. Kobayashi and S. Mackenzie (eds) Remaking Human Geography. London: Unwin Hyman, 116–33.Forbes, D. (1988) ‘Getting by in Indonesia: research in a foreign land’, in J. Eyles (ed.) Research in Human

Geography. Oxford: Blackwell, 100–20.Ford, L. (1994) ‘Sunshine and shadow: lighting and colour in the depiction of cities on film’, in S.C.Aitken

and L.E. Zonn (eds) Place, Power, Situation and Spectacle: A Geography of Film. Lanham, MD: Rowman &Littlefield, 119–36.

Fotheringham, S. (1997) ‘Trends in quantitative methods. 1. Stressing the local’, Progress in HumanGeography, 21: 88–96.

Fotheringham, S. (1998) ‘Trends in quantitative methods. 2. Stressing the computational’, Progress in HumanGeography, 22: 283–92.


386


Fotheringham, S. (1999) ‘Trends in quantitative methods III. Stressing the visual’, Progress in HumanGeography, 23: 597–606.

Fotheringham, S., Brunsdon, C. and Charlton, M. (2000) Quantitative Geography: Perspectives on Spatial DataAnalysis. London: Sage.

Fotheringham,S. and O’Kelly,M. (1989) Spatial Interaction Models: Formulations and Applications. London:Kluwer.Foucault, M. (1986) The Care of the Self: The History of Sexuality. Volume 3. New York, NY: Pantheon Books.Foucault, M. (1991) ‘Governmentality’, in G. Burchell et al. (eds) The Foucault Effect: Studies in

Governmentality. Hemel Hempstead: Harvester Wheatsheaf, 87–104.Fowler, F. (1988) Survey Research Methods (revised edn). Beverly Hills, CA: Sage.Fowler, F. and Mangione,T. (1990) Standardized Survey Interviewing. London: Sage.Fowler, R. (1991) Language in the News. London: Routledge.Franklin, B. and Murphy, D. (1990) What News? London: Routledge.French,W.W. (1940) ‘Local survey at Tiverton’, Geography, XXV: 76–84.Friedman, J. (1990) ‘Being in the world: globalization and localization’, in M. Featherstone (ed.) Global

Culture: Nationalism, Globalization and Modernity. Newbury Park, CA: Sage, 311–28.Fuller, D. (1999) ‘Part of the action, or “going native”? Learning to cope with the “politics of integration”’,

Area, 31: 221–7.

Galbraith,V.H. (1961) The Making of the Domesday Book. Oxford: Oxford University Press.Gale, S. (1972) ‘Inexactness, fuzzy sets and the foundations of behavioural geography’, Geographical Analysis,

4: 285–322.Gamson,W., Crotequ, D., Hoynes,W. and Sasson,T. (1992) ‘Media images and the social construction of

reality’, Annual Review of Sociology, 18: 373–93.Gandy, M. (1996) ‘Visions of darkness: the representation of nature in the films of Werner Herzog’,

Ecumene, 3: 1–21.Geertz, C. (1973) The Interpretation of Cultures. New York, NY: Basic Books.Geertz, C. (1983) Local Knowledge: Further Essays in Interpretive Anthropology. New York, NY: Basic Books.Genthe, M.K. (1912) ‘Comment on Colonel Close’s address to the purpose and position of geography’,

Bulletin of the American Geographical Society, 44: 27–38.Gerard, R.W. (1969) ‘Hierarchy, entitation and levels’, in L.L. Whyte et al. (eds) Hierarchical Structures.

New York, NY:American Elsevier, 215–30.Gibbons, M., Limoges, C., Nowotny, H., Schwartzman, S., Scott, P. and Trow, S. (1994) The New Production

of Knowledge. London: Sage.Giddens,A. (1981) A Contemporary Critique of Historical Materialism. Volume 1. Basingstoke: Macmillan.Giddens,A. (1984) The Constitution of Society. Cambridge: Polity Press.Giddens,A. (1985) The Nation-state and Violence. Cambridge: Polity Press.Gilbert, D. (1995) ‘Between two cultures: geography, computing and the humanities’, Ecumene, 2: 1–13.Gilbert, D. (1999) ‘Sponsorship, academic independence and critical engagement: a forum on Shell, the

Ogoni dispute, and the Royal Geographical Society (with the Institute of British Geographers)’, Ethics,Place and Environment, 2: 219–28.

Gilbert, E. (1998) ‘Ornamenting the facade of hell: iconographies of nineteenth-century Canadian papermoney’, Environment and Planning D: Society and Space, 16: 57–80.

Gilbert, E.W. (1972) British Pioneers in Geography. Newton Abbot: David & Charles.Gilbert, N. and Mulkay, M. (1984) Opening Pandora’s Box. Cambridge: Cambridge University Press.Glennie, P. (1990) ‘Distinguishing Men’s Trades’: Occupational Sources and Debates for Pre-census England.

Historical Geography Research Series 25. Bristol: Institute of British Geographers.Gober, P., Glasmeier, A., Goodman, J., Plane, D., Stafford, H. and Wood, S. (1995) ‘Employment trends in

geography’, The Professional Geographer, 47: 336–46.Goddard, S. (ed.) (1983) A Guide to Information Sources in the Geographical Sciences. London: Croom Helm.Godkin, M.A. (1980) ‘Identity and place: clinical applications based on notions of rootedness and uproot-

edness’, in A. Buttimer and D. Seamon (eds) The Human Experience of Space and Place. London: CroomHelm, 73–85.

Godlewska,A. and Smith, N. (eds) (1994) Geography and Empire. Oxford: Blackwell.

REFERENCES

387


Goffman, E. (1968) Asylums: Essays on the Social Situation of Mental Patients and other Inmates.Harmondsworth: Penguin Books.

Gold, J. (1985) ‘From Metropolis to The City: film visions of the future city, 1919–39’, in J. Burgess andJ. Gold (eds) Geography, the Media and Popular Culture. London: Croom Helm, 123–43.

Gold, J. (1994) ‘Locating the message: place promotion as image communication’, in J. Gold and S.Ward(eds) Place Promotion. Chichester:Wiley, 19–37.

Goldman,A. and McDonald, S. (1987) The Group Depth Interview: Principles and Practices. Englewood Cliffs,NJ: Prentice Hall.

Goodchild, M. (1984) ‘Geocoding and geosampling’, in G.L. Gaile and C.J.Willmott (eds) Spatial Statisticsand Models. Dordrecht: Reidel, 33–54.

Goodchild, M. (1994) ‘Geographical information systems (GIS)’, in R.J. Johnston et al. (eds) The Dictionaryof Human Geography (3rd edn). Oxford: Blackwell, 219–20.

Goodchild, M. and Longley, P. (1998) ‘The future of GIS and spatial analysis’, in P. Longley et al. (eds)Geographical Information Systems: Principles, Techniques, Management and Applications (2 vols). Cambridgeand New York, NY: Geo Information International, 567–80.

Gorden, R. (1987) Interviewing. Homewood, IL: Dorsey Press.Gormley, N.B. (1994) ‘Letter from the field: reflections half way through’, in E. Robson and K.Willis (eds)

Postgraduate Fieldwork in Developing Areas: A Rough Guide. IBG Developing Areas Research GroupMonograph 8. London: IBG, 103–10.

Goss, J. (1993) ‘Placing the market and marketing place: tourist advertising of the Hawaiian Islands,1972–92’, Environment and Planning D: Society and Space, 11: 663–88.

Goss, J. (1995a) ‘Marketing the new marketing: the strategic discourse of geodemographic information sys-tems’, in J. Pickles (ed.) Ground Truth: The Social Implications of Geographic Information Systems. New York,NY: Guilford, 130–70.

Goss, J. (1995b) ‘We know who you are and we know where you live: the instrumental rationality of geo-marketing information systems’, Economic Geography, 71: 171–98.

Goss, J. (1996) ‘Introduction to focus groups’, Area, 28: 113–14.Goss, J. (1999) ‘Once-upon-a-time in the commodity world: an unofficial guide to mall of America’, Annals

of the Association of American Geographers, 89: 45–75.Gould, P. (1979) ‘Geography 1957–1977: the Augean period’, Annals of the Association of American

Geographers, 69: 139–51.Gould, P. (1981) ‘Letting the data speak for themselves’, Annals of the Association of American Geographers,

71: 166–76.Gouldner,A. (1967) Enter Plato. London: Routledge & Kegan Paul.Graham, E. (1997) ‘Philosophies underlying human geography research’, in R. Flowerdew and D. Martin

(eds) Methods in Human Geography. Harlow: Longman, 6–30.Greenbaum,T. (1988) The Practical Handbook and Guide to Focus Group Research. Lexington, MA: Lexington

Books.Gregory, D. (1978a) Ideology, Science and Human Geography. London: Hutchinson.Gregory, D. (1978b) ‘The discourse of the past: phenomenology, structuralism and historical geography’,

Journal of Historical Geography, 4: 161–73.Gregory, D. (1980) ‘The ideology of control: systems theory and geography’, Tidschrift voor Economische en

Social Geografie, 71: 327–42.Gregory, D. (1981) ‘Human agency and human geography’, Transactions, Institute of British Geographers, 6: 1–18.Gregory, D. (1982) Regional Transformation and Industrial Revolution: A Geography of the Yorkshire Woollen

Industry. London: Macmillan.Gregory, D. (1985) ‘Suspended animation: the stasis of diffusion theory’, in D. Gregory and J. Urry (eds)

Social Relations and Spatial Structures. London: Macmillan, 296–336.Gregory, D. (1986) ‘Entitation’, in R.J. Johnston et al. (eds) The Dictionary of Human Geography (2nd edn).

Oxford: Blackwell, 129–30.Gregory, D. (1989a) ‘Presences and absences: time–space relations and structuration theory’, in D. Held and

J.B. Thompson (eds) Social Theory of Modern Societies: Anthony Giddens and his Critics. Cambridge:Cambridge University Press, 185–214.


388


Gregory, D. (1989b) ‘Areal differentiation and post-modern human geography’, in D. Gregory andR.Walford (eds) Horizons in Human Geography. London: Macmillan, 67–96.

Gregory, D. (1989c) ‘The crisis of modernity’, in R. Peet and N. Thrift (eds) New Models in Geography.London: Unwin Hyman, 348–85.

Gregory, D. (1994) Geographical Imaginations. Oxford: Blackwell.Gregory, D. (1995) ‘Between the book and the lamp: imaginative geographies of Egypt, 1849–50’,

Transactions, Institute of British Geographers, 20: 29–57.Gregory, D. (1996) ‘Areal differentiation and post-modern human geography’, in J. Agnew et al. (eds)

Human Geography: An Essential Anthology. Oxford: Blackwell, 211–32.Gregory, D. (2000a) ‘Structuration theory’, in R.J. Johnston et al. (eds) The Dictionary of Human Geography

(4th edn). Oxford: Blackwell, 798–800.Gregory, D. (2000b) ‘Contextual approach’, in R.J. Johnston et al. (eds) The Dictionary of Human Geography

(4th edn). Oxford: Blackwell, 110–12.Gregory, D. (2000c) ‘Compositional theory’, in R.J. Johnston et al. (eds) The Dictionary of Human Geography

(4th edn). Oxford: Blackwell, 103.Gregory, D. (2000d) ‘Model’, in R.J. Johnston et al. (eds) The Dictionary of Human Geography (4th edn).

Oxford: Blackwell, 508–10.Gregory, D. (2000e) ‘Empiricism’, in R.J. Johnston et al. (eds) The Dictionary of Human Geography (4th edn).

Oxford: Blackwell, 205–6.Gregory, D. (2000f ) ‘Positivism’, in R.J. Johnston et al. (eds) The Dictionary of Human Geography (4th edn).

Oxford: Blackwell, 606–8.Gregory, D. and Ley, D. (1988) ‘Culture’s geographies’, Environment and Planning D: Society and Space,

6: 115–16.Gregory, D. and Urry, J. (eds) (1985) Social Relations and Spatial Structures. London: Macmillan.Gregory, S. (1978) Statistical Methods and the Geographer (4th edn). London: Longman.Gregory, S. (1992) ‘Thinking statistically’, in A. Rogers et al. (eds) The Student’s Companion to Geography.

Oxford: Blackwell, 135–40.Gregson, N. (1995) ‘Classics in human geography revisited: Thrift, N. “On the determination of social

action in space and time”. Commentary 2’, Progress in Human Geography, 19: 527–8.Gregson, N., Brooks, K. and Crewe, L. (2000) ‘Narratives of consumption and the body in the space of the

charity shop’, in P. Jackson et al. (eds) Commercial Cultures: Economies, Practices, Spaces.Oxford:Berg, 101–21.Grigg, D. (1965) ‘The logic of regional systems’, Annals of the Association of American Geographers, 55: 465–91.Grigg, D. (1967) ‘Regions, models and classes’, in R.J. Chorley and P. Haggett (eds) Models in Geography.

London: Methuen, 461–509.Guelke, L. (1974) ‘An idealist alternative in human geography’, Annals of the Association of American

Geographers, 64: 193–202.Guelke, L. (1978) ‘Geography and logical positivism’, in D. Herbert and R.J. Johnston (eds) Geography and

the Urban Environment. Chichester:Wiley, 35–61.Guelke, L. (1982) Historical Understanding in Geography. Cambridge: Cambridge University Press.

Habermas, J. (1972) Knowledge and Human Interests. London: Heinemann.Habermas, J. (1984) The Theory of Communicative Action. Volume 1. Reason and the Rationalisation of Society.

London: Heinemann.Haggett, P. (1965) Locational Analysis in Human Geography. London: Edward Arnold.Haggett, P., Cliff,A. and Frey,A. (1977) Locational Models (Volume I); Locational Methods (Volume II). London:

Arnold.Haining, R. (1990) Spatial Data Analysis in the Social and Environmental Sciences. Cambridge: Cambridge

University Press.Hakim, C. (1980) ‘Census reports as documentary evidence: the census commentaries 1801–1951’,

Sociological Review, 28: 551–80.Halfacree, K. (1993) ‘Locality and social space, discourse and alternative definitions of the rural’, Journal of

Rural Studies, 9: 1–15.Hall, P. and Markusen,A. (eds) (1985) Silicon Landscapes. Boston, MA:Allen & Unwin.

REFERENCES

389


Hall, P. and Preston, P. (1988) The Carrier Wave: New Information Technology and the Geography of Innovation.London: Unwin Hyman.

Hall, S. (ed.) (1997) Representation: Cultural Representations and Signifying Practices. London: Sage.Hall, S., Hobson, D., Lowe,A. and Willis, P. (eds) (1980) Culture, Media, Language. London: Methuen.Hall,T., Healey, M. and Harrison, M. (2002) ‘Fieldwork and disabled students: discourses of exclusion and

inclusion’, Transactions, Institute of British Geographers, 27: 213–31.Hammersley, M. (ed.) (1993) Social Research: Philosophy, Politics and Practice. 2nd edn. London: Sage.Hammersley, M. (1995) The Politics of Social Research. London: Sage.Hammersley, M. and Atkinson, P. (1983) Ethnography: Principles in Practice. London:Tavistock.Hammersley, M. and Atkinson, P. (1995) Ethnography: Principles in Practice. 2nd edn. London: Routledge.Hammond, P. and McCullagh, P. (1974) Quantitative Techniques in Geography. Oxford: Clarendon Press.Hamnett, C. (1998) ‘The owner-occupied housing market in Britain: a north-south divide?’, in J. Lewis and

A.Townsend (eds) The North-south Divide. London: Paul Chapman, 97–113.Hannah, M.G. (1997) ‘Space and the structuring of disciplinary power: an interpretive review’, Geografiska

Annaler B, 79: 171–80.Hanson, S. (1988) ‘Soaring’, The Professional Geographer, 40: 4–7.Haraway, D. (1988) ‘Situated knowledges: the science question in feminism and the privilege of partial

perspective’, Feminist Studies, 14: 575–99.Haraway, D. (1989) Primate Visions: Gender, Race and Nature in the World of Modern Science. New York, NY:

Routledge.Haraway, D. (1991) Simians, Cyborgs and Women: The Reinvention of Nature. New York, NY: Routledge.Haraway, D. (1996) ‘Situated knowledges: the science question in feminism and the privilege of partial

perspective’, in J. Agnew et al. (eds) Human Geography: An Essential Anthology. Oxford: Blackwell,108–28.

Harley, J.B. (1988) ‘Maps, knowledge and power’, in D. Cosgrove and S. Daniels (eds) The Iconography ofLandscape: Essays on the Symbolic Representation, Design and Use of Past Environments. Cambridge:Cambridge University Press, 277–312.

Harley, J.B. (1989) ‘Deconstructing the map’, Cartographica, 27: 1–23.Harré, R. (1979) Social Being. Oxford: Blackwell.Harris, T., Weiner, D., Warner, T. and Levin, R. (1995) ‘Pursuing social goals through participatory GIS:

redressing South Africa’s historical political ecology’, in J. Pickles (ed.) Ground Truth: The SocialImplications of Geographic Information Systems. New York, NY: Guilford, 196–222.

Harrison, R.T. and Livingstone, D.N. (1979) ‘There and back again: towards a critique of idealist humangeography’, Area, 11: 75–9.

Harrison, R.T. and Livingstone, D.N. (1980) ‘Philosophy and problems in human geography: a presuppo-sitional approach’, Area, 12: 25–31.

Hart, J.F. (1982) ‘The highest form of the geographer’s art’, Annals of the Association of American Geographers,72: 1–29.

Hartshorne, R. (1939) The Nature of Geography: A Critical Survey of Current Thought in the Light of the Past.Lancaster, PA:Association of American Geographers.

Harvey, D. (1969) Explanation in Geography. London: Edward Arnold.Harvey, D. (1973) Social Justice and the City. London: Edward Arnold.Harvey, D. (1982) The Limits to Capital. London: Blackwell.Harvey, D. (1985) Consciousness and the Urban Experience. Oxford: Blackwell.Harvey, D. (1989a) The Condition of Postmodernity: An Enquiry into the Origins of Cultural Change. Oxford:

Blackwell.Harvey, D. (1989b) The Urban Experience. Oxford: Blackwell.Harvey, D. (1990) ‘Between space and time: reflections on the geographical imagination’, Annals of the

Association of American Geographers, 80: 418–34.Harvey, D. (1995) ‘A geographer’s guide to dialectical thinking’, in A.D. Cliff et al. (eds) Diffusing Geography:

Essays for Peter Haggett. Oxford: Blackwell, 3–21.Harvey, D. (1996) Justice, Nature and the Geography of Difference. Oxford: Blackwell.Harvey, D. (2000) Spaces of Hope. Edinburgh: Edinburgh University Press.


390


Hay, A.M.H. (1994) ‘Systems analysis’, in R.J. Johnston et al. (eds) The Dictionary of Human Geography(3rd edn). Oxford: Blackwell, 615.

Hay, I. (1998) ‘Making moral imaginations. Research ethics, pedagogy and professional human geography’,Ethics, Place and Environment, 1: 55–76.

Hay, I. (2000) Qualitative Research Methods in Geography (Meridian Series). Melbourne: Oxford University Press.Hepple, L. (1998) ‘Context, social construction, and statistics: regression, social science and human geography’,

Environment and Planning A, 30: 225–34.Hepple, L. (2000) ‘Spatial autocorrelation’, in R.J. Johnston et al. (eds) The Dictionary of Human Geography

(4th edn). Oxford: Blackwell, 775.Herbert, S. (2000) ‘For ethnography’, Progress in Human Geography, 24: 550–68.Hetherington,K. and Law, J. (2000) ‘After networks’,Environment and Planning D: Society and Space, 18: 127–32.Hill, M.R. (1981) ‘Positivism: a “hidden” philosophy in geography’, in M.E. Harvey and B.P. Holly (eds)

Themes in Geographic Thought. London: Croom Helm, 38–60.Hillis, K. (1994) ‘The virtue of becoming a no-body’, Ecumene, 1: 177–96.Hinchliffe, S. (2000) ‘Performance and experimental knowledge: outdoor management training and the end

of epistemology’, Environment and Planning D: Society and Space, 18: 575–95.Hodge, D.C. (ed.) (1995) ‘Should women count? The role of quantitative methodology in feminist

geographical research’ (a series of papers), The Professional Geographer, 47: 426–66.Hoggart, K., Lees, L. and Davies,A. (2002) Researching Human Geography. London:Arnold.Holbrook,B.and Jackson,P. (1996) ‘Shopping around: focus group research in north London’,Area, 28:136–42.Holloway, S.L., Rice, S.P. and Valentine, G. (eds) (2003) Key Concepts in Geography. London: Sage.Holstein, J. and Gubrium, J. (1997) ‘Active interviewing’, in D. Silverman (ed.) Qualitative Research: Theory,

Method and Practice. London: Sage, 113–29.hooks, b. (1997) Real to Reel. Chicago, IL: University of Chicago Press.Horvarth, R.J. (1971) ‘The “Detroit Geographical Expedition and Institute” experience’, Antipode, 3: 73–84.Howard, S. (1994) ‘Methodological issues in overseas fieldwork: experiences from Nicaragua’s northern

Atlantic coast’, in E. Robson and K. Willis (eds) Postgraduate Fieldwork in Developing Areas: A RoughGuide. IBG Developing Areas Research Group Monograph 8. London: IBG, 19–35.

Hubbard, P.J., Kitchin, R., Bartley, B. and Fuller, D. (2002) Thinking Geographically: Space, Theory andContemporary Human Geography. London: Continuum.

Huck, S. and Sandler, H. (1979) Rival Hypotheses: Alternative Interpretations of Data Based Conclusions.New York, NY: HarperCollins.

Hudson, R. (1995) ‘Making music work? Alternative regeneration strategies in a deindustrialized locality:the case of Derwentside’, Transactions, Institute of British Geographers, 20: 461–73.

Huff, D. (1973) How to Lie with Statistics. Harmondsworth: Penguin Books.Huggett, R. (1980) Systems Analysis in Geography. Oxford: Clarendon Press.Hughes,A. and Cormode, L. (eds) (1998) ‘Theme issue: researching elites and elite spaces’, Environment and

Planning A, 30: 2095–180.Hughes, A., Morris, C. and Seymour, S. (eds) (2000) Ethnography and Rural Research. Gloucester:

Countryside and Community Press.Hutchings, G.E. (1949) ‘The geographer as field naturalist’, School Nature Study, 44: 33–8.Hutchings, G.E. (1960) Landscape Drawing. London: Methuen.Hutchings, G.E. (1962) ‘Geographical field teaching’, Geography, XLVII: 1–14.

Jackson, P. (1983) ‘Principles and problems of participant observation’, Geografiska Annaler Series B: HumanGeography, 65: 39–46.

Jackson, P. (1985) ‘Urban ethnography’, Progress in Human Geography, 9: 157–76.Jackson, P. (1988a) ‘Definitions of the situation: neighbourhood change and local politics in Chicago’, in

J. Eyles and D.M. Smith (eds) Qualitative Methods in Human Geography. Cambridge: Polity Press, 49–74.Jackson,P. (1988b) ‘Street life: the politics of carnival’,Environment and Planning D: Society and Space, 6: 213–27.Jackson, P. (1989) Maps of Meaning: An Introduction to Cultural Geography. London: Unwin Hyman.Jackson, P. (1993) ‘Changing ourselves: a geography of position’, in R.J. Johnston (ed.) The Challenge for

Geography: A Changing World, a Changing Discipline. Oxford: Blackwell, 198–214.

REFERENCES

391


Jackson, P. (1998) ‘Classics in human geography revisited: Ley 1974 The Black Inner City as Frontier Outpost,commentary 1’, Progress in Human Geography, 22: 75–6.

Jackson, P. (1999) ‘Postmodern urbanism and the ethnographic void’, Urban Geography, 20: 400–2.Jackson, P. (2000) ‘Positionality’, in R.J. Johnston et al. (eds) The Dictionary of Human Geography (4th edn).

Oxford: Blackwell, 604–5.Jackson,P. (2001) ‘Making sense of qualitative data’, in M.Limb and C.Dwyer (eds) Qualitative Methodologies

for the Geographer: Issues and Debates. London:Arnold, 119–214.Jackson, P. and Smith, S. (1984) Exploring Social Geography. Hemel Hempstead:Allen & Unwin.Jackson, P., Stevenson, N. and Brooks, K. (1999) ‘Making sense of men’s lifestyle magazines’, Environment

and Planning D: Society and Space, 17: 353–68.Jacobs, J.M. (1996) Edge of Empire: Postcolonialism and the City. London: Routledge.James, P.E. and Mather, C. (1977) ‘The role of periodic field conferences in the development of geograph-

ical ideas in the United States’, The Geographical Review, 67: 446–61.Jameson,A.H. and Ormsby, M.T.M. (1934) Mathematical Geography. London: Sir Isaac Pitman & Sons.Jayaratne,T. (1993) ‘The value of quantitative methodology for feminist research’, in M. Hammersley (ed.)

Social Research. London: Sage, 109–23.Jensen, M.A.P. (1946) ‘The promotion of field studies’, Geography, XXXI: 153–6.Johnson, R. (1986) ‘The story so far: and further transformations?’, in D. Punter (ed.) Introduction to

Contemporary Cultural Studies. London: Longman, 277–313.Johnston, R.J. (1978) Multivariate Statistical Analysis in Geography. London: Longman.Johnston, R.J. (1979) Political, Electoral and Spatial Systems. Oxford: Clarendon Press.Johnston, R.J. (1997a) ‘W(h)ither spatial science and spatial analysis’, Futures, 29: 323–36.Johnston,R.J. (1997b) Geography and Geographers: Anglo-American Human Geography since 1945. London:Arnold.Johnston, R.J. (2000a) ‘On disciplinary history and textbooks: or where has spatial analysis gone?’, Australian

Geographical Studies, 38: 125–37.Johnston, R.J. (2000b) ‘Questionnaire’, in R.J. Johnston et al. (eds) The Dictionary of Human Geography

(4th edn). Oxford: Blackwell, 668.Johnston, R.J. (2000c) ‘Intellectual respectability and disciplinary transformation? Radical geography and

the institutionalisation of geography in the USA since 1945’, Environment and Planning A, 27: 971–90.Johnston, R.J., Hepple, R., Hoare,A., Jones, K. and Plummer, P. (2003) ‘Contemporary fiddling in human

geography while Rome burns: has quantitative analysis been largely abandoned – and should it be?’,Geoforum, 34: 157–61.

Johnston, R.J. and Pattie, C. (2000) ‘New Labour, new electoral system, new electoral geographies?A review of proposed constitutional changes in the United Kingdom’, Political Geography, 19: 495–515.

Johnston, R.J., Pattie, C. and Allsopp, J.G. (1988) A Nation Dividing? The Electoral Map of Great Britain1979–1987. London: Longman.

Jones, E. (1980) ‘Review of Ideology, Science and Human Geography by Gregory, and Humanistic Geography:Prospects and Problems, ed. Ley & Samuels’, The Geographical Journal, 146: 113–15.

Jones, J.P. III and Casetti, E. (eds) (1992) Applications of the Expansion Methods. London: Routledge.Jones, K. (1991) ‘Specifying and estimating multilevel models for geographical research’, Transactions,

Institute of British Geographers, 16: 148–59.Jones, K. and Duncan, C. (1996) ‘People and places: the multilevel model as a general framework for the

quantitative analysis of geographical data’, in P. Longley and M. Batty (eds) Spatial Analysis: Modelling ina GIS Environment. Cambridge: GeoInformation International, 79–104.

Jones,W.D. (1931) ‘Field mapping of residential areas in Metropolitan Chicago’, Annals of the Association ofAmerican Geographers, 21: 207–14.

Jones,W.D. (1934) ‘Procedures in investigating human occupance of a region’, Annals of the Association ofAmerican Geographers, 24: 93–107.

Jones,W.D. and Sauer, C.O. (1915) ‘Outline for fieldwork in geography’, Bulletin of the American GeographicalSociety, 47: 520–6.

Katz, C. (1992) ‘All the world is staged: intellectuals and the projects of ethnography’, Environment andPlanning D: Society and Space, 10: 495–510.


392


Katz,C. (1994) ‘Playing the field: questions of fieldwork in geography’,The Professional Geographer, 46: 67–72.Kearns, R. (2000) ‘Being there: research through observing and participating’, in I. Hay (ed.) Qualitative

Research Methods in Human Geography. South Melbourne: Oxford University Press, 103–21.Keat, R. and Urry, J. (1975) Social Theory as Science. London: Routledge & Kegan Paul.Keith,M. (1991) ‘Knowing your place: the imagined geographies of racial subordination’, in C.Philo (ed.) New

Words, New Worlds: Reconceptualising Social and Cultural Geography.Aberystwyth:Cambrian Printers, 178–92.Keith, M. (1992) ‘Angry writing: (re)presenting the unethical world of the ethnographer’, Environment and

Planning D: Society and Space, 10: 551–68.Kennedy, B.A. (1979) ‘A naughty world’, Transactions, Institute of British Geographers, 4: 550–8.Kennedy,C. and Lukinbeal,C. (1997) ‘Towards a holistic approach to geographical research on film’,Progress

in Human Geography, 21: 33–50.Kimble, G.H.T. (1938) Geography in the Middle Ages. London: Methuen.Kindon,S. (2003) ‘Participatory video in geographic research: a feminist practice of looking?’,Area, 35: 142–53.Kingman, M.M. (1969) ‘Field study for the non-sighted’, The Professional Geographer, 21: 199.Kingsbury, P. (2003) ‘Psychoanalysis, a gay spatial science?’, Social and Cultural Geography, 4: 347–68.Kinross, R. (1985) ‘The rhetoric of neutrality’, Design Issues, 2: 18–30.Kinsman, P. (1995) ‘Landscape, race and national identity: the photography of Ingrid Pollard’, Area, 27: 300–10.Kitajima, S. (1999) ‘Changing universities and the question of academic practice’, Ethics, Place and

Environment, 2: 254–6.Kitchin, R.M. and Hubbard, P.J. (eds) (1999) ‘Research, action and “critical geographies”’ (a series of

papers), Area, 31: 195–246.Kitchin, R.M. and Tate, N. (2000) Conducting Research in Human Geography: Theory, Methodology and Practice.

Harlow: Prentice Hall.Kneafsey, M. (2000) ‘Changing roles and constructing identities: ethnography in the Celtic periphery’, in

A. Hughes et al. (eds) Ethnography and Rural Research. Cheltenham: Countryside Commission, 52–65.Kobayashi,A. (2001) ‘Negotiating the personal and the political in critical qualitative research’, in M. Limb

and C. Dwyer (eds) Qualitative Methodologies for Geographers. London:Arnold, 55–70.Kong, L. (1993) ‘Ideological hegemony and the political symbolism of religious buildings in Singapore’,

Environment and Planning D: Society and Space, 11: 23–45.Kong, L. (1995a) ‘Music and cultural politics: ideology and resistance in Singapore’, Transactions, Institute of

British Geographers, 20: 447–59.Kong, L. (1995b) ‘Popular music in geographical analyses’, Progress in Human Geography, 19: 193–8.Kong, L. (1996) ‘Popular music in Singapore: exploring local cultures, global resources, and regional iden-

tities’, Environment and Planning D: Society and Space, 14: 273–92.Kong, L. and Goh, E. (1995) ‘Folktales and reality: the social construction of race in Chinese tales’, Area,

27: 261–7.Kreuger, R. (1988) Focus Groups: A Practical Guide for Applied Research. Beverly Hills, CA: Sage.Krippendorff, K. (1980) Content Analysis. London: Sage.Kunzru, H. (1997) ‘You are cyborg: for Donna Haraway, we are already assimilated’, Wired, 5

(www.wired.com/wired/archives/5.02/ffharaway.html).

Langton, J. (1972) ‘Potentialities and problems in adopting a systems approach to the study of change inhuman geography’, Progress in Human Geography, 4: 125–79.

Latour, B. (1991) ‘Technology is society made durable’, in J. Law (ed.) A Sociology of Monsters: Essays onPower, Technology and Domination. London: Routledge, 103–31.

Latour, B. (1993) We Have Never Been Modern. Brighton: Harvester Wheatsheaf.Latour, B. (1996) Aramis, or the Love of Technology. Cambridge, MA: Harvard University Press.Latour, B. (1999) Pandora’s Hope: Essays on the Reality of Science Studies. Cambridge, MA: Harvard University

Press.Laurier, E. (2001) ‘Why people say where they are during mobile phone calls’, Environment and Planning

D: Society and Space, 19: 485–504.Laurier, E. and Philo, C. (1999) ‘X-morphising: review essay of Bruno Latour’s Aramis, or the Love of

Technology’, Environment and Planning A, 31: 1047–71.

REFERENCES

393


Laurier, E. and Philo, C. (2003) ‘The region in the boot: mobilising lone subjects and multiple objects’,Environment and Planning D: Society and Space, 21: 85–106.

Law, J. (1986) ‘On the methods of long distance control: vessels, navigation and the Portuguese route toIndia’, in J. Law (ed.) Power, Action and Belief. London: Routledge & Kegan Paul, 234–63.

Law, J. (1991) ‘Introduction: monsters, machines and sociotechnical relations’, in J. Law (ed.) A Sociology ofMonsters. London: Routledge, 1–23.

Law, J. (1994) Organising Modernity. Oxford: Blackwell.Lee,R. (ed.) (1992) ‘Teaching qualitative geography: a JGHE written symposium’ (a series of papers), Journal

of Geography in Higher Education, 16: 123–84.Leighley, J. (ed.) (1963) Land and Life: A Selection from the Writings of Carl Ortwin Sauer. Berkeley and

Los Angeles, CA: University of California Press.Leslie, D. and Reimer, S. (2003) ‘Gender, modern design, and home consumption’, Environment and Planning

D: Society and Space, 21: 293–316.Levitas, R. (1996) ‘Fiddling while Britain burns? The “measurement” of unemployment’, in R. Levitas and

W. Guy (eds) Interpreting Official Statistics. London: Routledge, 45–65.Lewis, C. and Pile, S. (1996) ‘Woman, body, space: Rio Carnival and the politics of performance’, Gender,

Place and Culture, 3: 23–42.Lewis, P. (1985) ‘Beyond description’, Annals of the Association of American Geographers, 75: 465–77.Ley, D. (1974) The Black Inner City as Frontier Outpost: Images and Behavior of a Philadelphia Neighborhood.

Washington, DC:Association of American Geographers.Ley, D. (1977) ‘Social geography and the taken-for-granted world’, Transactions, Institute of British

Geographers, 2: 498–512.Ley, D. (1978) ‘Review of Prisoners of Space? by Graham Rowles’, Economic Geography, 54: 355–6.Ley, D. (1980) Geography without Man: A Humanistic Critique. School of Geography Research Paper 24. Oxford:

University of Oxford.Ley, D. (1981a) ‘Behavioural geography and the philosophies of meaning’, in K.R. Cox and R.G. Golledge

(eds) Behavioural Problems in Geography Revisited. London: Methuen, 209–30.Ley, D. (1981b) ‘Cultural/humanistic geography’, Progress in Human Geography, 5: 249–57.Ley, D. (1982) ‘Rediscovering man’s place’, Transactions, Institute of British Geographers, 7: 248–53.Ley, D. (1988) ‘Interpretative social research in the inner city’, in J. Eyles (ed.) Research in Human Geography:

Introductions and Investigations. Oxford: Blackwell, 121–38.Ley, D. (1998) ‘Classics in human geography revisited: Ley, D. (1974) The Black Inner City as Frontier Outpost:

author’s response’, Progress in Human Geography, 22: 78–80.Ley, D. and Cybriwsky, R. (1974) ‘Urban graffiti as territorial markers’, Annals of the Association of American

Geographers, 64: 491–505.Ley, D. and Mountz,A. (2001) ‘Interpretation, representation, positionality: issues in field research in human

geography’, in M. Limb and C. Dwyer (eds) Qualitative Methodologies for Geographers: Issues and Debates.London: Edward Arnold, 234–47.

Ley, D. and Samuels, M.S. (eds) (1978) Humanistic Geography: Prospects and Problems. London: Croom Helm.Leyshon, A., Matless, D. and Revill, G. (1995) ‘The place of music’, Transactions, Institute of British

Geographers, 20: 423–33.Leyshon,A., Matless, D. and Revill, G. (eds) (1998) The Place of Music. New York: Guilford.Lie, M. (1992) ‘Teknologi og kjonnsidentitet: Har datamaskinen kjonn?’, in A. Andenaes et al. (eds) Epler

Fra Var Egen Hage.Trondheim: Centre for Women’s Studies.Limb, M. and Dwyer, C. (eds) (2001) Qualitative Methodologies for Geographers: Issues and Debates. London:

Edward Arnold.Lindsay, J.M. (1997) Techniques in Human Geography. London: Routledge.Livingstone, D.N. (1992) The Geographical Tradition: Episodes in the History of a Contested Enterprise. Oxford:

Blackwell.Longley, P. (1998) ‘Foundations’, in P. Longley et al. (eds) Geocomputation: A Primer. Chichester:Wiley, 17–29.Longley, P., Brooks, S., McDonnell, R. and Macmillan, B. (eds) (1998b) Geocomputation: A Primer.

Chichester:Wiley.


394


Longley, P. and Clarke, G. (eds) (1995) GIS for Business and Service Planning. Cambridge: GeoInformationInternational.

Longley, P., Goodchild, M., Maguire, D. and Rhind, D. (eds) (1998a) Geographical Information Systems:Principles, Techniques, Management and Application (2 vols). New York, NY:Wiley.

Lorimer,H.and Spedding,N. (2002) ‘Editorial:putting philosophies of geography into practice’,Area, 34:227–8.Lowenthal, D. (1961) ‘Geography, experience and imagination: towards a geographical epistemology’,

Annals of the Association of American Geographers, 51: 241–60.Lutwack, L. (1984) The Role of Place in Literature. Syracuse, NY: Syracuse University Press.

MacKinnon, D. (2000) ‘Managerialism, governmentality and the state: a neo-Foucauldian approach to localeconomic governance’, Political Geography, 19: 293–314.

Macmillan,W. (1998) Computing and the Science of Geography. University of Oxford Economic Research GroupWorking Paper. Oxford: University of Oxford Economic Research Group.

Madge, C. (1993) ‘Boundary disputes: comments on Sidaway (1992)’, Area, 25: 294–9.Madge, C. (1994) ‘The ethics of research in the “Third World”’, in E. Robson and K. Willis (eds)

Postgraduate Fieldwork in Developing Areas: A Rough Guide. Developing Areas Research Group Monograph 9.London: RGS and the IBG, 91–102.

Madge, C., Raghuram, P., Skelton,T.,Willis, K. and Williams, J. (1997) ‘Methods and methodologies in fem-inist research’, in WGSG (ed.) Feminist Geographies: Diversity and Difference. Harlow: Longman, 86–111.

Maitland, F.W. (1897) Domesday Book and Beyond: Three Essays in the Early History of England. Cambridge:Cambridge University Press.

Malbon, B. (1999) Clubbing: Dancing, Ecstacy, Vitality. London: Routledge.Mallory, W.E. and Simpson-Houseley, P. (1987) Geography and Literature: A Meeting of the Disciplines.

Syracuse, NY: Syracuse University Press.Mann, M. (1986) The Sources of Social Power. Volume I. A History of Power from the Beginning to AD 1760.

Cambridge: Cambridge University Press.Marcus, G. (1986) ‘Contemporary problems of ethnography in the modern world system’, in J. Clifford and

G. Marcus (eds) Writing Culture: The Poetics and Politics of Ethnography. Berkeley, CA: University ofCalifornia Press, 165–93.

Marcus,G. (1995) ‘Ethnography in/of the world system: the emergence of multi-sited ethnography’,AnnualReview of Anthropology, 24: 95–117.

Marcus, G. (1998) Ethnography through Thick and Thin. Princeton, NJ: Princeton University Press.Marcus, G. (2000) ‘The twistings and turnings of geography and anthropology in the winds of millennial

transition’, in I. Cook et al. (eds) Cultural Turns/Geographical Turns: Perspectives on Cultural Geography.Harlow: Longman, 13–25.

Marcus, G. and Fischer, M. (1986) Anthropology as Cultural Critique: An Experimental Moment in the HumanSciences (1st edn). Chicago, IL: University of Chicago Press.

Marcus, G. and Fischer, M. (1999) Anthropology as Cultural Critique: An Experimental Moment in the HumanSciences (2nd edn). Chicago, IL: University of Chicago Press.

Marsh, C. (1988) Exploring Data. Cambridge: Polity Press.Martin, D. (1996) Geographic Information Systems: Socio-economic Applications (2nd edn). London: Routledge.Martin, D. (1997) ‘Geographic information systems and spatial analysis’, in R. Flowerdew and D. Martin

(eds) Methods in Human Geography. Harlow:Addison Wesley Longman, 213–29.Martin, D. (2000) ‘Towards the geographies of the 2001 UK Census of Population’, Transactions, Institute of

British Geographers, 25: 321–32.Martin, R. (2001) ‘Geography and public policy: the case of the missing agenda’, Progress in Human

Geography, 25: 189–210.Martin, R., Sunley, P. and Willis, J. (1993) ‘The geography of trade union decline: spatial dispersal or

regional resilience?’, Transactions, Institute of British Geographers, 18: 36–62.Marx, K. (1973) Grundrisse. Harmondsworth: Penguin Books.Mason, J. (1996) Qualitative Researching. London: Sage.Massey, D. (1984) Spatial Divisions of Labour. London: Macmillan.Massey, D. (1991a) ‘Flexible sexism’, Environment and Planning D: Society and Space, 9: 31–57.

REFERENCES

395


Massey, D. (1991b) ‘A progressive sense of place’, Marxism Today, June: 24–9.Massey, D. (1992) ‘Politics and space/time’, New Left Review, 196: 65–84.Massey, D. (1993) ‘Power-geometry and a progressive sense of place’, in J. Bird et al. (eds) Mapping the

Futures: Local Cultures, Global Change. London: Routledge, 59–69.Massey, D. (1994) Space, Place and Gender. Cambridge: Polity Press.Massey, D. (1999a) ‘Space-time, “science” and the relationship between physical geography and human

geography’, Transactions, Institute of British Geographers, 24: 261–76.Massey, D. (1999b) ‘Spaces of politics’, in D. Massey et al. (eds) Human Geography Today. Cambridge: Polity

Press, 279–94.Massey, D. (2000) ‘Practising political relevance’, Transactions, Institute of British Geographers, 25: 131–3.Matless, D. and Revill, G. (1995) ‘A solo ecology: the erratic art of Andy Goldsworthy’, Ecumene, 2: 423–48.Matthews, J. (1981) Quantitative and Statistical Approaches to Geography: A Practical Manual. Oxford: Pergamon.Matthewson, K. (1998) ‘Classics in human geography revisted, Duncan, J.S.“The superorganic in American

cultural geography”: commentary 2’, Progress in Human Geography, 22: 569–71.May, J. (1991) ‘Putting some space back into time: the spatio-temporal self and the turn to the postmodern’,

in C. Philo (ed.) New Words, New Worlds: Reconceptualising Social and Cultural Geography. Aberystwyth:Cambrian Printers, 168–71.

May,T. (1997) Social Research: Issues, Methods and Process. Buckingham: Open University Press.McCloskey, D. (1988) ‘The consequences of rhetoric’, in A. Klamer et al. (eds) The Consequences of Economic

Rhetoric. Cambridge: Cambridge University Press, 280–93.McDowell, L. (1983) ‘Towards an understanding of the gender division of urban space’, Environment and

Planning D: Society and Space, 1: 15–30.McDowell, L. (1988) ‘Coming in from the dark: feminist research in geography’, in J. Eyles (ed.) Research

in Human Geography: Introductions and Investigations. Oxford: Blackwell, 155–73.McDowell, L. (1989) ‘Women, gender and the organisation of space’, in D. Gregory and R.Walford (eds)

Horizons in Human Geography. London: Macmillan, 136–51.McDowell, L. (1992a) ‘Doing gender: feminism, feminists and research methods in human geography’,

Transactions, Institute of British Geographers, 17: 399–416.McDowell, L. (1992b) ‘Multiple voices: speaking from inside and outside “the project”’, Antipode, 24: 56–72.McDowell, L. (1995) ‘Understanding diversity: the problem of/for theory’, in R.J. Johnston et al. (eds)

Geographies of Global Change. Oxford: Blackwell, 280–94.McDowell, L. (1997) Capital Culture: Gender at Work in the City. Oxford: Blackwell.McDowell, L. (1999) Gender, Identity and Place: Understanding Feminist Geographies. Cambridge: Polity Press.McDowell, L. and Court, G. (1994) ‘Performing work: bodily representations in merchant banks’,

Environment and Planning D: Society and Space, 12: 727–50.McDowell, L. and Massey, D. (1984) ‘A woman’s place?’, in D. Massey and J.Allen (eds) Geography Matters!

A Reader. Cambridge: Cambridge University Press, 128–47.McDowell, L. and Sharp, J.P. (eds) (1997) Space, Gender, Knowledge: Feminist Readings. London:Arnold.McEwan,C. (2000) Gender, Geography and Empire: Victorian Women Travellers in West Africa. Aldershot: Ashgate.McKibben, B. (1989) The End of Nature. New York, NY:Anchor Books.Mead, W. (1981) ‘Review of The Human Experience of Space and Place, ed. Buttimer & Seamon’,

The Geographical Journal, 147: 94–5.Medick, H. (1976) ‘The proto-industrial family economy: the structural function of household and family

during the transition from peasant society to industrial capitalism’, Social History, 3: 291–315.Meinig, D. (1983) ‘Geography as an art’, Transactions, Institute of British Geographers, 8: 314–28.Mercer, D. and Powell, J. (1972) ‘Phenomenology and related non-positivistic viewpoints in the social

sciences’, Monash Publications in Geography, 1972–6. Melbourne: Monash University.Merrifield, A. (1994) ‘Situated knowledge through exploration: reflections on Bunge’s “Geographical

Expeditions”’, Antipode, 27: 49–70.Merton, R., Fiske, M. and Kendall, P. (1990) The Focused Interview (2nd edn). New York, NY: Free Press.Michael, J. (1982) The Politics of Secrecy. London: Penguin Books.Michael, M. (2000) Reconnecting Culture, Technology and Nature: From Society to Heterogeneity. London:

Routledge.


396


Milbourne, P. (2000) ‘Exporting “other” rurals: new audiences for qualitative research’, in A. Hughes et al.(eds) Ethnography and Rural Research. Cheltenham: Countryside and Community Press, 179–97.

Miller, E.J.W. (1969) ‘The naming of the land in the Arkansas Ozarks’, Annals of the Association of AmericanGeographers, 59: 240–51.

Miller, M. (1998) ‘(Re)presenting voices in dramatically scripted research’, in A. Banks and S. Banks (eds)Fiction and Social Research: By Ice or Fire.Walnut Creek, CA:AltaMira Press, 67–78.

Mills, C.W. (1970) The Sociological Imagination. Harmondsworth: Penguin Books.Mills, I. (1987) ‘Developments in census-taking since 1841’, Population Trends, 48: 37–44.Milner,A. (2002) Re-imagining Cultural Studies: The Promise of Cultural Materialism. London: Sage.Mitchell, B. and Draper, D. (1983a) Relevance and Ethics in Geography. London: Longman.Mitchell, B. and Draper, D. (1983b) ‘Ethics in geographical research’, The Professional Geographer, 35: 9–17.Mitchell, D. (1995) ‘There’s no such thing as culture: towards a reconceptualisation of the idea of culture

in geography’, Transactions, Institute of British Geographers, 20: 102–16.Mitchell, D. (2000) Cultural Geography: A Critical Introduction. Oxford: Blackwell.Mitchell,T. (1989) ‘The world-as-exhibition’, Comparative Studies in Society and History, 31: 217–36.Mitchell,W. (1986) Iconography: Image, Text, Ideology. Chicago, IL: University of Chicago Press.Miyares, I. and McGlade, M. (1994) ‘Specialisation in “jobs in geography”’, The Professional Geographer,

46: 170–7.Mohammad, R. (2001) ‘“Insiders” and/or “outsiders”: positionality, theory and praxis’, in M. Limb and

C. Dwyer (eds) Qualitative Methodologies for Geographers: Issues and Debates. London: Edward Arnold,101–17.

Molotch, H. (1996) ‘LA as design product’, in A. Scott and E. Soja (eds) The City. Berkeley, CA: Universityof California Press, 225–77.

Moon, G. and Brown,T. (2000) ‘Governmentality and the spatialised discourse of policy: the consolidationof the post-1989 NHS reforms’, Transactions, Institute of British Geographers, 25: 65–76.

Moore, D. (1997) ‘Remapping resistance: “Ground for struggle” and the politics of place’, in S. Pile (ed.)Geographies of Resistance. London: Routledge, 87–105.

Morgan, D. (1993) Successful Focus Groups: Advancing the State of the Art. Beverly Hills, CA: Sage.Morgan, D. (1997) Focus Groups as Qualitative Research (2nd edn). Beverly Hills, CA: Sage.Morgan, K. and Sayer, A. (1988) Microcircuits of Capital: ‘Sunrise’ Industries and Uneven Development.

Cambridge: Polity Press.Morgan, R. (1967) ‘Grass roots of geography’, The Geographical Magazine, 40: 145–6.Morley, D. and Robins, K. (1995) Spaces of Identity: Global Media, Electronic Landscapes and Cultural

Boundaries. London: Routledge.Moser, C. and Kalton, G. (1971) Survey Methods in Social Investigation (2nd edn 1985). London: Heinemann.Murdoch, J. (1995) ‘Actor networks and the evolution of economic forms: combining description and

explanation in theories of regulation, flexible specialization and networks’, Environment and Planning A,27: 731–57.

Murdoch, J. (1997) ‘Inhuman/nonhuman/human: actor network theory and the prospects for a non-dualistic and symmetrical perspective on nature and society’, Environment and Planning D: Society andSpace, 18: 731–56.

Murdoch, J. (1998) ‘The spaces of actor-network theory’, Geoforum, 29: 357–74.Murdoch, J. and Marsden,T. (1995) ‘The spatialization of politics: local and national actor-spaces in environ-

mental conflict’, Transactions, Institute of British Geographers, 20: 368–80.Murdoch, J. and Ward,N. (1997) ‘Governmentality and territoriality – the statistical manufacture of Britain’s

“national farm”’, Political Geography, 16: 307–24.Mutersbaugh, T. (1998) ‘Women’s work, men’s work: technology acquisition in a Oaxacan village

(Mexico)’, Environment and Planning D: Society and Space, 16: 439–58.Myers, G. (1996) ‘Naming and placing the other: power and the urban landscape in Zanzibar’, Tijdschrift

voor Economische en Sociale Geografie, 87: 237–46.Myers-Jones, H. and Brooker-Gross, S. (1994) ‘Newspapers as promotional strategists for regional defini-

tion’, in J. Gold and S.Ward (eds) Place Promotion: The Use of Publicity and Marketing to Sell Towns andRegions. Chichester:Wiley, 195–212.

REFERENCES

397


Nagar, R. (1997a) ‘Exploring methodological borderlands through oral narratives’, in J.P. Jones III et al.(eds) Thresholds in Feminist Geography. Oxford: Rowman & Littlefield, 203–24.

Nagar, R. (1997b) ‘The making of Hindu communal organizations, places and identities in postcolonial DarEs Salaam’, Environment and Planning D: Society and Space, 15: 707–30.

Nancy, J.-L. (1991) The Inoperative Community. Minneapolis, MN: University of Minnesota Press.Nash, C. (2000) ‘Performativity in practice: some recent work in cultural geography’, Progress in Human

Geography, 24: 653–64.Nash, C. (2003) ‘Cultural geography: anti-racist geographies’, Progress in Human Geography, 27: 637–48.Nast, H. (ed.) (1994) ‘Women in the field: critical feminist methodologies and theoretical perspectives’

(a series of papers), The Professional Geographer, 46: 54–102.Nast, H. (1998) ‘The body as “place”: reflexivity and fieldwork in Kano, Nigeria’, in H.J. Nast and S. Pile

(eds) Places through the Body. London: Routledge, 93–116.Nast, H. (2000) ‘Mapping the unconscious: racism and the Oedipal family’, Annals of Association of American

Geographers, 90: 215–55.Natter,W. and Jones, J.P. III (1993) ‘Signposts towards a poststructuralist geography’, in J.P. Jones III et al.

(eds) Postmodern Contentions: Epochs, Politics, Space. New York, NY: Guilford, 165–203.Natter,W. and Jones, J.P. III (1997) ‘Identity, space and other uncertainties’, in G. Benko and U. Strohmayer

(eds) Space and Social Theory: Interpreting Modernity and Postmodernity. Oxford: Blackwell, 141–61.Norcliffe, G. (1977) Inferential Statistics for Geographers. London: Hutchinson.Northedge,A. (1994) The Good Study Guide. Milton Keynes: Open University Press.

Oakley,A. (1979) From Here to Maternity: Becoming a Mother. Harmondsworth: Penguin Books.Oakley,A. (1984) Taking it Like a Woman. London: Fontana.Oakley,A. (1990) ‘Interviewing women’, in H. Roberts (ed.) Doing Feminist Research. London: Routledge &

Kegan Paul, 30–61.O’Brien, L. (1992) Introducing Quantitative Geography. London: Routledge.Ogborn, M., Blunt,A., Gruffudd, P., May, J. and Pinder, D. (eds) (forthcoming) Cultural Geography in Practice.

London:Arnold.Olsson, G. (1980) Birds in Egg/Eggs in Bird. London: Pion.Olsson, G. (1988) ‘The eye and the index finger: bodily means to cultural meaning’, in R. Golledge et al.

(eds) A Ground for Common Search. Santa Barbara, CA: Santa Barbara Geographical Press, 126–37.Olsson, G. (1991a) Lines of Power/Limits of Language. Minneapolis, MN: University of Minnesota Press.Olsson, G. (1991b) ‘Invisible maps’, Geografiska Annaler, 73B: 85–91.Olsson, G. (1994) ‘Heretic cartography’, Ecumene, 1: 213–34.Olsson, G. (1995) ‘Signs of persuasion’, in W. Natter et al. (eds) Objectivity and its Other. New York,

NY: Guilford, 21–32.Openshaw, S. (1984) The Modifiable Areal Unit Problem (CATMOG 38). Norwich: GeoAbstracts.Openshaw, S. (1991) ‘A view of the GIS crisis in geography, or, using GIS to put Humpty-Dumpty back

together again’, Environment and Planning A, 23: 621–8.Openshaw, S. (1996) ‘Fuzzy logic as a new paradigm for doing geography’, Environment and Planning A,

28: 761–8.Openshaw, S. (1998) ‘Towards a more computationally minded scientific human geography’, Environment

and Planning A, 30: 317–32.Openshaw, S. and Openshaw, C. (1997) Artificial Intelligence in Geography. Chichester:Wiley.Oppenheim,A. (1992) Questionnaire Design, Interviewing and Attitude Measurement. London: Pinter.Ord, J. and Getis,A. (1995) ‘Local spatial autocorrelation statistics: distributional issues and an application’,

Geographical Analysis, 27: 286–306.

Paassen, C.V. (1976) ‘Human geography in terms of existential anthropology’, Tijdschrift voor Economische enSociale Geografie, 67: 324–41.

Page, N. and Preston, P. (eds) (1993) The Literature of Place. London: Macmillan.Painter, J. (2002) ‘Governmentality and regional economic strategies’, in J. Hillier and E. Rooksby (eds)

Habitus: A Sense of Place.Aldershot:Ashgate, 115–39.


398


Palm, R. (1975) ‘Review of Ley, D. (1974) The Black Inner City as Frontier Outpost’, Annals of the Associationof American Geographers, 65: 572–3.

Palm, R. (1998) ‘Classics in human geography revisited: Ley 1974 The Black Inner City as Frontier Outpost,commentary 2’, Progress in Human Geography, 22: 77–8.

Panofsky, E. (1970) Meaning in the Visual Arts. Harmondsworth: Penguin Books.Parfitt, J. (1997) ‘Questionnaire design and sampling’, in R. Flowerdew and D. Martin (eds) Methods in

Human Geography. Harlow: Longman, 76–109.Parr, H. (1998a) ‘Mental health, ethnography and the body’, Area, 30: 28–37.Parr, H. (1998b) ‘The politics of methodology in “post-medical geography”: mental health research and the

interview’, Health and Place, 4: 341–53.Parr, H. (2000) ‘Interpreting the “hidden social geographies” of mental health: ethnographies of inclusion

and exclusion in semi-institutional spaces’, Health and Place, 6: 225–37.Parr, H. (2001) ‘Negotiating different ethnographic contexts and building geographical knowledges: empir-

ical examples from mental health research’, in M. Limb and C. Dwyer (eds) Qualitative Methodologies forGeographers: Issues and Debates. London:Arnold, 181–95.

Parr,H. and Philo,C. (2003) ‘Introducing psychoanalytic geographies’, Social and Cultural Geography, 4: 283–93.Pattie, C. and Johnston, R.J. (1993) ‘Surface change but underlying stability? The geography of the flow of

the vote in Great Britain, 1979–1992’, Area, 25: 257–66.Peck, J. (1999) ‘Grey geography?’, Transactions, Institute of British Geographers, 24: 131–5.Peet, R. (1977) ‘The development of radical geography in the United States’, Progress in Human Geography,

1: 240–63.Peet, R. (ed.) (1978) Radical Geography: Alternative Viewpoints on Contemporary Social Issues. London: Methuen.Peet, R. (1998) Modern Geographical Thought. Oxford: Blackwell.Penslar, R. (ed.) (1995) Research, Ethics, Cases and Materials. Bloomington, IN: Indiana University Press.Philbrick, A.K. (1957) ‘Principles of areal functional organisation in regional human geography’, Economic

Geography, 33: 299–336.Philip, L.J. (1998) ‘Combining quantitative and qualitative approaches to social research in human geogra-

phy – an impossible mixture?’, Environment and Planning A, 30: 261–76.Philo, C. (1984) ‘Reflections on Gunnar Olsson’s contribution to the discourse of contemporary human

geography’, Environment and Planning D: Society and Space, 2: 217–40.Philo, C. (ed.) (1991a) New Words, New Worlds.Aberystwyth: Cambrian Printers.Philo, C. (1991b) ‘Delimiting human geography: new social and cultural perspectives’, in C. Philo (ed.)

New Words, New Worlds.Aberystwyth: Cambrian Printers, 14–27.Philo, C. (1992) ‘Neglected rural geographies: a review’, Journal of Rural Studies, 8: 193–207.Philo, C. (1994) ‘Escaping Flatland: a book review essay inspired by Gunnar Olsson’s Lines of Power/Limits

of Language’, Environment and Planning D: Society and Space, 12: 229–52.Philo,C. (ed.) (1995) Off the Map: The Social Geography of Poverty in the UK.London:Child Poverty Action Group.Philo, C. (ed.) (1998) ‘Reconsidering quantitative geography’ (a series of papers), Environment and Planning

A, 30: 191–332.Philo, C. (2000) ‘More words, more worlds: reflections on the “cultural turn” and human geography’, in

I. Cook et al. (eds) Cultural Turns/Geographical Turns: Perspectives on Cultural Geography. London: PrenticeHall, 26–53.

Philo, C., Mitchell, R. and More,A. (1998) ‘Reconsidering quantitative geography: the things that count’,Environment and Planning A, 30: 191–202.

Pickles, J. (ed.) (1995) Ground Truth: The Social Implications of Geographic Information Systems. New York, NY:Guilford.

Pile, S. (1991) ‘Practising interpretative geography’, Transactions, Institute of British Geographers, 16: 458–69.Pile, S. (1996) The Body and the City: Psychoanalysis, Space and Subjectivity. London: Routledge.Pile, S. and Thrift, N. (eds) (1995) Mapping the Subject. London: Routledge.Pirie, G. (1984) ‘Review of Western, J. (1981) Outcast Cape Town’, Economic Geography, 60: 257–9.Platt, R.S. (1935) ‘Field approach to regions’, Annals of the Association of American Geographers, 25: 153–72.Plummer, P. and Sheppard, E. (2001) ‘Must emancipatory economic geography be qualitative?’ Antipode, 33:

194–9.

REFERENCES

399


Pocock, D. (1979) ‘The novelist’s image of the north’, Transactions, Institute of British Geographers, 4: 62–76.Pocock, D. (ed.) (1981a) Humanistic Geography and Literature. Beckenham: Croom Helm.Pocock, D. (1981b) ‘Imaginative literature and the geographer’, in D. Pocock (ed.) Humanistic Geography and

Literature. Beckenham: Croom Helm, 9–19.POD (1969) ‘Research’, in Pocket Oxford Dictionary (POD). Oxford: Clarendon Press, 703.Pool, I. (1957) ‘A critique of the twentieth anniversary issue’, Public Opinion Quarterly, 21: 190–8.Poole, R. (1988) ‘Deconstruction’, in A. Bullock and O. Stallybrass (eds) The Fontana Dictionary of Modern

Thought. London: Fontana, 205–6.Popkin, K. (1998) ‘Corporate imaginations of rural idyll and community.’ Unpublished PhD thesis, School

of Geographical Sciences, University of Bristol.Powell, R.C. (2002) ‘The Sirens’ voices? Field practices and dialogue in geography’, Area, 34: 261–72.Pratt, G. (1989) ‘Quantitative techniques and humanistic-historical materialist perspectives’, in A. Kobayashi

and S. Mackenzie (eds) Remaking Human Geography. London: Unwin Hyman, 101–5.Pratt, G. (1992) ‘Spatial metaphors and speaking positions’, Environment and Planning D: Society and Space,

10: 241–4.Pratt, G. (1994) ‘Poststructuralism (including deconstruction)’, in R.J. Johnston et al. (eds) The Dictionary of

Human Geography (3rd edn). Oxford: Blackwell, 468–9.Pratt, G. (2000a) ‘Focus group’, in R.J. Johnston et al. (eds) The Dictionary of Human Geography (4th edn).

Oxford: Blackwell, 272.Pratt, G. (2000b) ‘Post-structuralism’, in R.J. Johnston et al. (eds) The Dictionary of Human Geography

(4th edn). Oxford: Blackwell, 625–6.Pred,A. (1995) Recognizing European Modernities. London: Routledge.Preston, P. and Simpson-Houseley, P. (eds) (1994) Writing the City: Eden, Babylon and the New Jerusalem.

London: Routledge.Prigogine, I. (1984) Order out of Chaos: Man’s New Dialogue with Nature. London: Heinemann.Prince, H. (1980) ‘Review of Humanistic Geography: Prospects and Problems, ed. Ley & Samuels’, Annals of the

Association of American Geographers, 70: 294–6.Proctor, J.D. (1998) ‘Ethics in geography: giving moral form to the geographical imagination’, Area,

30: 8–18.Pryke, M. (2000) ‘Tracing economic rhythms through visual and audio montage’ (available at

http://www.open.ac.uk/socialsciences/geography/research/bellin/).Pryke, M. (2002) ‘The white noise of capitalism: audio and visual montage and sensing economic change’,

Cultural Geographies, 9: 472–7.Punch, M. (1986) The Politics and Ethics of Fieldwork. Beverly Hills, CA: Sage.Punch, S. (2001) ‘Multiple methods and research relations with children in rural Bolivia’, in M. Limb and

C. Dwyer (eds) Qualitative Methodologies for Geographers: Issues and Debates. London:Arnold, 165–80.

Quiller-Couch, A.T. (1916) On the Art of Writing: Lectures Delivered in the University of Cambridge,1913–1914. Cambridge: Cambridge University Press.

Quinton, A. (1988) ‘Dialectic’, in A. Bullock and O. Stallybrass (eds) The Fontana Dictionary of ModernThought. London: Fontana, 225.

Rabinow, P. (1989) French Modern: Norms and Forms of the Social Environment. Cambridge, MA: MIT Press.Raco, M. (2003) ‘Governmentality, subject-building, and the discourses and practices of devolution in the

UK’, Transactions, Institute of British Geographers, 28: 75–95.Radcliffe, S. (1994) ‘(Re)presenting post-colonial women: authority, difference and feminisms’,Area, 26: 25–32.Reah, D. (1998) The Language of Newspapers. London: Routledge.Reichert, D. (1992) ‘On boundaries’, Environment and Planning D: Society and Space, 10: 87–98.Relph, E. (1970) ‘An inquiry into the relations between phenomenology and geography’, The Canadian

Geographer, 14: 193–201.Relph, E. (1976) Place and Placelessness. London: Pion.Relph, E. (1980) ‘Review of Humanistic Geography: Prospects and Problems, ed. Ley & Samuels’, The

Geographical Review, 70: 115–17.


400


Revill, G. (1995) ‘Hiawatha and pan-Africanism: Samuel Coleridge-Taylor (1875–1912), a black composerin suburban London’, Ecumene, 2: 247–66.

Richards, I. (1929) Practical Criticism: A Study of Literary Judgement. London: Kegan Paul, Trench,Trübner & Co.

Richards, P. (1986) ‘Risk’, in H. Becker (ed.) Writing for Social Scientists. Chicago, IL: University of ChicagoPress, 108–20.

Richardson, L. (1988) ‘The collective story: postmodernism and the writing of sociology’, Sociological Focus,21: 199–208.

Richardson, L. (1990) Writing Strategies: Reaching Diverse Audiences. Newbury Park, CA: Sage.Richardson, M. (1981) ‘Commentary on the “superorganic” in American cultural geography’, Annals of the

Association of American Geographers, 71: 284–7.Roberts, H. (ed.) (1990) Doing Feminist Research. London: Routledge & Kegan Paul.Robinson, G.M. (1998) Methods and Techniques in Human Geography. Chichester:Wiley.Robson, E. (1994) ‘From teacher to taxi driver: reflections on research roles in developing areas’, in

E. Robson and K.Willis (eds) Postgraduate Fieldwork in Developing Areas: A Rough Guide. Developing AreasResearch Group Monograph 9. London: RGS with the IBG, 36–59.

Robson, E. and Willis, K. (eds) (1994a) Postgraduate Fieldwork in Developing Areas: A Rough Guide. DevelopingAreas Research Group Monograph 9. London: RGS with the IBG.

Robson, E. and Willis, K. (1994b) ‘Introduction’, in E. Robson and K.Willis (eds) Postgraduate Fieldwork inDeveloping Areas: A Rough Guide. Developing Areas Research Group Monograph 9. London: RGS with theIBG, 1–4.

Rodaway, P. (1994) Sensuous Geographies: Body, Sense and Place. London: Routledge.Rogers,A.,Viles, H. and Goudie,A. (eds) (1992) The Student’s Companion to Geography. Oxford: Blackwell.Rorty, R. (1979) Philosophy and the Mirror of Nature. Oxford: Blackwell.Rorty, R. (1987) ‘Method, social science and social hope’, in M. Gibbons (ed.) Interpreting Politics. Oxford:

Blackwell, 241–59.Rosaldo, R. (1986) Culture and Truth: The Remaking of Social Analysis. Boston, MA: Beacon.Rosaldo, R. (1993) Culture and Truth: The Remaking of Social Analysis. London: Routledge.Rose, C. (1981) ‘Wilhelm Dilthey’s philosophy of historical understanding: a neglected heritage of con-

temporary humanistic geography’, in D. Stoddart (ed.) Geography, Ideology and Social Concern. Oxford:Blackwell, 99–113.

Rose,D. and Sullivan,O. (1996) Introducing Data Analysis for Social Scientists.Buckingham:Open University Press.Rose, G. (1990) ‘Imagining Poplar in the 1920s: contested concepts of community’, Journal of Historical

Geography, 16: 425–37.Rose, G. (1991) ‘On being ambivalent: women and feminisms in geography’, in C. Philo (comp.)

New Words, New Worlds: Reconceptualising Social and Cultural Geography. Lampeter: Social and CulturalGeography Study Group, 156–63.

Rose,G. (1992) ‘Geography as a science of observation: the landscape, the gaze and masculinity’, in F.Driverand G. Rose (eds) Nature and Science: Essays in the History of Geographical Knowledge. London: HistoricalGeography Research Group, 8–18.

Rose, G. (1993a) ‘Some notes towards thinking about the spaces of the future’, in J. Bird et al. (eds) Mappingthe Futures: Local Cultures, Global Change. London: Routledge, 70–83.

Rose, G. (1993b) Feminism and Geography: The Limits of Geographical Knowledge. Cambridge: Polity Press.Rose, G. (1994) ‘The cultural politics of place: local representation and oppositional discourse in films’,

Transactions, Institute of British Geographers, 19: 46–60.Rose, G. (1995) ‘Distance, surface, elsewhere: a feminist critique of the space of phallocentric knowledge’,

Environment and Planning D: Society and Space, 13: 761–81.Rose, G. (1997a) ‘Situating knowledges: positionality, reflexivities and other tactics’, Progress in Human

Geography, 21: 305–20.Rose, G. (1997b) ‘Performing inoperative community: the space and the resistances of some community

arts projects’, in S. Pile (ed.) Geographies of Resistance. London: Routledge, 184–202.Rose, G. (2000) ‘Psychoanalytic theory, geography and’, in R.J. Johnston et al. (eds) The Dictionary of Human

Geography (4th edn). Oxford: Blackwell, 653–5.

REFERENCES

401


Rose, G. (2001) Visual Methodologies: An Introduction to the Interpretation of Visual Materials. London: Sage.Rossi, P.,Wright, J. and Anderson,A. (1983) Handbook of Survey Research. San Diego, CA:Academic Press.Roszak,T. (1994) The Cult of Information. Berkeley, CA: University of California Press.Roth, J. (1966) ‘Hired hand research’, The American Sociologist, 1: 190–6.Round, J.H. (1895) Feudal England: Historical Studies on the XIth and XIIth Centuries. London: Swan

Sonnenschein.Routledge, P. (1996) ‘The third space as critical engagement’, Antipode, 28: 398–419.Rowles, G.D. (1978a) Prisoners of Space? Exploring the Geographical Experience of Older People. Boulder,

CO:Westview Press.Rowles, G.D. (1978b) ‘Reflections on experiential fieldwork’, in D. Ley and M.S. Samuels (eds) Humanistic

Geography: Prospects and Problems. London: Croom Helm, 173–93.Rowles, G.D. (1980) ‘Towards a geography of growing old’, in A. Buttimer and D. Seamon (eds) The Human

Experience of Space and Place. London: Croom Helm, 55–72.Rowles, G.D. (1981a) ‘Geographical perspectives on human development’, Journal of Human Development,

24: 67–76.Rowles, G.D. (1981b) ‘The surveillance zone as a meaningful space for the aged’, The Gerontologist,

21: 304–11.Rowles, G.D. (1983) ‘Place and personal identity: observations from Appalachia’, Journal of Environmental

Psychology, 3: 299–313.Rowles, G.D. (1987) ‘A place called home’, in L. Carstensen and B. Edelsten (eds) The Handbook of Clinical

Gerontology. New York, NY: Pergamon, 335–53.Rundstrom, R.A. and Kenzer, M.S. (1989) ‘The decline of fieldwork in human geography’, The Professional

Geographer, 41: 294–303.Ryan, J.R. (1997) Picturing Empire: Photography and the Visualisation of the British Empire. London: Reaktion

Books.

Saarinen, T. (1966) Perception of Drought Hazard on the Great Plains. Department of Geography ResearchPaper 106. Chicago, IL: University of Chicago, Department of Geography.

Sack, R.D. (1980) Conceptions of Space in Social Thought. London: Macmillan.Sadler, D. (1997) ‘The role of supply chain management strategies in the “Europeanization” of the auto-

mobile production system’, in R. Lee and J.Wills (eds) Geographies of Economies. London:Arnold, 311–20.Said, E. (1978) Orientalism. London: Penguin Books.Said, E. (1993) Culture and Imperialism. New York, NY:Alfred A. Knopf.Salter, C.L. (1969) ‘The bicycle as a field aid’, The Professional Geographer, 21: 360–2.Salter, C.L. and Meserve, P. (1991) ‘Life lists and the education of the geographer’, The Professional

Geographer, 43: 520–5.Saltmarsh, R. (2001) ‘A journey into autobiography: a coal miner’s daughter’, in P. Moss (ed.) Placing

Autobiography in Geography. Syracuse, NY: Syracuse University Press, 138–48.Saltmarsh, R. (2002) ‘From mining town to university: a class-cultural experience of an education in Wales.’

Paper presented to the Association of American Geographers’ annual conference, Los Angeles.Sandberg, L.A. and Marsh, J.S. (1988) ‘Focus: literary landscapes – geography and literature’, The Canadian

Geographer/Le Géographe canadien, 32: 266–76.Sanjek, R. (1990) ‘A vocabulary for fieldnotes’, in R. Sanjek (ed.) Fieldnotes. Ithaca, NY: Cornell University

Press, 92–121.Sauer, C.O. (1924) ‘The survey method in geography and its objectives’, Annals of the Association of American

Geographers, 14: 17–33.Sauer,C.O. (1956) ‘The education of a geographer’,Annals of the Association of American Geographers, 46:287–99.Sayer,A. (1984) Method in Social Science: A Realist Approach (1st edn). London: Hutchinson.Sayer, A. (1989) ‘The “new” regional geography and problems of narrative’, Environment and Planning

D: Society and Space, 7: 253–76.Sayer,A. (1992) Method in Social Science: A Realist Approach (2nd edn). London: Routledge.Sayer, A. (1997) ‘Realism and geography’, in T. Barnes and D. Gregory (eds) Reading Human Geography.

London:Arnold, 112–24.


402


Sayer, A. and Storper, M. (1997) ‘Ethics unbound: for a normative turn in social theory’, Environment andPlanning D: Society and Space, 15: 1–17.

Schaeffer, F.K. (1953) ‘Exceptionalism in geography: a methodological examination’, Annals of theAssociation of American Geographers, 43: 226–49.

Scott, A.J. (1984) ‘Territorial reproduction and transformation in a local labour market: the animated filmworkers of Los Angeles’, Environment and Planning D: Society and Space, 2: 277–307.

Scott, A.J. and Angel, D.P. (1987) ‘The US semiconductor industry: a locational analysis’, Environment andPlanning A, 20: 1047–67.

Scott, J. (1990) A Matter of Record. Cambridge: Polity Press.Seamon, D. (1976) ‘Phenomenological examination of imaginative literature: a commentary’, in

G.T. Moore and R.G. Golledge (eds) Environmental Knowing. Stroudsburg, PA: Dowden, Hutchinson &Ross, 286–90.

Seamon, D. (1979) A Geography of the Lifeworld: Movement, Rest and Encounter. London: Croom Helm.Selkirk, K.E. (1982) Pattern and Place: An Introduction to Mathematical Geography. Cambridge: Cambridge

University Press.Selwyn,T. (ed.) (1996) The Tourist Image: Myths and Myth-making in Tourism. Chichester: Wiley.Semple, R.K. and Green, M.B. (1984) ‘Classification in human geography’, in G.L. Gaile and C.J.Willmott

(eds) Spatial Statistics and Models. Dordrecht: Reidel, 55–79.Serres, M. and Latour, B. (1995) Conversations on Science, Culture, Time. Ann Arbor, MI: University of

Michigan Press.Sharp, J.P., Routledge, P., Philo, C. and Paddison, R. (eds) (2000) Entanglements of Power: Geographies of

Domination/Resistance. London: Routledge.Shelley, F.M., Archer, J.C., Davidson, F.M. and Brunn, S.D. (1996) Political Geography of the United States.

New York, NY: Guilford.Sheppard, E. and Barnes,T. (1990) The Capitalist Space Economy: Geographical Analysis after Ricardo, Marx and

Sraffa. London: Unwin Hyman.Shields, R. (1996) Cultures of the Internet. London: Sage.Shortridge, J.R. (1991) ‘The concept of the place-defining novel in American popular culture’, The

Professional Geographer, 43: 280–91.Shurmer-Smith, P. (1998a) ‘Classics in human geography revisited: Duncan, J.S. “The superorganic in

American cultural geography”, commentary 1’, Progress in Human Geography, 22: 567–9.Shurmer-Smith, P. (1998b) ‘Becoming a memsahib: working with the Indian administrative service’,

Environment and Planning A, 30: 2163–79.Shurmer-Smith, P. (2000) ‘Review of Cook et al. (eds) (2000) Cultural Turns/Geographical Turns: Perspectives

on Cultural Geography’, Transactions, Institute of British Geographers, 25: 524–5.Shurmer-Smith, P. (2002a) ‘Introduction’, in P. Shurmer-Smith (ed.) Doing Cultural Geography. London:

Sage, 1–7.Shurmer-Smith, P. (2002b) ‘Methods and methodology’, in P. Shurmer-Smith (ed.) Doing Cultural

Geography. London: Sage, 95–8.Shurmer-Smith, P. (2002c) ‘Postcolonial geographies’, in P. Shurmer-Smith (ed.) Doing Cultural Geographies.

London: Sage, 67–77.Sibley, D. (1981a) Outsiders in Urban Societies. Oxford: Blackwell.Sibley, D. (1981b) ‘The notion of order in spatial analysis’, The Professional Geographer, 33: 1–5.Sibley, D. (1985) ‘Travelling people in England: a regional comparison’, Regional Studies, 19: 139–47.Sibley, D. (1995) Geographies of Exclusion: Society and Difference in the West. London: Routledge.Sibley, D. (1998) ‘Sensations and spatial science: gratification and anxiety in the production of ordered land-

scapes’, Environment and Planning A, 30: 235–46.Sibley, D. (n.d.) Spatial Applications of Exploratory Data Analysis (CATMOG 49). Norwich: Geo Books.Sidaway, J.D. (1992) ‘In other worlds: on the politics of research by “First World” geographers in the “Third

World”’, Area, 24: 403–8.Sidaway, J.D. (1997) ‘The production of British geography’, Transactions, Institute of British Geographers,

22: 488–504.Silk, J. (1979) Statistical Concepts in Geography. London: George Allen & Unwin.

REFERENCES

403


Silk, J. (1984) ‘Beyond geography and literature’, Environment and Planning D: Society and Space, 2: 151–78.Silverman, D. (1993) Interpreting Qualitative Data: Methods for Analysing Talk, Text and Interaction. London:

Sage.Simon, D. (1999) ‘Classics in human geography revisited:Western, J. (1981) Outcast Cape Town’, Progress in

Human Geography, 23: 423–5.Simpson, C.A. (1945) ‘A venture in field geography’, Geography, XXX: 35–44.Simpson, S. and Dorling, D. (1999) ‘Conclusion: statistics and “the truth”’, in D. Dorling and S. Simpson

(eds) Statistics in Society. London:Arnold, 414–20.Skelton, T. (2001) ‘Cross-cultural research: issues of power, positionality and “race”’, in M. Limb and

C. Dwyer (eds) Qualitative Methodologies for Geographers: Issues and Debates. London:Arnold, 87–100.Sloman, A. (1988) ‘Methodology’, in A. Bullock and O. Stallybrass (eds) The Fontana Dictionary of Modern

Thought. London: Fontana, 525–6.Smailes,A.E. (1964) ‘Urban survey’, in R. Clayton (ed.) The Geography of Greater London: A Source Book for

Teacher and Student. London: George Philip & Son, 202–21.Smith, C. (1982) ‘Review of The Human Experience of Space and Place, ed. Buttimer & Seamon’, The

Geographical Review, 72: 107–8.Smith, D.M. (1975) Patterns in Human Geography. Harmondsworth: Penguin Books.Smith, D.M. (1977) Human Geography: A Welfare Approach. London: Edward Arnold.Smith, D.M. (1979) Where the Grass is Greener: Living in an Uneven World. Harmondsworth: Penguin Books.Smith, D.M. (1988a) ‘A welfare approach to human geography’, in J. Eyles (ed.) Research in Human

Geography: Introductions and Investigations. Oxford: Blackwell, 139–54.Smith, D.M. (1988b) ‘Towards an interpretative human geography’, in J. Eyles and D.M. Smith (eds)

Qualitative Methods in Human Geography. Cambridge: Polity Press, 255–67.Smith, D.M. (1994) Geography and Social Justice. Oxford: Blackwell.Smith, D.M. (1995a) ‘Moral teaching in geography’, Journal of Geography in Higher Education, 19: 271–83.Smith, D.M. (1995b) ‘Against differential research funding’, Area, 27: 79–83.Smith, D.M. (1997) ‘Geography and ethics: a moral turn?’, Progress in Human Geography, 21: 583–90.Smith, D.M. (1998) ‘Geography and moral philosophy: some common ground’, Ethics, Place and

Environment, 1: 7–34.Smith, D.M. (2000) ‘Moral progress in human geography: transcending the place of good fortune’, Progress

in Human Geography, 24: 1–18.Smith, D.M. (2004) ‘Moral geographies’, in P. Cloke et al. (eds) Envisioning Human Geographies. London:

Arnold, 195–209.Smith, N. (1992) ‘History and philosophy of geography: real wars, theory wars’, Progress in Human

Geography, 16: 257–71.Smith, N. (1993) ‘Homeless/global: scaling places’, in J. Bird et al. (eds) Mapping the Futures: Local Cultures,

Global Change. London: Routledge, 87–119.Smith, N. (1996) The New Urban Frontier. London: Routledge.Smith, S.J. (1981) ‘Humanistic method in contemporary social geography’, Area, 13.Smith,S.J. (1984) ‘Practising humanistic geography’,Annals of the Association of American Geographers, 74:353–74.Smith, S.J. (1988) ‘Constructing local knowledge: the analysis of the self in everyday life’, in J. Eyles and

D.M. Smith (eds) Qualitative Methods in Human Geography. Cambridge: Polity Press, 17–38.Smith, S.J. (1994) ‘Soundscape’, Area, 26: 232–40.Smith, S.J. (2001) ‘Doing qualitative research: from interpretation to action’, in M. Limb and C. Dwyer (eds)

Qualitative Methodologies for Geographers. London:Arnold, 23–40.Soja, E. (1989) Postmodern Geographies. London: Verso.Sparke, M. (1994) ‘Escaping the herbarium: a critique of Gunnar Olsson’s “Chiasm of thought-and-

action”’, Environment and Planning D: Society and Space, 12: 207–20.Sparke, M. (1996) ‘Displacing the field in fieldwork: masculinity, metaphor and space’, in N. Duncan (ed.)

BodySpace: Destabilising Geographies of Gender and Sexuality. London: Routledge, 212–33.Stacey, J. (1988) ‘Can there be a feminist ethnography?’, Women’s Studies International Forum, 11: 21–7.Stacey, J. (1991) ‘Can there be a feminist ethnography?’, in S. Gluck and D. Patai (eds) Women’s Words: The

Feminist Practice of Oral History. London: Routledge, 111–18.


404


Stacey, J. (1997) ‘Can there be a feminist ethnography?’, in L. McDowell and J. Sharp (eds) Space, Gender,Knowledge: Feminist Readings. London:Arnold, 115–23.

Steinberg, J. (1985) ‘Verbal anaemia’, New Society, 19 July: 90.Stewart, D. and Shamdasani, P.N. (1990) Focus Groups: Theory and Practice. Beverly Hills, CA: Sage.Stoddart, D.R. (1986) On Geography and its History. Oxford: Blackwell.Stoddart, D.R. (1987) ‘To claim the high ground: geography for the end of the century’, Transactions,

Institute of British Geographers, 12: 327–36.Stoddart, D.R. (1991) ‘Do we need a feminist historiography of geography – and if we do, what should it

be?’, Transactions, Institute of British Geographers, 16: 484–7.Stoll, C. (1995) Silicon Snake Oil: Second Thoughts on the Information Highway. London: Doubleday.Storper, M. (1994) ‘The transition to flexible specialisation in the US film industry: external economies,

the division of labour and the crossing of industrial divides’, in A. Amin (ed.) Post-Fordism: A Reader.Oxford: Blackwell, 195–226.

Storper, M. and Christopherson, S. (1985) The Changing Location and Organization of the Motion PictureIndustry. Los Angeles, CA: UCLA Graduate School of Architecture and Urban Planning.

Strohmayer, U. and Hannah, M. (1992) ‘Domesticating postmodernism’, Antipode, 24: 29–55.Stutz, F. (1979) ‘Review of Rowles, G. (1978) Prisoners of Space?’, The Geographical Review, 69: 473.Sudman, S. and Bradburn, R. (1982) Asking Questions. San Francisco, CA: Jossey-Bass.

Taylor, P.J. (1990) ‘Editorial comment: GKS’, Political Geography Quarterly, 9: 211–12.Taylor, P.J. and Johnston, R.J. (1979) Geography of Elections. London: Penguin Books.Taylor, P.J. and Johnston, R.J. (1995) ‘Geographic information systems and geography’, in J. Pickles (ed.)

Ground Truth. London: Guilford, 51–7.Thompson. K. (ed.) (1997) Media and Cultural Representation. London: Sage.Thrall, G.I. (1985) ‘Scientific geography: report on “scientific geography” conference, Georgia, USA, March

1995’, Area, 17: 254.Thrift, N. (1983a) ‘Literature, the production of culture and the politics of place’, Antipode, 15: 12–24.Thrift, N. (1983b) ‘On the determination of social action in space and time’, Environment and Planning

D: Society and Space, 1: 23–57.Thrift, N. (1985) ‘Flies and germs: a geography of knowledge’, in D. Gregory and J. Urry (eds) Social

Relations and Spatial Structures. Basingstoke: Macmillan, 366–403.Thrift, N. (1991) ‘Over-wordy worlds? Thoughts and worries’, in C. Philo (ed.) New Words, New Worlds.

Aberystwyth: Cambrian Printers, 144–8.Thrift, N. (1995) ‘Classics in human geography revisited:Thrift, N.“On the determination of social action

in space and time”: author’s response’, Progress in Human Geography, 19: 528–30.Thrift, N. (1996) Spatial Formations. London: Sage.Thrift, N. (1997) ‘Cities without modernity, cities with magic’, Scottish Geographical Magazine, 113: 138–49.Thrift, N. (1999) ‘Steps towards an ecology of place’, in D. Massey et al. (eds) Human Geography Today.

Cambridge: Polity Press, 295–322.Thrift, N. (2000a) ‘Dead or alive?’, in I. Cook et al. (eds) Cultural Turns/Geographical Turns: Perspectives on

Cultural Geography. London: Prentice Hall, 1–6.Thrift, N. (2000b) ‘Non-representational theory’, in R.J. Johnston et al. (eds) The Dictionary of Human

Geography (4th edn). Oxford: Blackwell, 556.Thrift, N. (2000c) ‘Entanglements of power: shadows?’, in J. Sharp et al. (eds) Entanglements of Power.

London: Routledge, 262–78.Thrift, N. (2000d) ‘Afterwords’, Environment and Planning D: Society and Space, 18: 213–55.Thrift, N. and Dewsbury, J.-D. (2000) ‘Dead geographies and how to make them live’, Environment and

Planning D: Society and Space, 18: 411–32.Tickell,A. (1995) ‘Reflections on activism and the academy’, Environment and Planning D: Society and Space,

13: 235–7.Tickell,A. (1999) ‘On getting inside the project’, Ethics, Place and Environment, 2: 234–9.Tillyard, E.M.W. (1972) The Elizabethan World Picture. London: Penguin Books.Tivers, J. (1978) ‘How the other half lives: the geographical study of women’, Area, 10: 302–6.

REFERENCES

405


Tomaselli, K.G. (1988) ‘The geography of popular memory in post-colonial South Africa: a study ofAfrikaans cinema’, in J. Eyles and D.M. Smith (eds) Qualitative Methods in Human Geography. Cambridge:Polity Press, 136–55.

Tozer, H.F. (1897) A History of Ancient Geography. New York, NY: Biblo & Tanner.Tuan,Y.-F. (1972) ‘Structuralism, existentialism and environmental perception’, Environment and Behaviour,

3: 319–31.Tuan,Y.-F. (1976a) ‘Humanistic geography’, Annals of the Association of American Geographers, 66: 266–76.Tuan,Y.-F. (1976b) ‘Literature, experience, and environmental knowing’, in G.T. Moore and R.G. Golledge

(eds) Environmental Knowing. Stroudsburg, PA: Dowden, Hutchinson & Ross, 260–72.Turner, T. (1991) ‘Social dynamics of video media in an indigenous society: the cultural meaning and

personal politics of video-making’, Visual Anthropology Review, 7: 68–76.Turner,T. (1992) ‘Defiant images: the Kapayo appropriation of video’, Anthropology Today, 8: 5–16.Twyman, C., Morrison, J. and Sporton, D. (1999) ‘The final fifth: autobiography, reflexivity and interpreta-

tion in cross-cultural research’, Area, 31: 313–25.

Unwin, D. (1981) Introductory Spatial Analysis. London: Methuen.Upton, G. (1999) ‘Illuminating social statistics’, in D. Dorling and S. Simpson (eds) Statistics in Society.

London:Arnold, 400–13.Urry, J. (1990) The Tourist Gaze: Leisure and Travel in Contemporary Societies. London: Sage.US National Research Council (1997) Rediscovering Geography: New Relevance for Science and Society.

Washington, DC: National Research Council.

Valentine, G. (1989) ‘The geography of women’s fear’, Area, 21: 385–90.Valentine, G. (1995) ‘Creating transgressive space: the music of kd lang’, Transactions, Institute of British

Geographers, 20: 474–86.Van Hoven, B. (2003) ‘Using CAQDAS in qualitative research’, in N.J. Clifford and G.Valentine (eds) Key

Methods in Geography. London: Sage, 461–76.

Walford, G. (ed.) (1994) Researching the Powerful. London: UCL Press.Walford, N. (1995) Geographical Data Analysis. Chichester:Wiley.Wallis, R. and Malm, K. (1987) ‘The international music industry and transcultural communication’, in

J. Lull (ed.) Popular Music and Communication. Newbury Park, CA: Sage.Ward, D. (1978) Cities and Immigrants: A Geography of Change in Nineteenth Century America. New York,

NY: Oxford University Press.Warde,A. (1985) ‘Spatial change, politics and the division of labour’, in D. Gregory and J. Urry (eds) Social

Relations and Spatial Structures. London: Macmillan, 190–212.Warnke, G. (1987) Gadamer: Hermeneutics, Tradition and Reason. Cambridge: Polity Press.Warren, C. et al. (2000) ‘Writing the other, inscribing the self ’, Qualitative Sociology, 23: 183–99.Watts, M. (1999a) ‘Commodities’, in P. Cloke et al. (eds) Introducing Human Geographies. London: Arnold,

305–15.Watts, M. (1999b) ‘Privatization and governance’, Ethics, Place and Environment, 2: 256–7.Watts, M. (2001) ‘1968 and all that…’, Progress in Human Geography, 25: 157–88.Webb, S. and Webb, B. (1932) Methods of Social Study. London: Longmans Green & Co.Weber, M. (1962) Basic Concepts in Sociology. London: Peter Owen.Weber, M. (1978) Economy and Society: An Outline of Interpretative Sociology. Berkeley, CA: University of

California Press.Weitzman, E.A. (2000) ‘Software and qualitative research’, in N.K. Denzin and Y.S. Lincoln (eds) Handbook

of Qualitative Research (2nd edn). London: Sage, 803–20.Weitzman, E.A. and Miles, M.B. (1995) Computer Programs for Qualitative Data Analysis: A Software

Sourcebook.Thousand Oaks, CA: Sage.Wells, L. (ed.) (1997) Photography: A Critical Introduction. London: Routledge.West, R.C. (1954) ‘The term bayou in the United States’, Annals of the Association of American Geographers,

44: 219–21.


406


Western, J. (1978) ‘Knowing one’s place: “the coloured people” and the Group Areas Act’, in D. Ley andM. Samuels (eds) Humanistic Geography: Prospects and Problems. London: Croom Helm, 297–318.

Western, J. (1981a) Outcast Cape Town. Minneapolis, MN: University of Minnesota Press.Western, J. (1981b) ‘Review of The Human Experience of Space and Place, ed. Buttimer and Seamon’, Economic

Geography, 57: 275–8.Western, J. (1999) ‘Classics in human geography revisited:Western, J. (1981) Outcast Cape Town’, Progress in

Human Geography, 23: 425–7.Whatmore, S.J. (2002) Hybrid Geographies: Natures, Cultures and Spaces. London: Sage.Whatmore, S.J. and Thorne, L. (1997) ‘Nourishing networks: alternative geographies of food’, in D. Goodman

and M.Watts (eds) Globalising Food: Agrarian Questions and Global Restructuring. London: Routledge, 287–304.Wheeler, K.S. and Harding, M. (eds) (1967) Geographical Fieldwork. London: Bland Education.Wheeler, P.T. (1967) ‘The development and role of the Geographical Field Group’, East Midland Geographer,

4: 185–91.Whitehead, F. (1987) ‘The GRO use of social surveys’, Population Trends, 48: 45–54.Whittlesey, D.S. (1927) ‘Devices for accumulating geographic data in the field’, Annals of the Association of

American Geographers, 17: 72–8.Widdowfield, R. (2000) ‘The place of emotions in academic research’, Area, 32: 199–208.Williams, R. (1979) Politics and Letters: Interviews with New Left Review. London: New Left Books.Williams, R. (1981) Culture. London: Fontana.Williams, R. (1986) Intermediate Statistics for Geographers and Earth Scientists. London: Macmillan.Willis, P. (1977) Learning to Labour: How Working Class Kids Get Working Class Jobs. London: Gower.Willis, P. and Trondman, M. (2000) ‘A manifesto for Ethnography’, Ethnography, 1: 5–16.Wilson,A.G. (1974) Urban and Regional Models in Geography and Planning. London:Wiley.Wilson,A.G. (1981a) Catastrophe Theory and Bifurcation: Applications to Urban and Regional Systems. London:

Croom Helm.Wilson,A.G. (1981b) Geography and the Environment: Systems Analytical Methods. Chichester:Wiley.Wilson, A.G. (1986) ‘System’, in R.J. Johnston et al. (eds) The Dictionary of Human Geography (2nd edn).

Oxford: Blackwell, 476–7.Wilson,A.G. (1992) The Culture of Nature. Oxford: Blackwell.Wilson,A.G. and Bennett,R. (1985) Mathematical Methods in Human Geography and Planning. Chichester:Wiley.Wilton, R. (1998) ‘The constitution of difference: space and psyche in landscapes of exclusion’, Geoforum,

29: 173–85.Winchester, H. (2000) ‘Qualitative research and its place in human geography’, in I. Hay (ed.) Qualitative

Research Methods in Human Geography. South Melbourne: Oxford University Press, 1–22.Winchester, H. and Costello, L. (1995) ‘Living in the street: social organisation and gender relations of

Australian street kids’, Environment and Planning D: Society and Space, 13: 329–48.Wolcott, H. (1990) Writing up Qualitative Research. London: Sage.Wolfinger, N. (2002) ‘On writing fieldnotes: collection strategies and background expectancies’, Qualitative

Research, 2: 85–95.Women and Geography Study Group (WGSG) (1984) Geography and Gender: An Introduction to Feminist

Geography. London: Hutchinson.Women and Geography Study Group (WGSG) (1997) Feminist Geographies: Explorations in Diversity and

Difference. London: Longman.Wood, D. (1993) The Power of Maps. London: Routledge.Woods, M. (1998) ‘Researching rural conflicts: hunting, local politics and actor-networks’, Journal of Rural

Studies, 14: 321–40.Woods, M. (1999) ‘Power, professors and protest: reflections on the RGS–IBG debate and activism in 1990s

academe’, Ethics, Place and Environment, 2: 228–33.Woodward, R. (1996) ‘“Deprivation” and “the rural”: an investigation into contradictory discourses’, Journal

of Rural Studies, 12: 55–67.Wooldridge, S.W. (1955) ‘The status of geography and the role of fieldwork’, Geography, XL: 73–83.Woolf, P. (1988) ‘Symbol of the Second Empire: cultural politics and the Paris Opera House’, in D. Cosgrove

and S. Daniels (eds) The Iconography of Landscape. Cambridge: Cambridge University Press, 214–35.

REFERENCES

407


Wright, G. (1987) Gadamer: Hermeneutics, Tradition and Reason. Cambridge: Polity Press.Wright, J.K. (1947) ‘Terrae incognitae: the place of the imagination in geography’, Annals of the Association

of American Geographers, 37: 1–15.Wright, M. (1997) ‘Crossing the factory frontier: place and power in the Mexican maquiladora’, Antipode,

29: 278–302.Wrigley, N. (1976) Introduction to the Use of LOGIT Models in Geography (CATMOG 10). Norwich:

GeoAbstracts.Wrigley, N. (1985) Categorical Data Analysis for Geographers and Environmental Scientists. London: Longman.Wrigley, N. and Bennett, R. (eds) (1981) Quantitative Geography: A British View. London: Routledge &

Kegan Paul.

Yates, E.M. and Robertson, M.F. (1968) ‘Geographical field studies’, Geography, LIII: 53–67.Youngman, M. (1986) Analysing Questionnaires. Nottingham: School of Education, University of

Nottingham.


408


Index

Abler, Ronald 25abstraction 291–5access negotiation 15Acker, 370ACORN classification 273, 274action-based research 26actor-network theory (ANT) 95,

189–91, 301–4advertising 78–84advocacy geography 26Agger, B. 361Agnew, J. 186Aitken, Stuart 89, 90, 110, 111, 324alternative formations 99Amaral, D.J. 26Amin,A. 360–1Anderson, James 19Anderson, K. 183animation 114Annual Abstract of Statistics 44, 51–2architecture 111–13archives, ESRC 65art 109

Gainsborough 120–2, 326landscape 12, 13, 109, 120–2

articles 65artificial intelligence 278the artisan 310, 312–17artisanal marketization 97–8artistic movements 99Atkinson, M. 70, 156, 165attitude scaling 138, 142–3Augé, Marc 377authenticity 68–9author genius 311authorial intention 311authorship 68–9autocorrelation 259, 265autonomy 372availability samples 145axial codes 315

Bailey,T. 255Bakhtin, M. 167

Barnes,Trevor 21, 277, 278, 296,297, 330, 337

deconstruction 334maths 249, 252metaphors 360writing 348

Barnett, Clive 31Barthes, Roland 333Batterbury, S. 197Batty, M. 268, 270, 272, 275Baxter, J. 149Beauregard, R. 67Beaver, S.H. 124–5Becker, Howard 339, 341, 342, 345Bell, J. 131–2, 135, 136Benjamin,Walter 112, 361Bennett, Katy 197, 198Bennett, R. 218, 255, 268, 287Benn,Tony 69–70Berg, B. 157, 158, 159Berger, John 120, 325Berleant-Schiller, R. 125–6Berry, B.J.L. 233, 234Bertalanffy, Ludwig von 218bias

discursive 258non-official sources 69–70official information 60questionnaires 132–3

binaries 234–40, 244–5, 333–4, 347The Black Inner City as Frontier

Outpost 173–5Blakemore, M.J. 223, 224Bonsiepe, 354books 65Boots, B. 268Brandth, B. 79–80Bridge, G. 131, 132Bristow, J. 193Brown, L. 256, 257, 258Brown,T. 41Bunge,William 26, 86Burgess, Jacqueline 72, 159–60,

161–3, 331

Burgess, R. 197, 198, 199Burke, P. 309Burrough, P. 268Buttimer,Anne 25, 178, 329Bygott, J. 224

Callaway, H. 193Callon, M. 301, 304Campbell, David 26Carter, Paul 326cartography 223, 224

official information 52visualization 274–5see also maps

Casetti, E. 268Castells, Manuel 115catastrophe theory 283categorical data analysis 220categorization 181–5causality 260–3

explanation 288–9in/extensive research 290

census 36, 41, 42, 46, 47history 48–9, 52–3occupations 226

Central Statistical Office(CSO) 53

central tendency 255, 256Chamberlain, Paul 115–17chaotic modelling 268–70Chapman, G.P. 218Chicago School 23Chisholm, M. 286Chorley Committee 225Chorley, R.J. 218cinema 110–11circuit of culture 103civil rights organizations 50–1clarity 343, 346Clarke, G. 273Clarke, K. 274–7Clark, M.J. 280classification 221, 222, 232

ACORN 273, 274

3151-Index.qxd 3/3/2004 4:25 PM Page 409

classification cont.dichotomous 228enumeration 254

Cliff,A. 265Clifford, J. 359Cloke, Paul 135, 151, 173, 208,

232, 283, 296concrete research 295ethics 167–8, 374, 375filmic texts 89household indicators 255, 256meaning 22politics 365, 368–9, 370reflexivity 193

closed questions 130, 137close reading 311–12cluster analysis 263Coan,T.M. 182coding 135, 216

grounded theory 313–17hierarchies 240

coefficient of variation 255Coles, Robert 176, 194communities, categorization 181–5complexity 343, 344, 345computer-based analysis 252computers

electronic media 113–15games 114geocomputation 270–7grounded theory

software 315–17modelling 263–4presentation 362–3see also internet; technology

concrete research 294, 295confidentiality 50–1consensus 328–9consumption, geography of 113content analysis 254context 233–4

identities 244imaginative sources 117–18literature 105photography 110

the conversationalist 310, 326–30Cook, Ian 185, 190, 193, 194, 198,

202–3, 315, 316, 330Cope, Meghan 313, 314corporate marketization 98correlation 262–3, 278Corrigan, Philip 41, 45–6, 47, 48–9Cosgrove, Denis 109, 277, 324,

325, 326, 352–4Couclelis, H. 271counter-transference 331county maps 46covert research 194

Crang, Mike 111, 170, 185,216–17, 315, 316, 317

Crang, P. 180, 189, 198, 202credibility 68, 69Cresswell,Tim 66–7, 237, 324the critic 310, 311–12critical science 27Cromwell,Thomas 45–6cultural difference 165, 167cultural formations 98–9cultural forms 100–1cultural identifications 100cultural institutions 96–8cultural materialism 95–103cultural production 96–103cultural turn 95, 180–1, 183,

191, 263culture 107

circuit 103material 113superorganicism 183

Curry, M. 279, 372Cybriwsky, R. 298

Dandeker, Christopher 44Daniels, Stephen 109, 120, 296,

324, 325, 326Darby, H.C. 44, 56–8, 349–50Darden, J. 174–5data

definition 4description 360–2exploring 255–8inscription 358–60presentation 342–7provenance 219–20sifting and sorting 215–46, 339spatiality 259summarizing 255transcription 358–60types 219, 220, 221, 259

data construction 31–2ethnography 169–205imaginative sources 93–122non-official sources 62–92official sources 41–61‘talking to people’ 123–68

data interpretation 207–13enumeration 247–84, 309explanation 285–306representation 336–65sifting and sorting 215–46, 339understanding 307–35

Data Protection Act 51data source 93Davies, G. 188, 190, 191Davis, C.M. 18, 123–4Davis, D.H. 18, 19

Davis, Mike 63, 88, 113Dear, M. 236de Blij, H. 176, 177decoding 103deconstruction 236the deconstructionist 310, 333–4deduction 216, 287Demeritt, David 330–1, 371,

372, 373Denzin, N. 283depth models 240Derrida, Jacques 236, 249, 333–4de Saussure, Ferdinand 320descriptive statistics 253, 254–9, 277de Souza,A. 339, 343Deutsche, Ros 110de Vaus, D. 143, 144, 145Dewar, N. 176, 177Dewsbury, J.D. 7Dey, I. 248–9, 252, 254dialectics 241–2dialogics 167, 331diaries 67, 69–70, 196dichotomy 228, 236–9Dilthey,Wilhelm 309, 310, 335discursive practices 95, 96discussion groups 159–63

conducting 161strategy 161–3see also focus groups

Disney nature films 91dispersion measures 255Dixon, Deborah 232–4, 235, 277documentary media 66–7Dodgshon, R. 305Doel, Marcus 231, 334Domesday Book/survey 44,

45, 55–8Domosh, M. 9, 107, 352–4Dorling, D. 223, 224, 275, 281–3Dowler, L. 198drama

cultural forms 101cultural identifications 100

Driver, Felix 4, 7, 9, 11, 44, 46–7dualisms 238, 239Duncan, C. 268, 271, 281Duncan, James 181–2, 184, 322,

336, 360Dwyer, Claire 29, 186, 190,

192, 193, 197Dyer, Sarah 330–1dynamic linear modelling 268dynamic non-linear modelling 268

Eagleton 348Ebdon, D. 259‘economic activity’ 60


410

3151-Index.qxd 3/3/2004 4:25 PM Page 410

Economic and Social ResearchCouncil (ESRC) 65, 371, 373

The Economist 66education

cultural reproduction 102league tables 54

Edwards, K.C. 19either/or logic 228–9Ekinsmyth, C. 192electronic media 113–15email questionnaires 132emic codes 314–15empirical-analytic 309empiricism 250England, K. 167enrolment 302entitation 217–19, 221, 223, 229entity maps 230–1, 232Entrikin, N. 172–3, 176, 177,

178, 179, 296enumeration 20, 220, 247–84, 309epistemology 129

of the grid 232–3, 235ethics 25, 164–8, 376–7

ethnography 24explanation 291GIS 279–80

ethnic groups 77the ethnographer 310, 317–23ethnography 23, 24, 169–205, 359

history of 195–6‘new’ 180–96

ethnomethodology 244etic codes 315Evans, M. 199Exeter Domesday Book 57explanation 285–306

understanding distinction309, 335

exploitation 165, 166–7exploratory data analysis 258explorer-scientists 8–9, 15extensive research 127, 128,

150, 289–90Eyles, J. 126, 149

face-to-face questionnaires 132–3,146, 148

factor analysis 263factual sources 62–92, 93Fairbairn, D. 224, 273Family Expenditure Survey 53feminist geography 22, 28–30, 180

abstraction 294binaries 237–8dualisms 238–9ethnography 192gaze 16

feminist geography cont.identities 244writing 347

feminization, advertisements 81feudal system 55–6the field 38, 194

definition 4field-based research 1–16, 18,

21, 23–4see also ethnography

Fielding, J. 252Fielding, N. 252fieldnotes 196–204, 359fieldwork 18, 38, 196–204film 110–11

texts 89–91, 324Finch, Janet 164Fischer, M. 185, 187–8Flowerdew, R. 279, 281flow modelling 268Floyd, B. 248, 249focus groups 134, 159–60

ethics 164strategy 161–2, 163see also discussion groups

Foot, S. 252Forbes, D. 134–5Ford, Larry 111Forrest,Thomas 326Fotheringham, S. 250, 251, 263–4,

265–8, 272, 274, 281Foucault, Michel 49, 304Fowler, F. 146–7fractals 268, 271–2Freedom of Information Act

(1966) 51frequency statistics 256, 257funding 371–3fuzzy logic 241, 252

Gadamer, Hans-Georg 328, 329Gainsborough,Thomas 120–2, 326Gandy, Matthew 110gatekeepers 156Gatrell,A. 255gaze 16Geertz, Clifford 185, 191, 197,

282–3, 308–9, 317, 319–23gender

labour market 48power issues 164stereotypes 77see also feminist geography

generalization 294general linear model 265, 268, 278General Register Office 47, 52, 53genre questions 361genres 102

geocoding 223geocomputation 271–7geodemographics 273–4Geographical Field Group 12, 18geographical information science

(GISc) 251geographical information systems

(GIS) 223, 225, 251, 272–80,278–81, 372

geographical informationtechnologies (GIT) 225

geographical matrix 234geographical scale 240Getis,A. 268Gibbons, M. 304Giddens,Anthony 43–4, 49–50,

186, 187, 189Gilbert, D. 372Gilbert, E.W. 350Gilbert, N. 357–8Glaser, Barney 313Glennie, Paul 226–7, 244globalization, popular music 119glocal 240Gober, P. 251Godlewska,A. 9Goffman, Erving 157Gold, J. 110Goodchild, M. 223, 225, 271Goss, J. 67, 81–4, 160, 280Gouldner,A. 126Gould, P. 216, 249governmentality 49government data see official

informationGovernment Social Survey 53Graham, D.J. 287–8Graham, E. 250Gramsci,Antonio 49graphics 114graphs 256Great Domesday Book 57–8Greek tragedies 101Gregory, Derek 189, 227, 296,

297, 298–9, 329, 337architecture 111–12ethnography 187explanation 305, 335Habermas 27maps 46metaphors 360natural science 307positivism 250world-as-exhibition 16writing 348

Gregory, S. 221, 258, 262grounded theory 313–17group questionnaires 134

INDEX

411

3151-Index.qxd 3/3/2004 4:25 PM Page 411

Gubrium, J. 149, 150Guelke, Leonard 287, 330

Habermas, Jürgen 27, 309, 328Haggett, Peter 259, 264Hakim, Catherine 48Hamilton,W. 326, 327Hammersley, M. 70, 156, 165Hammond, P. 248Hanson, S. 343Haraway, Donna 188, 192Harley, J.B. 85, 91–2, 224harm 165, 166Harré, R. 127Harrison, R.T. 330Harris,T. 280Hart, J.F. 349Harvey, David 27, 103, 110,

290–2, 294architecture 111dialectics 242explanation 287morality 375politics 365quantitative revolution 249

Hay,A.M.H. 218Hepple, Les 259, 278Herbert, S. 169, 170, 192, 193hermeneutic circle 329–30hermeneutics 309–10Hertz, Rosanna 345Herzog,Werner 110Hetherington, K. 300heuristic searches 272hierarchies 234–40, 244–5Hinchliffe, S. 189historical-hermeneutic 309Hoggart, K. 169, 170, 188, 191–2,

194, 197, 200Holstein, J. 149, 150homosexuals, stereotypes 77household questionnaires 134Huck, S. 259, 260–2Huff, D. 259Huggett, Richard 217, 220–1humanistic ethnographies 171–3humanistic geography 22–5, 28,

105, 181, 186, 296–7, 312hundreds 55–6Hutchings, G.E. 12, 13, 125hypertext 114hypotheses

null 259rival 259, 260–1

the iconographer 310, 323–6iconography 109idealist human geography 330

identities 229–37, 242, 243–4identity management 194image analysis 325imaginative sources 62, 93–122indicative samples 145induction 216, 255industralisation 46–7inferential statistics 254, 258–63, 278inflation 54informed consent 165, 166Inquisitio Eliensis 56inscription 358–60instituted artist 96, 97intensive research 127, 128,

150, 289–90interaction, electronic media 114interactive relationships 129internet 113

newspaper reports 66QUALIDATA 65questionnaires 132

interpretative research 23, 24, 331intersubjectivity 130, 322

ethnography 170, 192interviews 150, 151, 152,

155, 157, 165interval data 220, 221, 259interviews 148–59

appropriateness 150–1checklist 155conducting 157–9ethics 164history of 123–4, 126interactive relationships 129involved relationships 129preparation 152–7questionnaire distinction 148–9questionnaire overlap 126–7voices 151

intressement 302intrusion 13–14involved relationships 129

Jackson, Peter 187, 198, 322, 330archives 63discussion groups 160, 161, 162ethnography 169, 173, 174,

175, 178superorganicism 182, 183–4, 185

Jayaratne,T. 131Johnston, R.J. 192, 247, 248,

251, 263–5politics 371questionnaires 130spatial analysis 249–50, 263

Jones, E. 176Jones, John Paul 232–4, 235,

268, 277

Jones, K. 268, 271, 281Jones,Wellington 19Joseph Rowntree Foundation

371, 373journals 65Junker, 198juxtaposition 350, 351

Karnaugh maps 228, 230Keat, R. 286, 287Kenzer, Martin 21Kingsley, Mary 107, 108Kinross, R. 354–6Kinsman, Phil 110Kitajima, S. 372Kitchin, R.M. 224, 225knowledge production 7, 304, 309Kong, Lily 111, 117–19Krippendorff, K. 254Kroeber,Alfred 182Kuleshov, Lev 362

Labour Force Survey (LFS) 59–60, 61labour market

gender differences 48unemployment figures 58–61,

223–4, 265, 267–9landscape 10, 46, 325–6, 347

art 12, 13, 109, 120–2description 349–54metaphors 115–17photography 109–10

Latour, Bruno 183, 189, 190, 303Laurier, E. 189, 190, 191, 244Law, J. 190, 300Leavis, F.R. 312Lee, Dick 117–19Lee, R. 252letters 67Levitas, R. 59, 60Lewis, C. 244Ley, David 24, 190, 192, 296, 297–8,

299, 300ethnography 173–5, 178, 179, 180humanistic geography 22interpretive research 23thick description 195

Leyshon,A. 107–9lifeworld 309Limb, Melanie 29, 186, 190,

192, 193, 197liminality, advertisements 81Lindsay, James 4, 216, 219–20,

221, 224literary critic 311–12, 319literature

imaginative sources 104–5searches 65


412

3151-Index.qxd 3/3/2004 4:25 PM Page 412

Little Domesday Book 57Livingstone, David 15, 224, 330, 370local geography 10, 12localization, popular music 119local modelling 265–8local point pattern analysis 268location

discussion groups 161ethnography 200–1interviews 157–8see also the field

logic 242fuzzy 241, 252

logocentrism 249, 333Longley, P. 268, 271, 272, 273Lorimer, H. 7Lowie, Robert 182Lucas, 260

McCloskey, D. 349McCullagh, P. 248McDowell, Linda 3, 5–7, 8, 28,

193, 195, 197binaries 237explanation 285gender division 28reflexivity 29, 31social constructions 191the therapist 331

McGlade, M. 251McKibben, B. 91Macmillan,W. 271Madge, Claire 192, 367, 374Malbon, Ben 189, 361, 362maps 67–8

codes 315, 316entity 230–1, 232official statistics 52photography 109Tudor England 46visual texts 85–9see also cartography

Marcus, George 185, 187–8, 190, 191marginality, advertisements 81markets, artists 97–8Marsh, C. 263Martin, D. 41, 63–4, 223, 271Marxism 25, 27, 186, 188Marx, Karl

abstraction 291–3dialectics 241–2

masculinity, tractor advertisements 79Mason, Jennifer 152, 153–5,

158–9, 164, 285Massey, Doreen 63, 103, 110,

183, 184, 235glocal 240metaphors 360

material culture 113Matless, David 109May,T. 364, 368mean 255, 256meaning

circuit of culture 103the deconstructionist 333the ethnographer 318, 322–3the iconographer 324interviews 149–50, 159non-official sources 68, 70the therapist 332understanding 307–35

media 98‘art’ 100documentary material 66–7texts 71–8, 324

median 255, 256Medick, H. 227memos 66Merton, R. 161Meserve, P. 11metaphors 360–1

landscape 115–17metaphysics of presence 333methodolatry 126, 127methodology 3

definition 5see also data construction

Michael, Mike 190Milbourne, P. 374Miles, M.B. 317Miller, 345–6Miller, Daniel 113Mills, 343minutes 36–7, 66missing data 68Mitchell, Don 182, 183, 184–5Miyares, I. 251mobilization 302mode 102, 255, 256modifiable areal unit problem

(MAUP) 265Mohammad, Robina 368, 369Molotch, H. 67Moon, G. 41Moore, Donald 323morality 374–5Morgan, D. 160–1Morley, D. 89Mountz,A. 190, 192Mr and Mrs Andrews 120–2, 326Mulkay, M. 357–8multilevel modelling 252, 268multimedia productions 114multiple methodologies 252multi-stage cluster samples 145multivariate analyses 263

multivariate spatial relationships 268Murdoch, Jonathan 190, 301, 303music 107–9, 117–19Mutersbaugh,Tad 323

Nagar, Richa 192, 203–4Nancy, J.-L. 332Nash, C. 189National Statistics 53Natter,W. 232natural science 307–8Naylor, S. 7Neil,Andrew 71neurocomputing 272newspapers 66–7, 71–8nominal data 220, 221, 259non-official sources 62–92non-parametric tests 259non-probability samples 144, 145non-representational theory

188–9, 299–305Northedge,A. 337, 338, 344, 346note-taking

in the field 196–204, 359interviews 152

Nuffield Foundation 371, 373null hypotheses 259

Oakley,Ann 164objectivity 17, 20, 126, 309, 357

politics 367–9rhetoric 354–60writing 347

O’Brien, L. 265observant participation 188, 189observation 8, 11–12

being there 8–16participant 10–11, 170, 171,

194, 195–6space possession 15

occularcentricism 15occupations 48, 226–7, 232, 233–4,

244, 265, 266–8Office for National Statistics

(ONS) 53Office of Population Censuses and

Surveys (OPCS) 53official information 41–61Official Secrets Act 51O’Kelly, 268Olsson, Gunnar 231–3, 234–7,

241–3, 334open coding 315open questions 135, 136, 138, 143Openshaw, C. 252Openshaw, S. 252, 253, 265, 279opinion polls 127–8Oppenheim,A. 131, 135, 137–8, 142

INDEX

413

3151-Index.qxd 3/3/2004 4:25 PM Page 413

oppositional formations 99order 277ordinal data 220, 221, 259Ord, J. 265, 268organization, cultural

production 102–3orientalism 348Other 192otherness, filmic images 90Outcast Cape Town 173, 175–7overt research 194

Palm, R. 173, 174, 175Panofsky, E. 325paradisal icons 81parametric tests 259, 263parish registers 46Parr, Hester 165, 170, 192, 197,

200, 331participant observation 10–11,

170, 171, 194research tips 195–6see also ethnography

patronage 96–7Pattie, C. 263Peet, Richard 22, 25, 208Peretti, 260performing arts 107–9personal documents 67personal pronoun 21, 151Philip, L.J. 252Philo, Chris 151, 180, 192, 330, 331

ANT 189, 190, 191ethics 374, 375ethnomethodology 244non-official sources 66quantitative revolution 248

photography 15, 78–84, 109–10Pickles, John 279Pile, Steve 165, 244, 331piloting, questionnaires 135, 145–6Pirie, G. 176, 177place-name research 125–6planning, discussion groups 161Platt, R.S. 9, 11, 18Le Play Society 12, 124–5Plummer, P. 249Pocock, D. 105, 106, 312poetics 348pointillism 231political geography, abstraction 294politics

geographical practice 364–75official information 54

Pollard, Ingrid 110Poole, R. 236positionality 21, 29, 30, 31,

129–30, 329–30, 365

positivism 20, 105, 172, 250,277, 278, 281

explanation 290–1, 295–6,297–8, 305

spatial science 286, 287postal questionnaires 132–3, 139–42,

146, 150–1post-artisanal marketization 97–8post-colonial geography 28, 29post-market institutions 96, 98postmodernism 232, 233, 236poststructuralism 188, 232, 235, 277Powell, R.C. 4powerful groups 193powerless groups 164, 193power relations 29, 31, 130, 164–5,

192, 193, 372practical criticism 311–12pragmatism 281Pratt, G. 160, 188, 236, 252, 360preconstructed data 4, 32, 36–7

imaginative sources 62, 93–122non-official sources 62–92official sources 41–61

Pred,Alan 361–2pressure groups 66Prigogine, I. 268, 283primary data 4, 21, 219Prince, H. 178principal components analysis 263Prisoners of Space? 173, 177–80privacy 165, 166, 280probability samples 144, 145problematization 302product-moment correlation

coefficient 263professional marketization 98projective techniques 143–4promotional material 67Pryke, Michael 363psychoanalysis 165, 331–2publicity material 67Punch, S. 198purposive samples 145

QUALIDATA 65qualitative data/methods 17, 19,

125, 131, 220, 247–53computer-based analysis 252ESRC 65

quantitative data/methods 17,19–20, 21, 125, 220–1

‘downturn’ reasons 250, 251, 279enumeration 247–84ESRC 65official information 41–61questionnaires 130, 131revolution 104–5

questionnaires 123, 130–48, 150–1accompanying letters 146, 147carrying out 144–8construction 134–44ethnography 174example 139–42format 132–4interview distinction 148–9interview overlap 126–7piloting 135remote relationships 129response rates 132, 133, 146when to use 131–2

questionsdesign 135, 137–8types 135, 136–7, 159

Quiller-Couch,Arthur 311Quinton,A. 242Quoniam, S. 350, 352, 353quota samples 145

Rabinow, P. 47radical geography 22, 25–8Radical Statistics Group 281–2random samples 145rapport, interviews 157–8raster systems 271ratio data 220, 221, 259Reah, D. 72, 76, 77record keeping 43–4reference material 65reflexivity 22–31, 165, 167, 191–5regional description 349–54regional economies 49regionalization 19, 222Registration Act (1836) 47regression analysis 263,

264–5, 278relational analysis 235Relph, Edward 297, 332remote relationships 128–9representation, human geographies

336–65representativeness, non-official

sources 68, 70representative samples 144–5, 155–6reproduction, cultural production

101–2research

alliances 165definition 3

Research Assessment Exercise 373researcher as robot model 148research institutes 66research questions

in/extensive research 289–90interviews 152–3

research reports 65–6


414

3151-Index.qxd 3/3/2004 4:25 PM Page 414

response rates, questionnaires 132,133, 146

Revill, George 109rhetoric 350

objectivity 354–60writing 347–9

Richards, I.A. 311–12Richardson, L. 347, 348, 354Richardson, Miles 322–3Richards, Pamela 339–40Robins, K. 89, 360–1Robinson, G.M. 220robotic relationships 127–8Robson, E. 193, 196, 203roles, interviews 157Rorty, Richard 281, 328, 359Rosaldo, R. 194Rose, D. 255Rose, Gillian 129, 237

art 120, 122dualisms 238, 239film 110gaze 16the iconographer 324identities 234oral histories 70psychoanalysis 331, 332reflexivity 31writing 347

Rowles, Graham 24, 173, 177–80,189, 192–3

Rundstrom, Robert 21Ryan, James 15

Saarinen,T. 144Sadler, D. 66Said, E. 348Salter, C.L. 10–11, 124Saltmarsh, Rachel 193, 365, 366–7samples

discussion groups 161interviews 155–6size 64, 145snowballing 156

samplingbias 70errors 52questionnaires 144–5statistics 254

Sandler, H. 259, 260–2Sanjek, R. 359Saro-Wiwa, Ken 372Sauer, Carl 3, 5–7, 8, 9–10, 11,

17–18, 123, 182, 322Sayer,A. 127, 128, 131, 286, 287,

288–9, 290, 294–5, 361Sayer, Derek 41, 45–6, 47, 48–9Schaeffer, F.K. 286

Schutz,Alfred 177, 296–7science/scientific method 18, 19,

20, 25, 277–8, 287, 290,296, 307–8

Scott, J. 68, 69, 70Seamon, D. 178, 332secondary data 4, 21, 219selective coding 315self-constructed data 4, 32, 37–9

ethnography 23, 24, 169–205, 359‘talking to people’ 123–68

Selkirk, Keith 227, 228, 229, 238semantic differentials 174semiotics 319, 320set theory 227–9Shakespeare,William 115–17Sheppard, E. 249, 252Sherman, Cindy 110Shurmer-Smith, P. 180, 181,

182, 183, 184, 197Sibley, David 14, 232, 252, 277Sidaway, J.D. 193signals, cultural identifications 100significance levels 259signified 320signifiers 320signifying systems 102–3signs 320Simon, David 176Simpson, Charlotte 10Simpson, S. 282Skelton,Tracey 365Sloman,A. 5Smailes,A.E. 19Smith,Adam 188Smith, D.M. 26, 327, 377Smith, N. 9, 240, 279, 360Smith, Susan 129, 156, 178, 185, 187,

188, 189, 192, 198, 201–2, 322snowballing 156social actors 296social construction 191

maps 85newspapers 71

social context, imaginative sources 94social identity, GIS 280social science 307–8Soja, E. 350, 351soliloquy 100space-time forecasting model 268spatial analysis 232–3, 249–50, 272–7spatial autocorrelation 259spatial distribution 259spatiality of data 259spatial modelling 268, 269spatial process modelling

263–70, 277spatial referencing 222–3

spatial science 19–21, 22, 172, 286–8Spearman rank correlation 263specializing formations 99spectral analysis 268Spedding, N. 7spirituality 95sponsorship 97, 372Spradley, 201Stacey, J. 198standard deviation 255standardization, questionnaires 146–8Stanley, H.M. 370Statistical Abstract of the United

States 44, 51–2statistical societies 47statistics 220, 249

British deficiencies 282descriptive 253, 254–9, 277inferential 254, 258–63, 278public policy 47–8, 54Radical Statistics Group 281–2word origins 47

Steinberg, J. 343–4stereotypes 72, 77, 192Stoddart, D.R. 8, 9, 12–13Storper, Michael 111stratified samples 145Strauss,Anselm 313stress index 174structuralism 233, 235structuration theory 186–7, 188Stutz, F. 178subjectivity 17, 20, 309

cultivated 312explanation 295–9writing 347

Sullivan, O. 255superorganicism 181–5, 322surrealism 242surveys 16–22

fatigue 130official statistics 52–3see also questionnaires

synthesis research 294systematic samples 145systems 309systems analysis 217, 218

‘talking to people’ 123–68tape-recording 152Tate, N. 224, 225Taylor, P.J. 251, 278technology

official information 45state information 50–1see also computers; internet

telephone questionnaires 132–3, 146television 110–11

INDEX

415

3151-Index.qxd 3/3/2004 4:25 PM Page 415

tertiary data 219therapeutic alliance 165the therapist 310, 330–3thick description 89, 164, 283,

308, 309, 317, 319, 323thin description 283, 309Thorne, L. 189Thorpe, Mark 113Thrift, Nigel 183, 186, 187, 188,

189, 235actor-network theory 301non-representational theory

299, 300, 304writing 337

Tickell,A. 372time-series 268Tomaselli, K.G. 110topologics 231totalitarian states 50tourist advertising 78–84Tozer, H.F. 8tractor advertisements 79–80tradition, cultural reproduction 102transcription 358–60transculturation, popular music 119transference 331transformations, newspapers 72travel writing 105, 107Trondman, M. 186, 197Tuan,Y.-F. 296Twyman, C. 193, 198

understanding 307–35unemployment data 58–61, 223–4,

265, 266–8univariate spatial relationships 268universal laws 287

Upton, G. 275, 276urban populations 46–7urban surveys 19Urry, John 286, 287, 305

van Hoven, B. 315vector systems 271–2Venn diagrams 228, 229, 231vessel-of-answers approach 149video 111vignettes 143–4village studies 125virtual reality 114–15visual arts 109visualization 273–7visual textsmaps 85–9

‘nature documentaries’ 89–91tourist advertising 78–84

voting data 41voyeurism 90

Warburg,Aby 324–5Ward, D. 47Warde,A. 360Warnke, Georgia 328–9Warren, C. 198Watts, M. 292, 293–4, 365, 372Weber, Max 44, 309, 335Weir, Peter 110Weitzman, E.A. 317welfare states 50Wessels,Victor 176, 194Western, John 173, 175–7, 178,

183, 194Whatmore, S.J. 189Wheeler, P.T. 19

Whitehead, F. 53Whittlesey, D.S. 123Williams, Raymond 47, 95–103, 104Willis, K. 196Willis, P. 186, 187, 197Wilson,A.G. 91, 218, 252, 268Winnicot, Donald 331Wisner,W.B. 26Wolcott, H. 341–2Wolfinger, 197, 198, 199, 201Woods, M. 301–4, 372Woodward, R. 255Wooldridge, Sidney 11–12Woolf, P. 112–13world-as-exhibition 16Wrigley, N. 220, 254, 265writing

academic jargon 343, 345–6composition 361–2as a conversation 346genre questions 361‘good’ 343–7record-keeping 43–4representation 336–65research diaries 196rhetoric 347–9, 350, 354–60style 361–2

Zonn, Leo 110, 111

Indexed by Caroline Eley


416

3151-Index.qxd 3/3/2004 4:25 PM Page 416