An obsession with meaning: a critical examination of the pictograph sites of the Lake of Woods

.~.

An Obsession with Meaning: A Critical Examination of the Pictograph Sites of the Lake of the Woods

Alicia]. M. Colson Department of Anthropology

Mc Gill University

June 2006

A Thesis submÏtted to McGill University in partial fulfi1ment of the requirements of the degree of Doctor of Philosophy

© Alicia J. M. Colson, 2006

1+1 Library and Archives Canada

Bibliothèque et Archives Canada

Published Heritage Branch

Direction du Patrimoine de l'édition

395 Wellington Street Ottawa ON K1A ON4 Canada

395, rue Wellington Ottawa ON K1A ON4 Canada

NOTICE: The author has granted a non­exclusive license allowing Library and Archives Canada to reproduce, publish, archive, preserve, conserve, communicate to the public by telecommunication or on the Internet, loan, distribute and sell theses worldwide, for commercial or non­commercial purposes, in microform, paper, electronic and/or any other formats.

The author retains copyright ownership and moral rights in this thesis. Neither the thesis nor substantial extracts from it may be printed or otherwise reproduced without the author's permission.

ln compliance with the Canadian Privacy Act some supporting forms may have been removed from this thesis.

While these forms may be included in the document page cou nt, their removal does not represent any loss of content from the thesis.

• •• Canada

AVIS:

Your file Votre référence ISBN: 978-0-494-32167-6 Our file Notre référence ISBN: 978-0-494-32167-6

L'auteur a accordé une licence non exclusive permettant à la Bibliothèque et Archives Canada de reproduire, publier, archiver, sauvegarder, conserver, transmettre au public par télécommunication ou par l'Internet, prêter, distribuer et vendre des thèses partout dans le monde, à des fins commerciales ou autres, sur support microforme, papier, électronique et/ou autres formats.

L'auteur conserve la propriété du droit d'auteur et des droits moraux qui protège cette thèse. Ni la thèse ni des extraits substantiels de celle-ci ne doivent être imprimés ou autrement reproduits sans son autorisation.

Conformément à la loi canadienne sur la protection de la vie privée, quelques formulaires secondaires ont été enlevés de cette thèse.

Bien que ces formulaires aient inclus dans la pagination, il n'y aura aucun contenu manquant.

Errata:

Appendix 6 commences on page 551 instead of page 550 as stated in the Table of Contents. Page 550 do es not exist.

Table of Contents Abstracts .............................................................................................................................................. vi Acknowledgements ........................................................................................................................ vüi Note on spelling .............................................................................................................................. xüi Note on Kl.EIO lAS ...................................................................................................................... xiv List of Figures ................................................................................................................................... xv List of Tables ................................................................................................................................. xxiv Chapter 1 - Introduction ....... ....................................... , ................................................................... 1 Where are these sites? .......................................................................................................................... 1 Overview of rock image studies ......................................................................................................... 2 The use of informants ......................................................................................................................... 4 \Vhat is an interpretive appraach? ..................................................................................................... 5

The culture-historical approach (Level 1): .......................................................................... 6 The contextual approach (LeveI2): ...................................................................................... 6 The intuitive approach (Level 3): .......................................................................................... 7 The analogical approach (Level 3): ....................................................................................... 9 The homogical approach (Level3): .................................................................................... 10

The debate between the "lumpers" and the "splitters" ................................................................ 11 The structure of this thesis: ............................................................................................................... 16 Chapter II - Lake of the Woods: Regional Setting of the Sites ......................................... 17 Section 1: Geohistory and geography of the region ...................................................................... 17

(a) Geology ............................................................................................................................. 17 (b) Changing lake levels ........................................................................................................ 23

Section 2: History and ethnography ................................................................................................ 27 (a) Who lived in this regian during the early historical times? ........................................ 28 (b) The unresolved debate between historians and archaeologists over ethnic identification at the time of contact with Europeans ....................................................... 35 (c) Understanding the archaeological record of the Boreal Forest ............................... .48 (d) Can biological evidence resolve the problem of ethnic identity? ............................. 62

Section 3: The images created and used by the Algonquian-speaking peoples ......................... 74 (a) Is establishing the ethnic origins and its objects possible? ........................................ 76

.~. (b) Images as a means of communication .......................................................................... 97 Conclusion: ........................................................................................................................................ 123 Chapter III: A Review of the Literature concerning the Study of Rock Images Sites in the Lake of the Woods in the Canadian Shie1d prior to 2001. ................................... '" ..... 127 Section 1: The earliest investigation of these images .................................................................. 127 Section 2: More recent literature ................................................................................................... 128 Section 3: Techniques used to record the pictographs ............................................................... 136

(a) Subdividing the sites into sections ............................................................................... 136 (b).Terminology ................................................................................................................... 161 Cc) The development of the techniques used ta record the images ............................ 163

Section 4: Technical Analysis .......................................................................................................... 168 (a) Technical issues involved in recording and interpreting rock image sites ............. 168 (b) Conservation and Analytical Techniques ................................................................... 167

Section 5: Culture-Historical Approach .............................................. ......................................... 173 Section 6: Contextual Approach .................................................................................................... 191 Section 7: Literature concerned with meaning ............................................................................. 194

Page ii

(a) Intuitive Approach: ........................................................................................................ 194 (b) Analogical Approach: ................................................................................................... 197 (c) Homological Approach ................................................................................................. 197

Conclusion ......................................................................................................................................... 209 Chapter IV - My Work on the Rock Image Sites ................................................................. 213 Section 1: Numbers of images ...................................................................................................... 217 Section 2: Defining sites .................................................................................................................. 218 Section 3: What methods were used? ............................................................................................ 218

(a) Examination of maps from different periods ............................................................ 218 (b) Methods for coilecting data ......................................................................................... 223 (c) Image analysis of potential sites in the field ................................ ; .............................. 232

Section 4: The method used for describing shapes and how does one consider time: .......... 235 (a) Exfoliation ...................................................................................................................... 235 (b) White mineraI deposits ................................................................................................. 236 (c) Lichen and rock tripe growth and encroachment .................................................... 237

Section 5: Assessing age .................................................................................................................. 238 Section 6: Description of a shape .................................................................................................. 239 Section 7: The tools used to aid the analysis of the data ........................................................... .242

Ca) IZLEIO lAS ................................................................................................................... 242 (b) Image analysis using Adobe Photoshop 6.0 and VIPS ............................................ 244 Conclusion ............................................................................................................................ 245

Section 8: Vocabulary used to de scribe each type of shape found ........................................... 245 Section 9: Possible relationships between rock image sites and other archaeological sites .. 252 Conclusion: ........................................................................................................................................ 253 Chapter V: Culture-History Approach .................................................................................... 255 Section 1: Blurring/overlay discussion and clarity of images ................................................... .254 Section 2: Distribution of images ................................................................................................... 267 Section 3: Analysis of the rock image style in relation to Northern Ontario and to adjacent

areas ............... : ...................................................................................................................... 268 Section 4: The techniques used to make paintings ..................................................................... .277 Section 5: Discussion of the offerings left at sites ....................................................................... 279 Section 6: Discussion of dates ........................................................................................................ 297 Conclusion ........................................................................................................................................ 387 Chapter VI - Contextual Approach .......................................................................................... 391 Section 1: ............................................................................................................................................ 391

Caves: ............................................................................................................................. 412 Summary for Section 1: ...................................................................................................... 414

Section 2: ........................................................................................................................................... 415 Summary for Section 2: ...................................................................................................... 417

Section 3: ........................................................................................................................................... 417 Summary for Section 3 ....................................................................................................... 435

Conclusion ......................................................................................................................................... 439 Chapter VII - Application of Approaches Concerned with Meaning ........................... .441 Section 1: The two useful approaches concerned with meaning ............................................. .441

(a) Analogical approach: .................................................................................................... 442 Cb) Homological approach: .............................................................................................. .444

Section 2: An exarnination of four birch bark scroils ................................................................ .452

Pageiii

(a) How were the seroIls examined? ............................................................................... .455 (b) Culture-historical Approaeh: ....................................................................................... 456

Comparison of the images found on bireh bark seroIls and on the roek image sites ................................................................................................................................. 458

(e) Contextual Approaeh: .................................................................................................. 460 The eontext of the information on the seroIls: ..................................................... .460

SeroIl1: .................................................................................................................... 460 SeroIls 2 and 3: ....................................................................................................... 460 SeroIl4: .................................................................................................................... 461

(d) Meaning: ......................................................................................................................... 461 Seroil 1: ......................................................................................................................... 462 SeroIls 2 and 3: ............................................................................................................ 463 SeroIl 4: ......................................................................................................................... 466

(e) Coneluding eomments: ................................................................................................ 467

.~ .. Seetion 3: Efforts to treat roek image eombinations as stories: ............................................... .467 (a) DiKm-4: .................................................................................................................. 468 (b) DhKm-3: ................................................................................................................ 469 (e) DgIZtn-1: ................................................................................................................. 470 (d) DhI<.:n-1: ................................................................................................................. 470 (e) DhKo-1: .................................................................................................................. 471

Coneluding Remarks: ........................................................................................................ 472 Seetion 4: The eontinuing importance of the images and the sites .......................................... .4 72 Conclusion: ........................................................................................................................................ 473 Chapter VIII: There is no Holy Grail: An Obsession with Meaning . .......................... ..477 Introduction: ..................................................................................................................................... 477

What has been going on here? ................................................................................................ 481 Seetion 1: Four Goals ...................................................................................................................... 484 Seetion 2: The theoretieal and methodologieal implieations of this thesis: ............................ .497 Seetion 3: Further researeh - What should be done next and whieh way should the field

go? ............................................................................................................................................... 500 Appendices ...................................................................................................................................... 502 Appendix 1: Summary of the information available on the pietograph sites prior to 2001 .. 502 Appendix 2: What some of the Tools of Adobe PhotoShop 6.0 ean reveal. ......................... .525 Appendix 3: A Blank Copy of Form A and Form B for Sites on Cliffs and Caves ............... 531 Appendix 4: The Mod file used by KLEIO lAS ......................................................................... 534 Appendix 5: Uehen, Roek Tripe, Exfoliation, and White Mineral Deposits in eonjunetion

with Site Orientation .......................................................................................................... 542 Appendix 6: The Different Types of the Shape Loosely Categorised as a 'Creature' ............ 550 Appendix 7: Examples of the Different Types of Shapes in Each Group .............................. 557 Appendix 8: The Geology, Geomorphology, and Type of Land Mass of Eaeh Site ............. 565 Appendix 9: Possible Correlations between the Type of Track, Reserves, Offerings in 2001

and before, and Type of Site ................................................................................... 570 Appendix 10: Indian Reserves, Offerings in 2001 and before, Type of Site, Type of Land

Mass, and Orientation of Sites ................................................................................ 577 Appendix 11: Data regarding the Type of Aecess to the Pictograph Site, Vegetation at base

of site, water levels, and the number of eliffs used .............................................. 582 Appendix 12: Do ail the sites have the necessary physieal features for pietographs? ............ 585

Pageiv

Appendix 13: Sites grouped according to the number of physical attributes present ............ 588 Appendix 14: Sites grouped according to number of physical attributes present plus offerings .

..................................................................................................................................... 589 Appendix 15: Sites grouped according to number of physical attributes present and with

rocks at base of cliff face ......................................................................................... 590 Appendix 16: The "presence and absence" tables and the dendrograms from the Pearson

correlation and the Euclidean distance method ta consider the relationship· between the physical attribûtes of the sites, the different sites and the images, and the relationship between the sites and sorne of the largest shapè types .. .591

Appendix 17: The "presence and absence" tables for the images of the birch bark scrolls .. 635 Cited Bibliographic References ...................................................................................................... 637 Ethics Certificate .............................................................................................................................. 681

(Ail the figures in this the sis including the appendices are available for viewing on the CD in the envelope on the inside of the back cover.)

Page v

Abstract

Most researchers who study rock image sites tend to be interested in the meaning of

images, even though they could obtain more empirical information about these images and

their physical location. Furthermore, very little of the work done in the past on rock image

sites has been systematic. In this thesis I address the dearth of detailed information on the

images and their context. This thesis presents a thorough examination of the images of the

twenty-seven pictograph sites in the Lake of the Woods, in the Canadian Shield. These

pictograph sites were selected because they exhibit traits evident in rock image studies in

other parts of the world.

This study is based on data col1ected during three months of fieldwork conducted in

2001. Images were found on cliff faces and inside caves. New images and new sites were

found and identified.

Here, as elsewhere, the choice of theoretical approach influences the fieldwork,

analysis, and search for meaning. Each prescribes the types of questions asked and

determines the levels of understanding obtained about whichever form of archaeological

evidence is being considered. The different but complementary theoretical approaches

should be employed in a definite order. The same data must be examined in sequential order

using these different approaches to increase the potential quantity and quality of information

gained. Archaeologists should use the foilowing sequence of approaches: culture-historical,

contextual, foilowed by either the homological, or analogie al approaches, or a combination

of the latter two.

Classifying and describing any image is very difficult, since the level of description

given to an image affects the way in which it can be analysed, and heavily influences the

possible outcome of any discussion of perceived meaning. A rigorous examination of the

images of these sites was conducted to (a) identify the possible vocabulary of images, (b)

de termine whether combinatory rules exist, (c) reconstitute the life history of each site, and

(d) as certain whether the images can be related to other indigenous images to determine if

this can provide information about the meaning(s) of the rock images. In assessing the

meaning of the rock images, the images of a few birch bark scroils were considered, since it

was posited that a detailed investigation of the scroils, the ethnographic record, and their

pictographs might provide some answers regarding the meanings of the images found on the

rock faces.

Page vi

Résumé

La plus grande partie des chercheurs qui étudient les sites rupestres tendent à

s'interesser à la signification propre des images bien qu'ils puissent en tirer plus

d'informations empiriques sur les images et leurs cadres physiques. En outre, très peu de ces

études ont été systématiques. Cette thèse se met à considerer le grand manque de détails sur

les images dans leur contexte par le moyen d'un examen détaillé, rigoureux et objectif de

vingt-huit sites rupestres dans la région du Lac des Bois sur le Bouclier canadien. Ces sites

ont été choisis.car ils démontrent des traits communs qui se trouvent parmi bien d'autres

sites rupestres à travers le monde. Les données ont été recueillies au cours de trois mois de

travaux sur le terrain entrepris en 2001 et comprennent de nouveaux sites, falaises et abris, et

de nouvelles images identifiés pour la première fois pendant cette période.

Ici, de même qu'ailleur, la méthodologie théorétique choisie pour l'étude détermine la

direction des travaux sur le terrain, leur analyse et les questions qui mènent à la signification

des résultats. Chaque méthodologie ordonne les questions posées et le niveau de

compréhension des réponses selon la catégorie de l'evidence examinée. Des approches

différentes mais complémentaires devraient être employées dans une succession bien

définie.en examinant les mêmes données à fin d'assurer et augmenter la quantité et la qualité

des résultats. Les archéologues devraient déployer ces approches dans l'ordre suivant: sur le

plan culturel-historique; dans le contexte des données; puis par des enquêtes homologiques

ou analogiques, seule ou en combinaison.

La classification et description d'une image est difficile car le niveau de description

influe sur la façon de l'analyser de même que sur l'issue de sa signification. Un examen

détaillé et minutieux des images a été entrepris dans le but de: a) identifier un vocabulaire

possible des images; b) determiner s'il existe des règles combinatoires; c) reconstituer

l'histoire de chaque site; et d) apprendre si ces images pourraient avoir un rapport à d'autres

images hors des sites rupestres, ce qui pourraient aider à éclairer la signification des images

rupestres. Puisqu'on a proposé qu'une enquête détaillée sur les images posées sur écorce de

bouleau pourrait peut-être mener à des réponses aux questions sur la signification de l'art

rupestre, quelques rouleaux d'écorce ont été examinés.

Page vii

. r--'-.

Acknowledgements

My PhD committee is headed by Professor Bruce Trigger and consists of Professors

Toby Morantz, Michael Bisson, and Andre Costopolous from the Department of

Anthropology, McGill and Professor José Igartua from Département d'histoire, Université

du Québec à Montréal. l must thank Bruce Trigger, my supervisor, for the numerous

opportunities to play with ideas and explore their possibilities especially when he may have

thought that they were not sensible but he let me discover for myself that they were really

not very useful. l must thank him for his guidance, his patience, and for always being there

whenever l needed his help and advice and for giving me the opportunity to think carefully,

clearly, critically, and to evaluate everything with an open mind and the widest possible

perspective. l have tried at aIl times throughout this degree to follow his advice regarding

data to make sure that l squeezed the orange until nothing more could be extracted.

Imust thank Bruce Trigger, Toby Morantz, and José Igartua for aIl their thoughtful

editing and writing advice. l would like to thank José Igartua for his advice and assistance

with regard to the computing tools utilised and for his advice when writing about something

technical in computing so that l used as few technical words as possible but still got the ideas

and concepts across to the reader. l must thank Andre Costopolous for his advice and

technical assistance and Michael Bisson for his thoughtful comments. l must thank Toby

Morantz for her support and help during the final year of this PhD.

l must thank Professors Bruce Trigger and Toby Morantz for financial support

whilst writing up my thesis. l am immensely grateful for their help .

l would like to thank various members of the Anthropology department at McGill

with whom l took courses including Professors Colin Scott, Toby Morantz and Jerome

Rousseau for the opportunity to explore and obtain a solid grounding in the field of

anthropology as it is perceived in North America

l must thank Dr. Scott Hamilton (Department. of Anthropology, Lakehead

University, Ontario) for his advice on various drafts of the conclusion of my thesis.

l must thank Robert von Bitter (Ontario Ministry of Culture and Communication,

Toronto, Ontario); Dr. Stewart Baldwin, Dr. Scott Hamilton, and Dr. Joe Stewart (professor

Emeritus) of the Department of Anthropology, Lakehead University, Ontario), C. S. Reid

(formally Regional Archaeologist, Ontario Ministry of Culture and Communications, Kenora,

Ontario), Dr. Jean-Luc Pilon, Curator of Ontario Archaeology, Canadian Museum of

Page viii

I~

.,.--.. ..

Civilization, Hull, Québec), Dr. J. Peter Denny (professor Emeritus, Department of

Psychology, The University of Western Ontario), Dr. Richard Preston (professor Emeritus,

Dept. of Anthropology, McMaster University, Ontario), Dr. Rhonda Telford (Historical

Research and Consulting, Hamilton, Ontario), W. A. Ross (formally Regional Archaeologist,

Ontario Ministry of Culture and Communications and currently Ross Archaeological

Research Associates, Thunder Bay, Ontario), Dr. Martin Magne (Manager, Cultural Resource

Services, Western and Northern Service Centre, Parks Canada), Dr. Cath Oberholtzer

(Conjunct Professor, Department of Anthropology, Trent University, Ontario), Dena

Doroszenko (Ontario Heritage Trust, Toronto, Ontario), Dr. 1. N. M. Wainwright

(Analytical Research Laboratory, Canadian Conservation lnstitute, Ottawa, Ontario),

Margaret McMahon (Water Statistics Analyst, Water Resources Division, Ontario Power

Generation), Mike Larson (Minnesota Department of Natural Resources, Minnesota), Rick

Cousins (Senior Water Resources Engineer, Lake of the Woods Secretariat, Lake of the

Woods Water Control Board, Canada), Dr. Daniel Arsenault (Départment d'histoire de l'art,

Université du Québec à Montréal, Québec), Dr. Beth A. S. Shook (Department of

Anthropology, CSU Chico, California, USA), Dr. Tim Jones (Executive Director,

Saskatchewan Archaeological Society, Saskatoon, Saskatchewan), Dr. lan A. Cowx

(International Fisheries lnstitute, University of Hull, UK), Dr. Charles A. Bishop (Research

Professor of Anthropology, Union College, Schenectady, New York, USA), Dr. John Long,

(Associate Professor, Faculty of Education, Nipissing University, Ontario), Dr. Jane Down

(Senior Conservation Scientist, Conservation Processes and Materials Research, Canadian

Conservation Institute, Ottawa, Ontario), Nancy Binnie (Conservation Scientist, Canadian

Conservation Institute, Ottawa, Ontario), Mark J. Dudzik (State Archaeologist, Fort Snelling

History Center, St. Paul, Minnesota, USA), Marie-Louise Perron (Library and Archives

Canada, Canadian Genealogy Centre, Library and Archives Canada, Ottawa, Ontario),

Jonathan Moore (Underwater Archaeology Service, Parks Canada Agency, Ottawa, Ontario),

Kari Layman (Hydraulics, USACE - St. Paul District, Minnesota, USA), Dr. Lisa Dillon

(Départment de demongraphiè, Université de Montréal, Québec), Dr. Sara Louri, and Hayley

Fitzsimmons for giving me references, information, archaeological reports, data of different

types, slides, files, and opportunities to discuss ideas.

l must also thank those archaeologists who had worked on sites in the Lake of the

Woods prior to myself and who kindly provided the references of their work: C. S. Reid

Page ix

I~

(formally Regional Archaeologist, Ontario Ministry of Culture and Communications, Kenora,

Ontario), Dr. Brian L. Molyneaux (Department of Anthropology, University of South

Dakota, Vermillion, South Dakota, USA), Dr. Jack Steinbring (Department of Anthropology,

Ripon College, Ripon, Wisconsin, USA), and William A. Fox (Field Unit Superintendent,

Western Arctic District, Parks Canada).

l must thank my field assistant Melinda Bell and my volunteer Elizabeth Campbell

for their help and assistance in the field. l must also thank W. A. Ross (Regional

Archaeologist, Thunder Bay, Ontario) for technical help and advice while in the field. l must

also thank David Fraser and the diving team from Sunset Diving, Kenora, Ontario for their

assistance while surveying and recording the pictograph sites. 1 must thank the Kast family

and everyone at Tomahawk Lodge and Marina, Wendal Dafcik and his guides of Crow Rock

Lodge, the owners and the guides of Monument Bay Lodge, Paul and Beverly Eliuk, Bill and

Billie-Jean Murphy from Sioux Narrows provided advice and expertise which made the

practical aspects of conducting field work possible.

l must also thank Professor Wendy Hall, Head of School, Professor Paul Lewis, and

Dr Kirk Martinez (Intelligence, Agents, Multimedia Group, School of Electronics and

Computing Science, University of Southampton, UK) for their help, guidance, and advice

when 1 used VIPS and for the experimental work 1 did in their laboratory. l must thank Dr

Mike Westmacott, Dr M. Faizal A. Fauzi, Dr David P Dupplaw, and Dr. Fazly Salleh Abas

who were doing research for their PhDs while 1 was there as well as Dr. David Millard and

Dr. Christopher Bailey for their technical assistance and for their questions because 1 realised

that the manner in which images are described fundamentally and directly affects its

subsequent interpretation.

1 must thank Dr. Andrew Sawyer (Hon. Research Fellow, Department of Museum

Studies, University of Leicester & mwr ltd) for teaching me how to use KLEIO IAS. l must

thank both Professor Manfred Thaller (professor in Computer Science for the Humanities,

Universitat zu KaIn, Germany) and Dr. Gerhardt Jaritz (Department of Medieval Studies,

Central European University, Central European University, Prague, Hungary) for help and

advice regarding the technical and theoretical aspects of KLEIO and KLEIO lAS.

l must thank Dr. Mima Kapches for permission to consult the Selwyn Dewdney

Papers in the New \Vorld Department Collection, in the Royal Ontario Museum, Toronto,

Ontario and Lori Nelson, the Director of the Lake of the Woods Museum in Kenora,

Page x

Ontario for letting me examine the papers of several early settlers in the Lake of the Woods

and using the photograph of the Massacre Scroll.

Numerous people gave me advice on how to track down a photograph of a birch

bark scroll and their combined efforts made the chapter on birch bark scrolls possible. l

must thank Moira McCaffrey (Director, Research and Exhibitions, McCord Museum of

Canadian History, Montréal, Québec); Dr. Jban Vastokas (professor Emerita, Department of

Anthropology, Trent University, Ontario), Dr. Jennifer S.H. Brown, (Director of The Centre

for Rupert's Land Studies/Department of History, The University of Winnipeg, Manitoba),

Dr. George Fulford (Department of Anthropology, The University of Winnipeg, Manitoba),

Dr. Sally Cole (Department of Anthropology and Sociology, Concordia University, Montréal,

Québec), Dr. J. c. H. King (Keeper, Department of Africa, Oceania and the Americas,

British Museum, London, England), Dr. Teresa Schenk (Center for the Study of Upper

Midwestern Cultures, University of Wisconsin, Madison, Wisconsin, USA), Lori Nelson

(Director, Lake of the Woods Museum, Kenora, Ontario), and finally Dr Heidi Bohdaker

(SSHRC Postdoctoral Fellow, Institute for Comparative Studies in Literature, Art and

Culture, Carleton University, Ontario).

l must thank several friends and colleagues who took cime from their own work to

read, comment, and discuss various versions of my thesis: Harry Lerner, Dr. Csilla Dallos

(Department of Anthropology, St. Thomas University, New Brunswick), Apryl

Wassaykeesic, Cecil Chabot, Loretta Flegel, Sheila Oakley, Yasir Khan, Dr. André LeBlanc

(Visiting Assistant Professor, History of Science and Technology Programme, University of

King's College, Nova Scotia) and Dr. Carolyn Podruchny (Department of History, York

University, Ontario). l must thank Elizabeth Campbell for editing this thesis, Francis Carson

who translated my abstract into French, and Neha Gupta and Rebekah Jobling who helped

convert sorne electronic fIles from one format to another.

l must thank the interlibrary loans department and the Blackadder-Lauterman,

Redpath, McClennan, Schulich libraries and particularly Rosa Orlandini and Joanna Hobbins

from the Walter Hitschfeld Geographie Information Centre for their assistance. l am grateful

for the continual assistance and guidance of Rose Marie Stano, Cynthia Romanyk, and Diane

Mann from the Department of Anthropology at McGill.

l must thank several people for technical support and advice regarding any computer

hardware problems that l had including Sean Colson (pC Consultants, Ryde, Isle of Wight,

Page xi

UK), University Bytes (Montréal, Quebec), Computech (Kenora, Ontario), Ian Shaw

(Managing Director of mwr InfoSecurity, Basingstoke, Hampshire, UK) and Joanna

Hobbins.

l must thank the following organizations for providing funding for m)' research: Max

Bell Fellowship for Canadian and Northern Studies, Sigma Xi Grant-in-Aid, Sigma Xi,

MacKenzie Ward Trust, Sioux Narrows, Ontario, Ontario Rock Art Conservation

Association, Canadian Historical Association, two Alma Mater Student Travel Grants, as weIl

as in-kind funding from the Ontario Ministry of Culture, Tourism and Recreation.

l must thank those friends who have not been involved in any aspect of my research

for having provided moral support: Alice Anderson, Denise Arthur, Andrea Cardarello, Sara

and Jim Busby, Nancy Chapman, Dr. Cristina Redko Meir, Dr. Caroline Gooch, Dr.

Doralice Barros Pereira, Angela Boville, Melissa Anderson, Ruth Greenaway, Gavin Taylour,

Judy Petch, Masami Waki, Rieko Sasa, Teresa deI Blanco, Nuzhat Dunn, Debra Irons, Sarah

Phillips, Lucy Braga, Rita Weidendorfer, Jennifer Famery, Heather Podruchny, Simone

Abuhab, Belinda Beaton, Alessandra Crippa, Eleanor Roff, and Dan Friesting. l come from a

large family and all of them have been incredibly supportive while undertaking this degree

but l must specifically thank a few of them: Professor Richard and Betty Preston, Finbar and

Susie Coison, my godparents Professor Henry and Mercedes Ettinghausen, my brother-in­

law Ian Shaw, my sister Beth Coison, and finally my parents Dr. R. Frank and Jean Coison

l would like to dedicate this the sis to two people: my paternal grandmother Kathleen

Cotter Colson and my maternaI grandfather MacKenzie Belfield Ward. My grandmother

taught me that it was important to stand up for the freedom of speech and thought no

matter how hard it rnight be to do this. My grandfather, MacKenzie Ward had a reputation

of being both strong rninded and highly opinionated. Both of us disagreed with each other

numerous times about many things but ultimately we both knew that it was better to agree to

disagree because our disagreements stemmed from each of us having different perspectives

and attitudes about the world and the way it functioned. He repeatedly told me that it was

better to stand up for what l believed, if l believed my perspective was weIl grounded and

researched, regardless of how hard it might be at the time than to buckle and back down

because of someone else's opinion.

Page xii

r--' ..

N ote on Spelling

The author of this thesis has used standard British English grammar, punctuation,

and spelling from the O:x.ford English Dictionary throughout her research and her thesis since

she started at McGill in September 1998.

Page xiü

Note on KLEIO lAS

The combination of the two letters 'xq' plus a number appear throughout several of the chapters in this thesis. This is not a special code. The letters 'xq' indicate that the result was provided as a consequence of a query numbered 'xq and a number'. It is a convention of KLEIO to use these letters and numbers as references to a particular database query.

Page xiv

List of Figures

Figure 1.1: Physical Location of the Lake of the Woods ............................................................... 2 Figure 1.2: Examples of shapes loosely categorised as 'crescents' and the sites that they come

from in the Lake of the Woods .................................................................................... 13-15 Figure 2.1: Cliffs in the southern end of Whitefish Bay with clearly defined high water levels ............................................................................................................................................................... 24 Figure 3.1: Map of the sites that Lawson recorded in 1885 ....................................................... 127 Figure 3.2: Lawson's published drawing of the images at DiKo-2, the petroglyph site. In:

Lawson, A. C 1885. Ancient Rock Inscriptions on the Lake of the Woods. Amm'can NaturalistXIX (7): 654-657.Pp.: 656 ................................................................................ 128

Figure 3.3: Lawson's published drawing of the images at DiKp-l, the pictograph site from 1885. Lawson, A. C 1885. Ancient Rock Inscriptions on the Lake of the Woods. American Naturalist XIX (7): 654-657. Pp: 657 ............................................................... 128

Figure 3.4: Annotated photograph of DiKp-l taken in June 2001 ........................................... 129 Figure 3.5: The four sites that Cameron recorded in 1978 ........................................................ 131 Figure 3.6: Fox's photograph of DhKm-l in 1974. Courtesy of Ontario Ministry of Culture

and Communications .......................................................................................................... 133 Figure 3.7: Dewdney's 15 July 1960 field record of Site # 92A, now called DhKm-l and

DhKm-4. © Royal Ontario Museum, Toronto Canada .............................................. .135 Figure 3.8: A field recording by Dewdney of two pictograph sites in Lake of the Woods, Site

#70 Sunset Channel (DiKp-l) and Site # 69 Picture Rock Island (DhKm-3). © Royal Ontario Museum, Toronto Canada ...................................................................... 136

Figure 3.9: Dewdney's field drawing of Site # 28, undertaken on 15 August 1953. © Royal Ontario Museum, Toronto Canada .................................................................................. 138

Figure 3.10: Face 2 and 3 of Site # 28 by Dewdney in 1953. © Royal Ontario Museum, Toronto Canada .................................................................................................................. 139

Figure 3.11: Dewdney's field drawing of Site #106 undertaken on August 19 1960. © Royal Ontario Museum, Toronto Canada .................................................................................. 139

Figure 3.12: The Close Proximity of DhKm-4 to DhKm-l ...................................................... 140 Figure 3.13: Annotated photograph of DhKm-l in 2001 by Cols on according to Dewdney's

divisions ................................................................................................................................ 141 Figure 3.14: Annotated photograph by toIson of DhKm-1 taken in 2001 according to

Lambert's divisions ............................................................................................................. 142 Figure 3.15: Dewdney's field drawing of Site # 69 undertaken on 8 August 1959. © Royal

Ontario Museum, Toronto Canada .................................................................................. 142 Figure 3.16: Part 1 of Dewdney's field drawing of Site # 91A and B, undertaken on 15 July

1960. © Royal Ontario Museum, Toronto Canada ....................................................... 143 Figure 3.17: Part 2 of Dewdney's field drawing of Site # 91A and B, undertaken on 15 July

1960. © Royal Ontario Museum, Toronto Canada ....................................................... 143 Figure 3.18: Setting with offerings and CD., Site # 29 by Dewdney. © Royal Ontario

Mùseum, Toronto Canada ................................................................................................. 144 Figure 3.19: Photôgraph of part of DjKn-1 called Site # 29 taken by Dewdney in 1959. ©

Royal Ontario Museum, Toronto Canada ...................................................................... 144 Figure 3.20: Part of DjKn-l,July 2001 ......................................................................................... 145 Figure 3.21: Dewdney's 1964 field drawing of Site # 197 (DgKl-l) and # 198 (DgKl-2). ©

Royal Ontario Museum, Toronto Canada ...................................................................... 146 Figure 3.22: Photograph of DgKl-l in July 2001 with Pastershank's 1989 subdivisions ...... 146

Page xv

Figure 3.23: DgKl-2 in 1989 bl' Pastershank. Courtes)' of Ontario Ministry of Culture and Communications ................................................................................................................. 147

Figure 3.24: Dewdney's published images from this site. In: Dewdney S. and K. Kidd. 1967. Indian Paintings if the Great Lakes. 2nd Edition Toronto: University of Toronto Press. p.l10 ..................................................................................................................................... 148

Figure 3.25: Three of the new images at DgKl-2 in July 2001 .................................................. 148 Figure 3.26: The new image at DgKl-2 not painted upon the older ochre images in July 2001

............................................................................................................................................... 149 Figure 3.27: Annotation by Coison in 2004 of DgKl-2 taken by Pastershank in 1989 ......... 150 Figure 3.28: Field recording of detail of Face l, DgKI-2, Pastershank in 1989. In:

Pastershank, G. 1989. An Archaeological Sun1ry if Sabaskong B'!)', Lake if the Woods. Ontario Conservation Archaeology Report #13, North-western Region. Kenora: Ontario Ministry of Culture and Communications. p. 56 ............................................ 150

Figure 3.29: Photograph of Face 1 by Pastershank in 1989. Courtesy of Ontario Ministry of Culture and Communications ............ ; ........................................................................... 151

Figure 3.30: Annotation by Coison in 2004 of photograph, of Face II by Pastershank in 1989. Courtesy of Ontario 1linistry of Culture and Communications .............................. 152

Figure 3.31: Field drawing with detail of Morph 4 at DgKl-2. In: Pastershank, G. 1989. An Archaeological S urvry if S abaskong B'!)', Lake if the Woods. Ontario Conservation Archaeology Report #13, North-western Region. Kenora: Ontario Ministry of Culture and Communications. p. 58 ................................................................................ 152

Figure 3.32: Field recording of detail of Face III. In: Pastershank, G. 1989. An Archaeological Survry if Sabaskong B'!)'J Lake if the Woods. Ontario Conservation Archaeology Report #13, North-western Region. Kenora: Ontario Ministry of Culture and Communications. p. 60 ...................................................................................................... 153

Figure 3.33: Field recording of Face IV of DgKl-2. In: Pastershank, G. 1989. An Archaeological Survry if Sabaskong B'!)'J Lake if the Woods. Ontario Conservation Archaeology Report #13, North-western Region. Kenora: Ontario Ministry of Culture and Communications. p. 61 ................................................................................ 154

Figure 3.34: DgKI-2, Detail of Face IV by Pastershank in 1989. Courtesy of Ontario Ministry of Culture and Communications ...................................................................................... 155

Figure 3.35: Pastershank's Field recordings of Faces V, VI and VII. In: Pastershank, G. 1989. An Archaeological Survry if Sabaskong B'!)'J Lake if the Woods. Ontario Conservation Archaeology Report #13, North-western Region. Kenora: Ontario Ministry of Culture and Communications. p. 62 ................................................................................ 155

Figure 3.36: U sing Pastershank's (1989) subdivisions of rock surface to divide the paintings in 2001 into Faces V, VI and VIl ..................................................................................... 156

Figure 3.37: Face l of DgKl-17 by Pastershank in 1989. Courtesy of Ontario Ministry of Culture and Communications ........................................................................................... 158

Figure 3.38: DgKl-17 _ Pastershank's FaceII and II in 1989. In: Pastershank, G. 1989. An Archaeological Survry of Sabaskong B'!)'J Lake of the Woods. Ontario Conservation Archaeology Report #13, North-western Region. Kenora: Ontario Ministry of Culture and Communications. p. 66 ................................................................................ 159

Figure 3.39: Pastershank divided this Face into NO parts called Panel i and ü in 1989. In: Pastershank, G. 1989. An Archaeological Survry if Sabaskong B'!)'J Lake if the Woods. Ontario Conservation Archaeology Report #13, North-western Region. Kenora: Ontario Ministry of Culture and Communications. p.68 .............................................. 159

Figure 3.40: DhKm-5 in June 2001 ............................................................................................... 160

Page xvi

.~ ..

Figure 3.41: The new part ofDhKm-5 ......................................................................................... 161 Figure 3.42: Dewdney working with a colleague on an improvised platform at Site #108. ©

Royal Ontario Museum, Toronto Canada ...................................................................... 164 Figure 3.43: The plastic being taped to the surface of the cliff face and the mineraI

precipitation and patination being recorded by a black pen by Rusak in 1991. Courtesy of Ontario Ministry of Culture and Communications .................................. 167

Figure 3.44: DiKp-l, Sunset Channel, 25th July 1975 © Canadian Conservation Institute . .170 Figure 3.45: The table drawn by Dewdney to illustrate how he classified the images at these

sites. In: Dewdney S. and K. Kidd. 1962. lndian Paintings of the Great Lakes. Toronto: University of Toronto Press, p. 18 ................................................................................... 174

Figure 3.46: The Massacre Scroll photographed in 1993 by J. Rusak. Collections of the Lake of the Woods Museum, Kenora, Ontario ....................................................... : .............. .190

Figure 3.47: Part of DhKn-l beside Obabikon Channel ........................................................... l92 Figure 3.48: The serpents and the turtles described by Dewdney in 1960 .............................. 199 Figure 4.1: Examples of cliff faces examined for pictographs during 2001 ............................ 220 Figure 4.2: DhKm-18 (an "oId" site) and DhKm-20 Ca "new" site) ....................................... .221 Figure 4.3: A map of where the diving took place ...................................................................... 222 Figure 4.4: Overview of DgKI-2 taken in 2001 ........................................................................... 224 Figure 4.5: Physical location of DgKI-2 in Sabaskong Bay, Lake of the Woods. C4nadian

Hydrographic Services. "Sabaskong Bay, Lake of the Woods, Southeastern Channel" [hydrographic chart]. 1 :40,000. Sheet 6214. Ottawa: Department of Oceans and Fisheries, 1989 .................................................................................................................... 225

Figure 4.6: Field drawing of DgKI-2 by M. Bell in 2001 ............................................................ 229 Figure 4.7: Red/orange lichen on the western end of Crowrock Island ................................. .233 Figure 4.8: Part of DjKn-l used as the control file ..................................................................... 234 Figure 4.9: DiKp-l in June 2001 annotated by Colson in 2002 to indicate locations of previously unknown images ......................... ; .............................................. 241 Figure 4.10: Map indicating location of other known archaeologïcal sites to pictograph sites ............................................................................................................................................................. 253 Figure 5.1: DkIZn-6 in 2001 ............................................................................................................ 256 Figure 5.2: The different shapes of 'creatures' at DhKn-l in 2001 .......................................... 257 Figure 5.3: Overview of DhKm-4 in Whitefish Bay in 2001 ..................................................... 257 Figure 5.4: DhKm-l has images affected by exfoliation, white mineraI deposits and rock tripe

and lichen encroachment in 2001 .................................................................................... .258 Figure 5.5: The exfoliated images at DhKm-l in 2001. © Elizabeth Campbell 2001 ........... .259 Figure 5.6: The smear at DhKm-l in 2001. © Elizabeth Campbell 2001 .............................. .259 Figure 5.7: The upper part of DiKn-1 in 2001. © Elizabeth Campbell 2001 ......................... 260 Figure 5.8: Severe deposition of white mineraI deposits at DiKq-l 0 in 2001 ......................... 261 Figure 5.9: AIl of the images at DhKm-3 in 2001. © Elizabeth Campbell 2001 ............ 262-263 Figure 5.10: Screenshot of some of the red images below the white mineraI deposits ......... 264 Figure 5.11: If this screen shot is compared with the screen shot above it is clear that the

accumulation of white mineraI deposits at DhKm-S is not uniform .......................... 264 Figure 5.12: Screenshot showing what occurs to the same im4ges using different tools when 1

mathematically tried to rem ove the white pixels deemed 'noise" ................................ 266 Figure 5.13: A map of the Lake of the Woods area showing the numbers of images at each

site ......................................................................................................................................... 267 Figure 5.14: D1Jn-l from Harmon Lake Courtesy of Ontario Ministry of Culture and

Communications ................................................................................................................. 270

Page xvii

Figure 5.15: DjKp-l0 on Moberly Lake. Courtesy of Ontario J\1inistry of Culture and Communications ................................................................................................................. 270

Figure 5.16: DjJp-l on Moberly Lake. Courtesy of Ontario J\1inistry of Culture and Communications ................................................................................................................. 271

Figure 5.17: Close up of a rectangle with the circle and the diagonallines fromDhKm-3 in 2001 ....................................................................................................................................... 276

Figure 5.18: The rectangle with the stick figure at DhKo-l in 2001 ........................................ 277 Figure 5.19: Sites where offerings were found in 2001 .............................................................. .281 Figure 5.20: Sites where tobacco was found as an offering in 2001 ......................................... 282 Figure 5.21: Sites where money was found as an offering in 2001 ........................................... 283 Figure 5.22: Offerings discovered by Fox in 1974 at DhKm-1. Courtesy of Ontario J\1inistry

of Culture and Communications ...................................................................................... 284 Figure 5.23: Sites where clothing was found as offerings in 2001 ............................................ 285 Figure 5.24: Sites where liquor botdes were found as offerings in 2001 ................................. 286 Figure 5.25: Offerings that Dewdney found at DjKq-l0 called Site # 95, Picture Rock Point

in 1960. © Royal Ontario Museum, Toronto Canada ................................................... 287 Figure 5.26: Offerings that Dewdney found at DjKq-l0 called Site # 95, Picture Rock Point

in 1960 .................................................................................................................................. 287 Figure 5.27: The lower part of the DiKn-l, Hayter's Point in 2001.. ....................................... 288 Figure 5.28: The largest pile of the offerings found at DgKl-2 in June 2001.. ....................... .289 Figure 5.29: Offerings stuffed in the crevice at DhKm-1 in 2001 © Elizabeth Campbell 2001

............................................................................................................................................... 290 Figure 5.30: The Bowl near DiKm-50 © Elizabeth Campbell 2001 ........................................ 291 Figure 5.31: The Marker's Mark © Elizabeth Campbell 2001 ................................................... 291 Figure 5.32: The Vial © Elizabeth Campbell 2001 ..................................................................... .292 Figure 5.33: The Hand Tooled Top © Elizabeth Campbell 2001 ............................................ 292 Figure 5.34: Close-up of Top © Elizabeth Campbell 2001 ....................................................... 292 Figure 5.35: The Lead Shot © Elizabeth Campbell 2001 ........................................................... 292 Figure 5.36: The Vial Opening © Elizabeth Campbell 2001 .................................................... .292 Figure 5.37: Outside of Pot © Elizabeth Campbell 2001 .......................................................... 293 Figure 5.38: Inside of Pot © Elizabeth Campbell 2001 .............................................................. 293 Figure 5.39: Side view of Pot © Elizabeth Campbell 2001 ........................................................ 293 Figure 5.40: Outside of Pot © Elizabeth Campbell 2001 .......................................................... 294 Figure 5.41: Inside of Pot © Elizabeth Campbell 2001 .............................................................. 294 Figure 5.42: Surface of Pot © Elizabeth Campbell 2001 .......................................................... .294 Figure 5.43: Base of Pot © Elizabeth Campbell 2001 ................................................................ 294 Figure 5.44: Rim of Pot © Elizabeth Campbell 2001 ................................................................. 295 Figure 5.45: Setting of Site # 197. © Royal Ontario Museum, Toronto Canada ................... 299 Figure 5.46: Face 1 Site # 197 by Dewdney in 1964. © Royal Ontario Museum, Toronto

Canada .................................................................................................................................. 300 Figure 5.47: Face 2, Site # 197 by Dewdney in 1964. © Royal Ontario Museum, Toronto

Canada .................................................................................................................................. 300 Figure 5.48: Face 1 of DgKl-1 and Face 2 of DgKl-2 taken by Pastershank in 1989. Courtesy

of Ontario J\1inistry of Culture and Communications .................................................. 301 Figure 5.49: Face 3 of DgKl-1 by Pastershank in 1989. Courtesy of Ontario J\1inistry of

Culture and Communications ........................................................................................... 301 Figure 5.50: Photograph of DgKl-2 in 2001 ................................................................... : ............ 302

Page xviii

Figure 5.51: Face 1 of Site # 198 by Dewdney in 1964. © Royal Ontario Museum, Toronto Canada .................................................................................................................................. 303

Figure 5.52: "Cave" at Site # 198 in 1964 by Dewdney. © Royal Ontario Museum, Toronto Canada .................................................................................................................................. 303

Figure 5.53: Pastershank's Face 1 by Pastershank in 1989. Courtesy of Ontario Ministry of Culture and Communications .................................... : ...................................................... 304

Figure 5.54: Faces II, III and IV by Pastershank in 1989. Courtesy of Ontario Ministry of Culture and Communications ........................................................................................... 305

Figure 5.55: Setting from distance of Site # 28, Sioux Narrows by Dewdney in 1953. Courtesy of Ontario Ministry of Culture and Communications .................................. 306

Figure 5.56: Face 2 and 3 of Site # 28 by Dewdney in 1953. Courtesy of Ontario Ministry of Culture and Communications ........................................................................................... 307

Figure 5.57: Close-up of DiKm-3 taken in 1991 ......................................................................... 307 Figure 5.58: Close-up of DiKm-3 taken in 2001 ......................................................................... 308 Figure 5.59: Face 2, Site # 106 by Dewdney in 1960. © Royal Ontario Museum, Toronto

Canada .................................................................................................................................. 309 Figure 5.60: Site # 106, Sabaskong-Obabikon Channel, Watercolour reproduction of Face II

by Dewdney. © Royal Ontario Museum, Toronto Canada .......................................... 309 Figure 5.61: Close-up ofleft hand part of DhKn-1 in 2001 ...................................................... 310 Figure 5.62: Close-up of right hand part of DhKn-1 in 2001 ................................................... 311 Figure 5.63: Setting "the hole" at Devil's Hole Site #92 by Dewdney in 1960. © Royal

Ontario Museum, Toronto Canada .................................................................................. 312 Figure 5.64: Setting with W. F. and offerings beside Site # 92 by Dewdney in 1960. © Royal

Ontario Museum, Toronto Canada .................................................................................. 312 Figure 5.65: DhKm-1 taken in September 1974 by W. Fox. Courtesy of Ontario Ministry of

Culture and Communications ........................................................................................... 313 Figure 5.66: DhKm-1 in June 2001 ............................................................................................... 314 Figure 5.67: Thompson's Drawings of the images at the pictograph site. © Royal Ontario

Museum, Toronto Canada ................................................................................................. 316 Figure 5.68: DhKo-1 by K. Campbell in 1951. © K.athleen Campbe1l2001 .......................... 317 Figure 5.69: Dewdney's field recording of Site # 108 undertaken on August 20 August 1960.

© Royal Ontario Museum, Toronto Canada .................................................................. 318 Figure 5.70: Dewdney's photograph of Site # 108. © Royal Ontario Museum, Toronto

Canada .................................................................................................................................. 319 Figure 5.71: Section of DhKo-1 taken by Molyneaux. In: Molyneaux, B. L. 1987. The Lake of

the Painted Cave Archaeology 40 (4) p.20 ..................................................................... ;319 Figure 5.72: Overview of DhKo-1 in 2001 .................................................................................. 320 Figure 5.73: Dewdney's Site # 93 wruch he called an "offering site" in July 1960. © Royal

Ontario Museum, Toronto Canada .................................................................................. 321 Figure 5.74: Setting of Site # 93 by Dewdney in 1960. © Royal Ontario Museum, Toronto

Canada .................................................................................................................................. 321 Figure 5.75: Jerry Corpine at Site #93 by Dewdney in 1960. © Royal Ontario Museum,

Toronto Canada .................................................................................................................. 322 Figure 5.76: DiKm-4 by Moll'neaux. In: Molyneaux, B. L. 1980a. Landscape Images Rotunda

13 (3): 10 .................................................................. ; ........................................................... 323 Figure 5.77: DiKm-4 bl' Pelshea in 1978. Courtesy of Ontario Ministry of Culture and

Communications ................................................................................................................. 323 Figure 5.78: DiKm-4 taken in 1991 ............................................................................................... 324

Page xix

Figure 5.79: DiKm-4 taken in 2001 ............................................................................................... 324 Figure 5.80: Dewdney's field drawing of Site # 69 called Picture Rock Island undertaken on

8 August 1959. © Royal Ontario Museum, Toronto Canada ...................................... 325 Figure 5.81: Face lof Site # 69 in Whitefish Bay, Lake of the Woods taken by Dewdney in

1959. © Royal Ontario Museum, Toronto Canada ....................................................... 325 Figure 5.82: Face II of Site # 69 taken by Dewdney in 1959. © Royal Ontario Museum,

Toronto Canada .................................................................................................................. 326 Figure 5.83: Water colour reproduction of Face III of Site # 69 by Dewdney in 1959. ©

Royal Ontario Museum, Toronto Canada ...................................................................... 327 Figure 5.84: Black and White photograph of DhKm-3 taken by Molyneaux. In: Molyneaux,

B. L. 1980a. Landscape Images Rotunda 13 (3): 7 ........................................................... 327 Figure 5.85: The images at DhKm-3 taken by Pelshea in 1978. Courtesy of Ontario Ministry

of Culture and Communications ...................................................................................... 328 Figure 5.86: Left hand side of DhKm-3 taken in 2001 .............................................................. 329 Figure 5.87: Annotated section of DhKm-3 by Colson taken in 2001 .................................... 329 Figure 5.88: Right hand end of DhKm-3 ..................................................................................... 330 Figure 5.89: Face 1- Figure a & b, Site # 92A taken by Dewdney in 1960. © Royal Ontario

Museum, Toronto Canada ................................................................................................. 331 Figure 5.90: Left hand side of DhKm-4 in 2001 ......................................................................... 332 Figure 5.91: Right hand side of DhKm-4 in 2001 ....................................................................... 332 Figure 5.92: Dewdney's field drawing of Site # 105 undertaken on 19 August 1960. © Royal

Ontario Museum, Toronto Canada .................................................................................. 333 Figure 5.93: Close view of Site # 105 taken by Dewdney in 1960. © Royal Ontario Museum,

Toronto Canada .................................................................................................................. 334 Figure 5.94: Closer view of DhKm-5 by Pelshea in 1978 .. © Royal Ontario Museum, Toronto

Canada .......................................................... ; ....................................................................... 334 Figure 5.95: Molyneaux's photograph of DhKm-5. © Royal Ontario Museum, Toronto

Canada .................................................................................................................................. 335 Figure 5.96: Photograph of DhKm-5 taken in 2001. © Royal Ontario Museum, Toronto

Canada .................................................................................................................................. 337 Figure 5.97: Dewdney's field drawing of Site # 70 in Sunset Channel undertaken on August

12, 1959. © Royal Ontario Museum, Toronto Canada .................................... ; ........... .337 Figure 5.98: Photograph of DiKp-l taken by C. S. Reid in May 1975. © Royal Ontario

Museum, Toronto Canada ................................................................................................. 339 Figure 5.99: DiKp-l, Sunset Channel, 25thJuly 1975 annotated by A. Colson in 2001 ....... 340 Figure 5.100: Photograph of DiKp-l taken in June 2001 .......................................................... 340 Figure 5.101: Lichen growth and encroachment as marked by A Colson in 2001 .......... ~ ...... 341 Figure 5.1 02: The areas containing missing images at DiKp-l indicated by Colson ............. 342 Figure 5.103: DgKm-l in Splitrock Bay taken in 2001 ............................................................... 343 Figure 5.104: Left hand part ofDgKI-17 ...................................................................................... 344 Figure 5.105: Right hand part of DgKl-17 ................................................................................... 344 Figure 5.106: Cyclone Point (unknown photographer and unknown date). © Royal Ontario

Museum, Toronto Canada ................................................................................................. 345 Figure 5.107: DgKl-19 by Pastershank taken in 1989. Courtesy of Ontario Ministry of

Culture and Communications ........................................................................................... 346 Figure 5.108: DgKl-19 photographed in 2001 ............................................................................. 346 Figure 5.109: Annotation by Colson of Site # 95 taken by Dewdney in 1960. © Royal

Ontario Museum, Toronto Canada .................................................................................. 347

Page xx

Figure 5.110: Left hand side and right hand side of DiKq-l0 in 2001 ..................................... 348 Figure 5.111: "Setting of Site # 94 with P.D." Photograph taken by Dewdney in 1960. ©

Royal Ontario Museum, Toronto Canada ...................................................................... 349 Figure 5.112: Annotated setting of DjKr-23 in 2001 by Colson ............................................... 349 Figure 5.113: DjKr-23 in July 2001 ................................................................................................ 350 Figure 5.114: Dewdney field drawing of Site # 104 undertaken on 19 August in 1960. ©

Royal Ontario Museum, Toronto Canada ...................................................................... 351 Figure 5.115: Setting of Site#- 104 taken by Dewdney (© Royal Ontario Museum, Toronto

Canada) annotated by A. Coison on the right ................................................................ 351 Figure 5.116: Devil's Bay Site # 104 taken by Dewdney © Royal Ontario Museum, Toronto

Canada .................................................................................................................................. 352 Figure 5.117: Bird identified by Dewdney and Lambert beside the two vertical stick identified

in 2001 .................................................................................................................................. 353 Figure 5.118: Smear at DhKm-18 in 2001 ........................ ; .......................................................... 353

~ .. Figure 5.119: Oblong blob at DhKm-18 in 2001 ........................................................................ 354 Figure 5.120: Close up of Site # 270 Horseshoe Island in Whitefish Bay by Dewdney. ©

Royal Ontario Museum, Toronto Canada ...................................................................... 355 Figure 5.121: DiKm-51 taken in 2001. © Elizabeth Campbell 2001.. ...................................... 355 Figure 5.122: Annotated by Colson in 2001 ofDiKm-51 taken in 2001.. ............................... 356 Figure 5.123: Annotated by Colson in 2001 ofDiKm-51 taken by Dewdney ....................... .356 Figure 5.124: DkKn-7 in 2001 ........................................................................................................ 357 Figure 5.125: DkI<n-6 in 2001 ........................................................................................................ 358 Figure 5.126: Dewdney's field drawing of Site # 107 in Obabika Narrows undertaken on

August 191960. © Royal Ontario Museum, Toronto Canada .................................... 359 Figure 5.127: Part of DjI<n-l taken by Molyneaux. In: Molyneaux, B. L. 1980a. Landscape

Images Rotunda 13 (3): 9 ..................................................................................................... 360 Figure 5.128: Pelshea's photograph of the le ft hand side of DjI<n-l taken in 1978. Courtesy

of Ontario Ministry of Culture and Communications .................................................. 361 Figure 5.129: Pelshea's photograph of the right hand side of DjI<n-l taken in 1978. Courtesy

of Ontario Ministry of Culture and Communications .................................................. 362 Figure 5.130: Part of DjI<n-l in June 2001 .................................................................................. 363 Figure 5.131: Part of DjI<n-l in June 2001 .................................................................................. 363 Figure 5.132: Part of DjI<n-l in June 2001 .................................................................................. 364 Figure 5.133: Setting from a Distance of Site # 91A and B taken by Dewdney. © Royal

Ontario Museum, Toronto Canada .................................................................................. 365 Figure 5.134: "Face lB figure c + d" by Dewdney in 1960. © Royal Ontario Museum,

Toronto Canada .................................................................................................................. 365 Figure 5.135: "Face lA fig a + b" by Dewdney in 1960. © Royal Ontario Museum, Toronto

Canada .................................................................................................................................. 366 Figure 5.136: DiKn-l in J une 2001 ................................................................................................ 366 Figure 5.137: Blob beside exfoliated area at DiI<n-l in 2001 .................................................... 367 Figure 5.138: Dewdney's "Face lA fig a + b" at DiI<n-l in 2001.. .......................................... 367 Figure 5.139: Red images are visible although partially hidden by the lichen and rock tripe in

the upper half of DiKn-l in 2001 .................................................................................... 368 Figure 5.140: Shapes above the main collection of shapes at DiI<n-l ..................................... 368 Figure 5.141: DhKm-20 in June 2001 ........................................................................................... 369 Figure 5.142: DhKm-19 in 2001 .................................................................................................... 370 Figure 5.143: DgKm-2 in July 2001 ............................................................................................... 371

Page xxi

Figure 5.144: DiKm-49 in June 2001 .................................................... : ...................................... ;372 Figure 5.145: DiKin-50 in June 2001 ............................................................................................ 373 Figure 5.146: DhKrn-20 created by R, Bird and signed in 1982 ............................................... 374 Figure 5.147: Two of the modern solid spray painted images at DgKl-2 ................................ 374 Figure 5.148: The third modern image which is solid at DgKl-2 .............................................. 375 Figure 5.149: The modern hollow spray painted image at DgKl-2 ........................................... 375 Figure 5.150: Modern spray paint image at DhKrn-20 ............................................................... 376 Figure 6.1: DhKm-l photographed August 27'h 2003 ................................................................ 398 Figure 6.2: Annotated photograph by Colson of the holes which are called caves at DgKI-2

in the rock above the water level.. ................................................................................... .403 Figure 6.3: Inside the cave at DgKl-2 ........................................................................................... .403 Figure 6.4: Map of the pictograph sites that have deep holes in 2001 .................................... .404 Figure 6.5: Map of pictograph sites with large cracks in 2001 .................................................. .405 Figure 6.6: Map of pictograph sites with crevices in 2001 ........................................................ .406 Figure 6.7: Map of pictograph sites with overhangs in 2001 ................................................... ..407 Figure 6.8: Map of pictograph sites with rock shelves, or benches or ledges in 2001 .......... .409 Figure 6.9: Map of pictograph sites with rocks at their base in 2001 ...................................... .411 Figure 6.10: DiKm- 50 on Horseshoe Island, Whitefish Bay ................................................... .412 Figure 6.11: The black marks on the surface of the rock's surface ........................... ..413 Figure 6.12: Sites with smears ......................................................................................................... 419 Figure 6.13: Sites with blobs ........................................................................................................... 421 Figure 6:14: Sites with unidentifiable shapes .............................................................................. .423 Figure 6.15: Sites with vertical stick figures ................................................................................. .424 Figure 6.16: Sites that the shape of the 'upper torso of a vertical stick figure' ...................... .425 Figure 6.17: Sites that had the shape of hand prints .................................................................. .425 Figure 6.18: Location of sites that had both a left and a right hand print shape ................... .426 Figure 6.19: Sites with shapes of loosely categorised creatures ............................................... ..427 Figure 6.20: An example of a shape called 'one vertical and one horizontal short line that

cross each other at ninety degrees' ................................................................................... 430 Figure 6.21: Sites with one vertical and one horizontal short line that cross each other at

ninety degrees ...................................................................................................................... 431 Figure 6.22: An example of a shape called 'one vertical and one horizontal short line that

cross each other at ninety degrees with short lines at the end of each of the vertical and horizontallines' ........................................................................................................... 431

Figure 6.23: Sites with the shape called one vertical and one horizontal short line cross each other at ninety degree angle with short lines at the end of each of the vertical and horizontallines .................................................................................................................... 432

Figure 6.24: Sites with groups of four diagonallines ................................................................. .433 Figure 6.25: Sites with shapes called 'three lines joined at one end' ........................................ .434 Figure 7.1: The left hand part of scroll number 1. Courtesy of Lake of the Woods Museum,

Kenora, Ontario, Canada ................................................................................................... 452 Figure 7.2: The right hand part of scroll number 1. Courtes)' of Lake of the Woods

Museum, Kenora, Ontario, Canada ................................................................................ .453 Figure 7.3: One of the scrolls that Densmore photographed. In: Densmore, F. 1910.

Chippewa Music I. Washington Government Printing Office. Srnithsonian Institution, Bureau of American Ethnology. Bulletin 45. p. 110 ................................ .453

Page xxii

Figure 7.4: One of the serolls that Densmore photographed. In: Densmore, F. 1910. Chippewa Music I. Washington Govemment Printing Office. Smithsonian Institution, Bureau of Ameriean Ethnology. Bulletin 45. p. 110 ................................ .453

Figure 7.5: The scroll published by Dewdney in 1970. In: Dewdney, S. 1970c. Ecologicai Notes on the Objiway shaman-artist. Artscanada. August: 17-28. P.: 27-28 .............. .454

Figure 7.6: Dewdney's tracing of the scroll in figure 7.5. In: Dewdney, S. 1975b. The Sacred Scroils of the Southem Ojibwqy. Toronto and Buffalo: University of Toronto Press: Figure 146 Song record with herbaI references (ROM -3), on p. 141 ....................... .455

Figure 7.7: The nine differences between the Dewdney's drawing and the original seroll .. .455 Figure 7.8: Scroll number 2 annotated by Colson from left to right in 2005 ......................... .464 Figure 7.9: Mnemonic for the Song of Good Medicine. In: Densmore, F. 1910. Chippewa

Music 1. Washington Government Printing Office. Smithsonian Institution, Bureau of American Ethnology. Bulletin 45. p. 1 00 ....................................................................... .464

Figure 7.10: Mnemonic for the Song of the Crab Medicine Bag. In: Densmore, F. 1910. Chippewa Music 1. Washington Government Printing Office. Smithsoruan Institution, Bureau of American Ethnology. Bulletin 45. p. 102 ..................................................... .464

Figure 7.11: Mnemonie for the Song of the Fire-Charm. In: Densmore, F. 1910. ChippelJJa Music 1. Washington Government Printing Office. Smithsoruan Institution, Bureau of American Ethnology. Bulletin 45. p. 103 ....................................................................... .465

Figure 7.12: Scroll number 3 annotated by Colson in 2005. In: Densmore, F. 1910. Chippewa Music 1. Washington Government Printing Office. Smithsonian Institution, Bureau of Ameriean Ethnology. Bulletin 45. p. 106 ....................................................................... .465·

Figure 7.13: Mnemonic for the Song of the Owl Medicine. In: Densmore, F. 1910. Chippe1va Music 1. Washington Government Printing Office. Smithsoruan Institution, Bureau of American Ethnology. Bulletin 45. p. 1 04 ....................................................................... .465

Figure 7.14: Mnemonic for the Song of Starvation. In: Densmore, F. 1910. Chippewa Music 1. Washington Govemment Printing Office. Smithsoruan Institution, Bureau of American Ethnology. Bulletin 45. p.l 04 ........................................................................ .466

Figure 7.15: Mnemonic for the Song of the Man who Succeeded. In: Densmore, F. 1910. Chippewa Music 1. Washington Govemment Printing Office. Smithsonian Institution, Bureau of American Ethnology. Bulletin 45. p. 98 ....................................................... .466

Figure 7.16: Seroll number 4 annotated by Coison in 2005 ...................................................... .467 Figure 7.17: The shapes at DiKm-4 according to Ann Seymour ............................................. .468 Figure 7.18: The turde, snake, and the first lattice discussed by Seymour at DhKm-3 ........ .469 Figure 7.19: The thunderbird, line, handprint, lattice, and the human figure with hands raised

upwards at DhKm-3 identified by Seymour. ................................................................. .470 Figure 7.20: The annotated photographs of images at DhKn-l .............................................. .471 Figure 7.21: An annotated photograph ofDhKo-l ................................................................... .471

Page xxiii

List of Tables:

Table 1.1: Sequence in which approaches must be utilised ............................................................ 6 Table 6.1: Hierarchical Cluster Analysis - dendrogram using Average Linkage (Between

Groups) using Euclidean distance method to establish whether a relationship exists between the sites and the different features .................................................................. .401

Table 6.2 Hierarchical Cluster Analysis - dendrogram using Average Linkage (Between Groups) using Pearson correlation method to establish whether a relationship exists between the sites and the different features .................................................................. .401

Table 6.3 Hierarchical Cluster Analysis - dendrogram using Average Linkage (Between Groups) using Euclidean distance method to determine the nature of the relationship between smears and sites ................................................................................................. 420

Table 6.4: Hierarchical Cluster Analysis - dendrogram using Average Linkage (Between Groups) using Euclidean distance method to determine the nature of the relationship between smears and sites .................................................................................................. 420

Page xxiv

,~.

Chapter 1 - Introduction

While most people who study rock image sites tend to be interested in the meaning

of images, much more empirical information can be obtained about these images and their

physical location. Very little of the work done in the past on rock image sites has been

systematic. Hence, in this thesis l will address the dearth of detailed information on the

images and their context. l will subsequently suggest a new approach to studying rock images

that is as systematic, rigorous and objective as possible using a group of sites in the Canadian

Shield. The goal of this thesis is to rigorously conduct a detailed and thorough examination

of the images of the pictograph sites in the Lake of the Woods region to (a) identify the

possible vocabulary of images, (b) determine whether combinator)' mIes exist, (c)

reconstitute thelife history of each site, and (d) ascertain whether the images can be related

to other images outside the rock image sites to de termine if this can provide information

about the meaning(s) of the rock images.

Where are these sites?

The rock image sites of the Lake of the Woods region exhibit traits evident in rock

Image studies in other parts of the world. Here, as elsewhere, the choice of theoretical

approach influences the fieldwork, analysis, and search for meaning. This study is based on

data collected during three months of fieldwork conducted on the pictographs in the Lake of

the Woods in 2001. Petroglyph sites were not included in this study because the high water

levels in the field season of 2001 inundated these sites. l did not examine petroforms since

this study only concerns images on rocks and not images created using rocks. Only the

pictograph sites were examined.

l chose the sites of the Lake of the Woods area, in Northwestern Ontario, because

they appear representative of the data resulting from the archaeological surveys conducted

on the rock images of the Canadian Shield. The map below, in figure 1.1, indicates the

physicallocation of the Lake of the Woods with respect to the Great Lakes and Hudson Bay.

Page 1

,-MANtTOBA ~~~4

(

:

/ ......... . .'

.,,,,,CO

lKE OF THE WOODS

.' /

Figure 1.1: Physical Location of the Lake of the Woods

QUE BEC

The majority of the scholars interestèd in rock images of the Canadian Shield (for

example: Dewdney and Kidd 1962, 1967, and 1975; Fox 1974; Lanoue 1989; Lambert 1983,

1985, n.d.; Molyneaux 1975; Rajnovich 1989, 1994; Reid 1977) examined the sites in the Lake

of the Woods from 1885 (Lawson 1885) onwards. The individu al pictograph sites associated

with this lake vary in the number and range of their images. Most of the known pictographs

occur at the base of vertical granitic rock walls, either immediately beside the water or several

metres from the water's surface. None of these sites can be dated using absolute·

chronometrie techniques. l based ml' research on archaeological experience gained from

conducting fieldwork in the Canadian Shield region since 1990.

My views on rock image studies in general

In preparation for my thesis, l carried out a comparative survey of rock image studies

around the world. l examined the effects both of different theoretical approaches, and the

interpretative themes used by each scholar. The term 'rock art' is applied world-wide to

images that are placed on the surfaces of rocks. 1t occurs in many different places and

settings: Australian rock shelters, the surfaces of boulders in the Jordanian desert, vertical

Page 2

r---..

rock faces or rock outcrops on the Canadian Shield, the sides of the stone passages of New

Grange in Ireland, and the walls of deep caves in France and Spain. 'Rock art' also covers

features created using rocks of different sizes to produce 'rock,' or 'boulder alignments.'

The term 'art' is problematic because it suggests that these images have primarily a

decorative value and no intrinsic value or meaning of their own. lt also implies classification

of these images according to Western notions of high or low art, or, perhaps, a craft. These

terms have loaded meanings, since they impose the analyst's conventional values. Rock

images should not be considered within such a perspective, since, evidently, the cultural

context of the 'reader' or 'viewer' influences perception and classification. This prejudgement

affects how images are understood (Blocker 1994; Conkey 1987; Price 1989).

Not:withstanding these continuaI dangers, scholars interested in the se images

continue to use the term 'art'. Images similar to those found on rock surfaces of the

Canadian Shield are encountered in other cultural contexts (Densmore 1974 [1928]; Phillips

1999; Ritzenthaler and Ritzenthaler 1970). The designation of all these sorts of images as

'art', 'folk art' or 'hanrucrafts' has created problems for their analysts.\ There are no clearly

delineated guidelines indicating the way in which the meaning of such images should be

unpacked. Yet, whether or not these images were intended as 'art', they are a form of

communication, and, therefore, they challenge us to understand them. l rusagree with

Whitley's (2001: 22-23) argument that the term 'rock art' should not be changed since a

western intellectual tradition has used it for over one hundred years. The use of a term for a

long period of time should not justify its continued usage, particularly if the users

acknowledge that problems exist with it. Continuing such a practice or 'tradition' merely

Ieaves the arena open for continuaI disputes and discussions over whether these images are

art or not. l therefore propose that these representations should be termed rock images, or

petroglyphs and pictographs.

Rock image sites cannot be studied using the same techniques as are applied to other

archaeologicai sites. The theoretical approaches used and the questions asked may be the

same but the data sources are rarucally different and generally far more limited. These images

cannot be excavated using the techniques for recovering, cataloguing, and analysing data that

1 Frank (2000: 1-18) in her introduction clearly surrimarises the intellectual arguments, and debates concerning the development of the different theories that have led to different objects and images being termed "decorative" or "flne" arts or handicrafts. Dondis (1973), Gombrich (1984), Lay ton (1991), Morphy (1989), and Rapaport (1997) considered the question of what can be defmed as art, crafts, and visualliteracy. Haselberger (1961) proposed a method for dealing with ethnological art.

Page 3

archaeologists apply to 'conventional' archaeological sit~s. The area surrounding such images

may be excavated but the physical context of the site often provides little or no information

about the meaning(s) of the images themselves.

The subjective beliefs and ideas held by the people who created these images did

more to shape them than technological processes or the economic or political systems in

which these people lived. Therefore, the archaeologist must rely to an unusual degree on a

range of non-archaeological sources in order to establish the meaning of the images. It is

very difficult to access this information for a group whose past is available only through the

archaeological record. The difficulties in accessing the symbolic knowledge of a group of

people through the inherent attributes and physicallocation of such images may explain why

these sites have often been ignored, or merely described, in contrast to similar images found

on birch bark scrolls. Fieldwork and archivaI work must be considered as equally important

in this study, since information must be drawn from a wide range of disciplines, induding

archaeology, anthropology, history, art history, geology, and geography.

It is time to re-examine the fundamental daims regarding rock images and their

study. Rock images have fallen into 'the domain of archaeology because they are found

outside on the surface of rocks, and not on clothes, canvases, portable objects or documents

in archives. My own fieldwork has demonstrated that, contrary to common beliefs, sites in

the Lake of the Woods region did not always exist in conjunction with a body of water. It

also shows that rock images occasionally are found in caves rather than on cliff faces.

Suppositions about the known physicallocations of the pictographs have had a fundamental

influence upon their discovery, and how and by whom they are interpreted (see Brown and

Brightman 1988: 55-56; Dewdney and Kidd 1962: 13-14; Kohl 1985 (1860]: 415; Wheeler

1977a, 1977b, 1979; Rajnovich 1989: 185-186).

The influence of divergent disciplinary perspectives and paradigms of thought upon

the analysis of images in general must be understood. The contrasting ways in which images

are regarded has affected their interpretation.

The use of informants

The question exists whether or not to rely upon informants for information

regarding both the name and meaning of a shape. l decided not to rely on informants for

ethnographic information for several reasons. Firstly, several archaeologists in Northwestern

Page 4

Ontario informed me that archaeologists do not have a good relationship with the local

indigenous peoples. lt would be very diffic~lt, if not impossible, to consider undertaking

ethnographie fieldwork in conjunction with archaeological fieldwork. In addition, 1

discovered that information regarding the meaning of these images is a politically sensitive

topie. Indeed, 1 established a connection between sorne of the new images at DgKl-2 and a

court case regarding charges of incest at one of the reserves in the Lake of the Woods.

However, W. A. Ross, who was the regional archaeologist in 2001, asked me not to provide

any more details regarding the images and why they were created (Ross, Pers. Comm. July

and August 2001). My third reason is that indigenous peoples currently living on the local

reserves are currently using the rock image sites, in severalland daims, as evidence that they

have always lived in the Lake of the \Voods region. Indeed, DgKm-1, in Splitrock Bay near

Blueberry Island in Sabaskong Bay, was to be used in 2001 as a means of claiming that the

Ojibwa had always lived in this region. Finally, anthropological evidence suggests that the

local inhabitants may not always have" spoken Ojibwa. The first section of Chapter II

addresses the problem of identifying who spoke Cree and Ojibwa in the large region which

includes the Lake of the Woods.

What is an interpretive approach?

Archaeologists, unlike scholars from other disciplines, often cannot utilise tex tuaI

data. Instead, they must interpret how different groups of people thought, behaved, and

interacted by drawing upon the remaining material evidence of these groups. Five different

interpretive approaches exist in archaeology: culture-historical, contextual, intuitive,

analogical, and homological. Archaeologists use these approaches to infer human behaviour

and meaning from archaeological data. Each approach dictates the types of questions used,

and determines the levels of understanding obtained about whatever archaeological evidence

is being considered. Analysts must examine the same data using these different approaches

sequentially, since this increases the potential quantity of information gained concerning

these images. The different theoretical approaches are complementary rather than exclusive,

and should be employed in a definite order (fable 1).

Page 5

CULTURE-HISTORY

Levell: Establishes the shape, location, and date of the images.

j CONTEXTUAL

Level 2: Relates various types of empirical information to the site where they were

found, and considers a broad set of associations and relations among the images themselves,

and between the images and their physical settings. A contextual approach involves a search

for patterns relating different images and combinations of images within sites.

l INTUITIVE and/or ANALOGICAL and/or HOMOLOGICAL

Level 3: Assigns meanings to images on the basis of three methodologies which are

employed together or separately.

Table 1.1 Sequence in which approaches must be utilised

The culture-history approach (Levell):

Archaeologists ought to apply the culture-history approach prior to interpreting these

images because it permits them to establish the form, date, and location of the images. This

approach provides archaeologists with the basic data that they need to perform higher-Ievel

operations. Archaeologists should concentrate on assembling detailed information about

each individual site and consider each image as a separate entity at this level.

The contextual approach (Level 2):

Those wishing to connect the culture-historical information with the totality of the

images and the natural features of the site itself subsequently take the contextual approach. l

use the word contextual in a broader and less technical sense than Hodder (1991) intended.

At this level, the scholar can propose low-Ievel generalisations about relating their data,

without searching for explanations of what these data mean. The contextual approach

Page 6

permits archaeologists ta consider a broader set of associations and relationships among the

images themselves, the images and rock surfaces, and the generallocation of the site. Prior to

proceeding to Level 3, it is imperative that scholars search for patterns in their data.

Archaeologists must use this approach before attempting to assign meaning to images. Sin ce

relatively little work has been devoted to studying problems of this sort, it is evident that

archaeologists are more interested in assigning meanings or explanations to the images

themselves, or to groups of images, than in systematically analysing the images and their

formaI relationships. Archaeologists seeking to assign meanings to different images must

employ the analogical, the homological, or the intuitive (narrative or constructivist) approach,

or a combination of these.

The intuitive approach (Leve! 3):

The intuitive (narrative, constructivist, or so called 'humanist') approach associated

with post-processual archaeology developed as a reaction against the positivism of the

processual archaeologists during the 1970s and 1980s. Hodder (1985) had problems with

many of the key te nets of positivism and processualism. He, as weIl as Shanks, Tilley, Ucko,

and Thomas, were concerned by the lack of interest with cognitive factors. Yet, it was

Hodder's students, Shanks and Tilley, who really developed this approach into a force with

which to be reckoned. For the extreme relativist, there is no possibility that one person's

interpretation is better than another's. These scholars maintain that it is impossible to

demonstrate empirically that one interpretation is right while another interpretation is wrong.

The scholar frequently reaches conclusions based upon thought processes or information

known only to him. They argue that, because gender, class, cultural background, and racial

prejudices so strongly influence the interpretation of archaeological data, it is difficult to

accept such interpretations. Strong relativists and archaeologists, such as Thomas, Shanks,

and Tilley, also reason that it is only possible to understand segments of a culture in relation

to its entirety. Each culture must be understood hermeneutically because it constitutes a

unique system of independently defined entities. Meaning, they reason, must be contextual,

especially in historical hermeneutics, if the archaeologist is determined to understand another

world. Hence the development of the fourfold hermeneutic circle: archaeologists must work

through the hermeneutic of their own discipline first, then the hermeneutic where they must

operate and live as human beings, thirdly the hermeneutic of understanding a dramatically

Page 7

different culture from their own, and finally the hermeneutic of moving between the past

and the present (Shanks and Tilley 1994: 107-108).

Shanks and Tilley (1987: 76) asserted that universals did not exist because culture and

rustory both represent contingency. Cross-cultural studies could be undertaken only by

"translating" precisely how members of the culture being studied understood that culture.

Bence the difficulty in undertaking any type of cross-cultural analysis. Cross-cultural

comparisons are rughly problematic because things only have meanings within their own

cultures. Strong relativists believe that cultures are incomprehensible to anyone outside of

that culture.

So, what can we conclude?

Weak relativists, such as Wylie, stress that archaeological interpretation cannot be

free from bias. Weak relativists contend that beliefs held by individu al archaeologists

influence how they choose and excavate a site, and how they identify, classify and interpret

artefacts. They argue that what passes as an archaeologist's interpretation is influenced by her

beliefs and, therefore, it is difficult to be certain about that interpretation's validity. This is

because the archaeologist, with her bias, unconsciously has the power to dis tort

interpretations.

The intuitive or narrative approach is popular among petroglyph and pictograph

scholars because it enables them to draw conclusions, even when there is a paucity of

detailed textual records pertaining to the images' meaning. The question remains whether

trus position is usefuL The strong relativists load their theoretical discussions with tortuous

vocabulary. Their texts invariably claim a great deal but, in reality, fail to advance beyond the

subjective. The tone of these interpretations is always authoritative and assertive but they

remain exercises in "navel gazing" or introspection. Readers, who cannot verify how

interpreters reached their conclusions, find themselves at the whim of each scholar's

intuition. The applications of trus approach cannot be duplicated, since researchers rarely

explain how their conclusions were achieved. If an archaeologist uses the intuitive approach,

it is not feasible for another archaeologist to apply any of his interpretations in another

context. Therefore, a scholar interested in establishing the meaning of rock images in a more

rigorous and persuasive fasruon should adopt the analogical or homological approach (see

Table 1).

Page 8

The analogical approach (Leve1 3):

The analogical approach, associated with processual archaeology, was a reaction by

positivists against the culture-history approach during the 1960s. Proponents argued that

behaviour was predictable from material culture because many uniform connections existed

between the various components of socio-cultural systems, material culture, and hum an

behaviour. As Lyman and O'Brien (2001: 303) observed, the use of general comparative

analogies reflected the theory of orthogenesis. Yet general comparative analogies were not

new in the 1960s, sin ce archaeologists had used them, as weil as homologies (specific

historical analogies), to examine the archaeological record of North America during the late

nineteenth and the first part of the twentieth century (ibid.: 304).

Biologists define analogies as similar features of different specles without close

evolutionary relations. The similarities have resulted from natural selection operating to adapt

different species to a similar environment (Abercrombie, Hickman and Johnson 1985: 20).

An archaeological analogy is a likeness or partiallikeness assumed to exist as a consequence

of convergent development under comparable conditions. Interpretation using analogies

allows scholars to use strong cross-cultural regularities between behaviour and material

culture in systemic contexts to attribute behavioural correlates to material remains recovered

from archaeological contexts.

Binford (1977: 288) maintained that ethnographic analogies are useful sources of

middle-range generalisations. An analogical interpretation is based upon the existence of

strong functional correlations between specific aspects of behaviour and specifie aspects of

material culture. Given such a correlation, the presence of similar material culture in the

archaeological record allows an archaeologist to presume that similar or associated

behaviours existed in the pasto Watson, LeBlanc and Redman (1971: 51) argued that

analogies provided archaeologists with the "richest sources of hypothesis" available.

What are the types of bridging arguments used for justifying an analogical

interpretation?

Scholars who privilege an analogie al approach conclude that it is only worth

employing regularities in human behaviour that are universal. They presume that correlations

exist between past and present day cognitive and behavioural capabilities of human beings.

Proponents assume that, if similar behavioural characteristics can be established between

specifie aspects of material culture and behaviour in the modern (present day) world,

Page 9

scholars can extend them to cover the same, or similar aspects of material culture in the

archaeological record (Binford 1981). Scholars adopting analogical approaches use universal

generalisations, rather than concepts specifie to individu al or historically related cultures. It is

impossible to include the idiosyncratic facets of an historical tradition if analysts consider

only universal correlations relevant.

The homological approach (Leve13):

An archaeological homology is a similarity in two or more cultures occurring as the

result of shared historical origin unobscured by adaptation to different cultural

environments. Archaeological homologies result from diffusion as well as common descent,

they differ from biologie al ones. In biology, reproductive isolation ensures that homologies

occur only among species descended from a common ancestor. Consequently, homological

similarities imply evolutionary relationships among organisms (Abercrombie, Hickman, and

Johnson 1985: 145).

Scholars establish homologies by tracing cultural continuities through time in a single,

or a series of historically related cultural traditions. Interpretations employing homologies

frequently utilise the direct historical approach. This approach identifies parallels between

culturally specifie beliefs and their material expressions during the eariy historical period, and

employs material culture to trace these beliefs back into prehistoric times. The bridging

arguments for establishing homologies between the present and the past are culturally

specifie. According to Watson, LeBlanc, and Redman (1971: 50), homologies are

advantageous in regions of strong cultural continuity and where the same techniques and

implements have been utilised for a long period of time. Written records, oral heritage, and

ethnographie observations are probably the strongest types of evidence used when devising

bridging arguments (Trigger 1995: 452). Scholars exploit these materials to establish which

beliefs existed in specifie cultures.

The direct historical approach connects archaeology with other disciplines, providing

information concerning the practices and beliefs of specifie, or groups of, historically related

societies. By relying on a wide variety of evidence, archaeologists can attempt to establish

whether continuity exists in the practices and beliefs in question. The homological approach

allows in-depth analysis of a specifie group and its material culture. It permits archaeologists

to consider what human beings might have thought about in the past, and the meaning of

Page 10

specific symbols and objects. It may be possible to determine that similar images used in a

variety of different mediums held similar meanings. Scholars intending to use homological

interpretations must have a detailed understanding of the skills required to use

archaeological, historical, and ethnographic materials from the region in question. These

materials are multidisciplinary in nature. Those taking this approach must verify how cultures

establish entrenched beliefs, and how susceptible these are to evaluation in terms of that

culture. Any scholar intending to use the homological approach must be aware of different

cultural perspectives. He must acknowledge that using the direct historical approach is

neither straightforward nor simple. Continuity of form does not necessarily imply continuity

of meaning. Meaning can also change, especially since kons tend to be polysemie. Material

symbols can ob tain new meanings and become examples of what Davis (1992: 25) called an

"iconographic disjunction." Continuity or discontinuity in beliefs over time is not correlated

simply with material culture.

In sum, culture history and the contextual approach seek to answer questions such as

what mate rial there is to study, when images were created, and who created them. The

analogical, homological, and intuitive approaches seek to ascertain for what purposes they

were created, and what their meanings were to their creator(s). Inadequacies in each of the

first two stages limit the effectiveness of the stage of analysis. Therefore, an archaeologist

studying a rock image site must first implement the culture-historical approach, then the

contextual approach, and finally the intuitive, analogical or homological approach.

The debate between the "lumpers" and the "splitters"

As discussed earlier in this chapter the philosophical approach dicta tes the questions

asked of and applied to the data. But the taxonomy applied by the archaeologist determines

the degree of resolution or the level of aggregation used to collect, describe, and classify the

data, which in turn affects the conclusions. In archaeology, two methods of classifying

information and creating taxonomies are prevalent: those of "lump ers" and "splitters". It is

important to understand the differences between splitters and lumpers because it affects how

archaeological evidence and information is drawn upon and used. It affects how the origins

of early man are considered. Indeed, Clark (1999: 2029-2032) argued that archaeological

research concerned with the examination of modern human origins in Europe was

epistemologically naïve. It could he maintained, be summarised, in terms of two paradigms:

Page 11

/~"

the replacement and the continuity paradigm. What is important about this debate for my

argument is that both paradigms define and consider variables that are used by both models

in distinctly different manners. Scholars utilise different taxonomies and phylogenies

according to the paradigm that they invoke. Therefore advocates of the replacement

paradigm use 'splitter' taxonomies and dendritic phylogenies and argue that a distinction

exists between archaic Homo sapiens and neandertals (ibid.). Advocates of the continuity

paradigm employ "lumper" taxonomies and reticulate phylogenies. These researchers also do

not distinguish between archaic Homo sapiens and neandertals because they maintain that

the neandertals were one of several geographical clades that existed amongst archaic Homo

sapiens (ibid.). This split between lumpers and splitters is not unique to this discussion sin ce

it occurs exists in other discussions in archaeology as weIl as in other disciplines (see Hall and

Chase-Dunn 1993; Lieberman, Wood, and Philbeam 1996; Nurse 1997).

Essentially people termed "lumpers" look for similarities and group whatever is being

classified according to whether they share a few major characteristics, and thus work with a

limited number of categories. People called "splitters" classify things into different groups

based on minor characteristics and thus use a finer categorization. l had during the initial

phases of my research, as discussed in chapter IV, used the VIPs software to determine the

physical nature and the boundaries of each image. It became clear that the images at aIl of the

pictograph sites were not solid and physically different from each other once closely

examined. This argument that the images are different is easier to understand if the example

of the 'crescent' is used. This shape occurred twelve times at seven sites: DgKl-l, DgI<J-2,

DhKm-S, DiKn-l, DjKn-l, DiKm-3, and DiKm-S1. One could argue that they should aIl be

loosely categorised as crescents because if they are described according to each of their

physical attributes, a different technique of grouping them emerges. Each crescent below, in

figure 1.2, is slightly different.

Page 12

A vertical crescent at DgKl-2

~ ..

Two horizontal crescents at DgKl-l

A horizontal crescent at DiKm-51

Page 13

A horizontal crescent and a diagonal crescent at DhKm-5

Two horizontal cres cents at DjKn-l.

Two horizontal crescents at DiKm-3

Page 14

Figure 1.2 Examples of shapes loosely categorised as 'crescents' and the sites that they come from in the Lake of the Woods.

It should be evident from these photographs above (figure 1.2) that the red images have

suffered from the passage of time and numerous climatic changes. Sorne of them are quite

pale and difficult to see. Furthermore, none of the images is probably exactly the same image

that it was when originaIly created. Sin ce aIl of the se images were clearly different from each

other in sorne respect, the "splitter" approach was utilised when the data was described

rather than the "lumper" approach. A nested classification system was attempted and it is for

this reason that each type of shape was examined in loose groups. Therefore each shape was

examined as a shape and loosely categorised as type within a specific group so as to try to

find general patterns. Shapes could be loosely categorised as a type within that larger group.

Therefore loosely categorised shapes could be compared with distributions of other loosely

categorised shapes. As discussed in Chapter IV in considerable detail it is important that the

archaeologist should describe, not interpret, the evidence when recording a rock image site.

This operation involves thinking very carefuIly about how one describes images, since any

description will affect subsequent analysis. This should explain why groups are referred to in

Chapters V, VI, and VII as shapes loosely categorised as types of 'crescents' for example. But

it was discovered when the data was analysed and discussed in Chapter VI that using a nested

Page 15

classification system was not possible because of the variability of the data. This discovery is

supported by the results presented in the dendrograms in Appendix 15.

The structure of this thesis:

This thesis has seven chapters, beginning with this introduction.

Chapter II discusses the regional setting of the pictograph sites examined. This

chapter will consider the ethnicity of the people who lived in this region, establish which

ethnic group was responsible for the creation of these images, and determine which bodies

of ethnographic evidence can be drawn upon when the question of these images' meaning

arises. This chapter ends with a discussion of the range of images used by the Aigonquian­

speaking peoples and the Assiniboine to establish whether it is possible to examine these

images as a means of communication.

Chapter III presents and discusses earlier research on the pictograph and petroglyph

sites in the Lake of the Woods area.

Chapter IV explains the basis for my decision-making while conducting fieldwork,

and how l identified and classified the images found. My goal was to systematically collect

detailed information about rock image sites over a portion of the Canadian Shield in order to

better understand the nature of the available evidence.

Subsequent chapters deal with the archaeological data according to each of the

approaches that must be applied to this data as outlined in Chapter 1. Therefore Chapter V

applies the culture-historical approach to this data, and seeks to establish the total range of

the images. Chapter VI presents the conclusions that are drawn after applying the contextual

approach, and considers whether combinatory rules exist regarding the relation of the

different sorts of images to one another. Chapter VII uses mainly the homological approach

to consider how to establish the meaning of these images, and whether they are related to

other images outside the Lake of the Woods rock image sites. Chapter VIII summarises the

findings, discusses whether and if the goals established at the beginning of the thesis were

achieved, considers the theoretical and methodological implications of this research, and

posits some thoughts for future research.

Page 16

Chapter II: Lake of the Woods: Regional Setting of the Sites

This chapter discusses the regional setting of the pictograph sites that this thesis

examines. It considers the ethnicity of the people who lived in this region, hence, which

ethnic group created these images. It aims to determine which body of ethnographic

evidence can be drawn on when the question of the meaning of these images arises.

This chapter has three sections grouped according to different factors. The first

section examines how the physical geography, geology, and climate influenced the

production and survival of rock image sites as well as the history, ethnographic record, and

the physical anthropology of the peoples who lived in the Lake of the Woods area located

between Lakes Winnipeg and Superior. The second part considers the debates about the

history and the ethnography of the indigenous people who lived there. Problems include: (a)

the inability to create a clear chronological framework for the cultural and ethnic identities of

former inhabitants; (b) understanding the archaeological record of the Boreal Forest; (c) the

unresolved debate between historians and archaeologists concerning the ethnic identification

of those peoples who were in early contact with the Europeans; and finally (d) whether

biological evidence can help resolve the problem of the precise ethnic identity of these

peoples. The third section of this chapter has two goals. The first is to establish the ethnicity

of all the different images and the objects upon which they exist, so that it can be determined

which images the Algonquian-speaking peoples created, and which the Assiniboine made.

The second goal is to examine these images as a means of communication, since similar

images were used throughout this geographical region.

Section 1 - Geohistory and geography of the region.

(a) Geology

Climate, soils, physical geography, and hydrology have influenced the settlement

patterns, movements, and the size of populations inhabiting the Boreal Forest and the Great

Lakes-St. Lawrence Forest. The Lake of the Woods is part of a watershed that flows from

the Boundary Waters into the Winnipeg River and northwards to the Arctic Ocean. It is a

remnant of glacial Lake Agassiz, and has two distinct parts. The bedrock geology and glacial

action have caused the differences that distinguish the northern and southern portions of

Page 17

Lake of the Woods. Therefore, the history of the geology, the soils, climate, and hydrology

must be discussed.

This region is also part of the Canadian, or Precambrian Shield, one of the major

physiographic units of the North American continent (Gardner 1981: 6). The effects of

glaciation on the Canadian Shield are evident. Both thin soils and bedrock outcrops are

abundant as a consequence of much of the area being scoured by the massive ice sheets. The

sediment that does exist is glacial-Iacustrine connected to the glacial retreat (Meyer and

Hamilton 1994: 96-97). The Wisconsin maximum occurred between 19,000 and 16,000 years

BP and deglaciation occurred between 14,000 years and 12,000 BP (Bird 1973: 165 and Hare

1976: 509). The beginning of the Holocene period, about 11,000 years ago, resulted in a

sudden withdrawal of the ice margin. The ice margin coincided with the southern edge of the

Canadian Shield about 10,500 BP and vast glacial lakes such as Lake Agassiz developed

between 11,600 and 8,400 BP (Hamilton 1996: 12). Lake Agassiz covered parts of

Saskatchewan, Manitoba, Ontario, Northern Dakota, and Minnesota (Winterhalder 1983:

10). Bryson and his colleagues (1969: 8) argue that, immediately following the Cockburn

glacial phase around 8500 BP, the ice sheet rapidly disintegrated in the Hudson Bay region. A

saltwater incursion formed the Tyrrell Sea, later called Hudson Bay, which covered the area.

By 7,000 BP Lake Agassiz drained northward through the Nelson channel, into the Tyrrell

Sea (Elson 1965) . Post-glacial warming reached its maximum about 7,000 to 5,000 years ago

in the Hypsithermal Interval. Tt was warmer in this period than at present but initially it was

drier but later on it became wetter (Terasmae 1970). Wright (1976: 592) argues that this

climate period was subsequently followed by the Neoglacial, which was a period of cooling.

A climatic optimum, partially corresponding to the Neo-Atlantic episode, was

reached around AD 1090-1230, when the northern hemisphere experienced a period of

warmer weather (Bryson and Wendland 1967: 294). The climate subsequently became colder

but during the late 1400s the temperatures improved. The "Little Ice Age" occurred between

1550 and 1850 du ring which the region experienced considerably colder temperatures. Hare

and Thomas (1979) stated that between 1885 and the 1940s, the Northern Hemisphere

experienced a warming trend, so that by 1970 this region had achieved half of the post-1885

temperature gain. However, larger ranges of variability occurred in the northern regions of

Canada.

Page 18

Much of the crystalline bedrock in this region is exposed or close to the surface

(Gardner 1981). MineraIs such as silver, gold, platinum; associated sulphides of iron, copper,

zinc, and lead have been found in conjunction with the sedimentary and igneous rocks

(Minning, Cowan, Sharpe and Warman 1994: 12). Hematite, the principal constituent of iron

ore, is readily available as a primary constituent as well as a secondary product in the

sedimentary and igneous rocks of this region (Bates and Jackson 1984: 233).

Glaciers covered the Subarctic region for most of the past 20,000 years; consequently

the soil is thin and not very extensive. Chemical and biochemical breakdown and mechanical

weathering is slow in the Subarctic. Thin podzol soils coyer the Shield area leaving rock

outcrops and peat deposits. Consequently these processes affect neither the bedrock nor any

unconsolidated material to create a soil profile (Gardner 1981: 13). The soil profile typically

consists of a thick leaf mulch, a leached (ash colour) horizon, a second horizon with deposits

of organic colloids leached from the first horizon and the clays of prehistoric Lake Agassiz.

On the southern margins of the Shield, where the Boreal Forest vegetation meets that of

Great Lakes-St. Lawrence Forest, the soils become "brown-wooded and grey-wooded"

(ibid.). Both of these soils, in comparison to the podzols farther north, can be put to

agricultural use with careful management and proper drainage.

Two vastly different major forests co-exist in this region: the Boreal Forest, and the

Great Lakes-St. Lawrence forest. The Great Lakes-St. Lawrence forest occupies a large part

of northeastern North America, reaching its western extremity in southeastern Manitoba. It

covers the northern half of Minnesota and in northwestern Ontario it is found in a narrow

east-west strip extending about eighty-five kilometres north of the international border

(Meyer and Hamilton 1994: 97; Gardner 1981: 12). The softwoods found in this region

include eastern white and red pines, eastern hemlock, and yellow birch, while hardwoods

include red oak, sugar and red maples, basswood, and white elm (Rowe 1972: 11). This pine­

hardwood forest provides a variety of potential food sources including nuts, berries, and sap.

Wild rice was a major foodstuff for the peoples of the Late Prehistoric period in the Great

Lakes-St. Lawrence fore st zone. The wild rice stands, the great variety of fish (especially

sturgeon), and other aquatic animaIs were vital for survival. The sturgeon contains isinglass, a

natural binder that could have been used in the manufacture of the paint used for

pictographs. Isinglass, a transparent gelatine prepared from the air bladder of the sturgeon,

can be used as an adhesive and clarifying agent. The fauna of this forest included white-tailed

Page 19

deer, moose, woodland caribou, and waterfowl, the latter frequently found amongst the wild

rice stands.

Rowe (1972: 6) identifies the Boreal Forest as a conifer rich zone distinct from the

hardwood-rich Great Lakes-St Lawrence forest to the south and the tundra to the north. He

noted that deciduous forests decreased northwards, becoming more open in the forest­

tundra transition. The Boreal Forest is in a state of "quasi-equilibrium", since it is connected

to the summer and the winter positions of the Arctic front (Bryson 1966: 257). Long-term

variations in climatic patterning, essentially shifts in the win ter or summer positioning of the

Arctic front, cause the expansion or contraction of the Boreal Forest region (Bryson 1966).

Furthermore, the floral and faunal resources of the Boreal Forest on the Pre-Cambrian

Shield are not nearly as rich as those of the Great Lakes-St. Lawrence Forest. Winterhalder

argued that this fore st was in reality a "mosaic of vegetation habitats or patch-types" (1983:

28). This region is covered by dense spruce forest, open bogs, lichen woodlands, tundra­

covered ridge tops, alpine tundra, and wet forested depressions or river valleys (Gardner

1981: 12). According to Gardner (ibid.) and Winterhalder (1983: 32), the vegetation, under

the influence of moisture conditions, climate, soil conditions, and topography, defined a

habitat. Winterhalder (ibid.) argued that the vegetation also had an "internaI rhythm of

disturbance and succession" abetted by snow, wind, and more importantly cyclical fires that

provided the forest with a short-term historical time frame.

Fires have beneficial effects on the soils of coniferous forests since the natural rate of

decomposition is slow and the build up of litter hinders rapid plant growth (Kelsall, Telfer

and Wright 1977: 15). Forest fires enable mineraIs and other nutrients to be released for use

by other plants (ibid.). Fire increases the variety and range of plants and animaIs in an are a

making it an important factor in the creation of this mosaic community. Areas that do not

experience forest fires inevitably undergo a sharp decrease regarding both the number and

variety of their flora and fauna (ibid.: 24). Forest fires also affected the human populations of

this area. Wright (1981: 86) postulates that forest fires invariably forced inhabitants to shift

their hunting regions. This diversity of different ecological communities enables varying

levels of usable biomasses to coexist. Closed coniferous forests are examples of ecological

communities that provide limited or unusable food resources. Rich ecological communities

include aquatic areas such as rivers and lakes with wild rice, wild fowl and fish; and recently

burnt out areas undergoing colonisation by new plants and animaIs.

Page 20

Today the region between Lakes Winnipeg and Superior has up to 120 frost-free days

and its climate is considerably less harsh than the area to its north (Anfinson 1990: 150;

Meyer and Hamilton 1994: 99). Gardner (1981: 10) argues that the maximum period of frost­

free days in this northern region number between forty and sixty. Snow is an important

ecological factor in the Boreal and Great Lakes-St. Lawrence Forests, but it is far more

significant for the inhabitants of the Boreal Forest, since accumulations are heavier and snow

cover can last for at least six months. Additional significant ecological factors for this region

include the density, hardness, structure, and cover-temperatures of the snow (ibid.). The

topography, distribution and structure of vegetation, air temperature, wind speed, and wind

direction affects both the depth and the duration of the snow cover. Tt is possible that the

different ecological communities of the Boreal Forest invariably influence the movement and

activity of the snow. Gardner (1981: 10) argues that snow was an "effective mediator" for the

water and energy balances in the Subarctic.

Lakes of different sizes, bogs, and large numbers of streams and rivers covering the

Shield region drain either into the Arctic or Adantic Oceans. The Lake of the Woods, and

Red Lake, Thief Lake, and Mud Lake in Minnesota are remnants of glacial Lake Agassiz

(Anfinson 1990: 150). Rivers and lakes played a vital role in the seasonal transportation and

communication of regional inhabitants. AIthough they provided the only feasible avenues of

transport through most of the year, the freezing and breaking-up of the ice on these rivers

and lakes impeded transportation. Birch bark canoes were used in the summer months; in

winter people used snowshoes on the snow-topped ice. Food resources were closely tied to

these bodies of water, sin ce fish was a major food resource, as was caribou. Wright claimed

that humans could not have survived solely on the migra tory fowl or other seasonal

resources; therefore fish and caribou were fundamental parts of their diet (1981: 90).

Drainage patterns and large bodies of water served as focal points for setdements,

subsistence, and commercial activities.

The size of prehistoric populations is difficult to assess but was probably connected

to climate, vegetation, and the ability to travel through the region. Formulas exist based upon

ethnographic data (see Brose 1970; Leacock 1973; Rogers 1972) but they must be used with

caution, since archaeologists must extrapolate from ethnographic data without a detailed

knowledge of the conditions of the pasto

Page 21

The utilisation of aquatic foodstuffs and the fauna is difficult to gauge in this region,

since very little organic evidence survives in the highly acidic soils. Wright notes that bone is

rarely found (Wright 1981: 90), although the neutralising effects of the ash from the hearths

found at the Heron Bay site, a Laurel site on the north shore of Lake Superior, enabled bone

tools and refuse to be preserved (ibid.). It is likely that both wild rice and the resources

provided by aquatic environments were vital to the subsistence economies of Late

Prehistoric populations (Meyer and Hamilton 1994: 98).

Although Clelland's (1982) discussion focused on the fisheries of the Upper Great

Lakes region, he made some interesting observations that could be applied to the area west

of Lake Superior. He maintained that archaeologists and anthropologists have not properly

appreciated the value of the inshore fishery (ibid.: 761) and that the archaeological record

suggests an increased understanding of aquatic resource availability from the Late Archaic

period to the Late Woodland (ibid.: 777). He also postulated that, while the Late Woodlands

people resided in areas that hindered the domestication of food products because of the

short growing season, these regions had a high degree of residential stability and

concentration of population (ibid.: 756). Increased reliance on the fish may have aided this

population growth. Clelland posits that the Late Woodland peoples observed the

concentration of fish-spawning in streams, stream mouths, and coastal shallows before they

noticed that in the late autumn fish concentrated in shoals offshore (ibid.). Arthurs (1984,

1986) contended that the Laurel and Blackduck inhabitants of the Long Sault site, or

Manitou Mounds, along the Rainy River, probably relied heavily on the seasonal fish spawns.

Anfinson contended that, during the early historie period, one third of the northern

region of northern Minnesota was wet prairie, with patches of aspen and oak woodlands

(ibid.). Wild rice grew in this region and in the border lakes of northern Minnesota but not in

the Lake Superior shore region (ibid.: 150-151). Archaeologists Johnson (1969a, 1969b) and

Gibbon and Caine (1980), reasoned that the abundance of wild rice provided a stable, rich,

and perennial - if annually variable - resource base, enabling permanent large village sites.

Johnson (1969a) postulated that perhaps the populations increased between AD 1000 and

AD 1400. Johnson (1985) argued that the Cooper, Vineland and Wilford sites are examples

of such Late W oodland settlements. The type of resource utilisation practised there

permitted a large population size compared to the subsistence and settlement patterns of the

Page 22

Siouan groups, who lived in this reglOn at the time of contact and du ring most of the

"colonial period" (Whelan 1990: 82).

Gibbon and Caine (1980: 57-72) claimed that intensive harvesting of wild rice caused

a dramatic increase in population between the Middle and the Late Woodland cultures in

eastern :Minnesota. La Vérendrye wrote about the abundant wild rice in the Lake of the

Woods in a letter of 1732 sent to the Marquis de Beauharnois, Governor General of France

(Burpee 1927: 97). Rajnovich (1984: 198) asserted that this document is the first written

record of abundant wild rice in the Lake of the Woods; unfortunately, the rice crop of the

following year failed.

Using palynological evidence from Northern Ontario and Northern Minnesota,

Rajnovich determined that in the Lake of the Woods region wild rice was harvested as early

as the start of the Laurel cultural tradition (ibid.: 204). She based this conclusion on

McAndrews' (1969) analysis of the sudden increase in Gramineae pollen approximately 2450

±100 years ago. The majority of this grass pollen was identified as wild rice pollen and she

concluded that wild rice was present at the beginning of the cooler, wetter climatic stage

about 2450 years ago, or 1935 years if the chronology obtained by the sediments from Rice

Lake is used (ibid.: 1678). Rajnovich (1980a) observed that this period corresponded to the

beginning of the Laurel cultural tradition in Minnesota and the Lake of the Woods region.

She (1984: 204) asserted that the lack of archaeological evidence from local sites was

probably a consequence of recovery and identification techniques rather than evidence of "a

real lack of utilisation" (ibid.). She also concluded from examinations of the setclement

patterns of the Middle and Late Woodland periods that a tendency existed for wild rice areas

to cluster around the archaeological sites. However, she warned that this relationship must

only be viewed as tentative until confirmed by the retrieval, through the use of soil flotation

techniques, of wild rice grains in a major archaeological excavation (ibid.).

(b) Changing lake levels

The question of changing lake levels must be considered because archaeologists in

northern Ontario have traditionally found rock image sites by examining large expanses of

cliff faces along the shores of lakes and rivers. The consequences of changes in water levels

could be important for the existence and discovery of pictograph sites, since some could be

Page 23

completely or partially submerged, or exist along former waterways that have dried up after

waterways have changed.

Despite the association of pictographs with water, archaeologists have tended to

neglect the possibility that water levels fluctuate for any river or lake. They must pay

attention to the past use of the area in question and consider whether the water levels and

the shapes of the bodies of water have changed.

Lake of the Woods water levels have fluctuated over time and these changes can

affect the discovery, visibility, condition, and recording of rock image sites. Dramatic

changes in the Lake of the Woods water levels are indicated on the rock surfaces. Horizontal

bands of different shades of grey visible in the figure 2.1 below indicate the different water

levels that the lake has experienced.

Figure 2.1: etiffs in the southem end ofWhitefish Bay with c1early defined high water levels.

The photograph above (figure 2.1) was taken in late summer 2003, when the reglOn

experienced a major drought and previous high water levels were clearly visible. The topmost

and the strongest of the thin white bands visible on the granite cliffs at Whitefish Bay and

throughout the Lake of the Woods indicates the most recent highest water level. The bottom

edge of the lichen above that is the highest marker. White and Meyer (1916: 166), the chief

engineers of the survey that took place between 1913 and 1916, maintained that no exact

record existed of the length of time required for "standing water to remove lichen growth

Page 24

sufficiently to create a water mark." They argued that, based upon observations made while

surveying, a high water mark could be created within a single season. Once a mark was

created, it required a number of years to fade, since the slow lichen growth rate required

several years of lower water for it ta grow downwards "to rem ove the sharpness of the edge

of the mark" (ibid.).

The "natural outlets" of the Lake of the Woods were unchanged, according to the

local informants, until 1879 (International Joint Commission 1917: 16), when a headrace was

constructed as an overflow channel for power purposes in a depression in the extreme

western end of Portage Bay, Keewatin. In 1885, the power intake was deepened and used by

a sawing and planing mill until 1906, when the plant closed down (ibid.). In 1887-88 the

water level was changed again by the construction of the Rollerway Dam, on Tunnel Island,

at the western outlet of the Lake of the Woods (International Joint Commission 1917: 16).

Portage Bay and the western outlet were about one mile apart. This dam raised the water

level by an average of o. 45m (1.5 feet) and above the previous high water levels by O. 30 m

(1 foot) between 1893 and 1898, for which continuous water level records are available (ibid.:

17).

Before 1896 the area had suffered from low rainfall levels but it was difficult to

establish the level of the Lake of the Woods until 1899 when a gauge was set up at Warroad

(ibid.: 22 & Meyer and White 1916: 104). Between 1894 and 1895, the Norman Dam was

built in the Winnipeg River about a mile below the Rollerway (White and Meyer 1916: 21).

The effect of the Norman Dam on lake level was more marked than the Rollerway since it

raised the mean by 1.07 m (or 3.5 feet) (International Joint Commission 1917: 18). The

authors of the International Joint Commission Report noted that the increase of the levels of

the Lake before the construction of the dams was between 22.86 cm (0.9 feet) in 1899 to

1.83 m (6.0 feet) in 1913 (ibid.). White and Meyer (1916: 106) also observed, while surveying

the Lake of the Woods region, that changes occurred in the lake level by a foot or more as a

consequence of shifts in direction of the wind. Both argued that it was important to keep this

type of change in mind given the proposed regulation of the lake.

However, this region had been the subject of several previous treaties concerning

question of the changing water levels, who controlled them and who could use these bodies

of water. The most notable was the Ashburton-Webster Treaty signed in 1842 between the

governments of Great Britain and the United States which created a "workable and lasting

Page 25

boundary" between Canada and the United States (Carroll 2001: 286). Article II of this

Treaty enabled the citizens and subjects of the British Empire (Canada) and the United States

to use the region from Grand Portage, on the shore of Lake Superior, to Pigeon River

(Gibbons 1929: 8).

The International Joint Commission of the United States and Canada controls and

monitors the water level of the Lake of the Woods today. This commission was established

by Article VII signed on 11 th January 1909, of the Treaty of Ashburton-Webster, at the

request of the Dominion of Canada and the United States of America (Gibbons 1929: 8;

Keeper 1912: 3). Concern regarding the use of water resources caused the US. to pass an

Act in 1902 to report upon the condition of the water of the Lakes as weIl as the water

between the St. Lawrence River and the Atlantic Ocean. Prior to the signing of the 1909

treaty, no agreement existed regarding the use or the diversion of boundary waters or streams

crossing this international boundary (ibid.). Today, the water flows into the Lake of the

Woods from both sides of the border and its waters are used for a wide range of purposes

including power, sanitation, irrigation, navigation, and transportation. The International Joint

Commission of the United States of America and Canada uses dams to maintain the level of

the water year-round so that these activities can continue undisturbed.

The consequences of this agreement are readily evident to archaeologists. The

building of the dams, from 1888 onwards, meant that the water level rose and was

maintained. However, the increase in the water level also caused the shape of the Lake of the

Woods to change and, possibly, completely or partially submerged pictograph sites. It seems

likely that known pictographs that are visible today while sitting in a boat would have been

located physically higher up the cliff faces. Yet if the water level had fluctuated as much as

seven feet prior to keeping careful records, the positioning of images could have reflected the

differing high water levels at the time they were painted.

The geology of the rock on which the images exist must be considered because

archaeologists have argued that pictograph sites in the Precambrian Shield region always

occur on granite cliffs immediately beside water. The sites exist on granite arguably because it

is a major component of the bedrock outcrops which are visible because massive ice sheets

scoured much of the Shield. Clearly the sites occur in a locale where the geology and the

geomorphology are considerably more complex. However, none of the archaeologists have

Page 26

considered whether the geology and the geomorphology in the immediate vicinity of the site

is the same. This is a question that must be considered.

The changing water levels must be considered sin ce these affect the images both with

respect to their physical deterioration, and the ability to view and create them from a boat. It

is important to consider whether images are accessible by boat because archaeologists argue

that pictographs always are visible while sitting in a canoe and that they exist near to the edge

of the water. Clearly, a possible relationship exists between water level, the images, the

physicallocation of the sites and their immediate geology and geomorphology.

Section 2: History and ethnography.

The second part of this chapter considers the debates in the history and ethnography

of the indigenous peoples of this area in prehistoric and historical times. Four sets of

problems have been identified: (a) the inability to establish a clear framework for the cultural

and ethnic identities of those who lived here during the historic period; (b) understanding the

archaeological record of the Boreal Forest; Cc) the unresolved debate between historians and

archaeologists concerning the ethnic identification of those peoples at the time of contact

with the Europeans, and finally (d) whether bioanthropological evidence can resolve the

problems of the ethnic identity of these peoples.

To establish the ethnicity of the people who lived in this part of North America is

fundamental because anyone concerned with a homological approach to interpreting these

pictograph sites must establish which ethnic groups produced them. Knowledge of the

ethnicity of different indigenous groups at specific times assists in identifying the precise

body of data required to interpret these rock images. Unfortunately, it is difficult to identify

past ethnic groups from archaeological records. It is easier to gather information about

different ethnographie cultures from the ethnohistorical record. To determine the ethnic

identity of a group of people from a specific geographical region during the historical period

a scholar carefully consults historical records, if they exist. These records begin with the

arrivaI of the Europeans in this region, sometime in the mid-eighteenth century. The

archaeological record must be connected with the historical record in order to determine

which indigenous groups resided in a specific region at specific times. Both historians and

archaeologists have collected a vast amount of information about the peoples who resided in

this area. Many ethnohistorians also have discussed how land tenure, settlement patterns, and

Page 27

.r'.

mobility changed during the historical period (Bishop 1976; Cooper 1939; Hickerson 1966;

Rogers 1963; Rogers and Black 1976). Ali of these issues influence how archaeologists

interpret the archaeological record.

(a) Who lived in this region during the early historical times?

The first of the four problems is connected to the inability to create a clear

chronological framework regarding the "Northern Ojibwa". This ethnological question has

occupied numerous scholars (for example: Bishop 1976; Dawson 1987; Greenburg and

Morrison 1982; Schenk 1994; Valentine 1995), who have drawn on a wide range of

disciplines including history, anthropology, linguistics, and ethnohistory. Different arguments

apply, according to which documents scholars use to substantiate their conclusions.

Unfortunately, there is no agreement about a clear demarcation between the Cree

and the Ojibwa groups or who their ancestors might have been. The designation of the Oji­

Cree causes additional problems for scholars trying to establish the identity of the peoples at

the time of contact in northwestern Ontario, northern Minnesota, and southeastern

Manitoba. It is evident that individuals cali themselves Oji-Cree. This term causes confusion

both for other indigenous groups and for anthropologists, archaeologists, and historians.

Valentine (1995), in her discussion of the historical background of the Severn Ojibwa

groups, suggested why confusion might arise over the identity of the Oji-Cree. She noted

that very little historical documentation existed for the northern Ontario Ojibwa groups

(ibid.: 24). The early documents concerned with the region calied the people of this region

"Cree" and this term was understood as referring to the specific linguistic and cultural group

Cree, whether they were Moose or Swampy Cree (ibid.: 25). Valentine argued that an error

occurred here: many contemporary Severn Ojibwa cali themselves Cree, although they

recognise that their language is different from that of the Cree, who live north and east of

them in James and Hudson Bay (ibid.). The two languages, she stated, are closely related and

share many linguistic features including aspects of morphology, some discourse particles, and

some lexical items (ibid.) . Valentine noted that early historical records from the French J esuit

missionaries and the Hudson's Bay Company indicate that the Ojibwa were in the area

around Lake Superior north of Lake Huron and east to Georgian Bay (ibid.: 24). The

geographic location of these Ojibwa, she asserted, appeared to correspond with that of the

southern Ojibwa dialect. Documentation exists, Valentine observed, that the Ojibwa had

Page 28

moved westwards into Manitoba and Saskatchewan, which in turn indicated that it was the

dialectical differences between the Severn Ojibwa and the southern Ojibwa that caused the

southern Ojibwa to believe that the Severn Ojibwa dialect was a type of Cree (ibid.).

Despite the brevity of his fieldwork, Father John Cooper (1936), an anthropologist,

obtained valuable information concerning the Ojibwa and the Cree in the Lake of the

Woods. He was informed that the Ojibwa were not the original inhabitants of the area but

had driven out the Sioux (ibid.: 2). He also discovered that sorne of the older people spoke

Cree as a result of the Ojibwa-Cree mixture "of long ago" (ibid.). However, no one spoke

Sioux. One of his informants mentioned that the Ojibwa spoken in the Lake of the Woods

differed from that spoken around Lake Nipigon (ibid.).

Greenburg and Morrison (1982: 83) postulated that the individual called "La

Marteblance" by La Vérendrye in 1729 was the Cree Captain "Martin" of the Hudson's Bay

Company. They (ibid.) observed that "La Marteblance" was captain of the Lake of the

Woods Cree, who according to La Vérendrye spoke a "corrupted Cristinaux" and was allied

with the Monsoni of Rainy Lake. Greenburg and Morrison maintained this information

suggested that these Cristinaux were the same individuals who, along with the Monsoni in

1678, had taken furs to the English at Port Nelson to avoid the Ottawa middlemen who had

pillaged them on Lake Superior (Greenburg and Morrison 1982: 83-84 quoting Margry 1888

6: 30-31). Greenburg and Morrison postulated that the Lake of the Woods Cree had not

moved anywhere by the 1730s, since they came from trus part of northwestern Ontario

(Greenburg and Morrison 1982: 84 based upon Burpee 1927: 164). They argued that their

"disappearance" was made explicable by the arrivaI of the French in their lands which meant

the Cree no longer needed to travel to Fort Albany to trade (ibid.).

Both suggested that the descendants of the Lake of the Woods Cree were called

Ojibwa, based upon Warren's designation of the area as being inhabited by people of the

goose totem. Warren, writing during the nineteenth century, stated that those peoples

belonging to the goose totem lived on the remote northern boundaries of the Ojibwa

territory amongst the "Muskeegoes" and the "Bois Forts" (Warren 1984: 45). Greenburg

and Morrison did not dispute both Hickerson's (1956 and 1962) and Hallowell's (1936 and

1955) opinion that during the eighteenth century the "Ojibwa" or "Saulteaux" moved into

the prairies. Greenburg and Morrison agreed with Wheeler's hypothesis (1977b: 123) that

both groups assimilated with other Saulteaux speakers or "Cristinaux" that they met en

Page 29

.~ ..

route. Greenburg and Morrison argued this would account for the statements Cooper (1936:

2) obtained from informants during the 1930s that the older residents of the Lake of the

Woods spoke "Cree", which reflected a mixture of Ojibwa-Cree that had existed long ago.

Greenburg and Morrison's article examined the dominant hypothesis concerning the

migration and emergence of the peoples called the Northern Ojibwa. Researchers postulated,

based on documentary evidence and fieldwork amongst northern peoples, that the groups

known today as the Northern Ojibwa had inhabited the area since contact with the

Europeans (ibid.: 76-78). No migration, or population movement, had occurred sin ce, but

only a diffusion of the term "Ojibwa" to ethnic groups that at the time of contact with the

Europeans were known by a range of different names. These names included: Kilistinon or

Cree, Monsoni, Muskego, and Gens de Terres (ibid.: 76). Greenburg and Morrison offered

an alternative view of Ojibwa culture history to explain this phenomenon. These authors

asserted that the anthropologists Bishop and Hickerson and the geographer Ray had made

two "unwarranted assumptions" about the words "Kilistinon" and "Ojibwa" (ibid.: 78).

Firstly, "IZilistinon" should not be equated with contemporary Algonquians speaking the

language termed "Cree" (ibid.). Secondly, those scholars should have not equated the term

"Ojibwa" with Ojibwa-speakers (ibid.). Greenburg and Morrison argued that Bishop, Ray

and Hickerson applied "misleading and invariant linguistic and cultural categories to

historical documents" (ibid.). Greenburg and Morrison posited that the terms "IZilistinon"

and "Ojibwa" must not be taken at face value because the ethnic groups named in

seventeenth century records were ancestral to those inhabiting the same regions. They

asserted that detailed investigations of the Boreal Forest, stretching from north-central

Quebec to eastern Manitoba, would reveal that many of the hunters spoke a "mixed" dialect

of Ojibwa at the time of contact (ibid. 87).

Greenburg and Morrison suggested that the emergence of the Northern Ojibwa

could be eXplained (ibid.: 90), as it was known from historical records. They noted that La

Vérendrye stated that the groups of hunters who resided north and west of Lake Superior

spoke "corrupted Cristinaux" or "bad Saulteux" (La Vérendrye: Appendix A in Burpee

1927). Greenburg and Morrison argued that since these same people travelled to James and

Hudson Bays they were called "l(ilistinon". Greenburg and Morrison subsequently

postulated that these people started to spend their summers at Sault Ste. Marie either just

before or just after contact with the Europeans in order to fish and take part in intertribal

Page 30

trade and ceremonies (ibid.: 90-91). They agreed with Hickerson (1966) that these groups had

the opportunities to meet other peoples who lived in clan villages along the shores of Lakes

Superior and Huron, yet during the winter they returned to their own lands to hunt and trap.

Greenburg and Morrison hypothesized that during the latter half of the seventeenth century

the hunters who resided north and west of Lake Superior stopped coming to Sault Ste. Marie

because the French and English fur traders built large numbers of trading posts on Lake

Superior and Hudson Bay (1982: 91). This explained, the authors argued, Warren's (1984: 85)

observation, recorded during the nineteenth century, that the Northern "division" of the

Ojibwa, also called Muskegoes (or Muskego or Muskeegoes) and Bois Forts, had been

separated from the southern Ojibwa for at least eight generations, since neither of these

groups had been together at Sault Ste. Marie for at least two centuries. Greenburg and

Morrison (1982: 91) speculated that terms "Ojibwa" and "Saulteaux" accompanied the

Midewiwin ceremony, which Hallowell (1936) and Dewdney (1975b) posited had expanded

along the southern shores of Lake Superior, throughout the prairies and back towards Lake

Winnipeg and the headwaters of the Albany River. Furthermore, Greenburg and Morrison

argued that the only "Northern Ojibwa", who did not practice the Midewiwin ceremony,

used the Severn Ojibwa dialect and considered themselves Cree (1982: 91).

Greenburg and Morrison (ibid.: 91-92) concluded that the application of the blanket

tribal designations of "Ojibwa" or "Cree" produced either erroneous impressions of cultural

homogeneity or discreteness mas king local ecological and social variability of ethnic groups.

What occurred, in reality, was that the blanket designation "Cree" and "Ojibwa" had diffused

among a wide range of Algonquian speakers who already inhabited the Boreal Forest region

but were gradually recognised by the Europeans and therefore appeared to increase in

number over time. Greenburg and Morrison concluded that this explanation was more

consistent with "the dynarnics of population in hunting and gathering groups" (ibid.: 91).

This explanation dealt with Ray's (1974: 102) question of why "Ojibwa" hunters would settle

in regions supposedly abandoned by the "Cree" hunters who left the area because it was

depleted of fur-bearing animaIs. Greenburg and Morrison maintained that these "Cree" and

any other peoples who lived in this region had not left but had merely become "Ojibwa".

They (ibid.: 92) concluded that the use of the term "northern Ojibwa" for the peoples who

lived north and west of Lake Superior created more problems than it solved, and preferred

the term "Boreal Forest Ojibwa-Cree" or Rogers' (1963) term "Cree-Ojibwa".

Page 31

Rogers and Taylor (1981: 213), like Greenburg and Morrison, argued that insufficient

data hindered any attempts to create rigid boundaries for the area of land inhabited by the

Northern Ojibwa at any time in their history. Rogers and Taylor postulated that these

boundaries were blurred by population movements, intermarriage with surrounding groups

and cultural traits originating with other lndian groups and Europeans (ibid.). Both scholars

suggested that the peoples called Northern Ojibwa have never considered themselves as a

distinct group. Rogers and Taylor documented that many names existed that attested the

Northern Ojibwa were either a part or a small subdivision of a larger entity (ibid.) The term

ocipwe was used as a self-designation for a band that lived north of Sault Ste. Marie which

Andrew Graham, a fur trader reported in 1775 was the Nakawakuck Nation (Williams 1969:

204). Graham discovered this term from the Keischatchewan, or Plains Cree, who also told

him that these people were few in numbers, caught their own furs, were excellent beaver

hunters and caught a variety of fish. Graham (ibid.) also established that the Keischatchewan,

although their language was not very different from that spoken by the Nakawakuck Nation,

could not understand it. Rogers (1963: 65) objected to this use of information provided by

Graham's informants, postulating that, since they inhabited the coastal area between York

Factory and Fort Severn and were not Plains Cree, this term would merely imply people who

lived to the south of them. He (ibid.: 66) maintained that the term "Nakawakuck" had no

linguistic, cultural or tribal significance.

Like Greenburg and Morrison, Schenk (1994), an historian, debated the

methodological problems involved in the identification of the Ojibwa during the historie

period. She observed that most archaeological studies proceeded from two premises: (a) that

the groups called Ojibwa were originally Ojibwa and (b) that all native cultures are static.

Schenk observed that the term Ojibwa was used beyond its original boundaries in the

seventeenth century. She proposed it was possible to identify the Ojibwa based on historical

records of Ojibwa tradition, to trace the use of the names from the seventeenth century

through the early period after contact and to explain why people who were not originally

Ojibwa came to be called Ojibwa (ibid.: 395). The people, who came to be known as the

Ojibwa, were first mentioned in Du Peron's letter of 1639 as the nation of the Sault, found at

the rapids or the falls between Lake Superior and Lake Huron (Thwaites 1959 (15): 155).

During the seventeenth century Sault Ste. Marie was at the centre of a vast circumference

filled with Algonquian nations that for a long period of time behaved as separate entities

Page 32

(Schenk 1994: 403). By 1640, Le Jeune referred to the Ojibwa as the Saulteurs, meaning

those peoples who resided at the rapids (Thwaites 1959 (18): 231). The term, Saulteurs,

subsequently became the term used to refer to the Ojibwa throughout much of the French

period (Schenk 1994: 395).

Schenk maintained that, while historians, anthropologists, and archaeologists have

confused these different groups, the Ojibwa peoples themselves knew who they were. She

concluded that the extended use of this name was by outsiders and not the people

themselves, the Ojibwa and their allies calling themselves "the Anishinabeg". Hence, it is

likely that the use of this term Ojibwa is a recent phenomenon. Since these people were once

few in number, scattered over a wide area, and without an overarching political system, it is

probable that the people referred to themselves using localised names rather than one

general name. U sing a single name for small scale societies implies social cohesion of the

people involved as well as self-identity.

Several principles, Schenk argued, existed for naming Amerindian groups. Firstly, the

name varied according to time, place, and who was doing the naming. Secondly, the names

that people had for each other were often different from those given to them by outsiders.

People were either distinguished as enernies (Nadouesioux) or as allies (Algoumequin), or by

where they lived (Winnebago, Saulteur). Thirdly, Schenk argued, most Amerindians do not

even use their name, since they see no need to identify themselves as a group. They refer to

themselves as "the people". She hypothesized that names could be linguistically transformed

to the point where their original meaning was lost.

Peers, also an historian, in 1994 debated the identity of the Ojibwa and the Cree

during the early historie period. She postulated that the names utilised by Europeans were

related to the different relations between themselves and the rival European trading

companies (ibid.: xv). The French, who encountered the ancestors of the western Ojibwa at

Sault Ste. Marie, called them Saulteurs, later Saulteaux and Outchibouec, and still later

Ojibwa or Chippewa (ibid.). Peers asserted that the development of fur trade posts and

missions throughout the interior may have eroded ethnie differences between local groups

connected by kinship and trade, as well as by language and cultural practices, which in turn

may have influenced the widespread use of the terms Saulteur and Ojibwa (ibid.). Peers

contended that the term Saulteur indicated a common heritage rather than a place of origin.

Frequently the same people had different names, or different written versions of the same

Page 33

name, if they traded with both the French and the Hudson's Bay Company (ibid.: xvi). Peers

noted that the Saulteur were calied "Eecheepos" and Oachiapoia lndians" in the Fort Albany

journals of the 1720s and the 1730s. She maintained that these words "were renderings of

ocipwe or Ojibwa" whom the French had identified as the Outchibouec a hundred years

earlier (ibid.). Peers (ibid.) suggested that the Hudson's Bay Company servants used terms for

the Ojibwa which reflected the rivalry between the Hudson's Bay Company and the North

West Company. Hence, different labels, or names, were applied by the Europeans either

because other lndians calied them that, or because the Europeans grouped them according

to their own criteria of dialect, dress, and socio-economic lifestyle.

The inability to create a framework for the peoples who lived in this reglOn ln

historical times stems from two larger problems, namely a limited understanding of the

archaeological record and the continuous changes in the European naming of the people

who lived here. The names applied to the groups who lived between Lakes Superior and

Winnipeg in historical times are unreliable and reflect the various organizing principles of

missionaries, and fur traders rather than linguistic ties. Rhodes and Todd (1981: 60), two

linguists, noted that government documents listed groups by the names that each group had

officialiy been assigned. Anthropologists and linguists, they argued, in contrast, labeled

groups by the name of a group's setdement and language or cultural affiliation. lndeed, the

people that are referred to as Sioux cali themselves as either the Lakota, the Nakota, or the

Dakota. Rhodes and Todd asserted that the reconstruction of prehistoric populations and

movements solely from linguistic evidence was incredibly difficult: the nomadic hunter­

gatherer activities of the ancestors of the modern Ojibwa and the Cree created huge

problems for scholars seeking to associate a particular group with a particular point on the

map (ibid.). Goddard argued that it was impossible to determine accurately how long ago

Proto-Algonquian was spoken, although he guessed that it was between 2,500 and 3,000

years ago (1978: 586). He (ibid.) accepted Siebert's (1967) proposition that the most probable

homeland of the Proto-Algonquian language was the area between Georgian Bay and Lake

Ontario. Siebert had located this homeland by pinpointing the multiple intersections of the

names of significant species reconstructed in proto-Algonquian and the actual distributions

of those species (1967: 40).

Rhodes and Todd noted that the Proto-Algonquian language had numerous

descendants, two of which were Cree and Ojibwa (1981: 60). Both linguists maintained that

Page 34

dialect srufting in both languages was a setlous problem for anyone interested in

reconstructing prehistoric population movements using linguistic evidence. Rhodes and

Todd stated that it was possible for a group to speak a different dialect from that of their

ancestors, rendering any conclusions based on modern dialects worthless (ibid.: 64).

The final problem was the confusion caused by how anthropologists and linguists

used these terms (ibid.: 66). Linguists have problems establisrung the history of each

language, especialiy given how dialects have changed. The problems evident from the

changing of groups' names might only be resolved by the careful use of historical documents

to gain a better understanding of the synonyms that existed. Perhaps scholars should

consider an examination of the names of the different groups based upon changes in place

names over time, in a similar manner to Hartley's (1980) study of the Lake Superior region.

Continuous changes in the naming of the people who lived in this region rend ers it difficult

or impossible to de termine who lived in this region in historical times.

The examination of only one body of data exacerbates the problem. Different

arguments exist according to the type of documents used to substantiate conclusions. The

designation of the Oji-Cree causes additional problems for scholars trying to establish the

past ethnic identity of the peoples who cali themselves Oji-Cree. This term causes confusion

both for other indigenous groups and for anthropologists, archaeologists, linguists, and

historians. Although establishing the ethnicity of the different indigenous groups in this

region is problematic, it is important because it helps to indicate the ethnicity of the possible

creators of the images. This information in turn assists the researcher in identifying the body

of data required to interpret them.

(b) The unresolved debate between historians and archaeologists over ethnie

identification at the time of contact with Europeans

Historians and archaeologists have hotly debated the ethnic identification of the

peoples who lived in the Lake of the Woods region at the time of contact with Europeans

(Bishop and Smith 1975; Bishop and Ray 1976; Bishop 1976; Dawson 1977, 1987; Evans

1961; Hickerson 1966; Mayer-Oakes 1970; Rogers and Smith 1973; Wheeler 1977b; Wilford

1941, 1955; Wright 1968b, 1968c). The positions taken by scholars will be examined to

determine how and why their opinions have differed.

Page 35

Archaeologists have spent a considerable amount of time defining reccurnng

combinations of design traits on ceramics from different periods in search of typological

characteristics that have general cultural significance (see Wright 1967; Stoltman 1973; Reid

and Rajnovich 1991; Lugenbeal 1976). Archaeological research has examined the spatial and

temporal distribution of the different prerustoric cultures, foilowing the standard

archaeological practice of placing past societies in a culture-rustorical context. Historians,

unlike archaeologists, rely upon information provided about the identities and movements of

specifie peoples from written documents created by fur traders, missionaries, or explorers.

Archaeologists rely upon visible changes in the ceramic and litruc assemblages as weil

as other changes that are possibly symbolic expressions of social identity in the

archaeological record. Ceramic and litruc assemblages are especiaily important in this region

as the rest of the cultural material is destroyed by the high acidity, poor preservation, and

lack of stratigrapruc clarity of the archaeological record. The ceramic assemblage is

considered more important than the litruc one for establisrung the identity of peoples prior

to European contact because changes in the construction, materials used, and decorative

elements serve as crucial markers of possible change. Different styles are readily visible and

change more quickly than is the case with litrucs (Meyer and Hamilton 1994: 96 and

Hamilton 1981: 41).

Archaeologists classify ceramics found in the region into three major types: Laurel,

Blackduck, and Selkirk. Each of these types is associated with a cultural tradition composed

of a number of temporaily and regionaily defined phases, such as the Rainy River Composite,

wruch consists of the Sandy Lake, Duck Bay, Bird Lake, and Winnipeg River complexes.

Considerable contributions have been made with respect to understanding the archaeological

record in the larger geograprucal area witrun wruch the Lake of the Woods exists (e.g.

MacNeish 1958). Furthermore, the several taxonomie systems have been employed to

organise the material (Wilford 1955; Syms 1977; Reid and Rajnovich 1990; Lenius and Olinyk

1990) unfortunately complicate any understanding of this larger region. At the same time,

archaeologists often apply different terms to the same thing. Lugenbeal (1978: 45) observed

that Wilford defined the Laurel Focus based upon rus 1933 excavations of Mound 4 in the

Rainy River region, wruch became the Laurel Tradition according to Wright (1967) or,

according to Stoltman (1973), the Laurel Culture.

Page 36

However, if each cultural tradition is ordered chronologically for the area in question

- Laurel, Blackduck, and Selkirk - the problem exists that each successive one could not have

sim ply expelled or erased the previous cultural group responsible for the different types of

ceramics and lithics evident within the archaeological record. Radiocarbon dates from

individual site components date the Laurel period between 900 B.P and 2,000 B. P.,

Blackduck sites have been dated between 400 B. P. and 1400 B. P. while Selkirk materials

date from 300 B. P. to 1,000 B. P. Each period overlaps with the others. Each has its own

early period of dominance based on these radiocarbon dates, which are themselves

considered problematic. The radio carbon dates indicate that there is some sort of overlap

but, to date, archaeologists have recovered very little transitional or mixed ceramic material

that cross-cuts typological boundaries and demonstrates exactly how these ceramic traditions

co-existed.

1t is difficult to equate archaeological cultures identified in assemblages with ethnic

boundaries of specific groups especially if have problems exist in trying to recognise or

identify these symbolic signatures if they are still present. Stark (1998: xvüi) realised that

conventional methods of stylistic analysis used by North Americanists were ineffective in her

attempts to analyse the iconological or symbolic aspects of undecorated goods that

"circulated in very restricted systems". More productive results could be obtained she

asserted by taking alternative methods to the analysis of stylistic variation. She (ibid.) asserted

observed that different scales of social boundaries existed, the ceramics of the Tonto Basin,

in central Arizona, that otherwise remained hidden in the archaeological record. Her edited

volume of papers considered the relationship of style, as it had been defined, to social

boundaries and whether it was possible to "detect" the conditions under which social

boundaries and mate rial culture might concur. Stark maintained (ibid.: 2-3) that scholars

should question why they desire a "unified theory of style" and consider how material

cultural patterning is examined. She observed that the archaeologists who responded to her

request could be divided into two groups based on two basic premises: that "Anglo­

American archaeologists had refined" a range of techniques for conducting studies of spatial

analysis and interpreting distributional patterning and while the French archaeologists had

devised a "useful set of methods studying formaI variation". 1ndeed, Chilton's (1998), Dietler

and Herbich's (1998), Hegmon's (1998), and Goodby's (1998) studies are particularly useful

because each discussed, utilising different archaeological assemblages in a North American

Page 37

context, how to determine social boundaries and material culture. Chilton's (1998) article is

probably the most useful for the region between Lakes Superior and Winnipeg. She

examined the ceramic assemblage so as to consider the relationships between the technical

choices, historical context and society for three late Woodland (AD 1000-1600) sites in the

northeastern United States. She (ibid.: 159) concluded based on the variations in the ceramic

assemblages of these sites that both the Algonquian and the lroquoian peoples, in this

region, were active agents of their own social change who made choices regarding their

subsistence, settlement and social structures.

Shennan, an archaeologist, (1989) has argued that the "question" of the origin of

ethnie groups should be "analytically" separate from the "nature of archaeological cultures"

(1989: 6). He postulated that archaeological culture should not be equated with specifie

ethnie groups in the past because this assertion was "positively misleading". Archaeological

entities should be perceived instead as "summary descriptions of patterns of spatial analysis,"

while ethnicity should refer to "self-consciousness" (ibid.). He outlined four arguments,

suggested by other scholars, why the notion of archaeological cultures as entities should be

rejected. Shennan (ibid.: 13) rejected the idea that archaeological 'cultures' are equivalent to

ethnie groups, demonstrating that distributions in the archaeological record are the outcome

of a number of disparate processes. He postulated that, if archaeologists used quantitative

analytical techniques rather than "presence-absence information," a wide variety of cross­

cutting patterns would be revealed, rather than readily identifiable "bounded entities" (ibid.).

He also maintained that entities such as "cultures" did not exist but were unpredictable

interrelations of various distributions produced by different processes. Concerns with

ethnicity in archaeology were for Shennan (ibid.: 18) becoming "refocused around the

concept of style".

Hodder has argued that, despite interaction across tribal boundaries, distinctions in

material culture persisted for some artefacts (Hodder 1982). He considered (ibid.: 185) how

matetÏal culture influenced, and was a "product" of various groupings of people who created,

used and manipulated it. Hodder asserted (ibid.) that a simple link could not exist between

degrees of interaction, scales of production, and material culture patterning sin ce other

factors intervene such as ideologies, attitudes, intentions, and strategies.

It appears that in the archaeological literature from the region between Lake

Winnipeg and Lake SupetÏor the term ethnicity implies ethnie group. However, Isajiw (1974)

Page 38

discovered, from a survey of sixty-five sociological and anthropological studies on ethnicity,

that only thirteen studies included a definition of some sort. The remainder of these studies

did not clarify what was meant by "ethnicity".

The definition of this term is not easy. Jones (1997: 55) proposed that henceforth

scholars should employ "a processual analysis of ethnicity" rather than the previous rigid

categories of 'race', 'culture', 'society', and 'tribe' (ibid.). Jones maintained (ibid.) that these

rigid categories had various ideological meanings for different groups of people in different

parts of the world. The notion of ethnicity is complex and possibly elusive and not a clear­

cut theoretical tool to be used by archaeologists. Nevertheless, a detailed examination of the

archaeologicalliterature of this region demonstrates that archaeologists are concerned to find

symbolic expressions of social identity in the archaeological record.

Wilford (1941, 1955), an archaeologist, studied the prehistoric cultures of Minnesota

and identified the Late Woodland Blackduck focus and defined the characteristics of

Blackduck ceramics. Other cultural traits included triangular projectile points, long-bone

fleshers, unilateral barbed bone points, bird-bone whistles, antler flakers, and several types of

side scrapers. These people built mounds containing flexed burials with ochre and grave

goods in cylindrical pits (1941: 313-315, 1955: 136). He postulated that the people of the

Blackduck cultural tradition in the Rainy River region might be the ancestors of historie

Assiniboine, who spoke a Siouian language (1955: 136).

MacNeish (1958) identified the Manitoba Focus at a number of sites in southeastern

Manitoba dating AD 1000 to AD 1750 and argued that this focus was similar to the

Blackduck Focus of northern Minnesota (1958: 76-77). MacNeish noted that the Selkirk

Focus occurred between AD 1350 and AD 1750 but did not develop from the Manitoba

Focus. He postulated that similarities between the two were due to "natural interchange"

resulting from contact rather than a common origin (ibid.: 77). These two traditions differed

in terms of burial traditions, the decoration of the ceramics, and other parts of the

archaeological assemblage (ibid. 73). MacNeish postulated that, during the fourteenth

century, the Cree had moved southwards and replaced the Assiniboine between the Lake of

the Woods and the southern part of Lake Winnipeg (ibid.).

Evans (1961), another archaeologist, argued that the burial practices were critical to

the ethnic identities of these peoples. He agreed with Bushnell (1927) that the Assiniboine,

during the seventeenth century, separated from the Yanktonai Dakota. They left the

Page 39

headwaters of the Mississippi to move northwards to the Lake of the Woods and

subsequendy to Lake Winnipeg. Evans disagreed that sites classified as Blackduck were

Assiniboine because they were located outside the present limits of Siouan groups (1961:

271-275). Wright (1963, 1965, 1968b, 1968c), agreed with Evans (1961) that the Blackduck

sites were not occupied by prehistoric or early historie Assiniboine. He asserted that the

presence of Selkirk and the Blackduck ceramics on the north shore of the Great Lakes was

explicable since it reflected the mobility of women, who made the pottery and who moved

residencies as a consequence of "patrilocal residence requirements" (Wright 1968b and

1968c: 107). Both maintained that Algonquian speakers occupied the area alongside the

modern day international border when contact occurred with the Europeans.

However, Wright disagreed with Evans over their precise ethnie identification.

Wright posited that the Ojibwa occupied the heardand of the Blackduck area while Evans

asserted that the Cree were living in this area. Wright (1968b) maintained that

anthropologists would have a difficult task to distinguish between southern Cree and

northern Ojibwa based upon the associations and differences between the Selkirk and

Blackduck foci ceramics (ibid.: 23). He contended that this task would be easy if the Selkirk

focus had belonged to the Cree and the Ojibwa were responsible for the Blackduck ceramics

(ibid.). He also acknowledged, however, that the use of the terms 'Cree' and 'Ojibwa' created

pitfails and argued it was probably more appropriate if these archaeological assemblages were

cailed "Boreal Forest Algonkians" and "Great Lakes-St. Lawrence Forest Algonkians" after

the forest classification. Yet, assigning new names to these archaeological assemblages does

not aid an understanding of ethnicity. It merely confuses the situation further by making the

terminology more complex.

Skinner (1911: 117), Hailoweil (1955: 114-115), Dunning (1959: 3-4), and Hickerson

(1966:4 and 1967: 45), ail of whom were anthropologists, disagreed with Rogers, another

anthropologist (1963), and Wright (1968b), postulating that the Ojibwa arrived in this region

during the historie period. Hickerson (1967), an ethnohistorian, argued that the history of

this region was "indissolubly linked with the history of European exploration and history"

(ibid.: 47). He asserted that the Cree, who were influenced by the fur trade, first occupied this

area. By the late eighteenth century, groups of Chippewa, otherwise cailed Ojibwa, that

originaily inhabited clan-named villages along the north shore of Lake Huron displaced these

ab original Cree, and their close relatives the Monsoni-Cree (ibid.: 45, 47), from the are a to

Page 40

the west of Lake Superior. Hickerson (ibid.: 44) argued from historical documents that, by

1660, the Cree inhabited the are a north of Lake Superior and that between 1660 and 1680

intertribal conflict broke out among the Dakota, the Chippewaj Ojibwa, the Cree, and other

groups in the region west of Lake Superior. The Chippewa and the eastern Dakota made

peace in 1679 that lasted until 1736. Hickerson maintained that at this time the Cree and

their allies, the Assiniboine, were the chief occupants of the area to the northwest and west

of Lake Superior. Consequently, the French traders and officiaIs, who arrived in this region

between 1717 and 1721 with their goal of finding the Western Sea, had to deal with the Cree

and their allies, who controlled Rainy Lake and the Lake of the Woods.

Hickerson's (1967) view, based on La Vérendreye's reports, is different from that

held by his colleagues. La Vérendreye wrote in his letter of June 1736 to the Marquis de

Beauharnois, Governor General of New France, that the Cree and the Monsoni-Cree came

to see him at Fort Charles in the Lake of the Woods to ask about the massacre of the twenty­

one men from Fort St. Charles (Burpee 1927: 213-219). La Vérendreye wrote earlier in the

same letter that he was aware of a group of Sioux in the Lake of the Woods "who were

100 king for Cree and who might perhaps do them sorne harm" (ibid.: 216). Rogers' (1963)

view supported Wright's (1968b) claims; based on information from informants from Round

Lake Ojibwa, northwest of Lake Superior, Rogers (1963: 65-66) asserted that the Ojibwa had

always lived in the interior region north and west of Lake Superior and that farther north

they blended into the Cree.

Bishop and Smith, both anthropologists, (1975: 59) disagreed with Evans'

conclusions (1961: 271-275), describing his arguments as "unsubstantial and weak". Evans

doubted that the Blackduck sites in northern Minnesota were Assiniboine, since the burial

practices evident in the excavated mounds would have had to change twice and the

Blackduck sites extended eastwards beyond the historically known range of the Siouan

groups. Evans based his arguments on Bushnell's daim (1927) that during the seventeenth

century the Assiniboine separated from the Yanktonai Dakota, left the headwaters of the

Mississippi and moved northwards to the Lake of the Woods and then to Lake Winnipeg.

For Evans, burial practices are crucial for identifying each cultural group. He argued

that the Mille Lacs Aspect, assumed to be Siouan, which lies 75 miles southwest of the west

end of Lake Superior, contained secondary burials. The Headwaters Lake Aspect, that lies to

Page 41

the north of the Mille Lacs Aspect, assumed to be a site of Algonquian-speaking Cree, has

primary burials containing Blackduck ceramics.

Essentially, Bishop and Smith (1975: 54- 61) believed Evans did not use the historical

data very weIl. Aware of the divergent views held by archaeologists and historians, they (ibid.

59) postulated that the central question was the affinity of the Yanktonai and the

Assiniboine. Lowie (1910), an anthropologist, had demonstrated that the separation of the

Yanktonai and the Assiniboine languages predated the historie period. Bishop and Smith

(1975: 59), therefore posited that, if the Assiniboine occupied Blackduck sites, the

archaeological evidence should support this claim. Furthermore, they argued that, following

the arrivaI of the fur trade, the cultural differences between the Yanktonai and the

Assiniboine led to animosity between these two peoples. They suggested that the Cree might

have resided in the area and, occasionally, during the early historie period, they and the

Assiniboine could have occupied the same camps; therefore it was highly likely that inter­

tribal trade occurred, possibly involving foodstuffs in ceramic vessels. They proposed that

this might explain the co-existence of Selkirk and Blackduck assemblages at some

archaeological sites.

Bishop and Smith (ibid.) argued that little evidence substantiated the claims that the

Blackduck and Manitoba Focus sites were Cree in origin. They disagreed with Wright's

(1965, 1968b) conclusions that the Blackduck focus was Ojibwa. Wright had based his

conclusions upon direct archaeological and historical evidence with supporting data from

linguistics, ethnology, and physical anthropology. Bishop and Smith (1975) agreed with

MacNeish's (1958), Wilford's (an archaeologist) (1955), and Hlady's (an archaeologist) (1970:

108) conclusions that Blackduck was Assiniboine, as explained below.

Bishop and Smith (1975) argued that archaeological data and the early historical

references supported the view that the Assiniboine, at the time of contact, about AD 1620,

lived in the region from Lake of the Woods, to Lac Seul, to the mouth of the Kaministikwia

River, near present day Thunder Bay. This region was also the heartland of the Blackduck

focus ceramics. Historical documentation led Bishop and Smith (ibid.) to reason that Cree

subsequently moved west beyond the Pigeon River, since the Cree were reported to be

"visiting on the north Shore of Lake Superior" in the seventeenth century. The Ojibwa were,

at the time of earliest European contact, living as far west as Michipicoten Bayon the

northeastern shore of Lake Superior. The Cree lived to the north and the west of the Ojibwa.

Page 42

The Assiniboine lived westward from the Kaministikwia River on Lake Superior in Northern

Ontario, Northern Minnesota, and Manitoba. Bishop and Smith concluded that the early

historicalliterature both supported their findings and explained the discontinuity between the

archaeological and historical materials. Nevertheless, they warned that it was problematic to

"draw a one-to-one relationship between ceramic fod and their ethnic manufacturers" (ibid.:

61).

Articles by Dawson (1975; 1977), an archaeologist, suggest that he violently disagreed

with the perspective of the historians regarding the ethnic identities of the peoples of this

region. He argued that the Algonquian speakers were not recent arrivaIs to northwestern

Ontario (1975: 158 and 1977: 67) and agreed with MacNeish (1958) and Wilford (1941, 1955)

that the Terminal Woodland, or the Late Woodland, period of the Blackduck Tradition could

be equated with Assiniboine (Dawson 1975: 158; 1977: 72). The main strength of his

arguments rejects the reliance by historians on "second hand information" recorded in early

French written records and cartographic evidence of this region (Dawson 1975: 159-169;

1977: 67-73). Dawson (1987: 145) later argued that the "identity of the peoples resident in

the area at the beginning of the contact period cannot be stated unequivocally, for records do

not begin until about the middle of the period". In this paper he appears to imply that the

archaeological record shouid be the sole provider of information because of the lack of good

historical records for the contact period (ibid. 1987: 151).

Bishop and Ray posited that the inability to resolve the ethnic identity of who lived in

this region of North America arose from "methodologicai and conceptual biases" (1976:

117). This "generai failure" occurred because scholars found it difficult "to appreciate" the

transformations consequentiai to European-Canadian influences over the past three to four

centuries (ibid.). Problems would be avoided if scholars considering the protohistoric and

prehistoric period used the archivaI materials relevant to the region in conjunction with

anthropologicai studies.

Bishop and Ray subsequently discussed the problems invoived in attempts to

combine archaeological, ethnographic, and archivaI data to "produce an accurate and

meaningful culture history" using the direct historical approach (ibid.: 123). They argued that

the Iack of knowledge of the early contact period hindered attempts to establish a secure

baseline, and to make assumptions about the rate of acculturation of the lndian groups living

in this region (ibid.). However, they concluded that the available evidence, both historicai and

Page 43

ethnographie, supported their view that the Assiniboine and the Cree occupied the southern

margins of the northwestern Ontario prior to European intervention (ibid.: 137). Wheeler, an

archaeologist (1977b), disputed the notion that the historical facts could "speak for

themselves" and argued that interpretation of these historical sources must be undertaken

with great care (ibid.: 123). He contested the interpretations of the documents used by

scholars to assert that the Assiniboine were the occupants of the region between Lake

Nipigon and the Lake of the Woods when the Europeans arrived (ibid.: 122). Unfortunately,

he failed ta state whether northern Minnesota or southwestern Manitoba were included in

this statement. Wheeler believed that confusion would continue if scholars continued to

identify specifie historical groups with specifie terri tories based on these "vague and

ambiguous records" (ibid.: 123). He asserted that insufficient evidence existed to argue that

the Assiniboine resided in the Lake Forest and the Southern Boreal Forest from Lake

Nipigon to the Lake of the Woods. Wheeler (ibid.: 122) supported the notion that a major

westward expansion of the Ojibwa had occurred through the Boundary region as proposed

by Rogers (1963) and Hailowell (1955). However, he (1977b: 122) stated that, unlike his

colleagues, he was "most dubious of accepting a Cree-Ojibway expansion and Assiniboine

withdrawal" from the Boundary Waters region prior to 1720.

Meyer and Hamilton (1994), unlike their predecessors in this debate, refrained from

violendy disagreeing with previous views. As archaeologists, they realised that the entire

culture-history of this region was based on changing ceramic styles. They asserted that the

peoples from the Great Lakes-St Lawrence Forest and the Boreal Forest interacted with

those who lived in the Plains region to the west and to the south of them throughout

prehistory (ibid. 127). These interactions included the movement of ideas as weil as of

people. They posited that Selkirk ware was produced by northern Algonquians who later

were called Cree (ibid.). Sandy Lake ware, in contrast, was connected to the Siouans, while

sorne of the northerly material was probably created by predecessors of the people

historically called Assiniboine (ibid.). Evidendy, they agreed with Arthurs' conclusion (1986:

263-64) as to the crea tors of the Sandy Lake material on the northern fringes of its

geographical distribution. Meyer and Hamilton maintained that this argument was possible if

proponents correlated the protohistoric distribution of Selkirk and Sandy Lake ware with the

distribution of lndian groups at the time of contact. These archaeologists argued that the

occupants of northwestern Ontario, southeastern Manitoba, and northern Minnesota made

Page 44

Duck Bay, Selkirk, and Sandy Lake ware. Unfortunately, Meyer and Hamilton did not discuss

the identity of those peoples who created the Duck Bay ware. Both acknowledged the

problematic nature of the assumption that a specifie sort of archaeological artefact equated

to a specifie ethnie entity (1994: 125).

Yet, Duck Bay was one of three archaeological complexes that belonged to the Rainy

River composite, along with the Winnipeg River and Bird Lake complexes. Lenius and

Olinyk (1990: 82) argued that these small complexes shared common ceramic traits, and that

the peoples associated with them may have shared similar social, political, and religious

activities. They existed at the same time as the burial mounds, dated to the Late Woodland

period, were being built along the banks of the Rainy River (ibid.). Lenius and Olinyk (ibid.:

83) argued that the Rainy River composite was created by a coalescence of the Blackduck

and Laurel cultures. It was very difficult to find a precise definition of what is implied by the

terms "composite" or "complex" within trus body of literature; however, it is likely that a

composite is probably a larger entity which encompasses two or more complexes. These

composites appear to have a shorter time scale and possibly a smaller geograprucal scale than

the cultural traditions called Laurel, Blackduck, or Selkirk.

Lugenbeal (1976) proposed that the variation ln Laurel pottery of northern

Minnesota was directly linked to temporal change and that Laurel assemblages could be

dated loosely by the changing proportions of the stylistic types of pottery. Reid and

Rajnovich (1991) argued that regional expressions existed witrun the Laurel cultural tradition,

wruch lasted 1,000 years, implying geopolitical divisions. Lenius and Olinyk posited that from

AD 700 to 1000 the Blackduck and the Laurel cultural tradition co-existed but maintained

their own distinctive pottery forms. They speculated that about AD 1000 the ceramics that

belonged to the Rainy River composite emerged as a consequence of a coalescence of the

Blackduck and Laurel cultures. Neither author clarifies why this would have occurred. Trus

change, Lenius and Olinyk asserted, was accompanied by a diminution in the size of the

mounds built in trus region in the Late Woodland period (Kenyon 1970: 66; Arthurs 1986:

19). Lenius and Olinyk noted that mounds belonging to the Rainy River composite and the

Laurel cultural tradition existed together on the same site (ibid.).

Kenyon acknowledged that, although it was difficult to establish whether the

occupation of trus region was continuous, it was used intermittently for 8,000 years (Kenyon

1970: 68). He was confident that this region had probably been used continuously for the

Page 45

previous 2,000 years. Kenyon's conclusions were based on his excavation of the Armstrong

Mounds, which he dated to the Late Laurel cultural tradition and argued were built about

1,000 years ago (ibid.). Unfortunately, Kenyon did not provide any detailed discussion in his

1970 report of the ethnic affiliation of the peoples. It is likely that his conclusions were based

on the excavation of five mounds in the same locality (Kenyon 1986). He argued that the

principal goal of these mounds was to denote sacred places and to verify territorial

occupation. He postulated that (ibid.: 80), after spending the winter in small groups, large

numbers of people gathered there for ceremonial purposes in either the spring or the

autumn, when sufficient food resources were available.

Kenyon's (1970, 1986) opinion regarding the use of these mounds is consistent with

the osteological research conducted by Tobenson and his colleagues (fobenson et. al. 1992;

1994) and the archaeological analysis conducted by Johnson and Ready (1992). Kenyon

(1986) supported his arguments with his descriptions of the burial practices evident from the

skeletal remains and detailed information about grave goods. Kenyon (ibid.: 112-116) found

a large number of small ceremonial ceramic vessels, which Lenius and Olinyk (1990: 83) later

argued belonged to the Rainy River composite and were rare outside the region. These

vessels are of interest to the author of this the sis since a similar vessel was found in the

summer of 2001 at the base of a pictograph site in the Lake of the Woods, at a depth of fifty­

two feet below the present day water level. The discovery of this vessel supplements Lenius'

and Olinyk's (ibid.) observation that only two such vessels were known from outside the

Rainy River composite. Lenius and Olinyk (ibid.: 101) subsequendy argued that the Rainy

River composite had evolved out of the Blackduck and Laurel cultures. Meyer and Russell

(1987: 22) postulated that the Blackduck and Laurel were also the base of the Selkirk

composite tradition. Hence it was believed that both the Rainy River and the Selkirk

composites had developed from the same distant ancestors and shared a common ceramic

technology. Wright (1971: 22-23), Dawson (1976b: 3) and Meyer and Russell (1987: 25-26)

posited that, probably, ancestors of Algonquian-speaking groups made the ceramics of both

composites.

Yet, what was the ethnie identity of these people associated with the Laurel and

Blackduck cultures? Lenius and Olinyk tried to identify each group using a study based upon

glottochronology developed by Voorhis (1978). Both Lenius and Olinyk (1990: 101) and

Syms (1980: 5) disagreed with linguists who rejected glottochronology, since these linguists

Page 46

ignored absolute chronology in favour of identifying sub-group relationships within language

families on the basis of similar patterns of innovation. Syms maintained that

glottochronology was invaluable and therefore linguists ought to reassess its value.

Voorhis (1978) determined that a Cree-Ojibwa split occurred between AD 987 and

AD 1079. Lenius and Olinyk asserted that this date range was similar to the date suggested

for the emergence of the Selkirk composite, which represented the northern "Cree" people,

and the Rainy River composite, which represented the more southerly "Ojibwa" people

(1990:101).

Syms argued that the Blackduck should be reconsidered as belonging to a number of

Algonquian alternatives such as Ojibwa, Cheyenne, or Arapaho (Syms 1982: 7). He figured

that if scholars preferred an Assiniboine identification for the Selkirk culture then the Sandy

Lake and the closely related northerly Wanikan complex, were related to another group. He

asserted that archaeologists should come to a consensus concerning the terminology used.

Glottochronology, although it is considered problematic, provides support for both Lenius'

and Olinyk's (1990) and Syms' (1982) conclusion as to the ethnic identify of these peoples

and how they were related.

The debate about who lived west of Lake Superior remalnS unresolved and as

problematic as ever. Two groups of scholars are embroiled in an academic argument over

which ethnic group resided in the region first. Unfortunately, most of the archaeological

literature for this region is dominated by inventories. Yet, the archaeologists are concerned

with identifying traits so they can be connected to each other and possibly to an ethnic

group. At the same time there is growing evidence of the problematic nature of the

assumption that an archaeological feature is equivalent to a specific cultural entity. The vast

quantity of descriptive literature results from most archaeologists having employed a culture­

historical approach which, although important, remains highly empirical. The culture­

historical approach only requires an archaeologist to provide detailed descriptions of the sites

in question. The entire situation is further complicated because the archaeologists working in

this region rely solely on one primary diagnostic, ceramics.

Archaeologists have agreed that three major cultural types exist: Laurel, Blackduck,

and Selkirk and each is called a tradition, culture, or focus. All these terms refer to different

things. All archaeologists agree that, temporally and regionally, each tradition consists of a

number of phases. The Laurel tradition, for example, has the Rainy River Composite, which

Page 47

consists of the Sandy Lake, Duck Bay, Bird Lake, and Winnipeg River complexes. Each

tradition, or culture, is presumed to be chronologically ordered. However, each successive

cultural group could not have simply expelled or erased the previous one responsible for the

different types of ceramics and lithics. This argument is supported by the radiocarbon dates

which indicate an overlap, but very little transitional or mixed ceramic mate rial has been

recovered to support the co-existence of these traditions. Unfortunately, archaeologists used

different terms for the same thing, and none of them agree who the ancestors of the Cree

and the Ojibwa were during each cultural tradition. It is impossible to consider whether the

Cree or the Ojibwa were related to one specifie culture. Indeed, each culture was composed

of composites and complexes that themselves had links to other groups within the

geographical region in question. It is easier to state whether the Algonquin or the

Assiniboine were in the region. It appears that both groups were, in the region at different

times, but it is unclear precisely which group was where at any specifie times. Clearly it is

very difficult to consider the precise ethnicity of the people who may have created the rock

images of the Lake of the Woods.

(c) Understanding the archaeological record of the Boreal Forest

Archaeologists have argued over how hard it is to find and excavate sites in the

Boreal Forest. Successful archeological research there requires understanding the problems

inherent in the archaeological record of that part of the Subarctic. Hamilton has asserted that

the current interpretation of the archaeological record of this region is a "rather nasty, messy,

and simplistic culture-history" (Hamilton pers. comm. October 2000). Much of the synthesis

of this region is heavily dependent upon the work of MacNeish (1954, 1958), G. E. Evans

(1961), L. A. Wilford (1941), ]. V. Wright (1967, 1972a), E. N. Lugenbeal (1976), W. ].

Mayer-Oakes (1970),]. B. Stoltman (1973) and K. C. A. Dawson (1974, 1976a, 1976b). The

relatively small quantity of information is probably a consequence of the vegetation and the

climate.

Reid (1988) has identified and discussed eleven problems associated with site

excavation in this region, which he subdivided into "syndromes" and "conditions". He

defined "syndromes" as the problems connected to the approachesj explanations created by

archaeologists (ibid.: 188). Eight syndromes exist which he identifies as follows: Telephone

Booth, Telephone Book, Small Site, Sparse Site, Cigarette Package, Flat Earth, Immaculate

Page 48

Conception, and Superficial Analysis. A final syndrome was also a condition: the 49th ParaUel

Syndrome/Condition. "Conditions" were defined as the problems germane to the nature of

habitation sites within the Boreal Forest and to the modem environment of the Boreal

Forest. Two were identified: Collapsed Stratigraphy and Remote Access. To appreciate the

precise nature of each syndrome, it is crucial ta examine them one by one.

Reid argued that Telephone Booth syndrome was evident in terms of how

archaeological excavation was undertaken in the Boreal Forest. Essentially, archaeological

sites were excavated using metre square grids where each individu al was responsible for their

one-metre square and excavated and recorded it at his or her own speed (ibid.: 189). This

syndrome became more prevalent, Reid (ibid.: 190) asserted, as a consequence of cultural

management resource studies conducted by archaeological contractors for profit under

environmental protection legislation. An example of such legislation was the Province of

Ontario's Environmental Assessment Act (ibid.).

Reid asserted it was possible to compare these contractors' reports to a telephone

book because, like a Telephone Book, conservation archaeological reports consisted of

"endless lists of sites discovered on surveys complete with artefact catalogues but little

interpretation and even less Gustified) hypothesis formulation" (ibid.).

The problems with this approach were compounded by archaeological terminology.

Artefacts, whether lithics or ceramics, were measured, counted, classified according to types,

and assigned to a specifie cultural tradition, but sites were described as having a number of

cultural affiliations.

The frequency of sites having multiple cultural attributions causes problems for

archaeologists trying to establish the distribution of specifie cultural traditions and ethnic

groups through time. Halverson (1988), for example, undertook a survey of Cameron, Flint,

Kakagi (Crow), and Stephen Lakes, about 120 kilometres south of Kenora in Northwestern

Ontario. A typical conclusion for a site in this report is as follows: DhKk-22 was culturally

affiliated with "the Selkirk or Blackduck and modern use" (Halverson 1988: 29). Halverson

reached this conclusion based upon the evidence excavated from test pits at this site, which

consisted of four utilised local chert flakes, five fabric impressed body sherds, one primary

unused local chert flake, three unused local chert secondary flakes, fourteen local chert

tertiary flakes, one used quartz secondary flake, and one unused local rhyolite secondary flake

(ibid.: 29).

Page 49

Sorne surveys do not provide an in-depth analysis of the material collected, while

others present the information in a format that makes it impossible to use (for ex ample:

Haywood 1983; Riddle 1982). The description provided for the cultural affiliation of the

Ouimet Site, FbIx-2, on Attawapiskat Lake is an example of a pie ce of archaeological

information that cannot be used anywhere else than within the published report. It was

classified as having "cultural affiliations to the Archaic, Initial Woodland, Terminal

Woodland and heavy modern use" (Riddle 1982: 24-25). Riddle, like many archaeologists,

classified sorne sites as being prehistoric of an unknown affinity, historie and modern. But

the division of the terms Initial and Terminal Woodland, according to Wright (1981: 91), was

for archaeological convenience rather than any indication of a major culture change or event.

Sometimes archaeologists cause additional problems by creating statistics that are

useless and incomprehensible other than within the context of the report itself. Haywood

observed (1983: 32) that only three, or 11.1 %, of the twenty-seven sites he surveyed had

Paleo-Indian artefacts. Rock image sites are treated as separate units of analysis and not

interpreted in conjunction with nearby archaeological sites (see Pastershank 1989 and

Halverson 1988).

Reid argued that the difficulties that exist in accessing the Boreal Forest because of

the Remote Access Condition compounded the Telephone Book syndrome, resulting in little

research being conducted in this region. The Small Site Syndrome was, according to Reid, a

people problem caused by absence of universities in a region the size of Europe, and by

institutions elsewhere not training archaeologists at the graduate level for working in the

Boreal Forest. Only two universities exist in Northern Ontario, and neither institution had a

graduate programme in archaeology and anthropology. Reid argued that the lack of

universities in northern Ontario with graduate and comprehensive research programmes

compounded the logistical problems of working in the Boreal Forest region, since the nearest

universities involved in research were in Manitoba and southern Ontario (1988: 191).

Consequently, the training provided to students in Canadian institutions elsewhere caused

"an almost unavoidable bias towards Iroquoian [big site] 'mind sets' at least for the

Woodland period" or for the Siouian area of the east central Plains (ibid.). Hence the number

of hunter-gatherer sites of the Boreal Forest of Ontario was assumed to be "small" . Very few

large sites have been excavated in northern Ontario, Manitoba and Minnesota. Reid (1988:

Page 50

191) was concerned that current researchers refused to acknowledge that recent literature

discussed examples of large hunter-gatherer habitation sites in the Boreal Forest.

The Sparse Site Syndrome occurred, Reid asserted, because researchers believed that

even small sites occurred only sporadically. Reid relates this syndrome to the Telephone

Booth Syndrome, the Small Sites Syndrome, and the "resultant self-fulfilling prophetic nature

of the associated research designs, methodologies, and interpretations" (ibid.: 192).

The Cigarette Package Syndrome, like the Sparse Site Syndrome, was also a people

problem. Reid maintained that it afflicted the "macho northern archaeologists [the MNA, an

identifiable species which shall be the subject of a suitable paper] who having braved the

Remote Access Condition through sheer possession of the "Right Stuff', beset by black flies,

mosquitoes, keep their field notes on the back of their Players cigarette packs" (ibid.: 193).

He asserted that the existence of this syndrome meant it was impossible to resolve the

Compressed Stratigraphy Condition.

The Flat Earth Syndrome, like the Cigarette Package Syndrome, was for Reid another

problem created by people rather than by the archaeological record. He argued that this

syndrome occurred if archaeologists misunderstood the meaning of the word 'level' since it

did not mean that the archaeologicallayer was flat. If archaeologists did not understand that

this word 'level' meant a cultural episode, the consequences were severe because the

archaeologist could unwittingly cut through several strata (ibid.: 193).

The Immaculate Conception Syndrome was also a people problem since it occurred

when theories and hypotheses "spring full blown from the fertile minds of researchers"

without any serious background research or justification (ibid.). Reid posited that

archaeologists will use whatever hypothesis is currently in vogue and consequently this

problem was an "incurable" one as weil as being very difficult to detect by the "average

reader." He observed that in almost every case where this syndrome existed, very little field

investigation had occurred. If it had occurred, then the resulting fieldwork was superficiaL

He posited that the Immaculate Conception Syndrome was difficult to identify because "the

hypotheses always conform in style."

Reid (ibid.: 195) contended that another people problem, the Superficial Analysis

Syndrome, was closely connected to the Telephone Booth, Small Site, Sparse Site, Cigarette

Package, and Immaculate Conception Syndromes. This syndrome Reid (ibid.) directly

connected to the fact that the examination of the archaeological record of this region did not

Page 51

belong to Willey and Sabloff's Explanatory Period. Archaeological methodologies and

research designs for the Boreal Forest belonged, Reid argued (ibid.: 187), to what Willey and

Sabloff (1980: 130) had termed the "Classificatory-Historical Period: Concern with Context

and Function." Reid maintained that researchers should not be "lazy" and should concern

themselves with moving beyond the preliminary conclusions of the pioneers in the discipline.

The 49th Parallel Syndrome/Condition, he argued, was a people and a geo-political

problem and arising when archaeologists only read the literature pertinent to the region in

which they worked, regardless of the provincial, international, or state boundaries that

defined these regions. Resolution occurred only if archaeologists ignored the current

geopolitical boundaries, read the pertinent literature, and commurucated with their colleagues

in neighbouring regions.

Two "conditions" existed according to Reid: the compressed stratigraphy and the

remote access condition. Conditions, unlike syndromes, were physical problems rather than

problems caused by archaeologists. Reid (1988: 191) maintained that the problem of

compressed stratigraphy, or "collapsed stratigraphy" according to Syms (1977), was

characterised by "multiple long term occupations without sterile soil horizons between them,

compressed into a very shallow site". The problem of a compressed soil horizon was caused

by two principal factors: the slow rate of soil development in the Boreal Forest and the fact

that sites were not occupied on a year round basis but behaved as "way stations" (ibid.).

The second condition was Remote Access. Archaeologists who work in the Boreal

Forest experience severe logistical problems because there is very litde road access to mu ch

of this area. Indeed, most of the region can be only explored by bush plane or canoe.

Valuable and limited resources are frequendy "eaten up" by the sheer expense and time taken

to travel. Travelling by foot is problematic, given the vegetation mosaic of the Boreal Forest,

which includes impenetrable muskeg and bush. Reid argued that the high financial costs

involved in conjunction with no graduate programmes in universities in Northern Ontario

compounded the problem. It is likely that many sites in this region remain unmapped and

unknown as a consequence of difficult logistical problems. He postulated that this remote

access condition intensified the Superficial Analysis, Telephone Booth, Small Sites, Sparse

Sites, Cigarette Package, and Immaculate Conception Syndromes.

He reasoned that if archaeological examinations of this reglOn were to develop

beyond the Classificatory-Historical stage, two factors must be considered. Pirst, realistic

Page 52

research design parameters were required. Second, field methodologies that were pertinent to

the Boreal Forest should be considered because each field methodology had its own unique

problems. He maintained that archaeologists needed ta think carefully about their field

methodologies since no one achieved anything if they sim ply borrowed field methodologies

from areas with radically different cultural and physical environments. Reid (ibid.: 197 - 202)

subsequently considered nine minimum principles that could be "translated directly" into

specific field applications for use while excavating "conventional" archaeological sites.

Reid (ibid. 202) concluded that the archaeological investigation of the Boreal Forest

was on the verge of a new paradigm. He implemented the guidelines he had suggested at the

site DkKp-8 (Reid and Rajnovich 1987). This study has had sorne impact on the excavation

of sites elsewhere in Northwestern Ontario. Halverson (1992) used Reid's (1988) guidelines

for the excavation of DgKl-3, situated on the eastern shore of the Lake of the Woods.

However, although Pastershank (1989) was trained by Reid in archeological field techniques,

as was Halverson (1992), she did not use these guidelines.

The information collected by archaeologists often appears as "in-house" research

reports, frequently referred to as "grey" literature, since they are not published in refereed

journals but by provincial governments or individuals or remain as unpublished licence

reports or even reports to funding bodies. These unpublished reports are difficult to ob tain,

although they are often valuable sources of information, enabling researchers to establish the

level of data collection and analysis in a specific region.

It is evident that the methods by which archaeological information is disseminated

have changed over time. Sorne unpublished reports of the Ontario Ministry of Culture,

Tourism and Recreation, were reproduced in limited numbers and distributed on an

individual basis to select archaeological researchers who were "actively engaged in field

work". The Conservation Archaeology Reports were intended to provide a quick and

economical means of informing those archaeologists involved in regional archaeological

conservation and resource management. These archaeologists were involved in salvage and

emergency excavation, and in creating archaeological inventories under the Ontario Heritage

Act and the Environmental Assessment Act (1983). Archaeologists conducted surveys for

research, conservation, and cultural resources according to regional guidelines, and wrote

reports for each salvage, mitigation, and assessment or inventory project. Vast quantities of

data were generated according to regional guidelines, making them difficult to utilise in any

Page 53

.~.

other format. These reports may appear superficially dull and uninteresting but, despite their

limitations, they are valuable. They provide opportunities to establish whether a particular

region has been surveyed and to search for comparable collections of data.

Since these guidelines ensure that archaeologists apply the same fieldwork methods

and analytical techniques to rock image sites as to other archaeological sites, examinations of

these images are "trapped within a formula". This is a serious limitation, because the practical

methods of obtaining the data are radically different and far more limited; these images

cannot be excavated using the techniques for recovering, cataloguing and analysing data that

archaeologists apply to "traditional" archaeological sites. Sometimes, archaeologists can

excavate the area surrounding these images, but the physical context of the site rarely

provides information about the cultural context of the images themselves. Therefore the

archaeologist must rely to an unusual degree on a range of non-archaeological sources to

study and establish the "meaning(s)" of these images. The difficulties in accessing the

symbolic knowledge of a group of people through the attributes and physicallocation of the

images they produced alone may explain why many archaeologists have often ignored, or

merely described, these images in a perfunctory manner. But the interest in images also

accounts for why others have studied them.

Archaeologists in this region have spent a considerable amount of time examining

the spatial and temporal distributions of different prehistoric cultures. Unfortunately, much

of the work in this region has taken the form of inventory and salvage projects. This makes it

difficult to establish the geographical or temporal extent of these cultures so that scholars

can examine them in relation to other contemporaneous groups.

Hamilton argued that previous attempts to understand the temporal and spatial

distribution of prehistoric cultures have made implicit use of the concept of "edge" or

"boundary" to clarify the relationships between archaeological entities (Hamilton 1988: 39).

The meaning of this term, Hamilton postulated (ibid.), was analogous to the situation that

occurred in an ecological zone when "two disparate biotic communities come into contact"

(ibid.). Like Bird (1961), he argued that these boundaries were not clearly defined but rather

discontinuous cultural boundaries between human populations (Hamilton 1988: 39-40). He

contended that the concept of rnixed elements, which enabled boundaries to exist between

different communities, empowered archaeologists in this region to discuss several artefacts

Page 54

with design traits associated with several different cultural entities. It also provided an

interpretative framework to discuss cultural interchanges and contact (ibid.).

The evidence presented in the gray literature implies that transition periods were

quick. Yet archaeologists have sought to identify possible "transitional pots" in the artefacts

obtained during excavations, based on detailed knowledge of the ceramics that belong to

each cultural tradition of this large geographic area. These pots provide stylistic traits

belonging to more than one cultural tradition and therefore demonstrate clear evidence of

traditions overlapping one another.

It is agreed by archaeologists that a pot or a sherd is to be classified as Laurel if the

body is either tapered or conically shaped and has a slightly outflaring lip profile. Although

their decorative characteristics may vary, they were usually decorated with linear stamps and

serrated working edges. A wide range of decorative effects are found on these smoothed

vessels involving the use of a combination of stamps and variable motifs (Hamilton 1988:

41). Blackduck ceramics are characterised by globular bodies with constricted necks and

wide, outflaring rims. These pots exhibit a wide range of decorative elements done with a

linear cord-wrapped stick as well as punctate impressions on the rim and neck and textile

impressed body surfaces. Selkirk ceramics are distinguishable by their large globular body,

constricted neck, and an upright, straight-sided rim and lip. Surface decoration is limited to a

narrow range of cord wrapped and punctate impressions on the rim, above a textile

impressed body. These textile impressions are distinct from those of the Blackduck ceramics.

As previously stated, these three ceramic traditions spatially and temporally overlap

one another; yet each had its own period of dominance, if one considers the radiocarbon

dates. Yet very little ceramic material has been recovered to date that cross-cuts typological

boundaries and clearly demonstrates how these ceramic traditions interrelated. Hamilton

(pers. comm. October 2000) and Ross (pers. comm. October 2000) assert that very few

sherds exist to explain what happened as the result of the overlap between the Laurel and the

Blackduck periods. Yet it seems possible that both types were used at the same time. A sherd

that is deemed "transitional" might have a shape and/or decorative characteristics that are

identified with two different periods. Reid and Rajnovich discovered from the excavations of

House 1 at DkKp-8 that Blackduck and Laurel ceramics lay side-by-side (Reid and Rajnovich

1983: 6).

Page 55

Several archaeologists, induding Dawson (1976a), Gibbon and Caine (1980), Goetlz

(Goetlz personal communication in Halverson 1992: 57), Halverson (1992) and Koezur and

Wright (1976), daim to have identified examples of "transitional ceramics". Clearly very little

evidence exists for the existence of these types of pots. Goetlz found a "transitional" vessel

on Big Rice Lake in Cass County, Minnesota. Koezur and Wright provided detailed

descriptions of the nine "transitional ceramic type" vessels found at the Potato Island Site,

EIKc-l, near the headwaters of the Albany, English, and Berens river systems in northern

Ontario (1976). The authors (ibid.: 38-39) postulated that these vessels represented the

transition from the Laurel to the Blackduck period. However, the situation was further

complicated by the presence of a Selkirk trait, twined fabric impressions, on the nine

transitional vessels as weil as the Blackduck ones. To Koezur and Wright this suggested that

the Selkirk and the Blackduck periods and the preceding Laurel cultural tradition had a

"clinal aspect in their spatial relationships and that the traditions were not completely dis crete

and unrelated developments" (ibid.: 39). However, only two of the occupations of this site,

affiliated to the Terminal Woodland and Archaic, were excavated, presumably because

disturbance of the site was so extensive (ibid.: 16-18).

Dawson (1977) identified vessels belonging to the assumed transitional period

foilowing the Laurel culture in the Lake Nipigon area at several sites: DiJc-l, DiJc-2, DkJc-l,

and EaJf-1. He (ibid.: 17) observed that transitional Laurel ceramics were found at sites only

in conjunction with late Traditional Laurel ceramics as weil as Blackduck ceramics that were

designated as belonging to the "later Western Area Algonkian culture". Ceramics were

apparently also found in the same region, that were deemed as a "late manifestation for

south of the height of land mixed attributes on Laurel rim forms," since the Laurel Period

occurred prior to the Transitional period (ibid.: 78). He asserted that the Blackduck period

ceramics, whose attributes were evident in the Transitional period, marked the beginning of

the Terminal Woodland Period, which Dawson (ibid.) termed the "Algonkian culture". He

also contended that the existence of mixed Blackduck-Laurel ceramics, and Blackduck

ceramics with fabric-impressed exteriors rather than the Laurel plain exteriors, or Blackduck

cord-maileated exteriors, hinted at an "early close affinity with Selkirk fabric-impressed

wares" (ibid.). Dawson maintained that Blackduck traits developed from mixed ceramics and

marked the beginning of the Terminal Woodland Period Algonkian Culture (Dawson 1976a:

77).

Page 56

Halverson (1992) identified one vessel, from DgKl-3 on the eastern shore of the

Lake of the Woods, as possibly representing the hypothesized cultural shift between the

Middle and Late Woodland periods. The coil technique used to manufacture the pot was

interpreted as a Middle Woodland trait, while the overall cord impression on its exterior

surface was considered a Late Woodland trait (ibid.: 56-57).

Gibbon and Caine (1980: 57-72) examined the transition from Middle Woodland to

Late Woodland cultures in eastern Minnesota. They contrasted the data from four Middle

Woodland complexes with data from transitional St. Croix phase sites, and from early

Woodland phases. Gibbon and Caine (ibid.: 61) considered the St. Croix phase, not the St.

Croix site, to be transitional between the Middle to Late Woodland cultures stretching from

the north-western corner of Wisconsin across eastern and central Minnesota, through the

Boundary Waters Region into the Red River Valley. The dating for this phase was based on

stratigraphie relationships at four habitation sites where St. Croix phase pottery was

associated with Brainerd ware, two radio carbon dates of AD 340 ± 135 and AD 600 as weIl

as the relationship between the St. Croix phase and the Arvilla burial complex. Gibbon and

Caine (ibid.: 62) used this evidence to argue, that the period AD 300 to 800 was the most

probable time range of the St. Croix phase. The ceramic type belonging to the St. Croix

phase retained many Middle Woodland ceramic characteristics, including dentate stamping

and general vessel shape, but St. Croix sites demonstrated an increased variability in size and

quantity in comparison to Middle Woodland sites. They postulated that the transition from

Middle to late Woodland in the eastern Minnesota region involved several major shifts in

subsistence emphasis and in directly related cultural practices. These scholars asserted that

several interrelated variables eXplained this transition: firstly, the increased availability of wild

rice in the major lakes of Minnesota after AD 600-800 (ibid.: 64-65); secondly, climate

change since the appearance of a distinctive Late Woodland subsistence-settlement pattern

corresponded with the beginning of late warmer and mois ter neo-Atlantic climatic episode

(ibid. 65); and finally an increase in population sizes and a resulting increased pressure on

traditional food resources (ibid.). Gibbon and Caine also concluded that, after AD 900, an

increase occurred in population in the Carolinian-Canadian transition zone that correlated

with a way of life based on the intensive harvesting of wild rice.

Vessels that provide evidence of cultural transitions are important despite their rarity.

It seems it is the transition between Laurel and Blackduck with which archaeologists are

Page 57

concerned rather than the transition between the Blackduck and Selkirk cultural tradition.

Why archaeologists are interested in the earlier transitional phase rather than the later one is

unclear. It is agreed that a mixture of decorative traits identified with two cultural traditions

indicates that a pot could be classified as transitional. This leads one to postulate that the

presence of different sorts of traits may indicate that different groups of people were making

ceramic vessels. More evidence should exist to explain what occurred.

Hamilton (1988: 41) observed that the small quantity of sherds that crosscut

"typological boundaries" has caused speculation concerning the relations between the

cultures represented. Does the adoption of technologies reflect the diffusion of new

technologies from one group to another over a wide area, with each group incorporating

innovations into the designs? It is important that archaeologists are relying upon stylistic and

decorative traits, visual elements to identify these ceramics. Hamilton has repeatedly asserted

that these speculations illustrate sorne of the shortcomings of archaeological research in this

region (ibid.). He argues that archaeological sites must be excavated using good stratigraphie

controls, that good survey and sampling strategies are required for locating the sites

themselves, that detailed inter-site comparisons are needed, and finally that solid sociological

explanations are required to explain what brought about the overlaps between the different

cultures.

Arthurs (1978: 58) observed that Sandy Lake ware, which Birk (1977) argues is

diagnostic of the recently defined Wanikan Culture, was corn mon in Northwestern Ontario.

Initially, Sandy Lake Ware was found from northwestern Wisconsin westwards across east­

central ~1innesota into the extreme ~ssissippi River headwaters region. Cooper and

Johnson (1964) found it associated with Clam River focus pottery and Blackduck ceramics at

several sites in northern Wisconsin and northern ~nnesota. The Sandy Lake pottery is

clearly different from Blackduck type ceramics both in its construction and treatment

(Cooper and Johnson 1964: 475; Arthurs 1978: 59). Cooper and Johnson (1964: 479)

postulated that, although more data were required, Sandy Lake ware probably followed

Blackduck in northern ~nnesota.

Arthurs noted that almost all the sites in Northwestern Ontario with Sandy Lake

ware were within a one hundred mile radius of the "hub" of the Wanikan Culture of

northern-central ~nnesota (1978: 59). All of the sites were located either beside or within

close proximity to major waterways that provided access to the ~ssissippi headwaters (ibid.).

Page 58

Unfortunately, although Sandy Lake ware was found associated with Blackduck ware and

historie trade goods, it remains difficult to establish the relations between those cultural

groups. Slight differences existed in the Ontario Sandy Lake ware when compared to that

found in Minnesota (ibid. 59-60). However, the late-prehistoric/ early historie context of the

Wanikan Culture caused speculation concerning its ethnohistoric identity. Wright (Koezur

and Wright 1976: 39) argued that it was Assiniboine, while Johnson (Rackerby 1975: 479)

asserted that it represented the proto-historie Eastern Dakota. Dawson, on the other hand,

stated that it represented the incursion or influence of Mississippian peoples from the south

into Northwestern Ontario. Arthurs refrained from drawing any conclusions as to which

ethnie groups created the Sandy Lake ceramics. Instead, he concluded that future research

should consider its distribution and its stratigraphie and temporal relations with the

Blackduck, Selkirk, and European cultures; and with similar pottery types in Minnesota and

Manitoba.

As ever, the questions concerning the ethnic affiliation of each cultural tradition

remain unanswered. Archaeologists have deemed specifie ceramic elements to be indicators

of cultural traditions. Yet, it is difficult to establish the cultural affiliation of sorne ceramics if

elements of two different and possibly not contemporaneous cultural traditions are present.

Wright (1968b) examined the Ojibwa bands who occupied the northern shores of the Upper

Great Lakes. He argued that it would be "anticipated" that the Ojibwa hunters and gatherers

of that region differed in a number of important aspects from their kinsmen to the west and

the south (ibid. 98). Their kinsmen to the west occupied a richer gathering environment,

while their kinsmen to the south relied substantially on agriculture. Unfortunately, he fai!ed

to identify the specifie bands of the Ojibwa in question. Wright (ibid.: 105) noted that the

evidence of the Ojibwa along the northern shores of the Great Lakes was sparse and recently

obtained. However, these weaknesses in the information provided by the archaeological

record could be "compensated" by the stratified sites containing debris from seasonal

occupations accumulated over a considerable period of time. The evidence provided by these

sites enabled Wright to examine the culture history of the Ojibwa from the historical period

back to approximately the tenth century (ibid.).

Wright's argument is riddled with unsubstantiated claims. He suggested that the

isolated location of the northern Aigonquian speakers meant that their culture had remained

intact until it was "effectively studied by trained ethnographers" (ibid.: 106). Wright assumed

Page 59

that these peoples were not affected either directly or indirectly by the changes that occurred

among the other groups around them. This statement presumes that these peoples did not

move around and that they had always resided in this region. Furthermore, Wright fai!ed to

provide any ethnographie evidence that cultures remained intact over a long period and that

the physical environment remained statie. Instead, he assumed that the evidence indicated

that ceramics were either a borrowed trait, or an erratic foreign element that was weakly

integrated into the indigenous culture. Evidently, two or more different ceramic traditions

were mixed in single components, while the frequency and the presence of the various

ceramic traditions varied through time at single stratified sites. Apparently, according to

Wright, each individual ceramic tradition had a "heart" outside the region in question where

it represented a unified and developing cultural element.

Wright argued that homogeneous lithic assemblages found at Ojibwa sites were,

unlike the ceramics, not subject to spatial and temporal inconsistencies. He maintained that

this "paradoxical" situation could be eXplained in terms of the social behaviour of the

Ojibwa. Wright based his subsequent proposition on McPherron's (1967) and Owen's (1965)

claims that the archaeological situation around the Upper Great Lakes could be eXplained if

one accepted the notion that, in areas of low population densities, exogamous patrilocal

groups acquired women from considerable distances, resulting in a stable male culture and a

mobile female culture (ibid.: 107). Wright argued that, as pottery manufacture was a

manifestation of female culture, the mixture of traits reflected the high mobility of women

amongst different groups. Wright asserted that the relative stability visible in the lithic

assemblages was a natural expression of male cultures. Although Wright assumed that men

created the lithics and women created the ceramic objects, he provided no evidence to

substantiate why he drew these conclusions. The lithic assemblage, Wright maintained,

possessed a "local character" that was associated with the drainage systems in the region and

hence they could be used to make broad spatial comparisons. This conclusion was supported

by the paucity of exotic siliceous stone and other stone trade items in the archaeological sites

in question. The women who used different ceramic traditions in adjacent areas

accommodated to the patrilocal residence pattern which existed in each region. Wright

claimed that the men were closely bound to their specifie localities since they required an

intimate knowledge of the local faunal resources to ensure the survival of the group (ibid.

Page 60

107). This statement is arguably supported by the ethnographie studies conducted by Rogers

(1963) and Bishop (1976).

Culture change, for Wright, was evidently graduaI and he believed this reflected the

semi-isolated development of conservative hunting and fishing bands who, because they

maintained "sufficient contacts" with each other, remained "broadly similar". This article

substantiates the argument that it is possible to discuss the archaeological record of this part

of North America without clearly delineating cultural boundaries and correlating cultural

periods with specifie ethnic entities.

Even examinations of mortuary practices can be useful. Johnson and Ready's (1992)

study, although it is unconcerned with the physical remains themselves, is concerned with

mortuary practices of a specifie group of people. If the mortuary practices are established in

datable archaeologicallayers, it is possible to as certain the mortuary practices associated with

specific cultural traditions. They examined four burial masks from the MdGnstry Mound 2

belonging to the Blackduck period. They concluded that the masks were probably associated

with rites and ceremonies related ta the Earth Diver, Water Panther, or other spirits

discussed by Hall (1979), who considered Menominee or Potawatomi mythical divers.

Johnson and Ready (1992), like Tobenson and his colleagues (Tobenson et. al. 1992; 1994),

did not consider the ethnic or biological identity of the individuals who lived during the

Laurel and the Blackduck periods. These results are use fuI for cultural identification but do

not provide information regarding ethnie or biological identity or any convergence between

the two.

A disconnection exists in the relationship of the identity of the peoples who used the

habitation sites, camp sites, and were buried in this geographical region. It is impossible to

correlate the ethnie, cultural and biological identities of the peoples whose lives are indicated

by evidence in the archaeological record. This is caused by the numerous difficulties that

exist in connecting these different identities. The complex nature of the archaeological

record of the Boreal Forest region probably needs more detailed investigation.

Unfortunately, even if it was possible to obtain more information about the culture-history

of the region, it remains unlikely that we would know more about the ethnie, social, cultural

and biological identity of the people who created rock image sites in this region.

Page 61

(cl) Can biologieal eviclence resolve the problem of ethnie iclentity?

Osteological, biological, and genetic evidence gained from analysing bones, blood

groups, DNA, and mitochrondrial DNA informs the archaeologist about the genetic

relationships, not about the cultural affiliations, of the groups involved. Despite these

limitations, genetic and biological evidence offers sorne hope of aid for those seeking to

determine whether a common ancestry existed, to what ex te nt neighboring groups have

interbred, or even from where sorne cultural elements were derived. Biological anthropology

itself can be subdivided according to whether the scholars were concerned with the

osteology, human blood group analysis, or mitochrondrial DNA.

Three groups of researchers conducted osteological examinations of the burials from

the Rainy River Region of Minnesota and Ontario (Ossenberg 1974; Tobenson, Aufderheide

and Johnson 1992; Tobenson, Langsjoen and Aufderheide 1994). Osteological examinations

of bones recovered through excavation of the burials can aid the identification of specific

groups since skeletal traits are often associated with particular archaeological cultures.

Ossenberg, in 1974, attempted to trace the lineage of certain historic Plains people

back to the Woodland period, and to test the hypothesis that they were descended from

prehistoric Hopewell groups (1974: 15). The frequencies of twenty-six discrete

morphological traits of the skull were analysed for a sample of 942 individuals from 19

locations. It employed a multivariate estimate of biological distance to explore the relative

"affinities" within an archaeological and ethnographic framework. Ossenberg drew her

samples from archaeological complexes classified by archaeologists as burial mounds. The

assumption was that a complex represented the material culture of a single group of people

over time and that successive generations of this population are represented genetically as

belonging to different phases, cultures, or foci. The results demonstrated that the data taken

from the Laurel, Blackduck, Arvilla, Manitoba, Melia, and Devil's Lake complexes

represented a cluster of closely related populations ancestral to the Dakota, Assiniboine,

Cheyenne, and possibly the Blackfoot. Although they were closely related to several

Woodland and historic groups, their affinity to Illinois Hopewell was weak. She concluded

that comparisons needed to be extended to include Archaic period burials and those

associated with other Woodland areas within the Great Lakes Region to try to trace both the

antiquity and the geographic ramifications of this putative relationship (ibid.: 37).

Page 62

Two groups could be distinguished: the southern and the northern Blackduck

population. Ossenberg concluded that "southern" skeletons were similar to the historie

Assiniboine materials and the "northern" Blackduck skeletons were closer to the Chippewa

or Cheyenne (ibid. 37 and 38). She argued that more skeletal data were required for the

Laurel culture.

Analysis of the morphological traits was a useful beginning. lndeed, Lugenbeal (1978:

49), an archaeologist, argued the possibility that the skeletal differences observed by

Ossenberg between the "northern" and "southern" Blackduck populations could be

correlated with the Blackduck ceramic phases from the Smith site, in northern Minnesota.

Lugenbeal (1978: 49) observed that Ossenberg had noted the differences in the skeletons

from the Osufsen, and Shocker sites of the Mississippi headwaters region and from Hungry

Hall 2 and McKinstry Mound 2 of the Rainy River region in northern Minnesota. Ossenberg

(1974: 37, 38) had concluded that despite the similarities between the "southern" skeletons

and the historie Assiniboine materials, the "northern" Blackduck skeletons were closer to the

Chippewa or Cheyenne. He recognised ~bid.: 50) that the ceramics of the McKinstry Mound

2 burials were related to the later Smith site Blackduck phases, as were numerous rims found

in the fill of Hungary Hall Mound 2. Furthermore, the Osufsen site, which was the source of

the majority of the "southern Blackduck skeletons" had more ceramics that resembled those

belonging to the earlier Blackduck phase of the Smith site (ibid.).

Other scholars in archaeology or pathology and medicine (Tobenson et. al. 1992;

Tobenson et. al. 1994) postulated that specifie mortuary practices were related to specifie

culture groups in the Boundary Waters region of Northern Minnesota and western Ontario.

Perhaps differentiating between different mortuary practices could aid those scholars trying

to distinguish biologie al groups. Evidence of the mortuary practices included skulls lacking

occipitals, punctured long bones, eut marks, and the presence of red ochre. Pathological

changes in the bones suggested that infection had resulted in significant morbidity for these

peoples. Other pathologies found could have been caused by a variety of traumas,

degenerative joint diseases, and congenital abnormalities.

Tobenson, Aufderheide, and Johnson (1992) examined 929 long bones from 727

adults and 202 sub-adults buried in Mound 4 built by the Laurel people of the Middle

Woodland period. Four intrusive burials, or thirty-four additionallong bones, also in Mound

4, dated to the Blackduck period. Since none of these long bones was punctured, with

Page 63

intentionally placed holes, they were excluded from the study (ibid.: 507). Blackduck sites,

dating from AD 800 to 1400, co-exist along sites dating to the Laurel tradition in this region.

Holes were found only on the humeri, femora and the tibiae belonging to the Laurel

culture rather than the Blackduck culture (ibid.: 507). Tobenson, Aufderheide, and Johnson

postulated that social differentiation probably existed in the Laurel culture because not all of

the individuals were selected for bone puncturing (ibid.: 511). They concluded that a simple

non-achieved characteristic that included both sexes and adults and sub-adults would be

inherited group affiliation (ibid.: 512).

Tobenson, Langsjoen, and Aufderheide (1994) analysed the hum an remains from

Smith Mounds dating to the Laurel culture. These researchers examined the remains to

determine burial practices, describe more fully the skeletal collection, and generate sorne

important demographic values. The skeletal collection from McKinstry Mound Two served

as the control for their study. The results revealed that the people of the Laurel Culture built

the mounds around 95 Be. They postulated that each episode of burial contained the dead

from a one-year interval of time. Women and sub-adults under the age of three were

underrepresented and were, presumably, interred elsewhere. Life expectancy at birth was 20

years. The average male reached 177 centimetres in height while the average female reached

160 centimetres. Examination of the Harris lines suggested increased stress in the first years

of life. Red ochre was not significantly associated with gender but was much less likely to be

applied to adult skeletons.

This information is useful because it is possible that different biological groups have

distinct mortuary practices. Genetic data may indicate biological similarities or differences

among living groups of lndians. Biologists and geneticists se arch for genetic differences in

the hope that it is possible to solve the bigger question of how exactly different modern

North American lndian tribes are related to each other.

Many researchers wondered whether it was possible that genetic drift rather than

natural selection could explain the genetic differences between different groups of Ojibwa

(Matson, Koch, and Levine 1954; Szathmary and Reed 1972; Szathmary, Cox, Gershowitz,

Rucknagel, and Schanfield 1974; Szathmary, Mohn, Gershowitz, Lambert, and Reed 1975).

Szathmary and Auger (1983) considered the idea that these populations were descended

from one source. Although they considered three tribes in the Pacific Northwest, Ward and

his colleagues (1993), highlighted a valid point that linguistic differentiation could occur

Page 64

much faster than the rates of biological differentiation. However, they (ibid.: 10667)

concluded that it was equally possible to assume that population differentiation had preceded

linguistic differentiation.

Several studies were undertaken to examine the genetic markers among the Ojibwa

or Chippewa (Matson, Koch, and Levine 1954; Szathmary and Reed 1972; Szathmary, Cox,

Gershowitz, Rucknagel, and Schanfield 1974; Szathmary, Mohn, Gershowitz, Lambert, and

Reed 1975). These studies appear to build upon each other. This line of research provides a

complementary, but different, source of information for archaeologists and historians trying

to establish ethnic differences.

Matson, Koch, and Levine (1954) wished to identify the inherited blood factors of

the Chippewa Indians of Minnesota by examining the genetic markers. They discovered that

blood group 0 was the large st blood group and no individu al in the group tested had the

blood group B or AB if they were "full-blooded" Chippewa (ibid.: 416). They also examined

the genetic frequencies of this group to discover and compare the Chippewa genetic makeup

with that of the resident white population, Swedish in origin, to establish whether genetic

drift had occurred. Matson and his colleagues discovered that some genetic drift had

occurred amongst the Chippewa (ibid.: 425). These scholars also noted that, if the

characteristic differences in genes 0 and A were excluded, the genetic information gained

concerning the Chippewa was comparable to that of the Blood and Blackfeet in Western

Canada (ibid.). However, Eull-blooded Chippewa could be differentiated from members of

the Blood tribe by a higher R2, the absence of fj and a higher Ff (ibid.). They argued that the

discrepancy evident in the MN-Ss system among the Chippewa and the Blood tribes was an

expression of the higher M gene in the Blood tribe type. Matson and his colleagues identified

that the Chippewa had the highest frequency of R2 in comparison to other human

populations in the world. The genetic makeup of this group is significant for those

concerned with genetic markers.

Eighteen years later, Szathmary and Reed (1972) discussed their analysis of blood

taken from two widely separated Ojibwa communities living on reserves in Ontario, and

critically examined the results. They were also interested in genetic markers. Genetic data

were taken from the Pikangikum band in northwestern Ontario which they knew had been

exposed to continuous European contact since 1933 (ibid.: 657) and from the Wikwemikong

band on Manitoulin Island, northwestern Lake Huron (ibid. 656) which had been in contact

Page 65

with Caucasians for over three hundred years. These researchers built upon work conducted

to estimate the proportion of Caucasian genes in Indian populations. One of the factors

causing gene variability in the northern American Indian population was the entry of

Caucasian alleles into these gene pools. Szathmary and Reed (ibid.: 667) discovered that

European admixtures in the Pikangikum band in north-western Ontario were insignificant.

However, the authors argued that the sam pIe size taken from the Wikwemikong band on

Manitoulin Island was also too small to be able to "detect any selection" of alleles (ibid.).

The authors discovered that Rh was the most valuable blood group system for

measuring Cauca sian gene flow among North American Indians (ibid.). Szathmary and Reed

(1972: 668) asserted that the GM system was "the most informative of all the genetic

systems" if scholars intended to calculate the rate of gene flow into North American lndian

populations. Genetic markers were crucial for scholars wishing to calculate gene flow.

Two years later Szathmary collaborated with Cox, Gershowitz, Rucknagel, and

Schanfield to compare the serum protein and red cell enzyme systems of the Wikwemikong

and the Pikangikum bands as examples of different branches of the Ojibwa "tribe"

(Szathmary, Cox, Gershowitz, Rucknagel and Schanfield 1974: 49). Their experiments

demonstrated the presence of the genetic markers Al Naspaki and Cc Chippewa among groups

and Tf Dchi only in the Pikangikum band. Szathmary and her colleagues (ibid.: 53-62) were

shocked to discover that neither group had the Mongoloid marker Cm 1, 3, 5, Il.

Ceruplasmin was equal in both groups and all the individu aIs tested were B homozygotes

(ibid.).

Szathmary and her colleagues asserted that genetic flow was not entirely responsible

for the heterogeneity observed between the groups. Instead, genetic drift and the founder

effect had probably played substantial roles in shaping the genetic makeup of the groups

tested. They postulated that this had occurred because, at European contact, the Ojibwa

were not a single homogeneous group; rather they were small groups scattered over a large

region. Yet, these researchers realised that the small size of their samples, the short time

period of only 200 years which had elapsed for the bands to have a separate evolution, and

the 900 miles separating the Wikwemikong and the Pikangikum bands prevented the

researchers from statistically demonstrating the effects of other genetic pressures (ibid.: 63).

Several years later, Szathmary, Mohn, Gershowitz, Lambert, and Reed (1975) again

analysed the blood from 105 Ojibwa from the Wikwemikong band and 96 Ojibwa from the

Page 66

Pikangikum band. Szathmary and her colleagues (1975: 352) considered whether genetic

differentiation had occurred between the two populations in the period during which their

ancestors had lived in geographical proximity. Furthermore, the authors speculated that

genetic divergence and selection might have occurred in both groups after this population

split. They observed that both populations lacked the r" and the R 0 gene identified among

the cognate Chippewa in Minnesota. Szathmary and her colleagues (ibid.: 358) were surprised

at the extent to which the Wikwemikong and Pikangikum bands differed quantitatively and

qualitatively concerning the genetic characteristics of their blood groups. The Wikwemikong

band had the genes called Dia, r, Lua and B which the Pikangikum band did not have (ibid.:

359). DifferentiaI frequencies occurred for the R NS) and MS gene complexes for both

bands. The authors argued that these differences contributed to the "extremely high genetic

and phenotypic chi-square between both populations which was a measure of their genetic

divergence" and explained these differences (ibid.).

Szathmary and her colleagues concluded that it was a matter of genetic drift rather

than rapid genetic divergence amongst the Ojibwa when they stopped being nomadic

hunting bands, settled on reserves and had little contact with the "extremes of the Ojibwa

continuum" (ibid.: 363). The authors (ibid.: 365) postulated that genetic drift, rather than

natural selection, accounted for the genetic differences between the two bands.

Szathmary (1983) examined the possibility of people originating from one place and

collaborated with Auger to consider the idea that those populations descended from a single

source continue share sorne biological, linguistic, and cultural traits acquired through a

common heritage. They (ibid.: 289) argued that in theory a relationship in one sphere could

predict a relationship in other spheres. Szathmary and Auger (1983: 291) argued that

Goddard's (1979) model of historical development from Proto-Algonquian provided a

framework for analyses to be undertaken utilizing biologie al data (ibid.). They also realised

that the Cree and the Ojibwa were closely related and were so closely related that Wright

(1968b: 23) and Dawson (1983: 76) considered their distinction to be an anthropological

convenience rather than "ethnie reality". Szathmary and Auger observed that Dawson (1983)

disagreed with Rogers (1983) and Bishop (1983) but both groups suggested that Cree and

Ojibwa had, over a long period of time, lived adjacently in northwestern Ontario. Szathmary

and Auger (1983: 307-308) viewed this was an excellent case for them to examine by

measuring inter-lndian gene flow.

Page 67

No gene was identified to belong to the Aigonquian speakers in general, but

Szathmary and Auger (ibid.: 295) argued that a variant allele of the Gc system could be called

an Ojibwa marker gene. The presence of this allele was not detected in any other American

lndian population, induding the four Algonquian speaking groups: the Blackfoot, Cree,

Naskapi, and Montagnais. Szathmary and Auger posited that the restriction of trus variant

suggested that a mutation had occurred prior to the Ojibwa territorial expansion. lndeed, its

restriction to the Ojibwa suggests that this mutation is relatively recent, since the gene has

not had time to diffuse beyond the Ojibwa (ibid.).

Szathmary and Auger conduded that genetic data demonstrated that the central

Algonquian Cree and Ojibwa were similar to each other, although they displayed sorne

genetic differences. They explained this genetic similarity as "recent common descent from a

single joint ancestral population". Detailed examination of these differences indicated that

extensive genetic flow between the groups explained these data. They refrained from stating

whether the pattern exhibited by Ojibwa and Cree similarities and differences dearly resolved

the argument among Bishop (1976), Bishop and Smith (1975), Rogers (1983), and Dawson

(1983). However, Szathmary and Auger asserted that the geograpruc presence of the variant

allele of the Gc system, the Ojibwa marker gene, pointed to a recent population movement

by the descendants of the group in wruch the mutation first occurred. These results support

the daim that the Ojibwa resided in the northern Lake Huron-eastern Lake Superior region

prior to their westward expansion three hundred years ago (Szathmary and Auger 1983: 311-

312).

Ward, Redd, Velanica, Frazier, and Piiiibo (1993) raised a crucial point concerning

linguistic and genetic variation in the Americas, wruch is applicable to the debate about the

ethnic identity of the peoples in the central Subarctic. Ward and rus colleagues noted that

world studies of human population had conduded that because "major linguistic phyla

exrubit a pattern of dustering that is congruent with genetic dusters defined by genetic

markers" (ibid.: 1 0663), linguistic evolution probably parallels genetic differentiation in

humans.

They appraised trus observation by evaluating the linguistic and biological

evolutionary affiliations of three tribes of the Pacific Northwest. The results of their analysis

indicated that the ruerarchy of genetic relationships failed to match the linguistic ruerarchy,

Page 68

implying that the time span needed for languages to change into phyla and families was

significantly shorter than that required for molecular differentiation (ibid.: 10667).

The authors concluded that the "tempo and mode" of linguistic and genetic

divergence was discordant if they were compared over a short period of time (ibid.). This

discrepancy occurred because linguistic diversity was generated in a fundamentally different

way from genetic diversity. Furthermore, if language change was a cultural phenomenon

driven by historical and cultural events, then languages should experience periods of change

and stasis. Molecular evolution, in sharp contrast to language evolution, develops at a

steadier rate. Research on the mitochrondrial DNA of North American lndian populations

has meant that the biological relationship between modern and ancient peoples can be

examined. This mitochrondrial DNA research could easily supplant earlier genetic research

as well as research on blood types and osteology since it is probably more reliable. However,

it is still in its infancy and its limitations in aiding archaeological concerns need to be

established before this type of research is accepted as the solution to archaeological

problems.

Five mtDNAs (A, B, C, D, X) have been identified as directly descending from five

founding Asian mtDNAs (Wallace and Torroni 1992: 406-412; Schurr et. al. 1990; Torroni

et. al. 1992). What is fascinating for scholars, regardless of discipline, is that haplogroup X

was found amongst the Chippewa (Smith, Malhi, Eshleman, Lorenz, and Kaestle 1999). Very

few North American lndian groups have this marker in their genetic makeup. AlI five of

these haplogroups are found in Asia, which is presumed to be the homeland of all Native

Americans.

These haplogroups occupy central nodes in a phylogeny of New World mtDNA

haplotypes and are widespread throughout North, Central, and South America (Smith et. al.

1999: 272). Researchers have used mitochrondrial DNA to study the phylogenetic history of

the human genetic population and archaeological and anthropological questions about the

colonization of the New World (Lorenz and Smith 1994, 1996, 1997; Torroni and Wallace

1995; Stone and Stoneking 1993; Wallace and Torroni 1992). Lorenz and Smith (1996)

argued that sorne correspondence exists between language group affiliations and the

frequencies of mtDNA haplogroups in sorne tribes, while geographical proximity perhaps

influences the genetic similarity among other tribes. Yet, Smith and his colleagues (1999) and

Lorenz and Smith (1996) postulated that genetic drift was not so widely distributed as to

Page 69

r-' ..

explain the uniformity in the mtDNA haplogroup distribution evident amongst many tribal

groups of modern North America. Researchers should be aware of this if they wish to use

modern data ta assess ancestorj descendant relationships between living populations and

prehistoric skeletal material (Smith et. al. 1999: 272).

These advances in genetic technology have enabled archaeologists and linguists to

test hypotheses about the temporal and cultural origins of Algonquian history. Although this

research has revealed interesting information, these scholars do not question the ethnic

identity of the inhabitants at the time of European contact. Some researchers were

concerned with establishing whether the Proto-Algonquian speakers had a western homeland

prior to arriving in northeastern North America (Scozzari et. al. 1997; Smith et. al. 1999;

Schultz, Malhi and Smith n.d.).

Recent regional studies have investigated genetic diversity on smaller regional scales

(Malhi et. al. 2001,2003,2004; Bolnick and Smith 2003, Eshleman et. al. 2004). Their results

have indicated that populations within a geographical region exhibit genetic similarity that

indicates that genetic flow limited genetic drift. This contradicts the early studies of genetic

differences and their distribution in North America which had argued that genetic

differences resulted primarily from genetic drift (e.g. Schurr et. al. 1990). Shook (2004: 3)

argued that these early studies were undertaken with "a limited understanding of the genetic

diversity and its distribution geographically and temporally across the New World". The

mtDNA patterns provided by these small regional studies strongly supports the hypothesis

that North America probably was colonised by small bands of people who were sufficiently

large to sus tain gene flow but small enough for significant genetic drift to be evident (Malhi

et. al. 2002). Genetic drift gradually declined as populations grew and bands became more

sedentary. Consequently specific regional genetic patterns emerged (Shook 2004: 4). She

asserted that variation in geographically similar populations in northwestern North America

possibly resonates with shared cultural traditions, language patterns, and even recent

demographic history. Ancient populations from two specific locations within northwestern

North America were compared with ones from southwestern Ontario and the Central

Illinois River Valley (ibid.: 9).

Shook (ibid.: 8) examined how mtDNA patterns changed in a large region over time

and geographical space by comparing genetic patterns of ancient populations. She compared

the genetic patterns of ancient populations from different time periods in the same

Page 70

. ~.

geographic locality with the same time period in two different geographic localities. The

hypotheses generated from the archaeological record and historical linguistics were also

contrasted to establish the nature of the role of cultural factors in genetic variation in the

Northeast. This region of North America was chosen because the mtDNA variations in

these populations represented multiple language families (Malhi et. al. 2001) and the Norris

Farms Oneota which is the largest ancient population studied to date (Shook 2004: 4).

Modern Algonquian speakers have high frequencies of haplogroups A, C, and X but

low frequencies of Band D (Lorenz and Smith 1996; Malhi et. al. 2001; Scozzari et. al. 1997,

Smith, et. al. 1999). Only 3 % of ail native North Americans belong to haplogroup X and

individuals belonging to this group are absent in native South America (Smith et. al. 1999) .

Shook and Smith (2004) knew that linguistic studies of Algonquian demonstrated

that at least one extensive expansion had occurred in northeastern North America over the

past 3000 years. The samples used for MtDNA analysis for Shook and Smith's study were

taken from prehistoric samples that represented this time period (2004: 5). However, the

Winnebago and another language family also existed in this are a, namely the Northern

Iroquoian.

The original speakers of the Proto-Algonquian language, whose descendants

currently inhabit almost a quarter of North America, arrived in this Central Canadian region

from the northwest approximately 3000 BP according to Denny (1989, 1991), who based his

estimate upon linguistic relationships between the Algonquin and several West Coast

language groups (Sapir 1916; Haas 1960, 1965; Goddard 1975, 1994). Fiedel (1987, 1990) had

speculated that language and possible cultural expansion and diversification of the

Algonquians occurred between 2600 BP and 1500 BP.

All of the samples that were used for mtDNA sampling came from populations that

had burial complexes. Shook (2004: 21) observed that genetic continuity was evident in both

localities tested and that both populations tested were biologicaily unrelated. However, the

mtDNA haplotypes demonstrated that lineages were shared between ancient and extant

individuals in each locality. She concluded that these lineages have been present in the

Northeast for the last 3000 years, that the Northeast exhibited genetic homogeneity and that

"variation was distributed clinicaily", which could be explained as "gene flow restricted by

geography" (ibid.: 22).

Page 71

Shook conduded that her study highlighted how important it was to consider the

genetic analysis of prehistoric populations to "determine when the patterns of population

genetic structure of extant populations were established" (ibid.). She maintained that

examining genetic data even from populations dating from 3000 BP was useful for those

concerned with the interpretation of the relationships between different archaeological

traditions. She also speculated that patterns of genetic variation could be eXplained by

linguistic and archaeological data.

Parr, from another research group based in Canada, rather than in the D.S., as were

Shook and her coileagues, directed a project where morphological analysis was conducted in

conjunction with paleogenetic data obtained through DNA from the Armstrong Mound first

excavated by Kenyon in 1966 (1970, 1986). The Armstrong Mound is part of a large complex

of ceremonial mounds in the Rainy River region. Parr and his coileagues intended to identify

the founding Amerindian haplogroup of these skeletal remains. These results will be

interesting but the report will only be published once a complete bioarchaeological

assessment on this burial mound has been completed.

However, Bianchi and Rothhammer (1995: 1238) warned that mtDNA research

might not be the "panacea that will resolve ail our anthropological doubts". lndeed, it is

important to keep in mind that such genetic data can inform only the archaeologists about

genetic relations, not about the cultural affiliations of the group involved.

Osteological, biologïcal, and genetic information informs the archaeologist about the

genetic relationships, not about the cultural affiliations, of the groups involved. This body of

information is incredibly useful if genetic analysis is used to question linguistic and

archaeological theories and data.

Osteological analysis by Ossenberg (1974) tested the assumption that a burial

complex represented the material culture of a single group of people over time. She

conduded that comparisons needed to be extended to indude Archaic period burials and

those associated with other Woodland areas within the Great Lakes Region to try to trace

both the antiquity and the geographic ramifications of this putative relationship (ibid.: 37).

She identified two groups: the southern and the northern Blackduck population. Ossenberg

conduded that "southern" skeletons were similar to the historie Assiniboine materials and

the "northern" Blackduck skeletons were doser to the Chippewa or Cheyenne (ibid. 37 and

38).

Page 72

Tobenson and bis colleagues (1992, 1994) tested the ide a that specific mortuary

practices were related to specific culture groups in the Boundary Waters region of Northern

Minnesota and western Ontario during the Laurel and the Blackduck period. They posited

that differentiating between different mortuary practices could aid those scholars trying to

distinguish biological groups in the archaeological record.

Matson, Koch, and Levine (1954); Szathmary and Reed (1972); Szathmary, Cox,

Gershowitz, Rucknagel, and Schanfield (1974); Szathmary, Mohn, Gershowitz, Lambert, and

Reed (1975) searched for the genetic differences to try to solve the bigger question of

precisely how different modern North American lndian tribes are related to each other. They

searched for genetic markers to identify whether genetic drift rather than natural selection

explained the genetic differences between different groups of Ojibwa. Matson and bis

colleagues identified that the Cbippewa had the bighest frequency of R2 in comparison to

other human populations in the world.

Szathmary and Auger (1983) examined the possibility that different populations

descended from a single source continue share sorne biological, linguistic, and cultural traits

acquired through a common heritage. Hence a relationsbip in one sphere could predict a

relationsbip in other spheres. They concluded that genetic data demonstrated that the central

Algonquian Cree and Ojibwa were similar to each other, although sorne genetic differences

as a consequence of extensive genetic flow between the groups.

The work of Ward and his colleagues (1993) highlighted a key point that linguistic

diversity was generated in a fundamentally different way from genetic diversity. They stated

that if language change was a cultural phenomenon driven by bistorical and cultural events,

then languages should experience periods of change and stasis. Molecular evolution, in sharp

contrast to language evolution, develops at a steadier rate.

Examination of the mitochrondrial DNA of North American lndian populations

enabled the biological relationship between modern and ancient peoples to be considered. It

could supplant earlier genetic research, research on blood types and osteology because it is

probably more reliable. However, it is still in its infancy and its limitations in aiding

archaeological concerns should be acknowledged before tbis type of research is accepted as

the solution to archaeological problems.

Five mtDNA haplogroups (A, B, C, D, X) were identified as directly descending

from five founding Asian mtDNAs (Wallace and Torroni 1992: 406-412; Schurr et. al. 1990;

Page 73

Torroni et. al. 1992). Haplogroup X was found to exist predominately amongst the

Chippewa since very few North American Indian groups have this marker in their genetic

makeup (Smith, Malhi, Eshleman, Lorenz, and Kaestle 1999).

Lorenz and Smith (1994, 1996, 1997), Torroni and Wallace (1995), Stone and

Stoneking (1993), and Wallace and Torroni (1992) used mitochrondrial DNA to study the

phylogenetic history of the hum an genetic population and archaeological and anthropological

questions about the colonization of the New World. Sorne researches argued that sorne

correspondence exists between language group affiliations and the frequencies of mtDNA

haplogroups in sorne tribes, but geographical proximity may influence the genetic similarity

among other tribes. Others argue that genetic drift was not so widely distributed as to explain

the uniformity in the mtDNA haplogroup distribution evident amongst many tribal groups

of modern North America.

These advances are interesting and enable archaeologists and linguists to test

hypotheses about the temporal and cultural origins of Algonquian history, but this

technology does not allow us to determine the ethnic identity of the inhabitants at the time

of European contact. Shook's (2004) study concluded that researchers concerned with the

interpretation of the relationships between different archaeological traditions should examine

genetic data, even from populations dating from 3000 BP. She also speculated that patterns

of genetic variation could be eXplained by linguistic and archaeological data. Unfortunately, it

remains very hard to relate the results of these studies to the study of rock images in the

Lake of the Woods because archaeologists and historians remain unclear as to who precisely

lived in the region and the ethnic identity of the individuals who may have created these

1mages.

Section 3: The images created and used hy the AIgonquian speaking peoples

This section discusses the imagery created by the Aigonquian speaking peoples and

the Assiniboine. It has two goals. The first is to establish the ethnicity of different images

and the objects upon which they exist so that it can be determined which images were

associated with either the Algonquian-speaking peoples or the Assiniboine. The second goal

is an examination of the images as a means of communication to better understand the rock

image sites.

Page 74

It is widely acknowledged that visual representations play an important role in

communicating information to the viewer (Boone 1994; ]aritz 1993; Vastokas 1996). It is

likely that these images were polysemie in their meaning and pointers to further sources of

information. The individuals involved in the creation of the objects and their decoration

manipulated their meaning.

Images and the objects upon which they exist have been examined by scholars from

a range of disciplines, including art history, new art history, archaeology, and anthropology.

The attitudes, perceptions, and the discipline of each scholar also have influenced inevitably

the examination of AIgonquian images, whether etched, quille d, embroidered, drawn, or

beaded.

Sorne regions have been examined more than others. Faulkner, Prince and Neptune

(1998: 32) asserted that the bead working traditions of northeastern North America have

received "scant attention in comparison" to the quantity of work undertaken on the Plains

groups.

Moreover, sorne of these images exist on objects traditionally viewed as marginal

sources of information since the objects belonged to such categories as art, decorative crafts,

handmade objects, handicrafts, souvenirs, or garments (Densmore 1974 [1929]; Ritzenthaler

and Ritzenthaler 1970; Ritzenthaler 1972). Phillips (1999: 17) argued that this western

method of dichotomizing high art and crafts and souvenirs provided no framework for

understanding objects on which AIgonquian images appear, which she called "souvenirs".

She overcame this problem by combining anthropological and historical data in conjunction

with field, archivaI, and museum research to contextualize her analysis and discussion of the

meanings of the objects known as 'native souvenirs' and of the images upon them.

Overcoming these imposed terms is important. Brasser (1976), unlike Ritzenthaler

(1972), realised that garments made and decorated by Aigonquian speaking peoples could

only be understood as clothing if they were contextualised within an historical setting. Reid

and Vastokas (1984) hailed the exhibition entitled "From the Four Corners: Native and

European Art in Ontario 5000 BC to AD 1867" as a landmark in Canadian history because

both European and Native artistic traditions were given equal attention in terms of a single

chronological framework, and were examined as a mutually interacting aesthetic system in

response to a common set of historical, geographical, and cultural circumstances. Reid and

Vasto kas (ibid.: 9) reacted against the old categorisations of native art as "primitive and

Page 75

lac king a documented history." They asserted that these divisions only supported an

'apartheid status' for native art. They argued that, henceforth, these objects should no longer

be perceived as handicrafts but as important, although difficult, to understand.

The objects that concern us come from a very wide graphical region which

encompasses the Lake of the Woods area between Lakes Winnipeg and Superior, specifically

Northwestern Ontario, northern Minnesota, and southeastern Manitoba. The geographic

region in question must be expanded westward and eastward to include the Great Lakes

region to the east of Lake Superior and the Plains region west of the Lake of the Woods. The

Algonquian speaking peoples, who were probably residing in the region described ab ove at

the time of contact, were being pu shed westward and were influenced to some degree, both

direcdy and indirecdy, by different indigenous groups and recendy arrived non-native

peoples. It is likely that some Algonquian groups were pu shed into the Plains region in

conjunction with the Assiniboine.

(a) Is establishing the ethnie origins and its objects possible?

It is evident from earlier sections in this chapter that, at least by the early seventeenth

century, the Assiniboine resided seasonally on the western half of the Lake of the Woods and

would do so continually until the middle of the eighteenth century (Heidenreich 1987: Plate

37). The Algonquian speaking peoples, throughout this time period, had lived in the eastern

portion of the Lake of the Woods. They gradually took control of the region as the

Assiniboine moved westward into the Plains (Heidenreich and Noël 1987: Plate 40).

Essentially the Algonquian speaking peoples lived in this region for a long period of time.

Distinguishing between the images produced by the two groups is harder since

stylistic, cultural, and social influences can remain long after a group is no longer in direct

contact with the groups from whom it learned a particular technique and manner of doing

something. This task of distinguishing between the traits that are Assiniboine and those that

are Algonquian is limited to the nineteenth century, because much of the material was made

of perishable materials that frequendy fail to survive in the archaeological record.

Greene (2001: 1038) observed that for a long period of time quillwork was the major

decorative art form in the Plains. During the latter part of the nineteenth century, quill use

declined with the introduction of pony and, subsequendy, seed beads. Large simple

beadwork designs were created on a white background that stylistically resembles that made

Page 76

~ ..

by the Plains Cree (DeMallie and Miller 2001: 1041). Whereas the Assiniboine, in contrast to

many other Plains groups, used the overlay stitch, which results in a flat, rather than a ridged,

surface (ibid.).

Rock image sites also exist in the Plains region (Keyser and I<lassen 2001) but the

"biographic" and the robe and ledger styles are quite different from those of the Canadian

Shield. Images that belong to the "biographical" style were created "to advertise a person's

standing in society and the world" (Keyser and I<lassen 2001: 244). This pers on was usually a

warrior. Keyser and I<lassen (ibid.: 224) argued that images deemed as biographie were

"largely "narrative"" since they portrayed everyday events and recorded historical occasions.

Images invariably consisted of different scenes containing large numbers of figures. The rock

images sites that are biographic in style date between the early 1700s and the late 1800s. The

robe and ledger styles associated with this region are connected to bison robes and the ledger

book paintings used during the late eighteenth and early nineteenth century (Keyser 1977,

Greene 2001: 1042).

Skinner's (1915) article is useful, sin ce it indicates how a group maintained decorative

traits and designs which were part of the original Algonquian group mindset, while changing

and gradually becoming part of another cultural mindset. He based his claims on the Bungi

Ojibwa, a Plains group originally from the Woodland region, who had not abandoned all the

Woodlands traits but had adopted and amalgamated sorne Plains traits. Skinner (ibid.: 315)

observed that their "decorative art" was only found on painting on raw hide, quillwork, or

bead work and that they used rectangular and triangular designs to create complex figures.

However, the Bungi often (ibid.: 317) thoroughly mixed the floral designs from the "forest"

with the Plains geometric images, while their "decorative art was more inclined to

conventional flower forms" (ibid.: 316). He concluded that (ibid.: 318) the Bungi used sorne

Woodland features because they were recent arrivaIs in the Plains region and that their

culture was under pressure from "outside influences". The Bungi also were slowly adopting

Plains cultural traits, su ch as the Sun Dance and other "folklore elements" (ibid.), including

various facets of the Plains religious, social, and politicallife.

Sager (1994) considered eight pairs of beaded moccasins, collected in southern

Saskatchewan, as an example of cultural borrowing by the Cree and the Ojibwa (Saulteaux).

Several bands of Cree and Ojibwa (Saulteaux) had become permanent residents of the

grassland borders as a consequence of the establishment of reserves in western Canada

Page 77

during the late nineteenth century. The Plains Cree were permanent residents of this region

before the establishment of the reserves. Sager (1994) concluded that the decorative elements

of these moccasins demonstrated that their makers had adopted Plains aspects of material

culture. His (1995) paper examined the origins and significance of the "crooked nose"

moccasin design of the Northern Plains. He argued (ibid: 38) that this design was identified

with the Blackfoot tribe but should be considered as generally Northern Plains, although it

had existed longer amongst the Blackfoot than any other group. Irrefutable evidence, he

asserted (ibid.: 42), existed for the presence of quilled and beaded crooked nose type designs

on Ojibwa and Iroquois moccasins dating back ta the 1700s. Sager concluded that evidence

of the crooked nose design was found on early moccasin types from many groups, who

added the longitudinal seam to the terminus of their centre longitudinal seam.

He argued that attention ought to be given to the design of moccasins made by the

Cree. He maintained that evidence existed that one or more Eastern Algonquian groups,

including the Cree and the Ojibwa, as well as the Métis had introduced the notion of vamp­

like decorations on moccasins amongst the Plains and the Intermontane groups (Sager 2000).

He (ibid.: 68) asserted, however, that both types, the false vamp and the form fitting vamp

type of moccasin, had slighrly different origins in the Subarctic region. The form fitting vamp

moccasins emerged during the late eighteenth century somewhere east of Lake Superior and

were the more widely distributed of the two types in the regions south of the Subarctic. The

false vamp moccasin, which was a simpler design, also had its origins in the Subarctic region

and the Red River region. It began gaining wider distribution during the early nineteenth

century, when it was worn by Cree, Ojibwa, and Métis and traded by everyone regardless of

whether they were native or not. Its use eventually spread as far south as Minnesota, to the

Dakotas, and northwest to the Saskatchewan River.

Objects made by the Algonquian speakers had an influence far beyond their linguistic

boundaries. Sager (2002: 76) maintained that no clear line of demarcation existed between

the Woodland and Plains groups in what he termed "artistic expression". Indeed, he asserted

that there were a "significant number of designs which eventually appeared in Assiniboine

quill- and beadwork which were omnipresent in the East". He suspected that part of the

problem in identifying Assiniboine quill- and beadwork stemmed from the fact that very litrle

may have existed in the first place. A second reason for the lack of distinction was that the

Page 78

Assiniboine, although they had moved westward, were associated with their allies to the East,

and with those in Red River and the Lake Winnipeg region (ibid.: 77).

He (ibid.: 80) contended that design elements were "too fundamental to be readily

assigned to any kind of regional exclusivity" since they were found throughout the entire

continent. However, by the reserve period, specific features enabled Assiniboine objects to

be identified (ibid.). They used lane stitch beadwork; the backgrounds of their quiilwork were

often undyed white, or dyed red or bright yeilow, and there were more dark green

backgrounds in their beadwork in comparison to other groups. They also widely employed

the bicoloured lozenge feather symbol. Assiniboine moccasins were decorated with three

bands of lane stitch beadwork, the centre of which was a darker contrasting colour, which

frequently formed part of the designs (ibid.: 80-81). Design elements usuaily included stacked

low triangles, or complex designs such as "radiating opposed isosceles triangles" and

rhomboids (ibid.: 81). The Assiniboine also continued to make quilled objects, although the

Blackfoot had abandoned this technique du ring the 1880s (ibid.).

Sager (ibid: 81) claimed that it was difficult to establish the precise date when the

Assiniboine started to utilise floral designs in their beadwork and quillwork for a fuiler

discussion), since the documentation on specimens from the early nineteenth century was

vague. This made it difficult to determine whether the Assiniboine or another group made a

specific item, although it had been obtained from the Assiniboine. However, Sager (ibid.: 82)

stated that a "limited amount of work could be reliably attributed to" them, but that the

farther away from the region where floral work was done, the more abstract designs became.

Specific design traits that are pertinent to different groups in different parts of North

America result from indirect or direct contact possibly involving either native or non-native

groups. However, a large body of historical and ethnographical literature exists, enabling

some of these traits and trends to be established. Some scholars have examined the

development of images on specific items to determine whether certain traits are Algonquian

or Plains (for example Duncan 1991; Sager 1994, 1995, 1996, 2000, and 2002). Others have

considered the evolution of imagery, now recognised as belonging to the Great Lakes region,

regardless of the object in question (penny 1991; Phillips 1984 and 1991; Whiteford 1991).

Still others have summarised ail of the studies conducted on Ojibwa bead- and quillwork

(Garte 1985).

Page 79

Sorne have considered the influences of other indigenous lndians groups, European

traders, and settlers regarding the range and type of designs utilised, as weIl as how their

meanings were manipulated (for example Lenz 1986, Phillips 1999, Lanford 1984; Peers

1999). Peers' (1999) interest lay principally in how the meaning of one octopus bag, made in

the early 1840s in the Red River region had shifted and changed from the time it was made

to its becoming a museum artefact. Sorne of the designs used by the Algonquians have

developed and changed in the shapes of images used in bead- and quillwork. The least

amount of change has occurred in the images on birch bark scroIls. A connection also exists

between the designs created using birch bark biting, bead- and quillwork, and the images on

birch bark scroIls.

It is important to determine whether the use of different designs has changed over

time and whether shifts in meaning accompanied any change in design and shape. It is

equaIly crucial to include information as whether changes have occurred in the execution of

these designs because the techniques used might be meaningful. Yet it is frequently very

difficult to establish which designs belong to which groups, especially if first, the criterion of

time is taken into account, and second, a group of lndians has changed from using geometric

shapes to floral designs. Fortunately, this shift has been tracked by Hoffman (1888, 1889,

1891), Schoolcraft (1983 [1860]), Densmore (1910, 1913, 1941, 1974 [1928], 1979 [1929]),

and Coleman (1947).

lt should also be remembered, as Braaten and Ellingson (1992: 27) stated, that

choices were made by the individual involved in creating these images probably reflect their

cultural group. At the same time, King's (1991) arguments regarding western coIlecting biases

must be remembered. What has survived from previous centuries may only represent what

was used for special occasions and not what was worn everyday (see Braaten and Ellingson's

1992 examination of moccasins belonging to the Ojibwa and the Sioux from the late

nineteenth century).

Densmore's (1974[1928]: 183) publication, based on informants from the Vermillion

Lake, White Earth, Cass Lake, Mille Lac, Leech Lake, and Red Lake reservations in

Minnesota, the Lac Court Oreilles Reservation in Wisconsin, and the Manitou Rapids

Reserve in Canada, concludes that a developmental shift occurred in design use. Geometrie

and "line" designs were older than floral designs. She discussed in great detail the simplest

designs and how they became more complex. She (ibid.: 184) established that the geometric

Page 80

patterns found on birch bark transparencies were copied and used in woven bead bands.

However, her informants told her that "in the early times" beads were often used as edging

on leggings and moccasins, consisting of small white opaque beads usually in combination

with simple line patterns. The main geometric designs, or units, as she termed them (ibid.),

were the star and a cruciform "combined from various lines and angles". Geometric images

were often combined with images such as "a rose, a star, a heart and a butterfly" (ibid.: 185).

Densmore differentiated between patterns with angular outlines and those that had

geometric designs. She observed that patterns with angular outlines were believed to be

similar to those made by "the white race" (ibid.). Geometric designs were deemed very

suitable for woven beadwork while those patterns which had angular outlines were best

adapted for use in applied designs. She, like Speck (1937), observed that the double curve

pattern was used and that angular patterns represented leaves. Several informants told her

that the bird was often used as a dream symbol, while the star was a "favourite" pattern as a

dream symbol (1974[1928]: 185). She subsequently discussed how the double curve motif

had developed and reviewed the literature regarding its physical distribution amongst the

AIgonquian groups north and south of the St. Lawrence Seaway. She noted that three of her

informams used this motif for the fronts of moccasins and as paper-cut designs from birch

bark.

Densmore discussed floral designs and patterns in considerable depth. The "rose"

image occurred when the lines and angles of the geometric patterns were combined. The

wild rose could be considered as the representative flower of the Chippewa since it existed

"in profusion throughout the country" (ibid.: 186). She argued that one of the oldest floral

designs was the "round-bead pattern" and that the "grouping of the four dots with or

without a smaller dot in the center is a natural development of designs" (ibid.). Individuals

who wanted designs that imitated nature used these floral designs. However, her informants

assertion that the later objects that used these images had no artistic value indicated shifts ln

the development of these designs ..

Informants distinguished between the older patterns of flowers and leaves and those

that have been prevalent since 1860-1870. Densmore (ibid) postulated that this was possibly

the time period when the Chippewa came into extensive contact with the white populations.

Densmore ascertained that the earlier floral designs were deemed "truthful", while the

Page 81

r·,

subsequently "modem" Chippewa designs "lacked truth" as well as "freedom of expression"

(ibid.: 187).

Densmore also described at great length the images cut from birch bark using

scissors, which aided the women to create designs. She established that the early cut-outs

were made using a sharp fish bone; the pricks being placed close together and then the bark

eut along the lines with a knife (ibid.: 188-189). These eut-outs were principally used for

floral beadwork but could be used for any other decorative work (ibid.: 188). They included

leaves, flowers, angular patterns, and double-curve motifs. The women seldom exchanged

patterns, "but a good designer" often traced patterns for less gifted workers (ibid.:189).

Densmore discussed the applied beadwork at length, arguing that it was considered easier

and older than woven beadwork. She argued that "in early times" the bead embroidery was

placed on the broadcloth of the garment to be made. Later, the embroidery was applied on

black velvet which was subsequently applied to the garment (ibid.: 191).

She identified the three simplest patterns that she posited were probably used first:

"the jumping pattern, the zig-zag, and the ottertail pattern" (ibid.). Women used the designs

they created using the birch bark cut -outs. Densmore described at considerable length how

these designs were transferred to the fabric so that they couId be utilised. She maintained

(ibid.: 192) that the admiration of colour, as a consequence of using coloured rather than

white beads, had led the women to lose "their appreciation for graceful lines, and the art

value of their work is further impaire d, if not entirely destroyed". White beads, regardless of

whether they were opaque or opalescent, were used for backgrounds.

Woven beadwork was for Densmore the continuation of an early form of porcupine

quill work (ibid.) and she described in considerable depth how this was undertaken. She

ascertained that the person who wove the beadwork never followed a pattern while weaving

and rarely repeated a pattern (ibid. 194). She (ibid.) observed that sorne bead workers used

birch bark transparencies as suggestions, while others had "the patterns in their heads".

Densmore, (1974[1928]: 329-397) repeated Schoolcraft's (1857) observations and

comments about birch bark biting. Schoolcraft (1857: 631) had described this technique as

dental pictographs and noted that it made use of the fine inner layers of the bark of the birch

tree. Both he (ibid.) and Kohl (1860:213) described the designs used as flowers, human

beings, and fancy baskets. Yet Kohl's (1860: 413) remark about the creation of a design of a

tomahawk contradicts Schoolcraft's (1857:631) conclusion that it was a "pretty art" and an

Page 82

example of one of the first "imitative and aesthetic arts" amongst the Ojibwa. Kohl (1860:

412-414) visited a women who was a birch bark biter and observed that she used her only

remaining tooth on her upper jaw to create a wide variety of designs. He (ibid.: 413)

described her selecting a suitable piece of bark and putting it into her mouth. Kohl (ibid.)

discerned that she did not cut through the bark but only pressed it with her tooth so that

when the designs were held up to the light they resembled porcelain transparencies.

Densmore, noted (1974[1928]: 393) that Speck (1914) recorded modern examples of

this art form amongst the Montagnais and the techniques used by the women to create these

images. She determined from her informants (1974[1928]: 393) that the designs to be used

were in the mind of the pers on prior to biting the birch bark. One of her informants told her

that the piece of bark was placed between the upper and lower teeth, generally between the

eye teeth, while the bark was slightly twisted between the teeth (ibid. 393). Densmore (1974

[1928]: 394) discovered that if two or more pieces were indented simple patterns were used,

but if the designs were complex then only one piece of bark was bitten at a time. She (ibid.:

393-396) identifie d, described, and provided numerous examples of birch bark biting on

several plates in this publication. Densmore (ibid.: 396-397) established that sorne of the

designs were subsequently used as patterns for beadwork and that etching and "self­

patterns" were created on the surface of birch bark containers. Descriptions were provided

of the containers, how they were made, and how they were subsequently used.

Densmore (1941) provided additional information from her fieldwork in northern

Minnesota and southern Canada. She observed that only human teeth were used and that the

marks were made using the eye teeth and the side teeth (ibid.: 678). She asserted that it "was

not unusual for 12 thicknesses of bark to be inserted at a time" (ibid.: 678-679) but knew of

one example that had 241ayers of bark. She stated that (ibid. (1941: 679), according ta her

oldest informants, such work was undertaken only by women, in the maple sugar camp, in

the evenings once work was completed. These transparencies were done at this time of year

because the bark of the birch tree was soft and fresh. Since many of Densmore's informants

were more than 80 years old, this information dated this practice "back about a century"

(ibid.). Densmore observed that Schoolcraft and another unstated "early writer" had

mentioned birch bark bitings but that this art form had not been discussed by "traders and

explorers" (ibid.). She established that the designs included geometric, abstract, or

conventionalized natural shapes such as "roses and stars, and there were sorne animais and

Page 83

/--~

/'

Indians" arranged in what Densmore termed "running patterns" (ibid.: 679). Ber informants

told her that only older women eould do these transpareneies, since middle aged women

laeked "the mental concept and their work was not arf' (ibid.) (italies are Densmore's).

Densmore's (1941) article eonsidered another unnamed form of imagery using birch

bark. Bireh bark was eut with a sharp edge, possibly a fish bone or a knife, since points were

prieked in the bireh bark and then the bark was eut between the points to ereate a pattern.

Densmore (ibid.) deseribed the patterns as "angular" but she observed that the zig-zag

outlines ereated in this manner were used to make bead patterns on moccasin cuffs. The use

of scissors, Densmore (ibid.: 679-80) observed, made possible a new form of art using birch

bark to create "little pictures or silhouettes" but these, like the birch bark biting, were only

made for pleasure. Anything could be cut including "conventional flowers, and leaves,

changing the curves to straight lines" (ibid.: 680). She asserted (ibid.) that the Chippewa, in

northwestern Minnesota and southern Canada, argued that these designs were "interpretations

as natural forms and maybe regarded as art [.]" (italics are Densmore's). Ber informants told

her that these designs in their beadwork came after the geometric patterns.

This article supplied additional information regarding the development of the floral

beadwork. Densmore discovered from her informants that use of conventional flowers and

leaves as designs elements eould be viewed as "art" and was viewed as an interpretation of

nature. This is interesting since it appears that her informants are talking about "art" which

means that a western idea had become part of their culture. These patterns according to her

informants arose after the development of the geometric designs and before the modern

floral patterns which imitated nature. It could be eoncluded that she thought that her

informants possibly adopted, or imported the western idea of art.

Friedl (1944: 150) observed that information from her Chippewa informant, from

Lac Courtes Oreilles in Wisconsin, contradicted both Schoolcraft's (1860: 631) and

Densmore's (1929: 184-185) information that only women created these images. Friedl's

female informant was taught with her brothers and her sister, by their father, a birch bark

game which taught them to invent stories from interpreting the patterns and figures that

were used to create these birch bark transparencies (FriedI1944: 149). Friedl (ibid.: 150) also

established, from her informant, that it was easier to separate the "tissue-thin layers" of the

birch bark if it was opened in a heated room while still frozen.

Page 84

Fieber (1978: 56) argued that the origins of the images found on birch bark bitings

were obscure but the technique developed in a casual manner. He maintained that this skill

was practiced for the pleasure of creating the object. Fieber (ibid.) asserted that birch bark

biting was used to create designs for woven beadwork since the designs were adapted for use

on headbands, knee bands, or shirts. He posited that the designs and skills involved became

more developed and gained artistic value as they were handed down over generations. This is

an interesting but unfortunately unsubstantiated statement. Dewdney (1975: 15) observed

that the thinnest cork layers of the birch bark were folded several times and then bitten using

the eye teeth and stated that the resulting pattern was revealed once the folded birch bark

was opened.

Oberholtzer and Smith (1995) argued that birch bark biting was becoming a lost art

which they observed had been recorded in the writings of the early ethnographers. They

(ibid.: 311) asserted that birch bark biting as a "truly indigenous art form" was distributed

throughout the boreal forest from the Atlantic to the Plains. From the seventeenth century

onwards, the increasingly large numbers of non-native peoples settling alongside them did

not influence its development. This is an interesting daim, since Phillips (1984, 1991 and

1999), Whiteford (1991; 1997), Penny (1991), and Lanford (1984) were in the process of

establishing that the meanings and the images used in beadwork, quillwork and embroidery

were influenced by Europeans.

Oberholtzer and Smith (1995: 312), like Friedl (1944: 150), argued that men also

participated in this activity, thus contradicting Fieber's (1978: 56) assertions that only women

practiced this art as a form of social entertainment. Oberholtzer and Smith (1995: 312-314)

debated whether women were the sole users of these images because evidence existed that

women frequently used images resulting from dreams and divination practiced on bumt

animal scapulae. Both (ibid.: 314) observed that Speck (1937), Davidson (1941), Spier (1915)

and Rousseau (1956) had recorded the characteristic symmetrical formation evident in the

images in the birch bark bitings. Oberholtzer and Smith (1995: 314) maintained that these

simple patterns were the "basic features of the double-curve motif' that was very prevalent

amongst the imagery of various Algonquian groups. Both Oberholtzer and Smith (ibid.: 315-

316) conduded that the socialization that occurred when something was presented as

entertainment, essentially creating images, was an important means of encoding and recalling

information required by the Aigonquians. So, despite birch bark biting being an example of

Page 85

indigenous 'art for art's sake', it was a means of learning new skills and assessing designs, new

ideas and cultural values. Both substantiated their statements by drawing upon their own

fieldwork, as well upon information Friedl (1944) obtained from a Chippewa woman.

Coleman's (1947: xi) descriptive study was based on fieldwork which she conducted

at the White Earth, Leech Lake, Fond du Lac, Mille Lacs, Nett Lake, Grand Portage, and

Cass Lake reservations in northern Minnesota. She drew extensively upon work by a variety

of people, including Hoffman (1891), Mallory (1893), Copway (1851), Kohl (1986 [1860]),

Lyford (1943), Densmore (1929) and Warren (1984 [1885]). She identified all of the

geometric designs, which included angular and curvilinear motifs, and what she termed

realistic motifs. She considered how the different types of images were utilised in a variety of

contexts, including social, material, and religious culture, amongst the communities she

examined. Nevertheless, it is possible to posit the idea that the material and social culture are

essentially the same.

Coleman examined the role of images in the material culture of the Ojibwa and (ibid.:

45) concluded that designs on clothing and bags were floral and geometric. However, if floral

work occurred on clothing, it was possible occasionally to find the double curve. She (ibid.)

observed that, if floral and geometric designs were used for birch bark objects, they

invariably were "arranged in bands" and the designs were "simple isolated floral figures".

Geometrie and realistic designs were usually found on textiles and the designs in bands on

yarn and nettle fibre bags. She concluded that none of the images, except for those on the

nettle fibre bags, held any symbolic significance.

Coleman (ibid.: 64) made five generalisations concerning the usage of images in the

category she called "social culture". Social classification did not influence decorative designs

except for "gens-totem associations", where she argued realistic representations of certain

animaIs were employed. Floral and geometric designs were frequently used for decorating

clothes worn for public occasions, but the articles worn by the chief were usually more

elaborate. Coleman (ibid.) established that social convention determined the type of designs

used by men and women. W omen used both geometric and floral designs for their own

clothes but "more ritually defined realistic designs" on ceremonial objects (ibid.). She

observed that in "more recent times" women decorated their husband's clothing and

weapons with geometric and realistic floral designs, while men previously utilised geometric

and animal figures. She (ibid.) observed that marks on "implements for games" only enabled

Page 86

them to be identified and these marks included simple geometric figures and realistic motifs

of men and animaIs. Colour, Coleman (ibid.) established, was used only symbolically for the

body paintings of warriors, hunters, the ill, and the dead. She noted these images were simple

geometric designs that did not hold any symbolic meaning. Her fifth generalisation was that

animal figures occurred more often in decorations that she believed were part of the category

that she called "social culture" rather than material culture. She observed that few designs

belonged to the social category, which included human beings, natural phenomena, and floral

designs.

However, Coleman (ibid.: 89) concluded that designs had an important place in the

religious context of the Ojibwa and that most ceremonial objects had specifie images wruch

either symbolised sorne phrase or practice of their religion. "Representations" of the bear,

panther, lynx, thunderbird, birds, turtle, otter, fish, and serpent existed on Midewiwin

ceremonial objects and scrolls. Designs of hum an beings, especially those having received

"mystic power", were indicated by "wavy lines around the head or proceeding from the

mouth". She observed that creatures and human figures were common but that all the

designs lacked details when they were used in a religious context. Floral designs were not

present.

Coleman also considered the history of Ojibwa designs from 1830 to 1947, since

none of the objects that she studied was more than one hundred years old and little

archaeological evidence existed. She (ibid.: 91) asserted that almost all of the cultural

exchange experienced by the Ojibwa of trus region was with the whites, and so she posited

that sorne of the greatest influences on the creation of the designs used by the Ojibwa

occurred with the introduction of cloth, such as broadcloth and velveteen, wruch replaced

rudes (ibid.: 92). Clotrung styles also changed, particularly clothes such as the vest, gauntlet,

and shoulder pouch (ibid.: 93). The final reason for the changes was the introduction of

beads of different colours and sizes (ibid.: 92). From information obtained from her

informants and examining museum collections, she learned that many designs came from

patterns of embroidery or cross stitcrung provided by government schools, such as the

booklet written by Lyford (1943).

Lyford's descriptive publication drew upon published sources and museum

collections and contained a substantial quantity of information concerning the "handicrafts"

made by the Ojibwa. Her study was written to be used in schools where these handicrafts

Page 87

had begun to be revived. Coleman (ibid.: 94) ascertained that booklets such as Lyford's were

used for over two generations and that the women she spoke to, between the ages of fifty

and sixty years old, only used the designs learnt from their white teachers at school. Later,

pattern books on lace introduced by Episcopalian missions in northern Minnesota became

an additional source of designs (ibid.: 95), which the native women used to crotchet.

Coleman (ibid.) argued that the imagery from the 1870-1930 period (i) was floral, (ii) lacked

cohesion and unit y, and was (iii) elaborate, (iv) large and (v) showy. Each of these

characteristics was examined in detail and Coleman observed that they existed on various

items of clothing, wood, and birch bark. Several geometric motifs (ibid.: 101-104) exhibited

design traits from other native groups, such as the Western Sioux, Turtle Mountain Cree, and

Plains Ojibwa, including the Western Sioux coyote track and forked designs. One seventy­

three year old informant in 1943 told her that the Ojibwa were aware that they tended to

assimilate designs they deemed "worthwhile" from other groups (ibid.: 101-104). Coleman

indicated that she was not entirely convinced that ail Ojibwa designs had strong European

influences but she admitted that this could be a false presumption.

She claimed that, although floral designs were present on objects dating to the 1870-

1930 period, they were not as common as from 1830 to 1870. She established from her

oldest informants that images belonging to the 1830-1870 period were not created using

patterns, but they were always weil planned (ibid.: 107). The images from this period were,

according to Coleman, classified as (i) geometric, floral and zoic, (ii) cohesive, (iii) simplistic,

(iv) smail, and (v) having "soft" colouring. Once again she examined each feature with regard

to objects in museums. She asserted that these were aboriginal Ojibwa designs or, "at least

had no evidence of outside influences" (ibid.: 111). Coleman's final and third period was

1940 to 1943, which she argued witnessed a considerable change in the attitude of the

Ojibwa regarding their designs (ibid.: 113-117). There was, she believed a lack of interest

amongst the youngest members of the Ojibwa populations in learning the craftsmanship of

their parents and grandparents. This was a consequence of commercialisation, the lack of

interest in preserving old designs, as weil as the education of the younger people in

government schools.

King (1991), an art historian, highlighted the existence of two biases that prejudiced

western knowledge concerning sub-arctic clothing. Although Densmore (1979 [1929], 1974

[1938]) and Hilger (1992 [1951]) provided detailed descriptions of men's and women's

Page 88

clothing amongst the ChippewajOjibwa, they only described their summer clothes. Firstly,

very little was known about winter clothing and secondly most published collections

included a large percentage of highly embelli shed and unused summer garments. King

postulated that information amassed by different researchers while working in this

geographic region exacerbated this problem. He used the example of twined bags and

referred to Phillips (1989)'s and Whiteford (1977)'s argument that it was very difficult to

determine which tribe had produced each object. However, he (1991: 91) argued that it was

possible to distinguish the 'arts' of the Anishinaabe peoples of the Great Lakes from those of

the Athapaskan peoples to the northwest and the Algonquian peoples to the East. He (ibid.)

postulated that art of any type was an important ingredient in maintaining the equilibrium

between early peoples and the sacred world, and he asserted that the motifs used in the

decoration of these art objects carried sacred and mystical powers. Yet, an additional reason

for the lack of win ter clothing and unembellished summer clothing could that these items of

clothes were used to the point that they wore out or they were not perceived as important to

collect.

Brasser's (1999) article provided sorne useful pieces of information. He (ibid.: 49)

argued that highly elongated triangles were characteristic elements in early paintings of the

Great Lakes region. He found that on indigenous shirts in a collection from New France

long narrow triangles were often combined with shapes that were reminiscent of feathers,

forked bird tails, and arrows. He also observed images of birds worked in quills attached to

the bottoms of the bags in this collection.

Lanford, an art historian, (1984) examined several Winnebago bandolier bags also

called pony bags, shoulder bags or friendship bags. She used the term band olier bags since it

suggested that the most common use of these bags was ornamental. She contended that,

although this bag had several styles in North America, within the Great Lakes region it had a

decorative rather than a functional role. The bag's shape had evolved either from European

shot pouches or from pouches used by native foot travellers to transport provisions (ibid.:

30). Prior to the nineteenth century, the bases of these bags were dyed black-brown or dark

smoked tan and decorated with quillwork in the Great Lakes area and eastwards. Bags made

du ring the nineteenth century were covered in the floral beadwork typical of the Great Lakes

region. Lanford (ibid.) asserted that, by the mid-nineteenth century, these bags had become

part of the assemblage worn with one's best clothes. She arrived at this conclusion noting

Page 89

that bags belonging to this period had fragile cloth foundations, and that some bags did not

have a pocket, or had un-backed beadwork. Lanford (ibid.: 36-37) suggested that several

features of construction and decorative elements made bandolier bags easy to identify, and

that a predominance of certain colours existed in their woven beadwork. She (ibid.: 37)

stated that there was a distinct lack of symmetry in Winnebago bags in contrast to bags made

by Algonquian groups. Lanford concluded that sin ce relatively few differences were evident

in bandolier bags, it was possible to argue that each person who made these bags had a

distinct conception of the decoration before its creation. Whiteford (1986), like Lanford

(1984), examined four bandolier bags which could be dated since documentation exists as to

their provenience. He (1986: 35-36) discussed their size, proportions, materials, and the

techniques used to create them, as well as their structure, decoration, and the beaded designs.

Whiteford (ibid.: 38) compared these four bags with bandolier bags from other museum

collections to try to learn who might have made them. He (ibid.: 43) concluded that the four

bags represented a stage in the evolution from the small buckskin pouches of the early

eighteenth century to the large ornate bandolier bags that emerged between 1850 and 1860.

He argued that, by the 1870s, these bags had achieved the basic styles to be used for decades

in the Great Lakes region.

Lanford (1986) discussed the development of woven beadwork in the Great Lakes

area. She posited that the Chippewa, Potawatomi, Menominee, and Winnebago of the Upper

and Western Great Lakes areas, provided greater quantities of woven beadwork from the

nineteenth century onwards. During the late eighteenth and early nineteenth centuries,

geometric and straight-line motifs predominated. Lanford (ibid.: 62) observed that these

motifs could be found sewn and woven in the porcupine quillwork of this region dating back

at least to the early historic period. She also (ibid.: 63) observed that relations could be

posited between the geometric motifs of woven beadwork and those found on stone, pottery

and shell objects dating to the prehistoric period. Geometrie and zoomorphic forms and

human design elements, were incorporated, carved or incised on antler, wood, bulrush mats,

scraped birch bark containers, and porcupine quillwork from the historic period (ibid.). Both

Lanford (ibid.) and Lessard (1986) postulated that woven porcupine quillwork and fibre

weaving techniques contributed to the evolution of woven beadwork. Appliqué floral

beadwork replaced this woven beadwork in the Great Lakes area during the late nineteenth

century. Lanford (1984: 64) argued that the speed of this transformation demonstrated that

Page 90

styles could shift according to fashion in a very short period of time. He (ibid.: 64-65)

suggested that additional research was required to determine precisely when woven quillwork

declined and to pinpoint possible European influence upon the development of woven quill­

and woven beadwork motifs in the Great Lakes region. She hesitated to ascribe direct

assimilation of non-Indian motifs in the Great Lakes area, but stated that it was unlikely that

highly complex examples would suddenly appear.

Phillips (1984: 418) disagreed with Lanford (1984) regarding all the motifs and their

special co-ordinates. In her short paper, Phillips, attempted to understand the manner in

which these images carried and conveyed information. She developed the idea that an image

is polysemie and its meaning can be manipulated. Phillips (ibid.: 418) observed that the

geometric motifs, deemed to be the oldest types of images used by the peoples in the Great

Lakes region, existed on the fringes at the bottom of many transitional loom-woven bag

types, while large curvilinear floral designs in appliquéd beadwork were found elsewhere on

the bag. She posited that this mixture of images may have occurred because of the threat

posed to indigenous traditional religious institutions as a result of increased missionary

activities. In addition, the intensified White presence during the nineteenth century might

have produced this intentional marketing of visual imagery. She argued that the geometric

imagery probably was associated with the Midewiwin and other indigenous religious practices

that appear to have been in decline during the nineteenth century. Phillips (ibid.: 411) posited

the idea that it was doubtful that similar complex designs on clothing could be equally

meaningful.

Phillips (ibid.) stated that the 'geometric designs' on the birch bark scrolls belonging

to the Midewiwin had symbolic meaning. One strategy to determine the meaning of these

designs, Phillips posited, involved examining a group of quilled and woven designs from an

early period that might have been sources for the later woven beadwork. Hence she (ibid.:

416) analysed several designs on items from the Jasper Grant collection to demonstrate that

visual designs varied in time and space. She utilised this collection, despite difficulties in

specifically attributing various items because it consisted of images that were "iconographie

representations of cosmic spirits" (ibid.: 412) that were used "concurrently" during the first

decade of the nineteenth century, just before intensive White settlement which had

"accelerated disruptions of traditional religious systems" (ibid.). She claimed that people who

Page 91

"were and had been in close contact with each other" made these objects, which, therefore,

probably were shaped by a "shared and widespread" "world-view" (ibid.).

One of the several examples she considered was a group of early nineteenth century

finger woven bags with images that she stated alternated between "figurative design and

abstraction" (ibid.: 414). Phillips (ibid.: 415) claimed that the woven bags in this collection

"display a fascinating range of imagery". She (ibid.: 416) argued that shapes, or design

elements, as she termed them, such as "interlaced wavy lines," zig-zags, and wavy lines, were

"the common vocabulary" utilised to infuse "spirit power in the Great Lakes art". She

believed that these images dated from the early contact period to the first part of the

twentieth century. Unfortunately, she did not provide any pictures of the images. She

referred to them as "Thunderbird", "Underwater Panther", various Images of

"Misshipeshu", and a variety of zigzags, which she appeared to interpret as being "the chief

manifestations of the powers of the sky spirits, lightening and thunder" (ibid.). Although this

is interesting, she did not indicate her source of information for naming the shapes that she

identifies, or the ethnographie sources she drew on to provide the meaning of these images.

Later in her discussion, she (ibid.: 417) utilised Hoffman's descriptions of the imagery

used on the Midé costumes at the end of the nineteenth century. She concluded that, since

zigzag and wavy lines existed on their ritual clothes, their meaning had not changed despite

the changes in techniques and materials over time. These design elements "still expressed

symbolically the cosmic powers whose protection was reaffirmed" during the Midé

ceremonies. Other images, such as "crosses", wavy lines, human figures, and ho ur glass

images, were examined and meanings attributed. Crosses were interpreted as the four

cardinal winds that "blow across the surface of the earth from the four cardinal directions".

They signified the "transition between the upper and the underworlds" (ibid.). She argued

that images of Thunderbirds or human figures often appeared as geometric figures on the

woven bags from the Grant Collection (ibid.). Phillips suggested that bandolier bags dating

from the late nineteenth century could be "mapping the cosmos and the spirit forces", but

she realised that not all of the bandolier bags had beadwork and imagery similar to those of

the Grant collection. Many loomed bags dated around 1900 had large geometric floral

motifs.

Phillips argued that these abstract designs ought to be interpreted as having symbolic

rather than decorative meanings. She (ibid.: 418) postulated that the obscure symbolism and

Page 92

hidden meanings were perhaps what was left of a religious system forced to go underground.

The abstract images in the art of the Great Lakes region created during the nineteenth

century were a means of expressing the concept of severallevels of reality and the visionary

perception of this reality. This could explain, she believed, why the older geometric images

were used on the bottom of bags that had floral imagery. She concluded that religious and

spiritual concepts were expressed by means of "abstract motifs". She claimed that early

writers, who remain unnamed, asserted that geometric images were connected to personal

dream experiences. The images, she maintained, were created in a "concrete visual form" to

keep the "blessing of the guardian spirit" but the vision itself must remain private to retain

its power (ibid.). Phillips asserted that these images should be called "ambiguous" since they

were capable of altering their physical appearance, which she maintained was a "basic belief

of Great Lakes lndian belief' (ibid. 419). She concluded that during the nineteenth century

the different styles of abstract imagery used reflected not just numerous levels of reality, but

what she called "particular visionary perceptions of these realities" (ibid.).

Seven years after Phillips (1984), Whiteford (1991) postulated that it was possible to

distingwsh Anishinaabe material culture from that of other Algonquian groups and also from

that of Athapaskan tribes to the northwest of the Great Lakes. 'Anishinaabe', he argued

(ibid.: 74), was an Algonquian name used to describe the Ojibwa/Chippewa who arrived with

other Algonquian groups around the Great Lakes from the East during the early 1800s. He

wished to identify the key traits of the Anishinaabe arts and considered weaving, beadwork

and work on birch bark. Much of this information was already discussed at great length by

Lanford (1986), Lessard (1986), and Hoffman (1891). Whiteford maintained that

Anishinaabe designs were bolder in conception and highly distinctive in the use of strongly

beaded backgrounds with strong floral patterns compared to those of the Cree and the

Athapaskans (Dene). He (ibid.: 76-77) claimed that the sophisticated weaving techniques

used by the Anishinaabe combined with the range of designs for marked twined bags unique

to this cultural group. Furthermore, colour splitting designs and the 'protruding hairs' feature

found on the edges of the stems of Cree and Athapaskan floral patterns were relatively rare

in Anishinaabe beadwork (ibid.: 80). Whiteford drew heavily on Blessing (1963) and

Hoffman (1891). His aim of comparing the arts of one group with those of their

geographical neighbours was a valid attempt to deterrnine which motifs and styles were

executed on which items, and the specialties of each group.

Page 93

Duncan (1991), an art historian, considered two different types of bags: panel bags

and octopus bags. She argued that, despite the paucity of documentation for both types,

regional and tribal preferences were expressed in their form and decoration. Duncan

maintained that pre-nineteenth century roots existed for both types of bags, but she did not

provide any evidence to substantiate her claim. She examined each bag in its classic central

sub-arctic version, and described its geographic and diachronie distributions along with its

variants in the Great Lakes, Plains, Plateau, and Northwest Coast areas (ibid.: 57-65).

Duncan utilised a taxonomie approach to place the bags in a tree-like structure to indicate

the probable patterns of historical diffusion and stylistic development. The oldest

ornamented bags, Duncan (ibid. 57) argued, probably were sacred in contrast to the most

recent ornamental bags. She reached this conclusion because the more recent bags appear in

photographs taken at the turn of the nineteenth century and were "proudly displayed on

dress occasions" (ibid) as symbols of family pride, tribal affiliation, and personal prestige.

Penny (1991) would have disagreed with Whiteford's (1997) perspective concerning

these floral motifs. Penny (1991: 53) asserted that American lndian art history was concerned

with historical change and that this history should record and identify the events and

processes contributing to that change or innovation. He (ibid.) examined the historical

factors leading to innovations in dress style that accommodated the advent of the floral style.

Penny (1991) observed that Speck (1937), Barbeau (1930), and Brasser (an unpublished paper

given in 1987 that Penny used) had also considered the origins of this floral ornamentation.

Speck (1937), an ethnographer, argued that floral motifs grew out of indigenous double

curve designs prominent in the Penobscot decorative arts of the eastern seaboard, ex tant

prior to European contact, as prophylactic depictions of medical plants. Barbeau (1930), a

folklorist, maintained that floral designs were taught in mission schools as early as the

seventeenth century, and argued that their appearance on native clothing originated in this

formaI training. Brasser (1987) concluded, from a thorough lite rature survey, that litde

evidence existed for plant or flower motifs in the prehistoric art of the Great Lakes until

1800.

Penny (1991: 59) claimed that clothing and its decorative elements could be read as a

semantic system functioning on different levels. He argued that for the indigenous groups of

the Great Lakes, these images denoted gender, sexual attractiveness, social status, identity

and occasionally certain facets of information from their biography, particularly "in the

Page 94

realms of warfare and spiritual power". Floral ornamentation functioned as a sign on two

levels: mimetic and semiosic (ibid.). Penny defined the term "mimetic plane", or level, as the

representational function of sign; hence "its relationship to the thing that it represents"

(ibid.). This meant that, if sorne one asked their informants about floral ornamentation, they

would be told that the images represented flowers. He (ibid.: 60) defined the "semiosic" level

as referring to the social method by which meanings are created and bartered. Since neither

was the same, this process enabled (ibid.: 60) the meanings of signs to be debated, modified,

and intensified. He argued that the manner in which these signs were subject to historical

analysis enabled new meanings to occur that could be understood as historical incidents.

Penny (ibid.: 61) posited that this categorization of the world led to the existence of polarity

with respect to Indian attitudes regarding clothing, since a category developed that polarized

into White European style clothing versus Indian style clothing. Clothing used by white

Europeans signified civilization, while Indian clothing, such as moccasins, leggings, and

garments with 'exotic patterning', implied a state of savagery to Whites. Indian dress, from

the indigenous perspective (ibid.) signified cultural integrity and resistance to the White

domination. The term 'dress', Penny (ibid.: 54) argued, meant a special category of clothing

reserved for formaI social occasions. White people, as defined by Penny (ibid.: 62), were the

missionaries who yearned to convert the Indians, the government authorities who desired to

control them, and European neighbours who were frightened of them.

Three groups were pre-eminently associated with the making of floral style

embroidery from 1800 onwards: the Huron of Lorette, the Ottawa and the Chippewa of

Michigan, and the Métis (ibid.61-68). Missionary endeavours of the Catholic Church

influenced heavily ail three communities (ibid.). This suggests that the techniques and the

imagery of floral embroidery were taught in the schools they attended. Penny (ibid.: 68)

argued that in each case potential existed for disjunction between White and indigenous

interpretations of floral style decoration as a sign of social identity. The Huron, Ottawa,

Chippewa, and Métis artists used this instruction to manufacture handicrafts for sale as an

expression of ethnic identity to 'Christian and civilized' Indians, while the White observers

interpreted the embroidery as signs of Indianness, which translated into successful marketing

of these products.

After the mid-nineteenth century, Penny asserted (ibid.: 69), the native North

Americans of the Great Lakes used two fashion systems: White and Indian. Each system

Page 95

operated concurrendy and congruendy. Penny did not indicate what would have been worn

by someone who utilised the lndian fashion system. However, he argued that, during the

mid-nineteenth century in Wisconsin and Michigan everyday clothing used by the lndian

community was "largely indistinguishable from that of neighbouring (low income)" white

communities (ibid.: 68). The White dominant culture, he argued, used the White fashion

system as an expression of accommodation. Penny (ibid. 71) maintained that the use of

'formaI dress' by the lndian population grew to be one of the most important "symbols of

ethnic solidarity and persistence of tradition" during the reservation era. Essentially a native

North American artist used specific styles for Penny refers to as historically and socially

incidental motives. Hence such innovations in culture often become traditions (ibid.: 72).

Penny's (1991) article established that floral ornamentation existed on clothes for reasons

that went deeper than aesthetics, and that two sets of meanings were held for different

audiences, since these native artists were masters in manipulating not just the motifs, but also

their meanings.

Phillips (1991) observed that in art history and anthropology, the two disciplines that

usually examine native art, style and meaning were usually considered as separa te. Phillips

(1991) like Penny (1991) considered factors that led to the development of the floral motifs.

She argued that, since European cultures defined the floral motifs on which lndian ones

were modelled as secular, it was presumed that their significance remained the same once

they were adopted by native Americans (ibid.: 97). Phillips argued that, if the new floral

decoration was non-referential, the changes that occurred constituted a de-sacrilisation of

important categories holding great cognitive significance for hunting peoples. Both Penny

(1991) and Phillips (1991) concluded that the indigenous people of the Great Lakes

understood and operated within two conceptual frameworks during the nineteenth century.

Phillips (ibid.: 98) believed that the relationship of style to meaning in American art objects

must take account of this relationship. Phillips (1999) expanded and developed the ide a that

meanings and images could be manipulated in her analysis of the objects called 'souvenirs' of

northwestern North America.

Whiteford (1997) also considered the origins of floral beadwork in the Great Lakes

region. He (ibid.: 75) identified the different plants used in designs as part of the increasing

use of floral motifs on objects dating from the 1860s onwards. Like Phillips (1999), he was

aware of Penny's (1991) article which considered the various factors responsible for the

Page 96

development of the floral motifs. Whiteford (1997: 75) argued that evidence existed for the

origins of the plant motifs prior to European contact and for the subsequent development of

these motifs after European contact. This evidence (ibid.: 75-76) included: (a) simple patterns

of semi-florallinear motifs in native porcupine quiilwork that dated as early as 1800, (b) the

light coloured angular and curvilinear designs that existed prior to the eighteenth century on

the surface of birch bark containers and (c) the images created as a consequence of birch

bark biting. European influence, Whiteford (ibid.: 77) maintained, commenced with the

French fur traders and missionaries farther east during the seventeenth century. Direct

influences occurred when the Catholic and the Anglican mission schools were established in

the Lake Winnipeg-Red River region in 1824 (ibid.). InformaI and formaI influences were

derived from the European women on the frontier, the wives of officers of the government

and the Hudson's Bay Company, and Métis women. Whiteford (ibid.: 79) asserted that the

floral designs resonated with the awe of the natural environment, represented the

supernatural power of nature, expressed Indian solidarity and at the same time appealed to

the tas tes of those non-Indians who purchased the objects on which they existed.

(b) Images as a means of communication.

The images and the designs produced by the Algonquian-speaking peoples have

changed and been influenced by numerous groups of people. lt is possible to conclude that,

although ail images are a form of communication, not ail images changed radicaily in

appearance or experienced changes in meaning. Copway (1851: 132) observed that over two

hundred figures, or images, were generaily used for ail types of "correspondence" or

communication. He asserted that the Ojibwa used these images, what he termed "picture

representations," until they adopted the use of "European manners" (ibid.: 134). What,

precisely, he meant by this remains unclear. He stated that the Ojibwa only communicated

with other indigenous groups using messengers. Otherwise, the Ojibwa became guarded as

far as their own religion was concerned because "the white s, ridiculed it" (ibid.).

Coleman (1947: 63) observed that the totem of the individual buried was placed

upon grave markers. Landes (1968: 172) noticed that birch bark was used to record "mystic

songs" by means of pictographs. She stated that the patterns used in pictographs were

familiar to ail Ojibwa hunters, trappers and traveilers who used them to indicate boundaries,

trails, notices of different sorts, and grave markers.

Page 97

Dewdney (1975b: 12), based on information from informants, asserted that unnamed

Ojibwa communities from northem Minnesota created census records on wood using the

"totemic signs" to identify the clan of the person buried as weil as taily marks to indicate the

number in "his family". He (ibid.: 13) observed that "pictographic images" were used on

grave markers and headboards of bark canoes. Ali these examples fail into his category of

secular writing. Indeed, he earlier (1970c: 22) had identified three types of writing: secular,

tutorial, and visionary. Secular writing was defined as trail messages left on birch bark at a

portage to give useful information to those who were foilowing. Dewdney assented that

using an individual's totem was similar to a "signature". It was recognised over a large

geographical area and existed on each person's personal objects (ibid.: 22-23). Tutorial

writing enabled an initiate to le am by rote the traditions and practices of the Midewiwin from

his or her teacher, the Midé. These images occurred on the Midé scroils and the images, as

mnemonics and were "read" by those who had been initiated into the Midewiwin Society.

The third type of writing, defined as a dream symbol, was obtained from a person's dream

guardian during their puberty dream. This dream and its meaning could not be disclosed to

anyone without risk of losing the guardian spirit, but occasionaily the dream symbol was used

as an "identity mark" provided that it was disguised in some way so that its meaning could

not be detected (ibid.).

Those images that changed most radicaily in their physical appearance may have been

used in a social rather than a religious context. It appears that those images that did not

experience radical changes belonged to the religious life of the Algonquian speaking peoples.

Here, images behaved as mnemonics in a variety of scenarios, since they were designed to

convey information over several generations to specific individuals cailed Midé. A

considerable quantity of information exists concerning the images recorded on birch bark

scroils. A large number of people have been interested in these images for over a hundred

and fifty years (Schoolcraft (1851); Kohl (1986 [1860]); Warren (1984 [1885]); Hoffman

(1883; 1888, 1889, and 1891); Mailory (1894); Densmore (1910, 1913, 1974 [1928], and 1979

[1929]); Skinner (1915); Reagan (1927, 1928, and 1935); Cadzow (1926); Kinietz (1940);

Voegelin (1942); Coleman (1947); Blessing (1963); Dewdney (1970c and 1975b); Vennum

(1978); Nelson (1983); Fulford (1988, 1989, and 1990), and Vastokas (1984, 1986-1987, and

1996).

Page 98

Schoolcraft was an American explorer and topographer from Watervliet, New York

State. He served as a topographer for an expedition led by the American General Cass to

explore the Lake Superior region in 1820. This experience coupled with ethnographie

fieldwork on the Upper Mississippi and in the Lake Superior region led to his appointment

as the Indian Agent for the peoples of the Lake Superior region from 1822 to 1836. From

1836 to 1841 he was Superintendent of Indian affairs in Michigan, where he settled in Sault

Ste. Marie after his marriage to an Ojibwa. He devoted much of his life to the study of

American Indians.

Schoolcraft (1851: 351) observed that the Ojibwa distinguished between

"kekeeowin" and "kekeewin". The term "kekeewin" was defined as a system of marks that

represented Ojibwa clan symbols and other images that were readily recognised and

understood by everyone in the tribe, and were used to leave messages, and on grave markers

to provide information about the person who had died. However, if these images were used

on rocks, Schoolcraft stated that they were cailed "mussinabiks". He (ibid.) defined the term

"kekeeowin" as a system of symbolic notation "for the songs and incantations of the Indian

meda [Midé] and priests, making it appear to the memory for the presentation of language

and musical notes". Schoolcraft described (ibid.) the images on the birch bark scroils as

"picture writing". He (ibid.: 361) observed that the members of the Midewiwin used these

pictographs as mnemonics to remember the words of their chants. These chants were

performed inside the sacred midewegan, 'spirit lodge', during curing rites and initiations.

Schoolcraft recorded these chants and discerned they functioned in a similar manner to

prayers or hymns. He was impressed by the Midé's ability to recoilect both the words and the

music, based upon the images on the birch bark scroils and concluded that each chant had a

key symbol wruch acted as a trigger to a body of information residing in the memory of the

shaman. Unfortunately, he did not identify the key symbols on the birch bark scroils and was

unable to elicit support for future research from the Smithsonian Institution.

Kohl (1985[1860]: xiii) provided sorne information regarding the images on the birch

bark scroils belonging to the Ojibwa Indians who lived on the south shore of Lake Superior

in northern Wisconsin. A German geographer, ethnologist, and travel writer, Kohl spent

four months with them as they adjusted to life on four of the new reservations in northern

Wisconsin. He (ibid.: 145) noticed that the "northern savages" used birch bark sheets of

varying sizes on wruch to write. He described the Ojibwa practice of creating pictographie

Page 99

dream records and suggested that many song scroils had iconographic onglns ln the

Algonquian dream or vision quest experience. He (ibid.: 143-144) observed that, at

Chequamegon, the Ojibwa also le ft records of their dreams on trees, rocks, birch bark,

blankets, and clothes. He (ibid.: 145) coilected, purchased, and copied many of these objects,

sorne which were possibly birch bark scroils. Kohl (ibid. 143-166) established through his

interpreters that the scroils were related to the Midewiwin Society and depicted different

totems as weil as the migration movement of the Ojibwa from the Atlantic to Lake

Superior.5 Unfortunately, his efforts at transcribing the images were not very comprehensive;

perhaps a consequence ofhis attitudes towards the Indians. He wrote that he (ibid.: 165-166)

became annoyed with "these savages" after his attempts to buy scroils from two of them

failed; foilowing an attempt to bribe them with pouches of sugar or tobacco.

Warren (1984[1885]) worked among the Ojibwa of Minnesota and Wisconsin and

was of mixed Ojibwa and Euro-American heritage. He wrote very little about these images

but he stated (ibid.: 67) that the songs and rites of the Midewiwin were transcribed using

"hieroglyphics". The only ex ample of these "hieroglyphics" that he (ibid.: 89-90) discussed

was of sorne images inscribed on a copper disc which recorded the number of generations of

a family of the Loon totem who had lived in the La Pointe area. The indentations indicated

that eight generations had lived ln this region. Warren (ibid.: 26-27) argued in the

introduction of his volume that he intended to conduct a detailed examination and

interpretation of the images used to record the religious idioms and the songs sung, but

unfortunately he died from a haemorrhage before beginning the projected book on the

Midewiwin (Fletcher Williams 1984: 18). Although a practising Episcopalian, he believed that

the Midewiwin was central to the Ojibwa world view.

Three decades after Schoolcraft (1856), the Smithsonian Institution provided funding

for further research into Ojibwa pictures. Hoffman, a Colonel in the US Army, carried out

fieldwork for three months amongst the Ojibwa of northern Minnesota under the

supervision of Garrick Mailory, who remained in Washington D. C. (Hoffman 1891: 156).

Mailory (1893: 202) and Hoffman (1891: 156) were concerned with complex sequences of

symbols and rejected Schoolcraft's (1851: 351) idea that any of these symbols could be

5 The debate regarding the origins of the Midewiwin, the Grand Medicine Society, will not be considered because it is not pertinent to the discussion. However, it has been discussed in sorne depth by others including Angel (2002), Bhar (1991), Balikci (1956), Dailey (1958), Grirn (1983), Hallowell (1936), Hickerson (1962, 1963, and 1970), Hoffrnan (1891), Landes (1968), Meyer (1990), Vecsey (1983, 1984); Rogers (1973).

Page 100

deciphered using a key. They rejected his distinction between "kekeeowin" and "kekeewin",

believing that he exaggerated many of his observations. Hoffman's (1883) paper focused

principaily on comparing "Eskimo" pictographs with those created by other indigenous

groups. He argued that the Ojibwa used birch bark for mnemonic records and personal

exploits (ibid.: 128) and that the style of drawing used by the Alaskans was "very similar to

that of the Ojibwa" (ibid.: 135). He (ibid.: 135) examined the images found on the stem of a

pipe presumably made by the Ojibwa, but did not distinguish between the images that could

be used and read by ail the members of the Ojibwa community and those images that only

the Midé could use. He (ibid.: 139) claimed that the method of recording transactions and

events on birch bark scroils was "almost, if not entirely discontinued". He therefore posited

that it was crucial to coilect as many examples as were stiil being used, in order to increase

the quantity of information concerning these images. Nevertheless, he concluded that the

Eskimo images were "greatly superior" to those made by any other group, especiaily with

respect to the "delineation of natural objects, especiaily animate forms" (ibid.: 146). He

placed the Ojibwa bark records second among ail the indigenous peoples' imagery. He also

stated that only through a thorough familiarity with Ojibwa "ceremonies and mythology"

could any understanding of these mnemonic images be obtained (ibid.).

Much of the information in Hoffman's (1888) article was republished almost

verbatim in his (1891) report for the Bureau of Ethnology at the Smithsonian Institution.

Yet, this 1888 article discusses two other birch bark scroils that were not described in the

1891 report. Both scroils came from Red Lake, Minnesota, and, although he did not ob tain

assistance in transcribing them, he made a copy without the knowledge of the Midé to whom

the scroils belonged. He (1888: 216) observed that the characters on both scroils were

identical and that one appeared to be copied from the other. He also (ibid.: 217) claimed to

have seen an ancient cosmological chart that apparently had never been shown to any white

man or to any non-Midé Indians. Hoffman described this map, its physical condition, and its

images but unfortunately did not provide any analysis of the images themselves.

Hoffman (1889: 215) discussed the meanings of the images, the mnemonics, on a

Midé birch bark seroil that had not been shown to a "white man" before in Red Lake in

northern Minnesota. Whether this is the same bireh bark seroil diseussed in his previous

article is diffieult to determine. Apparently, this seroil was not shown to an Indian unless he

had paid the required fees, and then only once he had undergone the necessary training and

Page 101

instruction to be initiated into the Midewiwin Society (ibid.). This seroll belonged to

Skwekom' -ik who had received it from his father-in-Iaw who in turn had received it from

Ba-da'-san the Grand Shaman and chief of the Winnibigo'shish Ojibwa. Ba-da'-san himself

had obtained this scroll, which was a copy of the original, from the Grande Midé at La Point

where the Midewiwin was held annually and ceremonies were conducted strictly in

accordance with ancient and traditional rites. Hoffman (ibid.: 216-217) subsequently

examined this scroll and a few others. He realised that, if a shaman was only a second degree

Midé, he could not actively participate in ceremonies that went beyond Midé of the second

degree. It also meant that he was not privy to information known about these same images

by a Midé who was also higher than second degree. Hoffman (ibid.: 218) discussed at length

the differences in the colours used in the sacred poles used by the Midewiwin based upon

information from Sicos 'ige, a second degree Midé. The remainder of this article considered

the facial decoration used by initiates into different degrees, the ceremonial procedures, the

types of medicine bags used by the Midé of different degrees, what occurred at a Ghost

Lodge, and finally an event where a Midé treated and brought back from the Spirit Land a

dead young man.

Hoffman's (1891) monograph discussed the collection of twenty-four song scrolls

used in religious ceremonies of the Midewiwin at the White Earth Reservation in Minnesota

during the late 1880s. He published reproductions of twenty of these scrolls along with

several other examples of images. Hoffman considered the various types of shamans existing

among the Ojibwa (ibid: 165) and stated that the Midé and Jiissakid were the most highly

regarded individuals in the community for their curing and medicinal skills. Both skills were

achieved either by visions alone or by visions and purchase. Midé priests (ibid.: 165)

customarily preserved birch bark records with incised lines that pictorially represented the

ground plan of the ceremony and the number of degrees to which the owner of the scroll

was entitled when he conducted a ceremony. Severallevels of knowledge and understanding

existed because of the hierarchy within the Midewiwin Society, and therefore some Midé

were more knowledgeable and powerful than others. This, in turn, affected the meaning of

the images depicted on birch bark scrolls. Hoffman (ibid.) established than these records

were sacred and could not be exposed to public view; they could be examined only after a

candidate had paid his initiation fee, fasted, and provided offerings of tobacco. These

pictographs were used by the Midé as mnemonic devices to remember the words of their

Page 102

chants, for the instruction of new members, to record oral traditions, and to correct

ceremonial procedures (ibid.: 191-193). Hoffman (ibid.: 267) established that the seroils

eould be read from left to right, right to left, and sometimes some seroils used both styles.

He observed that the Ojibwa used images to express ideas by shamans, hunters, and the

traveilers. However, a reeognised system of using these images that enabled one person to

completely comprehend the work of another did not exist. Essentiaily, an image acted as a

key symbol and behaved as a trigger of information, a mnemonic that resided in each

person's memory. He observed (ibid.: 289) that the songs sung by the shamans were based

upon the characters recorded on the seroils and in many cases words or phrases were not

recited. Unfortunately, he did not provide any analysis concerning the images themselves.

Densmore's (1910) publication discussed the song scroils coilected during 1907, 1908

and 1909 from the Chippewa lndians of White Earth, Leech Lake, and Red Lake

reservations in Minnesota, and from a Chippewa who lived in the Bois Fort reservation in

Minnesota, and from a Chippewa who lived in the Bois Fort reservation in Minnesota. She

discovered that the melody was more important than the words and that "in a succession of

renditions of a song" the words were often only used once (ibid.: 2). She determined that the

ide a of the song was firmly assoeiated with the melody and not with the words (ibid.).

Densmore observed that it was "permissible and customary" to compose new words for old

tunes, but she noticed that the words were "always similar in general charaeter" to the words

previously used. She concluded that although changes could occur and be introduced, the

identity of the songs remained unaffeeted. Densmore asserted that these tendencies did not

apply to songs used by the Midé.

Densmore (ibid: 8-11) considered the structure of the songs sung and established

that certain songs were harmonie and others were melodic in structure. The results she

presented in tables. She observed that the songs used by the Midé required a "special skill" to

translate (ibid.: 8) and that different songs were used when singing the Midewiwin (ibid.: 11-

12). She (ibid.: 13-14) also considered the beliefs of the Midewiwin, their goals, and aims, and

argued that their songs "represented the musical expressions of religious ideas" (ibid.: 14).

This meant that she often found translating the words very difficult. She established that a

Midé song was not considered fini shed unless it ended with the syilables "ho ho ho ho"

(ibid.: 15).

Page 103

Densmore (ibid.) established that the beliefs and the teachings of the Midewiwin

were the words of the songs, which often functioned as texts for discourses (ibid.). The

songs were only taught to those who paid for the privilege of learning them and the words

were recorded in mnemonics on birch bark. Densmore (ibid.: 16-17) ascertained that

members of the Midewiwin universally understood the images, or mnemonics, of the Midé.

However, she noted that these mnemonics could be used in other circumstances. She (ibid.:

20) also considered the role of the Jiissalcid (known as doctors or jugglers - a different type of

shaman amongst the Algonquians) who were different from the Midé. Densmore recorded a

wide range of songs, the medicinal songs which the Midé, rather than the Jiissalcid used. She

established that, for a charm to work properly, the correct song and the correct medicine

must be used. Therefore the person who bought the medicine from the Midé also bought the

song.

Densmore's (1913) publication continued from her earlier (1910) study of Ojibwa

music but only dealt with mate rial from the White Earth reservation in Minnesota. She relied

upon numerous informants for translations, including Mrs. Mary Warren, Rev. C. Beaulieu,

and Rev J. A. Gilfillian. Both ministers helped with the analysis and translation of the

Chippewa words. Rev. C. Beaulieu was a member of the White Earth Chippewa, while Rev J.

A. Gilfillian had lived on the reservation for twenty-five years.

She discovered that the rhythm was repeated more exactly than the melody. She also

discovered that a number of the songs did not have words and that, if songs had words, they

"suggested rather than expressed" the ide a of the song (ibid.: 2). If the words changed, they

conformed to the music, but the same word could be accented differently in different parts

of the song. She (ibid: 15) asserted that the principal reason for singing a song was directly or

indirectly to seek supernatural aid. However, a large variety of songs existed for different

reasons. She (ibid: 17) clarified that most of the old songs that had a major tonality included

the old "dream songs" used by the Midé, the practise of medicine, the pursuit of war, and the

essential nature of a dream associated with the idea of reliance upon supernatural help.

However, songs that expressed sadness, distress, or uncertainty used a minor tonality (ibid.).

Densmore's (1979[1929]) publication discussed "picture writing", which she (ibid.:

174) defined as "that form of thought writing which seeks to convey ideas by means of

picture signs or marks more or less imitative of the object of the idea". Significance was,

according to Densmore, an essential element of the pictographs and she maintained that they

Page 104

were "alike in that they all express thought, register a fact, or convey a message." She (ibid.)

observed that figures included "crude delineations" of men, animaIs, birds and other

"material objects" and that the Midé used symbols that represented the sky, earth, lakes, and

hills, and sounds including concepts such as "spirit power". Simpler symbolism was used for

messages, including signs that indicated days, direction and duplication of numbers.

Occasionally these were combined with possible ideographs to indicate "progressive action".

Densmore maintained that two types of writing existed: esoteric and non-esoteric.

The esoteric writing could only be understood by initiates, while the non-esoteric writing was

used by everyone to convey information. She argued that examples of esoteric writing

included the Midé records, writings and songs, stories of Winabojo, and drawings used to

work charms. Non-esoteric writing included "totem marks," messages left along a route and

maps used by travellers, illustrations for a narrative, records of time, as well names of

individuals (ibid.: 175). She developed and expanded upon each type of writing using

drawings.

Skinner's (1914) article briefly considered how the Plains Ojibway, known as the

Bungi, employed birch bark. He (ibid.: 317) observed that they used "picture writing", as he

termed it, on birch bark and wood, which in turn reflected this group's origins in the

Woodlands region. Skinner argued that this type of writing was connected with religion and

that the Midewiwin was important. However, he did not describe the images, who used

them, what they may have meant, or how their meanings were transferred from one pers on

to another.

Several very descriptive articles provide little analysis of the images on the scrolls (for

example Cadzow 1926; Kinietz 1940, Reagan 1921 and 1922). Reagan conducted fieldwork

while he acted as lndian Agent at Nett Lake in Minnesota. His (1921) article published some

medicinal recipes made from plants. George Farmer, Ne-bay-day-ke-shi-go-kay, "a medicine

man" of the Bois Fort in Minnesota allowed Reagan to copy them from his notebook.

Reagan (ibid.: 246) stated that the recipes were written in the Chippewa language using

Roman characters. Reagan also (1922: 332) published the words of the songs used by the

Midé, which he found were recorded by the same "medicine man". He (ibid.: 332) observed

that Farmer had used Roman characters, but that the "phonetic values were similar to

French". Reagan published a few translations of the songs, recorded in Farmer's notebook,

stating their meaning, and when in the Midewiwin ceremony the songs were used. At the end

Page 105

of his (1922: 366-369) article, he provided a brief description of the Midewiwin Society in

this region. Both articles are interesting, since it is possible to establish what these songs

meant. However, it would have been useful to know which song was connected to which

mnemonic and whether songs could exist without mnemonics.

Kinietz (1940) unsuccessfully sought birch bark records while conducting fieldwork

among the Chippewa of northwestern Michigan over several seasons. He (ibid.: 38) asserted

that the type of symbols used, the size of the record, and "the excellence of the delineation"

affected the type of record. He argued that "the excellence of the delineation" varied not just

with the purpose of the designs, but with the artists of the different bark images.

Unfortunately, he did not provide any indication as to his methodology or evidence that

could substantiate his claims. However, he distinguished two types of birch bark scrolls

based upon subject matter. The first, understood only by instructed persons called Midé,

were, according to I<inietz (ibid.), "ideographic, esoteric and mnemonic". The second type of

scroll he described as being descriptive in character, and was related to the sign language of

the Indians. Unfortunately, he provided little evidence to substantiate his daims. Otherwise,

he reiterated many of the findings of early ethnographers and preceding researchers

regarding both scrolls and the Midé.

Voegelin (1942: 44), unlike I<inietz (1940), stated that she was successful in her

research. Voegelin (1942: 44) conducted her fieldwork amongst the northern Minnesota,

Wisconsin, and eastern Ontario Ojibway. She maintained that pictographs were used to

convey messages, provide topographic information, and "illustrate narratives". She observed

that the Midé used pictographs to record esoteric history or songs. She (ibid.: 45) established,

based on information from informants, that, in post-contact times, a pointed piece of metal

was used to cut images into the birch bark while in "pre-white times" a wooden or bone

stylus was employed. She was also told that red paint made from blood root often filled in

the outlines of the incised figures (ibid.: 44-45). Voegelin (ibid. 45-46) described and

identified a number of images found on the birch barks scrolls using information gained

from her informants. She observed that when she asked her informants for the terms

"picture, photograph, pictographic representation and so on" (ibid.: 46) that the words given

for each of these terms were numerous and varied considerably between different dialectic

groups, especially on the edges of the Ojibwa and the Ottawa groups.

Page 106

Coleman's (1947) publication examined, discussed at length, and compared the

images from two birch bark scroils as part of her discussion of how the Ojibwa in northern

Minnesota used images in religious contexts. She established that these scroils were read

from left to right (ibid.: 79). One, a Midewiwin scroil, belonged to her informant Charlie

Fox, a Midé who also was the leader of the Midewiwin at Naytahwaush (ibid.: 77). The

second scroil belonged to a Mr Gustavson, an Ojibway from the White Earth reservation.

She observed that the images on the second scroil were larger and more finely delineated, but

that the details of the "faces were lacking and the manitos were not nearly so artisticaily

executed" (ibid.: 80).

Both Blessing's (1963) and Kidd's (1965) papers indicate that archaeologists were

also interested in birch bark scroils. Blessing's (1963) study of a coilection of birch bark

scroils differed from Kidd's (1965) and indicates that he was determined to establish the

meaning of these images on the birch bark scroils. Blessing (ibid.: 91) used a number of

informants who were Midé, sorne of whom were from Madeline Island, Wisconsin and had

since died. His principal informant was Thomas Shingobe, a Midé and member of the Miile

Lacs band whose ancestors originaily came from Madeline Island, down the St. Croix River

and into Minnesota. Blessing (ibid.: 93) argued that the Midé were members of a "cult" that

was "more or less an elaboration of earlier hunting societies which had been formed in the

pressing need for the organized gathering of food". He continued his line of argument by

stating that, as religious activities grew increasingly multifaceted, it became necessary to

devise a method of keeping records. Hence the individuals involved "utilized natural and

conventionalised forms" as mnemonic devices. He posited that the scroils from Minnesota

no longer served as memory aids but as a means of communicating information. These

mnemonic images remained "almost exclusively the property of the originator," but, later,

became part of the larger group. Among the Ojibwa (ibid.: 94), these images became more

complex and "approached the stage of hieroglyphics." He considered the scroils to be old.

Blessing further asserted that the largest number of scroils had been produced in Minnesota.

How he arrived at this conclusion is unclear. Blessing examined scroils from Minnesota,

Wisconsin, Michigan, and Canada, but it remains unclear where the Canadian scroils

originated (1963: 95). He established that the eIder Midé took time and effort in interpreting

the images on the scroils during the instruction of the younger Midé. It was in this manner

Page 107

that the meanings and significance of the images were transmitted from one generation to

another.

Blessing considered the Midé, the Midewiwin, and their activities in Minnesota in

considerable depth: from where and how the Ojibwa had come to settle where they did in

Minnesota, what were guardian spirits to the Ojibwa and the Midé, and how a scroil was

made and inscribed (ibid.: 100-109). Blessing argued that he had accumulated substantiai

quantities of information on the practices of the Midewiwin from two reservations: White

Earth sometime after 1946 and Miile Lacs in 1968. U sing informants, he (ibid 109)

established that the "minor" animal species were associated with lower degrees of the

Midewiwin and that the "rugher species" of animaIs were connected with the higher degrees.

For example, weasels, otters, and mink were associated with the first degrees wrule mytrucal

creatures were associated with the fifth to eighth degrees (ibid.: 109-110). Scroils from

different geographic areas exrubited slight differences.

Blessing (ibid.: 110) established that the scroils in his study consisted of three types:

30% were "instruction" scroils, 50% "ri tuaI" scroils, and 20% were "master" scroils.

Candidates used an "instruction" scroil to prepare for the initiation ceremony (ibid.: 110-

111), while the Midé who conducted the Midewiwin ceremonies employed a "ritual" scroil as

a "weil illustrated edition wruch served as a guide or a reference" (ibid.: 111). The "master

scroil" was cailed an rustoricai document but also indicated the four to eight degrees which

could be obtained by the Midé. Blessing subsequently examined seven scrolls wruch included

ail three of these types. He provided detailed drawings or tracings of seven scroils.

Unfortunately, by using the terms "natural" (ibid. 96) and "invented" to classify sorne

images, he demonstrated a potential problem in rus study since rus own judgement seems to

interferes in rus description.

Kidd's (1965) very descriptive article discussed two scroils discovered witrun an

archaeological context. Kidd wished to make archaeologists aware that the discovery of this

type of object could occur in regions where birch bark was used. P. Ryan accidentaily made

the first find in 1924, reporting it to Kidd in 1959 (ibid.: 480). The first group of birch bark

scroils were found when the water from the fire fighter's hose washed away the soil of a

mound on the north shore of Rainy Lake that covered them. Ryan recailed that it contained

probably around 200 to 300 hundred pieces of birch bark (ibid.: 481). W. T. Daigetty and

Bruce Littlejohn made the second find in 1962 in a cave on Burntside Lake in Quetico

Page 108

Provincial Park. Kldd examined both sites and both groups of bireh bark. He (ibid.: 483)

eoncluded that "seroils" and their assoeiated artefacts eould be found buried in the ground in

places seemingly "out of the way" in the woods, or hidden in caves. He eould not date when

this happened for either group of bireh bark objeets but maintained that arehaeologists

should be aware of the possibility of similar diseoveries in the future.

Chapter III eovers Dewdney's and Kldd's (1962, 1967) examination of the rock

image sites of the Canadian Shield in depth. However, Dewdney (1962: 110 -116) posited

that a eonneetion existed between the images of rock image sites and those made by the

Midewiwin, based upon their similarity. Dewdney eompared a few of the pietograph and

petroglyph images with those of the bireh bark seroils that were known to belong to the

Algonquians of the Great Lakes region. He eoncluded that a study of the depietions of

supernatural beings in the mythology of the Algonquians might aid the identification of these

images. Over ten years later, Vasto kas and Vastokas (1973) suggested that a eonneetion

existed between the engravings of the Peterborough petroglyphs, the images on the bireh

bark seroils, and Algonquian shamans (please see Chapter III for a detailed examination of

their study).

Landes (1968) eonsidered images on bireh bark seroils in her examination of the

Ojibwa religion and the Midewiwin. She (ibid.: 172) stated that only the shaman who ereated

the seroil understood the specifie meaning of the images used. It was the shaman's role, she

(ibid.: 172 - 173) asserted, to teaeh "this explieitly" to someone else "before the notation was

useable"; the shaman's intentions dietated the level of knowledge or seereey. She (ibid.: 224)

deseribed bireh bark seroils and their images, and stated that its intention and, frequently the

availability of bireh bark, determined the size of a seroil. She (ibid.) argued that the

"partieular message" of a seroil, essentiaily, what the pietographs meant, was probably

impossible to establish unless eaeh figure was labeiled or translated by the owner.

Dewdney's (1975) publication is valuable, although it was eriticised by Rogers (1975)

and Veesey (1976). Dewdney (1975: 21-22) posited that six categories of seroils existed based

on the variations in the themes he had identified in ail of the seroils that he examined. This

list included: (a) origin seroils, (b) migration eharts, (e) mas ter ritual seroils, (d) ghost lodge

and sky degree, (e) deviant seroils, and (f) enigmatie seroils. He asserted the importance of

remembering that the principal funetion of the bireh bark seroil was mnemonie. From 1966

to 1972 he undertook to identify and traek down over 150 bireh bark seroils in public and

Page 109

private collections in Britain, Canada, and the United States of America. He noted their

current physical location, type of scroll, its catalogue number, where it came from, who

collected it, its length in inches, number of sections, and a source code. He recorded the

images of sorne of the scrolls using a technique that he described in depth, whereby he traced

the scrolls, or, if a collection was deemed too large, only the "most important specimens"

(ibid.: 7) and sketched the remainder. He did not indicate the process by which he

determined which scrolls were more important. Dewdney (ibid.: 5) distinguished between an

instruction scroll or chart and what he called a song scroll. He (ibid.) argued that an

instruction scroll was a linear series of images to be utilised as memory aids "in rendering

either a sacred Midé song, or a profane personal one." The profane song was dreamed by a

"visionary shaman" for a various purposes, while a Midé composed a sacred song, which

could be sold or traded. However, he believed that the contents of the latter could only be

interpreted by the composer of the scroll or the owner. He (ibid.: 15-18) described in

considerable depth how scrolls were constructed from birch bark, and the possible

techniques used to draw the images on their surface. It remains unclear who made the

scrolls, but Dewdney maintained that the task of stitching of birch bark together to create a

larger scroll was 'traditionally - but not always - women's work." He did not provide any

evidence to support this statement. The stitching cou Id be "highly decorative" (ibid.). He

conducted an in-depth analysis of the styles used to depict the images, using terms such as

"rectilinear, abstraction, condensation, symbol conversion, atrophy, amputation" (ibid.: 18)

to describe the processes by which shapes changed as the scrolls were repeatedly copied.

Unfortunately, his terminology makes it very difficult, even impossible, to comprehend

precisely what information he wished to convey to the reader. His terminology is,

consequently, difficult to use. He also transcribed a number of the scrolls with the aid of

James Redsky from Shoal Lake Reserve, on the western portion of the Lake of the Woods,

relying on an interpreter when Redsky used Ojibwa rather than English.

Rogers (1975: 83) reviewed this book and asserted that Dewdney frequently allowed

his own perspective to colour the way in which he classified the spirit world of the Ojibwa.

Rogers argued that present day Ojibwa "do not" and never did adhere to the Euro-Canadian

worldview. According to Rogers (ibid.), the Ojibwa believe that the same spirits were helpful

in one context while they were injurious in another. He cautioned the reader of this book to

be wary of Dewdney's (1975) interpretations of the significance of the symbols on the

Page 110

/~.,

scrails. Rogers (1975) is considerably harsher than Vecsey (1976) in his review of Dewdney's

(1975) book but both reviewers severely criticised Dewdney's study for numeraus

weaknesses, which they ascribed to his laek of scholarly expertise.

Nelson (1983) diseussed a eoileetion of bireh bark scrails and objeets of unclear

origin. His article pravided a detailed review of previous work undertaken on these scroils by

Hoffman (1891) Sehoolcraft (1854), and Densmore (1979 [1929]). He also examined a

number of serails illustrated in various gailery catalogues. He (1993: 220-221) listed the

serails that Dewdney had not reeorded that were in other publications. Nelson subsequently

used Dewdney's (1975) publication in conjunction with Blessing's (1976, 1977) to identify

and de scribe at length the different types of serails that he had identified using Dewdney's

classification (1975: 21-22).

Vastokas (1984: 425), an art historian, acknowledged the work of her predecessors

but asserted that the interpretation of the images on the birch bark scrails stiil presented "the

most pressing chailenge to scholars" interested in Algonquian art history. She eould have

included archaeologists, anthrapologists, and historians. Vastokas (1984: 435) maintained

that the scrails were the most significant body of visual documents available for the study of

native North Ameriean art history. This paper outlined sorne preliminary and exploratory

observations in interpretation theory and methods whieh she maintained were applicable to

bireh bark serails. She (ibid.: 426) postulated that the style, iconography, and development of

the serails should be eonsidered from an art historical perceptive. Vastokas believed that,

since further fieldwork was impossible (ibid 427), two methods of researeh existed to

interpret these serails: (i) examination of the images depicted on the scrails themselves, and

(ii) external praof supplied by comparison with analogous or related visual systems.

She subsequently outlined each avenue of research and argued that data could be

obtained thraugh detailed examinations of the manner in which the motifs and elements

were organized on the scroils. Art historians, she maintained (ibid.: 428), eould use the

notion of style as a tool. Preliminary observations of the style in which these images were

executed, Vasto kas eontended, highlighted clear differences between the official Midewiwin

scroils and the personal song seroils that belonged to individual members, and these in turn

suggested significant differences in their funetions and meanings. She maintained that these

differences were observable without ailuding to the cultural eontext of each scrail. The most

intrinsic and explicit virtue (ibid.) was their narrative structure in which separate pictorial

Page 111

units were arranged to convey information to the beholder regarding mythological, historical,

or ritualistic events across time and space. This narrative role was crucial in any analytical

procedure aimed at interpreting the "form" or content of each scroil's images; the form

being the physical outline of an image. She maintained that narrative art differed from iconic

representations since, unlike iconic representations, narrative art could communicate events

in time and space. Narrative compositions she (ibid.: 430) defined as multiple motifs

organised in a "more perceptuaily scattered disposition ailuding sequence and movements to

the eye across the surface of pictorial ground." Furthermore, profile images denoted motion,

while frontal views indicated action. Vastokas (ibid.: 431) postulated that narrative art forced

the observer to detect the greater visual complexities of the composition. She concluded that

the birch bark scroils were "narrative visual wholes" to be interpreted in terms of their form

and context. She disagreed with early endeavours to decipher the images, such as Mailory's

(1893) attempts to examine them as writing systems, because these scroils were visual rather

than verbal documents.

Vastokas' (ibid.: 427) second avenue of research, outlined at the beginning of her

paper, was that external proof should be supplied by comparison with analogous or related

visual systems. The Midewiwin scroils play crucial roles as visual religious texts, as mnemonic

aids in the instruction of new members, and as a means to record oral tradition and correct

ceremonial procedure. Essentiaily, they recorded the history, origins, and ritual of the

Midewiwin Society. It was vital, therefore, to understand these scroils within the larger

context of Ojibwa shamanism and the numerous scroils that record the songs of individual

Midé. The song scroils contained, Vasto kas maintained, links with traditional motifs, since

personal scroils habituaily used traditional motifs that derived from personal vision quests.

Advances in the interpretation of the images, Vastokas (ibid.: 437) concluded, could occur if

theories and methods were utilised from a range of fields including art history, symbolic

anthropology, history of religions, and philosophical hermeneutics. Future examinations of

these images, she asserted, would be advantageous if they were undertaken in the context of

a circumpolar shamanic art tradition, since it had a cross-cultural system that was fairly

consistent.

Vennum's (1978) study could be placed in Vastokas' (1984: 427) second line of

research. He drew extensively upon the data and the songs of the Mide collected by Blessing

from the Mille Lacs and White Earth reservations. Vennum argued that larger numbers of

Page 112

Mide songs existed in this coilection than Densmore had coilected. He was interested in the

origin migration scroils of the Midewiwin, and so, he reiterated the legend of the Ojibwa

westward migration. He argued this had to be undertaken because the migration charts

"pravided the necessary dues for the identification and the interpretation of the origin­

migration" serails. Hence he pravided a considerable quantity of information about this type

of chart and about understanding its raIe within the Midewiwin. Vennum asserted that this

type of serail was distinguishable fram other serails beeause images behave as mnemonics

for the place names of "real or legendary resting places" used by the Ouer or Bear as they

traveiled westwards (ibid.: 761), while in other serails these places were presented by dots.

Serails were read from left to right, sin ce the Ouer, in the origin-migration story told by the

Midé, traveiled from the east to the west. Vennum argued (ibid.: 753) that the development

of the origin-migration story was linked with the migration of the Ojibwa as they moved

from the east to the west side of Lake Superior before the end of the seventeenth century.

Vennum diseovered that twelve migration charts existed in the Miile Lacs-White

Earth papers coilected and transeribed by Blessing (1963) and his informants. Vennum

(1978: 762-763) identified the place names, and was able to establish precisely where on Lake

Superior they had stopped. He argued that this information "chailenged" Dewdney's (1975:

9) statement about the indigenous knowledge of Lake Superior, while corroborating and

expanding the versions of the migration coilected by Warren (1984 [1885]) and Hoffman

(1891) during the nineteenth century. However, Vennum (1978: 764) could not identify ail

the images as place-names, and conduded that these were probably the "texts of songs". He

established that specifie motifs from sorne song scroils bore a striking resemblance to those

surviving in the song texts, suggesting that regularities also existed in their meaning. He drew

extensively upon Densmore, Schoolcraft, Kohl, Hoffman, and Landes (1968) to identify the

songs and discuss how they might be interpreted. Vennum (1978: 787) argued that these

migration serails may have been performed near the end of the Midewiwin ceremony, based

upon information in the Miile Lac-White Earth papers. He subsequently argued, based on

Blessing (1963: 93-94), Hoffman (1891: 290), and Densmore (1910: 26), that the images, as

mnemonics, "were less generalized and their meanings secretly guarded" (1978: 788).

Vennum (1978: 789) suspected that this could explain why Densmore (1910: 26) had

observed that two Midé may have had the same song, but posited that this occurrence was

possibly a coincidence. Based upon the Mille Lac-White Earth papers, Vennum believed

Page 113

(ibid.) that, in the post-1900 period, shared ownership of the songs began occurring in the

pictographs and the texts, but he was unsure as to whether the songs also shared melodic

similarities. He observed that Densmore had not considered whether her informants

recognised the songs by their texts, their images, or, as he termed them, pictographs, or by

their music. He (1978: 790) concluded firstly, that the Mille Lac-White Earth papers made

possible an understanding of the wide variety of ways in which the same Midé song could be

utilised, and, secondly, that a detailed examination of these songs was an excellent start for

anyone attempting to establish the meaning behind other "abstruse Ojibwa song scrolls."

Finally, he concluded that details found in the migration scrolls were useful for interpreting

Ojibwa pictography and song poetry in general.

Closs' (1986) article is the only one that built upon Dewdney's (1975) work with

birch bark scrolls. It examined the ritual importance and use of numbers, especially the

number 4, for the Ojibway, and in particular for the Midewiwin. He utilised the ethnography

collected by Densmore (1929), Dewdney (1975), Hoffman (1891), ]ohnston (1976), Landes

(1968), and Warren (1885). Closs (1986: 181) followed Dewdney's (1975: 12) definition of

pictography, which was used as "a generic term for any form of preliterate art-executed on

any available surface - that is known, or is assumed to have had communicative rather than

decorative or aesthetic intent". He (ibid.: 182) agreed with Dewdney and argued that this

pictography was mnemonic and "did not represent the written word." Closs applied the

information concerning the use of numbers to the drawings of six birch bark scrolls. He

(1986: 189) examined six Midé scrolls that he believed exhibited "graphic properties of the

ritual importance of the number 4" and "illustrate the usage of the number series in the Midé

cult." He argued that five of these scrolls were, according to Dewdney's (1975: 21-22)

classification system "master scrolls," while the sixth was a "possible migration scroll."

Closs' argument is interesting, but it remains unclear whether he examined the

original scrolls, or used drawings published by Densmore in her 1929 publication and

Dewdney in his 1975 publication. It is important to examine the actual scrolls rather than a

drawing or sketch of the scroll, because it is possible that the person who made the drawings

inadvertently made mistakes in copying the images.

Vastokas (1986-87) stated that any investigation of prehistoric and post-contact

native North American art presented new challenges for art historians, because it lacked

written documents, which, she acknowledged, were crucial for historians. New art historians

Page 114

(ibid.: 13) could examine these art works as primary data because they did not recognise the

lack of written documents as an obstacle. New art history allowed academics from different

scholarly areas to draw upon a wide variety of disciplines, themes, and methodologies.

Vastokas, like Rlfkin (1986), argued that new art history assumed the universality of human

visual expressiveness and adopted a cross-cultural perspective in which to explore art from a

point of view held by a particular social group, time period, or culture and instead of a

twentieth century perspective.

Vastokas (1996: 48) developed ideas introduced in her previous article. She argued

that images had been insufficiendy exploited in the writing of Aboriginal history and

suggested that these visual documents should be carefully drawn upon as "aids" when

writing Aboriginal history. She promoted the idea that much of Aboriginal art was produced

intentionally as "history", so that pictorial depictions should be recognised as writing.

Aboriginal artworks, she maintained (ibid.: 49-51) had different purposes in traditional

culture, and, since the arrivaI of the Europeans to North America, new functions were

attached to these objects (also see Phillips 1999).

Vastokas identified the problem as stemming from the viewpoint of European

historians and art historians who held that Aboriginal societies lacked a history despite the

fact that Aboriginal societies often recorded it. She posited that previous academics had

ignored the most important function of pictorial representation in traditional native culture

which was the recording of history. According to these scholars, Aboriginal peoples lived in

an "anonymous and static present". Vastokas argued that this attitude had affected both the

interpretation and presentation of the Aboriginal artworks and Aboriginal history.

According to her, there were four different ways by which Aboriginal art works

"served" the writing of Aboriginal history: firsdy, as "historical documents", in the same

manner as a textual document used by an historian with its own date and characteristics,

enabling the development of historical chronologies (ibid.) (an ide a she had introduced in her

1986-87 article); secondly, as "cultural expressions" that were visual embodiments of

Aboriginal beliefs, values, and society as recognised by practitioners in new art history;

thirdly, as "representations" of events taking place in time and space; and finally as "writing"

(1996: 52-53). She reiterated her idea that Aboriginal art in North America could be

subdivided into two dis crete functional types called iconic and narrative that she had

discussed at length in her 1984 article.

Page 115

Vastokas (1996: 53) supported Boone and Mignolo's (1994) daim that it was

important to get away from the modernist \Yj estern conception of art as something to be

appreciated strictly aestheticaily and to consider seriously the idea of visual representations as

communication. Vastokas (ibid.: 53-54) subsequently developed her argument regarding

Aboriginal art as "writing," stating that there was a good reason why aboriginal people

referred to pictographs and pictorial representations as "writing." She (ibid.: 54) daimed that

for Aboriginal North America the birch bark images, or "birch bark manuscripts," were the

most noteworthy expressions of the relations between image and word.

She described in considerable depth the structure and role of the Midé and of the

Midewiwin Society for the Ojibwa, the role of the birch records for the Midé, the range of

their physical dimensions, the range of images being used "from depictive to symbolic, to

entirely abstract in character" (ibid.: 54). Vastokas cited Boone and Mignolo's (1994) daim

that the terms writing and art, if applied to most of Aboriginal America, were problematic

since, unlike western and other phonological writing systems, the goal in the Americas was

not "visible speech." In further support of her supposition, Vastokas (1996: 57 ) cited Boone

and Mignolo's proposition that Pre-Columbian American art and writing were largely the

same thing.

For Vastokas (1996: 57), the Ojibwa birch bark scroils belong to "semasiographic

writing systems" which communicate ideas independently of language.6 She asserted that, if

scroils could be accepted as examples of "semasiographic writing systems," historians could

obtain an insider's view of native American history. Boone and Mignolo had argued that,

within these systems, a subset of iconic writing systems exist composed of representational

pictographic units that transmit meaning without employing speech. Anyone wanting to read

these pictographic units already knew the general meaning of individu al units and so, could

ascertain the meaning of the different combinations.

However, this notion does not distinguish between the fact that different groups of

people might have the knowledge to use certain images and not others. It is impossible under

Boone and Mignolo's definition that a general vocabulary of pictographic units exists that

can be subdivided into groups which are in turn recognised by the general population as

6 Warkentin (1999:3) argued that historians of writing divide sign systems into two groups: semasiographic language systems (pictography) and phonographic (language based) systems. She asserted that phonographic systems were further subdivided into logographic (where the sign is the equivalent to the word), syllabic (the sign is a syllable) and alphabetic systems.

Page 116

having sorne meanings. Each group of pictographie units is used by either ail or specifie

groups, within that society, for different purposes and reasons. Therefore, sorne of these

pictographie units are not understood by ail.

Boone and Mignolo argued that, not pictorial and conventionalized images, but the

placement and context of the images carried meaning. They (1994: 18) asserted that images

behaved as texts in semasiographic writing systems; however, these writing systems did not

have fuil running texts but a highly conventional set of symbols with meanings similar to

those associated with paintings.

Vastokas (1996: 57) maintained that Boone and Mignolo's (1994) conclusions

regarding Mixtec and Aztec writing systems were pertinent to the pictorial narratives of the

Algonquians, since the scroils were among a variety of visual narratives that historians could

use to examine native American history. Therefore, if the images on the birch bark scroils

were accepted as "semasiographic writing systems," these scroils could be accepted by

historians as one of a number of visual narratives that provide an insider's perspective on

native American history.

The consequences of this claim are interesting, since they mean that an image can be

treated as valid a primary document as is a piece of text. Vasto kas stated that considering the

scroils as documents "merely presented a different set of historical chailenges" (ibid.: 57).

She believed that the "most obvious" problem was "the absence of absolute dating" (ibid.).

She believed that additional research was required into the oral traditions that accompanied

visual documents, and into the contexts in which they were utilised. She acknowledged that,

although information had been lost, a considerable body of information still existed.

Vasto kas also lamented that "very few elders" among the Aboriginal community were "stiil

in touch with the traditions" that created these images (ibid.: 58). Vastokas remained

optimistic that fundamental research into these birch bark scroils could be conducted,

despite these considerable obstacles, especiaily if Native and non-Native historians workcdin

conjunction with each other.

Fulford's (1989, 1990) articles fail into Vastokas' first avenue of research. His

findings demonstrate that use fuI and interesting information can be obtained if a detailed

examination of these images is undertaken. He considered the shapes of the twenty-four

song scroils utilised by members of the Midewiwin from White Earth Reservation in

Minnesota (Fulford 1989). Hoffman coilected them during the late 1880s for the Bureau of

Page 117

American Ethnology. Fulford examined them for patterns of structural variations, and

outlined how three complex symbols or shapes, which he identified as the otter, bear, and

bird, had evolved from clan markers into pictographic markers. It was impossible (ibid.: 134)

to examine precisely how the interactive process influenced the form of the pictographs

because this required comparative observations to be made by potential members of the

Midewiwin as they re-transcribed the record.

Systematic variations were observable ln redundant forms once the White Earth

pictographs were arranged into concordances (indexes). Twenty-four scroils were reproduced

in three concordances, which were published in conjunction with this paper. Fulford (ibid.)

maintained that this method was similar to that used by historicallinguists to reconstruct the

protoforms of a spoken language, because it highlighted the changes introduced by iteration.

The subtle variations existing in redundant pictographs reflected the changes introduced in

the process of recopying the images. Over time, these changes become highlighted and

modulated directly with the rate of iteration, as weil as with the formaI complexity of the

pictographs. A delicate balance existed, Fulford suggested, between those pictographs, cailed

forms, that were too simple to be recognised, and those that were too complex to be stable.

He argued that a certain degree of formaI simplification or standardization assisted

communication. Complexity, he suggested, introduced changes into the system that increased

its adaptability, and enhanced the potential flow of new information.

He used two interrelated indices to evaluate the complexity of each form and image,

and to reduce the potential of projecting ethnocentric conjectures that the complexity of the

symbols encouraged. Two terms were used, "simple" and "complex", to refer to the

symbol's frequency of distribution throughout the sample, and its capacity for

transformation (ibid.: 135). Symbols were termed "complex" by Fulford if they were clearly

distinguishable throughout the sam pIe and had an obvious capacity for transformation, while

"simple" symbols had a homogeneous distribution and were incapable of being transformed.

However, a symbol could be complex because of the frequency of its distribution and its

capacity for transformation, but simple in the wider context of its meaning to Hoffman's

informants.

Fulford (ibid.: 149) discovered that the otter, bear and bird shapes were weil

distributed throughout each of these scroils, and indicated a "high capacity for

transformation". He concluded that the transformation that he found arose as a consequence

Page 118

of the re-transcription of the song scroils. He maintained that the grain of the birch bark

could not entirely explain the syntagmatic and associative relations between the shapes,

which he cailed pictographs, of the White Earth scroils.

Fulford (ibid. 150), based upon Vygotsky, posited that the movement of the stylus

and eye aeross the birch bark scroils was connected to higher psychologie al functions, and

that the formation variations introduced through iteration had potential symbolic

significance. He maintained that Midé symbolism achieved its most powerful exposition if

the relationship between the sequential linguistic elements and the complex symbols was

examined, since the particular transformation features of these symbols could be identified as

the projection of the human form into natural objects.

The patterns revealed in the three concordances, Fulford asserted, provided an

insight into the symbolic world of the Midewiwin. He argued that the propensity of the otter,

bear, and bird symbols to fluctuate as a consequence of iteration into distinct human forms

echoed the unconscious maintenance of earlier symbolic associations. He contended that the

elemental symbols used during the protohistoric period were linked to those used by the

Ojibwa clans of the late nineteenth century through Midé rites. Images (ibid.) that were once

associated with real human faces from the Ojibwa clans during the eighteenth century

subsequently developed into the ghostly figures of the Midé birch bark scroils during the

nineteenth century. He concluded that the Midé were able to distinguish the images of their

ancestors in the symbols of the otter, bear and bird, although the Ojibwa clan system had

disintegrated.

Fulford's study demonstrated that connections existed between different shapes,

which he cailed motifs or symbols, and ethnographie details, although the context of these

meanings had changed.

Fulford (1990) later considered the verbal structure of these Midé chants to establish

how they were pictographicaily encoded. He limited his discussion to seven song scroils, six

of which shared the otter symbol. The seventh scroil was included since it shared several

other formaI characteristics with the six otter scroils. He (ibid.: 127-128) retranscribed and

re-translated the chants that Hoffman had compiled before reworking the entire corpus to

develop glosses to improve the flow, or simply supply better connections between the words

and the pictographs. It was impossible, he maintained, to provide definitive transcriptions

and translations of the Hoffman material due in part to Hoffman's informants' poetic use of

Page 119

languages. Fulford concluded that the reproductions of the scroils were generaily accurate by

comparing Hoffman's published reproductions with the originals in the US National

Anthropological Archives. He subsequendy placed the changes of the seven scroils in a

concordance by typing the revised version of Hoffman's original transcriptions of the chants

and attaching photocopies of the related pictographs to the top of the transcription. This

information he put on file cards in horizontal rows where each row corresponded with the

sequence of pictographs and changes of one song scroil.

Fulford (ibid.: 128-135) believed that the recognition of the taboo that existed on

uttering the otter's name eXplained the relationship between the pictographs of otters and

hum an arms in the song scroils. He argued that the recognition of this taboo was clarified by

chants that accompanied otter pictographs, since they ailuded to incidents in the Midewiwin

creation story but did not directly describe the pictographs themselves. These differences

caused a mismatch to occur between the visual and the verbal forms and transformed the

meanings of the pictographs and the chants. Fulford (ibid.: 135) maintained that the verbal

transformation that occurred was the result of an iterative process that occurred in traditional

Midewiwin chants and mirrored the taboos connected with enunciating the names of "animal

brothers" linked with totemic ancestors.

Both of Fulford's papers complement each other and demonstrate that, with careful

deliberation, these images can be analysed. Both papers arguably support Vastokas' (1996)

argument that native artwork could be used as "historical documents".

Similar shapes to those on birch bark scroils were examined by Fulford (1992: 190)

who considered the ninety-two page account of Vincent Roy, a fur trader of Scottish and

Ojibwa ancestry. Roy administered the American Trader's Post at Vermillion Lake between

1835 and 1839. Roy's account book, which could be securely dated to the 1830s, contained

drawings that strongly resembled those on birch bark scroils, Fulford argued. He identified

forty of the forty-eight different types of the trade goods recorded in the account book and

twelve species of fur-bearing animaIs. He determined which pictographs signified debits and

credits of each account and the identities of sorne of Vincent Roy's customers.

In being able to identify Roy's customers, and that they were not local to the post,

Fulford could indicate that people were willing to travel vast distances to ob tain goods.

Fulford (ibid.: 192-193) established that sorne of Roy's customers used remote posts such as

Roy's, which was outside Schoolcraft's jurisdiction. Schoolcraft was the nearest

Page 120

Superintendent of lndian Affairs in Michigan, and he discouraged Americans from doing

business with Ojibway who sustained ailegiance with the British crown.

Fulford (ibid.: 192-194) believed that Roy was semi-literate but his two years of

schooling while living at La Pointe had taught him numeracy, literacy, and a Calvinistic

attitude to business. These basic reading, writing, and mathematical tools enabled him to deal

with suppliers in St. Louis or New York. However, the pictographs Roy utilised to record his

inventory were familiar to his Ojibwa contemporaries including those used Midewiwin song

scroils (ibid.: 194). Fulford argued that Roy provided evidence of an assimilation process that

occurred amongst the Ojibwa over two generations during the nineteenth century, as the

proliferation of schools encouraged reading, writing, arithmetic, and the Christian religion

among the indigenous populations. He ascertained from Roy's account book that, despite

this process, Ojibwa stiil used traditional names, and considered themselves as living in

harmony with the natural world, although by the 1830s the fur trade had destroyed their

traditional economic, social, and political structures. Fulford's article indicates that the

indigenous people of the Great Lakes area understood and operated within two conceptual

frameworks and confirms that these images, although they appeared on the birch bark

scroils, were also used elsewhere as carriers of meaning and information.

A question exists regarding the relevance of this body of information, especiaily since

this materiallacks written documentation. The Aigonquian-speaking peoples, rather than the

Assiniboine, used a specifie group of images. Unfortunately, knowledge of Assiniboine art,

and of other Plains groups from before the nineteenth century is limited, given that much of

it was made of perishable materials that frequendy failed to survive in the archaeological

record. Greene (2001: 1038) observed that quillwork was the major decorative art form in

the Plains over a long period of time. Yet rock image sites also exist in the Plains region

(Keyser and I(lassen 2001), but the "biographic" style associated with this region is

connected to bison robes and the ledger book paintings used during the late eighteenth and

early nineteenth century (K.eyser 1977 and Greene 2001: 1042).

It is evident that the images, made by the Aigonquian groups and used as a means of

communication, and the techniques used to create them, have shifted over time. However, it

is clear that not ail of these images changed radicaily in appearance and experienced changes . . ln mearung.

Page 121

It appears that three groups of images exist, based upon who used them and their

audience.

The first group of images have not undergone considerable changes in their shape.

The Midé used these images to record their religious and medicinal information, and this

group has been examined by Fulford (1989, 1990), and Vastokas (1984, 1986, 1996).

The second group consists of the early geometric shapes that changed to floral

designs examined by Densmore, Coleman (1947), Penny (1991), Phillips (1984, 1991), and

Whiteford (1986, 1991, 1997). Many of the people who analyse these images invariably use

terms such as 'Thunderbirds', and 'Underwater Panthers,' rather than describe the shape of

the images.

The final group consists of those images that everyone in the Algonquian-speaking

community used. Unfortunately, little evidence exists of the images associated with this

group, but perhaps some of these were used by Roy (Fulford 1992). Nevertheless, Landes

(1968: 172) stated that images were familiar to ail Ojibwa hunters, trappers, and traveilers,

who used them to indicate boundaries, trails, notices of different sorts, and on grave

markers. Coleman (1947: 63) observed that the totem of the deceased individual was placed

on grave markers. Warren (1984 [1885]: 89-90) discussed some images inscribed on a copper

disc that he maintained recorded the number of generations of a family of the Loon totem

who had lived in the La Pointe area. The indentations indicated that eight generations had

lived in this region. Although relatively little is known it is clear that this group of images was

important to the Algonquian speaking peoples as a means of communication

Although relatively little information exists about this last group of images, they must

be considered, because this group of images indicates that everyone within the Algonquian

speaking community used images. It is therefore likely that everyone knew the entire range of

images, and understood how they were utilised and combined. An individu al could perhaps

use one or several images to express an idea, something about an individual, or leave a

message. It appears from Landes (1968) and Coleman (1947) that anyone could use, make,

and understand them. The creator of these images could be an ordinary individual, and did

not need to be a Midé, or any other specialist within the community.

Since ordinary people used images as a means of communication, it is possible that

they were known by everyone, and that an image meant more than just one thing.

Page 122

Unfortunately, little information exists as to what these images looked like, how

many existed, their range, or how they were combined, but it appears from Coleman (1947)

and Landes (1968) that they occurred individually rather than as the image groups of the

birch bark scrolls. The fact that they were used by a wide variety of individuals implies

everyone knew how to use how to use them and that specialists did not exist. Hence, these

images retained their shape much like those used by the Midé regardless of the considerable

changes in the imagery used in beadwork and weaving. However, whether any shifts in

meaning occurred remains impossible to establish.

Conclusion:

It evident that the information needed to contextualise the pictograph sites in the

Lake of the Woods derives from a range of disciplines. There is considerable agreement

about the climatic changes, and the geological history of the region. But the historians and

archaeologists find it difficult to agree who lived there, in part because each discipline uses

different types of evidence to solve their problem. Ethnohistorians (Bishop, Ray, and Smith)

argue that historical documents are useful for tracing ethnicity in the early historie period.

Many archaeologists, who pre fer to equate ethnicity with archaeological cultures, do not

share this opinion. There is sorne agreement that the Sandy Lake materials around the

Winnipeg River are possibly Assiniboine.

With the exception of Meyer and Hamilton (1994), archaeologists tend to assume

that the area was occupied by only one ethnie group at a time. While sorne archaeologists

adamantly as sert that the Ojibwa had already arrived in this region by the early historie

period, historians argue that several groups of people from different ethnie backgrounds,

Cree, Ojibwa, and Assiniboine resided here.

The inadequate archaeological data, in conjunction with the insufficient historical

references to the individual groups who lived in this region prior to contact, at contact, and

du ring the early historie period, inhibits scholars from identifying the terri tories of various

groups. The problems resulting from different names for the same groups could perhaps be

resolved by the careful use of the historical documents and the identification of synonymous

designations.

Page 123

1t also, so far, has proved impossible to ascertain which specific groups of

Algonquians, Cree or Ojibwa, lived in the region between Lake Superior and Lake Winnipeg.

However, both groups shared a similar symbolic system.

From the letter written by La V érendreye, in 1736 scholars know that Siouan­

speaking peoples were in the region at the same time as the Algonquian speakers. Since the

early part of the eighteenth century, the Assiniboine were allies of the Cree against the Sioux,

although the Assiniboine are linguistically closely related to the Sioux (DeMallie and Miller

2001: 572).

The first reference to the Assiniboine in the J esuit Relations for 1657-1658 describes

them as living about 100 miles from Lake Nipigon, east of the Lake of the Woods (DeMallie

and Miller 2001: 573). Jean-Baptiste Louis's map, dated 1697, indicates that the Assiniboine

resided north and west of the Lake of the Woods while the Sioux were in the south in the

area now called Minnesota (DeMallie and Miller 2001: 572). 1ndeed, the Lake of the Woods

is referred to as Lake of the Assiniboine. This calls into question Wright's suggestions (1987:

Plate 9) that the Ojibwa, Cree, Algonquian, or Montagnais peoples inhabited the Lake of the

Woods between AD 500 and European contact.

By the early seventeenth century, the Assiniboine resided seasonally on the western

half of the Lake of the Woods, and would do so continually until the middle of the

eighteenth century (Heidenreich 1987: Plate 37). Algonquian-speaking peoples, throughout

this period lived in eastern portion of the Lake of the Woods and gradually took control of

the region as the Assiniboine moved westwards (Heidenreich and Noël 1987: Plate 40).

The question of changing lake levels was considered, since archaeologists in northern

Ontario have traditionally found sites by examining the large expanses of cliff faces along the

shores of lakes and rivers. The graduaI but dramatic increase in the water level starting at the

end of the nineteenth century caused the shape of the Lake of the Woods to change, and

possibly completely or partially submerged pictograph sites. 1t seems likely that pictographs

that are known today were once physically located higher up the cliff faces. Yet, if the water

level had fluctuated as much as seven feet prior to keeping careful records, the positioning of

images could have reflected the differing high water levels at the time they were painted.

1t is very difficult to create a chronological framework regarding the precise ethnie

identify of those who lived in this region in prehistoric times. Historians and archaeologists

have had severe problems in attempting to connect the archaeological and the historical

Page 124

record to determine which indigenous groups resided in a specifie region at specifie rimes. It

is relatively easy to use historical records to determine the ethnic identity of a group of

people from a specific geographical region during the historical period. However, prior to the

arrivaI of the Europeans in this region, such records do not exist. This is caused parrially by

continuous changes in the naming of the people who lived in this region, and the difficulties

of identifying which archaeological complex was associated with which ethnic group.

Osteological information is interesting but, like other archaeological information,

bones do not indicate the ethnic identity of the individuals. However, it cannot be excluded

from a discussion of the identities of individuals. Examination and information regarding

burial practices possibly provides information about the social, religious and cultural

identities of the individuals concerned but not about their ethnic identity. Examination of the

bioanthropological evidence (osteology, human blood group analysis or mitochrondrial

DNA) indicates that considerable advances have been made, but these sciences inform the

archaeologist about the genetic relationships, not about the cultural affiliations, of the groups

involved. Specific traits were shared by the Cree and the Ojibwa.

Many of these images have the same physical appearance despite the passing of time.

Several scholars have demonstrated that similar stylistic and design traits could be shared

with others, and may have belonged to specific groups (Duncan 1991; Lanford 1984; Phillips

1984 and 1999; Sager 1994, 1995 and 1996).

Solid evidence regarding whether the images of the pictograph sites in the Lake of

the Woods region were created by the Cree or the Ojibwa does not exist. It is easier to

conclude that these images were made by the Aigonquians rather than trying to ascertain

whether they were created by the Cree or the Ojibwa. It is possible to posit that they were

probably created by Algonquian-speaking peoples rather than Assiniboine-speaking peoples,

since the rock images in the Lake of the Woods visually strongly resemble those of sites

elsewhere in northwestern Ontario where groups of Algonquian-speaking peoples resided.

lndeed, the images of the rock images created by the Assiniboine are visually very different

from those created by the various groups of Algonquian-speaking peoples in northwestern

Ontario. A specifie group of images were used by the Algonquian-speaking peoples rather

than the Assiniboine.

The physicallocation of the sites in the Lake of the Woods also fits in to the general

Algonquian pattern of the rock image sites in northwestern Ontario. This warrants using the

Page 125

/ '

Algonquian ethnographie data ta interpret the meaning(s) of these images and, hence,

justifies using homologies for eonsidering the meaning of the Lake of the Woods images.

Bence without either the arehaeologieal or the ethnographie proof for warranting the ethnie

origins of this imagery, but based upon eulture-historieal data, 1 argue that a basis exists for

using general Algonquian cultural data to interpret the rock images 1 am studying in this

thesis.

Page 126

Chapter III: A Review of the Literature concerning the Study of Rock Image Sites in the Lake of the Woods in the Canadian Shield prior to 2001.

The collection and interpretation of data are intergenerational processes in which

each new generation amplifies, modifies, and builds upon the work of previous generations

of researchers. A critical evaluation of previous work is an essential part of planning and

executing any new research project. It is important to acknowledge, at the same time, that

most investigators work within their own paradigm. However, this acknowledgement should

not prevent their work from being critically and carefully examined.

The goal of this chapter is ta present and discuss the work undertaken prior to my

own by previous researchers concerned with the pictograph and petroglyph sites in the Lake

of the Woods area. 1 intend to establish which approaches have been the most popular, what

previous researchers have thought of pictograph sites, and how they examined that in which

they were interested. This information is important to examine sin ce it provides a standard of

comparison, a yard stick against which to compare, examine, and contrast the new data that 1

collected. Although this chapter is specifically concerned with the analysis of the pictograph

sites of the Lake of the Woods area, sometimes 1 will make reference to analogous studies of

sites elsewhere in the Canadian Shield.

The literature in this chapter is divided into seven sections. The first discusses the

earliest investigation of these images with specific reference to the Lake of the Woods.

Section two examines the technical literature that has discussed how pictograph sites have

been studied with special emphasis on Ca) subdividing the sites into sections, (b) the

vocabulary used to describe the images and Cc) the development of the techniques used to

record the images. Section three is concerned with the considerable body of literature that

presents the issues involved in the technical analysis of these sites. It is divided into two

parts: Ca) recording and interpreting the pictographs and petroglyphs and (b) the

conservation of the pictographs and the different methods used to analyse them. The last

three sections critique the literature that belongs to one of the five different approaches that

exist in archaeology: culture-historical, contextual, and intuitive, analogical, or homological.

Since these approaches are different, and a clear order exists Csee discussion in Chapter 1), it

is important for the literature belonging to each approach to be examined in sequence.

Therefore, section five considers the literature belonging to the culture-historical approach,

where archaeologists have considered the shape, location, and date of the images. Section six

Page 126

deals with the literature classified as belonging to the contextual approach, which involves

the search for patterns relating different images and combinations of images both between

and within sites. The seventh and final section reviews the literature in which archaeologists

assign meanings to images on the basis of three methodologies which are employed either

together or separately. Since three alternative approaches exist, this section contains three

parts: (i) the intuitive approach, (ii) the analogical approach, and finally (iii) the homological

approach.

Section 1. The earliest investigation of these images

Mackenzie (1793), Schoolcraft (1851-1857), Bell (1879-80), and Lawson (1885)

recorded the presence of images either painted or engraved on rocks in the Canadian Shield

prior to the twentieth century. While examining the geology of Molson's Lake to the west of

Hudson Bay in 1879, Bell (1881: 7c) found a pictograph site on the north side ofPai-Musk­

taban Sipi near its mouth. The painted figures existed on a gneiss cliff that "measured up to

100 feet" (ibid.) in the form of "small figures in red ochre, said to have been painted by the

father of the present chief of the colony of lndians which removed a few a years ago from

Little Playgreen Lake to Fisher River, on the west side of Lake Winnipeg". The largest

images were "not more than one foot high" (ibid.), and most of them were not more than

eight inches. According to Bell, among them were "represented[ed] a boat, canoe, tortoise,

bird, deer, oUer, lndian, pipes, etc." He stated that the local lndians described them "as being

much more wonderful than they really are".

The geologist Lawson's (1885) publication was the first to consider the meaning of a

pictograph (DiKp-1) and a petroglyph (DiKo-2) site in the Lake of the Woods. It is difficult

to establish which rock image site was the first to be recorded in the Lake of the Woods area,

but Lawson's detailed descriptions and drawings of DiKp-1, and a nearby petroglyph site,

DiKo-2 probably hold that distinction.

• Site \inde!' diseuss! on

Figure 3.1: Map of the sites that Lawson recorded in 1885.

Page 127

,~

Lawson commented on the groupings of the images at the petroglyph site, DiKo-2 (figure

3.2) and posited that there was "little hope" of establishing any meaning or narrative because

the groups of images, or archaeological sites, were so isolated from each other (ibid.: 656).

· ~ , . ~K .. 1: HU

~ "

'>i 8 I~ 1.

Y r.- Ct, """ 4;fr~"'~

U " 'J< <1-'" · F';, 0 ... C<$ ~~

" 0 :D"t ,.":7l#'M;,S. ... " ~ • ~7

· • b .. t •

Figure 3.2: Lawson's published drawing of the images at DiKo-2, the petroglyph site.

Lawson compared the images that he described in DiKp-l and DiKo-2 to engravings

on large rocks found in the Alagoas, Brazil, described by John Brenner in a previous issue of

the American Naturalist (ibid.). Lawson also noted the similarity of the images at DiKo-2

with the painted images at the pictograph site (DiKp-l) painted on hard greenstone nearby

(figure 3.3).

Figure 3.3: Lawson's published drawing of the images at DiKp-1, the pictograph site from 1885.

Lawson concluded that the similarity between the engraved images and the painted

images suggested that they were "closely related in authorship." He was obviously interested

Page 128

in who might have created them, sin ce he concluded his paper with: "[1]he Indians of

present day have no traditions about these inscriptions beyond the supposition that they

must have been made by the "old people" of long ago" (ibid.: 657). However, if the

pictograph site in 2001 is compared with how Lawson drew it in 1885, it is clear that he did

not see or record the red images to the right and above the ones he published. The

photograph of DiKp-1 below demonstrates that what Lawson saw differed from what

existed in 2001. It is possible that the other images were painted later.

Figure 3.4: Annotated photograph of DiKp-1 taken in June 2001.

Section 2. More recent literature.

More recent examinations of rock image sites are published in a variety of places.

Sorne of these publications have summarised the discoveries of pictograph and petroglyph

sites in the Canadian Shield (Brenner 1994; Conway, T. 1979; Dewdney 1963, 1965, 1977,

1979a; Molyneaux 1977, 1980a; Nute 1948; Wellmann 1979a). A few have considered issues

such as tourism (Arsenault 1996) and political and ideological conflicts (Arsenault 1997;

Nelson and Hinshelwood 1998). Others have discussed trips that involved searching for,

finding or visiting, sometimes recording sites in the Canadian Shield, and also speculating on

the meanings of images found (Cameron and Cameron 1979; Dewdney 1958, I<nowlton

1992; Macfie 1992; Mallory 1961; OIson 1961; Pettipas 1991a). Sorne briefly discussed and

referenced rock image sites as being widespread throughout the Canadian Shield within a larger

discussion of sorne other related topie. One of this group, Dickason (1972: 13), used part of

DhKm-3 as her example of the pictographs of the Canadian Shield. Several archaeologists were

Page 129

determined to find the oidest pictographs or petroglyphs in the Canadian Shieid (pohorecky

& Jones 1968a; Steinbring 1993, 1999).

During the 1970s and the 1980s, archaeologists concerned with pictograph sites were

interested in two related issues: the selection of a place in the lands cape and the orientation

of the site. Reid (1980) observed from the information on the pictograph sites in the West

Patricia region north of the Lake of the Woods that pictograph sites tended to occur on

straight shorelines, or coastlines that were not broken into parts by bays, islands or points

and that the creators of these sites preferred rock faces that faced in a south-to-east

direction. Rajnovich (1980b: 34) developed this ide a further in her discussion of pictograph

sites in Cuttle Lake, arguing that straight shorelines were predominately selected "to increase

the spectacular effect" of the images because these shorelines enabled the images to be

visible both from nearby and from afar. She argued that, based on her examination of sites

from the Rainy River region (south of the Lake of the Woods), south-to-east facing rock

walls were selected as light conditions here were best for creating and seeing paintings. The

"rising and noontime sunsrune reflected from the water" beside the cliff face caused

"dancing light sparkles over the work heightening its dramatic value" (ibid.). She asserted

that cliff faces which faced west-to-north only experienced "shadows or the weakest sunset

light." She assumed that pictograph sites always occurred on a cliff face immediately beside a

body of water close enough to reflect light on any images painted upon the rock's surface.

Licence reports are useful for finding sites and establisrung the range of work

conducted at different sites in the Lake of the Woods (for example: Cameron 1982,

Molyneaux 1978, 1979, and 1981). Cameron (1989) described fieldwork activities in the Lake

of the Woods where she recorded and described the images at four petroglyph sites:

Kennedy Island, called DjKp-4, Sunset Channel, called DiKo-2, Tranquil Channel, called

DjKp-3, and Mud Portage, called DkKr-4. She referenced these sites only by their local

names and did not use their Borden numbers. Such practise makes trac king what sort of

archaeological fieldwork was conducted at each site and by whom difficult for subsequent

archaeologists.

Page 130

• Sites _der m!IeU!llsi61l

• Picmgra,ph shes • Petmglypb !J'tes

Figure 3.5: The four sites that Cameron recorded in 1978.

Molyneaux (1978) wrote in his report to the Ontario Heritage Foundation that his

intention was to create a photographie record of two petroglyph sites when the water level

was abnormally low. He noticed that Lawson (1885) had described the lake levels as varying

over a range of 10 feet (ibid.: 1) but that the lake's flow and therefore its level was now

artificially controlled. Molyneaux, like Cameron (1989), did not use Borden numbers, only

local names: Kennedy Island, or DjKp-4, and Sunset Channel, or DiKo-2. Molyneaux (1979)

for his subsequent study produced a detai!ed photographie record of six of the seven

pictograph sites in Whitefish Bay, in the south eastern portion of the Lake of the Woods:

DiKm-3, DiKm-4, DiKm-l, DhKm-3, DhKm-4, and DhKm-5. Molyneaux hoped to use

photography to record the images systematically, accurately, and efficiently given the

"difficult field conditions" (ibid.: 4) and also to minimise human error and interpretation.

Other reports by Molyneaux discuss sites in areas neighbouring the Lake of the

Woods (Molyneaux 1981, 1982, 1985a; 1985b). Both his 1974 and 1975 publications resulted

from projects conducted under the auspices of Trent University and the Canadian

Conservation Institute. The Trent University team headed by Molyneaux examined eleven

pictograph sites and twelve rock image sites in Northwestern Ontario. Eight of these sites

existed in the Lake of the Woods, including DiKm-3, DhKm-5, DhKm-l, DhKm-4, DhKn-

1, DiKo-l, DjKn-l, and the now extinct lichenoglyph site called DhKn-2. However, the

team considered only one Lake of the Woods site, DiKp-l on Cliff Island, in the 1975 study.

Still other studies only mentioned or briefly described pictograph and petroglyph

sites in the Lake of the Woods (Fox 1974; Reid 1976 and 1977). Reid surveyed numerous

archaeological sites, among which were fifteen pictograph sites. These sites were examined as

part of an inventory that was conducted at the request of the Regional Lands Co-ordinator,

Northwestern Region, Ministry of Natural Resources (1976: ü). The goal of this

Page 131

archaeological project was to coliect data for the Lake of the Woods General Land Use Plan

and its principal concern was the "dense st current and potential development within the

northern half of the Planning Area" (ibid.). A number of specific archaeological sites were

tested at the request of the Senior Lands Planner. A total of seventy-one archaeological sites

were recorded in May 1975. Ali the pictograph sites were aliotted Borden numbers and

assessed as to their level of "interpretability" to the public as weli as their level of

"representability" (ibid.: 13-15). What these terms meant is unclear. Reid provided the

precise geographical location of, and specified how to reach each site. He also postulated

their "cultural affiliation," essentialiy their age, and sought to determine whether the site had

any religious significance to the local indigenous population. The physical condition of each

site was evaluated, and Reid established whether the general public knew of its existence and

whether it had been discussed in print. He did not describe the images but photographed and

determined the size each site.

Fox (1974: 3) described only DhKm-1, a pictograph site calied Devil's Hole, in the

southern part of Whitefish Bay, Lake of the Woods. He observed that bundles of clothing

were stuffed into a cleft in the rock surface of this site. An informant told him of rumours

that these bundles had been "placed there by local native people attempting to affect a cure

for a sick child." He considered that perhaps the offerings of these bundles were closely

connected to the "offering tree" which Ritzenthaler (1970) had described as existing among

the southern Chippewa. Fox found similar items of clothing such as buttons, buckles, and

suspender components at this site which he presumed to be evidence of earlier offerings. He

also found lead sinkers, a corroded metal bowl, and numerous coins (1974: 3.). The presence

of these objects was recorded and they were left where they had been found. He speculated

that the clothing bundles had not been disturbed since they were left as offerings. Birch bark

scrolis were not found at this site, although fragments of birch bark with stitching punctures

were found one to two metres back in the large cleft at this site. Fox posited that these

fragments might have been part of a birch bark container used to hold food offerings and

sent these fragments to the Canadian Conservation Institute for preservation. Before Fox

finished recording this site, ali the adjacent clefts in the rock were checked and photographs

were taken (ibid.: 4).

Page 132

Figure 3.6: Fox's photograph of DhKm-1 in 1974.

Lambert (n.d.) used the same procedures as Reid had in 1975 (1976) to record eight

pictograph sites (DkKn-7, DkKn-6, DjKn-l, DhKm-5, DjKn-2, DhKm-18, DhKm-4, and

DhKm-l) in the Lake of the Woods. This report was never published. Ali the pictograph

sites he recorded in the Lake of the Woods were photographed, recorded using direct 'dot­

for-dot' tracing (ibid.: 20) and described in considerable depth. These sites were discussed in

the findings of the third Northwestern Ontario Rock Art Project, conducted in 1985-1986.

This project considered some of the pictograph sites in the Lake of the Woods. The other

reports dealt with pictograph sites elsewhere in Northwestern Ontario (Lambert 1983, 1985).

Lambert (n.d. 201) argued that classification of the images from the pictographs examined in

the 1985-1986 project permitted him to distinguish between sub-areas of the Winnipeg River

drainage basin and the Bloodvein River drainage basin (ibid.: ii). The abstract of his third

survey indicates that he viewed these images not as 'art' but as a method of communication.

Lambert (n.d.: 201) argued that the largest number of animal creatures occurred in

the Bloodvein River Region, north of the Winnipeg River drainage basin. He asserted that

his results indicated that sympathetic hunting magic was an important function of the rock

images in the Bloodvein drainage basin, since more images of creatures were present in this

region. In the Winnipeg River region, images of canoes and anthropomorphs occurred more

Page 133

often. He (ibid.: 199-200) observed that there were more hand prints, and pairs of handprints

in the Winnipeg River region than in the Bloodvein River drainage basin.

For two sites in the Bloodvein River region, Lambert (ibid.: 202) posited that a

connection existed between the orientation of the site and its function. Almost ail of the

images at both sites were hand prints and both had similar orientations. He suggested that

both sites "signposted a possible avenue, by way of a portage" to another lake (ibid.: 202).

The pictograph sites he cailed the Rushing River sites, DjKn-1 and DjKn-2, in the Lake of

the Woods region acted in a similar manner, since they both used to exist on rapids or fails

prior to the raising of the Lake of the Woods water level (ibid.). He argued that pictograph

sites had "a function ta guide the traveiler" especiaily when they occurred in narrow

"situations" since those site areas were difficult to avoid. This meant for Lambert (ibid.) that

these sites could be interpreted as "message boards, maps or navigation aids." Unfortunately

he was unclear whether the message was to "guide or identify, or warn of fast water or

malevolent spirits".

He (ibid.: 204) argued that the images he examined were related to various secular

activities such as hunting, fishing, and horticulture. Specifie images he argued (ibid.: 204-

205), although unstated, were connected to hunting ungulates and sturgeon, others indicated

a life-threatening disaster connected to horticulture, and that sorne images were connected to

the Midewiwin. It is evident, upon close examination of Lambert's discussion of DhKm-1

and DhKm-4, that he has included and treated the images which comprise the site DhKm-1

as part and parcel of DhKm-4.

Page 134

Faœl-f)2A ~~ $'

t \ el ~If

VI 1'.4/sm, -/"

5ITE#!J2A "()eyih /fal" P

os' Whitd'lsh l>~ Lake "f th<r bods

Q ",.

'l.f·~J""f"'. 1 i , \) i ) ."."

1 ~~ :S"

.,.t.rf • ...r N=Jf fi2tJ(Z" Dl auf JtlI1 IS; IJ.tJ

Figure 3.7: Dewdney's 15 July 1960 field record of Site # 92A, now called DhKm-1 and DhKm-4.

This may have occurred because Dewdney, in 1960, called DhKm-4 Site # 92 and

DhKm-1 Site # 92A. It is clear from his field drawings that the images on Faces II and III of

Site # 92A are part of the pictograph site subsequently classified as DhKm-1. Face I of Site

# 92A is now called DhKm-4.

Pictograph and petroglyph sites were frequently recorded as part of larger

archaeological surveys concerned with other types of archaeological sites in the Lake of the

Woods. One notable example in the Lake of the Woods is Pastershank's (1989) report, in

which she considered pictograph sites as separate units of analysis, therefore not to be

analysed in conjunction with the archaeological sites examined at the same time. These

studies, aimed to establish that sites existed and to describe them, did not analyse them in

detail. This treatment occurred elsewhere in the Canadian Shield (Conway 1975, 1984;

Dawson 1973; C. Halverson 1988; Hill 1982, 1983; Lemaître 1995; Smith 1981). Occasionally

archaeological surveys were conducted for the sole purpose of surveying for pictographs and

petroglyphs but not for "conventional" archaeological sites that could be excavated (Lambert

1983, 1985, and n.d.; Rusak 1992a; Scott 1980; Pelshea 1980; Pelleck 1981; Rajnovich 1980b,

1981 b). The quantity of literature published in other government reports indicates that

archaeologists were concerned with these images (Conway n. d. c; Friend 1983; Jones 1966,

1981a; Petch 1991; Pohorecky and Jones 1968b; Rajnovich 1980c, 1980d, 1981c; Rusak 1991,

1992b; Steinbring and Iwacha 1982; Steinbring 1998; Tassé 1977a). These sites would have

Page 135

been recorded and discussed at sorne length in the reports required by the relevant

government bodies to inventory and analyse the archaeological sites in a given region.

Section 3. Techniques used to record the pictographs.

Subsequent methods used to record and analyse rock images are derived from

Dewdney's work. Indeed, Dewdney is rightly considered to be the "father of Canadian rock

art studies" since he was the first person to undertake a comprehensive study of these

images. The initial method for determining the size of sites appears to have emerged as

Dewdney (Dewdney and I<idd 1962) developed rus techniques for sketching and measuring

pictograph sites.

The technical lite rature is divided into three parts: Ca) subdividing the sites into

sections; (b) describing the images; and Cc) recording the images.

a) Subdividing sites into sections

It is crucial to understand the problems involved in Dewdney's use of the terms 'face'

and sometimes 'panel'. If the area covered with ochre is large, sites are invariably subdivided

into smaller components to aid recording, photography, and description. The initial method

for determining the size of a site emerged as Dewdney developed rus techniques for

sketcrung and measuring pictograph sites.

5itfE #10' 5uns"r ChclrJlleI­Lake of thRWoodr Ailg./Z'S9

(perroglyplzs of 5" .. 1171 sbawn ds~wh.-n:)

rwa>'/!f

Figure 3.8: A field recording by Dewdney of two pictograph sites in the Lake of the Woods, Site # 70 Sunset Channel (DiKp-l) and Site # 69 Picture Rock Island (DhKm-3).

Page 136

Dewdney's sketches of the rock features of the different parts of a new site probably

enabled him to identify the physical location of various groups of images (ibid.). Yet

Dewdney did not identify different parts of a 'face' as 'panels' if they had different

orientations to each other. It is possible that this practise developed as a result of other

fieldworkers beginning to take compass bearings of rock faces. It was observed that the

surface of the rock used for paintings faces the rising or the setting sun or in a southerly

direction (ibid.: 6). Hence the pictograph sites that Dewdney examined faced towards the

east, the west, and the south. Indeed, Dewdney (ibid.) stated that he had "seen only three

sites on wruch the sun never shines" where lichen grew, but he did not state in wruch

direction the surface of the rock was faced.

Difficulties exist with the meaning of the terms 'face' and 'panel' since researchers

often do not clarify their meanings. Unlike Pastershank (1989), who used these terms but did

not define them, Rusak (1992: 1) defined a 'Face' as a single figure or group of figures on a

common rock plane. She argued that the "F" must have its first letter capitalised to

distinguish it from its common meaning. It appears that she used the concept of panels to

subdivide large sites, especially if two areas containing pictographs were physically separated

on the same face. These panels were numbered sequentially if the panels had different

orientations.

It seems likely that the concepts of faces and panels were used to divide the rock face

into surfaces that could be discussed and analysed easily. In rus unpublished report (n.d.: 21),

Lambert developed the ide a of splitting sites even further into manageable parts, if cliff faces

with paintings were "sufficiently large to warrant supplementing the standard Face and panel

designations with seriaI, horizontal data points." Lambert maintained that seriaI horizontal

points were useful if separations appeared on the rock face. He (ibid.) applied it at Bloodvein

1 Pictograph, EiKs-l, in the Bloodvein drainage basin, but maintained that trus technique

would have been useful in the 1982 and 1984 Northwestern Ontario Rock Art Project

(Lambert 1983, 1985) studies. However, he acknowledged that he had not discovered many

sites where there were "continuous paintings for long, relatively flat, clifflengths" (ibid.: 23).

While archaeologists use this technique to subdivide pictograph sites for the purpose

of making them easier to describe and quantify, it can cause problems for subsequent

researchers, since precise rules do not exist concerning how trus technique should be

employed. Twelve pictograph sites were divided into parts when they were first recorded.

Page 137

This group includes DiKm-3, DhKm-3, DhKm-1, DhKm-4, DiKn-1, DhKn-1, and DjKll-1.

Each site will be examined in sequence. It is clear that both Dewdney and Pastershank

invariably subdivided sites into parts if they were the first individuals to examine a site. It is

therefore easy to subdivide this group into two smaller groups according to whether

Dewdney or Pastershank recorded the site. Dewdney examined DiKm-3, DhKm-3, DhKm-

1, DhKm-4, DiKll-1, DhKll-1, and DjKll-1, while Pastershank examined DgKl-19 and

DgKl-17 first. However, Pastershank contradicted Dewdney's divisions of DgKl-1 and

DgKl-2 when she examined them in 1989. DhKm-5 will also be examined because, rather

than being subdivided, it grew in size.

DiKm-3:

Dewdney used the crack visible in the middle of the field drawing (figure 3.9) and

photograph (figure 3.10) below to differentiate between the two parts, or Faces, of the site

called Site # 28. tt.: ---

about thrre md6 S)E. of 510ux St=lôALSO:

Narrowr- t'fu: I"t<;fc/'*hing i-km~d"po{,ft,.dafth'sr;,~,I;'field Book2

thfT rv.sl.dlJt2 of Q 'Pf'4!:P" t'rlH-"Fou"d hli!t""i~ now il! f1I", ROMdfrnogroph{c:coIJochon

faC', l

"

ettiUSTOPHIR DIWON..,

Figure 3.9: Dewdney's field drawing of Site # 28, undertaken on 15 August 1953.

Dewdney's first Face is the left side of the rock tripe growing vertically across the

pictographs, which is evident on the extreme left hand side of the photograph below.

Page 138

Figure 3.10: Face 2 and 3 of Site # 28 by Dewdney in 1953.

Unfortunately, it is unclear whether Reid, who examined the site in 1975, and Molyneaux,

who photographed and examined it in 1979, used Dewdney's subdivisions in their analysis.

DhKn-l:

Dewdney subdivided this site into Face l and II probably using the large vertical

crevice that exits in the cliff face that separates the two clusters of paintings in figure 3.11.

r", fU1'~" '---

IJ'q~

18"'''

a6<W12'waftzr Auy,Ij.l5'i/{J .

!

Sire #/OG 06abikonNofTi:JW5, /VOFtt7@a'of ObobikonlJO'/{. S'outil of AU!I71?atl Pel7il7su(<1, Lake oft~ W'oodr

Figure 3.11: Dewdney's field drawing of Site # 106 undertaken on August 19 1960.

l t is not clear whether Reid in 1975 and Molyneaux who examined it in 1974 and

Page 139

1981 followed Dewdney's method of dividing this site into two parts.

DhKm-l and DhKm-4:

Dewdney's field drawing of Site #92 A and Site # 92 indicates that the two sites

DhKm-1 and DhKm-4 were considered one site (see figure 3.7). Dewdney found and

recorded both in 1960, classifying them as one site. He recorded only half of the encire site,

possibly because he did not think that smears of ochre were paincings. However, why

Dewdney only recorded the images from part of DhKm-1, which he called Site # 92,

remains unclear. Dewdney's Face l is DhKm-4, while Faces II and III are part of DhKm-1.

DhKm-4 is clearly not immediately beside DhKm-1 as Dewdney's field drawings (figure 3.7)

of both sites imply. However, why Dewdney only recorded the images from part of DhKm-

1, which he called Site # 92, is unclear. The sites are more than a few metres apart from each

other as the annotated photograph in figure 3.12 below indicates.

Figure 3.12: The Close Proximity of DhKm-4 to DhKm-l.

Dewdney missed an image now classified as a smear as indicated by the annotated

photograph and Dewdney's photograph of the site below.

Page 140

Figure 3.13: Annotated photograph of DhKm-1 in 2001 by Colson according to Dewdney's divisions.

Lambert, in his (1985) unpublished study of sorne of the pictograph sites, made the

mistake of considering sorne of the paintings of DhKm-1 as part of DhKm-4, despite Reid

(1976) in 1975 reporting that the two sites were separate entities with their own Borden

numbers. Why this confusion existed for Lambert is difficult to determine, since he would

have known that Reid had already designated DhKm-1 as a separate entity from DhKm-4.

His own records show that Lambert (n.d.) knew of Dewdney's 1973 publication and had

probably seen Dewdney's 1960 field drawings of DhKm-4 and DhKm-1. The images of

Lambert's Face II of DhKm-4 are images of DhKm-1 that he recorded, if one examines the

annotated photograph below in figure 3.15. Lambert recorded the images that he deemed

were the site. He applied a technique of classification developed during his field survey of

sorne pictograph sites in the Upper Severn region of Northwestern Ontario in 1982 (1983).

Page 141

Figure 3.14: Annotated photograph by Colson of DhKm-1 taken in 2001 according to Lambert's divisions.

DhKm-3:

Dewdney was the first person to record this site. He subdivided it into three sections

as seen below in rus field drawing (figure 3.16).

#69 P,cTure Rgd; /,/<7nd WI)!fefish Boy,Lake of the Woods Aug855'

16,50 Fow$

CHili STOPHtR otWONEY

Figure 3.15: Dewdney's field drawing of Site # 69 undertaken on 8 August 1959.

It is unclear whether Molyneaux in 1974 and 1979, Reid in 1975, or Pelshea in 1978

paid any attention to Dewdney's subdivisions of the paintings.

Page 142

DiKn-1:

According to the two field drawings below (figures 3.16 and 3.17), Dewdney divided

this site, initially called Site # 91, into five parts. Several of these parts were subdivided again .

• rth<:ilsttip of Ha'{A:r Penil/sale; '/letfn e/ltranœ toWlJtlefishfJar,ta/œoftil<?/Iiâ"ds :j

lB·

\ 'tl ;i' [~:~~I-::~ lB'

,7:1..170

Face 1

[ '\,

Facclb Fac,z Il/:'

CH~U STOPHER DEWONtV

Figure 3.16: Part 1 of Dewdney's field drawing of Site # 91A and B, undertaken on 15 July 1960.

wntinued)

/t(!'1'1IttÇ ,;'own"

OruVII wc/ur kw/

or; cr /u/,/ 15; INO

{ace V

'"

,,' ./,,,"'" "",du·

50' _ ~ __

Ct11\1 STOPHi'R DEWDNEY

Figure 3.17: Part 2 of Dewdney's field drawing of Site # 91A and B, undertaken on 15 July 1960.

Page 143

DjKn-1:

Dewdney recorded this site as Site # 29 in 1958. From examining two photographs

of trus site, rus field notes in New World Arcruves at the Royal Ontario Museum, and

Dewdneyand Kidd's publication (Dewdney and Kidd 1962: 27 and 1967: 29) it seems that it

contained two components.

Figure 3.18: Setting with offerings and C.D., Site # 29 by Dewdney.

The second part of the site was published in Dewdney and Kidd's publication

(Dewdney and Kidd 1962: 27 and 1967: 29).

Figure 3.19: Photograph of part of DjKn-1 called Site # 29 taken by Dewdney in 1959.

Unfortunately, it remains unclear whether Molyneaux in 1974, Reid in 1975, or

Pelshea in 1978 used Dewdney's method of dividing the site into two parts. However,

Page 144

Lambert (n.d.) considered the site as two separate parts, recording two parts of this site as

Face l and Face II. Rusak in 1993 expanded the site to include a new section which she

called Face III. This was the location of a hand print that existed higher up the same cliff as

part of the same site but not near the water's edge. Unfortunately, l was unable to ob tain

coptes of Rusak's photographs, so l have provided my photograph of this image (figure

3.20).

Figure 3.20: Part of DjKn-1, July 2001.

Dewdney recorded and examined the following two sites ln August 1964.

Pastershank challenged his subdivisions in 1989. Both sites are close to each other, but differ

in terms of size, and in what has occurred since Dewdney first, then Pastershank, examined

them.

DgKl-1:

The water below this site is shallow, and the site is partially protected from the main

channel by two of the three small, neighbouring islands. It lies on a south facing spit of land

on the main travel channel, between red marker S56 and S60 at the northern end of

Sabaskong Bay. Dewdney identified two groups of images which he called Face l and II of

Site # 197 (figure 3.21).

Page 145

/lit orm d' 5a&/$konq i3aq, Cqk~ orthe Wood~

Sd!(~ #('/7 uod#198

"<" ... :, '!.i.... r'"

-. ~1 q~~~~~~~~~.'~~'~_~'~_~~" __ ' __ ~_~A_

'\ CMfUSTOPHIft lIlWf)NW

Figure 3.21: Dewdney's 1964 field drawing of Site # 197 (DgK1-1) and # 198 (DgK1-2).

Pastershank in 1989 (1989: 51) contradicted his method of subdividing the images. She

found images on a surface to the left of his Face 1 and, hence, renamed his Face 1 as her Face

II and his Face II as her Face III. The 'new' images to the left of her Face II were cal1ed Face

1. The site was clearly subdivided as indicated in the photograph in figure 3.22, into three

parts, cal1ed Faces, based upon the physical structure of the cliff face.

Figure 3.22: Photograph of DgK1-1 in July 2001 with Pastershank's 1989 subdivisions.

Page 146

DgKl-2:

Pastershank disagreed with Dewdney's subdivision of DgKl-2 when she recorded it.

Her divisions, unlike Dewdney's of 1964, did not depend on the identification of specifie

images, which she termed motifs, of DgKl-2.

Figure 3.23: DgKl-2 in 1989 hy Pastershank.

DgKl-2 is the larger of the two sites that Dewdney examined. The decisions

Pastershank made in subdividing it in 1989, when she recorded the site again with her crew,

are harder to understand and replicate than with DgKl-1. Dewdney recorded the part of

DgKl-2 called Face 1 by Pastershank in 1964, but since then the number of paintings present

at this site had increased in number. The image in figure 3.24 is the image that Dewdney

recorded at DgI<J-1 although it was published 90 degrees counter clockwise in Dewdney's

1962 publication with Kidd.

Page 147

Sabaskong Bay, Site #198

Figure 3.24: Dewdney's published images from this site.

l also discovered four additional and obviously new images at DgKl-2 in 2001 (figure

3.25). The presence of these images reflects the issue that Pastershank's subdivisions of the

site may have affected her description of the site. Therefore, it was important to replicate

Pastershank's subdivisions of DgKl-2.

Figure 3.25: Three of the four new images at DgK1-2 in July 2001.

Unfortunately, almost ail the images Pastershank identified, except for two

recognisable handprints and sorne parallel lines, were very pale and next to impossible to

distinguish. The disappearance or fading of the images recorded by Pastershank in 1989

could be a consequence of the exfoliation of the images, possibly in conjunction with the

deposition of mineraIs. Mineral deposition often causes a white deposit to form on images

painted using ochre.

Page 148

The new paintings found in 2001, in figure 3.25 above, were produced using modern

paint instead of ochre. The new paintings including an type of creature, a six pointed star and

a vertical crescent with its points facing right, were both superimposed on older paintings

and painted directly on the rock surface beside paintings Pastershank (1989:54-63) had

described, recorded, and photographed. The fourth new image was created not on any of the

designated "Faces" but was placed beside the area called a "site" by Pastershank. This image

is radically different from any of the new images since it is hollow rather than solid. Large

quantities of offerings also were discovered in the 2001 field season. .

Figure 3.26: The new image at DgKl-2 not painted upon the older ochre images in July 2001.

Pastershank (1989) designated Face I as being at right angles to the rest of the site on

the right hand side of figure 3.27. The large overhang was divided into Faces V, VI, and VII.

The rock underneath the overhang was divided into Faces II, III, and IV.

Page 149

Figure 3.27: Annotation by Coison in 2004 of DgKl-2 taken by Pastershank in 1989.

Face 1, first recorded by Dewdney in 1964, is relatively easy to relocate from its clear

images and Pastershank's published field recording in figure 3.28 below.

Figure 3.28: Field recording of detail of Face l, DgKl-2, by Pastershank in 1989.

Page 150

Figure 3.29: Photograph of Face 1 by Pastershank in 1989.

However, the remaining Faces are not allocated to the surface of the rock in a way

that can be identified easily, since Pastershank described their location on the rock surface in

relation to specifie morphs that she identified. Unfortunately, by 2001, these morphs were no

longer recognisable. Consequendy, it was difficult to determine the precise location of each

Face. Her published report designated Face II as the area of rock underneath the overhang

where she observed "an overpowering red ochre smear" (ibid.: 55). The image below is her

field recording of Face II (ibid.: 57). Pastershank argued that, when she recorded this "Face",

the clearest morph was a "zoomorphic bird (Morph 4)" found at the bottom corner on the

north side of the panel (ibid.).

Page 151

~ ...

Figure 3.30: Annotation by Colson in 2004 ofphotograph of Face II by Pastershank in 1989.

Figure 3.31: Field drawing with detail of Morph 4 at DgK1-2 by Pastershank in 1989.

Unfortunately, by 2001 this morph, which she called a "thunderbird," was not as

bright or distinct as Pastershank had described it. Consequendy, it was difficult to establish

its precise location. Pastershank argued that the base of this bird image, was "in sorne danger,

as it is sitting on the top of the middle of the roof-laminating caverns" shown in figure 3.31.

Page 152

The bird image she is concerned with is the arca of concentrated dots in figure 3.31. She

asserted that distinct handprints, called Morphs 5, 6, 8, and 10, were also on Face II. Three

"finger strokes" were recorded as existing in the central area of the red ochre smear cailed

Morph 7 (ibid.: 59). Pastershank found fresh tobacco left as an offering, in a smail crack

fifty-three centimetres below Morph 10. She stated that trus part of the site suffered from the

largest quantity of rock spalling. There was crumbling over the whole site and smail colonies

of lichen were growing there. She argued that mineraI seepage was evident both above and at

the sides of the red smear and throughout the he ad of the bird morph (Morph 4).

10cm

+ 1m22ahw 2m30aw.

Figure 3.32: Field recording of detail of Face III by Pastershank in 1989.

Pastershank maintained that Face III was located twenty-eight centimetres left of

Morph 10, and forty-seven centimetres below it (see figure 3.32). The only recognisable

morph on trus Face was a "possible hand print" called Morph 11 (ibid.: 59). However, by

2001 it was difficult to identify the morphs that Pastershank had used in 1989 to subdivide

the site. Pastershank (ibid.) located Face IV farther south, i.e. below ail of the Faces that she

had already designated as protected by the roof of the shelter from any "weathering agents."

Page 153

10cm

Figure 3.33: Field recording of Face IV of DgK1-2 by Pastershank in 1989.

Several morphs were identified, including the outline of an image she described as an

"upside-down man with his elbow bent on the right as if to place one's hand on this rup"

(ibid.) (figures 3.33 and 3.34). She maintained that the depiction of trus image was indicative

of "its importance and possibly predicts the death of some-one prestigious" but,

unfortunately, did not provide any evidence to substantiate her conclusion. Pastershank

identified another morph, a smear that apparendy measured "the size of a handprint" to the

left of the upside down stick figure. She identified nine handprints as existing at trus site and

maintained (ibid.: 63), as Lambert (1985:123-125) had done, that the hand was "an indicator

of death of a warrior." She argued that farther left of trus smear was a similar sized but

"indistinguishable" smear of red ochre. Pastershank photographed the same part of the site

where Morph 12 was found. However, by 2001 this morph was no longer distinguishable.

Page 154

Figure 3.34: DgKl-2, Detail of Face IV by Pastershank in 1989.

Faces V, VI, VII were stated to exist on the overhang created by the rock shelter.

These faces were contiguous and the morphs were adjacent to each other on the curved

overhang. According to Pastershank, Faces V and VI had one hand print each, called

Morphs 14 and 15 respectively, while Face VII had a horizontal undulating red line

measuring twenty-seven centimetres, called Morph 16 (ibid.) (figure 3.35).

Figure 3.35: Pastershank's Field recordings of Faces V, VI, and VII in 1989.

Pastershank observed that ail three faces suffered from mineraI deposition which she

argued was indicative of heavy water leaching that was caused by precipitation and lichen

growing above the red images.

It was difficult to utilise the arbitrary framework that Pastershank used to divide

DgI<1-2 into manageable units. Pastershank's subdivisions superimposed on the photographs

in figure 3.36 do not foilow the physical structure of the rock but depend on the

identification of specific morphs. Furthermore, problems exist for subsequent archaeologists

Page 155

Slnce the images fade and decay. The deterioration of the morphs underneath the rock

shelter made it impossible to identify the precise boundaries between Faces II, III, and IV.

Figure 3.36: Using Pastershank's (1989) subdivisions of rock surface to divide the paintings in 2001 into Faces V, VI, and VII.

Some of the difficulties in establishing the precise location of Pastershank's morphs

may be the results of the techniques by which they were reproduced for the final

archaeological report. The published field recordings are the final product of the whole

Page 156

~ .. 1

process to which the field recordings were subjected, once they had been taken to the

archaeological laboratory. The field copies of the sites were subsequendy traced onto acid

free albene paper in the laboratory. The field copies were reduced considerably in size using a

process called photo-mechanical transfer, so that they could be readily published using 8.5 x

11 inch paper. Therefore, it is possible that tiny details, which enabled Pastershank to readily

identify morphs, became difficult to determine.

Pastershank interpreted DgKl-2 as follows:

The environmental setting and the extensive ochre red paintings have made this site overpowering. Death may be the subject of Face IV as seen in the form of an upside down man, positioned over the third cavern, perhaps the death of an important person. A total of nine handprints were recorded at this pictograph site. The interpretation of the hand being an indicator of death may be valid (Lambert 1985: 123-125). To make this site even more spectacular is the placement of the powerful spiritual figure, the Thunderbird, over the middle cavern.

Pastershank (1989: 63)

Pastershank clearly considered the location of the images in this cavern-like setting as highly

significant. Her naming this site "Three Caverns Pictograph Site" (ibid.: 54) seems to reflect

this fa ct. Yet, only two caverns existed at this site since the second "entrance" is in reality a

hole in the wall of the larger of the two caverns. Pastershank (ibid.: 70) identified this site,

and the three other sites in Sabaskong Bay, as "tentatively called 'Boreal Forest' Aigonkian."

She argued that they were dated circa 1600, which she maintained was "characteristic" of the

majority of the rock image sites in the Lake of the Woods. This is an interesting conclusion

but she does not provide any evidence to support it.

DgKl-17:

Pastershank recorded the site in 1989 and subdivided it into three parts designated

Faces l, II, and III. Face l did not exist nearby Face II and III, which are beside each other.

Pastershank (1989: 63) described Faces II and III as "approximately seven metres away from

the eastern start of the bedrock cliff and Face l is an additional 5.4 metres away just from a

southward bend." She stated that the images at this site occurred in two "components"

(ibid.: 67). The first group, or component, if her word is used, are in the photograph below

in figure 3.37.

Page 157

Figure 3.37: Face 1 of DgK1-17 by Pastershank in 1989.

The images in this part of the site are quite faint, although she described them in

sorne detail. She asserted that this part of the site had "paired horned anthropomorphs, and

one hand print" (ibid.). Pastershank divided the images of her second component at this site

into two parts which she called Face II and Face III, although they are immediately beside

each other. Unfortunately, a photograph does not exist of this part of the site, nevertheless

her field recording in figure 3.38 provides sorne indication of the images present in

Pastershank's Face II and Face III.

Page 158

Figure 3.38: DgK1-17 - Pastershank's Face II and II in 1989.

No-one recorded this site between 1989 and 2001.

DgK1-19:

Pastershank (ibid.: 67) described this site as "only 18 metres west of DgKI-17" but,

because the images had "different styles, elevations, and hues of pigmentation," this site was

"not contemporaneous" with DgKl-17. She (ibid.) asserted that this site could be divided

into two smal1er parts since sorne of the images existed on a "perpendicular, south-facing

hiatus ofbedrock," as shown in the field recording in figure 3.39.

!il!' (,

Figure 3.39: Pastershank divided this Face into two parts called Panel i and ii in 1989.

Page 159

. ,r-

DhKm-5:

It appears that this site was considered as a single entity and not subdivided into

several parts by any of the archaeologists who recorded it including Dewdney in 1960, Reid

in 1975, Molyneaux in 1975 and 1979, Pelshea in 1978, and Lambert in 1985.

Figure 3.40: DhKm-5 in June 2001.

However, when l visited this site in the 1991 with Rusak, an archaeologist who at the

time was conducting fieldwork in the White Otter Lake/Turtle River system (1992a), she

pointed out that DhKm-5, in figure 3.40, consisted of not only the rock face recorded by

Dewdney and his subsequent colleagues, but an additional part in figure 3.41 that existed to

the left of the rock surface that was already termed DhKm-5 .

Page 160

Figure 3.41: The new part of DhKm-5.

Clearly, if a site could be subdivided into smailer units, faces and panels it could also

be increased.

b) Terminology

Pastershank's (1989) identification of specific morphs at DgI(J-2 inadvertendy

caused her descriptions to become blurred with low-Ievel interpretations. Identifying the

images was a crucial part of Pastershank's description of the sites and her subdivisions

enabled her description of each site to have a structure and become manageable. The

boundaries for each Face were dependant on her ability to recognise specific images that she

labeiled as morphs. Yet, by 2001 many of the images that she had described in 1989 were

impossible to identify.

The vocabulary used to describe the images of pictograph sites is important, since it

enables the archaeologist to classify the images, and to create an inventory of ail the images

at each site. The vocabulary used to describe each image influences how it is examined, both

as a separate entity and within a group. The words used to describe an image can and do

influence how the shape and ultimately the style is considered. Archaeologists are interested

in style because different modes of depiction are often equated with different ethnic groups,

while similarities are interpreted as evidence of those interactions.

Page 161

Pastershank's (1989) classification of all of the readily identifiable morphs on each

Face enabled her ta group them by type, and by whether they were hollow or not. In doing

this, she followed the technique of analysing these ochre images advocated by Dewdney

(Dewdney and Kidd 1962, 1967), Vastokas and Vastokas (1973), Lambert (1983, 1985 and

n.d.) and Rajnovich (1980b, 1980c, 1980d, 1981a, 1989, and 1994). Pastershank (1989: 70),

who was trained by Rajnovich, determined in her archaeological survey of four pictograph

sites in Sabaskong Bay that 76 identifiable morphs were found on 15 Faces. DgKI-1 had 26

of ail of the identifiable morphs found at these four sites, while DgKI-2 had 17, DgKI-17 had

30, and DgKI-19 had 3 morphs. Unfortunately, she did not describe the shapes of these

images. The names she gave to different morphs were also problematic.

Pastershank's verbal descriptions of each type of morph cause problems because they

imply meaning to the reader according to their physical shape. As Dewdney (1979b) had

noted earlier, this problem was widespread amongst archaeologists and others interested in

these images. He observed that numerous difficulties ensued when archaeologists described

visual images verbaily and admitted that his own work suffered from these flaws and

problems (1979b: 326). In one case, he had argued that at over half the sites he had recorded

images that bore:

no recognizable likeness to any known form and l designate them as abstractions. Many of them are single strokes occurring in groups or series that suggest tally marks. The remainder range from simple to relatively complex forms.

The other half of the symbols subdivided roughly into five groups: miscellaneous man-made objects, hand-prints, other human subject matter, animaIs, and composite-presumably mythological-creatures.

Dewdney and Kidd (1967: 18)

Dewdney (1979b) realised that verbal descriptions of images were prone to subjective

variation (ibid.: 326-328). He (ibid.) subsequently suggested that a series of graphic and visual

models could provide the level of rigour needed to classify images because words were

insufficient. Dewdney argued, drawing upon a number of examples from the literature on

rock image sites from both the Canadian Shield and elsewhere in North America, that he and

his colleagues had been careless "in assigning captions to our illustrations" (ibid.: 330) in the

publications of their research on rock image sites.

He (ibid.: 333-336) also questioned the manner ln which his coileagues had

reproduced rock image sites. He harshly criticised the practice of examining these images one

at a time, often extracting a single image and connecting it with similar images from another

Page 162

~ ..

group. He insisted that images which formed a group should be examined in conjunction

with each other as weil as in relation to their original location, and that a site must be studied

in relation to its local environment. Not doing so was the equivalent of treating sites and

images at "the level of an artifact picked up in a ploughed field."

Unfortunately, Dewdney's (1979b) hopes were not realised quickly. Sorne

archaeologists conducting field surveys in the Boreal Forest have continued to define the

images at rock image sites over-narrowly. For example, Haywood (1983) wrote that thirty

pictograph sites depicting "moose, caribou, deer, heron, human figures, canoes, handprints

and abstract symbols" (ibid. 17) existed in Quetico Provincial Park. He was concerned with

describing to someone who was not visiting each site what the images resembled. ]. Conway

(1979: 19-32) labeiled a range of images at several sites she recorded in northern Ontario as

including: "thunderbird," "abstract paintings," "beetle," "turtle man," "taily marks," "human

figure," and "fish." Both the perception and value judgements of the archaeologist have

affected the description of these images. Dewdney (1979b: 336-337) concluded that those

concerned with rock image sites should develop new recording standards independent from

the conventional methods of attributing meaning to images.

(c) The development of the techniques used to record the images

The development of how the images were recorded commenced with Dewdney in

the 1950s (Dewdney and Kidd 1962). After he had sketched the images, Dewdney measured

the distances between the physical places on the rock surface where the images existed and

designated them as individual parts using Roman numerals (1962: 8-9). He took colour

photographs, and recorded the compass bearing of the site, the depth of the water nearby,

and the height of the cliff (ibid.: 8-9). At first, a three-inch grid was lightly chalked on to the

rock to ensure an accurate scale copy of the images at each site (ibid.). He washed the chalk

off afterward. This grid enabled Dewdney to determine the location of a key point on a rock

face. The distances to key points in a painting were measured from string co-ordinates across

the rock face. The use of thin ]apanese rice paper employed for block printing, which

became transparent when it was wet, soon superseded this technique, as visible in the

photograph below.

Page 163

Figure 3.42: Dewdney working with a colleague on an improvised platform at Site #108.

Sometimes Dewdney built physical structures beside the site to enable him to record

them. The photograph in figure 3.42 demonstrates that Dewdney used a log probably

attached by rope to a tree growing above the cliff face at DhKo-1. Two rectangular shaped

grey areas exist beside and in the middle of the images. The one in the middle of the

paintings immediately in front of the standing individual is probably rice paper. The nature

of the second rectangular shape is unknown and appears perhaps to have been applied to

coyer something up.

Dewdney used a sponge roller to wet the paper, which also caused it to adhere to the

rock surface. The images below and visible through the paper were subsequently recorded on

the rice paper using Conté chalk allowing him to record all but the faintest of images.

Dewdney also made notes directly on the paper regarding cracks and the height of the

images above the water level. This development permitted Dewdney to reduce the time

taken in recording these images. However, as Dewdney noticed, the rice paper both

obscured the recorder's view of the fainter images and lighter details of the stronger coloured

motifs, especially if the stronger images were superimposed on larger, fainter images.

Dewdney's site numbering sequence reflected the order in which he recorded the site rather

than anything else. He argued that, as he recorded, his accuracy increased "due to practice"

(ibid.: 9).

Pelshea (1979: 24) expanded the range of colours used so that red recorded ochre

and green was used for lichen, frost spalls, and rock level changes. He also started to replace

Page 164

or substitute the Conté chalk with Dixon Lumber crayons since they did not cause the wet

rice paper to tear.

Jones (1968: 110) believed that the images should be outlined using felt pens, rather

than duplicated using a method called "the solid-line technique." The latter resulted in the

"character and the distinctiveness" of rock images in different geographical regions being

omitted. According to Jones (ibid.: 110), solid-line recording was the most common and

widespread technique for recording pictographs. It was no t, however, a good method for

"conveying actual appearance of painted symbols, and, as importantly, the char acter (his italics]

and distinctiveness of the rock art of a particular area." Jones was concerned that key aspects

of the images of different regions were being ignored and simplified, since that made "the

symbols more subject to gratuitous classification" and "interpretation." He subsequently

advocated that the best alternative was to record every aspect of these "symbols and their

settings that set them off from other styles" (ibid.: 111). Evidently, according to Jones a

recording technique had to render any distinctions in the style of the images evident in the

field recordings.

According to Jones (ibid.), Dewdney initiated the method of recording whereby the

edges of images were outlined. Jones either crosshatched or shaded the interiors of the

images to indicate gaps in the pigments of faded areas. Pictographs were recorded on

transparent plastic using different coloured pens to differentiate between the pictographs and

lichen encroachment and damage to the rock surface. Jones advocated that notes should be

made on the plastic where spalling, differential fading of the images, and unusual features

existed. Later in the laboratory, a stencil copy was made of this field recording using a light

celluloid paper and inking equipment. The resulting image was checked against photographs

so that the colours could be verified, ensuring a good reproduction of the original image. The

technique of Jones' recording as many precise details of these images as possible likely led to

the development of the dot-for-dot technique.

Questions regarding image reproduction had already been raised in the minds of

several archaeologists. Since Dewdney was aware that rice paper prevented faint paintings

from being discerned, he quickly adopted Jones' (1968) and Pohorecky's (1968) refinement

of the technique, using Saran Wrap (thin clear clinging plastic usually used for wrapping le ft­

over food) that was electrified with a brush, and grease or felt pens for tracing the images. It

is possible that Pohorecky (1968) advocated Saran Wrap because of its tendencies to stick or

Page 165

"ding" to surfaces. This new recording technique did not obscure the images, enabling

recorders to salvage faint images invisible under the wet rice paper. The water required to

adhere the rice paper to the rock surface possibly also reduced future visibility of the

pictographs, since it left mineraI deposits on the surface of the site. However, Steinbring and

Elias (1968a) questioned Dewdney's rapid adoption of this technique. They argued that,

sometimes, it was difficult to record a site using Saran Wrap, and that, in these instances,

field researchers should record the images using metre grids. The representational advantage

of the dot-for-dot recording technique is evident, given that a number of researchers on

pictograph sites in Northwestern Ontario (see C. Halverson 1988; Lambert 1983, 1985, n. d;

Rusak 1992a), including Pastershank (1989) used it for several decades. EssentiaIly each dot

of ochre visible on the rock surface was recorded by a dot, using a red felt tip pen, upon

clear plastic that was held by sticky tape to the rock face. Lambert (n.d.: 20) asserted that this

technique enabled the "duplication of the cultural, as weIl as the equaIly important non­

cultural information." He argued that recording both the ochre and the colour-coded "linear

and areal symbolics" conveyed "contextual associations for both types of information." This

type of record, he asserted, enabled archaeologists to provide more detailed tmage

descriptions and analyses of specific paintings. Lambert (ibid.) argued that the plastic

recording was a good idea and, since it used a 1:1 ratio, rendered "an almost exact, complete

reproduction." He maintained that this record was important for preservation purposes

because the pictographs that he examined suffered from erosion.

However, changes in pigment could not be replicated easily using the dot-for-dot

technique. Superimposed images were recorded with dots positioned close to each other if

they were dark, or, if the ochre was faint, relatively far apart. The relative distancing and

spacing of the dots appears to have been at the discretion of the recorder. Rusak advocated

that the 2 millimetre thick plastic be cut to size, and fastened to the rock face with tape.

Whatever was visible below Rusak reproduced immediately above it on the plastic using

waterproof felt-tip pens (Rusak 1992a). Rusak, as weIl as Lambert, did not use Saran Wrap

but did not inform the reader why it was not used. Red denoted ochre and blue, minerai

precipitation and patination. Any organic matter, such as lichen, rock tripe and moss

encroachment, was recorded using a green felt tip pen, while a black felt tip pen recorded

cracks and fissures on the rock surface (ibid) (figure 3.43). Pohorecky and Jones (1966: 104)

coded their felt tip pens according to colours on the MunseIl soil colour chart.

Page 166

Figure 3.43: The plastic being taped to the surface of the cliff face and the mineraI precipitation and patination being recorded by a black pen by Rusak in 1991.

Pastershank did not darify which colours were used to denote which features for recording

the pictograph sites in Sabaskong Bay, Lake of the Woods, but it appears from her 1989

report that she used the dot-for-dot recording technique. It is probable that Pastershank

(1989) used the same colours as Rusak (1992a), because Pastershank trained Rusak as a crew

member for the archaeological survey during which the four pictographs sites in Sabaskong

Bay were recorded. Field recordings, created using this technique, were once again traced

onto acid free albene paper in the laboratory and reduced in size considerably so that they

could readily be printed on 8.5 x 11 inch paper by a process called photo-mechanical transfer

(Rusak 1992a: 23).

This process of physically recording these images is problematic despite Pohorecky

and Jones's (1966) daim that this technique required "no artistic skill" and allowed relatively

few opportunities for the recorder to make mistakes. Both the field recording process and

the creation of a duplicate in a laboratory are laborious. Mistakes can accur if the recorder or

a copier is a novice, since the image produced could reflect the impression of the images in

their minds rather than being an accurate reproduction of the images. The recording of faint

and superimposed images is an especially difficult and possibly subjective activity.

Superimposition of images occurs when images are painted one top of another at different

Page 167

times and consequendy the colour of the image varies in intensity. During the intervals

between repainting, these images could have faded, weathered, or become partially covered

in lichen, rock tripe, or mineraI deposits. Photography provides a detailed record of the

images themselves. Dewdney (Dewdney and Kidd 1962:8-9) took colour photographs of the

sites he described. Only Molyneaux (1979) considered the use of the photographie record as

the sole means of documenting the pictograph sites of the Lake of the Woods. Others

continued to use black and white and colour photographs and slides as additional means of

recording the images (see Pastershank 1989 and Reid 1975). Still other researchers working

on sites elsewhere in the Canadian Shield maintained that photography was too problematic

because photographs were often "distorted" as a consequence of "many factors". They did

not list these factors, other than to state that one of them was the distance of the camera

from the site. It was for these unstated reasons that Pohorecky and Jones considered detailed

field recordings and tracing of the images more important than photographs, whieh,

nevertheless, were important for checking details of the images in question. Pohorecky and

Jones (1967: 305) argued that photography "allowed too much artistic licence" and "too litde

accurate draughtsmanship."

Section 4: T echnical analysis

(a) Technical issues involved in recording and interpreting rock image sites

Sorne archaeologists have discussed the technical issues involved in recording and

interpreting pictographs and petroglyphs (Brand 1979; Chisholm 1982; Pelshea 1979;

Pohorecky and Jones 1966, 1967; Pohorecky 1981; Pufahl 1990; Steinbring and Elias 1968a).

Pufahl (1990) considered that computer enhancement of the images from the Eagle Lake

area was a good idea because he had been forbidden, by the EIders and the Chief of the

Eagle Lake Reserve nearby to touch or trace the images at each local rock image site. Yet

Pufahl was aware that both the EIders and the Chief were concerned about the severe

physical deterioration of these sites because they were sacred to the local indigenous

community and "routinely receive[d] offerings of tobacco" (ibid. 19). Consequendy

photographs were the oilly means allowed by which these sites could be recorded.

Pufahl's initial efforts at using photography were thwarted because of slow shutter

speed, insufficient naturallight, and the need to take all the photographs from a boat. Pufahl

resolved to use polarizing filters with both his camera lens and the flash unit in conjunction

Page 168

with graphically enhancing the images on the computer. He proposed replacing those pixels

where the lichen had encroached upon red images with matching red ochres from the same

area (ibid.: 20). This meant that the image became completely visible instead of partially

obscured. He was aware that potential problems could occur if this process substantially

distorted these images. Pufahl was not only enhancing these images but also manipulating

them. In "replacing" the grey or non-red bits of images, he perhaps unconsciously changed

their shapes. Therefore his daim that his was a radical breakthrough technique for recording

images is slightly misleading.

(b) Conservation and analytical techniques

The concern of sorne researchers lay solely in the conservation of pictograph sites

(Tassé 1979); others desired to conserve petroglyph sites, particularly the petroglyphs in

Petroglyph Provincial Park in southern Ontario (Bahn, Bednarik and Steinbring 1996;

Wainwright, Sears and Michalski 1997). Molyneaux (1974) and a te am from the Department

of Anthropology at Trent University visited twelve pictograph sites in Northwestern Ontario

du ring the 1974 field season to document them and determine the general condition of each

site using "previously defined weathering variables" (ibid.: 1). The reader of this report is not

provided with any means of determining what these variables were, but they dearly were

important enough to warrant a detailed description of the sites for conservation and

protection purposes. Occasionally he described the images that he found at each of the sites.

Seven of the twelve pictograph sites covered in the 1974 study were in the Lake of the

Woods region. The team examined the following Lake of the Woods sites: DiKm-4, DhKm-

5, DhKm-4, DhKm-1, DhKn-1, DhKn-2, DhKo-1, DjKn-1, and DiKn-1. All of the sites

discussed were referenced using Selwyn Dewdney's numbers and their Borden numbers.

A year later, Molyneaux and his colleagues (1975) from the Canadian Conservation

Institute and the Department of Anthropology, Trent University, sought to document in

detail fifteen pictograph sites in the Canadian Shield in terms of physical conditions such as

moi sture access, seepage deposit formation, seepage deposit flaking, exfoliation, and

biological growth. They assessed each site to establish its physical condition with regard to

these criteria. Topographical maps illustrating site locations and scale drawings of the

pictographs were also recorded for each site. One site, DiKp-1, called Sunset Channel in

figure 3.44, in the Lake of the Woods was induded in this study.

Page 169

Figure 3.44: DiKp-1, Sunset Channel, 25th July 1975

The site DiKp-1 was assessed as follows:

The most significant factor affecting the condition of the site appears to be the heaving undercutting of the face by a fracture. The resulting instability will ultimately lead to the destruction of the site when the slab fractures away from the cliff face. Other weathering factors have a moderate effect, although a progressive decrease in pigment density appears to have occurred. The deterioration of the lower sections of the morphs 1 c and 1 e may be related to their proportional1y greater accessibility to wave splash and high water submersion.

(Molyneaux 1975: 4)

Other studies were concerned with the conservation of the paints used ln

pictographs, the conservation of pictograph sites in the Canadian Shield, and the different

ways of examining them (Myers and Taylor 1974; Taylor, Myers and Wainwright 1974, 1975;

Wainwright 1990, 1997; Wainwright and Taylor 1977).

Steinbring and Elias (1968a) advocated establishing the pH level of the mineraI

washes on the surface of pictographs to as certain whether the acidity of the soil above the

site affected the images below. They maintained that the acidity of the soil could affect the

visibility of the images.

Goods (1992) analysed the paint of one morph from the Rice River Pictograph Site in

Manitoba to determine the micrographical, elemental and mineralogical composition, the

constituents, and the paint binder. He discovered the presence of hematite, quartz, potassium

oxide, and aluminium oxide. He argued that organic matter was not detectable either in the

vehicle or the binder.

Page 170

Two additional studies were conducted to discover the composition of the paint, so

as to aid the conservation of the rock image sites in the Canadian Shield. Establishing the

nature of the paint aided conservation since the archaeologists would know what to expect in

the way of deterioration. The discovery of the paint recipes and the organic binders could

also have important consequences for archaeologists investigating pictographs world-wide

regarding their conservation, meaning, and the motives behind their creation. Archaeologists

involved in these experiments argued that the results demonstrated that the artists who

created these paints probably thought about what they were doing (Buisson et. al. 1989;

Clottes et. al. 1990a, 1990b; Clottes 1993; Couraud 1983, 1988; Leroi-Gourhan 1982:84-85;

Lorblanchet et. al. 1990; Menu and Walter 1991; Pepe et. al. 1991). An important implication

of these studies is how the results affect the procedures taken by conservators.

Taylor, Meyer, and Wainwright (1974: 32) wished to establish the causes of the

natural deterioration of pictograph sites in the Canadian Shield, prior to devising any

conservation treatments. Previous researchers had attributed the deterioration of these

paintings to physical and chemical weathering of the rock surfaces. Seven samples were

examined to determine the formation and the composition of the paint from the Agawa

pictograph site, on Lake Superior in Ontario, and a site on Lac Wapizonke, Quebec (ibid.).

Litcle information is provided as to the sampling strategy used and the size of the samples

removed. X-ray diffraction was used with a Debye-Scherrer camera equipped with a

Gandolfi attachment for single particle analysis. Cross-sectioned samples were examined

using a scanning-electron microscope equipped with an x-ray energy analyser under

magnifications of up to Xl0, 000 (ibid.: 33). The results of these experiments revealed that

the paint was anhydrous iron oxide. They observed that the pigment was "firmly attached to

and intimately mixed with the [white] mineraI deposit in a sandwich structure" instead of

adhering to the surface of the rock (ibid.: 40). The white deposit found on these cliffs coats

the images and provides a "protective coating from the leaching action of the rain and

erosion by wind" (ibid.) as well as preventing the pigment from dissolving away. Taylor and

his colleagues also argued that the graduaI accumulation of this white mineraI deposit on

pictographs led to "the faded appearance" of many of the images.

The authors noticed that two scatter patterns were evident within the samples. One

scatter pattern was identified as hematite and the other as calcium carbonate, a white mineraI

(ibid. 34). They argued that the scatter patterns indicated that the pigment was not direccly

Page 171

attached to the surface of the rock but was intimately mixed with calcium carbonate deposit,

on the rock's surface, in a sandwich-like structure (ibid. 40). The calcium carbonate deposit

consequently acted as a protective coating for the pigment against the leaching actions of the

rain and wind erosion. They asserted that these results eXplained why wet surface chemical

field tests conducted on the paintings failed to identify the hematite. The results also found

that the calcium carbonate prevented the pigment from dissolving away (ibid.). Taylor, Myers

and Wainwright asserted that the rate at which these mineraI deposits accumulated had to be

constant, but that it was impossible to establish this fact with certainty. They established

from microscopic analysis that although brighter red images often appear not to suffer from

white mineraI accumulations they still accumulated white mineraI deposits on their surface

because they were affected, although to a lesser extent, by groundwater seepage. Essentially

paintings on surfaces that are dryer suffer less from the accumulation of white mineraI

deposits than images that are continuously affected by groundwater seepage (ibid.: 40).

The information gained from examining these samples aided conservators ln

designing the methods by which pictographs might be conserved. The white mineraI deposit

on the surface of the images is a crucial component of the sites. Any efforts to conserve

these sites must take these results into consideration. The white deposit will accumulate

through time until it completely obscures the painting. The results also demonstrated that

exfoliation was not the principal cause of the deterioration at other sites.

This study subsequently prompted another examination of paint samples to

determine the structure of the pigments used in the pictographs of the Canadian Shield.

Wainwright and Taylor (1977:31) used scanning-electron microscopy to determine the

structure of the pigment of two samples. One was from a painting on Lac Wapizagonke,

Quebec, and the other from a painting on Mazinaw Lake in Ontario. Here again, information

is not provided concerning the sampling strategy and the size of the samples involved.

Analysis of cross-sections of the samples revealed distinct, almost parallel layers of pigment

(ibid.). Wainwright and Taylor posited that a chemical or a physical process separated the

original pigment into distinct layers and concluded that the pigment must have migrated

within the deposit because the samples were taken from images that appeared not to have

been repainted.

Both experiments demonstrate that conservators must undertake detailed studies of

the pigment prior to attempting to conserve rock paintings, especially in the Canadian Shield.

Page 172

Wainwright and Taylor's (1977) study demonstrated that pigment analysis revealed that a site

had been repainted or that the pigment may have migrated. However, these results must be

considered as preliminary, since only two samples were tested. Nevertheless, they

demonstrate that several layers of pigment may exist at pictograph sites which appeared to

have been painted only once.

Archaeologists were concerned to establish detailed information on the physical state

of the rock surface upon which these images existed, and technical information concerning

their physical composition and structure, that might be useful for the conservation,

preservation, and possible interpretation of the pictograph sites.

Section 5: Culture-Historical Approach

Archaeologists employ the culture-historical approach if they want to investigate the

shape, location, and date of the images with which that they are concerned. lndeed,

establishing the age of Shield rock images is central to the culture-historical approach and for

some rock image specialists establishing the styles of these images is a means by which the

age of an image can be determined (Grant 1967; Dewdney 1970c, Pettipas 1982; Pohorecky

1981; Rajnovich 1980b; Reid 1979; Rogers 1962; Steinbring 1977, 1979a, 1990, 1993;

Steinbring and Simpson 1986; Steinbring, Danziger and Callaghan 1987; Steinbring and

Callaghan 1985). Approximately half of these authors, with the exception of Grant (1967),

Rogers (1962), Rajnovich (1980b), and Dewdney (1970c) considered the petroglyph sites in

the Lake of the Woods.

Grant (1967: 145) considered the style of these images within this geographic region of

North America as "Northern Woodland" which "continued westward" into the prairies of

western Canada as far as the Columbia-Fraser Plateau (ibid.). He maintained that some

"simple abstract elements" existed but that they were subordinate to the naturalistic style

(ibid: 20) and that this style of painting was:

confined to the regions dominated by a nomadic hunting economy. The paintings are mainly simple, rather erude representations of men and animais. In sorne areas the drawings are done in red alone, though black and white were sometimes added.

(Grant 1967: 20)

Other images included: "men in canoes, moose, elk and buffalo," Ojibwa mythological

creatures such as the "water panther," and the "thunderbird," animaIs such as deer and

Page 173

mountain sheep, fish and buffalo, an abstract element which was a semicircle with rays, and

"human forms" of different types (ibid.). The pictographs were "painted (naturalistic)" (ibid:

20), while the petroglyphs existed on horizontal surfaces in Ontario and northern Minnesota

(ibid: 149) and had "crudely pecked animaIs, thunderbirds, and humans" which Grant

asserted were similar to those found at unnamed examples of petroglyph sites in southern

Minnesota. Much of his discussion in his chapter on the Northern Woodland rock image

sites reflects his consultation with Dewdney who was recording sites at the time that this

work was written and published. lndeed, he clearly stated in his acknowledgements (ibid.:

viii) that he could only have discussed the Canadian work after consulting with Dewdney,

who had been actively involved in surveying and recording sites in the Canadian Shield.

Grant clearly recognised a wide range of images. Unfortunately the vocabulary he used to

describe these images implies sorne meaning to the reader. He did not indicate the precise

technique he used to establish the identity of an image. Furthermore, he did not consider

local stylistic traits specifie to different regions in the Canadian Shield.

A site was deemed to exist if Dewdney recognised a clearly defined image existing

there. Dewdney obviously had thought careful1y about how to de scribe these images so that

they could be classified into groups. He classified images according to the categories in the

pie chart in figure 3.45 published in both editions of the book that he co-authored with Kidd

(Dewdney and Kidd 1962: 18, 1967: 18).

Figure 3.45: The table drawn by Dewdney to illustrate how he c1assified the images at these sites.

Page 174

~ ..

Unfortunately, Dewdney's description of these types of images inadvertedly implied their

meaning to either himself or to others. Dewdney probably classified the images using art

historical terms laden with Western cultural baggage such as "naturalism" in his 1962 and

1967 book with Kidd. This was probably because he had been trained as an artist and

illustrator. He stated (ibid.) that his own work suffered from these flaws and problems.

Dewdney (1979: 326) later realised that verbal descriptions of images were prone to

subjective variation and he argued that others were equally guilty of this problem. He

subsequendy suggested that a series of graphic and visual models could provide the level of

rigour needed to classify images because words were not sufficient.

When he commenced searching for pictographs, Dewdney felt that the presence of

red iron ochre on the rock surface was insufficient to designate a physical location in the

landscape as a site. Smears of ochre were "not shapes," from Dewdney's viewpoint in the

1962 publication with Kidd, since they were not recognisable symbols. Older images that had

decayed to the point of becorning amorphous shapes were also excluded as shapes. For

Dewdney, a shape must have a defined edge and be "recognisable."

Images such as blobs and smears, which could have been recognisable images blurred

by time, exfoliation, the accumulation of white mineraI deposits, and by the growth of rock

tripe and lichens, he ignored or left undescribed. Indeed, Dewdney (1975a:4) acknowledged

that he omitted the images that he considered "ambiguous" "when in doubt, leave it out."

Smears differ from blobs since they are usually not only larger than blobs but they could

have been placed on the rock's surface intentionally. Images could fade becoming blobs, but

if Dewdney's approach was used, places with only smears and blobs were not important and

not recorded.

Perhaps the lack of smears, washes, or shapes without clearly defined edges in his

field drawings reflects the fact that Dewdney did not consider these shapes as images. Faint

pale images probably would not be present on his field recordings using paint and rice paper

since the rice paper would probably obscure them when placed on top of them. Perhaps

Dewdney's perspective regarding smears stems from his training as an artist accustomed to

paintings and art objects created in the ]udeo-Christian perspective. However, the index of

his 1967 edition of his publication with Kidd indicates that "smearing of paint on rock" was

considered important since this term was included and it had not appeared in the index or in

the text of the previous publication (1967: 190). Indeed, Dewdney (ibid.: 107) wrote in the

Page 175

text of this edition that "the vague hints of frequent over painting suggest that this was a

long over-used site". The second edition of his publication with Kidd indicates that he had

changed his opinion as to the value of the smear as an image. Dewdney (1979: 326) later

realised that verbal descriptions of images were prone to subjective variation and he argued

that others were equally guilty of this problem.

Rajnovich (1981 b: 286) acknowledged areas of ochre as washes or smears. lndeed,

she argued that a style of image existed in the Canadian Shield region whereby the edge of

the image was painted first and the shape was subsequently filled in with an ochre wash.

Rusak (1992a: 124) observed that only two sites, DhJx-9 and DgJx-2, of the seventeen

pictograph sites that were examined during her 1991 field season in White Otter Lake

(twenty-seven sites are located in this lake) in 1991, did not possess an "ochre wash or

smear". She stated that

the location of washes and smears along prominent naturalledges, in small cubby holes, along noticeable cracks and crevices on the rock surface, indicates the ochre's intentional placement. It is therefore hypothesized that these washes and smears designate the suitable location to leave offerings for the powerful Manitous within [the rock].

(Rusak 1992a: 124) Smears and ochre washes were obviously important images for Rusak smce she

argued that their presence confirmed that these places were important in the world view of

the Aigonquian-speaking peoples. Rusak, like Pastershank (1989), viewed them as individual

images in themselves while Rajnovich (1981 b) perceived washes as a part of an image. Two

years after Reid, Pelshea in 1978, photographed and recorded the images of DiKm-4, DjKn-

1, and DhKm-3 in the Lake of the Woods. Unfortunately, it is impossible to determine what

he and Reid (1976, 1977) thought of the stylistic qualities of the images. Molyneaux (1979)

did not consider the stylistic aspects of the six pictograph sites that he considered. This is

because he (ibid.: 2) argued that "regional interpretations based upon the style typologies and

style areas and supported by ethnographie ethnologies are expedient but largely

indefensible." Rajnovich (1981 b) had discovered several styles of images from either her

involvement in or reading the results of a large number of field surveys in Northwestern

Ontario. lndeed, Dewdney (Dewdney and Kidd 1967: 176) had speculated that more than

one style of paintings existed in the Shield region and that only detailed studies would reveal

the regional styles.

Page 176

Rajnovich (1891 b: 286) maintained that four styles of images existed in Deer Lake:

"open or outlined morphs", "open or outlined morphs subsequently filled in with ochre

wash", "finger lined, or stock morphs" and "closed morphs". She argued that the lines of the

first style of image could signify "internaI organs," a stylistic feature frequently used by Cree

and Ojibwa artists during the twentieth century. She (ibid.: 286) claimed that the images at

the Deer Lake pictograph sites were "distinctive" since a large number of sites occurred in a

relatively small area with "the complete range of morph styles known to Shield rock art". She

(ibid.: 282) claimed that these sites were the sorne of most northerly sites to have been

discovered in Ontario.

Her system of classification of these images indicates that the images and the centres

of the images were not the same. Fortunately she provided a few examples of each style of

image so it is possible to compare the images she recorded with other examples elsewhere.

She reviewed work by her colleagues (Smith 1981 and Pelshea 1980) who also had conducted

research in the West Patricia district, and Dewdney, and concluded that geographical patterns

did not exist regarding the distribution of the morphs in northwestern Ontario. However,

she argued (1981 b: 286) that "open style morphs" were usually found north and south of an

imaginary line between the Lake of the Woods and Lake St. Joseph. She argued that the most

common style of image was the "open and closed style".

Rajnovich (ibid. 287) maintained that the images at Deer Lake were "abstract" rather

than "realistic" based on Dewdney's (Dewdney and Kidd 1967: 19) use of these terms. The

term 'abstract' was more appropriate because she argued that the creators of these images

were more concerned with their meaning rather than their physical shape.

It is interesting to observe that, although Pastershank (1989), Rusak (1992b), Smith

(1981), and Pelleck (1981) each recorded at least one pictograph site, none of them discussed

the possible style of the images.

Rajnovich (1980b and 1981 b) also considered the idea that images occurred in

groups. She observed that occasionally single images occurred at sites such as those in Deer

Lake (1981 b: 287), at the Pukamo Island, the Cuttle Lake, and the Jackfish Lake sites (1980b:

34). She inferred that the presence of repeatedly occurring images indicated that the

"pictograph artists did not perceive the Images as single, isolated figures but drew the

morphs within the framework of a "story"" (ibid.). The presence of a single shape image

Page 177

implied "an object, thought or character" whereas if two images of the same type were

present, then a relationship existed between the two shapes.

Conway (n.d. a. and n. d. b) considered three morphs, from various pictograph sites

in Northeastern Ontario, which occurred in conjunction with each other: an open armed

man, his canine companion, and an animal pelt. Another morph, which Conway identified as

a beaver, existed above the man and his animal companion. He identified these images as

specific star constellations. Rajnovich (1980b: 35) argued that Conway's (1978) identification

of Orion and Canis Major (see discussion page 199ff) at different sites across the Canadian

Shield was problematic. Bear images, she asserted, either occurred alone or in pairs as she

had observed in the pictograph sites of both Rainy Lake and Lake of the Woods. Pairs of

images also existed at Cuttle Lake (Rajnovich 1980d). She did not discuss canine images

although Conway had identified a canine image existing with a human figure.

She also observed that images described as a "canoe-with-passenger" motif existed

throughout the Canadian Shield but that two styles of this shape existed at Pukamo Island

and Jackfish Lake, two sites in the Rainy Lake region. She (ibid.) argued that the image from

the site on Pukamo Island had "stick" passengers while a similar image on the pictograph site

at Jackfish Lake, had "open armed passengers." Unfortunately Rajnovich did not provide any

photographs of either of the images under discussion. She (ibid.) posited that the differences

occurred since the passengers in the canoe were men in one canoe and Maymaygwayshi in

the other. Rajnovich (1980b: 35) maintained that the different style of canoe reflected either

different cultural affiliations of those drawing the se images or time periods in which the

images were drawn.

Rajnovich was clearly concerned with the style of the images in the Canadian Shield,

but she did not pay any attention to Dewdney's advice (Dewdney and Kidd 1967: 176) about

resolving the problem of dating pictograph sites. She asserted that a correlation existed

between Blackduck ceramics and images of hand prints in the Rainy River region. She

established that Blackduck ceramics were found in association with one pictograph site. This

information, in conjunction with information from her informant that sorne of the images

were created by the Midé, led Rajnovich to posit that these images dated to the Blackduck

period and provided a date for the origins of Midewiwin Society. Pastershank (1989: 71), like

Rajnovich (1980b: 35), observed that paired images could be important since pairs of images

were present at each of the pictograph sites she examined in Sabaskong Bay.

Page 178

At two pictograph sites where the images, he believed, conformed ta Grant's (1967)

designation of the Northern Woodland style, Pohorecky (1981) categorised and assigned to

different groups ail the images deemed to be human figures. However, he argued that the

standard categories of classification available for examining images could not be applied to

the "hum an figures" at both sites. Pohorecky did not indicate how he concluded that any of

the figures were human rather than supernatural. He (1981: 12) was principaily interested in

the "human" figures at the two sites on Smith Narrows between Hickson and Maribel1 Lakes

in northern Saskatchewan. He believed that these represented the northwestern tip of this

style, the epicentre of which occurred in Northwestern Ontario. Grant (1967: 147) had stated

that the large concentrations of pictographs had been described in Ontario, specificaily, in

the Lake of the Woods area. The criteria Pohorecky used were whether the figures had

upraised arms, arms hanging down, or one arm up and one arm down or whether the images

were hum an faces with no hands (1981: 13). He stated that the figures in the third category

had box-like shapes.

Lambert (n.d) examined DhKm-l, DhKm-4, DhKm-18, DhKm-5, DhKn-7, DhKn-

6, DjKn-l, and DjKn-2 during the field season of 1985. He applied a technique that he had

developed du ring his field survey of some pictograph sites in the Upper Severn region in

northwestern Ontario during the field season of 1982 (1983) for dealing with images. He first

subdivided images into twenty and, later, into thirty-five classes (n.d.) that he devised based

upon his own recognition and identification of each image. Therefore ail the images that he

classified as anthropomorphs he put in Class A; hand-prints in Class B, finger-prints in Class

C, canoes in Class D, pipes in Class E, crescents in Class F, taily-marks in Class G, arrows in

Class H, animaIs in Class l, struck animaIs in Class J, fish in Class F, birds in Class M,

lightening in Class N, turtles in Class 0, crosses in Class P, chevrons in Class Q, simple lines

in Class R, double lines in Class S, and unclassified images in Class T (1983: 113). Lambert

asserted that each image placed into each class ought to be examined "for its ability to

convey alternative meanings through differences in morphological attributes and associations

in the context of narratives" (ibid.: 111). Therefore images were placed into different but

larger groups. This grouping ailowed assignation of some level of meaning, according to

whether an image was identified as abstract, naturalistic, symbolic, or realistic. He believed

these terms "were normaily used to classify art and its elements" (ibid.: 110) and applied

them as defined by the Webster Third International Dictionary, published in 1976. Lambert

Page 179

(ibid.: 111) guessed at the meaning given to the image both by the creators of the images

themselves and the intended audiences.

The question of the authorship of these images remained for Lambert (1983: 1 09) an

unresolved issue. He assumed, however, that they were created by human beings rather than

foilowing "the contemporary belief' that they were created by the Maymaygwashi. He (ibid.:

110) acknowledged that the interpretation of these images was "a difficult process for those

who did not author them," due to the inherent problem whereby the observer, the researcher

needed to be as objective as possible to classify and describe the images for comparative

purposes.

The foilowing papers concern themselves with dating the images rather than with

style. Dewdney's (1970a) publication reveals his concern to resolve the dating problem for

paintings in the Canadian Shield. This paper had four parts, through which Dewdney

approached the resolucion of the problem. The first part discussed information concerning

the measurement of physical features, lichen, rock tripe and mineraI deposits to see whether

paintings from two different periods could be distinguished at the Hickson-Maribelli site in

northern Saskatchewan. Dewdney (ibid.: 8) argued that the site was in "traditional

Chipewayan territory" and had a "very impressive setting" (ibid.: 10) as weil as a deep crack

large enough to "hide a smail canoe." He examined the orientation of the rock surface; the

"protection of the site" (ibid.: 13), in which he considered the overhangs and cracks that

diverted the water seeping and drips from above; the height of the paintings above the

water, and whether rock benches existed at the base to prevent water splash from below. He

(ibid.) measured the level of lichen, mineraI deposits, and determined the "nature of the rock

surface," or type of rock, which he maintained affected the vulnerability of the paintings to

the weather. He argued that granite was more resistant to weathering than chIo rite schist. He

examined the colour and pigment impairment, the manner in which the paint was affected by

cime, but subsequently decided that the categories were too subjective. He concluded that he

could not measure variables objectively due to the difficulty in finding a variable that had

some significance as a dating device.

The second part of this paper presents conclusions that Dewdney obtained from

reviewing the world literature to see how other researchers throughout the world tackled the

question of dating rock image sites. He realised that the specialists from France, Spain, and

USSR had attempt to deal with the question of dating these images by excavating the area

Page 180

immediately adjacent ta the images themselves, a technique impossible in the Shield region

since images existed immediately beside water. Dewdney (ibid.: 14) maintained that offerings

could not provide any "stratigraphic issue" and they reflected a "cult" that could be of a

recent development. He argued that the water level mark could indicate something and that

analysis of the white mineraI deposit, which was calcareous, might be use fuI (ibid.: 15).

Through reading the world literature, he determined that examining the paintings for

superpository evidence was interesting, since this was common in Shield region. He asserted

that, where it was possible to establish which image had been superimposed, a sequence

could be established. He realised that, even if this technique was possible, too few examples

of this existed, such as on Face X of Site #181, and too few paintings showed what he called

"sufficient evidence of stylistic affinity with any other painting" (ibid.: 17).

Another method that could aid the dating of these images would be determining

stylistic sequences. Yet Dewdney argued that it was very difficult to define the style of a

painting, as evident from the literature from elsewhere in the world. His review of this body

of literature led Dewdney (ibid.: 18) to condude that relatively few numbers of images were

similar. He also (ibid.: 19) asserted that it was too early to discuss in any depth the

"complexities of sorting out style features" for the rock image sites of the Shield region.

Dewdney conduded that the "greatest common style denominator" of the images of the

Shield region was "their individuality". He argued that the use of "[L]aboratory techniques"

enabled the examination of pigment samples for their organic binder that caused the paint to

adhere to the rock surface. He (ibid.: 21) stated that considerable developments had occurred

using chromatography (e.g. Denniger 1971), spectroscopy of different types, and analysis

using scanning-electron microscopy. However, he identified the biggest problem as the

difficulty in obtaining a large enough sample to guarantee that most of the binding agent had

not disappeared. This he believed would cause researchers to scrape pigment samples from

an undetermined number of paintings.

Dewdney posited that patination, although it rarely occurred in the Shield region,

offered sorne hope of dating. Patination occurred, he argued, when the colour and texture of

the rock's surface changed when the surface broke away either as a consequence of

exfoliation or weathering (1970a: 23 -24). He asserted that oxidation of the pigment could

occur whereby the pigment became redder, oxidised as it was exposed to the sun. This is an

interesting idea, but Dewdney did not provide any evidence to substantiate this daim. He

Page 181

posited that lichenometry was a useful dating technique but he stated that more information

was required as to their growth rate and why lichen grew where it grew.

Dewdney reviewed the ethno-historic literature written by researchers elsewhere in

the world. He argued that by combining the information from informants, ethnohistoric

data, ethnographic data, and archaeological data that it might be possible to establish a secure

continuous cultural sequence. Therefore, according to Dewdney, it might be possible to

establish which images occurred when, especiaily given the depictions and/or impressions of

specific European innovations at seven sites in the Canadian Shield, such as flags on poles,

European forts, and guns. He maintained this, although he had discussed this topic earlier in

Dewdney and I<jdd (1967: 50), where he observed that several images represented as

European, by their painters, were very similar to those of buildings drawn by the Midé on

birch bark scroils.

The other key reference that Dewdney used as a means of dating was Schoolcraft's

information regarding the pktographs at Agawa rock image site on the north shore of Lake

Superior. Schoolcraft (1857: 405-411), Dewdney (1970a: 28) argued, "reliably" dated the site

as having been painted by Mygeangun, a shaman practitioner from the south shore, as

approximately 1800 given or taken twenty years. Dewdney asserted that prehistoric dues to

dating these images induded the recognition of objects such figures with objects such as

bows and arrows, and medicine bags.

Dewdney conduded from the world literature that rock image sites were not reliably

dated. He (ibid.: 30) posited that examining each site, its setting and its painting" in as mu ch

detail and as much precision as routinely employed in an archaeological excavation" afforded

the only method by which to date sites.

This paper's third section deals with his new approach, involving fieldwork to

develop a more intensive examination of the variables that caused or increased the "physical

impairment of the paint and rock substrate" at sites in Manitoba and Northwestern Ontario

(ibid.: 3). He (ibid.: 30) chose three pictograph sites that he argued "promised maximum

return for such detailed examination" namely: Agawa since its paintings were recent and

could be dated, as discussed above, a site on Frances Lake, and a site with an "early origin"

on Fishing Lake (ibid.:31). He (ibid.) argued that early sites should combine "evidence of

considerable paint impairment," "maximum of protection from weather," and "a

Page 182

combination of factors that were weather resistant," but did not state or clarify what he

meant by these statements.

Dewdney (ibid.) believed that the pictograph sites at Fishing and Frances Lakes

shared similar painting styles and context different to other sites in the Shield region, and

that both were "impressively protected from weathering." Why and how remains unclear. He

also (ibid.) argued that both sites showed "obvious impairment to suggest a substantial age."

He did not state what these impairments might have been. Dewdney's (ibid.: 36) subsequent

fieldwork had four objectives: (a) to review and extend the understanding of all the variables

involved in the "vulnerability and impairment of all paintings," (b) to record all the data that

could be useful for dating purposes, (c) "to test a variety of equipment and instruments and

the practicability of accurate recording," and (d) "to collect samples of both painting and

rock substrate for laboratory examination and analysis".

At each site, Dewdney (ibid.: 39) measured the quantity of moisture experienced

from three main sources: rainwater, groundwater and shore water. He argued that wind

driven rain still affected sites with an overhang or a bench at the base, because neither

physical feature protected the site. His experiments revealed that sites with a northerly

exposure, especially those that faced northeast, suffered most from wind driven rain. Those

cliff faces facing south, southeast, and east-southeast received the least amount of wind

driven rain. He calculated the maximum effect of direct sunlight available annually

experienced by a site to establish the temperature differences that occurred on a cliff face so

as to determine the possible rate of exfoliation.

Dewdney (ibid.: 40) estimated the quantity of ground water seepage because he

asserted that those physical features that may have protected the images from rain "ensured"

groundwater seepage. Dewdney observed that blank streaks on the surface of the rock face

indicated groundwater seepage while groundwater seepage was not constant if "tracks of

mineraI precipitates" were present on the rock surface. Unfortunately, the precise meaning of

the words "blank streak" remains unclear. He (ibid.: 40) devised the "time-of-wetness test"

using an instrument called a "splash test device," which he himself invented (1972: 1) so that

he (1970a: 40) could measure the elapsed time for a droplet to run down the surface and

evaporate.

He concluded that the largest physical affect of water on a painted surface and rock

substrate was frost action. Dewdney posited that perhaps thermographs could be set up near

Page 183

selected sites allowing information regarding the tempe ratures to be collected at each Face.

Results could be compared with maximum and minimum temperatures at the nearest

weather stations. This, he argued, might make it possible to assess the vulnerability of the

rock to frost action. He also argued that ice erosion was a possible problem for sites.

Dewdney (ibid.: 46) studied the lichen at Agawa, Frances, and Fishing Lakes to

establish the level of encroachment. He and his assistant recorded the dimensions of the

lichen thaluses. Dewdney also tried to establish the level of patination by comparing the

colour of the "mother rock" in a deep fracture. He stated that he was no longer a sceptic

concerning the colours of the pictographs, and argued that the colour of the pigment may

have changed according to the degree of impurity present or the extent of weathering. He

argued that it was possible that colour could increase because of the interaction between

solar radiation, moisture, and radiation. Unfortunately, it is unclear how much Dewdney

knew about the relationship between the solar radiation, moisture, and radiation.

Dewdney concluded that the problem of dating the pictographs in the Canadian

Shield could be resolved only by using what he termed a "three forked approach." Field-,

laboratory, and desk work were conducted simultaneously and considered equally important.

He argued that fieldwork included determining the importance of the pictograph sites to the

local indigenous peoples. Research ought to be conducted on petroglyph sites, sin ce he

believed that examination of these images could offer insights into pictograph sites. He

argued that future studies of these images must always include examination of the paint and

the rock itself, so samples must be removed for scientific analysis.

"Desk" research had three goals: (i) a stylistic analysis of the images supplemented by

an analysis of the images found on hide and birch bark scrolls; (ü) an examination of the

ethnohistoric literature for recent archaeological surveys and developments, analytical

techniques for field samples and comparative purposes; and, finally, (üi) to continue to

communicate with those colleagues also interested in the rock image sites of the Canadian

Shield.

There has been much controversy about the dating of rock image sites from the Lake

of the Woods area. Steinbring, an archaeologist from the University of Winnipeg, has

undertaken the largest number of studies of the petroglyph sites in the Lake of the Woods.

He argued that the pecked images of the petroglyph site called the Mud Portage Site (DkK.r-

4) on Clearwater Bay belonged to the Lake-of-the-Woods Style (Steinbring 1977). He

Page 184

described the site as containing ninety-four petroglyphs, twenty of which were found

through the excavation of up to one metre of archaeological deposits (ibid.: 139). The

images, or glyphs as Steinbring described them, were found upon "a grey-ish green [the

colour of the rock] basaltic formation which is essentially horizontal" (ibid.). Steinbring and

Callaghan (1985: 8-9) had observed that changes in the water levels had caused erosion of

these soils and started to expose the petroglyphs. They argued that, although this site had

been disturbed by rodent tunnelling and roots, the archaeological deposits were stratified,

and demonstrated a long continuum of intermittent occupations starting at 9,000 B.P., and

ending around the time of European contact in this region, which was postulated to have

occurred around 1732. Excavation at this site revealed more petroglyphs but these

archaeologists acknowledged that both radiometric and typological dating did not permit

clear projections of any artistic activity to before 7,000 B.P.

Steinbring and Callaghan knew that the highly acidic soils of the Shield region meant

that bone artefacts did not survive and hence one of the principal sources for radio carbon

dating was non-existent. Wood charcoal from levels 6 and 8 of the excavation provided dates

in the range of 2000-2500 B.P., which Steinbring asserted indicated the "substantial antiquity

for the deeper positions" (ibid.). Levels 8 to 12 of the excavation provided a large quantity of

archaeological objects which were deemed to be Archaic and to belong to the "Oxbow­

McKean" complex prevalent in this part of the sub-Arctic and dated to about 5000 B.P.

Steinbring argued that these layers covered the petroglyphs at several places throughout the

area. Beneath level 12, cultural materials existed that were not of "the usual type", that is, not

objects, lithics, that were readily identifiable as belonging to a known cultural type (ibid.).

Although Steinbring had discussed the style of the images in his (1977) publication, his later

publications do not clarify how an image can be classified by someone other than himself as

belonging to the "Lake-of-the-Woods-Style" (1979a: 13). Steinbring argued that images

belonged to the "Lake-of-the-Woods-Style" if they demonstrated something he termed a

"gross form" and if they were similar to an image depicted "in the nature of a classical art

style." Unfortunately, he is not clear as to what constitutes either "gross form" or "a classical

art style" (ibid.). Steinbring, likewise, did not elaborate on the difference between solidly

pecked and lightly pecked forms. He observed that "vague linear" forms, which remain

undefined, belonged to a different stylistic group and were evident at the peripheries of the

main panels of the Mud Portage Site (DkKr-4) and the Kennedy Island Site (DjKp-4). This

Page 185

style, as yet unnamed, was evident at two other nearby petroglyph sites in the northwestern

part of the Lake of the Woods, called the Dowse site, DrKr-2, and Macrun Point, DjKr-1

(ibid.: 16).

Steinbring concluded that those images classified as "Lake-of-the-Woods-Style" were

similar to those from the ]effers Petroglyph site in northern Minnesota, dated as Archaic in

age, since some of the images depicted copper projectile points and atl-atls. It is likely that in

doing so he drew upon his Ph.D. research on early Copper artefacts (Steinbring 1975). He

concluded that those images of "Lake-of-the-Woods-Style" belonged to the same time

period as those at the ]effers Petroglyph site, because the glyphs resembling bird tracks were

solidly pecked in both central and peripheral locations and consequently were not

"assignable" to either the pre-ceramic or ceramic horizons (ibid.). However, Steinbring did

not state precisely how sites could be subdivided into "central" and "peripheral" parts. It is

likely that he used rus own cultural perspective to deterrnine the identity of the images visible

on the rock surface. Description of these images is clearly intermixed with low-Ievel

interpretation and his approach to the classification of these images cannot be replicated.

It was his later work that established precisely how he found the images were

dateable, where specific styles were found, and how they were identified. Studies by

Steinbring and Nielson (1986) had established that some of the litruc materials came from a

quarry nearby, wruch was part of an early litruc complex associated with the rugh water levels

of glacial Lake Agassiz, at least 1,000 years before the initial occupation of Mud Portage,

DkKr-4. Steinbring posited that these early peoples had migrated to the present lakeshore

when glacial Lake Agassiz drained away and that their cultural materials were characterised

by unifaces and used platform flakes. This combination, Steinbring noted, was classified as

the earliest cultural remains in North America and was either termed the "Pre-Projectile

Point Stage" or "Early Litruc" (ibid.: 141). Steinbring argued that it was impossible to exclude

the idea that people who had lived in trus region during trus early time period had created the

petroglyphs. He argued that rus estimate of the date 7,000 BP "was conservative" and was

based on the presence of the materials from the Oxbow-McKean complex evident in Levels

8 to 12, which were at least three levels above the petroglyphs.

Once again Steinbring observed that the petroglyphs at trus site, DkKr-4, could be

conveniently divided into two styles based on their location (ibid.). The most prominent style

for Steinbring was the "Lake-of-the-Woods-Style" found on the crown of rock formations

Page 186

that he asserted "took advantage of the smoothest and most commanding locus". He argued,

based on these observations, that this was possibly the oldest style and that this conclusion

was confirmed by his excavation of the Mud Portage site. The images that belonged to this

style were "naturalistic and representational". Also belonging to this style were images that

were "highly eroded unclassified masses," which Steinbring argued could sometimes form

"solid iconic masses". Furthermore, Steinbring maintained that it was these masses that

linked the Lake-of-the-Woods-Style" to the "Clearwater-Bay-Style".

The second style of petroglyphs, the Clearwater Bay Style, in contrast, was found in

images created on the peripheral edges of the exposed rock surfaces and "virtuaily absent

from excavated surfaces" (ibid.). Only one example of a morph in this style existed beneath

the excavated layers of the site at Mud Portage, DkKr-4. Images belonging to this style were

linear and reflected iconic incongruence, since, invariably they were created using the glacial

striations that exist upon the surface of the rocks in this region. Steinbring maintained that

the striations were "exaggerated by pecking, and lines or curves were added". Linear bird-like

forms and insectiforms, such as a "dragonfly", were present in this style.

Steinbring (ibid.: 142- 143) conduded that both styles of images exist at the Mud

Portage Site, DkKr-4. He provided drawn examples, but did not provide photographs of

each style, so that others could observe precisely what he meant by each style, even though

he had argued that photographs are always more use fuI than drawings to a novice in this

aspect of archaeological research.

Steinbring and Cailaghan (1985: 3) contrasted the "Clearwater Bay Style", where

images were linear and outlined, with the "Lake-of-the-Woods Style" (ibid.: 2), where images

were "solid", because their centres were entirely pecked, and not outlined. Both authors were

clearly satisfied as to the pre-ceramic context for the "Lake-of-the-Woods Style" in

Northwestern Ontario, as weil as with the idea that the images at DkKr-4 were created over

a period of at least 5,000 years (ibid.).

Two years, later Steinbring, Danziger, and Cailaghan (1987: 3) dated the style of the

petroglyphs at the Clearwater Bay site, Lake of the Woods by the overlying archaeological

deposits in the soil mande immediately above the petroglyphs. The three archaeologists

reiterated a previous daim published by Steinbring and Cailaghan (1985) that the "Lake-of­

the-Woods-Style" dated to 9,000 B.P., using a radio carbon date based on one charcoal

sample taken from the archaeological deposit immediately above the petroglyphs at Mud

Page 187

Portage, DkKr-4. They speculated that these images, belonging to the "Lake-of-the-Woods­

Style" at Mud Portage, indicated "possible movements by mobile pre-Archaic populations

into the Canadian Shield" (ibid.: 9). However, this published date of 9,000 BP is not the

same as the result provided by the laboratory that dated this sample (BGS 1101, Brock

University Geological Society, Radiocarbon Laboratory, 1986). The laboratory report from

Brock University [Radio Geological Sciences Radiocarbon Laboratory] dated the sample

submitted by Steinbring and Callaghan and analysed by Melville on April 26th 1986 to 1,320

± 80 years.

Steinbring and his colleagues claimed, in print, that this sample was taken from soil

immediately above the petroglyphs but the diagram drawn on the form submitted to Brock

indicates it did not originate at, but above, the base of the soil horizon. If the sample was

taken from the base of the horizon, there should not have been any organic materials.

However Melville, the laboratory assistant who ran the test, wrote on the sheet that

accompanied the result of the test that were there numerous penetrating rootlets in the

sample to be tested and that he had removed as many as possible. Such an observation by

the scientist running the experiment suggests that perhaps the sample may have originated

from somewhere in the rootlet zone. Steinbring wrote in the section of the form that

required information concerning the "state of preservation, factors that may result in

anomalous age (present or past geological environment, sampling, storage, etc") that both

archaeologists speculated that the sample came from a feature such as a hearth, rather than

modern root burn. When asked for the "regional stratigraphie setting, geological or other

significance or probable age," Steinbring wrote that the sample in question was taken from a

deposit slightly below a deposit that had yielded a date of 2,550 ± 140 years.

Steinbring's (1993 and 1999) publications highlight the fact that he was determined to

use this "new" date of 9,000 BP, rather than the date obtained from the radiocarbon

laboratory at Brock, of 1,320 ± 80 years, since it confirmed his published hypothesis about

the se images in earlier works. Steinbring's (1993) publication argued that the images at DrKr-4

that were deemed to be of the "Lake-of-the-Woods-Style" were definitely Archaic in age.

Steinbring (1999) contended that the date for the petroglyphs at Mud Portage provided "a

base line for attempting a review of early rock art in the North American Mid-Continent."

Clearly, Steinbring wished to be the archaeologist who had discovered the oldest rock image

site in North America.

Page 188

Reid (1979), on the other hand, attempted to establish the age of several petroglyph

sites in the Lake of the Woods by relating the images there to those found on other

pictograph sites, and on the birch bark Massacre Scroll that related the murder of Father

Aulneau, the son of La V érendreye, and their companions. He disagreed with Steinbring's

attempts to date the images of DkKr-4 and discussed the newly discovered images from

DjKp-3, called Tranquil Channel, in conjunction with images from the other petroglyph sites

in the Lake of the Woods. He examined the relations among the different petroglyph sites,

and realised that three images were found at DiKo-2 (Sunset Channel), DkKr-4 (Mud

Portage), DjKp-4 (Kennedy Island), and DkKq-24 (Clearwater Bay): fertility symbols, stick

figures, and turtles. Reid (ibid.: 246) observed that none of these sites had images which were

deemed as "typically Archaic projectile points, or atl-atls", which Lothson (1976: 31) had

used to date the] effers Petroglyphs in Minnesota, but a number of similar animal and human

forms were present. He reasoned that the excavation of the Meek Site, DjKp-3, which had a

habitation site as well as petroglyphs, provided sufficient evidence for testing two

hypotheses: (a) that the majority of the rock image sites on the Lake of the Woods were

"temporally placed" in the Middle and Late Woodland period and were "a product" of the

Algonquian speaking peoples, and (b) that the petroglyphs at DjKp-3 (Tranquil Channel)

were created by the Cree during the tenth to the fifteen centuries AD.

Reid maintained that these petroglyphs indicated an Algonquian cultural affiliation of

these sites. Based on two radiocarbon samples, this archaeological site belonging to the

Selkirk culture was dated to 935 AD. and 1420 A D (ibid.: 252). Reid (ibid.: 253) reasoned

that majority of the Lake of the Woods sites were interrelated, based on their shared imagery.

He argued that the date from the habitation site supported his hypothesis that the Lake of

the Woods sites were relatively recent Algonquian ones, because the fertility images at

Tranquil Channel were connected to the nearby archaeological site.

If the ]effers Petroglyph site and DrKr-4 were connected, they belonged to "an

extremely long cultural" period and that the relationship between the petroglyph and

pictograph sites in the Lake of the Woods must be considered (Reid 1979: 247). He noted

that the Canadian sites shared similar animal and human forms with the American site. Reid

reasoned that several images at petroglyph sites DjKp-3, DiKp-2, DrKr-4, DiKp-4 and

DkKq-24, in the Lake of the Woods, could be identified as fertility symbols, stick figures,

and turtle shaped images. He concluded that, since the petroglyph sites in the Lake of the

Page 189

Woods shared the same images, they were probably connected to each other as weIl as to the

pictograph sites. Reid (ibid.: 247-250) asserted that the offerings left at DiKm-1, DhKm-1,

DiKp-1, DjKl-2, DkKn-7, DkKr-4, DhKm-3, and DhKo-1 were additional evidence of a

connection between the pictograph sites and DiKo-2, a petroglyph site. These offerings

were, for Reid, an indicator of religious importance and provided a sense of "the sacred" for

local indigenous communities.

One particular morph, the key factor for Reid (ibid.: 250), demonstrated a link

between these sites and the belief system of the Midé of the modern Ojibwa. This morph

was identified by Redsky, a Midé practitioner from Shoal Lake (part of the Lake of the

Woods) as a "paisq", which he roughly translated as a "bird of omen". This image was

identified at two of the petroglyph sites: DiKo-2 and DjKp-3, three times on the Massacre

birch bark scroIl in figure 3.46 and on an Ojibwa birch bark scroIl caIled a "migration chart"

held in the National Museum of Canada.

Figure 3.46: The Massacre Scroll photographed in 1993 by J. Rusak.

Reid (1979) has taken a different approach to dating the images of the petroglyph

sites since he did not rely on absolute dating techniques. He argued that none of the

petroglyph sites from the Lake of the Woods had shapes that resembled the images at the

]effers Petroglyph Site in Minnesota dated to the Archaic period. Steinbring and his

coIleagues (Steinbring and CaIlaghan 1985; Steinbring, Danziger and CaIlaghan 1987)

subsequently advanced an Archaic date for the Algonquian occupation and creation of

DkI<r-4 (Mud Portage), one of the sites that Reid had considered. They argued that DrI<r-4

was connected to the ]effers Petroglyph site since Archaic projectile points and atl-atls were

depicted at this site.

Page 190

Two articles debated about how a method might be developed so that images from

different sites could be related to each other. Whelan's (1983) study built upon the work of

Maurer and Whelan (1977). They sought to develop a method applicable to other forms of

"pictography within this region" (ibid.: 196), by which sites and paintings could be classified

and catalogued according to the subject matter of each image. Each image, termed data, was

placed into different levels: class, subclass, and attribute (ibid.).

Maurer and Whelan observed that problems existed in trying to conduct this type of

analysis since the descriptions of the images were invariably affected by the personal

interpretation of the observer. They discovered that difficulties existed in developing a

format that was relatively adaptable to a computer programme: how could an archaeologist

de termine and identify what constituted a class, a subclass, and their respective attributes,

and how could they identify specific symbol types and names (Maurer and Whelan 1977:

199).

Maurer and Whelan were determined to deal with the increasing quantity of data and

to clearly describe and classify these images without imposing meaning upon them during the

culture-historical approach. They understood the issues of unintentionally imposing meaning

to an image in the course of describing its attributes.

Section 6: Contextual approach:

This approach will be used, only after the application of the culture-historical

approach, to relate various types of empirical information to the site where the images were

found. It considers a broad set of associations and relations among the images themselves,

and between the images and their physical settings. This approach involves a search for

patterns relating different images, and combinations of images both between and within sites

Archaeologists inevitably describe the immediate physical landscape. Petroglyphs

exist on the surface of a rock while pictographs exist on rock surfaces either on rock walls or

in caves. It was traditionally believed that pictographs exist at the base of vertical granitic

rock walls, either immediately beside the water, or near the water's surface. Pohorecky (1968:

103) stated that:

the micro-environment of the rock paintings is easily categorised. The paintings occur either singly, or, more commonly, in small groupings on fiat, vertical and sloping rock faces or cliffs just above the surface of sorne body of water, and can be reached usually from sitting or standing position in a canoe, and sometimes by stepping onto ledges just back from the water. AlI paintings are visible from the vantage point of a canoe. This

Page 191

kind of setting, where paintings lie near the edge of the water in summer or ice in winter, is itself one of the most distinctive features of rock paintings throughout the entire area.

Yet, pictograph sites are not always found immediately near water. This statement is

supported by the pictograph site DhKn-1, first recorded by Dewdney as Site # 106 on 19th

August 1960 in figure 3.11. It is on the southwest facing shore of Aulneau Penin sula

immediately opposite Wisakode Island and north of Rabbit Island. It lies beside one of the

channels that can be used to enter Obaibikon Bay to the north. The paintings here can be

reached by standing on a narrow, long rock ledge immediately behind the large boulders and

the birch trees visible in figure 3.47. These boulders are a few metres from the water's edge,

where various grasses, trees, and short plants grow.

Figure 3.47: Part of DhKn-l beside Obabikon Channel.

Two studies from neighbouring regions exemplify the contextual approach because

the authors sought to ascertain how rock image sites were related to other features in their

vicinity (Noble 1968; Sweetman 1955). Noble described and discussed the pits, cairns, and

petroglyph sites found in close proximity to each other at Rock Lake, in Algonquin Park in

southern Ontario. He (ibid.: 63) contended that the rock cairns behind the rock lined pits

were not definitely connected with each other, although they were related in some manner.

The pits were interpreted as '''vision pits'" (ibid.: 62), while the cairns were interpreted, based

upon comparison with cairns at Red Sucker Point, on Lake Superior, as possibly being

""tobacco-drops", dedication cairns erected to a guardian spirit after a successful vision"

Page 192

(ibid.: 63). The petroglyphs were interpreted as being both archaeological and artistic in

nature and "of probable socio-religious import" (ibid.). He asserted that they were not as

complex, numerous, or detailed as those elsewhere in the Canadian Shield, discussed by

Sweetman (1955) and Dewdney and Kidd (1962). Noble also concluded that it was

impossible to infer the cultural affiliations of the people who used the pits.

Sweetman (1955) described how the images at the Peterborough Petroglyph site were

mapped and recorded. The images, called "glyphs" (ibid.: 101), were recorded using plaster

moulds, drawn, and photographed. Test pits were dug near the petroglyphs to establish

whether the people who had created the glyphs had left their tools nearby, or camped in its

vicinity while they created the images (ibid.: 101-104). Nothing but tree stumps or windfalls

were found in the test pits and these could not provide any information about the creation of

these images. An archaeological site of an Iroquoian village, which he called the Quakenbush

site was found nearby (ibid.: 104). Sweetman argued that it was unlikely that the inhabitants

of this village had created the petroglyphs, but he did not indicate why he arrived at this

conclusion. He (ibid.: 108) concluded that it was impossible to arrive at any clear conclusions

regarding the age and meaning of the images.

Arsenault and his colleagues have examined in considerable detail eight pictograph

sites in Quebec (Arsenault 1997; Arsenault, Gagnon, Martijn and Watchman 1995; Arsenault

and Gagnon 1998). Arsenault (1998) examined in considerable detai! the sacred space, and

physicallocality of eight pictograph sites in northern Quebec. They are in different regions of

the province: one is in the James Bay area, two are in Abitibi area, one is in the

Témiscamingue, two in the Outaouais region, one in Mauricie, and one in Haute-Côte-Nord.

Nevertheless, he drew considerably upon ethnographic data from western Ontario. He

argued from the substantial ethnographic record of the Ojibwa, that rock image sites in

Quebec were sacred places used by shamans.

He drew on ethnographic data collected from throughout Canadian Shield rather

than just specific regions in Quebec. This is surprising, especially since the Algonquian

groups of the western Canadian Shield probably had different cultural experiences from

those in Quebec; although these images were created by Algonquian-speaking peoples, they

were created by different groups. It is therefore possible that there were specific differences

in the religious practices and perhaps the use of images between the Ojibwa, living farther

west, and the local groups of Algonquian speaking peoples living in these regions of Quebec.

Page 193

Arsenault considered that pictograph sites may have acoustic properties that explain

why some rock faces rather than others were utilised. Archaeologists, elsewhere in the world,

have investigated whether site selection occurred as a consequence of the special sound

effects that could have been produced in specific spaces (Rainbird 2003, Waller 1993).

Both Dewdney and Kidd (1967) were obviously concerned with identifying what

constituted a site. Yet, in the same breath as they (ibid.: 6-8) delineated the properties of a

"typical site," they stated that it was difficult to generalise about the properties of a typical

site. Evidently, the visual landscape was not considered important for understanding these

images. Nor is much information recorded as ta how, precisely, sites were located. It appears

that Dewdney relied upon information provided by inhabitants of the areas where he sought

to find rock art sites (ibid.: 7-8).

Section 7: Literature concerned with meaning

The three remaining approaches are aU concerned with establishing the meaning of

these images. For a detailed discussion of each approach see Chapter I.

(a) Intuitive approach:

Archaeologists who have adopted the intuitive approach to explain rock image sites

in the Canadian Shield include Granzberg and Steinbring (1995), Lanoue (1989), Pettipas

(1982), Steinbring (1976, 1979b, 1982), Steinbring and Elias (1967, 1968b), Steinbring,

Granzberg and Lanteigne (1995), Steinbring and Steinbring (1978) and Vickery (1991). Only a

few of them have referred to sites in the Lake of the Woods.

Steinbring and Steinbring (1978) postulate that the images found at Machin Point,

DjKr-1, in the northwest part of the Lake of the Woods, were created as a consequence of

shamanism, based on a study of shamanism in connection with a petroglyph site in British

Columbia by Walker in 1977, and ethnographic data from the Winnipeg River watershed

(Steinbring and Steinbring 1978: 13). Both archaeologists speculated that a "template effect"

existed in the distribution of the glyphs of DjIv-1, which "seemed to be closely related to"

the naturallines evident in the rock upon which the images existed (ibid.: 13). The authors

concluded that the natural lines guided the creation of the images in a manner related to

Algonquian scapulimancy.

Page 194

Some of the images were identified and discussed according to the identification of

both archaeologists rather than merely described. According to the authors, careful scrutiny

of the images indicated that "vandalism" had occurred at this site. The green paint in some

of the "authentic-appearing" images, which had been created using a variety of pecking

techniques, was considered as evidence of vandalism. The authors maintained that they were

able to deduce that "fraudulent glyphs," made by short pecked lines, also existed at this site.

Low-level interpretation is embedded in the descriptions of these images, such as the

"dragonfly" petroglyph discussed by Steinbring and Steinbring (ibid.).

Steinbring (1982) later drew upon a variety of ethnographic sources to substantiate

his arguments that the images found in pictographs and petroglyphs in the Canadian Shield

existed as a consequence of Ojibwa and Algonquian shamanistic activities and dreams. The

petroglyphs of the Lake of the Woods sites were only briefly mentioned. The goal of his

paper was to argue that the art of contemporary native artists, with Algonquian backgrounds,

had the same origins as the crea tors of the petroglyphs, created during the Archaic Period.

He overlooked demonstrating precisely how this was possible, since he neither provided

evidence that he had consulted or interviewed contemporary native artists, nor suggested

who these peoples might be.

Steinbring (1976) speculated about the origins of the copper technology and lunar

cultism in the ancient Great Lakes region. He (ibid.: 164) argued that creseent shaped forms,

which he eailed lunar motifs, were present at rock image sites and on bireh bark seroils

created by the Ojibwa. He asserted that substantial ethnographie details existed concerning

lunar symbolism in the ethnology of the Great Lakes Algonquians, but provided no

references for his readers. He subsequently identified erescent shaped motifs at the

Peterborough Petroglyph site and disagreed with the Vastokases' (1973) identification of

these images as boats. He argued that some of them were probably not boats with people in

them but eould "actuaily represent copper erescentie symbols" (Steinbring 1976:. 164). He

subsequently argued that similar images are found at a number of sites in the Lake of the

Woods and elsewhere in Northwestern Ontario and eastern Manitoba (ibid.: 164-165). For

eaeh example given of a erescent shape at a site, he disagreed with its interpretation by

previous archaeologists, arguing that "many crescents have probably been misinterpreted as

boats or eanoes, the most obvious of these being the "upside down" ones (ibid.: 156).

Page 195

Granzberg and Steinbring (1995) classified glyphs by type usmg the intuitive

approach. They argued that line, tree, and circle motifs were "core symbols with a

transcultural base line of archetypal usage" based on examples taken from sites in the

American Southwest and from northern Manitoba and Northwestern Ontario (1995: 52).

They postulated that their data suggested, despite the lack of ethnographic references, that a

certain level of trans-culturally valid meaning existed, allowing the authors to identify distinct

types. However, both authors maintained that considerable research was necessary before

any categories could be securely identified (ibid.).

Lanoue's (1989) study considered the petroglyphs of the Lake of the Woods. He and

Richard Doble, from the University of Winnipeg, exarnined and recorded an unknown

number of petroglyph sites in 1981. Lanoue asserted that these rock image sites were

"stylistically Algonkian". He argued that these images, which could not be placed "in the

tradition of 'great art'" (ibid.: 7), were "the product of similar intellectual processes and

formally the same everywhere at all times" (ibid.: 9). Lanoue drew (ibid.: 42) upon

Steinbring's work (Steinbring et. al. 1987) to postulate that the petroglyphs were at least

5,000, and possibly even 9,000, years old. He maintained that his own examination was

flawed because he did not have the full body of data available to him since the full repertoire

of images in the Lake of the Woods was not yet established. He posited this was because (a)

not all the petroglyph sites had been found and (b) sorne were covered up with water and

therefore not visible.

Using CUITent knowledge concerning hunter-gatherer and tribal societies, Lanoue

proposed that, if the style used to create the images was related both to their meaning and

content, useful information regarding the "central paradox of territoriality" might be

obtained. He maintained that the Ojibwa material from Rama Reserve, northwest of

Toronto, should be used to illustrate his answer rather than analyse his data. He was

particularly interested in the animal images, referring to them as "symbolic mediators of a

paradox which characterises territorial hunting and gathering societies" (ibid.: 64). Images

were stylised since their meaning occurred as a consequence of its association with different

"symbols and symbol sets" instead of being a point of reference (ibid.: 67). Two sub-styles

were identified in the depiction of the animal images that were not easily recognisable. He

claimed that those naturalistic petroglyphs that were "stylized through simplification"

represented the paradox of location and mobility (ibid.: 68), while those animal images that

Page 196

were simple in design but distinguishable from the first group because traits had been added

that were "clearly not natural" (ibid.: 69), represented the intellectual paradox. The second

group had arisen as a consequence of the creation of boundaries which separated human

beings from animaIs, on the one hand, and the required dependency on the world of animaIs

to create a social world. Therefore, if specifie design traits indicated that an image was not

normal, or "natural," it meant that the authors of the images referred to ideas about the

supernatural rather than natural beings. Lanoue maintained that the people who occupied the

Lake of the Woods during the historie period were "preoccupied" with questions of

territoriality and desired to situate themselves not just using "a social and spatial grid" but

also by involving "abstract expression."

(b) Analogical Approach:

Cailahan (1998) has applied the neuropsychological "model", developed by Lewis­

Williams and Dowson (1988) to this geographical region of North America but only to sites

in Minnesota (Cailahan 1998). The rarity of this approach is interesting, given its popularity

elsewhere in the world.

(c) Homological Approach:

This is the most popular approach for archaeologists interested in determining the

meaning of rock image sites since meaning was drawn by comparing archaeological data with

related historie and ethnographie cultures (Conway n.d. a, n. d. b; Conway and Conway 1989,

1990a, 1990b; Dewdney 1970b, 1970c; Lévesque 1968; Lipsett 1970; Molyneaux 1980b,

1980c, 1987; Tassé 1977b; Turner 1979; Vastokas and Vastokas 1973).

Lipsett (1970) compared the images of specifie pictographs with images created and

used by the Ojibwa, as discussed in Hoffman's (1891) report on the Midewiwin Society and

the different types of religious practitioners among the Ojibwa. Molyneaux (1987) took a

similar approach to the images he found at several pictograph sites in Northern Ontario,

including DhKo-l and DhKm-5 in Whitefish Bay, in the Lake of the Woods. Dewdney

(1970a) developed the idea that by comparing the style of rock images and those made by the

Midewiwin on birch bark scroils, it might be possible to distinguish between historie and

precontact images. He and Kidd (1967: 171) had argued that the images on the birch bark

Page 197

scroils would probably provide clues to the meaning of the images on the rocks ln the

Canadian Shield.

Dewdney and I<idd's (1962, republished 1967) study was arguably the first

comprehensive investigation of rock image sites of the Canadian Shield. Dewdney examined

twenty pictograph and petroglyph sites in the Lake of the Woods. He also posited a

connection between the images on rocks and the Midewiwin, based on the similarities in

morphs executed at rock image sites and on the birch bark scroils (1967: 106-110).

Numerous sites were described in varying degrees of detail (ibid.: 22). Unfortunately, detailed

interpretations of each site were not provided, although much of this information was

retained in Dewdney's private archive, which is now kept by the Royal Ontario Museum.

The age of the images was estimated from their style. Dewdney and I<idd speculated

as to why the sites were painted and who the artists were (ibid.: 16-20, 102-116). They

hypothesised that sorne of these sites depicted articles of European origin such as "an early

European fort" and "the f1ag" (ibid.: 48). The distinction between explicit "description" and

low-level "interpretation" is blurred in Dewdney and I<idd's descriptions of the different

images at each site. An example of this exists for site # 106 on the shores of Obaibikon

Channel, on Sabaskong Bay, now cailed DhKn-1, in the Lake of the Woods (figure 3.11).

Dewdney wrote about this site:

Speeding up the channel from Sabaskong Bay we caught a glimpse of red through the trees high up on the east shore. On shore, expecting to find another example of iron stains, we were happily astorushed to discover the paintings shown here: two serpentine figures one with antlers the other with horns, symmetrically facing a large turtle. To the left, rather crudely painted on very rough granite, was a serpent fifteen feet long, with open mouth, ears and three large flippers - a veritable Ogopogo.

A deep cleft between the ledge we stood on and the rock was almost filled with dirt and rubble. Lying on the ground was an ancient, weathered overcoat, and various rags that had rotted beyond recognition.

(Dewdney and Kidd 1962:43)

The "veritable Ogopogo" is visible in figure 3.47 while his "serpentine figures" are visible in

figure 3.48 below.

Page 198

Figure 3.48: The serpents and the turtles described by Dewdney in 1960.

Dewdney often classified the images using Western art historical terms such as

"naturalism", perhaps because he had been trained as an artist and illustrator. His cultural

background clearly influenced rus verbal descriptions of the different sites. Another example

relates to a pictograph site called Painted Rock in Lac La Croix:

For sheer naturalism there are no other paintings of moose that l have seen in the Shield country ta compare with the three on this site. Ail are surely by the same hand, as is the little antelope-or deer. Unique, too, are the pipe-smoking figures: one beside an hour-glass figure and tracks, the other not far from the initiaIs "L.R. 1781".

(Dewdneyand Kidd 1967: 26-27)

Dewdney and Kidd each dealt with separate components of this study. Dewdney

located, recorded, and described the pictograph sites and occasionally obtained sorne

ethnographie details (ibid.: 11-14), while Kidd considered and examined the related

anthropological data (ibid.: 159-177). However, trus split between the work undertaken by

Dewdney and Kidd is misleading: it may simply reflect academic traditions at the time their

book was written. Yet, throughout, Dewdney speculated as to the authors of these images,

why the images existed, their meaning, and the paint ingredients, prior ta discussing the

different sites and their respective paintings. Kidd (ibid.: 159-177) summarised the rustory of

the interest in these images, the archaeology and the early rustory, and considered the

Algonquian-speaking peoples in the Canadian Srueld. He, like Dewdney, speculated as to

why the images were created, the possible authors of these images, the paint's constituents,

Page 199

and the techniques of applying these images, as well as different possibilities for establishing

their meanings. Neither I<idd nor Dewdney considered that the meanings of these images

might be polysemic, despite I<idd's discussion of the Midewiwin Society and its members

(ibid.: 167-168).

The second (1967) edition contains considerably more information, but suffers from

the same problems. Dewdney and I<idd's study represents the first major step in the

homological analysis of these images. Although Dewdney was not an archaeologist, his work

on these images was immensely influential.

Dewdney (1970b) used ethnographic evidence concerning shamanic practises among

the Ojibwa to postulate the reasons for the creation of pictograph sites. He illustrated his

arguments using individual images from four sites in Whitefish Bay, Lake of the Woods:

DiKm-3, DhKm-3, DhKm-S, and DiKm-4. He drew upon information regarding the

worldview and religious practices of the Algonquian-speaking peoples collected by

anthropologists, and from his numerous conversations with indigenous informants, including

Norval Morrisseau, the Ojibwa artist, to try to determine how and why these paintings were

created, and what they meant (ibid.).

Dewdney argued that the pictographs, like the other images created by the

Algonquians on wood, birch bark, and hide were all forms of expression (ibid.: 22) that he

subdivided into three categories: secular, tutorial, and visionary. The images on the birch

bark were deemed secular and possibly included those le ft at portages for others, to give

practical information such as an individual's 'signature', a representation of his totem or his

name" (ibid.). He maintained this mark or image would have also existed on personal objects

that belonged to individuals. Yet, since Algonquian society was collective, an individual

would only have left a "mark" for territorial reasons. Tutorial images were defined as those

that were used between a shaman and his clients or those used as mnemonics on the birch

bark scrolls (ibid. 23). Dewdney deemed sorne images as visionary, based upon his interviews

with older generations of Ojibwa.

Dewdney (1970b) asserted that the description and classification of images grouped

together by their physical shape was problematic, since images should not be examined

separately if they were originally part of a larger group from a site. This daim occurs in an

article that seems to have been part of a larger research project that he was undertaking prior

Page 200

r-..

~ ..

to ms death. It seems evident from reading ms research and field notes3 that Dewdney may

have been synthesising the information he had collected. He had noted that twenty-six

percent of the sites that he had recorded were dry sites and seventy-four sites were wet sites.

What he meant by the terms "wet" and "dry" is undear. It is also dear that he sought to

identify ail the different variations of each type of image he had recorded. He provided the

number and gave the name of the pictograph site, if it had one, where each image existed.

What precisely Dewdney intended to do with this analysis remains unknown.

He was also interested in whether the type of image he had identified at each site

existed only at certain orientations, or in certain parts of a pictograph site and whether tms

changed according to where in the Canadian Smeld the site was physically situated. This

work is possibly related to some ideas that Dewdney (197 Sa: 1) expressed in an

Archaeological Newsletter that perhaps understanding how the stylistic variations were

distributed throughout a region provided chronological dues to dating the images of the

Shield region. He argued that he initially posited that using a computer provided the most

objective means of sorting and correlating the images which he had collected over the course

of ms research. He posited that he had approximately 3,000 figures which he termed morphs

(ibid.: 2). He realised that although he had used a computer that "an inadequate typology,

loosely defined coding values and inconsistencies (Plus too many errors)" led mm to

condude that higher standards were required in the initial phase of a study (ibid.). He

concluded ms short piece by reminding his colleagues that it was important to be as

meticulous and objective as possible and that he was guilty of being too subjective when

recording and describing ambiguous images (ibid.: 4).

It is dear from ms notes, held by the ROM, that he was conducting sorne systematic

analysis using a computer with colleagues at the University of Western Ontario. Another line

of research that concerned mm in ms notes is mentioned in an ROM Archaeological

Newsletter dated 1971. He discussed his observed correlations between specific pictograph

sites and "vision pits" which were special places that he believed were chosen by shamans

for their vision quests (ibid: 3). He posited that DhI<m-3 was possibly one such example of a

place that a shaman had chosen as a "shaman's retreat," especially given its physical

characteristics: the images were beside a bench "large enough to pitch a tent on, and far

3 Dewdney's field notebooks, field drawings, slides and photographs are held by the Department of Anthropology, Royal Ontario Museum, in Toronto Ontario, Canada.

Page 201

above the water" (ibid.: 4). However, he concluded that further research was required to

substantiate his opinion that pictographs sites were probably connected with "vision pits"

(ibid.). Clearly severallines of research were being explored and developed for some purpose

that is now unclear and difficult to establish.

Vastokas and Vastokas (1973) attempted to understand the mearung of the

Peterborough Petroglyph Site. These images were pecked from crystalline lime stone, one of

the softest rocks in the Canadian Shield, and ground out using gneiss hammerstones (ibid.:

14-18). The petroglyphs were speculatively dated, based on available data, to the end of the

Woodland period, between A. D. 900 and 1400 and attributed to prehistoric Aigonquians

(ibid.: 20-27). The Vastokases postulated that a connection existed between the engraved

images on the rock faces and the images on birch bark scroils created by the members of the

Midewiwin Society (ibid.: 40-47). They surmised, from ethnohistorical and anthropological

sources, that the images were located where individual shamans believed they could consult

the resident spirits (ibid.: 47-54) and that unusuaily shaped large rock outcrops, rocky hiils

and boulders with crevices and holes were the "dwelling-places of the manitous and mythical

creatures such as the Maymaygwayshi" (ibid.: 48).

Vasto kas and Vastokas discussed at considerable length how the Algonquian universe

was organised and how the physical locations of the rock images in the landscape were key

loci, since these places were where the different worlds co-existed and hence these places

were "the seat[s] of visionary experience and super natural counsel" (ibid.: 54). They (ibid.:

29-40) discussed the nature of Algonquian hunting and gathering practices and the different

objects upon which similar images appeared, and maintained that the manner in which

shamanism, Algonquian cosmology, manitous, vision quests, and guardian spirits were

interrelated provided the information required to understand the petroglyphs. They

postulated that a detailed understanding of Aigonquian use of pictographs to convey

information would aid modern acadernics to understand these images.

Images were analysed once they had been classified by type and placed in different

broad groups according to subject matter (ibid.: 55-129). The Vastokases noted that,

although over nine hundred glyphs existed at the site, only three hundred were readily visible.

The others were deemed "vague and their identity is difficult to de termine owing to the

degree of overlap and superimposition, wear and erosion of the limestone surface" (ibid.:

55). They subsequently asserted that the variety and extent of the images' meaning was only

Page 202

accessible once detailed analysis of this site was completed (ibid.). Interpretation and

discussion of the possible meanings of each group could occur once the identifiable images

had been classified by shape and subject matter, and with reference to ethnographic and

ethnohistoric data. Three years earlier, Dewdney (1970b) had asserted that the description

and classification of images that were grouped by type (their physical shape) was problematic,

since images should not be examined separately according to their physical shape; especially

if they were originally part of larger group of images which were different.

It was possible that the meanings of the different images were interrelated. Based on

the changes in the styles of the images, the Vastokases (ibid.: 131-134) concluded that the site

had probably been used repeatedly over a long period of time. They speculated that these

stylistic changes were possibly connected to technological, cultural, or temporal factors. They

(ibid.: 137) asserted that the images of the Peterborough Petroglyph site should only be

compared with other petroglyph rather than pictograph sites to establish stylistic affinities.

However, this statement is problematic, because it is possible that the crea tors of these

images did not differentiate between images ground into stone and those painted on stone.

The Vastokases concluded that the rock itself, the landscape, was crucial to the meaning for

these images and to understanding why it was used for these images. Although this study is

over thirty years old, many of its conclusions remain important to the analysis of these

images today.

Both of Conway's (n.d. a. and n. d. b) papers examined sites in northeastern Ontario.

He (n.d.a) was concerned specifically with three morphs, from various pictograph sites in

northern Canada, which occurred in conjunction with each other: an open armed man, his

canine companion, and an animal pelt. Another morph, which Conway identified as a beaver,

the pelt, existed above the man and his animal companion. He identified these images as

specific star constellations. The open armed man was identified as Orion since the paintings

had anatomical attributes that were similar to those of the constellation. The position of

canine companion, identified in the pictograph sites as either a dog or wolf, concurred with

that of Canis Major. The bilateral symmetry and location of the beaver matched that of

Gemini to Orion (ibid: 1). Conway (ibid. 1-2) drew upon ethnographic information collected

by Schoolcraft, Speck and some other unnamed individuals, and from his own informants, to

identify specific constellations as animaIs and supernatural heroes. Schoolcraft was important

for Conway, because he was particularly interested in native astronomy and cosmographic

Page 203

~ ..

information. Conway Qbid.: 3) maintained, based on unknown sources, that native

astronomy depicted at pictograph sites had "considerable antiquity" since according to him

such paintings had ceased to be created after 1650 A. D. by Ojibwa shamans and medieine

people. He asserted that few sites showed historie scenes involving Europeans, but provided

little evidence to substantiate this daim. He argued that the style changes and presence of

constellation groups in the pictograph sites indicated that native cosmographie information

had existed since the prehistoric era. Fourteen constellations were identified, but Conway

was specifically interested in the Orion trio whieh he identified as existing on a birch bark

scroll made for Schoolcraft by a Lake Superior shaman as weIl as in some unnamed

"examples of Ojibwa rock art sites" (ibid.: 4).

Conway (ibid.: 4-5) asserted that the "Orion trio of paintings" was explicable using

information from popular native folklore. However, he argued that it was possible to put

forth another interpretation of these images, since the constellation paintings behaved as a

mechanism by which seasonal changes were indicated to the local indigenous population.

Orion was interpreted as a celestial marker for "a shift from late fall to winter camps" (ibid.:

5). The depiction of constellations at pictograph sites indicated that these sites were "sky

maps" which enabled information about seasonal migrations through the landscape to be

transferred from one generation to another. Conway conduded that rock image sites should

be considered as "one of the few permanent markers" in the lands cape that provided

knowledge to the transitory Ojibwa hunter-gathers of this region.

Another paper by Conway (n. d. b) is perhaps a later draft of the paper previously

discussed (n. d. a) as its content is similar. In this second paper, he hypothesised that specific

images were based upon the constellations recognised by prehistoric Algonquians. These

images originated from sites in an area occupied by the historie Ojibwa, Cree and other

Algonquian bands (ibid.: 5). Conway argued that this combination of morphs was first

identified at a site by himself in Matagamasi Lake, in northem Ontario (ibid.: 7). He

subsequently briefly discussed Dewdney's published works, the Vastokas and Vastokas

(1973) publication, as weIl as various site reports from throughout Ontario in an attempt to

establish whether this triad of images existed. Conway maintained it was easier to identify

whether these three images existed together if the painted figures were "not as dustered"

(ibid.: 8). This triad of images existed in only two clusters in the Temagami stylistic and

Shield Edge stylistic areas (ibid.: 9). It was absent elsewhere in the Upper Great Lakes area

Page 204

and north central Ontario. Unfortunately, Conway did not define what constituted the

Temagami or the Shield Edge styles. Two types of canines were identified to exist on their

own or with a human figure: a dog and a wolf. Two examples of the dog morph were

identified at two sites in the Lake of the Woods: DiKo-2 and DhKm-3. Wolf morphs were

identified at one undisclosed site in the Lake of the Woods. But Conway did not establish

how wolf and dog morphs were different from each other.

Conway's conclusion had four points. Firstly, two distinct canine morphs existed in

the rock images sites in Ontario, which Conway identified as a wolf and a dog. Nevertheless

both images could denote "the same Algonkian concept" (ibid.: 12). Secondly, both morphs

had comparable limited geographical distributions, namely the northeastern edge of the

Canadian Shield in Ontario and north and west part of the Lake of the Woods in north

western Ontario. Thirdly, dog canine morphs often occurred with hum an figures while wolf

canine morphs existed with human figures that had their arms open. Conway's fourth

conclusion was that "an Algonkian triad" existed at the majority of sites that had the wolf

morphs, and consisted of "a human figure or shaman figure with outstretched arms, a pelt or

a cross" (ibid.: 12). Conway's conclusion is interesting since it demonstrates that his own

Eurocentric cultural perspective was influential in his choice of images to examine, despite

his utilising the pertinent ethnographie evidence from the region where the sites existed.

Fol1owing Hodder (1987), Rajnovich (1994: 160) argued that it was possible to arrive

at sorne conclusions that could not "be absolutely proven" (Rajnovich after Hodder 1987: 6)

but were plausible. Rajnovich (1989, 1994) repeatedly drew on the Lake of the Woods as a

source of examples for her discussion of rock image sites in the Canadian Shield. She

postulated that archaeological, ethnographie, and anthropological evidence should be

combined with linguistic and oral evidence to establish the link between the different types

of Ojibwa shamans, Midé and Jiissakid, to obtain an understanding of these images. She

(1994: 19-22) asserted these images were signs, probably mostly polysemie metaphors, since

Algonquian languages are metaphorical. She (ibid.: 22-39) discussed the Algonquian world

view, the practice and use of indigenous plant medicines, and the different types of religious

practitioners. Rajnovich (ibid.: 29) maintained that only the Midé and Jiissakid, two of three

types of religious practitioners, rather than the Wabeno, created rock images.

She appears to base her arguments on two rock image sites, providing a field drawing

of each site: Dewdney's field drawing of Painted Rock Island, DhKo-1, in the Lake of the

Page 205

Woods and a petroglyph site in Uskik Lake in Saskatchewan. She (ibid.: 41-55) debated the

age of these images and concluded that it was only possible to use relative dating techniques

and to draw upon ethnographie, and related archaeological data from nearby peoples to posit

the age of individual sites.

She (ibid.: 57-63) related the images present on the Qjibwa birch bark song scroils to

those on the surface of rocks by drawing on the work of Garrick Mailory (1880, 1881 and

1884), Tanner Games 1956), Hoffman (1888, 1891, 1896), Kohl 1985 [1860], and Densmore

(1910, 1913, 1926, 1979), as weil as sorne unpublished studies in the Smithsonian

Institution). Rajnovich (1994: 57-63) postulated that the metaphorical nature of the

Algonquian languages provide clues to the interpretation of the paintings and engravings.

She (ibid.: 65-66) argued that the pictographs were standardised in the sense that other

people could recognise the images and interpret them but that different and legitimate

meanings could be attached to specifie symbols.

To demonstrate why and how these images were polysemie, Rajnovich first discussed

the importance of the cliff faces and then a range of images with reference to ethnographie

sources and informants (ibid.: 66- 143). Each type of image was analysed and interpreted

after it had been grouped into a broad shape type. Meaning was ascribed to an individual

shape through references to data from ethnographie and ethnohistoric materials, as weil as

from informants.

She ignored Dewdney's (1970b) stricture that images should not be interpreted

individuaily from the other images with which they exist. Her practice of first dividing images

into broad groups also may cause problems. This activity is probably related to the manner in

which rock image sites are treated by archaeologists accustomed to coilecting data according

to archaeological conventions and guidelines. It is surprising that her discussion of these

images did not draw on the conclusions about these images that she had published earlier in

government reports (for example Rajnovich 1980b, 1981b).

Rajnovich (ibid.: 145) observed that a common explanation for the presence of these

images was "hunting-magic", (ibid.: 145- 157) and discussed at great length what this could

mean with reference to the Algonquians of the Canadian Shield. She subsequendy argued,

based upon ethnographie data, that rock images appeared on rocks where manitous lived

(ibid.: 160), and that these sites occurred beside water because that was where the sky, earth,

water, underground, and underwater meet. Places su ch as deep lakes, whirlpools, bases of

Page 206

.~.

lakeside cliffs, caves, and crevices were the locations used to cross from one world into

another (ibid.: 35). For Coleman (1937:34) these places were where the manitous lived and

where the manitous and the medicine people could meet.

Rajnovich asserted that paintings were done principaily using red ochre because this

mineraI was "a powerful medicine itself and used in sorne cures" (ibid.). Specific images were

painted because they referred to particular experiences of the medical practitioners of the

Algonquians, the Midé, since they created the images on the birch bark scroils. The images

were painted on rock surfaces, where these medical practitioner people wanted "to leave

lessons" (ibid.: 161). Rajnovich (ibid.: 163) postulated that the paintings in this region were

"intimately" connected to Algonquian beliefs, legends, and songs. She concluded that the

best approach for those wishing to establish the meanings of these images was for

information from the Algonquian people's history and philosophy to be combined with

ethnographic data. Although Rajnovich's (1994) study has sorne flaws, it remains the most

recent and sustained contribution to research to date on the meaning of rock image sites of

the Canadian Shield.

Sorne archaeologists have used Rajnovich's approach to analyse specific rock image

sites from other parts of the Canadian Shield, but not from the Lake of the Woods

(Hamilton 2000; Pettipas 1991 b). They have explored the concept that different bodies of

ethnographic information could aid archaeologists in answering specific questions regarding

these images. It is important that the archaeologist also acknowledges the problems of using

ethnographic evidence from present day informants to interpret images from the

archaeological record. Jones (1981b: 46-47) maintained that ethnographic information from

Northern Saulteaux and the Cree groups aided the "empirical concerns of archaeologists"

seeking to establish the binder used to create the paint to make pictographs. Wheeler (1977a,

n. d.) argued that attention should be paid to studies of the structure of the Cree and Ojibwa

languages where the May May Quah Sao (often cailed Maymaygwayshi), who were spirit

helpers, were associated with stones in the sacred stories told by Algonquian-speaking

peoples (n.d.: 1). The word "stone" belonged, Wheeler asserted, "in an animate grammatical

category in Algonkian languages" that was "linguisticaily distinguished as that which interacts

with man and that which does not" (ibid.). Wheeler (n.d.: 5) recorded a rock image site on

the Semple River near Oxford House in northeastern Manitoba including the legend

connected to this site by informants. Information existed, Wheeler noted, regarding the

Page 207

relationship of these spirit helpers and rock image sites ln the ethnographie and

anthropologicalliterature such as such as that produced by Dewdney and Kidd (1962; 1967)

and Landes (1968), while others (Stevens 1971) had recorded stories that connected these

spirit helpers to rock image sites.

Two more in-depth studies of two large rock image sites must be mentioned since

they are found in the Canadian Shield, although neither site is in close proximity ta the Lake

of the Woods (Conway and Conway 1990b; Vasto kas and Vastokas 1973). Conway and

Conway were archaeologists based in the Sault Ste Marie region of Ontario and, so, this

study reflects the ethnographic and historical details of this area. Bath archaeologists asserted

that the paintings at the Agawa rock images site on the north shore of Lake Superior were

created by "Ojibwa Indian shaman artists" (ibid.: 7) who in the indigenous communities were

the link between the human world and the spirit worlds. They (ibid.) argued that Agawa was

"an exceptional site" with several components.

Conway and Conway (1990b) drew upon information from a variety of sources

including Schoolcraft's published date for the images, and from several Ojibwa elders,

including (a) Fred Pine, who Conway identified as a "noted Ojibwa tribal elder" born in 1897

and Shingwauk's great-grandson (ibid.: 19), (b) Chief Norma Fox from the Cockburn Island

band of Lake Huron (ibid.: 21), and (c) Bill Sheskekwin who lived in the Dog Lake area of

the Upper reaches of the Michipicoten River. Conway and Conway argued that their

informants enabled them to establish that several individuals had painted the images at

Agawa, including: a 17,h century Amikwa band leader called Myeengun (the Wolf) and a 19th

century shaman called Shingwauk, or 'The White Pine', from Lake Superior.

Conway and Conway argued that, because the water level of Lake Superior was

higher than the painted are as of the cliff over three thousand years ago, the paintings had to

be younger. They estimated that the most visible paintings were created during the last five

hundred years. Conway and Conway (ibid.: 7-8) posited that the images were created for

religious reasons and were the "end products of religious experiences such as vision quests,

group ceremonies, and acknowledgement of spiritual assistance," and did not directly

chronicle episodes from the pasto

The images are examined in seventeen "panels" (ibid.: 15), and as one hundred and

seventeen individual images (ibid.: 8, 15). Conway and Conway asserted (ibid.: 15) that the

panels were "numbered from north to south (left to right)" if an individuallooking at the site

Page 208

was in a canoe in front of the cliff face. They established the meanings of the pictographs of

Agawa Bay with reference to ethnographic and ethnohistoric data as weil as information

from Fred Pine, Chief Norma Fox, and Bill Sheskekwin. The images on each panel were

described and identified in considerable depth and each panel was interpreted. It is, however,

unclear whether the division of the Agawa rock image site was based on information gained

from their indigenous informants or not.

The examination of pictograph and petroglyph sites has developed and advanced

over the last hundred years. Considerable time, effort, and money have been spent on the

development of the techniques to physicaily record and date them, and many have tried to

interpret the se images. People want to find out how to discover more of them, satisfactorily

record the images, and preserve them for posterity. Yet, the most important goal is to

establish what these images may mean.

Different people have clearly taken different approaches, although it is clear that the

homological approach is more popular amongst those determined to establish the meaning

of these images. The current understanding of these sites in the Lake of the Woods is that

they were created by shamans, ritual specialists, in key places in the landscapes where the sky,

the underground, the underwater, and the earth meet and where the traveling soul of the

shaman could cross from one world to another.

Conclusion

The goal of this chapter was to present and discuss the work undertaken prior to my

own by previous researchers concerned with the pictograph and petroglyph sites in the Lake

of the Woods area. l determined to establish the most popular philosophical approach, what

previous researchers have thought of pictograph sites, how they examined what they were

interested in, and to understand the origins and the reasons behind the terminology used to

identify the images found. This critical review of the literature demonstrated that a large body

of data ranging from conservation of the paints used in pictographs and the different ways of

examining them (e.g. Myers and Taylor 1974; Wainwright 1990, 1997), how to establish the

age and style of specific pictograph and petroglyph sites in the Lake of the Woods (e.g. Reid

1979 and Steinbring 1977), the context of the physical surroundings of the images (e.g.

Noble 1968), to whether particular images only exist together and how to establish their

meaning (e.g. Cailahan 1989; Conway and Conway 1990b). Much of the literature was

Page 209

obviously concerned with the sites of the Lake of the Woods but it is important to consider

analogous studies of sites elsewhere in the Canadian Shield. Sorne of the literature is

descriptive (e.g. Lawson 1885), and sorne is concerned with technical issues, but much of it

was taken using several of the five approaches that exist in archaeology: culture-rustorical,

contextual, intuitive, analogical, and homological. As I argued in Chapter I, it is important to

initialiy use the culture-rustorical approach, then the contextual approach, and, if one is

concerned to establish the meanings of the images then either the intuitive, the analogical or

the homological approach should be used. Several archaeologists have used only the culture­

historical approach (e.g. Grant 1967; Reid 1979; Steinbring, Danziger and Caliaghan 1987;

Steinbring and Caliaghan 1985). Others utilised the contextual approach (e.g. Arsenault 1997;

Pohorecky 1968; Noble 1968). Many were determined to establish the meaning of the images

they found. The homological approach is the most popular approach for archaeologists (e.g.

Dewdney and Kidd 1962; Rajnovich 1994, Vastokas and Vastokas 1973), wrule the intuitive

approach is the second most popular approach (e.g. Granzberg and Steinbring 1995; Lanoue

1989). Only one archaeologist utilised the analogical approach (Caliahan 1989).

A considerable quantity of fieldwork and research has been conducted on the

pictographs of the Lake of the Woods, and my field work in 2001 should contribute to trus

larger body of information. It was for trus reason that prior to conducting my own fieldwork

I summarised ali of the information available for each of the pictograph sites pertaining to it:

alternative names; whether photographs were taken; how was it recorded; measurements if

taken; location of field recordings and published works making reference to the site; the

name of each researcher who examined it, with the dates and rus profession; rock type;

location of the sites; whether offerings had been le ft, and whether any statements had been

made as to its possible cultural affiliations (Trus information is in Appendix 1.). Ali of this

information was required so that I knew who had conducted research, what type of analysis

they had undertaken, and at wruch pictograph sites in the Lake of the Woods.

Previous researchers have established several key pieces of information about

pictograph sites. Pictographs occur on cliff faces while petroglyphs occur on horizontal

surfaces throughout the Canadian Shield, both in close association with bodies of water.

Images in caves are relatively rare and therefore little information has been accumulated

about these sites. Previous researchers agree that the paint used was made from hematite

mixed with an organic binder wruch cannot be identified. It is also agreed that these images

Page 210

occur in places that usuaily have ail of the foilowing five physical features: rock benches

(sometimes cailed shelves), overhangs, large cracks, large crevices, and holes. Absolute dating

techniques cannot be utilised, therefore, relative dating techniques are usuaily used to

ascertain the age of these images. Sorne researchers have recorded the quantities of lichen

and rock tripe growth and encroachment, exfoliation, and the accumulation of white mineraI

deposits present at each site. This information is often compared with the orientations of

these sites. Pictograph sites on cliff faces tend to face south, east, and occasionaily west. Very

few sites face north. Previous researchers have described the images they found, and

attempted to organise systems of stylistic classification to aid in establishing local and

regional patterns in the Canadian Shield. But there is no consensus about which classification

scheme is best used to describe the images. It is acknowledged that the description of these

images is highly subjective, since the vocabulary used influences their interpretation, and each

researcher tends to use their own system. Site description has also influenced the manner in

which a site can be subdivided into parts, facilitating its description but causing problems for

subsequent archaeologists. Researchers have observed that pairs of the same image often

occur at the same site.

Ali researchers agree that it is unclear which ethnic group created these images, but

they agree that the images were created by the Algonquian-speaking peoples. Consensus

exists regarding the importance of the sites to the current local indigenous peoples, who

consider these places as important and integral to their own world view. Objects left at sites

as "offerings" have always been recorded by previous researchers. Consensus exists

regarding the polysemic nature of these images, and that they were used to convey

information to other people. Ali researchers concur that the images of the pictograph sites

are similar to those on birch bark scroils made by the Midé ritual specialists. Hence, everyone

has posited that a detailed investigation of the images on birch bark scroils, the ethnographic

record, and the pictographs might provide sorne answers as to the meanings of the images

on the rock faces.

Unfortunately, l was only able to examine the pictograph sites in 2001 because the

water level was far too high for me to survey and examine the petroglyph sites. It was equaily

important for my research to foilow the three levels outlined in Chapter 1. Therefore, l used

the culture-historical approach, then the contextual approach, and finaily the most pertinent

of ail the approaches that should be used to establish the meaning of these images. Each

Page 211

subsequent chapter de aIs with and discusses the data that l collected according to this

sequence. However, the next chapter, Chapter 4, presents the basis for making the decisions

that l made while conducting field work and how l identified and classified images that were

found. The goal of my fieldwork was to systematically collect detailed information about

rock image sites over a portion of the Canadian Shield in order to better understand the

nature of the evidence that was available. Each of the subsequent chapters will deal with the

archaeological data according to each of the different philosophical approaches that must be

applied to this data in sequence as outlined in Chapter I.

However, l do not use the intuitive approach, sin ce l think that it is important ta

articulate to other researchers how my conclusions were reached. l do not think that readers

should be reliant on the whims of the author, and they should be able to understand

precisely how an interpretation was reached. l think it is important for discussions and

arriving at conclusions to use vocabulary that is easy to understand. If an archaeologist is

determined to establish the meaning of these images in a more rigorous and persuasive

fashion, the analogical or homogical approaches should be adopted (see Table 1 in Chapter

I). Chapter V presents how the culture-historical approach is applied to this data, and seeks

to establish the total range of the images. Chapter VI presents the conclusions that are drawn

once the contextual approach is described, and considers whether combinatory rules exist

regarding the relation of the different sorts of images to one another. Chapter VII uses the

most applicable approach concerned with meanings to consider how to establish the

meaning of these images, and whether these images are related to others outside the rock

images sites.

Page 212

Chapter IV: My Work on the Rock Image Sites.

This chapter presents the ground work required to collect the data I needed on each

of these sites. It presents the basis for my decision making while conducting field work, and

how l identified and classified the images found. Through fieldwork, I systematically

collected detailed information about rock image sites over a portion of the Canadian Shield

in order to better understand the nature of the available evidence. This required not only re­

recording the known sites, standardizing and verifying current knowledge, but also locating

as many unknown sites as possible. This data will enable me to achieve the four goals of my

thesis to (a) identify the possible vocabulary of images, (b) determine whether combinatory

rules exist, (c) reconstitute the life history of each site, and (d) ascertain whether the images

can be related to other images outside the rock image sites to determine if this can provide

information about the meaning(s) of the rock images.

Therefore, in 2001, l conducted three and a half months of fieldwork on the

pictograph sites in the Lake of the Woods area located between North latitudes 49° 00" - 49°

50" and West longitudes 93° 53" - 95° 09". I implemented a newly developed method of

collecting survey and field data on the pictograph sites in the hope that perhaps new

pictograph sites could be found. l based my research on my experience, gained conducting

fieldwork in the Canadian Shield region since 1990; and working on conventional