Davidson’s Objective - Minerva Access

Davidson’s Objective

2

A man breaking his journey between one place and another at a third place of no name, character, population or significance, sees a unicorn cross his path and disappear. That in itself is startling, but there are precedents for mystical encounters of various kinds, or to be less extreme, a choice of persuasions to put it down to fancy; until– ‘My God,’ says a second man, ‘I must be dreaming, I thought I saw a unicorn.’ At which point, a dimension is added that makes the experience as alarming as it will ever be. A third witness, you understand, adds no further dimension but only spreads it thinner, and a fourth thinner still, and the more witnesses there are the thinner it gets and the more reasonable it becomes until it is as thin as reality, the name we give to the common experience... ‘Look, look!’ recites the crowd. ‘A horse with an arrow in its forehead! It must have been mistaken for a deer.’

– Tom Stoppard, Rosencrantz and Guildenstern Are Dead

“The horse who knows he is a horse, is not. Man’s major task is to learn that he is not a horse.”

– Elie Wiesel, Souls on Fire

3

Davidson’s

Objective

LANGUAGE &

THE CONCEPT OF OBJECTIVITY

* * *

Ross Barham

Submitted in total fulfilment of the requirements of the degree of Doctor of Philosophy

School of History, Anthropology, Philosophy & Social Studies

The University of Melbourne, Australia

4

Thesis Title: Davidson’s Objective: Language and The Concept of ObjectivityStudent Name: Ross Campbell Barham Student No: 57399 Qualifications: BA (Hons) Grad. Dip. Ed. (Secondary) MA (Philosophy) ORCID ID: orcid.org/0000-0003-0847-8044 Subject: PhD – Dept. of Philosophy (part-time) Course Code: 161-701 Submission: 14th February, 2018 Institution: The University of Melbourne, Parkville, Victoria, AUSTRALIA 3010 Faculty: Arts School: The School of History, Anthropology, Philosophy and Social Sciences (SHAPS) Department: Philosophy Supervisors: Dr François Schroeter

Dr Laura Schroeter Dr John Armstrong

Publications: NA

This is to certify that: – the thesis comprises only my original work towards the PhD; – due acknowledgement has been made to all other material used; and, – this thesis is fewer than 100,000 words in length.

__________________________

Ross Barham

5

In Loving Memory of

Dr Grant Lester Barham MBBS

(1946–2007)

* * *

Knowing how close we were

Is how I find myself

6

Abstract* * *

The thesis critically examines Donald Davidson’s claim that language plays a non-trivial role in explaining possession of the concept of objectivity. After showing that a priori arguments do not establish this claim, different versions of Davidson’s triangulation argument are developed and found wanting. Drawing on Michael Tomasello’s work on joint attention and Robert Brandom’s Inferentialist account of objectivity, the final chapter suggests how a naturalistic account of language and triangulation may nonetheless be essential to explaining possession of the concept of objectivity.

7

Preface

* * *

To help get our heads around what this thesis is about, permit me to tell a brief children’s story that most will likely already be familiar with in some guise or another:

A child is entrusted to care for her neighbours’ goldfish while they are away. Unfortunately, the fish dies and so the girl surreptitiously replaces it with an identical looking substitute. The subterfuge is discovered and the child subsequently comes to appreciate the importance of honesty.

The relevance of the above story to this thesis is how it so effortlessly and yet so strikingly conveys the idea of the concept of objectivity. For the reader to understand that the child is able to replace the original fish with an undetected surrogate, requires not only an appreciation of the neighbours’ susceptibility to not discern any difference between the two fish, but an understanding that the replacement fish is not the fish that the neighbours mistakenly presume it to be. The concept of objectivity is in play insofar as the story conveys a ‘disconnect’ between the way the neighbours take the world to be and the way the world actually is.

To further emphasise what’s at stake here, let us compare the above story with another: here, a child is entrusted with a house key that she subsequently loses and replaces with an identical copy. In this new story, irrespective of whether the child does or does not confess the original key was lost, the replacement of the original key doesn’t have the same significance as it did in the case of the fish. It’s easy to see why this is so if we turn our attention to the lock to which both keys open. After all, the lock doesn’t know or care if it is

8

one key or the other that engages its mechanism; so long as whatever is inserted can open the lock, then that key is effectively ‘the key’ to the lock. With humans, however, things are different. The owners of a pet goldfish may care and, so, might want to know whether the fish in their bowl is the fish that they take it be. (Equally, the owners of a key may have some sentimental attachment to it, such that they would care if it were replaced, even if the lock remained indifferent.)

Of course, as the story makes plain, humans are not immune to being mistaken with respect to what is objectively the case, but even so, the possibility of such a mistake arises only insofar as the concept of objectivity in play. That is to say, to have been tricked, the holiday-goers must think of the fish in the bowl as their fish in particular, regardless of whether it is in fact the fish they take it to be. We all do this – we readily apply the concept of objectivity to our possessions, relationships, values, beliefs and much more. We do so insofar as we are able to think as we do, that ‘This is my home’, ‘S/he’s “the one’’’, ‘Murder is wrong’, ‘Water is H2O’, etc., etc.

In light of this capacity we share, a question arises: what is it that makes humans in this way different from locks and the like? This is, then, very broadly the issue I will be concerned with in this thesis. Putting it another way, I am interested in asking:

How is it that we possess the concept of objectivity?

Happily, numerous thinkers have already made substantial progress towards, both, better framing this question and exploring various ways it might eventually be answered. One philosopher in particular – namely, Donald Davidson (1917–2003) – has undertaken some of the most germinal (albeit not uncontentious) work in this regard, and it will be primarily his philosophy that I will look to in attempting to further develop our understanding of these matters.

9

Acknowledgements

* * *

My supervisors, Laura, François and John.

For support, advice and criticism: Zach Weber, Vicki Macknight, Kemran Mestan, Tama Coutts, Tristram Oliver-Skuse, Dave Ripley, Larelle Bossie, Mike & Maria Ashcroft, Kai Tanter, Chloe Mackenzie, Kate Phelan, Shannon Brick, Michael Tomasello, Joseph Rouse, Endre Begby, Graham Priest, Brian Scarlett, Barry Taylor, Damon Young, Dylan E. Wittkower, Tim Chandler, Keith Windschuttle, Adrian Currie, Derek Turner, Raimond Gaita, Tamas Pataki, Peter Slezak, Greg Restall, Karen Jones, Dana Goswick, Richard Paul Hamilton, Gearoid Brinn & Georgina Butterfield, Paddy & Chrissy Leahy, Anna Welch, Ralph Pettman, Yeshe Colliver, Raffaele Rufo, Aaron Goldbird, Catherine Schnell, Blair Mahoney, Sam Crocket, Flip Rusden, Fran Neal, Elizabeth Fitzgerald, Luke Thompson, Sam Bryant, Amanda Carroll, George Marotous, Dan Vine, Janet Prideaux, Monica Bini, Lenny Robinson-McCarthy, the UMMPG, various AAP audiences, and a handful of anonymous referees of papers published outside of the scope of this thesis.

For employment and support: Melbourne High School and my ever-patient students.

For proofreading: Elisa Webb.

My family, especially Joy & Grant Barham, Jack Barham, Brenda & Graham Webb, Elisa Webb, and Coen & Orlo Barham.

* * *

I acknowledge the Wurundjeri of the Kulin nation as the traditional owners – past, present, and future – of the land on which the research and writing of this thesis took place.

10

Contents

* * *

Preface 7

Acknowledgements 9

Contents 10

Introduction 11

PART ONE ­– Thought & Talk

1. Surprise 33

2. Error 65

PART TWO –Word & World

3. Content 91

4. Interaction 125

PART THREE – Saying & Sharing

5. Cooperation 146

6. Inference 176

Conclusion 229

Bibliography 237

11

INTRODUCTION

* * *

… once we come to recognise the distinction between appearance and reality, and the existence of objective factual or practical truth that goes beyond what perception, appetite, and emotion tell us, the ability of creatures like us to arrive at such truth, or even to think about it, requires explanation. An important aspect of this explanation will be that we have acquired language and the possibilities of interpersonal communication, justification, and criticism that language makes possible.

– Thomas Nagel, Mortal Questions

What’s in a name? I remember reading Jacques Derrida’s Who’s Afraid of Philosophy? and thinking that he had started with the book’s title but had somehow managed to work his way backwards from there. Although this thesis is not fashioned in the Deconstructionist mould (if indeed there can be such a thing), I, too, would like to start out by looking back to the title of this thesis so as to unpack its double meaning – albeit with the resolute aim of thereby moving the discussion forward.

I. DAVIDSON’S AIM

The most natural reading of the title, ‘Davidson’s Objective’, is of a goal or aim had by the late American philosopher, Donald Davidson (1917–2003). In this work I will indeed be

12

concerned to further the discussion of what Davidson dubbed the problem of objectivity, to which he devoted much time and thought. In this light then, this thesis is aptly named ‘Davidson’s Objective’ insofar as it rehearses and takes up for its own a goal had by Davidson – namely, better understanding the problem of objectivity. I shall explain in the next section what exactly Davidson took this problem to consist of, and moreover, how it will be both construed and pursued subsequently in this investigation, but for now let us dwell a moment longer on the title’s initial meaning.

To call a thesis that focuses solely on the problem of objectivity, ‘Davidson’s Objective’, might seem to imply it is Davidson’s sole or primary aim. However, as anyone who has encountered Davidson’s work will acknowledge, this was at least prima facie not the case, nor do I wish to argue that it might be substantively otherwise. Although there are certainly many interrelated recurring themes and issues in Davidson’s decades-long career, it would, I feel, be a great disservice to the breadth and ingenuity of one of the Twentieth Century’s greatest philosophers to suppose he had only a single goal or aim in mind. As Davidson himself described the meandering trajectory that his work covered [Dav00:65]: ‘One thing led to another; I have never thought I had a program, nor have I intended to produce a system’. In this light, Davidson’s hope of addressing the problem of objectivity is to be understood as only one of many. It is worth making this point explicitly at the outset, as those familiar with the seemingly interrelated nature of Davidson’s work will undoubtedly feel (as I have done throughout) that recourse to Davidson’s many other ideas might be advisable to do justice to the arguments that pertain explicitly to the problem of objectivity. However, for reasons of word limitations and pragmatic restrictions to the volume of relevant literature digestible within a single thesis, little to no mention will be made of Davidson’s other more well known ideas, including: mental events, anomalous monism, truth-theoretic semantics, T-sentences, radical interpretation, the indeterminacy thesis, the omniscient interpreter, Robinson Crusoe, paratactic same-saying, akrasia, The Third Dogma of Empiricism, the Principles of Charity, conceptual schemes, decision theory, malapropisms, metaphors, etc., etc. While these and many more of Davidson’s ideas are certainly worthy of extended and careful consideration, herein it has been deemed necessary to consider only the arguments both directly and explicitly pertaining to Davidson’s problem of objectivity. In passing defence of such a constraining strategy, some reassurance may be found in Friedrich Nietzsche’s aphorism, ‘The Error of Philosophers’ [Nie1879/1996:261]:

13

The philosopher believes that the value of his philosophy lies in the whole, in the building: posterity discovers it in the bricks with which he built and which are then often used again for better building: in the fact, that is to say that the building can be destroyed and nonetheless possess value as material.

Such a reading accords well also with the broader intention for this thesis, for, while Davidson’s own treatment of the problem of objectivity forms the argumentative backbone, it is not intended to be of a primarily exegetical nature – this work is not meant as an addendum to the History of Philosophy. Rather, by undertaking a critique of Davidson’s philosophy and evaluating its many shortcomings, I hope to make my own contribution towards furthering Davidson’s original project. The aim of this thesis, then, is to take stock of the remarkable contribution Davidson made towards answering the problem of objectivity and, in doing so, to make further progress towards Davidson’s intended goal, rather than merely mapping the terrain that he managed to gain.

It is, further, prudent to acknowledge from the outset that surrounding themes and possible approaches to the problem of objectivity (or at least related variations of it) abound in Philosophy. Notable thinkers as distant and diverse (though, unforgivably, all male) as René Descartes, John Locke, Immanuel Kant, Friedrich Nietzsche, Georg Hegel, Edmund Husserl, Ludwig Wittgenstein, Michael Dummett, Wilfrid Sellars, Tyler Burge, Robert Brandom, Michael Haugeland, Thomas Nagel, Crispin Wright, Saul Kripke, Richard Rorty, John McDowell, Robert Stalnaker, Fred Dretske, P. F. Strawson, John Searle, Hans-Georg Gadamer, Günter Figal, Jerry Fodor and Michael Tomasello among many others, have relevant and insightful things to say regarding the question at hand, and the work of any one of these thinkers on these matters is certainly deserving of its own extended treatment. However, while all of the above have informed my thinking about the problem of objectivity, and while I do draw explicitly on the insights some of these thinkers (in particular, Burge, Tomasello and Brandom), once again, mainly for pragmatic reasons of space and research constraints, I have deemed it necessary to focus primarily on Davidson’s own contribution. It would be truly wonderful to have the freedom and competency to undertake a comprehensive review of the entire literature on the problem of objectivity and its many proposed approaches and solutions, but such a majestic project will have to wait for another occasion, if not another researcher.

14

That being said, my choice of Davidson as the primary focus is not at all arbitrary. For one, as will been seen in the next section, Davidson explicitly outlined a very specific version of the problem of objectivity and, as shall be explored throughout this thesis at large, he made various ingenious and insightful efforts towards addressing it. But, I think, the major reason Davidson deserves the limelight (at the very least, in the modest context of a thesis such as this), is that, while most of his original arguments are commonly recognised to have critically faltered if not outright failed, I nonetheless intend to demonstrate that recent thinking regarding the problem of objectivity is largely continuous and congruent with the approach of which Davidson was a very significant pioneer. Current progress being made towards resolving the still persistent problem of objectivity is largely part of Davidson’s remarkable and broad philosophical legacy, and so, his work warrants our serious and sustained consideration, such as I aspire to undertake herein.

II. THE PROBLEM OF OBJECTIVITY

The alternative reading of this work’s title is of Davidson’s notion or conception of what is taken to be ‘objective’ – or more accurately, of objectivity. In the ensuing section, I will consider some of the various senses commonly meant when speaking of objectivity, but in order to better understand how these conceptions stand in relation to Davidson’s own take on objectivity, we must first consider his proposed problem, for it will form the driving force of the investigation throughout.

1. Origin

According to Davidson, a central concern of Western Philosophy since Descartes has been what Davidson calls the problem of knowledge (see [Dav95/04:03]).1 As is well known, René Descartes, in his Meditations on First Philosophy, not only argued for the absolute certainty of

1It should be duly acknowledged that Christia Mercer [Mer16] has shown that Descartes’ thinking is heavily indebted to the work of Teresa of Ávila

15

the existence of presently occurring first-person experience – made famous by the dictum, ‘Cogito ergo sum’ – but did so via the method of hyperbolic doubt, whereby one treats as wholly false any belief derived from a ‘source’, such as reason or the senses, that conceivably admits of any degree of doubt whatsoever (see [Des1641/1996:17]). Perhaps unsurprisingly, a particularly insidious side-effect of Descartes’ sceptical method was the articulation of what came to be the full-blown challenge of global scepticism, wherein the existence of the mind-independent world is theoretically called into doubt. And it has been in response to this challenge, suggests Davidson, that Western philosophers have traditionally been concerned to establish as knowledge our common-sense belief in the mind-independent existence of the world. Davidson paraphrases the problem as a question: ‘How can we justify our belief in a world independent of our minds, a world containing other people with thoughts of their own, and endless things besides?’ [Dav95/04:3].

As Davidson sees it, the problem of knowledge arises as a result of mistakenly privileging first-person knowledge over the knowledge that common sense reveals we do possess regarding both the existence of other minds and of the mind-independent world. Once such a privilege has been assumed, thinks Davidson, it is unavoidable that we should be drawn to ‘attempt to derive knowledge of the external world from it’ [Dav91/01:206]. But, as Davidson laments [Dav91/01:206]: ‘the elaboration of such reductive proposals, and the demonstration of their failure, constitutes much of the history of philosophy from Descartes to the present’. It is therefore, in light of this ‘grudging tribute to the apparent impossibility of unifying the three varieties of knowledge [i.e. of self, other, and world]’ [Dav91/01:206] that Davidson advises we abandon the problem of knowledge in favour of a new line of inquiry.

Davidson proposes that rather than persisting in vain with the problem of knowledge, we should instead ask, ‘How did we come by the concept of objectivity in the first place?’ [Dav95/04:3] Or, as he alternatively phrases it: [Dav95/04:7]:

The problem is to account for our having the concept of objectivity – of a [concept of] truth that is independent of our will and our attitudes. Where can we have acquired such a concept? We cannot occupy a position outside of our own minds; there is no vantage point from which to compare our beliefs with what we take them to be about.

16

This, Davidson calls the problem of objectivity – though it would be more accurate to call it the problem of the concept of objectivity, for the aim is not, like the problem of knowledge, to demonstrate the existence of the external world, but rather, to explain how we came to possess such a notion in the first place.

Davidson points out that the problem of objectivity is conceptually prior to the problem of knowledge and so, is distinct from it insofar as the problem of knowledge couldn’t have arisen if it weren’t for our having the concept of objectivity in the first place. Davidson’s hope is that by investigating this new problem, not only will answers be more readily forthcoming than they have been to the obstinate problem of knowledge, but furthermore, perhaps the traditional problem will be seen to vanish with its resolution. As he puts it [Dav95/04:4]:

How have we come to be able to appreciate the fact that our beliefs may be false, that there is a basic difference between what we believe and what is the case? This is the topic of my lectures: What explains our grasp of the concept of objective truth? It may be that the epistemological question will be solved if we can answer the apparently simpler question how [objective] thought is possible.

For my own part, I hope to show via the arguments and findings developed herein that, while the problem of knowledge will not be answered to the satisfaction of the skeptic, nonetheless, having a better, more empirically grounded grasp of how it is that we possess the concept of objectivity at least takes the sting out of sceptical challenges, if not permitting one to tell the sceptic to ‘get lost’ (as Richard Rorty famously summarised the upshot of Davidson’s work (see [Ror80])).

2. Nature

When considering Davidson’s particular take on the problem of objectivity, it is important to emphasise that he has no interest in humans as such. Rather, Davidson is concerned to understand how any creature or entity could come to possess the concept of objectivity. That is to say, he is interested only in the general necessary conditions that make the concept

17

of objectivity possible. In the same breath, Davidson hopes to do away with any particular interest in an historical, evolutionary account of the emergence of the concept [Dav95/04:7]: ‘I am not concerned with the scientific explanation of the existence of [the concept of objectivity]; my interest is in what makes it possible.’ And further [Dav95/04:7]: ‘… why evolution has produced creatures that can entertain propositions … is a matter for the speculation or discovery of scientists.’

This rather anti-empirical, non-evolutionary approach ought, I think, to strike us initially as being somewhat odd, especially considering that only two pages prior, Davidson makes the following claim [Dav95/04:5]:

This attitude and method [of rejecting the Cartesian first-person perspective as primary] has sometimes been called naturalism because naturalism starts by accepting common sense (or science) and then goes on to ask for a description of the nature and origins of such knowledge.

Here we see the strong influence exerted on Davidson by Willard Van Orman Quine, who, if nothing else, is famous for arguing for a rather strong scientific naturalism, in which science determines ontology. As Quine himself puts it [Qui81:21]: ‘…it is within science itself, and not in some prior philosophy, that reality is to be identified and described.’ Although here Quine’s reference to ‘reality’ signals that his focus is primarily metaphysical (or anti-Metaphysics, as the case may be), nonetheless, to a contemporary audience where empirical research increasingly informs our best philosophical theorising, the sentiment at least may readily be generalised. Indeed, Davidson himself is to be acknowledged for his significant influence in occasioning this general shift of contemporary Western Philosophy to at least a more liberal naturalism (see [Sto06&07] and [Lep07]2).

In light of the above, an explanation seems called for as to why Davidson would not see fit to at all countenance an evolutionary or scientific account of the emergence of thought in possession of the concept of objectivity. The reason for this unwillingness lies in further consideration of what has already been said regarding Davidson’s rejection of the Cartesian 2 This infamous exchange surrounds the metaphilosophical implications of Davidson’s philosophy, although it is, I suspect, commonly felt that Lepore & Ludwig won the exchange in favour of the legitimacy of the more traditional philosophical approach.

18

legacy of treating the first-person perspective as primary. That is to say, although some commentators, such as Ernest Lepore and Kirk Ludwig (see [Lep05:399]), take Davidson to be endorsing the third person as primary, Davidson explicitly states that neither the first, second, nor third can claim precedence over any other. Nor may any one of these three epistemic standpoints be reduced to any other (see [Dav91/01:214]). As he colourfully illustrates it [Dav91/01:220]: ‘The three sorts of knowledge form a tripod: if any leg were lost, no part would stand.’

A consequence of the irreducibility of these three varieties of knowledge is that Davidson doesn’t hold out much hope for the discursive capacity to seamlessly describe the emergence of the intentional from the realm of the non-intentional [Dav97/01:127]:

In both the evolution of thought in the history of mankind, and the evolution of thought in an individual, there is a stage at which there is no thought followed by a subsequent stage at which there is thought. To describe the emergence of thought would be to describe the process which leads from the first to the second of these stages. What we lack is a satisfactory vocabulary for describing the intermediate steps.

And, as he continues [Dav97/01:128]:

It is not that we have a clear idea what sort of language we could use to describe half-formed minds; there may be a very deep conceptual difficulty or impossibility involved. That means there is a perhaps insuperable problem in giving a full description of the emergence of thought. I am thankful that I am not in the field of developmental psychology.

Davidson even goes so far as to claim that the seeming limitations (if not outright failures) of his arguments intended to show the non-intentional conditions sufficient for intentionality to arise, actually support his irreducibility thesis [Dav01:15]:

19

Our failure to provide an analysis of the concept of error or, as we could as well say, of the concept of objectivity, or thought itself, does not mean there is something hopelessly mysterious about these concepts; it only reflects the fact that intentional phenomena cannot be reduced to something simpler or different.

While it is not my aim herein to attempt to refute Davidson’s claim to the discursive incommensurability of intentional and non-intentional phenomena, a significant endeavour for this thesis will be to show that Davidson’s dismissal of scientific and evolutionary accounts of the emergence of distinctly human thought unduly undermines the force of his subsequently ‘academic’ arguments. I hope to show that by drawing on empirical research – much, admittedly, not yet available to Davidson at the time he was writing – at least some of the various insights Davidson brought to the table regarding the requisite conditions for the possession of the concept of objectivity are made all the more plausible. To share a sentiment expressed by Michael Tomasello when discussing these very matters [Tom14:151]:

[Evolutionary theory] does not solve Davidson’s (1982) problem of a common theoretical vocabulary that spans from “no thought” to “thought”, but it does narrow the distance to be traversed by any one step significantly.

Indeed, at least in some places, Davidson himself appears to be mildly receptive to such a possibility [Dav01:294]:

There cannot be said to be proof of [the] claim [that possession of the concept of objectivity is dependent upon linguistic communication causally embedded in an intersubjectively shared world]. Its plausibility depends on a conviction which can seem either empirical or a priori; a conviction this is a fact about what sort of creatures we are. Empirical if you think it just happens to be true of us that this is how we come to be able to speak and think about the world; a priori if you think, as I tend to, that this is part of what we mean when we talk of thinking and speaking.

20

We can see that Davidson’s dismissive characterisation of the empirical conviction – i.e. that it ‘just happens to be true’ [Dav01:294 (italics added)] – already clearly indicates his subsequently stated preference for a priori necessity claims. However, while a diminishing number of scholars are still drawn to this latter approach (see, for instance, Christian Barth’s Objectivity and the Language-Dependence of Thought, especially [Bar11:27-8]3), herein, the ultimate aim is to salvage at least some of Davidson’s many theoretical insights by marrying them with our best understanding of empirical reality. As was mentioned above, much of the work to be undertaken throughout this dissertation (especially in Parts Two and Three) will be to argue that such empirical discoveries and thus-wise informed speculation can be mustered to support at least some of Davidson’s key insights as to how best to resolve the problem of objectivity. To invoke the authority of Ruth Millikan [Mil12:889]:

For a philosopher to speculate about animal cognition is implicitly to engage in theoretical psychology or theoretical neurology at a very high level of abstraction. As with all sciences, research in psychology and neurology need to be guided by speculative hypotheses, in these particular cases, by hypotheses about what kinds of functions, hence structures, it would be sensible to look for. We philosophers may be in a position to help, but we can’t expect armchair argument to go very far. In the end, all the questions are rock-bottom empirical.

To be clear: I am not endorsing a strong version of scientific naturalism here. Our aim isn’t necessarily to bridge or close the gap between the empirical and the normative (perhaps by attempting to reduce the latter to the former). It is only to narrow the divide, to the benefit of both camps.

3. Challenges

It is important to stress from the outset the two main challenges posed by a commitment to this approach. These are: i) the threat of begging the question; and, ii) the threat of

3 In support of the claim that such self-avowedly a priori approaches are on the downturn, see Pieranna Garavaso’s review of Naomi Scheman’s Shifting Ground: Knowledge and Reality, Transgression and Trustworthiness [Gar11].

21

triviality. Both of these shall be treated in turn below; and, as I shall argue, addressing (ii) is our best hope of countering (i).

i) The threat of begging the question looms large when considering the role played by language in our possession of the concept of objectivity. This is principally because the Davidsonian line of investigation I shall be tracing out herein holds that the solution to the problem of objectivity lies in our being language-users. Indeed, as we shall see in Part One, Davidson’s early arguments are to the effect that only with language is it possible to possess the concept of objectivity (and so, as Davidson sees it, any thought whatsoever). As Davidson puts it [Dav01:293]: ‘… language and thought are mutually dependent’. And again [Dav91/01:209]: ‘My thesis is … that a creature cannot have a thought unless it has language.’ And one more for good measure [Dav82/01:100]: ‘The source of the concept of objective truth is interpersonal communication. Thought depends on communication. This follows at once if we suppose that language is essential to thought….’ In this light, the general argumentative strategy herein, then, must be to show that the concept of objectivity arises simultaneously (though perhaps not necessarily) with the advent of language. The difficulty with such a strategy is in explaining how semantic content can arise from non-semantic phenomena (i.e. behaviour) without unwittingly ‘contaminating’ the given description of the non-semantic phenomena with the semantic content sought after.

An analogy might be helpful here. When looking for traces of life on Mars, one of the many problems researchers face is to ensure that the Earth probes do not contaminate the Martian scene to be examined. The difficulty here, let us presume, is twofold. Firstly, life on Earth is so prevalent as to make it very difficult to completely sterilise the equipment to be sent to search for life. But, secondly, if the Martian site does happen to become unwittingly contaminated, it could be extremely difficult if not impossible to determine after the fact whether the life detected either did or did not originate from Mars. We could imagine that this might be because, as no life on Mars has yet been encountered, it may be difficult (if not impossible) to conclusively distinguish it from life that had originated on Earth.

To draw the analogy back to the goal of providing a non-question-begging account of the role played by language in our possession of the concept of objectivity: given semantic content so profusely permeates our engagement with our environment, when it comes to describing the emergence of semantic content out of non-semantic phenomena, the chances of ‘contamination’ (i.e. begging the question) are extremely high. This danger – like the

22

difficulties of detecting Martian life – is the result of two factors. Firstly, the ‘tools’ with which we are equipped to provide such an account (i.e. language) are themselves inherently semantic, so that it is arguably impossible to ever achieve complete ‘sterilisation’. That said, unlike the Martian context, full sterilisation is not required; it is sufficient that the account developed be non-circular in illuminating some non-trivial aspect of the role played by language in our possession of the concept of objectivity, as I shall argue momentarily. Secondly, as we do not yet have an adequate account of the emergence of semantic content, one of the central challenges faced in providing one is to ensure that it picks out the point that language emerges to a sufficient degree of precision and with sufficient explanatory power. Like the case of Mars, the difficulty here arises when attempting to determine whether the emergence of semantic content can be attributed to some particular non-semantic phenomena without begging the question. Avoiding triviality, as I shall now explain, appears to be the way of overcoming this difficulty also.

ii) The threat of triviality arises specifically because, as was highlighted by the story of the surrogate goldfish given in the Preface, language-based communication can so readily convey the concept of objectivity. Davidson himself is also clearly cognisant of the need to provide a non-trivial account of the role played by language in this regard. He makes this explicit when he asks [Dav82/01:105]: ‘What would show command of the contrast [between subjective belief and objective truth]? Clearly linguistic communication suffices’. The threat of triviality looms, however, for language or linguistic communication ‘so clearly’ – i.e. trivially – suffices insofar as language is an uncontested means by which one may demonstrate their possession of the concept of objectivity. When it comes to establishing the role that language plays in possessing the concept (and not merely demonstrating one’s possession thereof) such instances are trivial.

Again, an analogy might be helpful. Consider the claim that in order to possess the concept of, say, Sherlock Holmes, one must be a language user. While such a claim is likely true, it is only trivially so. This is because there is nothing about the literary character of Sherlock Holmes in particular that makes possession of the concept of him any more dependent upon language than that of any other literary character one might care to mention (John Watson, for instance). That is to say, insofar as it is true that the general concept of purely literary characters depends on language, it is trivially true that any one particular character does so also.

23

In the same vein then, not only can we readily demonstrate that language suffices for the concept of objectivity, but moreover, even if it turned out to be the case that we could only imagine linguistic communication as doing so, there is still call for a non-trivial story to be given as to what it is about the nature of language that enables this particular concept to arise. As Davidson puts it [Dav82/01:102 (italics added)]: ‘What we need, then, in order to make a case [that only language users possess the concept of objectivity], is a characterisation of what it is that language supplies that is necessary for thought.’

Moreover, in light of the above discussion concerning the threat of begging the question, a non-trivial account will sufficiently address both of the challenges outlined. That is to say, on the one hand, the need to provide a non-circular account that illuminates the role played by language in our possession of the concept of objectivity (i.e. the ‘contamination’ issue) will be assuaged by a non-trivial account insofar as the story given must be sufficiently ‘interesting’ (i.e. not trivial) as to avoid the charge of circularity. On the other hand, the difficulty faced in pin-pointing the moment where language emerges to an explanatorily sufficient, yet non-question begging degree of precision (i.e. the ‘differentiation’ issue) will similarly be addressed by providing a non-trivial account, for the explanatory power of such an account ought to vindicate it against any fatal charges of begging the question.

This then is the primary aim of this thesis: to establish a non-trivial account of the role played by language in our possession of the concept of objectivity.

Before moving on to consider Davidson’s arguments proper, it is prudent to briefly consider what precisely is meant when we speak of ‘the concept of objectivity’.

III. THE CONCEPT OF OBJECTIVITY

The topic of objectivity has been a relatively popular one in Western Philosophy for at least the last couple of decades. Along with Davidson’s own 2001 collection, Subjective, Intersubjective, Objective, numerous other major works and collections have also been

24

published with objectivity as their explicit focus (see, for instance: Rou15; Tom14; Bar11; Bur10; Fig06/10; Das07; Noz01; Bra00&94; Wri92; Ror91; Pop72). Perhaps unsurprisingly then, it is somewhat problematic to speak of the concept of objectivity without qualifying precisely what one has in mind, for surely as many meanings abound as there are theories that purport to articulate exactly what’s at stake. In this section, two central senses of the concept of objectivity will be briefly characterised before turning, on the one hand, to locate where Davidson’s original interests lie, and, on the other, to where this investigation shall direct its focus.

1. Attitude-Transcendence

One particular sense of objectivity predominates in common parlance. When people speak of objectivity, they often do so in an epistemic key, such as in the context of verdicts being passed in Criminal Law, or empirical research being undertaken in scientific investigation. Robert Brandom provides a phrasing befitting this sense of objectivity when he speaks of attitude-transcendence [Bra00:189] – a notion clearly characterised by John Searle in the following passage [Ser95:8]:

Epistemically speaking, ‘objective’ and ‘subjective’ are primarily predicates of judgements. We often speak of judgements as being ‘subjective’ when we mean that their truth or falsity cannot be settled ‘objectively’, because the truth or falsity is not a simple matter of fact but depends on certain attitudes, feelings, and points of view of the makers and the hearers of the judgement. An example of such a judgement might be, ‘Rembrandt is a better artist than Rubens’. In this sense of ‘subjective’, we contrast subjective judgements with objective judgements, such as the judgement, ‘Rembrandt lived in Amsterdam during the year 1632’. For such objective judgements, the facts in the world that make them true or false are independent of anybody’s attitudes or feelings about them.

According to Searle, the inherent nature of certain judgements are subjective because the matters in question are inherently subjective – for instance, questions of aesthetic taste. Of course, a trained art critic will likely bring a more epistemologically objective ‘eye’ to a work of art in forming her judgement; but the judgement itself is best characterised as

25

subjective insofar as it depends importantly on the attitudes, feelings and points of view of the makers and hearers of the judgment. The truth or falsity of objective judgements, on the other hand, are (sufficiently, if not wholly) independent of such subjective factors. For instance, the height of Everest is independent of anyone’s beliefs about the matter. In fact there’s a riddle that illustrates this point perfectly: What was the Earth’s highest mountain before Mount Everest was discovered? Answer: Mt Everest! (Even though Kangchenjunga had previously been mistakenly considered the highest.)

The social and philosophical significance of objectivity in this epistemic sense cannot be overestimated, as is commendably articulated in the following epideictic passage by Tyler Burge [Bur10:550-1]:

... the growth of objectivity […] is one of the most impressive achievements of humankind. In the practical domain, there is the subordination of self-interest to morality. There is the rule of law and the egalitarianism of democracy. There is the rejection of differences of race, color, religion, gender, sexual orientation as irrelevant to principles governing rights and opportunities. There is the building of institutions, art, and communicative systems that transcend and outlive the individuals and individual perspectives that make them go. In the theoretical domain there are the abstractions and generalities of mathematics, the systematizations of natural science, and the retrospectives of history. There are even the reflections of philosophy, on the scattered occasions when they make progress. Mathematics, natural science, history, law, morality, philosophy – theoretical and practical – have repeatedly shown perspectives that had seemed final and absolute to be infected with egocentricity or provincialism.

Although Searle adds in conclusion of his own summary that ‘it should be obvious […] that the contrast between epistemic objectivity and epistemic subjectivity is a matter of degree’ [Sea95:08], Burge’s examples emphasise the benefits of judging matters more objectively than subjectively, while at the same time remaining silent on whether or not these matters are inherently what Searle would characterise as objective matters.

26

2. Mind-independence

This takes us nicely into introducing the other intimately related concept of objectivity, which might be best characterised as mind-independence. Searle, again, provides a useful characterisation of the notion of mind-independence in its ontological guise, when he writes [Sea95:8]:

In the ontological sense, ‘objective’ and ‘subjective’ are predicates of entities and types of entities, and they ascribe modes of existence. In their ontological sense, pains are subjective entities, because their mode of existence depends on being felt by subjects. But mountains, for example, in contrast to pains, are ontologically objective because their mode of existence is independent of any perceiver or any mental state.

Presumably, from this, he means to emphasise that the notion of ontological objectivity is the notion of something existing independently of ourselves and our awareness of it. It is not, however, clear from what Searle says that it is not merely tracking the difference between what is straightforwardly mental and what is not. That is to say, it is not obvious what more might be meant by saying that pains are ‘felt by subjects’ than by saying ‘pains are mental states’. This conflation arises, I think, because Searle’s characterisation doesn’t make it sufficiently clear that subjective entities also exist ‘objectively’ in the sense, for example, that your pain exists independently of my subjectively knowing about it. Robert Nozick, on the other hand, explains what is at stake in his following characterisation [Noz01:75]:

The notions of ‘objective’ and ‘subjective’ are contrasting notions, at least insofar as objects and subjects themselves contrast. Something is objective when (or to the extent that) it is determined in its character by features of an object; it is subjective when it is determined in its character by states such as consciousness, emotions, and desires that are intrinsic to being a subject.

This characterisation comes closer to capturing what the concept of objectivity in the sense of mind-independence amounts to. That is to say, something is subjective by virtue of some

27

dependency relation either to some particular subject or subjects. On the other hand, it will be more objective to the extent that its character is free from such subjective dependencies. A mountain is the way it is independently of how anyone thinks of it. The characteristics of a belief, on the other hand, depend on it being held by some subject as true. This, however, is not to deny a materialistic commitment to mind. Although the characteristic of a subject’s belief may also depend objectively on the physical state of their brain, the belief itself nonetheless remains a subjective entity insofar as it equally needs to be held as true by the subject who believes it to be true.

While I do not wish to deny that Searle and Nozick’s characterisations are insightfully illuminating in their own right, neither, I feel, sufficiently emphasises what we might loosely call the ‘transcendental’ aspect lying at the heart of the concept of objectivity. Such a notion was hinted at when Davidson wrote [Dav95/04:7 (italics added)]: ‘Where can we have acquired [the concept of objectivity]? We cannot occupy a position outside of our own minds…’ While the concepts of objectivity and subjectivity no doubt closely parallel and bear upon the concepts of subject and object (along with those of the mental and non-mental), nevertheless, essential to the concept of objectivity is the notion of a reality (epistemic, ontological, moral, etc.) that is intrinsically independent of ourselves and our thinking about it. The concept of objectivity is then at least in part a concept of something ‘transcendentally’ non-conceptual. In the epistemic key, it is a concept of a truth beyond opinions, attitudes and prejudices. In the ontological vein, it is the concept of a reality existing beyond our awareness of it (i.e. the opposite of Berkeley’s dictum, esse est percipi (see [Ber1710/1988:54]). But generally speaking, the concept of objectivity may be characterised as essentially the notion of that which is in one way or another transcendent of ourselves as subjects.

3. Conceptual Objectivity

In light of the above, we can better see that in formulating and attempting to address the problem of objectivity, Davidson was not so much concerned with accounting for our possession of the concept of objectivity in general, but, rather, with the supposed objectivity of our epistemic claims. To again quote his central articulation of the problem [Dav95/04:7 (italics added)]: ‘The problem is to account for our having the concept of objectivity – of a

28

[concept of] truth that is independent of our will and our attitudes’. Joseph Rouse clearly adopts such a line in arguing that Davidson (along with the likes of John McDowell, Kit Fine, Donna Haraway, Michael Haugeland and Robert Brandom) occasioned a shift away from the traditional epistemic notion of objective knowledge, to what Rouse calls, ‘conceptual objectivity’. As he explains it [Rou15:182]:

Davidson and these ‘left-Sellarsians’ relocate the question of objectivity. Instead of asking how knowledge could be objective, they ask how knowledge claims could even purport to be objective—that is, they ask what it is for our performances to be about objects and accountable to them at all. The issue then concerns objective conceptual content rather than objective knowledge. Epistemic conceptions of objectivity only come into play once some claim to knowledge has been formulated and recognized as a claim. The epistemic question is then whether that claim is true, or objectively justified, and objectivity is invoked to settle this question, whether at the ground level or through epistemic ascent. Davidsonians and Sellarsians think that understanding how our performances are meaningfully accountable to the world renders epistemic objectivity superfluous. With Fine and Haraway, they think epistemic questions can only be settled within ongoing inquiry, which answers holistically to norms of conceptual objectivity.

Much of this should sound familiar from the above introduction to Davidson’s problem of objectivity – specifically that Davidson eschews the traditional problem of knowledge or what Rouse here dubs ‘epistemic objectivity’. Instead, as Rouse explains, Davidson aims to account for how we may possess objective conceptual content; how our beliefs and utterances can be about and responsive to the objective world. Davidson’s answer, as Rouse hints at, is via some form of conceptual holism. To quote Davidson in support of this aspect of Rouse’s reading [Dav74/01:198]:

Truth of sentences remains relative to a language, but that is as objective as can be. In giving up the dualism of scheme and world, we do not give up the world, but re-establish unmediated touch with the familiar objects whose antics make our sentences and opinions true or false.

29

Davidson’s goal, it seems to me, is to establish that our words mesh with and are responsive to the world. His goal is not – as with the problem of knowledge – to establish that we have objective knowledge as such, but rather to provide an account of how our practices are grounded in reality. And again, with respect both to Davidson’s abandonment of the strong epistemic sense of objective knowledge and to his global lingualism (i.e. that thought requires language) [Dav82/01:105]:

The concept of intersubjective truth sufficies as a basis for belief and hence for thoughts generally … [and] having the concept of intersubjective truth depends on communication in the full linguistic sense.

But while Rouse is doubtlessly correct in at least his broad characterisation of Davidson as a ‘left-Sellarsian’, herein, by contrast, the aim will not be concerned with Davidson’s efforts toward establishing that our beliefs and knowledge claims do have objective content as such (i.e. that they actually and accountably respond to the ontologically objective world). Rather, our primary concern is to consider the extent to which Davidson provides a plausible and non-trivial account of the role played by language in our possession of the general concept of objectivity as a concept – not that the concept succeeds in representing objective reality. This is an important qualification to make at the outset, as a great deal of the commentary surrounding Davidson’s work focuses on either criticising or defending these further aspirations. While such matters are indeed legitimate and interesting in their own right (although, see the exchange between [Sto06&07] and [Lep07] for contrary views on this matter), the limited scope of this thesis is concerned only with the question of the non-trivial role (if any) played by language in our coming to possess the concept of objectivity in general, even though evidence of conceptual awareness regarding the more specific varieties would also count as satisfactory evidence for the more general concept as well.

30

PART ONE

* * *

Thought & Talk

Part One of this thesis focuses on Davidson’s ‘Main’ Argument. It has also been variously referred to as his ‘Main’ Argument Against Animal Thought [Lur15:§1.c.iii] or The ‘Main’ Argument for the Necessity of Language for Thought [Lep05:392]. Davidson’s ‘Main’ Argument is originally advanced to establish global lingualism (or what Christian Barth dubs universal conceptual lingualism [Bar11:1]), which holds that language is necessary for any and all thoughts whatsoever.

First articulated in his 1982 article, Rational Animals, the ‘Main’ Argument runs as follows [Dav82:102]:

31

The ‘argument’ has two steps […] First, I argue that in order to have a belief, it is necessary to have the concept of belief. Second, I argue that in order to have the concept of belief one must have language.

Each premise of the ‘Main’ Argument is given its own respective supporting argument: i.e. The Argument from Surprise and The Argument from Error. These two arguments will be considered in Chapters 1 and 2 respectively.

The Argument from Surprise is intended to establish that to have beliefs is to have the concept of belief. The force of this argument relies on the common presumption that the possibility of surprise as a basic condition of having beliefs arises as a result of meta-cognitively recognising one’s beliefs to have been false. This presumption, however, will be shown to be unfounded via an account of surprise not involving beliefs about beliefs.

The Argument from Error is intended to establish that the concept of error (and so of both belief and objectivity) is a condition of having language, such that we might have some reason to suppose that the reciprocal is also necessarily the case. But while Davidson points to the commonly recognised sufficiency of language to support the concept of error, he fails to provide any compelling argument for the claim that language is a necessary condition. The main objection considered herein against this supposition is that there are contexts where the concept of error would seemingly be readily applicable despite not involving language such that it is prima facie plausible that language is not necessary for the concept of error.

Both of Davidson’s arguments are exclusionary in nature. That is to say, they don’t so much aim to account for our possession of the concept of objectivity as to establish that no thought is possible without it. Objections against this thesis of global lingualism aim to show that languageless creatures can be (and are) credited as capable of thought even in the absence of the concept of objectivity. However, as both of Davidson’s arguments are generally recognised to be fatally flawed, the significance of these arguments for the purposes of this thesis is more about providing motivation for the ensuing discussion of Parts Two and Three. The overall aim of Part One is to suggest that, even though all agree that Davidson’s exclusionary aims fall short, nonetheless, the objections levelled against him do not establish the contrary thesis that the concept of objectivity is possible without language. That is, there

32

is good reason for supposing that necessity arguments – whichever direction they may incline – invariably hold little water, such that we will thereby be recommended to undertake the constructive analysis of Parts Two and Three. Specifically, I will argue that the fatal objection levelled against Davidson’s Argument from Surprise – namely, that, by Morgan’s Canon, one should not ascribe cognitive capacities above what is minimally required to account for a particular behavioural phenomenon – applies equally to the objections raised against Davidson’s Argument from Error. This means that the same general reasons for why we should not presume that thought is only possible in the context of language, apply equally to similar a priori suggestions that thought is possible outside of the context of language. This result then, is intended to motivate the discussion in Parts Two and Three, which aims to provide a substantive, non-trivial account of the role played by language in our coming to possess the concept of objectivity.

To say that the objections against The Argument from Surprise apply equally to the objections Against the Argument from Error is not, however, to imply that thought (or the concept of objectivity) is not possible outside the context of language. Rather, I only wish to suggest that in light of the mutual failings of these various a priori arguments both for and against the thesis of global lingualism, a more substantive approach is recommended, and it is precisely this tack that will be taken in Parts Two and Three of this thesis.

33

CHAPTER ONE

* * *

Surprise

I. INTRODUCTION

This chapter is concerned with rehearsing and evaluating what is now commonly referred to as The Argument from Surprise. This argument is intended to establish that one can have beliefs only if one has the concept of belief, and so, as was seen in the introduction to Part One, is intended to support the first premise of Davidson’s ‘Main’ Argument.

Davidson’s Argument from Surprise is intended to establish that one can have beliefs only if one has the concept of belief. Or as Davidson originally puts it: ‘… in order to have a belief, it is necessary to have the concept of belief’. [Dav82/01:102] Taken by itself, this seems like a rather counter-intuitive claim. It is not at all clear why one would need to know that one has a belief qua belief in order to have any beliefs at all. For instance, it seems

34

commonsensical that my dog can believe that I’m about to take her for a walk without her needing to conceptualise that her belief is a belief. Davidson, however, demurs. He argues that one has a belief only if one ‘commands the contrast between what is believed and what is the case’, [Dav82/01:105] and further, that it is the possession of the concept of belief that enables one to make such a distinction. Davidson’s appeal to the phenomenon of surprise is intended to make this claim seem intuitive and, so, more compelling.

In its original explication, the argument is articulated as follows [Dav82/01:104]:

[…] I think that surprise requires the concept of a belief. […] Suppose I believe there is a coin in my pocket. I empty my pocket and find no coin. I am surprised. Clearly enough I could not be surprised (though I could be startled) if I did not have beliefs in the first place. And perhaps it is equally clear that having a belief, at least one of the sort I have taken for my example, entails the possibility of surprise. If I believe I have a coin in my pocket, something might happen that would change my mind. But surprise involves a further step. It is not enough that I first believe there is a coin in my pocket, and after emptying my pocket I no longer have this belief. Surprise requires that I be aware of a contrast between what I did believe and what I come to believe. Such awareness, however, is a belief about a belief: if I am surprised, then among other things I come to believe that my original belief was false. I do not need to insist that every case of surprise involves a belief that a prior belief was false (though I am inclined to think so). What I do want to claim is that one cannot have a general stock of beliefs of the sort necessary for having any beliefs at all without being subject to surprises that involve beliefs about the correctness of one’s own beliefs. Surprise about some things is a necessary and sufficient condition of thought in general.

In light of the above, we can more formally construe the argument as follows:

TheArgumentfromSurprise

1.Ifonehasbeliefs,onemustbesubjecttosurprises;

2.Ifoneissurprised,itisbecauseone(suddenly)realisesone’spriorbeliefstohave

beenfalse;

35

3.Ifonebelievesthatone’sbeliefshavebeenfalsethenonepossessestheconceptof

belief;

4.Therefore,tohavebeliefsentailsthatonehastheconceptofbelief.

The first premise is supposed to hold given that ‘a belief is something that “can be true or false”’ [Glo03:54]. As Davidson writes of this commitment [Dav95/04:6]: ‘[…] what defines [beliefs] is the possibility of truth or falsity…’. Thus, given that beliefs are inherently either true or false, Davidson presumes that a false belief at least has the potential to result in surprise. In this, Burge concurs [Bur10:267]:

This claim seems fairly plausible. Ordinary beliefs, as well as perceptions, are associated with anticipations that may not be realized. For example, if a person has a belief that that is a solid body, and reaches out to touch it, and there is only a feel of thin air, there will be surprise. If solidity is not felt, the individual is surprised.

It would verge on begging the question if Burge were here to claim that if one has beliefs then one must be subject to surprises insofar as a false belief will result in a surprised reaction when it is realised as false. This is because such an explanation teeters on asserting that surprises require beliefs about beliefs – the intended upshot of the second premise. In the above quotation, Burge is careful not to fall into this trap, and nor should we. Let us then be equally careful to point out that at this stage of the argument it is purposefully unstated as to why or how it is that false beliefs are supposed to give rise to surprises. Explaining how and why they do is central to the point of The Argument from Surprise, and is the primary focus of its second premise. Presumably we are to take it on our pre-theoretical understanding of personal past experiences that surprise is an unavoidable yet entirely typical result of our recognising false beliefs as having being mistaken. Indeed, given that the intended dialectical force of The Argument from Surprise is to make its apparently counterintuitive conclusion seem more intuitive, it is perhaps understandable that here Davidson offers no further explicit explanation linking the realisation of false beliefs to the physiological reaction of surprise. Rather, the argument is supposed to appeal to our intuitive understanding of how and why we come to be surprised. As Carruthers explains [Car08:5-6]:

36

We correctly interpret the feeling of surprise for what it is: a state that is caused when a perception gives rise to conflict with prior belief or expectation. And when we report on the feeling, it is in just such terms that we couch our description. We say, for example, ‘I felt surprised because I was expecting P but saw Q instead’.

Exactly how the perception, Q, conflicts with the belief or expectation, P, such that it will result in surprise is beside the point at this stage in the argument. All that we are meant to concede at this point is that it is the nature of any and all beliefs that conflicting perceptions, beliefs, etc., may result in surprises. As we shall see in Section II, however, Lepore & Ludwig argue that this supposed intuitive acceptance fails to hold up in the light of day, and indeed, even conflicts with other oft-held intuitions. Although Lepore & Ludwig’s points do not refute (1), they will be shown to work to significantly undermine the supposed intuitiveness of Davidson’s argument such that its dialectical force is much weakened.

The second premise arises from Davidson’s peculiar characterisation of surprise. According to Davidson, if one can be surprised then one must be capable of being surprised by the meta-cognitive realisation that one’s past belief was false. This characterisation of surprise is taken to stand in contrast to, both: i) the phenomenon of being startled, and ii) changing one’s beliefs without thereby being surprised.

i) According to Davidson, one can be startled without recourse to belief – for instance, by a loud noise or a bright flash of light. To be startled, one need not have entertained any beliefs or expectations. To be startled is merely to trigger an unthinking reflex. For instance, healthy newborn babies who presumably have no beliefs whatsoever will nonetheless be startled by sudden loud noises or the sudden sensation of falling. The capacity to be startled by either of these two particular stimuli is known as the Moro reflex, and neonates who are not in this way startled are thereby identified as having a profound disorder of the motor system, not false beliefs.

ii) On the other hand, one might change one’s mind from a false to a true belief without necessarily having any beliefs about the new or old beliefs; for instance, this may occur over a long period of time perhaps with regard to a dispositional belief. But as Davidson sees it, it is by recognising that one’s past belief was false that one will be surprised. That is to say, in

37

order to trigger a surprise, one must have a belief about belief; namely, that the latter belief was false. Both Glock and Carruthers initially concede that prima facie this seems right. Carruthers writes [Car08:4] :

[Davidson’s characterisation of surprise] seemed plausible enough, since, for sure, when we feel surprise we do characteristically believe that one of our prior expectations has been overturned.

Thus, according to Carruthers, it appears that we are surprised when we come to believe that another of our beliefs (here, ‘our prior expectations’) had been false. Glock similarly writes [Glo03:55 ]: ‘In one respect, Davidson’s focus on surprise is promising. The vast majority of our beliefs are implicit expectations of which we become aware only when they are disappointed.’In either case, the key to Davidson’s characterisation of surprise is that it involves beliefs about belief. Davidson appeals to our intuitions regarding past experiences in claiming that, without the relevant beliefs about belief, either we wouldn’t have been surprised or our seemingly surprised physiological reaction would only count as being startled. As Davidson himself puts it [Dav82:104]: ‘Surprise requires that I be aware of a contrast of what I did believe and what I come to believe. Such awareness, however, is a belief about belief…’.

In Section III, we will consider (and subsequently develop) Carruthers’ account of surprise requiring no meta-cognitive beliefs about belief. Given the success of this alternative first-order account, Davidson’s Argument from Surprise will thereby be seen to critically fail.

The final premise of The Argument from Surprise is unstated in the above quoted passage where Davidson outlines his argument. He only writes about coming to believe that one’s beliefs were false. Clearly, however, Davidson must assume that a belief that one’s belief was false requires having the concept of belief if he hopes to establish the intended conclusion of The Argument from Surprise. In order to appreciate how it is that Davidson thinks this inference follows, it is important to note that Davidson explicitly reserves the term ‘concept’ for a highly intellectualised and atypical meaning. As he explains [Dav95/04:8]:

38

There are many people, including philosophers … who identify the ability to discriminate items having a certain property with having a concept – with having the concept of being such an item. But I shall not use the word ‘concept’ in this way.

Davidson then goes on to explain what he means by ‘concept’ [Dav95/04:8]:

I should … like to reserve the word ‘concept’ for cases where it makes clear sense to speak of a mistake, a mistake not only as seen from an intelligent observer’s point of view, but as seen from the creature’s point of view.

And further [Dav95/04:9]:

To apply a concept is to make a judgment, to classify or characterize an object or event or situation in a certain way, and this requires application of the concept of truth, since it is always possible to classify or characterize something wrongly.

Thus, according to Davidson’s stipulation, to have a concept ‘x’ requires that one classify or characterise it in such a way that the concept of truth applies to it. For example, if one has the concepts of ‘Harry’ and ‘Happiness’ such that one can believe that ‘Harry is happy’, then, according to Davidson, to hold this belief regarding Harry is to hold that it is true that Harry is happy insofar as one thereby implicitly recognises that Harry might have otherwise not been happy (e.g. if he were sad) such that one could have at least possibly been mistaken regarding Harry’s wellbeing. From this it follows that if one has the belief that one’s past belief was false (and one has not just merely changed one’s mind), then one must have the concept of belief such that one can be capable of applying the concept of falsity to it. As Davidson explains immediately following his original articulation of The Argument from Surprise [Dav82/01:104]:

Much of the point of the concept of belief is that it is the concept of a state of an organism which can be true or false, correct or incorrect. To have the concept of belief is therefore to have the concept of objective truth. If I believe there is a coin

39

in my pocket, I may be right or wrong; I’m right only if there is a coin in my pocket. If I am surprised to find there is no coin in my pocket, I come to believe that my former belief did not correspond with the state of my finances. I have the idea of an objective reality which is independent of my belief.

For Davidson, to have a belief about the truth or falsity of another belief such that one can be surprised if it turns out to be false, is to have the concept of belief. In this, Lepore & Ludwig concur [Lep05:395]: ‘… it can be granted that to be surprised about the correctness of one’s own beliefs is in part to think that one’s beliefs were false, and so requires that one have the concept of belief.’ Glock disagrees, however. Glock argues that the work Davidson wants the concept of belief to do requires that it meet conditions that cannot be satisfied by the supposed connection between the second to the third premise. As we shall see in Section IV, recent work by Elisabeth Camp has shown there are good reasons to allow that the requirements for possession of a concept be slackened. Nonetheless, it will be concluded that, granting Caruthers’ account of first-order surprise, further supporting arguments would be needed in order to warrant the inference to the third premise.

With respect to the overall argument, Davidson admits that his ‘reasoning can be challenged at several points’ [Dav82/01:102]. But as has been indicated in the above summary, matters are much worse, as we shall see in detail in the following.

II. BELIEF

The first premise of Davidson’s Argument from Surprise holds that if one has beliefs, one will be subject to potential surprises. As he puts it [Dav82/01:104]: ‘What I … want to claim is that one cannot have a general stock of beliefs of the sort necessary for having any beliefs at all without being subject to surprises...’. Davidson, however, offers no clear argument to explain this supposed basic link between beliefs and surprise. It seems obvious that, given the second premise of the argument aims to establish that surprise involves beliefs about belief, we cannot appeal to this line of reasoning in order to support the first

40

premise without begging the question. Glock, however, reads Davidson as doing exactly this, as we shall see presently.

1. The Argument Without Surprise

Glock points to the fact that we can intelligibly recognise that a prior belief was wrong without necessarily being surprised by it. He claims, for instance, that if I am ‘mindful of the fact that I am a notorious scatterbrain’ [Glo03:287], I may simply shrug my shoulders when I come to realise that a past belief was false. But, according to Glock, such a scenario satisfies the sufficient conditions that Davidson outlines for surprise. In this light, Glock claims [Glo03:287]:

[Davidson] seems to conceive of being surprised independently of any specific behavioural manifestations, and as involving beliefs about beliefs by definition. Given this stipulation, the soundness of [the first step of the argument from surprise] hinges straightforwardly on the question of whether a creature can only have beliefs if it is capable of also believing that a prior belief was false.

Glock thereby dismisses the original role that surprise was supposed to play in The Argument from Surprise as being beside the point. According to Glock, the force of the argument depends only on whether one needs to have beliefs about beliefs in order to have beliefs. Recourse to surprise is then, at best, rhetorical, or at worst, question-begging. The first and second premises should therefore be combined, removing any recourse to surprise whatsoever. Such a premise would presumably read as follows:

1´.Ifonehasbeliefs,thenonehasbeliefsaboutbeliefs.

However, as was canvassed in Section I, the dialectical force of the Argument from Surprise relies on it appealing to our intuitive understanding so as to make what is a rather counter-intuitive conclusion seem more intuitive. Davidson’s original argument does indeed ultimately lead to the inference that one can have beliefs only if one has beliefs about beliefs

41

(and so the concept of belief); the crucial step that is supposed to bring us to this conclusion is that our commonsensical understanding of surprise supports a peculiar and yet apparently universal understanding that false beliefs (and so potentially any beliefs whatsoever) give rise to surprises by way of our coming to believe that our previous beliefs had been false. Without this appeal, The Argument from Surprise falters. To consider generally (as Glock would have us do) whether or not any beliefs whatsoever necessitate beliefs about beliefs certainly seems like an unpromising avenue to go down, and Glock himself argues that to do so is futile. Rather, for Davidson, the significance of focusing on surprise is that it appeals to an intuition that if one has beliefs, then invariably some of them will be false beliefs, and false beliefs render us susceptible to surprises. Davidson puts it in such a way as to suggest that he, at least, does not regard surprise as synonymous with a belief that one’s belief was false [Dav82/01:104]: ‘I do not need to insist that that every case of surprise involves a belief that a prior belief was false…‘. Although in no way does this demonstrate the justification for the first premise, it should give us some impetus to resist Glock’s advised move away from surprise as it is certainly not dialectically equivalent to the original argument.

2. Language Without Surprise

Against the supposed justification for the first premise, then – whether based on intuition or empirical evidence – Lepore & Ludwig have much to say, and provide two cases in point. We shall examine each in turn.

Lepore & Ludwig ask of the first premise [Lep05:395]:

… why do we have any more reason to think that one can have a general stock of beliefs only if one can be surprised than we do to think that one can have any thoughts or beliefs at all only if one has the concept of belief?

And further, they state [Lep05:395]:

42

… it is not obvious that attributing beliefs to someone requires thinking of him as having the concept of belief. It is hard to see how the support offered by [The Argument from Surprise] advances the case, for it is no more obvious that attributing beliefs to someone requires thinking of him as being capable of being surprised.

That is, they take issue with the dialectical force of Davidson’s first premise in relation to the argument as a whole. In order to establish the conclusion that one must have the concept of belief in order to have any beliefs, the supporting premises must be more clearly recognised and understood; the explanans needs to be clearer than the explanandum. Lepore & Ludwig suggest that this is not the case with the first premise in relation to the argument as a whole. Indeed, Davidson himself seems to acknowledge this point of weakness in only going so far as to suggest, ‘perhaps it is equally clear that having a belief … entails the possibility of surprise.’ [Dav82/01:104 (italics added)] Lepore & Ludwig claim that this is not far enough, however.

Against the obviousness of Davidson’s claim that in order to have a general stock of beliefs one must be subject to surprises, Lepore & Ludwig suggest that individuals sometimes seemingly have a general stock of beliefs such that they can make mistakes without thereby recognising their mistakes, let alone being surprised by them.

Firstly, they point to reports made by child psychologists (i.e. [Gop98] and [Per87]) to the effect that ‘children (who have already acquired some language) pass through a developmental stage in which they are unable to recognise that they have had false beliefs.’ [Lep05:395] If one were to accept Davidson’s Argument from Surprise, it must be concluded that these children ‘do not have beliefs, if that requires being capable of being surprised.’ [Lep05:395] Such a conclusion seems untenable, however, for it seemingly posits two antithetical claims:

i) they have some language mastery, so they have propositional attitudes; and,ii) they are incapable of being surprised, so they do not have any propositional

attitudes.

43

That Davidson is committed to the first claim is highly plausible, for Davidson is well known as further holding that one can only have beliefs (and so be susceptible to surprises) only if has a language. Indeed, as was seen in the introduction to Part One, The Argument from Surprise is intended to support the first step of Davidson’s ‘Main’ Argument, the overall upshot of which is that language mastery is necessary for propositional attitudes, such as belief. This then is presumably what Lepore & Ludwig are dialectically appealing to when they claim that to have some language mastery is to have some beliefs.

Before evaluating this objection in its own right, it is interesting to note what Davidson might say in response, for presumably he would seek to take refuge in the difficulties inherent in accounting for and characterising the emergence of thought, either ontogenetically, in individual development, or phylogenetically, across the evolutionary history of a species (as outlined in Section II.2 of the Introduction). By itself this would be an inadequate dodge, but it does highlight an avenue by which Davidson might seek to undermine the force of this particular counter-example made by Lepore & Ludwig. That is, Davidson may wish to claim that the developmental stages of language acquisition provide something of a blind-spot as to whether or not low levels of language mastery do actually indicate propositional attitudes of the child. As Davidson asserts [Dav97/05:139]:

The infant may never say “Mama” except when its mother is present, but this does not prove conceptualization has taken place, even on a primitive level, unless a mistake would be recognized as a mistake.

Clearly this again assumes too much to be anything more than a straightforward avowal of Davidson’s broader philosophical position, but it does suggest which of the above two antithetical conclusions Davidson would see dropped first. That is, although Davidson admittedly holds that language is essential to the possession of beliefs, this is subject to the further claim that only the ability to recognise mistakes as mistakes guarantees that one actually has a language (and so, any beliefs whatsoever). This in turn suggests that to Davidson, while the capacity for surprise is, as he says, ‘a necessary and sufficient condition of thought in general’ [Dav82/01:104], the necessity of language remains primary.

That being noted, all of this obviously goes well beyond the scope of The Argument from Surprise; all that is supposed to be needed for the present context is that belief entails the

44

possibility of surprise. To introduce language mastery and the ability to explain errors at this stage would not only demand a much richer characterisation of surprise than anyone will be willing to concede, but moreover, would be question-begging against the second step of Davidson’s ‘Main’ Argument. Therefore, Lepore & Ludwig’s objection holds given the notion that the claim that language mastery entails belief is at least equally (if not more) commonsensical than the first premise of The Argument from Surprise. That is, commonsense tells us that it is more plausible to suppose that the child has beliefs but simply cannot yet recognise its mistakes, than it is to hold that it has no beliefs at all. If Davidson’s argument were to be able to withstand such an objection it would have firstly to establish something like the argumentative line sketched above and incorporate it into the very structure of The Argument from Surprise without begging the question. As it stands, Lepore & Ludwig’s first counterexample undercuts the dialectical force of Davidson’s position.

3. Beliefs Without Surprise

Lepore & Ludwig offer a further case against the claim that to have beliefs, one must be susceptible to surprises. They write [Lep05:395]: ‘Imagine, for a moment, an omniscient being. By hypothesis he has beliefs, but he is obviously incapable of surprise. If anything, the […] argument seems to take us further from our goal.’ The objection here is that it is seemingly intelligible that a creature could have beliefs without being surprised. That is, it is presumed that an omniscient being has beliefs, yet the same time it must be incapable of being surprised by virtue of knowing everything. We might add, so as to avoid contradicting the underlying assumption that beliefs are either true or false, that the omniscient being’s necessarily true beliefs are so only in contrast to the never realised potentiality of being false. That is, true beliefs are true to the exclusion of the innumerable ways of them being made false.

Taken in conjunction with their previous objection that one may have language mastery without subsequently being capable of recognising mistakes, the force of this objection can be more clearly seen as revealing that it isn’t obvious why surprise is essential to belief. The primary role of belief is to believe that something is the case; surprise seems entirely incidental to it. With this is mind, Lepore & Ludwig’s objection against the dialectical force of the premise that to have beliefs one must be susceptible to surprise is well made; even if

45

Davidson’s argument were to hold, its recourse to surprise cannot sustain the intuitive appeal that it was hoped to, such that the counterintuitive conclusion it lends support to will not seem any more commonsensical thereby.

III. META-COGNITION

The second premise of The Argument from Surprise claims being surprised requires that one have beliefs about beliefs; namely, that one’s belief was false. As Davidson puts it [Dav82/01:104]:

Surprise requires that I be aware of a contrast between what I did believe and what I come to believe. Such awareness, however, is a belief about a belief: if I am surprised, then among other things I come to believe that my original belief was false.

It is by this characterisation of surprise as involving beliefs about belief that Davidson hopes to have differentiated surprise from being startled or merely changing one’s beliefs either unconsciously or without surprise. That is, it is implied that having a belief that one’s prior belief was false must be capable of triggering a surprised response, and so explains why merely changing one’s beliefs does not in itself result in surprise.

It is important to briefly address the following qualification that Davidson makes of this premise; he writes [Dav82/01:104]: ‘I do not need to insist that every case of surprise involves a belief that a prior belief was false (though I am inclined to think so).’ Presumably by this caveat, Davidson isn’t ruling out the possibility of surprise occurring, not because one came to believe that a past belief was false, but for some other meta-cognitive reason, like the realisation that one’s beliefs are odd or quaint or some other unexpected property besides falsity. Such alternatives would be permissible (at least by Davidson’s lights) in that they involve higher-order notions that inherently presume the capacity of being able to hold a belief true or false. As Davidson argues elsewhere [Dav75/84:155-6]: ‘… belief is central

46

to all kinds of thought. If someone is glad that, or notices that, or remembers that, or knows that, the gun is loaded, then he must believe that the gun is loaded.’ Analogously, if one is surprised by the fact that one’s belief turned out to be, say, odd, they must thereby believe that they have the belief in question such that they can view it as odd.

On the other hand, Davidson surely cannot allow that surprises might result from false beliefs where one doesn’t also believe that the belief is false. To demonstrate that belief entails the concept of belief requires that surprise cannot merely arise from a non-meta-cognitive change of beliefs in the face of contradictory perceptions. If Davidson weren’t to insist on this, then The Argument from Surprise would critically falter. The very force of the argument must be that the minimal degree of conceptuality demanded of surprise is a belief that one’s belief was false, because surprise only ever arises from such meta-cognition. If not, then the argument would offer no compelling ground for countenancing the alternative (commonsense) view that one might well be surprised without beliefs about beliefs, and furthermore, that one could therefore have beliefs without the concept of belief.

Carruthers argues exactly this point in proposing an account of surprise that involves no recourse to meta-cognitive beliefs about beliefs [Car08]. Carruthers claims the physiological effects characteristic of surprise can indeed be triggered by first-order cognitions that involve no explanatory recourse to meta-cognition. He proposes a convention by which to make explicit the minimal cognitive processes required, not only for surprise, but also for further claimed instances of meta-cognitive processes in non-human animals. By applying Carruthers’s own convention more systematically to the case of surprise, it will be seen that Davidson’s second premise is ultimately untenable.

1. First-Order Surprise

Carruthers treats Davidson’s Argument from Surprise as little more than ‘a cautionary tale’ – once potentially ‘plausible’, but now nothing more than a warning for those who may have failed to appreciate the relevant application of Morgan’s Canon (‘Roughly: don’t attribute to animals cognitive processes more complex than is necessary’ [Car08:2]). The account Carruthers gives of The Argument from Surprise is as follows [Car08:4]: ‘To be

47

surprised, [Davidson] argued, involves coming to believe that one of your previous beliefs is false, and hence presupposes meta-cognitive thoughts about your own belief states.’

Although Carruthers acknowledges that such a story seems credible enough in that this is indeed how we typically feel and speak about surprise, Carruthers nonetheless disagrees, stating [Car08:5]: ‘The trouble with all this is that the initial claim is false. Surprise, itself, is a purely first-order phenomenon.’ By this, Carruthers means a number of things; namely:

i) the physiological effects characteristic of surprise are not cognitive, let alone meta-cognitive;

ii) the feeling of surprise ‘involves representations of bodily states, not psychological ones’ [Car08:5], and so only involves first-order cognitions that are in no way meta-cognitive; and,

iii) the cognitive processes required to trigger the physiological effects of surprise are not meta-cognitive either.

It is the lattermost point (iii) that needs further explanation, especially in that it runs contrary to commonsense. As Carruthers’ acknowledges [Car08:4]:

… we would naturally report on our surprise in just this sort of way: ‘It gave me a real surprise. I had been confident that it would turn out one way, but then I saw that the opposite had occurred.’

Carruthers thus explains that our ‘characteristic’ and ‘natural’ tendency to suppose that our surprises are meta-cognitive is misguidedly based on the fact that [Car08:6]:

… humans are inveterate mind-readers, in the first person as well as in the third. We correctly interpret the feeling of surprise for what it is: a state that is caused when perception gives rise to conflict with prior belief or expectation. … Because we conceptualize and report that state in meta-cognitive terms, we are inclined to think of ourselves as merely expressing our conscious awareness of a meta-cognitive state.

48

According to Carruthers, we are surprised before we come to believe that we are surprised. That is to say, it is only the explanatory interpretation of a first-order conflict of beliefs that is meta-cognitive, but this post hoc understanding is always subsequent to the surprise itself. This is clearly at odds with Davidson’s argument in that it means that one could have a general stock of beliefs, be surprised, and yet not thereby have beliefs about beliefs or the concept of belief. One could be surprised and not know why. Imagine, for instance, experiencing a dream in which something surprising occurred, but the subsequent physiological effect woke you from your sleep, such that you never knew what it was that surprised you. Here the question remains as to whether the initial surprise could occur without meta-cognitively recognising that one’s previous belief had been false, but at least the scenario provides a counterexample to the otherwise intuitive notion that, if one is surprised (as opposed to being merely startled), one must know what it was that one found surprising.

Indeed, Carruthers’ account appears more plausible if we consider the likely evolutionary advantage provided to creatures possessing such a capacity even for first-order surprise. Creatures whose expectations are not met would prima facie seem to have an evolutionary advantage if they were to straightforwardly experience the kind of physiological changes common to surprise that might make them less prone to danger. Of course, being able to meta-cognitively recognise what it was that one was mistaken about such that one came to be surprised would be even more advantageous insofar as it might better allow one to determine the likely cause and so potentially dangerous results of one’s thwarted expectations. But the general point stands: this latter capacity which humans do possess, nonetheless appears to be subsequent and so, independent of the first-order phenomena outlined by Carruthers.

For Davidson, the physiological effects of being surprised must arise as a consequence of meta-cognition; specifically from the belief that one’s past belief was false. Thus, if Carruthers’ account is right, The Argument from Surprise is seemingly refuted.

In order to make his argument for first-order surprise, Carruthers introduces a novel convention (which he employs not only in respect to Davidson’s Argument from Surprise,

49

but throughout in his treatment of various ‘live’ contenders of meta-cognition in animals). As Carruthers introduces this convention [Car08:5]:

I shall use capitals to represent the mental states and attitudes that a creature has (belief, perception, desire, etc.), using square brackets the represent the contents of those states and attitudes.

And he adds, parenthetically [Car08:fn#5]:

… only representations attributed to the creature are those that figure within the square brackets. It is I, as theorist, who represents the creature’s beliefs and other attitudes. The creature itself just has those.

In this light, Carruthers’ explicit account of the how surprise is triggered without recourse to meta-cognitive processes is as follows [Car08:5]:

… surprise normally arises when a creature has an activated BELIEF [P] together with a PERCEPTION [Q], where the latter then gives rise to a novel activated BELIEF [not P]. The mechanism that gives rise to surprise is one that takes the contents [P] and [not-P] produced in this sort of way as input, and which produces a suite of reactions as output: releasing chemicals into the bloodstream that heightens alertness, widening the eyes, orienting towards and attending to the perceived state of affairs [Q], and so forth. And it is the detection of these changes in ourselves that constitutes the feeling of surprise.

To make Carruthers’ account slightly more formal (and in keeping with his own usage throughout the bulk of his article), we may interpret the cognitive process of surprise as follows:

AFirst-OrderAccountofSurprise

1.BELIEF[P]

50

2.PERCEPTION[Q]

3.BELIEF[not-P]

4.PERCEPTION[surprise]

Implicit in this process is the physiological reaction that takes place between (3) and (4). As Carruthers puts it [Car08:5]: ‘…it is the detection of [the physiological] changes in ourselves that constitutes the feeling of surprise.’ And further, as he explains in a footnote [Car08:fn#6]: ‘Note that the feeling of surprise itself involves representations of bodily changes, not psychological ones, just as other sorts of feeling – e.g. pain – do [...]. These feelings are purely first-order in character.’

According to Carruthers, typically in humans it is (4) – the feeling of being physiologically surprised – that, gives rise to the subsequent meta-cognitive identification and reports of what caused the surprise; namely:

5.BELIEF[thechangefrom(1)to(3)caused(4)]

(5) is clearly meta-cognitive in that it is a belief about beliefs (1) and (3), and so, must necessarily be only ever subsequent to the first-order process of surprise such that it is essentially extraneous to the surprise itself.

From this it should be apparent that Carruthers’ central contention is that surprise is essentially a feeling arising from a particular sequence of beliefs. Moreover, it is a process, the ‘genesis and nature’ of which requires no recourse to meta-cognition. That surprise is a particular process culminating in a particular feeling is apparent upon noting that (5) need not occur for all given instances of (4). Moreover, we can thereby suppose that non-human animals may well experience the feeling of surprise without having to commit ourselves to their having beliefs about beliefs, let alone the concept of belief. Carruthers readily acknowledges that we humans, as ‘inveterate mind-readers’ [Car08:6], naturally interpret the cause of (4) in ourselves, others, and non-human animals in terms (5), but he denies that it is anything like (5) that causes (4). Rather, according to Carruthers, it is the first-order cognitive process from steps (1)–(3) that gives rise to (4), and that (5) is but a contingency.

51

2. Unsurprising Changes in Belief

From the above, it appears that the second premise of Davidson’s Argument from Surprise is in poor standing. However, before the final verdict is passed, there remains a further issue to be addressed. If we take Carruthers’ account at face value it may appear that it is unable to explain the difference between surprising and unsurprising changes in belief. That is, Carruthers’ account of the phenomenon of surprise without recourse to meta-cognition does not transparently explain the difference between normal changes in beliefs and those that trigger surprise. Consider the following two scenarios.

SurprisethatOne’sPocketisCoinless

1.BELIEF[Ihaveacoininmypocket]

2.PERCEPTION[mypocketisempty]

3.BELIEF[Idon’thaveacoininmypocket]

4.PERCEPTION[I’msurprised]

This first scenario is borrowed directly from Davidson’s original explication of The Argument from Surprise, and moreover fits Carruthers’ account. That is, if (3) results from (2) in light of (1) without meta-cognition, then the physiological affects of surprise and (4) will both occur. However, the supposed ‘mechanism that is sensitive to conflicts between [beliefs]’ [Car08:5] such that surprise arises as the result isn’t immediately apparent here. To make this clear, consider the following account of an LED light changing colours from green to red:

AnUnsurprisingChangeintheColourofaLightEmittingDiode

1.BELIEF[thelightisgreen]

2.PERCEPTION[thelightisred]

3.BELIEF[thelightisnotgreen]

52

Initially, one has the belief that the light is green, but a change in the perceived colour of the light (presumably brought about by an actual change in the colour of the light being emitted by LED itself) subsequently brings about the contrary belief that the light is not green. But no surprise ensues. One simply has the belief that the light is green, then the contrary belief that it isn’t.

Both of the above scenarios have an identical generic structure from (1)–(3), and yet it is unclear from the two sequences as they stand why the former should trigger a surprised reaction while the latter does not. From what Carruthers has provided alone, the mere fact that in both cases (3) contradicts (1) does not explain why surprise arises in some instances and not in others. That is, although Carruthers’ claims that ‘all that [surprise] requires is a mechanism that is sensitive to conflicts between [beliefs]’ [Car08:5], exactly what that mechanism is remains obscure.

When Davidson insists that surprise requires one to have beliefs about beliefs, he may be indirectly appealing to meta-cognition as what differentiates unsurprising changes in beliefs from surprising ones. That is, he claims that surprises occur when we realise that a belief that we had held was false. In Carruthers account, on the other hand, it is not at all clear what is doing the work in differentiating between surprising and unsurprising conflicts in belief. Carruthers’ account can be likened to an opaque box: put in two conflicting beliefs and in some instances, though not all, a surprise will occur. In Davidson’s account the box is transparent, for it is only when one meta-cognitively realizes that the two beliefs are in conflict with one another that one will be surprised. So although Carruthers has given a first-order account of surprise, further explanation is needed to differentiate it from unsurprising conflicts in belief.

It is important to note that these considerations, taken by themselves, are no repudiation either of Carrurthers’ account or its bearing on The Argument from Surprise. Davidson claimed that surprise necessarily involves meta-cognition; Carruthers has provided a first-order account of surprise without recourse to meta-cognition. Case closed. The burden falls to Davidson, not Carruthers, to show why Carruthers’ account will not suffice. However, if Carruthers’ account can be readily adapted to account for the difference between surprising

53

changes in belief and unsurprising ones (as I think it can), then all the better for Carruthers’ position.

3. Expectations

Presumably, it is the difficulty of accounting for the difference between surprising and unsurprising changes in belief that Carruthers obliquely refers to when he writes that it is ‘the contents [P] and [not P] produced in this sort of way…’ that triggers the physiological reaction of surprise [Car08:5 (italics added)]. Promisingly, Carruthers also later states that surprise ‘is a state that always involves some sort of conflict between prior expectation and current belief’ [Car08:6 (italics added)]. In this light, Carruthers’ general account of surprise might be understood as implicitly ascribing to any belief an associated expectation about how long the state of affairs or event will endure. That is to say, for a newly-acquired belief to be capable of occasioning a surprised reaction, one’s original belief must have involved a tacit belief regarding the expected duration of the state of affairs or event, such that the novel belief is not only incompatible with the explicit content of the original belief, but also conflicts with its tacit expectation.

For instance, in the case of the belief that there is a coin in my pocket, I expect that it will hold true until such time as an event occurs that would explanatorily alter the relevant state of affairs (e.g. my removing it from my pocket). The perception that my pocket is empty is surprising precisely because this perception alone does nothing to offset my expectation that my pocket should contain a coin. Had I emptied my pockets and found that I did indeed have a coin (as was expected), then the equally true state of affairs that I now don’t have a coin in my pocket would be not in the least bit surprising (because I have emptied my pockets of its contents, including the coin).

In the case of the unsurprising change in the colour of an LED, the perception and novel belief that the light is red and therefore not green is unsurprising despite the fact that I had only moments before believed ‘the light is green’. The explanation here is that, unlike the coin, the associated expectation of my original belief was presumably temporarily limited to something like, ‘…until it changes colour’. Although I may believe the light to be green

54

now, it may be the case that have no expectations whatsoever as to whether or not it will be green, red, or otherwise, at any time other than the present moment in which I believe it to be green.4 Obviously, here one must presumably understand that at least some LEDs are capable of emitting more than one colour of light, otherwise, if one had believed that a light’s colour must remained fixed and unchanging, then presumably one’s first encounter with a colour-changing LED would be surprising. Such background knowledge affecting one’s expectations need not be theoretical. All that is required is familiarity with the changeable nature of the item.

In the unsurprising change in colour of an LED, we see that, although (2) gives rise to (3), surprise does not ensue, because (3) simply replaces (1) and so does not conflict with it in the right way. This is because (1) involved no expectation that it would endure beyond the moment for which it was held. (Of course, if one was subsequently shown that one’s belief had been false for the time that one held it, say, via the play-back of a video recording, or the revelation of colour-blindness, etc., then one would surely be surprised.)

With these considerations in mind, Carruthers’ account can be developed further so as to more clearly indicate when and why incompatible beliefs give rise to surprises and when they do not, e.g.:

ATemporally-Indexed,UnsurprisingChangeintheColourofaLightEmittingDiode

1´.BELIEF[thelightisgreent1]

2.PERCEPTION[thelightisred]

3´.BELIEF[thelightisnotgreent2]

ATemporally-Indexed,SurprisethatOne’sPocketisCoinless

1´.BELIEF[Ihaveacoininmypockettn]

2.PERCEPTION[mypocketisempty]

4 It is perhaps tempting to presume that a more elegant solution would be attribute expectations to all beliefs, but, while some explanation remains in accounting for why some beliefs may have expectations of likely durations where others do not (presumably this has something to with the subject’s familiarity with the kind of content their belief is about), nonetheless, I consider it more in keeping with Morgan’s Canon not to overburden beliefs with the necessary condition of such expectations.

55

3´.BELIEF[Idon’thaveacoininmypockett2]

4.PERCEPTION[I’msurprised]

In this way, we seem to have successfully accounted for why cases like the above do not result in surprise. By adopting Carruthers’ passing suggesting that surprise ‘is a state that always involves some sort of conflict between prior expectation and current belief’ [Car08:6 (italics added)], we see that for a surprise to result, the change in belief must conflict with the prior belief via its tacit expectations. Examples of beliefs that typically do not carry with them significant expectancies include, ‘It is 4.11 pm’, ‘I am pleased’, or ‘The LED is green’. It is not difficult to imagine scenarios where subsequent, though incompatible beliefs (e.g. ‘It is 4.12 pm’, ‘I am disappointed’, or ‘The LED is red’) would result in surprise. This is because such initial beliefs do not carry with them significant expectations as to the duration of their holding true. This means that novel beliefs, despite being incompatible with the original, do not conflict with them in the sense of falsifying them. Often, as in the case of the LED, they simply replace the original.

Other beliefs – such as ‘There is a coin in my pocket’, ‘My old friend, Harkanwal, is an amiable fellow’, or ‘2 + 2 = 4’ – carry with them significant expectations such that if it turns out there is no coin in my pocket, or Harkanwal has become fractious, or two plus two somehow now equals anything other than four, then I will certainly be surprised. This is because the novel belief conflicts in the right way with the original belief, not merely in the sense of replacing it, but falsifying it. The new belief directly calls into question the truth of the previous belief. This should be clear also in the case of the missing coin: the fact that one’s pocket is empty is surprising not merely with regard to the particular moment that one realises that this is the case, but also with regard to the time preceding the realisation. That is, (2) perceiving that one’s pocket is empty conflicts with (1) the belief that one has a coin in one’s pocket in such a manner as to call (1) into question with respect to the entire duration leading up to the moment one realised it was false. This is appreciable from the natural thought process that might follow such a surprise, where one wonders whether or not the belief had ever been true (e.g. ‘Did the coin even make it into my pocket?’; ‘Might I have missed when I went to put it in there?’; ‘Did I just assume that I would have put it in there?’) or whether it had only become erroneous at some later stage (e.g. ‘Have I been pick-pocketed?’; ‘Is there a hole in my pocket?’). Either way, the upshot is that the original belief is directly contradicted by the latter belief in such a way as to call into question the truth of the original, and this is what we find surprising. The notion of expectation is

56

thereby able to account for the difference between surprising changes in belief and unsurprising ones.

In this light, Carruthers has provided good theoretical reasons for rejecting not only Davidson’s second premise, but also the entire Argument from Surprise outright. If surprise can occur without one having beliefs about beliefs, then we have little reason to suppose that the concept of belief (and, so, the concept of objectivity) is a necessary concomitant of belief. Davidson’s goal then of establishing – via The Argument from Surprise – that it is only if one has the concept of objectivity that one can be said to have any beliefs whatsoever appear to be dashed.

IV. CONCEPTUALISATION

Despite Davidson’s Argument from Surprise being shown to have critically failed via Carruthers’ first-order account of surprise, it is prudent to nevertheless consider the third and final premise of the argument, which holds that if one has the belief that one’s prior belief was false, then one must have the concept of belief. Lepore & Ludwig regard this move as perfectly acceptable. As they explain [Lep05:395]: ‘… it can be granted that […] to think that one’s beliefs were false [...] requires that one have the concept of belief.’ Glock, however, disagrees.

As was discussed in Section I, Glock’s construal of the argument regards the original focus on surprise as beside the point. In this light, Glock treats the third premise as resting solely on the issue of whether the recognition of a mistake entails that one has the concept of a mistake. Glock sees the concept of a mistake as tantamount to the intended conclusion of Davidson’s original argument (i.e. possession of the concept of belief), ‘since the concept of a mistake at issue is that of a mistaken belief.’ [Glo03:286] Thus, although Carruthers’ first-order account of meta-cognition means that we have little reason to entertain the third premise as it was originally construed, Glock nonetheless deems it worthwhile to consider whether the weaker, first-order claim might nonetheless be capable of entailing the concept

57

of a mistake and so, of belief. Unsurprisingly perhaps, he argues that it isn’t. Glock writes [Glo03:288]:

Even if a is capable of recognizing a mistake, and, in that sense, of understanding the possibility of being mistaken, it remains to be shown why a needs to have the concept of a mistake, as Davidson has it. What a must be able to recognize is that an object x which it initially took to be F, is not F after all. However, there is as yet no need for a’s recognition to display a grasp of the general concept of a mistake, a concept that covers not just a’s own misapplication of the specific concept F, but also anyone else’s misapplication of any other concept, including concepts that a lacks. For a chimpanzee to recognize that it is dealing with dorylus rather than macrotermes, it does not need to recognize that its own previous belief falls under the same concept – that of a mistake – which also applies to all other mistakes, e.g., Columbus’ belief that he had reached India. Neither does it need to recognise that its previous classification falls under the same concept – that of a misclassification – as that of someone who classifies whales as fish.

To reiterate: by ‘recognition’, Glock (in keeping with Carruthers) is only countenancing first-order cognitions. To make this perfectly clear, consider Glock’s earlier characterisation of recognition [Glo03:288]:

For example, after barking up the oak tree for a while, [a] dog shows signs of disappointment and frustration; it then turns round and looks for the cat [that it has been chasing] somewhere else; finally, on spotting the cat, it starts barking up the pine tree with renewed vigor. It has not simply spotted the cat on the pine tree, but recognised that the cat went up the pine instead of the oak. However, that is not necessarily to recognise that its prior belief that it went up the oak tree was mistaken, even though the fact that the cat did not go up the oak tree does entail that that prior belief was false.

That is, Glock by no means regards the recognition of a mistake – including ‘the understanding of the possibility of being mistaken’ – to be a meta-cognitive process. All that is required is that one responds appropriately to perceptions contrary to one’s standing beliefs with frustration, confusion or surprise. Such an ability to recognise one’s mistakes

58

does not entail the full-blown concept of a mistake, which Glock implicitly takes as needing to fulfil Gareth Evans’ generality constraint. As Evans himself states this condition [Eva82:104]:

If a subject can be credited with the thought that a is F then he must have the conceptual resources for entertaining the thought that a is G, of every property of being G of which he has a conception.

That is, for any particular entity and property represented as a state of affairs (eg. ‘Harry is happy’), for one to have the concepts of Harry and happy, one must be able to represent these elements in alternative states of affairs, such as ‘Harry is sad’ or ‘Harkanwal is happy’. Similarly, in the case of the concept of mistake, one must have the potential to be able to apply it to other relevant instances of a mistake. Glock clearly takes such a potential capacity to go far beyond what is required for the mere recognition of one’s own particular mistakes. For instance, if a psychologist were to attempt to establish that a subject possessed the concept of a mistake by the subject’s ability recognise mistakes, then a test would have to be devised so as to reveal that the subject was, for instance, problem solving by way, not of the particular content of their mistaken beliefs, but by the fact that their beliefs had been mistaken at all. Even so, such a test would only establish by way of the subject’s capacity to recognise mistakes that it had a concept of mistake, not that it had the concept of a mistake in virtue of being able to recognise its mistakes. Thus we see the disconnect between the recognition of a mistake and the concept of mistake. If this is right, then Glock’s critique will prove to be the final nail in the coffin of Davidson’s Argument from Surprise, for, even if we were to put aside Carruthers’ refutation of the second premise, the third and final premise still cannot get off the ground.

That said, it is perhaps worth noting that Camp has made a compelling proposal to re-evaluate the application of Evans’ generality constraint as the deciding measure of conceptualisation. Ultimately, Camp concludes that there are good reasons to warrant the acknowledgment of differing levels of strictness for what is to be regarded as a concept and what isn’t. As she puts it [Cam09:305]:

59

I don’t believe that the theoretical and practical considerations […] warrant imposing full generality as an absolute condition on conceptual thought. […] Instead, we should view conceptuality as a matter of degree.

In this light, Glock’s argument might be seen as going too far in objecting that the recognition of a mistake cannot be reasonably thought to entail the ‘concept’ of a belief. After all, Glock would likely want to endorse Camp’s characterisation of what she calls a ‘minimalist view’ of concepts [Cam09:280]. Such a minimalist view regards concepts as only needing to require ‘cognitive, representational abilities that are [at least] causally counterfactually recombinable’ [Cam09:302], but they do not demand that such recombinations be stimulus-independent. For instance, with respect to the scenario above: when the dog, upon finally locating the cat, switches to barking up the pine tree, we assume that it believes it is still barking at the same cat that it had moments ago recognised wasn’t up the oak. Thus, by Camp’s view, it is seen as having at least a minimalist concept of the particular cat in question. Higher-level conceptual abilities might well require ‘cognitive, representational abilities that are systematically recombinable in an actively self-generated, stimulus-independent way’ [Cam09:302], or – higher again – ‘representation abilities whose epistemic status the thinker can reflect upon.’ [Cam09:302] But such higher-order conceptual capacities are clearly not what Glock has in mind with regards to concept of a mistake (or belief). According to the minimalistic view, all that the concept of a mistake requires is that one can track one’s mistakes in novel (though by no means, stimulus-independent) circumstances. If this is so, then a prime example of a creature having the concept of a mistake would be a lab-rat’s capacity to identify the visual markers (e.g. colour-coding) of a pathway that had misled them once before in a different maze. Glock would likely admit that, here, the recognition of a mistake will most likely always amount to having the minimal concept of a mistake. (Whether or not such capacities follow of necessity would be a matter for empirical investigation, however.)

That being said, if we take a moment to look back to the original discussion leading up to Davidson’s Argument from Surprise, it is, however, undoubtedly only the higher-order conceptual abilities that Davidson seeks. This is clear in the following [Dav82/01:103 (italics added)]: ‘What I think is clear is that if [one] is surprised, [one] does have reflective thoughts, and, of course, beliefs.’ What Davidson here means by ‘reflective thoughts’ is the capacity for meta-cognition. This, however, goes far beyond the minimalistic ‘conceptual’ abilities that were described by Camp and are presumably attributable to Glock’s scenario of

60

the dog barking up the wrong tree. Given that such high-level conceptual capacities were thought by Davidson to be the upshot of his Argument from Surprise, then despite Camp’s potentially accommodating hierarchy of concepts, Glock’s objection is seen to stand: it is highly unlikely that the recognition of mistakes entail one to have anything more than a minimalistic concept of belief. Davidson’s third premise then also appears undermined.

V. CONCLUSION

Davidson argues that to have any beliefs whatsoever, one must have the concept of belief (and so concomitantly, the concept of objectivity also). The Argument from Surprise is offered to arrive at this conclusion by the following claims: (1) if one has beliefs, one must be susceptible to surprises; (2) to be surprised demands that one believes that one’s past belief was false; and (3) a belief that one’s past belief was false entails that one has the concept of belief.

As we have seen with respect to the first premise, although Glock points out that surprise is not the only possible response to a false belief, the supposed dialectical force of a pre-theoretical understanding of surprise suggests that the argumentative focus should remain fixed on the phenomenon of surprise. Lepore & Ludwig, however, directly attack the supposed common sense of the original argument, citing empirical evidence that (even by Davidson’s own lights) suggests that one can have beliefs and yet may not might be able to recognise one’s mistakes, let alone be surprised by them. Lepore & Ludwig further provide a brief though intuitive thought-experiment of an omniscient and therefore infallible believer to illustrate the apparent incidental role that we normally suppose surprise to play in the possession of beliefs. In this way, the notion of beliefs without surprise appears readily intelligible. While none of these objections amount to an outright refutation of the claim that in order to have beliefs one must be subject to surprises, they nevertheless work to significantly undermine the dialectical force of the Davidson’s argument. The burden then falls heavily on the shoulders of those sympathetic to Davidson to show that surprise does arise from the possession of beliefs by necessity such that it can be said that ‘surprise about some things is a necessary and sufficient condition of thought in general.’ [Dav82/01:104]

61

With respect to the second premise, a slightly amended version of Carruthers’ account of first-order surprise, which manages to distinguish surprising changes in belief from unsurprising ones, is shown to offer the most damning objection against Davidson’s claim that to be surprised requires that one have beliefs that one’s past belief was false. Davidson’s original argument hinges on the notion that surprise requires beliefs about beliefs. Given that Carruthers has offered a compelling account of surprise as a first-order cognitive process, the force of The Argument from Surprise is seen to falter. But, we may ask, is it refuted? As was noted in Section III, in his original explication, Davidson qualifies his statement as follows [Dav82/01:104 (italics added)]:

I do not need to insist that every case of surprise involves a belief that a prior belief was false (though I am inclined to think so). What I do want to claim is that one cannot have a general stock of beliefs of the sort necessary for having any beliefs at all without being subject to surprises that involve beliefs about the correctness of one’s own beliefs.

If Davidson acknowledges that some surprises can be of the first-order variety (such as Carruthers’ account suggests is reasonable to suppose), but nonetheless maintains that if one has beliefs, one will necessarily be susceptible to higher-order surprises as well, then some may be tempted to suppose that Carruthers’ account misses its mark. But what reason has Davidson given us to believe that such higher-order meta-cognitive surprises must potentially occur if one has any beliefs whatsoever? None, unless The Argument from Surprise shows that meta-cognition is essential to any and all surprises. As Tyler Burge points out [Bur10:268]:

[The second premise begs] the critical question. It is certainly not ‘clear’ that if an individual is surprised, the individual has reflective thoughts. Davidson’s claims may … introduce an unargued transition between an ordinary ‘first-order’ notion of surprise and a higher-order notion requiring reflective thoughts. The transition needs argument. Davidson gives none. He simply skates over the distinction.

It is, however, interesting to note a further claim made by Burge in a footnote. He emphatically states (against Davidson’s arm-chair tendencies) that ‘it is an empirical question whether having beliefs always goes with having beliefs about beliefs.’ [Bur10:268] Given our

62

also sympathetic commitment to naturalism, perhaps Carruthers’ account alone cannot provide the final word in reducing Davidson’s Argument from Surprise to nothing more that ‘a cautionary tale’. Not only must we have a coherent first-order account of surprise, we also require actual empirical data that fits such an account. Unfortunately Carruthers provides none nor cites any.

To this possible line of counter-objection, however, Carruthers would presumably respond that empirical data may well provide evidence either for or against the claim that beliefs entail beliefs about beliefs, but only to the extent that such data is interpreted. The function of an account like the one he gives is to suggest that first-order interpretations can indeed be given where all too often meta-cognitive interpretations are the presumed default. As Carruthers writes of the moral given in light of Davidson’s (presumed) failed argument [Car08:6] :

… we should be wary of our tendency to assume that meta-cognitive classifications are classifications of meta-cognitive states, pushing the meta-cognitive character of the categorization process ‘downwards’ into the state itself. […] There is good reason to think that [this moral has] often been ignored in the burgeoning literature on meta-cognition in animals.

That is, although the final verdict must be grounded in empirical data (a claim with which both Burge and Carruthers, and perhaps even Davidson in some moods, would concur), Carruthers stresses that raw data must always be interpreted, and we have a tendency to unthinkingly and unnecessarily ascribe meta-cognition where there is none. Determining when and where such misleading interpretations have occurred will be a matter for theory and dialogue together. This then, is what Carruthers’ account aims to pave the way for with regard to the phenomenon of surprise.

Finally, although the third and final premise of Davidson’s Argument from Surprise is accepted by Lepore & Ludwig, in light of Carruthers’ objection against the second premise, then Glock’s objection that the recognition of false beliefs does not necessarily entail the concept of belief appears reasonable, at least to the extent that conceptualisation is taken to require meta-cognitive reflection. Recent discussion by Camp, however, suggests that what counts as ‘conceptual thinking’ should be regarded as a matter of degree such that even

63

Glock’s first-order account of the recognition of mistaken beliefs might nonetheless be deemed conceptual. And yet, even if Camp’s loosened constraints are accepted, Davidson’s explicit characterisation of surprise entailing ‘reflective thoughts’ goes far beyond what might be reasonably inferred by the phenomena of surprise. In this light, the final step of Davidson’s Argument from Surprise is also seen to critically falter.

* * *

As was outlined in the introduction to Part One, Davidson’s Argument from Surprise is advanced in support of his ‘Main’ Argument. The intended conclusion of the Argument from Surprise is held by Davidson to help subsequently establish the more radical claim that to think at all is to have the concept of our beliefs as being potentially either true or false, and so to possess the concept of objectivity. Unfortunately, and as we have now seen, the Argument from Surprise critically fails, and so provides little to no support for the more radical thesis.

As was also explained in the introduction to Part One, Davidson’s ‘Main’ Argument is exclusionary in nature. That is, it is intended to establish Davidson’s global lingualism; namely that thought – as basically requiring belief – is not possible without language. With the failure of The Argument from Surprise, however, we see this peculiar thesis crumble. Given we have little reason to suppose that beliefs are not possible without the concept of belief, it is improbable that language will be necessary for belief. Perhaps this is unsurprising, for as the likes of Jason Bridges argues [Bri06], it is counterintuitive in the extreme that the both lay and scientific attribution of beliefs to animals without language is just a useful, yet technically unsupported heuristic for our making sense of their behaviour. As Bridges puts it [Bri06:309]:

[There is] a very general conceptual framework for thinking about animal life, one whose details will be filled in quite differently for different species. But it does not take much scrutiny of our everyday thought in terms of this framework, or of the various sciences that attempt to improve upon and systematize that thought (ethology, ecology, comparative psychology, animal cognition), to realize that our ascriptions on this front are holistic in a way analogous to the holism Davidson famously brought to our attention in the interpretation of speakers.

64

That is to say: while our theoretical considerations do not have to slavishly follow science, nonetheless, following the Millikan quotation included in the Introduction (see [Mil12:889]), it does have to at least be sensitive to what our best science has to say. As Huw Price puts it [Pri08/13:4]: ‘Science tells us that we humans are natural creatures, and if the claims and ambitions of philosophy conflict with this view, then philosophy needs to give way.’ In such a naturalistic light, regardless of whether or not they are capable of surprise (or indeed any other singular capacity), the claim that animals don’t have beliefs is untenable. At any rate, without the Argument from Surprise in hand, if language is indeed to play a non-trivial role in our coming to possess the concept of objectivity, we should no longer (assuming we ever did) expect that such a finding might suggest either language or the concept of objectivity is necessary for any and all thought whatsoever. Rather, our concern from here on in is exclusively with the role played by language in our possession of the concept of objectivity.

65

CHAPTER TWO

* * *

Error

I. INTRODUCTION

The Argument from Error – or, more accurately, The Argument from the Concept of Error – is intended to establish the second premise of Davidson’s ‘Main’ Argument – i.e., that to have the concept of belief, one must have language. Taking up where The Argument from Surprise was hoped to have left off, The Argument from Error aims to demonstrate, by way of the concept of error, a necessary relationship between belief, the concept of belief, and language. As Davidson puts it [Dav75/84:170]:

Someone cannot have a belief unless he understands the possibility of being mistaken, and this requires grasping the contrast between truth and error – true belief and false belief. But this contrast, I have argued, can emerge only in the

66

context of interpretation, which alone forces us to the idea of an objective, public truth.

In this quotation, Davidson moves from belief simpliciter, to the possession of the concepts of both belief and error, and all the way to the phenomenon of interpretation lying at the heart of language. However, given that it was specifically Davidson’s Argument from Surprise that was intended to establish the relationship between belief and the concept of belief, herein The Argument from Error is to be understood as starting exclusively from the higher-order concept of belief. That is to say, given the criticisms levelled against the Argument from Surprise in the previous chapter, we have little reason to suppose that having beliefs requires one to possess of the concept of belief. Moreover, while it was conceded that in order for first-order surprises to occur, a creature must be able to minimally recognise that it has made a mistake, this was deemed insufficient to attribute the higher-order concepts needed for Davidson’s commitment to global lingualism to get off the ground. By contrast, our overarching concern is to ascertain a non-trivial relationship between language and the concept of objectivity. In this regard then, The Argument from Error remains of interest insofar as it promises to establish the following relationship between the concept of belief and language use:

TheArgumentfromTheConceptofError

1.Ifonehastheconceptofbelief,onehastheconceptoferror;

2.Ifonehastheconceptoferror,oneisalanguage-user;

3.Therefore,ifonehastheconceptofbelief,oneisalanguage-user.

The structure and findings of this chapter are as follows:

In Section II, I explore what precisely the concept of error is thought to consist of in light of the constraints imposed above. Specifically, I argue that the concept of error that Davidson’s argument requires is the meta-cognitive concept of a mistaken belief.

In Section III, it is conceded that the ‘reciprocal’ of the second premise – i.e. that language readily supports the concept of error – is uncontroversial. However, Davidson’s original

67

claim that only in the presence of language can the concept of error emerge is found wanting. In particular, Lepore & Ludwig provide three languageless contexts where the concept of error may find ready application. That being noted, the dialectical force of Lepore & Ludwig’s critique is seen to be mitigated given that it is directed at what they claim to be Davidson’s fundamental commitment to the primacy of the third-person point of view, whereas, by contrast, the scope of our interests herein are not so ambitious.

In Section IV, I propose three novel applications of the concept of objectivity which appear to be dependent entirely on language. It is not my wish to in any way suggest that these three particular contexts establish Davidson’s claim that the concept of objectivity is necessarily dependent on language. Rather, I seek only to shift the dialectic burden even further away from the kind of necessity claims explored in Part One generally, instead commending the kind of substantive, non-trivial, empirically grounded accounts I hope to pursue in the remainder of the thesis.

II. THE CONCEPT OF ERROR

To establish that the concept of belief entails the concept of error, Davidson emphasises that, insofar as it is the inherent nature of beliefs that they may or may not be true, beliefs are therefore liable to be erroneous (see [Dav82/01:104]). Lepore & Ludwig elaborate this line of reasoning as follows [Lep05:397]:

To have the concept of a belief, one must have the concept of error, or, what is the same thing, of objective truth (that is, a way things are independent of how one believes them to be). This is to say that having the concept of belief is having the concept of a state which is capable of being true or false; to understand this is to understand the contrast between the merely subjective, what one merely believes, and how things really are, the objective world […]

Let us unpack this more carefully: On the one hand, a person who may be said to believe p at least implicitly holds p to be true. On the other, it is the nature of belief that what one believes may be either true or false – this can be appreciated in the difference between

68

beliefs, which may be true or false, and knowledge, which only counts if true (as Socrates emphasised in Plato’s Gorgias (see [Pla94:95-100])). Taken together, these two points mean that the mere fact of believing p does not guarantee p is true. Furthermore, it means that if what one believes is not the case, then, although one takes (and implicitly hopes) it to be so, insofar as it is believed, one will be in error in holding this false belief to be true. In such light, the concept of belief can be seen to entail not only the concepts of truth, falsity, and error, but the concept of objectivity also. This is because, given that the concepts of truth and falsity are required to appreciate the epistemically contingent nature of belief (i.e. that one might believe p, even if p is not the case), the concept of objectivity must also be in play such that one can appreciate that the way things are objectively is at least in some respect independent of the subjective standing of one’s beliefs.

While the above summation supports Lepore & Ludwig’s interpretation of the first premise of Davidson’s Argument from Error, in light of the findings of Chapter One, we may pause to wonder if the matter is as straightforward as it first appears. Specifically, given that a creature can have beliefs and be able to recognise mistaken beliefs without thereby being said to possess a second-order concept of belief, the question arises as to whether such creatures might then have at least a lower-order, minimalistic concept of belief such that they can be said also to possess a lower-order concept of error that may or may not entail some conception of objectivity.

In order to better appreciate what’s at stake here, let us consider one of Davidson’s own examples in light of our best current ethological understanding. Davidson writes [Dav87/04:116]:

… think of a tribe of monkeys the members of which respond to the threat of danger by emitting a certain cry. Other monkeys, hearing this cry, respond as to danger… To explain the behavior of the monkeys we do not need to attribute intentions or beliefs to them (I am not arguing that they don’t have intentions or beliefs). And so nothing in their behavior as described has to count as making a mistake. Of course it can happen that a monkey lets out the ‘danger cry’ when danger is not present, and his fellows may react as to danger. From our point of view this may seem a mistake. But unless the monkeys believe there is danger when there is not, no error has been committed; they have simply responded to a stimulus that usually, but not always, accompanies danger.

69

Davidson holds that, insofar as we do not need to attribute intentions or beliefs to the monkeys in question, nothing in their behaviour ‘counts’ as having made a mistake. This position rests on his subsequent claim that ‘it is only where intention and belief are present that the concept of a mistake can be applied’ [Dav87/04:116]. But even if we allow this latter point, it is, as was argued in the conclusion of the last chapter, essentially a matter for our best empirically-based ethology to determine whether or not the monkeys’ behaviour can be adequately explained without appeal to intention or belief. As we saw Bridges argue, it does not seem that this is at all likely, not only in the scenario described by Davidson, but elsewhere also. Indeed, against Davidson’s claim that the application of the concept of a mistake is dependent upon the presence of intention and belief, Burge writes in regard to perceptual systems as conceptualised in the empirical sciences [Bur10:270]: ‘…a mistake “as seen from the creature’s point of view” [to quote Davidson] is simply a non-veridical perceptual (or other type of) representation by the creature’, where, according to Burge’s take on current ethology, ‘perceptual accuracy is [merely] success in fulfilling perception’s representational function’ [Bur10:379]. Here then it seems that at least the possession of some higher-order concept of belief is irrelevant to the scientific application and attribution of the concept of mistake.

That being said, Davidson’s above imagined scenario of a tribe of monkeys admittedly does not trade on non-veridical perceptions – the caller does make the alarm call that we are tempted to say its conspecifics react to ‘in error’ regardless of whether they have been misled intentionally or not. Perhaps Burge’s qualification that certain types of non-veridical representations (other than solely perceptual ones) may equally warrant application of the concept of mistake, might also be seen to allow for the particular notion of mistake being appealed to by Davidson when he says that, ‘from our point of view this may seem a mistake’. However, we would do well to consider if Davidson is attempting to point out a distinction that he takes to be significant. That is to ask: is there a significant difference between creatures that possess a minimalistic concept of mistake (presumably attributable to the actions and perhaps even conceptualisations of creatures that have beliefs but at the same time lack a higher-order concept of belief) and creatures that possess the entire gamut, including the higher-order notion as well?

In thinking about this question, empirical research has revealed far more fascinating and compelling phenomena in the behaviour of monkeys than was envisaged in Davidson’s

70

imagined hypothetical scenario. Indeed, the fact of these real-life phenomena, I suspect, better reveal what was supposed to be at stake in Davidson’s hypothetical scenario. Specifically, Michael Tomasello invokes empirically-based ethological explanations of the behaviour of vervet monkeys, when he writes [Tom08:15–16]:

When a vervet monkey hears a ‘snake alarm call’, it knows that a snake is nearby; when it hears an ‘eagle alarm call’ it knows that an eagle is nearby. Vervet monkey recipients thus extract referentially specific information from alarm calls, and this has been demonstrated repeatedly by playback experiments in which the call is played over a loudspeaker when no predator is nearby – and recipients still engage in the appropriate predator specific avoidance behavior.

Tomasello, however, points out that ‘in stark contrast to this picture of flexible comprehension, monkeys and apes do not learn to produce their vocal calls at all, and they have very little voluntary control over them’ [Tom08:16]. Most tellingly, continues Tomasello, the caller instinctively gives the alarm call irrespective of the whereabouts of its conspecifics [Tom08:18–19]:5

Evidence that the caller typically ignores the audience comes from the fact that vervet monkeys quite often persist in giving their alarm calls even when all the individuals of the group are already in some safe position looking at the predator.

Looking back to Davidson’s initial scenario, then, it seems fair to say that, as we have already conceded, the caller has indeed made a lower-order mistake if it gives the alarm call on the basis of a non-veridical perceptual representation (i.e. it mistakenly thinks it sees a snake). Nonetheless, it is not clear that the conspecifics who then react to this mistaken call can be said to have made an error as such. To be clear: by this I am not saying that the monkeys’ reaction to the false alarm call is not in error in the sense that Davidson holds – i.e. ‘unless the monkeys believe there is danger when there is not, no error has been committed’ [Dav87/04:116]. As Tomasello explains: they will believe that danger is present 5 That said, more recent studies (see [Cro17]) have found that chimpanzees make fewer alarm calls if they have heard the alarm call made recently before – that is, as the authors conclude, if they believe that their conspecifics are already aware of the threat. This capacity, however, has little bearing on our discussion here.

71

when they hear the call, regardless of whether the call itself is mistaken or not. Rather, the point is that, insofar as the caller gives the alarm call instinctively, but not intentionally for the sake of communicating to its conspecifics, the latter have not then made a mistake in reacting to the call as they normally would. As Davidson puts it [Dav87/04:116]: ‘…they have simply responded to a stimulus that usually, but not always, accompanies danger’. Indeed, Tomasello is seen to endorse this view when he writes [Tom08:19]:

The problem is that such ‘comprehension’ skills are not specialized only for communication; they are merely general skills of cognitive assessment. Thus, when a monkey learns that a certain alarm call of a certain bird species, or even of its own species, predicts the presence of a leopard, it is not clear that this is best thought of as the comprehension of a communicative act. The monkey has simply learned that one thing predicts another, or even causes another, in the same basic way as many other phenomena in their daily lives.

By contrast, in Aesop’s ‘The Boy Who Cried “Wolf”’ ([Aes03]), when the townsfolk are misled by the boy’s first cries, the townsfolk not only temporarily form a false belief that a wolf is present, but they further mistake the meaning of the boy’s cry as being intended to honestly convey the presence of a wolf and not to have purposely misled them for his own amusement. This capacity to mistake the meaning of an utterance depends as much upon the townsfolk’s capacity to see the boy’s call as a communicative act intended to convey a particular meaning, as it does the boy’s own capacity to mislead.6 As this example shows, language is readily capable of generating mistakes where before there had been none. In the case of the monkeys, the mistakes (such as they are) occur only insofar as both the monkey who gives the alarm call and those who hear its call think that there’s a snake where there isn’t one. By contrast, in the case of the boy who called ‘Wolf!’, the townsfolk are mistaken, both, with respect to their believing there to be a wolf when there isn’t one, but also in taking the boy’s cry of ‘Wolf!’ to be a genuine expression of his belief (mistaken or otherwise) that a wolf was present. As Davidson puts it: ‘I take it as obvious that linguistic behaviour is intentional and so requires belief’ [Dav87/04:116]. In this way, (at least a certain range of) linguistic behaviour is taken to suffice for the particular application of mistake that Davidson has in mind.

6 An extended discussion concerning the significance of deception is undertaken in Chapter 4.

72

But, while it may be obvious that language sufficies, given that we have rejected Davidson’s insistence that the concept of mistake is applicable it is only where we cannot explain behaviour without attributing intentions or beliefs, it remains unclear from the above as to how and on what grounds we are to differentiate first-order mistakes from what are implied to be the particular, higher-order concept of error needed to establish a dependency relation on the concept of belief. After all, it would beg the question to suppose that our common sense recognition that language suffices for possession of the higher-order concept of error could be invoked to support the first premise of The Argument from Error, given that the intended conclusion of the argument is to establish that language is necessary for the concept of belief.

My suggestion here, then, is that, while the lower-order concept of error need not require the concept of belief as such, the higher-order one does. That is to say, the lower-order concept of error requires only that a creature is able to recognise its mistakes. (See Chapter 1, Section IV for discussion of what such recognition may amount to.) The higher-order concept of error, on the other hand, necessarily involves both the capacity to recognise one’s mistakes coupled with an application of the concept of belief. To reappropriate Davidson’s original (albeit unsupportable) characterisation of surprise [Dav82/01:104 (emphasis added)]: ‘[The higher-order concept of error] requires that I be aware of the contrast between what I did believe and what I come to believe. Such awareness … is a belief about a belief’. Claudine Verheggen makes much the same point when she writes [Ver97:367]: ‘…there is a gap between the claim that someone has made a mistake and the claim that he has the concept of a mistake, and hence the concept of things being thus and so independently of his thinking that they are thus and so’. In such a light we might define the concept of error as follows:

The concept of error is the concept of a mistaken belief.

To illustrate the point with respect to the contexts discussed so far: in the case of the monkeys responding to a mistaken alarm call, the hearers of the call are not capable of being in error about the meaning of the alarm call because, to them, it is, as both Davidson and Tomasello agree, just a matter of ‘simply respond[ing] to a stimulus that usually, but not always, accompanies danger’ [Dav87/04:116]. In the case of the boy who cried ‘Wolf!’, on the other hand, the townsfolk are tricked by the boy insofar as they possess the concept of belief, which they apply to the boy in viewing him as truly believing there is a wolf present.

73

It is a higher-order error, not merely because the boy has devious intentions, but moreover because the townsfolk presume that he does not. The force of this interpretation is, I suspect, made most apparent in the second of the boy’s three calls. In the first case, the townsfolk are closer to the monkeys in unthinkingly responding to the call as a mere stimulus. In the third, they again (wrongly) presume to know the boy’s deceptive intentions. In the second case, however, they consciously trust that the meaning of the boy’s cry is intended to convey his genuine belief that a wolf is present – they choose to believe him, expecting that their earlier reprimand would have its intended effect. While I think the dispositional belief of the first case is sufficient for the higher-order concept of error to be applicable, it is seen to be so especially in light of the occurrent instance of the second.

Assuming, then, that we have adequately captured the difference between the higher and lower-order concepts of error (a point still to be ‘stress-tested’ with regard to the subsequent inferences of The Argument from Error), it should, nonetheless, be clear that this definition does not beg the question against the conclusion of The Argument from Error. Recall that earlier, concern was expressed that by appealing to Tomasello’s conclusion that the behaviour of vervet monkeys who seek refuge in response to an alarm call being made by a conspecific is not ‘best thought of as the comprehension of a communicative act’ [Tom08:19], it would beg the question, then, to presume that the difference between the ability to comprehend communicative acts as being communicative acts or not must rest on the creature themselves being a language user. By contrast, our subsequent supposition that the higher-order concept of objectivity is distinct from its lower-order counterpart in that it further requires application of the concept of belief, does not beg the question in this way.

That being said, it is prudent also to consider whether this definition might not beg the question against the first premise of The Argument from Error and, in particular, the reading that was given of Davidson’s (as well as Lepore & Ludwig’s) justification for it.

At first glace, the definition of the concept of error that has been advanced herein does not beg the question because the entailment runs contrariwise to that the first premise. That is, it is claimed that to possess the concept of error one must have the concept of belief (and also be capable of recognising mistakes), but the entailment of the first premise runs the other way in holding that if one possesses the concept of belief, then one must have the concept of error. Recall, however, that Lepore & Ludwig endorsed Davidson’s claim insofar

74

as ‘having the concept of belief is having the concept of a state which is capable of being true or false’ [Lep05:397]. Now, from what has been said previously, a state that is capable of being true or false, does not – as Lepore & Ludwig otherwise imply – require the concept of objectivity. It is sufficient for a belief that it be a representation that is either veridical or (albeit less frequently) non-veridical. Allowing, then, for such a minimalist view of the concept of error, it no longer appears that possession of the concept of belief straightforwardly entails the higher-order concept of error.

This is clearly a problem insofar as Davidson’s ‘Main’ Argument hoped to establish the global lingualist thesis that language is necessary for any beliefs whatsoever. However, given we have already jettisoned this aspiration along with The Argument from Surprise, we can instead merely stipulate that the conception of belief posited in the first premise of The Argument from Error be the higher-order one. That is to say, the first premise holds that if one has a concept of belief as being a state that may be true or false irrespective of one taking it to be true, then the higher-order concept of error is entailed. Nor does this stipulation water down the significance of The Argument from Error, for our concern herein is principally to discern the relationship between language and the concept of objectivity. If, as The Argument from Error suggests, this connection can be made via either the concepts of belief or of error, then we will be only too obliging to direct our attention there. As Davidson claimed [Dav75/84:169]: ‘…the concepts of objective truth, and of error, necessarily emerge in the context of interpretation’.

III. LANGUAGE

Allowing that we have established the entailment relation between possession of the higher-order concept of belief and the higher-order concept of error, the second premise of The Argument from Error, then, states that it is only in the presence of language that the concept of error is possible. Again, as Davidson puts it [Dav75/84:170 (italics added)]:

Someone cannot have a belief unless he understands the possibility of being mistaken, and this requires grasping the contrast between truth and error – true

75

belief and false belief. But this contrast … can emerge only in the context of interpretation, which alone forces us to the idea of an objective, public truth.

The reciprocal claim – i.e. that language entails the concepts of error, belief, and objectivity – is also made by Davidson [Dav75/84:169-70]:

The distinction between a sentence being held true and being in fact true is essential to the existence of an interpersonal system of communication, and when in individual cases there is a difference, it must be counted as error. Since the attitude of holding true is the same, whether the sentence is true or not, it corresponds directly to belief. The concept of belief thus stands ready to take up the slack between objective truth and the held true, and we come to understand it just in this connection.

Davidson has his own not uncontroversial background reasons in support of this argument. However, the general point is sufficiently well accepted to not require our investigating Davidson’s own rationale in any detail here. Instead, we may defer to the authority of Lepore & Ludwig, who readily endorse the general approach [Lep05:401]:

It seems clear that communication involves not merely the concept of belief, but of meaning, intention, action, desire, language, communication, and a rich pattern of other psychological and linguistic concepts. The particular application Davidson emphasizes arises from the usefulness of the contrast between belief and truth, which allows the interpreter to sometimes count a speaker wrong, in order to provide overall a smoother interpretation of his words, and to make better sense of him as a rational being.

For good measure, one may also look to Claudine Verheggen, who, though typically occupying a contrary position to that of Lepore & Ludwig, nonetheless, is in general agreement with them on this matter [Ver97:365]:

76

… possession of a language requires possession of the concept of objectivity, for the simple reason that one could not have a language unless one could distinguish between correct and incorrect applications of words. […] I take it that there is nothing over-intellectualist in the claim that one can have a language only if one has the concept of objectivity.

But while the reciprocal claim that language engenders the concepts of error, belief and objectivity, may be a relatively uncontroversial one, this is not the entailment demanded by the second premise of The Argument from Error. Rather, for the argument to succeed it needs to be shown that only language can support these concepts. This demand for strict necessity arises again from the exclusionary nature of Davidson’s ‘Main’ Argument. That is to say, he intends to establish the global lingualist thesis that without language there can be no beliefs. But, even putting aside the fact that in light of the critical failing of The Argument from Surprise we have found little reason to suppose that the possession of beliefs requires such demanding conditions (and indeed we have good reason to suppose it does not), our express interest here is not so ambitious. It will suffice for our present purposes to discern the role (if any, so long as it is a non-trivial one) that language may play in our possession of the concept of objectivity. It doesn’t matter so much here if the concept of objectivity might arise by means other than language. Indeed, Lepore & Ludwig propose three contexts seemingly outside of language where the concept of error (and so, by extension, of objectivity) may be taken to have ready application.

The first of these languageless contexts suggested by Lepore & Ludwig points out that memory allows us to compare our past selves with our present self such that we may thereby judge the beliefs of our past self to have been in error. They illustrate the point as follows [Lep05:402]:

Accustomed to the persistence of solid objects, one glances at a table and comes to believe there is a book lying there. Looking up moments later, one sees nothing there where one had thought there was a book. One may achieve a smaller overall adjustment of one’s theory of the world by supposing that one was mistaken the first time around, rather than supposing solid objects go out of existence without explanation.

77

The second context is one we are already familiar with from the last chapter (see Section IV) and involves the interpretation of the behaviour of languageless creatures [Lep05:402]:

Take the case of the dog barking up the wrong tree. Without the contrast between true and false beliefs, we are at a loss to make sense of the dog’s behavior, given that we believe that the cat is up the elm and not the oak tree. Taking the dog to have false beliefs restores his sanity and our sense of order.

And, finally, the third context involves purely theoretical reflection [Lep05:402]:

In thinking over a proposed solution to a problem in geometry or a proof in number theory, one may come to believe that one made a mistaken inference, or employed a false principle, because, for example, one sees that it is incompatible with something else one regards as more certain.

Looking to the first of these contexts, we may well be tempted to again emphasise a point now already well established: the recognition of a mistake need not entail the concept of a mistake. But, while this principle is adequate (and perhaps needed) to ward off misreading Lepore & Ludwig’s scenario as implying that matters might be otherwise, recognising this, nonetheless, does not adequately engage with the intended force of Lepore & Ludwig’s hypothetical scenario, for as they explain further [Lep05:402]:

…the usefulness of the contrast between true and false belief can be seen to lie in part in our ability to provide an overall smoother account of the nature of the world around us by sometimes attributing false beliefs to our past selves.

That is to say, the case of remembering an instance where one changed one’s beliefs is itself a context where ‘there would be a point to deploying the concept of error’ [Lep05:402 (italics added)], and this is all Lepore & Ludwig require to make their point – that is, only that there would be a point in such a context to deploy the higher-order concept of error. Such a minimal dialectical condition is available to Lepore & Ludwig’s analysis of Davidson’s argument because, as they more broadly contextualise it, The ‘Main’ Argument (and so,

78

The Argument from Error to boot) is to be evaluated in light of what they call Davidson’s commitment to the primacy of the third-person point of view. As they explain it [Lep05:387]:

… we are concerned principally with the question of what support can be found in Davidson’s work for what we have called his most fundamental assumption, namely, that we can properly treat psychological and linguistic concepts as theoretical concepts whose purpose is to enable us to systematize and keep track of behavior neutrally described – or, to put it another way, the assumption that the third person stance … is conceptually basic in understanding meaning and psychological attitudes.

Now, putting aside the question of whether or not it is appropriate to judge Davidson’s philosophy according to this supposed ‘fundamental assumption’ (see esp. [Sto05a&b]), the relevance that the above has for Lepore & Ludwig’s treatment of The Argument from Error is this: they thereby charge Davidson’s ‘Main’ Argument with the further burden of answering the challenge of global scepticism. This further burden is made apparent in what they call the strong interpretation of The ‘Main’ Argument. On this reading the argument needs to establish, not merely that a creature can subjectively take itself to be (or have been) in communication with what it takes to be mind-independent others, but that it has actually been so. As Lepore & Ludwig characterise the strong interpretation [Lep05:401]:

The strong interpretation holds that thinking of the creature as deploying a concept requires thinking of it as correctly deploying it in some of its activities, in the sense of applying it correctly to objects. On the strong interpretation, to deploy correctly the concept of a spatial object, for example, would require there to be spatial objects of which one thinks correctly that they are spatial objects. This is the objective application of the concept.

By contrast, then, the weak interpretation requires only that ‘we must think of ourselves in communication with others in order to have the concept of belief, truth, and objectivity’ [Lep05:403 (italics added)]. But, as Lepore & Ludwig go on to point out, it is commonly acknowledged that any transcendental argument which ‘attempts to move from observations about the conditions under which we can have certain concepts to objective features of the

79

world’ typically faces insurmountable difficulties [Lep05:402]. In particular, even if such transcendental arguments are sound, they nevertheless ‘fall short of establishing that we ever bring actual objects under our concepts…’ [Lep05:402 (italics added)].7 And so, as they conclude specifically with respect to the second premise of The Argument from Error [Lep05:403] :

Even if there were scope for the correct application of the concept only in the context of actual communication, if possession of the concept requires only thinking of oneself as being in a situation in which the application of the concept has a point, we could not conclude that we have ever actually been in communication with others from knowledge that we possess the concepts of belief, truth, and objectivity. A gap remains between thinking of ourselves as being in communication with others and actually being so.

Returning then to the question of the three contexts in which Lepore & Ludwig have suggested the concept of error may find ready application outside of language use, the significance of their framing the evaluation of Davidson’s Argument from Error in terms of his alleged fundamental assumption of the primacy of the third-person point of view means that the dialectic only minimally requires that they provide contexts where ‘there would be a point to deploying the concept of error’ [Lep05:402]. More specifically, this means that they are not beholden to provide an account of how the concept of error and objectivity arise. As Lepore & Ludwig themselves explain [Lep05:401-2]:

It is clear that, in asking what grounds the attribution of the concept to a creature, we are not asking for a story about how the creature could acquire the concept. It would be beside the point to ask for a brute causal story. And it could not be a request for a story about how a creature with beliefs could acquire the concept of error by engaging in certain sorts of activities. For the aim of the argument is to show that without the concept of error, and so the concept of belief, one cannot have beliefs at all.

7 See [Mal02] for a more sympathetic reading of both transcendental arguments in general, and Davidson’s in particular.

80

So, on this reading, the three contexts only need to amount to ‘a description of an activity which gives point to the application of the concept of error’ [Lep05:402]. Or again [Lep05:401]: ‘What is required is simply that we describe some activity an agent could engage in which would involve making mistakes, but not involve essentially communication’. And, it must be admitted, all three contexts succeed in this vein such that ‘it seems dubious that communication is the only context in which there is scope for the application of the concept of error’ [Lep05:402].

That being said and duly noted, herein, we are not concerned to evaluate the primacy of the third-person point of view or Davidson’s fundamental commitment to it, or any potential anti-sceptical arguments that may be invoked in support of either. Rather, insofar as we are concerned only with providing a non-trivial account of the role that language plays in possession of the concept of objectivity, it sufficies for our purposes to assume from the outset a metaphysically lightweight, ‘small-n’ naturalistic position not concerned to address the ostensible challenge of global scepticism. Again, as Davidson characterises this form of naturalism [Dav95:05]:

… we have a roughly correct view of our surroundings and of the existence of other people with minds of their own. I do not question that we are, often enough, justified in these beliefs: we know there are mountains and seas, fish and serpents, stars and universities. […] This attitude and method have sometimes been called naturalism because naturalism starts by accepting common sense (or science) and then goes on to ask for a description of the nature and origins of such knowledge.

In this light, the dialectical force of Lepore & Ludwig’s proposed contexts is significantly undercut with respect to the second premise of Davidson’s Argument from Error. Specifically, all three of Lepore & Ludwig’s contexts now stand in danger of violating Morgan’s Canon. This is to say, that if we are now to take the scenarios proposed by Lepore & Ludwig as actually requiring the application of the concept of error, Morgan’s Canon could be invoked to shift the dialectical burden back on Lepore & Ludwig to explain how their proposed contexts give rise to (or at the very least, necessitate) the concept of error. On the face of it, none of the contexts proposed necessitate that the concept of error be applied in order to explain the presupposed, lower-order recognition that mistakes had been made. Moreover, while it is plausibly the case that the Lepore & Ludwig’s languageless

81

contexts may find readily application for the concept of error, it may also be the case that such application is possible only if the subject in question already is a language user or possesses the concept of objectivity. The third context of someone engaged in geometry or proof theory seems particularly likely in this regard.

Once again, to be perfectly clear, no such difficulty faces Lepore & Ludwig’s proposed contexts when evaluating Davidson’s fundamental assumption regarding the primacy of the third-person point of view. For their purposes it suffices to merely indicate contexts where the concept of error would have application outside of language use. Once a commitment to a more liberal naturalism is presumed, however, the dialectical burden is inversely increased. Indeed, it seems reasonable to suppose that unless one holds oneself (or one’s dialectical interlocutor) beholden to the level of necessity required to potentially overthrow the challenge of global scepticism, then anything less than a non-trivial account of how a creature may come to deploy a concept must violate Morgan’s Canon.

Of course, this is not to suggest that the concept of error (and so, objectivity) has in any way been shown to arise only in the context of language. Rather, it is, I believe, to motivate a move, on the one hand, away from a priori necessity claims, and, on the other, towards the kind of non-trivial, empirically informed account that I hope to advance over the next two parts of this thesis.

Admittedly, there may be some who feel that, regardless of Lepore & Ludwig’s intended dialectical target, insofar as they have provided contexts outside of language where it is at all possible the concept of error may find ready application, they have thereby still shown at least that it is ‘dubious that communication is the only context in which there is scope for the application of the concept of error’ [Lep05:402]. While I have some sympathy with this doubt (and putting aside that it is incidental to the primary focus of this thesis), in the following section I offer three contexts of my own which suggest that at least some particular applications of the concept of objectivity may depend solely on language. I do not, however, intend to settle the debate in this way. Rather, my intention is only to further motivate a move away from such tit-for-tat a priori arguments, and instead champion the need for a more substantive account of the role that language might play in our possession of the concept of objectivity – for it is this very task that the following two parts of this thesis will begin to undertake.

82

IV. SUFFICIENCY

In this section, I introduce a number of novel examples of my own where the concept of objectivity is required to understand the scenario in question. Although I do not wish to claim that these in any way demonstrate that language is necessary for the concept of objectivity, I do wish to suggest that it at least appears to be true of these particular cases and presumably more like them. My motivation for airing these novel contexts, then, is to further counterbalance the dialectical weight from Lepore & Ludwig’s proposed contexts, which were intended to suggest that the concept is possible without language. Obviously, the aim here is not to settle the debate one way or another. It is, rather, to provide greater motivation for the task set for the remainder of this thesis: that is, to develop a non-trivial account of the role of language in our possession of the concept of objectivity.

i) The first example I would like to introduce is reappropriated from Davidson’s own Swampman thought-experiment. He writes [Dav87/01:19]:

Suppose lightning strikes a dead tree in a swamp; I am standing nearby. My body is reduced to its elements, while entirely by coincidence (and out of different molecules) the tree is turned into my physical replica. My replica, Swampman, moves exactly as I did; according to its nature it departs the swamp, encounters and seems to recognize my friends, and appears to return their greetings in English. It moves into my house and seems to write articles on radical interpretation. No one can tell the difference. But there is a difference.

In its original deployment, Davidson hopes this scenario will show that the meaning of our words depends on their being causally enmeshed by their past usage and acquisition. The supporting intuition that Davidson hopes people share in this regard, then, is that Swampman’s utterances – although identical to meaningful English – must be, in fact, meaningless sounds insofar as Swampman has no causal past. As Davidson puts it [Dav87/01:19]: ‘It can’t mean what I do by the word ‘house’, for example, since the sound

83

‘house’ Swampman makes was not learned in a context that would give it the right meaning – or any meaning at all.’

For our current purposes, however, this desired intuition is beside the point. Rather, Davidson’s Swampman thought-experiment is here invoked only as more sophisticated version of the children’s tale I told in the Preface about the surreptitiously replaced goldfish. There, the girl who had been entrusted to care for her neighbours’ fish presumably obtained the surrogate in the usual way from a pet store. By contrast, in the case of Swampman, a veritable miracle of ‘coincidence’ is invoked in place of any explanation of how Swampman could have possibly been brought about by lightning randomly striking a tree. The point for us is that, insofar as the story is deprived of any and all credible or informative explanatory elements, our attention is singularly drawn to the central point, which is that Swampman is not Davidson. As with the children’s story, this application of the concept of objectivity is readily and strikingly conveyed in language. And, I want to suggest, it is perhaps impossible for it to be otherwise conveyed. For instance, putting aside how one could ever know that both Davidson’s annihilation and Swampman’s creation had concurrently happened ‘by coincidence’ (as the thought-experiment stipulates), it appears that, assuming that one did somehow know it, there could be no other way except via language to convey to anyone else that the two events were not causally related – that is, that Swampman is objectively not Davidson. This is because the stipulation that the event happened entire ‘by coincidence’ and so, without any causal trace, there is literally nothing one could point to in the world to indicate that it was so. This peculiarity makes it distinct from the case of the surrogate fish, where presumably evidence could be mustered, such as the corpse of the original fish, or a video recording showing the surrogate fish being brought home from the pet store. Perhaps one could similarly convey the import of Davidson’s Swampman thought-experiment without linguistic language per se – say, by animation? – but, I posit that, if such a ‘communication’ were successful, then, whatever means by which one had managed to convey the difference must therefore qualify as language anyway.8

8 I appreciate the controversy surrounding the distinction between communication, which in the biological sciences counts as any physical or behavioral characteristics that influence the behaviour of others (e.g. a peacock’s plumage), and language which in philosophy and linguistics is commonly taken to have something to do with infinite recombinabilty. I am not here proposing necessary conditions for language, but rather pumping our intuitions that a sufficient condition might be application of the concept of objectivity.

84

ii) The second example I would like to introduce runs as follows: imagine a text-world9 in which the indubitable hero is denied and persecuted by all who inhabit the text-world, to the point that not only the hero, but the text-world itself seemingly denies their worth and, so, status as the hero.10&11 Here, the concept of objectivity readily finds application insofar as it is immediately apparent to the audience of the text-world that the hero’s worth is indubitable despite the mistaken opinion of the entire text-world, including all its inhabitants. It should also be clear that it is solely by way of language that such a text-world can be imagined in which the reading audience can appreciate objective aspects that nonetheless may remain entirely obscured to the text-world itself.

iii) The third and final example invites us to consider a Matrix-style scenario, except involving a languageless creature, such as a dog. Imagine the dog, having unwittingly spent the majority of its life in a virtual world chasing virtual cats and going for virtual walks around the virtual block, etc., but now is suddenly awakened into the non-virtual world. The question then, is if there is any way one might convey the significance of this turn of events to the dog that didn’t involve language? Consider, for example, attempting to show the dog that its previous virtual world was in fact virtual by way of a ‘Reality-Definition’ (as opposed to HD ‘High Definition’) screen. The dog would merely take this to be an impassable tunnel into aspects of what it would persist in taking to be the real world. Whatever languageless explanation one might try give about how the illusion had been created still seems to depend necessarily on the further, fundamental application of the concept of objectivity in understanding that the virtual world that one had taken for the real world was not in fact the real world. Indeed, as was the case with our appropriation of Davidson’s Swampman thought-experiment, if one could somehow concoct a way of conveying what had happened to the dog, such a method would itself constitute language and would thereby be seen to have brought the dog into the circle of language-users.

9 A ‘text-world’ is the universe described by a particular text, like, for instance, ‘Middle-earth’ of J. R. R. Tolkein’s Lord of the Rings (see [Tol68/91]). 10 This particular context was inspired by certain events of the Christian gospels, though there the connection ends. 11 The significance of the text-world itself seemingly denying the legitimacy of the hero is required as many text-worlds see otherwise normally causally indifferent phenomena (such as storms or luck) respond teleologically to the hero’s predicament. In the scenario outlined here, the text-world shows itself to be indifferent to the hero’s tragic plight. As the character of Jesus puts it (See Psalm22:2): ‘Eli, Eli, lama sabachthani’ (‘My Lord, why hast Thou forsaken me?’)

85

To emphasise the intended import of introducing these examples once more: I am not suggesting that they in any way settle the matter in favour of language being necessary for possession of the concept of objectivity. I am happy to concede that, even if one can show that there are certain applications of the concept of objectivity that at least appear to be wholly language dependent, one will not have thereby shown that all applications must be. Rather, the point is only to provide a counterbalance to the kind of a priori contexts offered by the likes of Lepore & Ludwig, as were introduced in the previous section. That is to say, the above is only meant to show that there appear to be cases which inherently and directly require the application of a particularly high-order concept of objectivity in order to understand them, which seem to be only possible in the context of language.

That said, one may further worry that my previous avowal of a commitment to a metaphysically lightweight, small-n naturalism is at odds with the invocation of such bizarre thought-experiments as Swampman. As Daniel Dennett objects in regard to Davidson’s original deployment of this thought-experiment [Den13:111]:

It is just as clearly physically impossible for the ‘traces’ of, say, Davidson’s memories to appear in the structure of Swampman’s brain as it is for a shark to form itself of cells containing cow DNA. Swampman may not be logically impossible, if only because cosmic coincidences of the sort imagined to produce things like Swampman are by definition not logically impossible, but they never happen, so who cares what we would say if they did? […] a general rule of thumb emerges from the comparison: the utility of a thought-experiment is inversely proportional to the size of its departures from reality.

In response, two things may be said of my reappropriation of the Swampman thought-experiment in particular (though similar considerations, I will suggest, apply to all three contexts). The first is that I am happy to acknowledge that is only prima facie that the Swampman story is in any way coherent. In my reappropriation, the point of Davidson’s ‘entirely by coincidence’ placeholder is only to draw our attention to the remarkable capacity of language to directly apply the concept of objectivity, as, for example, one might equally accomplish by straightforwardly saying, ‘Swampman is not Davidson’ (or, as in the case of surrogate goldfish: ‘This fish is not my fish.’). Any further consideration of how the scenario described by the thought-experiment is supposed to ‘work’ will invariably lead to

86

confusion and justified incredulity. As Dennett further stresses when talking about other similarly bizarre thought-experiments [Den13:111]:

The physicists would reply that if they were confronted with [such imaginary cases], they would have much more important things to worry about than what to call them. Their whole scientific picture depends on there being a deep regularity […] and the ‘fact’ that it is logically possible to break this regularity is of vanishing interest to them. […] If they find violations of the regularities, they adjust their science accordingly, letting the terms fall where they may.

Here, we are in full agreement with Dennett (as is Davidson, who also “regrets [his] sorties into science fiction” (see [Dav99b:192])).That said, I would again emphasise that the suggested examples need only be prima facie coherent for them to succeed in suggesting that at least certain applications of the concept of objectivity appear to be language dependent.

The second thing to say in response to Dennett’s proposed ‘general rule’ is that, although the content of my proposed examples is admittedly non-naturalistic, people do indeed say and imagine the kind of things described by my examples (regardless of whether they are coherent or not). Given that some people (including Davidson) presume that they may at least prima facie understand what is intended to be conveyed by such thought-experiments, I take this as sufficient grounds for assuming that the novel contexts I have proposed in this section are adequate to demonstrate that language is readily able to convey certain applications of the concept of objectivity where perhaps no other method possibly could.

As was said above, my intention in introducing these contexts is by no means to suggest that they provide adequate reason for supposing that language is necessary for the concept of objectivity. At most, these scenarios suggest that on the face of it, it looks as though certain applications of the concept of objectivity cannot be conveyed to other people (or dogs, as the case may be) outside of language.

87

V. CONCLUSION

Davidson’s Argument from Error sought to establish that language is necessary for not only the concept of error, but for all and any beliefs whatsoever. Given that The Argument from Surprise critically fails, The Argument from Error needed to be significantly reduced in scope from the outset. Instead, we sought only to consider whether Davidson’s argument could establish that language is necessary for the higher-order concept of error which stipulated possession of the concept of belief as well as the lower-order concept of error (i.e. the ability to recognise mistakes). Against even such a constrained reformulation of Davidson’s argument, while it was readily conceded that language sufficies for the concept of error, the stronger claim that only language may suffice was called into question by Lepore & Ludwig who offer three contexts outside of language use where the concept of error could nonetheless find ready application. In light of their broader dialectical context, where they seek to evaluate Davidson’s alleged commitment to the primacy of the third-person perspective, I argued that if we presume a commitment to a metaphysically deflated small-n naturalism, then the dialectical weight can be seen to shift away from merely providing possible contexts where the concept of error may have ready application, to instead paying greater observance to Morgan’s Canon in providing substantive, non-trivial accounts of the role played by language in our possession of concepts such as that of error and objectivity. As a result, the force of Lepore & Ludwig’s otherwise legitimate contexts was significantly undercut. Instead, the burden of proof rests with any who might seek to investigate the relation of phenomena such as language and the concept of error or objectivity to provide a non-trivial, empirically consistent, substantive account. Indeed, it is precisely such an undertaking that we shall endeavour to develop over the next two parts of this thesis. That being noted, the chapter was brought to a close by the introduction of three novel contexts which aimed to suggest prima facie that at least certain applications of the concept of objectivity might not be conveyed outside of language. These contexts were not intended to settle the debate in favour of Davidson’s claim that only language sufficies for possession of the concept of error and so, of objectivity. Rather, by their introduction, I sought only to reinforce the significance of the point made in the preceding chapter: if we hope to move the debate forward on what capacities are engendered by or unique to our being language users, then, while the various arguments explored in this chapter are perhaps useful in pumping our intuitions, nonetheless, what is needed is a substantive, non-trivial, empirically consistent account. Again, as was said above, this will be the principal focus of the following two parts of this thesis.

88

PART TWO

* * *

Word &World

Figure1.17thCenturyEtchingofTrigonometricalTriangulation[Mid11]

Parbir pictured a right-angled triangle, with vertices at the centre of the Earth, a point on the horizon, and his own eyes. He’d plotted the distance function on his notepad, and knew many points on the curve by heart. The beach sloped steeply, so his eyes were probably two full metres above sea level. That meant he could see for five kilometres…

Teranesia, Greg Egan

89

In Part One we saw that Davidson’s ‘Main’ Argument is unable to establish its principal thesis that to have beliefs, one must be a language user. At best, it can only support the weaker claim that if one has language then one possess the concept of objectivity. Unfortunately for Davidson, in order for the ‘Main’ Argument to succeed, it also needs to demonstrate the converse – that, in order to have the concept of objectivity, one must have language. Davidson himself openly acknowledges this latter necessity [Dav82/01:105]:

To complete the ‘argument’, however, I need to show that the only way one could come to have the belief–truth contrast is through having the concept of intersubjective truth. I confess I do not know how to show this. But neither do I have any idea how else one could arrive at the concept of an objective truth.

In the absence of any forthcoming argument, Davidson instead proposes that the process of geometrical triangulation might well serve as an analogy to better highlight the necessary conditions for both the concept of objectivity and the capacity for objective truth to arise.

In its normal context, geometrical triangulation allows one to calculate unknown distances on the basis of ones known. In this regard, it is extremely useful for determining the distance to locations otherwise unmeasurable. The process typically involves calculating the distance, d, by first forming a triangle between it and a baseline of a known distance, l. One then measures the angles, α and β, formed at the interstices of the baseline. These known values taken together permit one to calculate d by the following trigonometric equation: d = l ÷ ((1 ÷ (tan × α)) + (1 ÷ (tan × β))).12

As we can see in the following passage where the notion of triangulation is first introduced, Davidson is not so much interested in the mathematics of the process, as he is in the general underlying principle [Dav82/01:105]:

If I were bolted to the earth, I would have no way of determining the distance from me of many objects. I would only know they were on some line drawn from me towards them. I might interact successfully with objects, but I could have no

12 N.B. Nothing hangs on the reader’s understanding of this.

90

way of giving content to the question where they were. Not being bolted down, I am free to triangulate.

Our sense of objectivity is the consequence of another sort of triangulation, one that requires two creatures. Each interacts with an object, but what gives each the concept of the way things are objectively is the base line formed between the creatures by language. The fact that they share a concept of truth alone makes sense of the claim that they have beliefs, that they are able to assign objects a place in the public world.

This is a somewhat opaque passage that, nonetheless, seems to promise so much. Of particular interest to us here, is the emphasis that it places on explaining our sense of objectivity. Davidson here appears to be saying that the concept of objectivity depends on our engaging in language-based communication with others. Yet, as Ingvald Fergestad and Bjørn Ramberg point out [Fer11:221]:

Davidson’s triangulation figure is notoriously difficult to explicate. At issue is not only what the triangulation thesis is and how purported arguments for it stand up to closer scrutiny. There is disagreement, also, about Davidson’s purpose, about what his philosophical aims are, and what position the triangulation figure is supposed to occupy in Davidson’s pursuit of a coherent grasp of our ability, as natural creatures, to think and to communicate.

The aim of the following isn’t to resolve these difficulties. I point them out only to acknowledge the variety of competing interpretations surrounding Davidson’s triangulation arguments. Rather, in the following two chapters of this second part, I shall consider two widely recognised aspects of Davidson’s thinking. These are: 1) The Argument from Content Determination and 2) The Argument from Objectivity. In neither case shall we find Davidson’s recourse to the notion of triangulation satisfactory for reaching the conclusions Davidson hopes to establish. However, in considering the two arguments, we will lay the groundwork for Part Three of this thesis, so that we may then work to move the debate forward.

91

CHAPTER THREE

* * *

Content

I. INTRODUCTION

Davidson’s Argument from Content Determination aims to demonstrate that triangulation is necessary for belief. More precisely, Davidson argues that the solution to certain well known problems of causal theories of mental content minimally requires a triangular base of at least two creatures in mutual interaction with one another regarding their relevantly similar patterns of responses to a shared, mind-independent environment. At this stage, the nature of the mutual interaction need not be linguistic. As Davidson himself outlines the core of his argument [Dav91/01:212–3]:

...until the triangle is completed connecting two creatures, and each creature with common features of the world, there can be no answer to the question whether a creature, in discriminating between stimuli, is discriminating between stimuli at the sensory surfaces or somewhere further out, or further in. Without this sharing of reactions to common stimuli, thought and speech would have no particular content – that is, no content at all. It takes two points of view to give a location to

92

the cause of a thought, and thus to define its content. We may think of it as a form of triangulation.

In this chapter, one particular source of content indeterminacy – the problem of distance – will be introduced and I will explain the difference in Davidson’s overall approach to the problem from that of standard approaches. The Argument from Content Determination proper, then, will be explicated in its two initial phases: the pre-representational and the pre-linguistic. The question of full-blown semantic triangulation will be held over to Chapter Four (and beyond).

The first phase is predicated on the idea that if a reaction of a creature cannot be said to have some particular cause, then neither can the creature be said to have propositional beliefs of determinate content. In this vein, Davidson argues that the cause of a creature’s reactions cannot be determined either by the creature itself (taken in isolation), or by any external observation of solitary creatures. Rather, Davidson claims that it is only when two creatures are observed interacting with a shared environment (though perhaps not with one another) can the cause of their reactions be determined.

The second pre-linguistic phase then is to shift the focus onto the creatures themselves. Davidson argues that, although they may not necessarily interact with other, if a creature can come to correlate its conspecific’s relevantly similar responses to objects and events in a shared environment, it can thereby home in on the particular causes of its beliefs. Furthermore, he takes it that allows for the possibility for error to arise such that it might then overcome the problem of distance.

Davidson’s Argument from Content Determination will be shown to critically falter in much the same way as did his ‘Main’ Argument. In fact, the general diagnosis will be the same – that is, in order to possess beliefs it is not necessary that one possess the meta-cognitive awareness that one has beliefs as such. The particular criticism here, however, will be more illuminating in that recent empirical studies strongly suggest that many creatures do have perceptual and cognitive systems that can be known to objectively represent certain aspects of their environment. Such capacities then may be seen to form the true origins of objectivity, as Tyler Burge has argued at length (See [Bur10]). This is not to say that non-linguistic triangulation between creatures does not allow them to more precisely home in on

93

the causes of their belief; only that, as it stands, Davidson’s argument fails to show that some particular degree of precision is required such that one might therefore claim that triangulation is necessary in meeting such standards.

II. INDETERMINACY

Beliefs are propositional attitudes such that in order to have truth conditions, they must have determinate content. That is to say, in order for the propositional content of beliefs to be either true or false, beliefs must be about something. If a mental state is too vague to have truth conditions, then it can’t properly be counted as a belief. There are, however, numerous commonly recognised sources of ambiguity regarding mental content, and Davidson himself clearly picks out two such difficulties in the following passage [Dav97/01:129–30]:

The first ambiguity concerns how much of the total cause of a belief is relevant to content. [...] The second problem has to do with the ambiguity of the relevant stimulus, whether it is proximal (at the skin, say) or distal.

Although Davidson claims that his theory of triangulation will resolve both of these problems in one fell swoop (and perhaps others to boot), because he focuses almost exclusively on the latter problem, the discussion herein will be similarly constrained. Furthermore, given the overall goals of this thesis, it will not be considered whether or not triangulation might resolve any of the other problems of mental content. Such an undertaking is left for future research.

94

1. The Problem of Distance

The problem of distance concerns our capacity to pick out the salient point (or points) of the content of our beliefs from along a causal chain.

It is taken for granted that there are numerous causal chains (or more accurately, causal trees) connecting us to any given event in our environment. For instance, Lepore & Ludwig illustrate this point with an example of watching events on television [Lep05:409–10]:

There is the surface of the television screen, events in the stream of electrons being projected onto it, events in the cable, at the cable television company’s headquarters, in a satellite in geosynchronous orbit around the earth, at CNN in Atlanta, and at distant trouble spots around the globe, and so on.

In this way we better appreciate that for any given event or object to be epistemically accessible to us, there is a multitude of physical causes in the causal chains that link us. However, when we speak, for instance, of some event that we saw on the news last night, we commonly suppose that we are able to pick out the salient cause of our belief – which we take to be the event that the news was reporting on, and not the various means through which the news of the event reached us. Loosely speaking, explaining how we are generally able to accomplish such a feat constitutes the problem of distance. Davidson illustrates the problem in the context of ascribing beliefs to a dog that salivates at the sound of a bell. Of this scenario, Davidson asks [Dav92/01:118]:

… why say the stimulus is the ringing of the bell? Why not the motion of the air close to the ears of the dog – or even the stimulation of its nerve endings? Certainly if the air were made to vibrate in just the way the bell makes it vibrate it would make no difference to the behaviour of the dog. And if the right nerve endings were activated in the right way, there would be no difference.

Admittedly, one naturally presumes that the dog responds to the ringing bell itself and not merely the sound waves travelling from the bell to the dog’s ears, let alone its own nerve

95

endings. But this is the default presumption only insofar as we take ourselves to be readily able to pick out the bell as the object of our own beliefs, such that we then charitably extend the same common sense to the dog also. As Davidson puts it [Dav92/01:118]: ‘It seems natural to us [to say that the dog is responding to the bell] because it is natural – to us.’ However, the question may then be asked as to how it is that we ourselves are able to achieve this. That is, how are we able to determine the causal point relevant to the object of our belief from along the causal chain extending from (and even past) the object to our senses? Davidson illustrates the problem with regard to our own higher-order capacities [Dav99a:34]:

What is it that tells us that the stimulus (cause) of Alex’s ‘answer’ to the question ‘What’s the same?’ isn’t the activation of certain rods and cones in his eyes, or the firing of certain optic nerves, or the photons bouncing off surfaces we see as the same color? All of these causes, and endless more, are common to the cases where Alex emitted the sound, ‘Color’. We have no grounds for choosing one of these causes over the others.

To further illustrate the difficulty, let us briefly consider an analogous case of light travelling to Earth from a distant star. Before one learns that light has a fixed and finite speed, it is natural to presume that the stars we see in the night sky are now exactly as we observe them to be. But as we all learn soon enough, each star’s light is as many years old as the distance in light-years of the star from the Earth. Indeed, even the closest stars of Alpha Centauri are still over four light-years away. This means that from our vantage point on Earth we can only ever see these stars no less than they were four years ago (relatively speaking). But given this arrangement, it is natural to wonder how is it that we can look to the sky and suppose that we are able to pick out, from along the years-long trajectory of the light, the point from which the light was emitted from the star itself. The answer is, obviously, that we can’t – not at least without a theoretically and technologically sophisticated science of stellar parallax. With respect to the stars, when aided by the naked eye alone, we have no way of picking out where and when along the casual chain the supposed objects of our perceptions (or beliefs) actually lie. In this light, we may have more sympathy with the ancient conviction that the superlunary bodies were affixed to the equidistant celestial sphere.

96

If we bring the example back down to Earth, we can see that an analogous problem might equally be taken to hold with respect to the objects of our everyday local environments. For instance, when I believe myself to be reading a book, it may be asked how I am capable of accurately ascertaining where the book lies along the albeit comparatively short causal chain extending from the book itself to my central nervous system. That is to ask: am I really looking at (or thinking about) the book itself? Or perhaps what I take to be the book itself is merely the photon-packets of light travelling somewhere along the causal chain in between the book and my senses. Or am I only attending to the stimulation of my nerve endings? Indeed, uncanny phenomena such as the report of thunder arriving moments after a flash of lightening readily suggest to us how blindly trusting we are in our ability to perceive the objects of our environments as they are in themselves. To be clear: the issue is not that the two sensory reports can come apart. Rather, it’s that when they do, it draws our attention to the blind trust we place in supposing either the sound or the vision to justifiably determine the object of our belief. Admittedly, numerous plausible answers suggest themselves as to how we do overcome such limitations. However, we shall postpone systematic consideration of these answers-in-waiting for the sections to follow.

2. Varieties of Externalisms

Before moving on to examine Davidson’s Argument from Content Determination, let’s pause a moment to differentiate the more common approaches to the problem of distance from the precise nature of the task Davidson set himself to undertake. To do so is not necessarily to endorse Davidson’s rather idiosyncratic approach – such evaluations will come out in the wash, as it were. The hope rather, is merely to eschew any unnecessary confusion so that Davidson’s arguments may subsequently be judged in their own right.

* * *

Standard accounts of causal theories of mental content confront essentially the same problem of distance that Davidson concerns himself with. However, attention to detail reveals each is made with different emphasis. To illustrate this point, consider Adams & Aizawa’s entry on ‘Causal Theories of Mental Content’ in The Stanford Encyclopedia of

97

Philosophy. They explain that any satisfactory causal account of mental content must be able to answer the following problem (among others) [Ada10]:

In seeing a dog, there is a causal pathway between the dog through the visual system (and perhaps beyond) to a token of ‘X’. What in this causal pathway from the dog to ‘X’ constitutes the content-determining element? In virtue of what is it the case that ‘X’ means dog, rather than retinal projection of a dog, or any number of other possible points along the pathway? Clearly there is a similar problem for other sense modalities. In hearing a dog, there is a causal pathway between the dog through the auditory system (and perhaps beyond) to a token of ‘X’. What makes ‘X’ mean dog, rather than sound of a dog (barking?) or eardrum vibration or motion in the stapes bone of the inner ear? One might press essentially the same point by asking what makes ‘X’ mean dog, rather than some complex function of all the diverse causal intermediaries between dogs and ‘X’.

Although they do not name it as ‘the problem of distance’, it should be clear that Adams and Aizawa hereby identify the same source of ambiguity in the determination of mental content as Davidson. The difference, however, is that Adams and Aizawa take the standard line in asking, ‘What in this causal pathway from the dog to “X” constitutes the content-determining element?’ [Ada10]. The emphasis here on accounting for whatever feature it is that allows us to pick out the right cause is characteristic of the standard approach to the problem of distance. That is, they first suppose that we are able to represent objects in the environment, and then seek to account for whatever it is in the causal story that enables us to in this way pick out the correct object. For instance, on this line one may ask, ‘What is it about my perception that makes my belief that I am looking at a dog be about the dog that I see and not, for instance, merely the light reflected off an object?’ Jason Bridges nicely characterises this common approach toward the question of content determination [Bri06:291]:

[Davidson’s approach] differs importantly from the well-known theses in the philosophy of mind and language […]. Those theses are attempts to elucidate specific types of mental content by giving criteria for individuating contents of those types. They are motivated by considerations specific to the types of content in question…

98

Davidson’s idiosyncratic approach, by contrast, is to enquire how we are capable of picking out any point along the casual chain whatsoever. The question of content is beside the point. That is to say, Davidson is concerned with explaining our very capacity to pick out the cause of our beliefs and perceptions from along the causal chain at all [Dav01:7]: ‘The point is not to identify the norm, but to make sense of there being a norm...’. As quoted before, Davidson says in endorsement of his theory of triangulation [Dav91/01:212]:

... until the triangle is completed connecting two creatures, and each creature with common features of the world, there can be no answer to the question whether a creature, in discriminating between stimuli, is discriminating between stimuli at the sensory surfaces or somewhere further out, or further in. Without this sharing of reactions to common stimuli, thought and speech would have no particular content – that is, no content at all.

Davidson’s approach begins sceptically, without the presumption that we are able to pick out any points along the causal chain at all. By contrast, the common approach supposes that we are able to do so, but asks of whatever it is that we do pick out, what it is about that content in particular that we individuate. With this distinction in mind then, let us turn now to consider Davidson’s argument in its own right.

III. PRE-REPRESENTATIONAL TRIANGULATION

Davidson’s Argument from Content Determination is advanced as proceeding in two phases. The second phase deals with the question of the content of propositional beliefs with respect to pre-linguistic triangulation. In the first phase, however, the focus is on whether a creature’s reactions or responses to its environment can be said to have any particular salient cause. Davidson’s claim is that it is only when at least two creatures are observed interacting with a shared environment (though not necessarily with one another) that either creatures’ reactions may be said to have some particular cause. The creature in isolation cannot determine for itself the cause of its reactions. Nor can the observation of a solitary creature.

99

Only in the context of a minimally triangular relation between the observer and at least two creatures interacting with their environment (though, again, not necessarily with one another) may the question of the cause of their reactions be resolved. And so, outside of the context of triangulation then, Davidson suggests that the creature cannot be said to even perceptually (let alone propositionally) represent aspects of its environment objectively.

The nature of this first phase of the argument is primarily epistemological. That is, the emphasis is on what minimal conditions must be met in order to determine (in the sense of identifying) the cause of a creature’s reactions. Unlike in the later phase, it is not a constitutive claim. Davidson isn’t suggesting that the creature does not react to the causal connections it has with its environment independent of their identification; only that, outside of a minimally triangular relationship, no one particular cause can be said to be the cause of the creature’s responses. Here, we are concerned only with the explanatory model of accounting for the creature’s reactions to events in its environment, not the representational model of what the creature itself is capable of perceiving or believing.

Davidson’s purpose in considering these pre-representational matters is at least twofold. On the one hand, if it can be demonstrated in this first phase that the cause of a creature’s responses cannot be identified outside of a minimally triangular relationship between an observer and at least two creatures, then insofar as it is held that the creature itself must be meta-cognitively aware that its beliefs are beliefs (i.e. that they may be either true or false, as Davidson maintains), pre-triangular contexts thus appear unserviceable in resolving the problem of distance for the content of belief. As Davidson sets it up [Dav01:12]: ‘If thought were present, [identification of the cause of a creature’s reaction to its environment] would provide an obvious indication of the content of the thought’. But, what cannot be said of non-representational reactions cannot then serve as any basis for claims regarding propositional beliefs.

On the other hand, although the minimally triangular relationship claimed to be necessary for determining the causes of a creature’s reactions is not the same as the full-blown (pre-linguistic or semantic) triangulation that Davidson later argues is in fact required for beliefs of determinate content, the principles thought to be established in the first phase regarding pre-representational causes are intended to be carried over to the later phases of the argument also. In this way, Davidson seeks to induct his audience with a prior appreciation of the notion underlying his principal argument for when it is finally revealed. However, the

100

negative counterpart to this hope is, as we shall see, that any failings found in the first phase will have to be redressed in the later phases as well.

* * *

This first phase of the Argument from Content Determination is advanced in stages. We begin with a consideration of the solitary creature – first unobserved, and then observed. We then turn to consider the observation of two creatures interacting with a shared environment, but not one another. The aim at each stage will be to explicate Davidson’s assessment of the relevant stage and to evaluate it accordingly. Some further preliminary conclusions will then be drawn before moving on to consider the case for pre-linguistic triangulation itself.

For a reader with a grasp of Davidson’s overall position, this strategy may seem somewhat uncharitable in the sense that many of the critiqued passages were probably written to be read in light of the argument’s conclusion having been established. Although this may be the case, the strategy adopted herein is worthwhile insofar as it promises to illuminate where exactly the crux of Davidson’s argument lies.

1. The Solitary Creature

The circumstance furthest removed, yet while still remaining relevant to the intersubjective aspect of triangulation, is that of the purely solitary creature both unobserved and in complete isolation from any interaction with another creature. Davidson takes such a one to be incapable of picking out the salient cause of its own reactions from the multifarious causes that impinge upon it. As he puts it [Dav92/01:119]:

If we consider a single creature by itself, its responses, no matter how complex, cannot show that it is reacting to, or thinking about, events a certain distance away rather than, say, on its skin. The solipsist’s world can be any size; which is to say, from the solipsist’s point of view it has no size, it is not a world.

101

While the phrasing, ‘if we consider...’, may initially seem to imply that we are to imagine ourselves actually observing the creature, the latter portion of the quotation makes it clear that our consideration is intended to be of the creature alone, vainly trying to determine for itself the causes of its reactions (let alone thoughts).13 Obviously, Davidson doesn’t hold out any hope for the creature’s efforts.

There are a few things worth saying in criticism of this passage.

i) Even if we allow that from the perspective of the creature itself, its ‘world ... has no size [and so] is not a world’, this does not entail that we, as thought-experimenters, must similarly abandon our own ability to grasp ‘the size of the world’ that includes the creature in question. That is to say, despite the creature’s own epistemic limitations, it seems reasonable to presume that there remains a fact of the matter as to which objective features of the world it is responding to, even if they are not recognised by the creature.

ii) Although the creature is unobserved, this doesn’t entail that we ourselves cannot reasonably presume that the creature reacts to particular aspects of its environment. Indeed, given that the focus of this first phase was taken to be epistemological and not constitutional, the claim that creatures do react to particular causes in their environment has not been put in doubt by anything said thus far. From this, it follows that, even if we allow that a minimally triangular relationship between an observer and at least two creatures is necessary for identifying the particular cause of the creature’s reaction (see Section III.2), this does not by itself establish that it is the triangular relation which fixes the cause of any particular reaction. A point parallel to this has been made by Martin Montminy, from whom I appropriate the following quotation to my own purposes [Mon03:35]:

… it is not clear why there needs to be (or to have been) an actual interaction with a second person in order to determine the cause of [a solitary creature’s] response. Why not say that the solitary [creature’s unobserved response ‘X’ is still

13 Thinking back to Chapter Two, Section IV, to convey this scenario of imagining an entirely solitary creature might also count as another instance where the concept of objectivity might be supported exclusively by language.

102

a response to the object, X,], since an appropriately informed [observer] could establish this, if such an [observer] were to [observe it]? In other words, why not say that the possibility of triangulation with a [creature] enough like [it] is sufficient to determine what the solitary [creature] is reacting to?

While I may not be able to say of an unobserved creature what it is that they are reacting to in particular, this epistemic limitation does not compel me to conclude that they are not reacting to particular objective features of their environment. For, if I were to observe them (perhaps in the context of a minimally triangular relationship), it is reasonable to presume that I could then identify the causes of their reactions. The mere potential for an observer to pick out the cause of a creature’s reaction then is sufficient for saying that its reactions have particular causes.

The consequences of the above reasoning for the later phases of the argument are that, insofar as we might come to view triangulation as resolving the problem of distance, it nonetheless seems plausible at this stage of the argument that the potential to be triangulated with might well suffice for beliefs of determinate content also. While we might come to find that, so long as a creature’s behaviour can potentially be triangulated with, then it can be said to have beliefs, nothing so far indicates that it is because of the presence or absence of actual triangulation that the creature does or doesn’t have beliefs of determinate content. Indeed, the very same consideration is advanced by Kathrin Glüer when she writes [Glü06a:15]:

While it might be true […] that if a creature has language, and, thus, thought, triangular principles of content determination do apply, it [...] remains doubtful that they apply because of the triangulation.

iii) Finally and most significantly, Davidson has implicitly conflated the creature’s capacity to self-reflexively pick out the salient causes of its reactions from the question of whether its reactions have particular causes. That is, while one might concede the creature itself has no explicit conception that it is reacting to an objective environment per se, this doesn’t entail that it isn’t reacting objectively to an objective environment. Indeed, in this vein, empirical evidence weighs heavily against Davidson’s subsequent claim that the creature’s circumstantially solipsistic ‘world can be any size’. Burge for instance, draws on a wealth of

103

scientific research demonstrating that many animal species are equipped with perceptual systems above and beyond mere sensory systems, allowing them to perceive aspects of their environments objectively (see, for instance, [Bur10:421–30]). Perception in the main is characterised by a creature’s ability to represent aspects of its environment as being distinct from the creature’s own sensory stimulations. Scientific practice, claims Burge, has established that explanations appealing to conditions of veridicality uniquely serve and so are vindicated by their explanatory power. As Burge explains [Bur10:408–9]:

Perceptual constancies are capacities systematically to represent a particular or an attribute as the same despite significant variations in registration of proximal stimulation. […] It is understood that these capacities cannot be explained simply as generalised weightings of registration of proximal stimulation. They must involve principles for forming representation of specific environmental particulars and attributes. […] Perceptual constancies show up saliently in behaviour. They are recognised in empirical theory. They are the primary mark of perceptual objectification. […] Various constancies occur in the visual perceptual systems of animals. […] The biological and information-theoretic forms of explanation that apply to non-perceptual systems remain applicable to perceptual systems. But an additional type of explanation is explanatory, as well. This type appeals to representational content as marking conditions for veridicality – in this case, perceptual accuracy. Veridicality is success in fulfilling not a biological function, but a representational function. […] The applicability of this type of explanation is supported by the system and specificity found in objectifying capacities present in the perceptual constancies.

And further [Bur10:310]:

We know empirically that there are perceptual states with such accuracy conditions. There is extensive support for explanations in which the representational aspects of perceptual states are explanatorily central.

Burge then goes on to list the phylogenetic distribution of perceptual systems from the visual perception of jumping spiders to the sonar perception of whales (see [Bur10:419–21]).

104

Admittedly, perceptual content is still a far cry from propositional content. But, even so, Davidson is evidently wrong in claiming that ‘[the creature’s] responses, no matter how complex, cannot show that it is reacting to [...] events a certain distance away rather than, say, on its skin’ [Dav92/01:119]. Obviously the creature may itself still lack the concept of objectivity; as Burge suggests [Bur10:257]: ‘objectification can be carried out in perceptual subsystems of the individual’. But insofar as a creature’s perceptual systems allow it to perceive objectively, Davidson cannot in this way motivate his later principal claim that the creature lacks beliefs of determinate content.

In light of the above, Davidson’s first excursus towards establishing the necessity of triangulation for the determinate content of belief is far from conclusive.

2. The Creature Observed

The next stage of the pre-representational phase of Davidson’s Argument from Content Determination involves observation of a solitary creature. Even here, Davidson persists in denying that we as observers might legitimately pick out the salient cause of a creature’s reaction to its environment.

In the following passage, Davidson clearly restates the problem of content determination with respect to the observation of a solitary creature [Dav01:4–5]:

Since any set of causes will have endless properties in common, we must look to some recurrent feature of the gatherer, some mark that he or she has classified cases as similar. This can only be some feature of the gatherer’s reactions ... in which we must ... ask: what makes these reactions relevantly similar?

As Davidson sees it, this closing question can only be answered in such a way as to undermine the very undertaking that was its charge [Dav91/01:212]: ‘This criterion cannot be derived from the creature’s responses; it can only come from the responses of an

105

observer to the responses of the creature’. Any effort to attribute the cause of a creature’s reaction cannot help but project our own criteria for what we take to be the salient causes.

While most would acknowledge that, in order to pick out the cause of a creature’s reaction to its environment, we, as observers, must project our own criterion for what is the salient cause of the creature’s reaction to its environment, nonetheless, all would presumably want to qualify this concession with one of the following three caveats:

i) Our projections either can or will be minimised by principled, systematic processes that enable us to home in on what is invariant in the creature’s responses to its environment, or alternatively our projections may at least be identified and ‘bracketed’ so as to avoid any naive misattributions.

ii) Our initial reliance on the projection of what we take to be salient can or eventually will be wholly overcome, perhaps by the same principled, systematic processes that were referred to in the above caveat.

iii) At least in some instances, what we take to be the salient cause will be the same as what the creature in fact reacts to, consciously or not.

Davidson, however, takes our reliance on our own criterion to entail the following prohibition [Dav90/01:203]:

Without [triangulation] there would be no grounds for selecting one cause rather than another as the content-fixing cause. A non-communicating creature may be seen by us as responding to an objective world; but we are not justified in attributing thoughts about our world (or any other) to it.

If Davidson here were only pointing out, for example, that a dog’s reaction to its master coming home cannot be said to be about its master understood as ‘a master’ (in the same explicit conceptual manner that she might also be ‘the postmaster general’), then all would likely agree. As Davidson puts it [Dav01:11]: ‘We would be making a mistake if we were to

106

assume that because the creature discriminates and reacts in much the same way we do, that it has the corresponding concepts’. But while we might hypothetically allow that the creature has no beliefs, it is not only deeply counter-intuitive, but (as was seen in Section III.1) is at odds with the various facts of empirical investigation to suppose that we couldn’t justifiably attribute determinate causes for its reactions, albeit with recourse to our own grasp of concepts. After all, our projections might nonetheless be principled and explanatorily vindicated.

To this extent then, the second excursus of Davidson’s Argument from Content Determination is off to a shaky start also. There is of course, much more to say about this stage, but before moving to do so, let us first introduce the next stage, for what applies here will, I shall argue, equally apply there also.

3. Non-Interacting Creatures

Although Davidson denies our capacity to attribute particular causes of a solitary creature’s reactions to its environment without unduly projecting what we ourselves take to be salient, he does think that when a further creature is introduced into the environment, even if the two do not interact with each other, as observers we will be no longer inextricably bound to our own viewpoint, such that our attributions may thereby be justified. As Davidson explains [Dav91/01:212]:

... it is only when an observer consciously correlates the responses of another [second] creature with objects and events of the observer’s world that there is any basis for saying the [initial] creature is responding to those objects or events rather than any other objects or events.

And again [Dav01:6]:

If a second lioness joins the first in pursuit of [a] gazelle, I can eliminate [the] complete dependence on my own choice of salient object in this way: I class

107

together the responses of lioness A with the responses of lioness B in the same places and at the same times. The focus of the shared causes is now what I take to be the salient object for both lionesses. I no longer have to depend on my own choice of the relevant stimulus of the lionesses’ behaviour.

The thought here is that the similarity of responses shared between the two creatures allows us to overcome the undue projection of what we take to be the salient features of the environment, which we then suppose are the cause of the creature’s reactions.

It is worth pointing out that this arrangement of an observer and two creatures interacting with a shared environment but not each other, is neither the pre-linguistic nor the semantic triangulation Davidson holds necessary and sufficient for belief. All we are presently concerned with at this stage of the argument is the question of what, if anything, can be said to be the cause of a languageless creature’s reactions. That being said, as was noted at the outset of this section, the reasoning underlying Davidson’s claims concerning this preliminary stage parallel his considerations regarding triangulation proper, and so are intended to indirectly support his later claims.

The claim that the cause of a creature’s reactions can be only be identified when an observer is able to compare the reactions of at least two creatures in a shared environment (though they need not interact with one another) appears odd at best. To see how this is so, let’s pause for a moment longer on the case of the solitary creature.

Perhaps if the situation were the first-ever spotting of some previously unknown species of animal, we might be tempted to concede that our initial observations of its lonely and novel interactions with the environment must, more than normally, rely heavily on our own perceived correlations of the salient features of the environment and the thereby seemingly caused responses of the creature. But even then this would be to counterintuitively assume that we still could not legitimately draw on our past experiences of observing similarly anatomically and behaviourally constituted creatures (e.g. with unidirectional eyes and omnidirectional ears, etc.) in making certain reasonable assumptions about what the creature is attending to when, say, it directs its gaze toward an area of the environment that it then reaches its hand out to take hold of.

108

Even if we were to stretch the example from lions and gazelles to a first ‘encounter of the third kind’ with some bizarrely embodied extraterrestrial being, even then the singularity of the creature appears to be of no less consequence than if there were a host of such alien lifeforms. The reason being: strictly speaking, nothing more is gained by the observation of two different lions reacting to their environment ‘in the same places and at the same times’ than to observe a solitary creature over time and under differing circumstances. Indeed, systematically altering various aspects of the environment and observing changes in the reactions of the same creature is a more rigorous process by which to determine the causes of its reactions (and beliefs, if any), than by noting whether two creatures of the same species react in the same way as each other. The similarity of responses between two creatures can be instructive only insofar as it allows us to make sense of their differences; were we to observe strictly identical behaviour in any number of creatures placed in identical circumstances, we would be left in no better position than with a single creature.

Of course, when Davidson writes of similarity in terms of two creatures’ reactions at the same places and at the same times it is clear that he doesn’t mean to say the reactions of the creatures have to be strictly identical. By ‘similarity’ he means that the differences between the two have to be both close enough and regular enough to allow an observer to home in on the invariant features shared between the two responses. What then is invariant between the two creatures can thereby be legitimately said to be the cause of their individual reactions. And when the two creatures are said to ‘differ in exceptional cases’, the import is that the observational process of correlating what is invariant invariably breaks down, leaving no principled means by which to determine the relevant cause. Such a common sense reading of Davidson’s intended meaning is not possible here however, as the following considerations reveal.

Everyone will readily allow that observing two creatures interacting with a shared environment (even if they do not interact with one another) would augment the process of determining the exact cause of either creature’s responses. For instance, consider an ethological investigation of whether a lion’s ability to predict the direction a gazelle will likely flee is on the basis of the gazelle’s eye movements or just the direction of its head. In order to tell, one might envisage observing the respective reactions of two lions, only one of which is in a position to see the gazelle’s eyes. If the lion that can see the eyes of the gazelle pounces in the direction in which the gazelle’s eyes were focussed, while the lion that could not see the eyes of the gazelle instead pounces in the direction of the gazelle’s head, then we

109

would have some (albeit limited and here, idealised) evidence to think that the cause of the first lion’s reaction was the eye-movement of the gazelle. By comparing the relevant differences in the creatures’ reactions to differences in their available causal stimuli, we are thereby able to vindicate (or alternatively reject) our initial projections of what are the salient cause or causes of the creatures’ reactions.

However, to draw on such comparisons in narrowing the identified cause (or causes) of the creatures’ respective reactions, one must first assume of each lion individually that one can already determine (albeit perhaps to a comparatively coarse degree) the general causes of their responses to the objects in the environment. This is because, in order to make such comparisons, one must have something to compare with. (Similar considerations apply mutatis mutandis to the increasingly precise determination over time regarding the causes of a solitary creature’s reactions.) But insofar as Davidson has precluded the possibility of legitimately ascribing causes to a solitary creature, he has inadvertently further excluded doing so with respect to multiple creatures also. Of course, if Davidson’s point had straightforwardly been to illustrate that there are relative degrees of objectivity attainable in determining the cause of a creature’s reactions to its environment, such that Davidson might then want to claim (still seemingly by fiat) that in the absence of a second creature, the influence of our own projected determinations would be too strong to be legitimate, then almost everyone would have conceded the general spirit (if not the letter) of such a claim. But Davidson’s argument is not made in such a spirit. Davidson’ claim is that we can only determine the cause of any creature’s reactions in the context of a minimally triangular relationship of an observer and at least two creatures reacting to a shared environment (though not necessarily to one another). We now see that the latter portion of this formula is inconsistent with the former. That is, if we can’t say at least something informative about the cause of an individual creature’s reactions, then by what means could we come to correlate the reactions of two individual creatures’ reactions to a supposedly shared environment?

4. Some Preliminary Conclusions

The first, pre-representational phase of Davidson’s Argument from Content Determination falters at each step along the way. In the first instance, even the unobserved solitary creature

110

can potentially be said to be responding to objective causes in their environments. Once we are free to observe the creature then, although we will indeed have to make recourse to our own concepts, we are not thereby precluded from drawing on general experiences, and systematically observing differences in the creature’s behaviour over time. Indeed, in the end, the capacity to triangulate and so augment our observational investigations by introducing multiple creatures into a shared environment depends on our being justified in doing so.

Far from providing any compelling reason to suppose that it is impossible to have beliefs of determinate content outside of the context of triangulation, the immense bulk of empirical, theoretical and, indeed, commonsensical evidence supports the claim that at least the perceptual origins of objectivity are without question, alive and strong in many of our animal relatives.

As to the still-remaining question of whether triangulation is directly necessary for propositional beliefs, let us now turn to consider the case of pre-linguistic triangulation, which takes the form of what I shall call, ‘intra-reactive triangulation’.

IV. INTRA-REACTIVE TRIANGULATION

The second phase of Davidson’s Argument from Content Determination focuses directly on the problem of distance with respect to beliefs. Davidson argues that pre-linguistic triangulation is a necessary condition for beliefs of determinate content, which is to say by Davidson’s lights, for the very possibility of any beliefs at all.

Recall from Part One, however, that no such higher-order meta-cognitive understanding was shown to be necessary for the possession of beliefs. We will maintain that commitment here also. That said, the issue of the minimal requirements for possession of such beliefs about beliefs remains of the utmost significance insofar as the particular focus of this thesis is the concept of objectivity and the question of the role that language may play in its

111

possession. While Davidson’s claim that beliefs about beliefs are necessary for any beliefs will not be further countenanced here, his claim that pre-linguistic triangulation is a necessary condition for the possession of beliefs about beliefs remains pertinent to the present investigation. I myself am not so much interested in demonstrating that only humans can think as I am in considering the role of language in enabling us to think in some of the more unique and characteristic ways that we do – in particular, in our ubiquitous application of the concept of objectivity.

With the above in mind, Davidson outlines the pre-linguistic phase of his argument as follows [Dav91/01:212–3]:

[In triangulating,] each of two people is reacting differentially to sensory stimuli streaming in from a certain direction. Projecting the incoming lines outward, the common cause is at their intersection. If the two people now note each other’s reactions (in the case of language, verbal reactions), each can correlate these observed reactions with his or her stimuli from the world. A common cause has been determined.

As may be noticed here, the parenthetical qualification that the two peoples’ noting each other’s reactions may be linguistic serves to hint that Davidson does not take pre-linguistic triangulation to be sufficient for ultimately resolving the problem of distance (let alone supporting possession of the concept of objectivity). We can see this point being made more explicitly in the following passage [Dav90/01:202]:

Without one creature to observe another, the triangulation that locates the relevant objects in a public space could not take place. I do not mean by this that one creature observing another provides either creature with the concept of objectivity; the presence of two or more creatures interacting with each other and with a common environment is at best a necessary condition for such a concept. Only communication can provide the concept, for to have the concept of objectivity, the concepts of objects and events that occupy a shared world, of objects and events whose properties and existence is independent of our thought, requires that we are aware of the fact that we share thoughts and a world with others.

112

The question of the sufficient conditions required to resolve the problem of distance (along with the question of possession of the concept of objectivity) will therefore be postponed until the next chapter, which deals directly with the question of semantic triangulation. For the time being, the focus will be restricted to the question of the necessity of pre-linguistic triangulation and to its supposed insufficiency for beliefs of determinate content.

As in the previous section, the strategy here will be to systematically move in stages towards pre-linguistic triangulation, evaluating the steps along the way in their own right before drawing some more general conclusions in overview. We begin by considering two creatures observing one another (though not mutually interacting), and then turn, in Chapter Four, to mutually interacting creatures in pre-linguistic triangulation proper.

1. Intra-Reacting Creatures

The first stage of pre-linguistic triangulation consists of at least two creatures – here, lions – reacting both to a shared environment and to the reactions of the other creature. Davidson’s strategy is to shift the burden of identifying the relevant cause of the lions’ reactions to the lions themselves. As Davidson explains [Dav01:7]:

The challenge is to put the lions in a position to distinguish these [exceptional] cases [where their respective reactions differ]. To do this we have to eliminate my arbitrary (or interested) choice of relevantly similar responses on the parts of the lionesses. A further element enters when the lions cooperate to corner their prey. Each watches the other while both watch the gazelle, noting the other’s reactions to the changes of the direction.

Unlike in the previous phase of Davidson’s argument, the point of interest here is no longer to do with our own observational legitimacy. Rather, the suggestion is that this context of two creatures reacting, not only to their environment but to the other’s reactions to the environment as well, is minimally necessary for the creatures themselves to be able to

113

distinguish cases where their comparative reactions coincide or differ. Davidson’s claim is that pre-linguistic triangulation provides at least the potential for the creatures to become aware of what they are reacting to, and to this extent, may also thereby offer some scope for the possession of beliefs about beliefs.

Let us then carefully unpack Davidson’s account. Here is what he initially says on the matter [Dav97/01:128]:

[Triangulation involves] a threefold interaction, an interaction which is twofold from the point of view of each of the two agents: each is interacting simultaneously with the world and with the other agent.

Not only will the creature be attentive to the shared object in the environment, but it will also be attentive to the reactions of the other creature to, both, the same object and the observing creature itself.

There are three issues to note in the account thus far:

i) It is unclear how a creature might come to accomplish such a feat considering that Davidson had all but denied the capacity to determine the cause of a solitary creature’s reactions. (See Section III.)

ii) Even if the issued raised in (i) is resolved, it is still not obvious how this threefold reaction is relevant to the creature’s potential to be self-aware of the cause of its own reactions (let alone beliefs). While we could certainly imagine a creature reacting both to an object in its environment and simultaneously to another creature reacting to, both, that same object and the first creature’s reactions, it is not clear that the latter feature has, from the first creature’s perspective, any bearing on its own relationship to the object in question. Thus far, the second creature remains nothing other than another object of the environment, which then is equally prone to the problem of distance and other issues of content determination.

114

iii) Even if we allow that two creatures can somehow react, both, to a shared object and to one another’s reactions to the initial reaction and the subsequent reactions, exactly what this has to do with the capacity for beliefs about belief is also unclear. As Katrin Glüer asks [Glü03:11]: ‘why would what is necessary for monkey reactions to have determinate objects also be necessary for propositional thought?’

Davidson acknowledges that these are the very issues at stake [Dav92/01:119–20]:

Of course [merely reacting both to an object and another creature’s reaction to that same object] cannot be enough, since mere interaction does not show how the interaction matters to the creatures involved. Unless the creatures concerned can be said to react to the interaction, there is no way they can take cognitive advantage of the three-way relation which gives content to our idea that they are reacting to one thing rather than another.

Here then is Davidson’s proposal for how to resolve these obstacles [Dav92/01b:160–1]:

What makes the idea of [an objectively mind-independent] object available to us? Its availability requires, above all, the corroboration of others, others who are tuned to the same basic events and objects we are, and who are tuned to our responses to those events and objects. When the second person enters the scene, he or she can correlate my responses with his or her own responses; for the first time it makes sense to speak of responses being ‘the same’, that is, relevantly similar. And for the first time, there is reason to speak of responses being responses to external objects rather than to immediate sensations; when two (or, of course, more) creatures can correlate their responses, those responses triangulate the object. It is the common cause of the responses, a cause that must have a location in shared, interpersonal space. When we can notice that we share reactions, the possibility of a check or standard is introduced, the possibility of occasional failures of expected joint reactions and hence of error.

With respect to the first issue, (i), two creatures are said to be able to become at least minimally aware of the relationship between the object and the other creature insofar as

115

both creatures share similar responses. As Davidson points out [Dav99b:731]: ‘Many animals of the same species share the same discriminative abilities…’ And further [Dav90/01:202]:

… we group together the causes of someone’s responses, verbal and otherwise, because we find the responses similar. What makes these the relevant similarities? The answer again is obvious; it is we, because of the way we are constructed (evolution had something to do with this), who find these responses natural and easy to class together. If we did not, we would have no reason to claim that others were responding to the same objects and events (i.e. causes) that we are. It may be that not even plants could survive in our world if they did not to some extent react in ways we find similar to events and objects that we find similar. This clearly is true of animals; and of course it becomes more obvious the more like us the animal is.

Of course, what we take to be similar in the reactions of our fellow humans, animals and perhaps, as Davidson suggests, even plants, is a strictly empirical matter; there is nothing theoretically heavyweight here – no hint of anything normative in this notion of similarity. This point, however, has its objectors and so deserves a moment’s pause.

* * *

Peter Pagin charges Davidson’s account with begging the question [Pag01:204]:

… it looks as if part of the difference between just reacting and reacting to specific objects or events is having knowledge not only of those objects or events but also of other persons and their knowledge of the same objects or events and of oneself. On the surface, this is a clear case of invoking, in the account itself, those categories and capacities that were supposed to be accounted for.

Here Pagin argues that Davidson begs the question in suggesting that the solution to the problem of distance lies in the creature ‘having knowledge’ that the reactions of other creatures are relevantly similar and pertain to the same specific object that the creature

116

themselves is reacting to. This is because it seems to presuppose that the creature must know: i) that the other creature is reacting to something; ii) what it is specifically that the other creature is reacting to; and, iii) that the creature itself is reacting to that very same object. Obviously, if Pagin is correct, these conditions do beg the question in presupposing possession of the very capacities the account was intended to account for.

In response, Davidson moves to create some dialectical space whereby his account does not circle back on itself too tightly [Dav01:292]:

This description [of two creatures correlating the other’s reaction with the observed external stimulus] is obviously a description that can be made only by someone with the concepts I have just used, but nothing in what I have described implies that the interacting creatures are capable of describing what they are doing. Nor does it imply that the creatures are capable of describing anything at all; I am talking here about interactions among creatures which well may lack concepts. [...] My thesis is that the triangular arrangement I have described is a necessary condition of thought. It is not sufficient.

Were Davidson’s account straightforwardly sufficient, it certainly would thereby render itself question begging, as Pagin contends. But as Davidson sees it, the pre-linguistic stage of his Argument from Content Determination does not claim to credit the creatures with beliefs, nor with any other normative notions of ‘knowledge’, ‘recognition’, ‘similarity’, etc.. Maria Lasonen and Tomás Marvan, for their part, concur [Las04:188–9]:

A non-linguistic being’s responses can be grouped as similar according to a standard provided by a linguistic being, and this standard helps individuate a content-endowing cause. Pagin is right that this cause cannot endow thoughts with contents or words with reference before both beings are fully aware of the situation. But the process through which the content-endowing cause is individuated – before both beings are aware of the triangular situation it is just the cause that will endow thoughts with contents once thought is present – takes place at least in part before both beings actually have full-fledged thoughts.

117

In this manner then, Davidson avoids the objection that he attributes to newly-triangulating creatures the very advantages that triangulation is taken to be necessary in order to possess.

* * *

With the above in mind, let’s return to Davidson’s proposed answer to (ii) – i.e. how the other creature can come to be something more to the creature than merely another object in its environment. Davidson’s response rests with the supposed capacity to correlate the reactions of another creature with the creature’s own reactions. As Davidson explains [Dav97/01:129]:

In [triangulation] two (or more) creatures each correlate their own reactions to external phenomena with the reactions of the other. Once these correlations are set up, each creature is in a position to expect the external phenomenon when it perceives the associated reaction of the other. What introduces the possibility of error is the occasional failure of the expectation; the reactions do not correlate.

The learnt correlations of both creatures’ similar yet specific reactions to various objects is claimed to allow them to view the other’s reactions not only as yet another event in their respective environments, but also as responses to a particular object in a shared environment. As the above quotation continues, Davidson illustrates the significance of this capacity by suggesting it is in this way that a creature may come to expect that the relevant object is present upon observing certain behaviours of its conspecific that have now come to be associated with the presence of the object. For example, when the lioness sees another lion suddenly adopt a pose that is typical of the hunt, the former may well expect that there is prey nearby. The full significance of being capable of such correlations is revealed when the conspecific’s reactions do not correlate in the manner with which the creature expected them to. In this way, suggests Davidson, the possibility for error arises.

This last stage is subtle and potentially misleading, and so deserves careful attention. One seemingly plausible (though incorrect) reading of what Davidson takes to be revealed when the expected phenomena do not match the observed correlations of the two creatures’ responses, is that this might allow either creature to simultaneously entertain two

118

apparently contradictory expectations (if not outright beliefs). To see how this might be so, reconsider the scenario sketched above: a lioness sees its conspecific freeze with its eyes directed at the scrub ahead – a stance that the former has come to associate with the presence of prey. The former lioness is now in a position to expect that potential prey lies in the bushes ahead, although the lioness may not be able to see it herself. On the one hand then, it now has an expectation that there is prey somewhere concealed in the scrub, but on the other, it has no direct perceptual evidence of the prey’s presence at all. It might seem that here there is scope for the creature to come to appreciate that one or the other, or both of its current expectations might be false. This supposed simultaneity of contradicting expectations is distinct to what was characterised in both chapters dealing with Davidson’s ‘Main’ Argument as a recognition that the creature’s past belief had been false by an appropriate change in beliefs. Here the expectations of the creature do not alternate, but rather are simultaneously at odds with one another. As such, it might appear that insofar as triangulation enables the creature to simultaneously maintain contradictory expectations, it its minimally necessary for it to later have beliefs about beliefs, which as Davidson has maintained all along, is basic for possession of the concept of objectivity. As plausible as it may initially seem however, this cannot be the reading that Davidson has in mind.

One obvious reason why Davidson shouldn’t wish to endorse the above reading is that the same potential to simultaneously entertain contradictory expectations is readily applicable outside the context of pre-linguistic triangulation. Purely for the sake of illustration (and clearly without Research Ethics approval), we can imagine that if initially a dog had been conditioned to expect a reward upon depressing a button, but later the rewards were randomly interspersed with mild electric shocks, then, while the dog may nevertheless persist in pressing the button for the sake of sustenance, it might be in two minds regarding what to expect in response. The possibility of generating simultaneously-held contrary (and so by extension, contradictory) expectations seems to not be the exclusive domain of triangulation.

Rather, the correct interpretation of significance of ‘the occasional failure of the expectation [such that] the reactions do not correlate’ [Dav97/01:129] is to suggest that the possibility for error proper exists only in light of differences in reactions between behavioural similarities that have otherwise been correlated by the creatures themselves. Against the common sense presumption that errors are primarily to be understood as a relation between a creature’s expectation and the objective state of affairs to which the expectation is

119

pertinent, Davidson’s consideration of pre-linguistic triangulation instead claims that the source of the possibility for error necessarily lies in interactions between two creatures. In other words, according to Davidson, a creature’s atypical reaction to its environment does not by itself constitute an error, even if it is the result of a misperception. Rather, the source of the possibility for error rests in the creatures’ expectations that the other’s reactions will be similar to their own. When we again read the above quoted passage in this light, we see that the emphasis isn’t on the creature having an expectation that an object will be present, but rather that ‘the reactions do not correlate’. It is this then that is said to be the original source of error. That is, the two creatures’ otherwise correlated reactions may occasionally differ.

It is vital to notice, however, that at this stage Davidson is not yet wanting to claim that a frustrated expectation that the other creature’s reaction would correlate to the creature’s own is sufficient for error proper. Rather, as Davidson puts it [Dav97/01:129 (italics added)] ‘What introduces the possibility of error is the occasional failure of the expectation…’ That is to say in a more belaboured manner: the occasional failure of the expectation is necessary for the eventual advent of error to be possible. Admittedly, it does appear that Davidson believes otherwise; for instance when he writes [Dav94/05:124]:

If you and I can each correlate the other’s responses with the occurrence of a shared stimulus, however, an entirely new element is introduced. Once the correlation is established it provides each us of with a ground for distinguishing the cases in which it fails. Failed natural inductions can now be taken as revealing a difference between getting it right and getting it wrong, going on as before, or deviating, having a grasp of the concepts of truth and falsity.

That said, he straightaway continues on with the following [Dav94/05:124]:

A grasp of the concept of truth, of the distinction between thinking something is so and its being so, depends on the norm that can be provided only by interpersonal communication.

120

And as yet, nothing that Davidson has described at this pre-linguistic stage of triangulation counts as communication proper. Rather, as was suggested above, it can only plausibly be the case that Davidson is here claiming to have identified the necessary, though not sufficient, conditions for the possibility of error. As he puts it elsewhere [Dav97b/01:83]:

The ultimate source (not ground) of objectivity is, in my opinion, intersubjectivity. If we were not in communication with others, there would be nothing on which to base the idea of being wrong, or, therefore, of being right, either in what we say or in what we think.

But, if we don’t yet have error proper at this pre-linguistic stage of triangulation, what then do we have? Thinking back to what was said in the previous section with regard to the importance of both similarity and differences in the reactions of either two observed creatures or one creature systematically observed over time, here I take it that the crux of Davidson’s point must similarly rest on an ongoing process of correlation between the two creatures’ similar, though always slightly unique reactions to shared objects and events (perceived, of course, from always slightly different perspectives). In such a way, by pre-representationally correlating the similarities and differences between each creature’s responses, both, over time and in various circumstances they may thereby be able to increasingly home in on a salient cause of their own reactions.

In answer, then, to the above question (ii) of how pre-linguistic triangulation is necessary for a creature to be able to regard its conspecific as anything more than simply another aspect of its general environment, Davidson’s claim is that when the creature’s own reactions to certain objects and events in the environment are correlated with another creature’s similar reactions to objects and events in their shared environment, the reactions of both creatures have the potential to be seen as reactions-to-the-environment as seen from the creatures’ own perspectives. By contrast, a creature’s ability to correlate objects and events in the environment with other objects and events (such as dark clouds being suggestive of rain) cannot support a subsequent capacity to view its own reactions as reactions to the environment as such because such objects and events lack the necessary similarity of response to the creature’s own responses. It is only when the other creature’s responses are relevantly similar to the creature’s own responses that the first creature’s reactions to the environment can be correlated with the other creature’s reactions to the same objects and

121

events in such a way as to support the notion that the creature’s own reactions are (or more indicatively, mightn’t be) reactions to particular objects and events in the environment.

V. CONCLUSION

The above considerations are intended to serve as Davidson’s answer to the problem of distance. Pre-linguistic triangulation is claimed as necessary for a creature to be able to view its own reactions to the environment as having particular causes. Admittedly (and as was established in the previous section), the creature itself needn’t be able to solve the problem of distance in order for there to be particular causes of its reactions; there is a wealth of empirical evidence that creatures’ innate perceptual capacities give rise to veridicality conditions. In this light, Davidson went too far in his initial claim, as even he himself appears to acknowledge later on [Dav03:697]:

… it is not, and perhaps never was, politically correct to say dumb brutes don’t think. We know what they can do, and that certainly includes all sorts of things right-thinking, or right-talking people, call thinking. I’m caving in on the words ‘think’, ‘represent’, ‘cognition’, and so on. I hold out on propositional contents.

That said (and as was seen above), Davidson nevertheless persists in his claim that for the creature itself to be capable of recognising that its own responses have particular causes, it is minimally necessary that the creature triangulates with another creature of similar behavioural responses. Such a capacity for the creature to recognise that its own (and others’) responses are responses to particular causes is said by Davidson to be pertinent to the question of the necessary conditions for possession of the concept for objectivity. While I too am happy to allow that creatures can possess beliefs without the capacity for meta-cognition, accounting for the possession of the higher-order concept of objectivity nonetheless requires that the creature itself be aware that its beliefs are beliefs in the sense that they are not necessarily true. In this light, if to have the concept of objectivity, one must be able to recognise that one’s beliefs have determinate content such that truth conditions are (at least seemingly) applicable, then Davidson’s account of the necessary conditions for

122

content determination remain highly pertinent and, moreover, are suggestive that communication may play a vital role in the possession of the concept of objectivity (as shall be considered in the following chapter and beyond).

At this point, one may (again) wish to object that a solitary creature over time might equally be able to come to correlate their own similar reactions to relevantly similar objects and events such that it too could thereby home in on the particular causes of its reactions. This is a similar point to that which was raised earlier (see Section III.3 above) when asking why the observation of two creatures interacting with an object in a shared environment (though not necessarily interacting with one another) can, according to Davidson, relieve us of our reliance on the illegitimate projection of what we ourselves take to be the salient cause of the creatures’ reactions, whereas the systematic observation over time of a single creature allegedly could not. Echoing this, the question here is why a solitary creature can’t plausibly draw on memories of its own experiences in such a way as to correlate its reactions to various past and present objects and events in such a way as to home in on the particular causes of its reactions in much the same way that Davidson argues is possible only in the context of triangulation between at least two creatures. If this alternate method were possible, then the process by which the solitary creature would home in on the particular cause of its reactions to the environment may equally well provide the ground for it to possibly come to recognise that its reactions are reactions to the environment. What then, if anything, is the crucial difference between two creatures triangulating with one another, and one creature ‘triangulating’ with itself over time?

Davidson offers some reasoning against the possibility of taking the solitary creature to be capable of homing in on the causes of its own reactions [Dav94/05:124]:

No matter what the stimuli, your similar reactions will indicate that you found something similar in the situations; and apparently dissimilar responses to the same stimulus can equally be taken to show that you took the stimulus to be different, or that for you is a similar response. As Wittgenstein says, by yourself you can’t tell the difference between situations seeming the same and being the same.

I don’t think we should hold much stock by this, however, for two reasons:

123

i) Davidson’s account is either empirical or a priori in nature. If it is empirical, then the purely theoretical possibility of a creature having similar reactions ‘no matter what the stimuli’ is of little interest. It is presumed as requisite for the possibility of ethological investigation and, moreover, borne out by explanatory successes that the variation in a creature’s reactions shares at least some tangible relation to the range of variation in stimuli. Burge makes a similar point when he writes [Bur10:407]: ‘... the perceiver need not be able to distinguish bodies from philosophically contrived stand-ins. The empirical explanation of the perceptual discriminatory ability counts these cases as irrelevant non-alternatives’. While it is plausible to suppose that, empirically, there may be more diverse and tangible ‘friction’ between responses of two triangulating creatures than is available to a solitary creature drawing on its own memories, such that together they could more readily and precisely home in on the causes of their reactions, nevertheless, nothing has been advanced in Davidson’s account thus far to block the possibility that a lone creature might be able to draw on its own private past and present experiences to correlate its own reactions to objects and events in such a way as to thereby come to recognise that its reactions are reactions to a distal environment.

ii) If, on the other hand, Davidson’s account is strictly theoretical, then the sceptical possibility that we might have similar reactions to widely variant stimuli is incidental to the claim that triangulation is a necessary condition for content determination. The burden rests with Davidson to show that only triangulation can solve the problem of distance, but not that any other means (such as a creature triangulating with itself) are unreliable, for they may well be reliable also, even if only occasionally.

In light of these two considerations, then, Davidson’s claim that only triangulation between two creatures is able to solve the problem of distance must be abandoned. It seems plausible enough to suppose that a lone creature may equally be able to home in on the causes of its reactions, thereby allowing it to come to recognise its reactions as being reactions to the environment.

That said, the notion of pre-linguistic triangulation is only claimed to be a necessary condition for solving the problem of distance insofar as Davidson presumes any satisfactory solution demands possession of the concept of objectivity. While I am happy to concede that Davidson goes too far in such a demand – that is to say, a solitary creature may well come to inarticulately solve a lower-order problem of distance – it remains of interest to the focus of

124

this thesis that Davidson should claim that it is only triangulation that may solve a higher-order problem of distance and, with it, the higher-order problem of objectivity also. Davidson’s reasoning here would presumably run: linguistic triangulation is sufficient for possession of the higher-order concept of objectivity; pre-linguistic triangulation is minimally necessary for linguistic triangulation; therefore, pre-linguistic triangulation is necessary (though not sufficient) for possession of the higher-order concept of objectivity.

That Davidson openly takes pre-linguistic triangulation as insufficient for the concept of objectivity and – what he initially considers to be equivalent – beliefs, is clear in the following [Dav01:13]:

Mere similarity of response is obviously not enough for thought, however, not even when one animal’s responses to events and features of the world serve to touch off responses appropriate to those same events and features in other creatures. This triangular arrangement is a necessary, but not sufficient condition, of thought. What else must be added to make the conditions sufficient? ... What else must be added to the basic triangle of two or more creatures interacting with each other through the mediation of the world if that interaction is to support thought? The unhelpful answer is communication.

In this light, it appears that any answer to our third question (iii) of how pre-linguistic triangulation is supposed to bear upon possession of the concept of objectivity, will have to wait until the following chapter, which directly considers Davidson’s claim that semantic triangulation is both sufficient and necessary for possession of the concept of objectivity.

125

CHAPTER FOUR

* * *

Interaction

I. INTRODUCTION

This chapter is concerned with Davidson’s Argument from Objectivity.

In the previous chapter, we considered a minimal form of ‘intra-reactive’ triangulation, whereby one creature more precisely homes in on the likely causes of its own reactions to objects and events in its environment by correlating them over time with repeated instances of its conspecific’s relevantly similar reactions to their shared environment. Distinctive to intra-reactive triangulation is that the conspecific need not interact with or even be aware of the creature that is observing it – intra-reactive triangulation could conceivably take place via a one-way mirror. In this light, while intra-reactive triangulation appears to go some way toward resolving the problem of content determination, it fails to establish the meta–representational intersubjectivity demanded by Davidson.

126

In this chapter, the basic triangular relationship of at least two creatures with a shared environment will be further developed to incorporate interaction between and mutual awareness of both of the creatures in question. I shall call this more sophisticated triangular relationship, interactive triangulation. Davidson argues in what is known as The Argument from Objectivity (though it might more accurately be called The Argument from the Concept of Objectivity) that interactive triangulation is a necessary condition for possession of the concept of objectivity. Davidson’s original line of reasoning rests on his commitment that the possibility of a concept of error that necessarily entails the concept of objectivity – the only ‘concept of error’ Davidson will recognise as such – requires an intersubjective standard whereby one knows the content of the thoughts of others, be they true or false. As was concluded in Chapter Two, we’ve little reason to follow Davidson in his emphasis on the concept of error as such. However, The Argument from Objectivity remains of interest insofar as it promises to shed light on how the move from intra-reactive triangulation to interactive triangulation might allow for a corresponding shift from mere reaction to interpretation, and perhaps even communication. If light can be shed on the matter in this way, then Davidson’s claim that language is necessary for possession of the concept of objectivity will be in a much stronger position than it has been in our investigation thus far.

Davidson outlines below the role that interactive triangulation is thought to play in enabling two creatures to each know the thoughts of the other [Dav92/01:121]:

The only way of knowing that the second apex of the triangle – the second creature or person – is reacting to the same object as oneself is to know that the other person has the same object in mind. But then the second person must also know that the first person constitutes an apex of the same triangle another apex of which the second person occupies. For two people to know of each other that they are so related, that their thoughts are so related, requires that they be in communication. Each of them must speak to the other and be understood by the other. They don’t, as I said, have to mean the same thing by the same words, but they must each be an interpreter of the other.

In order for Creature A to know that: i) it is reacting to a particular cause; ii) that Creature B is reacting to a particular cause; and, iii) that A & B are both reacting to that same cause, then A must know that B has the same cause in mind as A, which in turn requires that B knows what A has in mind, also.

127

Davidson claims that such mutual awareness of what the respective other has in mind is only possible via communication – though it remains to be seen if it need be full-blown language. The rough idea here (and as I shall spell out in more detail below) is that the inherent intersubjectivity of communication will allow the content of A and B’s thoughts to be afforded a sufficient degree of invariance whereby the two may in fact have the same content in mind and, so, know that this is the case and not otherwise (i.e. they will possess the concept of objectivity). Thus, from triangulation to the necessity for communication (and perhaps even language), we are promised an account of how the concept of objectivity emerges.

Given the distinction thus far made between intra- and interactive triangulation, it may seem that Davidson goes too quickly when he claims that, in order for A to know something about B, B must equally know it about A (i.e. ‘But then the second person must also know…’). After all, it seems plausible that Creature A might be at least minimally aware of the second apex of the triangle without seeming to require that B knows anything whatsoever about the presence of the first. But, of course, Davidson has something else in mind.

Recall that Davidson claims one cannot have thoughts unless one has the intersubjective standards made available only via communication. To quote again [Dav94/05:124]: ‘A grasp of the concept of truth, of the distinction between thinking something is so and its being so, depends on the norm that can be provided only by interpersonal communication’. By this, Davidson claims that the only way of knowing anything whatsoever is if one is able to know that the other person has (or hasn’t) the same thing in mind.

Is this question begging? Well, it depends on the presuppositions one brings to the table. Certainly it is if Davidson hopes to win his case by presupposing the very conditions he is attempting to explain, but not if we take him to be suggesting something more substantive in the above argumentative sketch. As I see it, the crux of Davidson’s claim rests in supposing that it will be by virtue of explaining how one creature might come to be able to know the mind of the other that we might have any hope of explaining how the creature can know anything whatsoever. In this light, the above quoted passage ought not to be so much an argument per se, but, rather, it is more like a treasure map where ‘X’ marks the spot of where to look for the ‘philosophers’ stone’, so to speak, enabling us to turn reactions into

128

interpretation, sounds into meaningful words, brain-states into thoughts and recognition of mistakes into the concept of objectivity. As Davidson puts it [Dav97/01:128]:

…there is a perhaps insuperable problem in giving a full description of the emergence of thought. […]

Despite these pessimistic remarks, I do have some suggestions about how we might approach the problem of saying something intelligible about the emergence of thought. There is a pre-linguistic, precognitive situation which seems to me to constitute a necessary condition for thought and language, a condition that can exist independent of thought, and can therefore precede it. Both in the case of nonhuman animals and in the case of small children, it is a condition that can be observed to obtain. The basic situation is one that involves two or more creatures simultaneously in interaction with each other and with the world they share; it is what I call triangulation.

In this light, triangulation – or at least its interactive form – is taken by Davidson to be a promising avenue of accounting for the emergence of thought; that is, of being able to most accurately pinpoint where non-semantic interaction gives rise to meaningful interpretation and communication. To be clear: Davidson does not claim that even the most ideally comprehensive analysis of triangulation will be able to successfully bridge the divide between the two worlds – one unthinking, the other meaningful. Or, to put it another way, Davidson isn’t confident that we have (or could ever have) the linguistic resources necessary to describe in non-question-begging terms the emergence of thought where it occurs at the point of triangulation. However, it should be clear from the above that Davidson believes that looking to the triangular relationship of at least two creatures interacting with one another in a shared environment is the best hope we have of providing an account of the emergence of language and the concept of objectivity to boot.

* * *

As before, the explication of Davidson’s argument will involve adding to the level of sophistication in the triangular arrangement under consideration. More specifically, the triangular relationship will be further developed from being intra-reactive to being interactive. Unlike in intra-reactive triangulation, where it was not necessary that both

129

creatures were able to observe one another, in interactive triangulation, the creatures not only must both perceive the other, but must do so in such a way as to make their very reactions themselves the foci of their shared perceptions. Unfortunately, Davidson’s writing does not provide any significant detail as to how this move is to be made, so the following section therefore consists almost entirely of original work in order to move from intra-reactive to interactive triangulation.14

Once the minimal structural framework for interactive triangulation is in place, I will explore the possibility of interactive triangulation being at least minimally interpretive. I conclude Section II by conceding that, even though some minimal form of interpretation may seem to be possible here, it doesn’t appear to have anything to do with interactive (as opposed to mere intra-reactive) triangulation and, as such, is less than promising for Davidson’s claim that communication (let alone full-blown language) is necessary for possession of the concept of objectivity. In the remainder of the chapter, I pursue work undertaken by Monima Chadha [Cha07], which explores whether empirical evidence of tactical deception among languageless creatures (primates, in particular) is able to establish meta-representational states, but conclude, in agreement with Chadha, that it is not.

II. INTERACTIVE TRIANGULATION

Recall that in Chapter Three, we left off at the particular stage in the development of the triangular context where at least two intra-reacting creatures could come to correlate not only the other’s reactions to particular causes in a shared environment, but potentially their own as well. It was not, however, deemed to be a condition of intra-reactive triangulation that the creature should subsequently react to the other’s reaction to their own initial reaction; the one creature need only observe the other so as to correlate their relevantly similar reactions to the mutually shared environment. To illustrate, we can see that intra-reactive triangulation could still take place for at least one of the creatures even if the two were separated by a pane of one-way glass, just so long as both had access to the shared 14 Davidson does provide illustrations of mutually interactive triangulations, but does so only in the context of ostensive learning taking place between someone who already possesses language and a child who is thereby being introduced into the semantic circle. See, for instance, [Dav91/01:212-3]

130

cause of their respective reactions. This stage was characterised as intra-reactive triangulation, because a fully interactive triangle had not yet been mandated.

This, then, is where the next step in the systematic development of triangulation picks up: in accounting for the move between intra-reactive triangulation and interactive triangulation. That this further development needs to be undertaken is demanded by The Argument for Objectivity insofar as it states that mutual intersubjective awareness is a necessary condition for possession of the concept of objectivity. As Davidson puts it elsewhere [Dav91/01:209–10]: ‘… only communication with another [creature] can supply an objective check’.

* * *

It is plausible to presume that insofar as the creatures react to their environment at all, they will also react to the reactions of their conspecifics. (Of course, for the purposes of the argument, we may simply stipulate that they do so.) Given that this is the case, it is reasonable also to presume that these reactions may then come to be recognised as being reactions to the observing creature’s own actions and reactions. That is to say, Creature A will react, not only intra-reactively by correlating B’s reactions to their shared environment, but A will also come to correlate its own reactions to B’s mutually intra-reactive reactions to Creature A’s initial reactions. Or again: A reacts to B’s reaction to what both B and A come to recognise as A’s initial reaction to the environment.

The process by which this is made possible is the very same process that allowed intra-reactive triangulation to (non-exclusively) resolve at least lower-order problems of content determination. To reiterate: in intra-reactive triangulation, the creature was seen to be able to move beyond simply reacting to its environment (which included the presence of conspecifics simply as further objects and events in this environment) by virtue of being able to correlate its conspecific’s relevantly similar reactions to the environment. The creature could thereby home in on the particular cause of both creatures’ reactions and, so, was subsequently afforded at least the potential to come to view both reactions as reactions to a particular cause. The parallel development here, then, is that A’s previous focus on an object or event in a shared environment comes instead to be shifted onto B’s reaction to A’s correlated reaction to the environment. Creature A now additionally responds to what is

131

seen as B’s response to A’s response to its environment. That is to say, the creature can now recursively correlate its own subsequent reactions to the other creature’s reactions to its own initial reactions.

Although the approach adopted herein of systematically moving through the various increasingly sophisticated stages of triangulation means that a scenario wherein some external object that is independent of the two creatures is the natural successor to intra-reactive triangulation, strictly speaking, the object or events in the environment that the creatures mutually respond to need only be their own reactions. That is to say, even if the two creatures in question were the only two objects in the universe, triangulation between the two – with each other’s reactions being the events to which both react – would still remain viable.

Furthermore, insofar as A must have already formed intra-reactive expectations regarding the likely cause of B’s reactions in order for creature B to then form interactive expectations regarding Creature A‘s reactions as reactions to B’s behaviour, it is likely (though not necessary), then, that Creature A would also form such meta-expectations of B as well. After all, given that Creature A is already able to correlate both its own and Creature B’s reactions to particular causes in their shared environment, it would be inexplicable for it to not also come to minimally recognise that B is now reacting also with expectations of A’s behaviour.

This stage in the increasingly sophisticated development of triangulation is properly interactive and no longer merely intra-reactive for the following two reasons:

i) In order for Creature A to form expectations regarding creature B’s expectations of the likely cause of A’s behaviour, both creatures must each react to the other. That is to say, unlike intra-reactive triangulation, where only one of the two conspecifics was needed to react to the other, here, it is necessary that both creatures must react (i.e. ‘interact’) to one another.

Admittedly, in a natural environment it is likely that both creatures would naturally interact with the other such that interactive triangulation might simultaneously take place alongside

132

of intra-reactive triangulation. Strictly speaking, however, it is necessarily the reciprocal relation to the other, found alone in interactive triangulation, which distinguishes it from intra-reaction. This is so because, as has been indicated above, inter-reactive triangulation can take place without mutual reactions occurring between the conspecifics.

ii) Although we speak of objects ‘interacting’ with other objects in brute physical ways, this next stage of triangulation is interactive in the fuller sense of the creature not merely reacting to its conspecific as simply another aspect of the environment, but rather as now reacting to the conspecific’s reaction to its own reaction. Or more straightforwardly: Creature A now reacts to B’s-reaction-to-A.

One apparent difficulty with this overall picture, however, is that, in contrast to intra-reactive triangulation, where both creatures were taken to be responding similarly to the objects and events in their shared environment, in the case of interactive triangulation, the creatures’ reactions will presumably be not at all similar to that of one another. One creature’s response to its conspecific might in fact be utterly dissimilar to the initial behaviour of its conspecific. For instance, one creature’s show of aggression is likely to be met by a radically different response – such as cowering – in the other. Given this is so, the question arises as to how they could come to correlate their reactions as they did in the non-interactive case of having similar responses to a shared, independent environment.

The answer to this question rests on the presumption that the intra-reactive recognition and correlation of the likely causes of the conspecific’s reactions are established. This, however, does not mean that the correlative process of intra-reactive triangulation must have taken place outside of or prior to any possibility for interaction. It is likely that intra-reaction (either initial or ongoing) can occur simultaneously with interaction – the one feeding off and informing the other. The claim is only that the correlative process in forming expectations of the likely causes of a conspecific’s reactions is necessary for interaction to take place, though not vice versa. By this account, then, interactive triangulation depends on intra-reactive triangulation. This should be clear by noticing the main ‘structural’ change between strictly intra-reactive triangulation and interactive triangulation was that the shared focus on the object or event in question was shifted to the creatures’ reactions to one another.

133

To return to the question of interactive correlations being formed in the context of divergent reactions, we can now see that, given intra-reactive recognition has taken (or is taking) place, then, so long as the respective reactions of each creature are similar in the sense that there is a sufficient degree of underlying invariance to be found between instances of the creatures’ disparate reactions and their shared cause, this will suffice for the creatures to be able to correlate their responses in such a way as to home in on the particular cause of their reactions. To illustrate: in the case of two lions, it doesn’t matter that when one creature growls, the other cowers, as long as these different behaviours occur jointly so that the lion may recognise that the conspecific is reacting to its behaviour.

An analogous point holds regarding the question of whether both creatures must access their shared environment via the same corresponding sensory modalities. This is particularly pertinent to interactive triangulation insofar as the creature will likely perceive the events of its own reactions kinaesthetically and proprioceptively, whereas its perception of the subsequent reactions of its conspecific will most likely be perceived primarily in the visual and auditory modalities. Most illustrative examples typically talk of only one sensory faculty being used (i.e. sight), but this tendency is most likely only for the rhetorical sake of clarity, not factual accuracy. What is relevant in both the intra-reactive and interactive cases, however, is not that the two creatures have the same sensory or perceptual experience of the environment, but rather that they come to home in on the shared cause of their reactions by what is similar between their reactions. That is to say, the relevant feature of the creature’s experiences is what is sufficiently invariant over time between the cause of its reactions and the reactions themselves.

III. NON-SEMANTIC COMMUNICATION

As was explained above, a creature engaged with its conspecific in interactive triangulation views the other creature’s reactions as responsive to its own initial reactions to the conspecific. This capacity potentially means that the creature now has at least the structural capacity to further come to correlate its own subsequent reactions as, likewise, reactions to the conspecific’s behaviour towards it. That is to say, it has at least the structural potential to form an awareness, both that it is responding to its conspecific’s behaviour, and that the

134

conspecific is responding to the creature’s behaviour. If this is right, then such an awareness potentially allows the creature – in fact, both creatures – to be able to act so as to bring about a desired reaction in their conspecific (although, precisely what this might amount to with respect to Davidson’s Argument from Objectivity will require subsequent analysis in Section V).

The focus in this section then, is to consider whether interactive triangulation straightforwardly occasions at least non-semantic communication, and, in attempting to answer this question, to determine what behavioural criteria might satisfactorily indicate that such communication is taking place.

It is important from the outset to keep in mind that interactive triangulation does not necessarily involve full-blown intentionality. The creatures need only themselves react and, in doing so, correlate the other’s relevantly similar reactions to the environment; the only change from intra-reactive triangulation is that this process of correlation now involves creature A seeing creature B’s reaction to the environment as inclusive of both A itself and A’s reaction to the broader environment also. That interactive triangulation need not necessarily be intentional is significant insofar as one does not want to beg the question in explaining the emergence of communication by attributing to the creatures in question mental content that already presupposes the capacity for semantic communication.

With this firmly in mind, interactive triangulation nonetheless makes it at least structurally possible that the one creature, having formed correlative expectations regarding its conspecific’s reactions to its own initial behaviour, might subsequently come to act in such a way as to bring about some desired behaviour regardless of the initial broad environmental conditions to which the creature typically reacts.

An empirical example of such triangular interactions is given by Tomasello (see [Tom08:25–6]) where he describes a chimpanzee infant that need only lightly touch its mother’s back for her to lower it in response so that the infant may climb aboard. Tomasello argues that this form of interaction involves what he calls intention–movement gestural signals that have been learnt (or at least acquired) ontogenetically as opposed to being inherited as fixed, phylogenetic intention–movement gestures. An example of the latter is the way in which the instinctual growl of a wolf gesturally communicates its intention to bite whether or not

135

the wolf eventually bites. Tomasello’s reasoning here is that the chimpanzee ‘touch-back’ gesture derives from the previous non-meaningful behaviour of the infant who would have initially had to physically pull down on the mother’s back in order to be able to climb up on it. The subsequent ‘touch-back’ gesture, then, is now communicatively meaningful as a result of the interactively triangulated correlative expectation formed by both mother and infant, in that this particular action both expresses the intention of the infant to climb aboard its mother’s back, and elicits the reaction of the mother lowering her back in recognition of the infant’s expectation.

It is significant to our consideration of the role played by triangular interaction that this particular gesture is unlikely to have arisen as a result of imitation (see [Tom08:25–6] for evidence against the hypothesis that the ‘touch–back’ gesture may have arisen through imitation). This is significant insofar as imitation need only be intra-reactive, requiring no need for mutual interaction between at least two creatures; whereas the distinct individual roles played by the infant and the mother in the case of the ‘touch-back’ gesture and response are necessarily interactive. As Tomasello explains of the latter [Tom08:26]: ‘Intention-movements are thus created as two interactants anticipate and so shape one another’s behaviour dyadically over repeated instances of the same interaction’.

Even putting aside the presumption that infant chimpanzees are not plausible candidates for intentionality (given their human counterparts are similarly discounted), nonetheless, the meta-cognitive awareness required for full-blown intentionality to take place does not appear to play any necessary part in the above-described ‘touch-back’ interaction. It is enough that, at first, the infant, by practical necessity, had grabbed and pulled itself onto its mother’s back in order to be physically able to climb aboard, and that the mother now recognising the desired outcome, begins to move in anticipation of the action, leading further to the infant changing its behaviour (most likely, gradually) from a grab to a touch or pat. Indeed, it is the nature of this last, final step that is the most telling, for we need only appeal to the laziness of the infant (and perhaps, equally, the comfort of the mother) to account for the emergence of this minimally communicative gesture. Tomasello himself emphasises the minimal demands of the interactive involved in the production of the ‘touch-back’ gesture when he writes [Tom08:26]:

‘...because of the way ritualisation [(or what we have been calling ‘correlation’)] works, these gestures are only “one-way” (not bidirectional) communicative

136

devices in the sense that the communicator and recipient each learn it in terms of their own role only—without knowing the role of the other’.

Thus, it appears that we have potentially identified yet another hierarchical position inside interactive triangulation itself. The ‘touch-back’ gesture, although necessarily interactive insofar as the gesture itself could not arise without the albeit distinct roles played by the two participants, need not require that the two participants be aware of the other’s role – at least not from the point of view of the other. This level of unidirectional interactive triangulation does not entail that interpretation of the other takes place, and, so, is insufficient for the purposes of Davidson’s Argument from Objectivity insofar as it stipulates that a creature must ‘know that the other person has the same object in mind’ [Dav92/01:121]. In this light, then, such minimal forms of non-semantic communication are insufficient to the purposes of Davidson’s Argument from Objectivity.

IV. NON-SALIENCE

What more might be required for interactive triangulation to shift from the situation described thus far – which is essentially two unidirectional intra-reactive triangles overlaid by fiat – to a mutual relationship wherein each creature might come to know of the thoughts of one another (as demanded by The Argument from Objectivity)? In answering this question, let us first ask whether it might be possible in the former context of minimally interactive triangulation, could a creature act to manipulate a desired reaction from the conspecific even without the stimulus being present? This is given that the creature has already formed expectations regarding the likely reaction that its conspecific will elicit if it itself were to behave as though it were reacting to an otherwise correlated stimulus in their shared environment. Is it then reasonable to suppose a creature might act so as to make its conspecific behave in a particular way, even though the stimulus that typically brings about such a reaction is absent? And, as the counterpart to this question, is it plausible to presume that a creature manipulated in such a manner, might act accordingly?

137

As we shall see in Section V, these questions turn out to be important to a proper understanding of what more might required of interactive triangulation for it to meet the demands of The Argument from Objectivity. But, even at this stage, the question remains of interest insofar as we hope to better understand the extent to which an interactively triangulating creature might act to communicate some perceptually non-salient aspect to its conspecific.

If we think back momentarily to the ‘touch-back’ example, we took it as reasonable to presume that the genesis of this action should have arisen from a non-communicative necessity for the infant to have initially pulled on its mother’s back in order to physically climb aboard. This is in keeping with the possibility of a creature coming to act in such a way as to elicit some expected behaviour in its conspecific even though the stimulus to which the behaviour is typically a response is not currently present. But, while it is in keeping with the possibility of such behavioural manipulation, it is insufficient to establish that the creature can (or will) behave as though a stimulus were present so as to manipulate their conspecific to respond accordingly. This is because, at least in the case of ‘touch-back’, although the infant does not actually grab on to clamber aboard, the stimulus by which the expectation had been formed by the mother is not absent as such, but has only been changed and modified in tandem with the forming of the expectation in question. As was suggested in the discussion earlier, the ‘touch-back’ gesture is likely to have come about through a gradual change from pulling, to grabbing, to perhaps even stroking, to simply patting. That is to say, the stimulus in question is still present in the case of ‘touch-back’; only the gesture has been altered. It is not, therefore, sufficient to answer our question.

Again, however, related work undertaken by Tomasello et al. can be invoked in order to garner at least a partial answer to the question of whether a creature might act so as to manipulate the behaviour of another despite the absence of a stimulus to which the expected behaviour is the typical reaction. Tomasello, in his investigation of chimpanzee gestures, presents a range of findings, which, when taken together, I claim constitute a plausible if not definitive response to the question at hand.

i) Firstly, Tomasello gives the example of chimpanzees pointing to objects only indirectly (or, consequentially) related to the object or event that they are in fact interested in. As Tomasello writes [Tom08:36]:

138

… some human-raised apes point to a locked door when they want access behind it, so that the human will open it for them […] (Gómez 1990). Another common observation, based on my own personal experience interacting with young chimpanzees, is that they will […] grab a human’s hand and put it in or on his pocket, and wait for a good result.

In these cases, the desired object (i.e. whatever treat that may or may not be hidden inside the human’s pocket) or event (e.g. gaining access to the space lying beyond the closed door) is not in the immediate perceptual field of the gesturing creature. The chimpanzee is able to point to indicate an object or event that it cannot see. So far, so good.

ii) Tomasello also presents the phenomenon of chimpanzees being able to follow the gestural pointing made by humans to objects that the chimpanzee itself is not able to perceive. As he writes [Tom08:38-9]:

Great apes follow the gaze direction of others, even to hidden locations behind barriers (Tomasello, Hare, and Agnetta 1999; Bräuer, Call, and Tomasello 2005). If a human points and looks toward some food that an ape currently does not see, and by following the pointing/looking the ape comes to see the food, she will go get it.

iii) Thirdly and finally, Tomasello presents a range of different experiments conducted mainly with chimpanzees, which show that they ‘understand that others have perceptions’ [Tom08:47]. Most strikingly, Tomasello presents the finding that ‘when chimpanzees compete with one another for food they take into account whether their competitor can see the contested food’ [Tom08:47].

Taking these three capacities of chimpanzees in concert, they have at least the potential to partially answer our question in the affirmative. That is to say, if a chimpanzee can: a) point to an object or event that is not currently present in either its own perceptual field or that of the recipient of the gesture; b) follow the gestural pointing of others to objects (and perhaps events), again, not its own perceptual field; and c) be aware of the perceptual fields of

139

others (and limitations thereof); then it doesn’t seem like much of a stretch that the chimpanzee might point to objects or events not in the perceptual field of its conspecific. When taken together, these points suggest that the answer is affirmative with respect to chimpanzees, even if the matter needs to be backed by further empirical investigation for it to be conclusive.

Can the example of the chimpanzees be generalised to all cases of at least minimally interactive triangulation? I can’t see why not. No further sophistication than what has already been characterised as two overlapping intra-reactive triangles is required for Creature A to gesturally point (for instance) to an object that Creature B cannot perceive and, yet, B may still be able to follow the direction of A’s pointing as a result of B’s pre-formed expectations regarding the typically correlated causes of similar instances of A’s pointing.

V. DECEPTION

The above, however, only provides a partial answer to the question of whether a creature might act in such a way to bring about an expected reaction of its conspecific even if the stimulus that typically causes similar such reactions were not present. The partiality of this answer occurs because, the question being phrased as such, so far we have only considered whether a creature might act so as to bring about some expected reaction in its conspecific even though the stimulus that typically causes such a reaction is absent to the conspecific’s perceptual field. I have not addressed, however, whether the creature might spontaneously gesture in such a way to bring about an expected range of behaviour in its conspecific, even if the stimulus that typically produced such a reaction was physically and not only perceptually absent from the environment of both creatures. That is to ask: have we any reason to expect that an interactively triangulating creature will be capable of acting to directly manipulate the behaviour of its conspecific, say, by spontaneously gesturing to deceive others?

Chadha, in her 2007 paper, ‘No Speech, Never Mind!’, draws on a range of empirical research, particularly focussing on instances of ‘tactical deception’ among primates to argue

140

against Davidson’s position that, without beliefs about beliefs (and, so, possession of the concept of objectivity) one cannot be said to have mental states. Her central contention is that ‘the fundamental ability to interpret experience and behaviour is a criterion for attributing mental states’ [Cha07:654]. She supports this by arguing that attribution of such interpretive abilities regarding experience and behaviour is necessary for accounting for behaviour such as acts of ‘tactical deception’ among animal species that nonetheless lack language. Perhaps the most striking case she cites is of a juvenile baboon acting as though it has been attacked by its older sibling who is in possession of food, so that the mother will chase away the sibling leaving the juvenile alone to enjoy the now abandoned food (see Cha07:652). Referring to many similar such cases, Chadha thereby claims [Cha07:252]:

…so many cases of tactical deception cannot be explained away as chance reinforcement learning. The deflationary alternative is easily ruled out because there are just too many recorded cases of deception.

But, while she disagrees with Davidson to this extent, she does, however, pull back from endorsing the further conclusion – controversially made by other researchers (Chadha here cites [Byr91&92]) – that ‘chimpanzees have meta-representational states and a theory of mind’ [Cha07:652]. Instead, Chadha acknowledges that in the presence of alternative accounts (Chadha here cites [Hey98] & [Ste03]) that are sceptical of the ‘anthropomorphizing explanation of tactical deception’ [Cha07:652] she adopts a strategy proposed by Susan Hurley [Hur03] in ‘circumventing the issue by talking in terms of conceptual abilities rather than concepts’ [Cha07:653]. In this way, Chadha claims we are able to attribute mental states to creatures without at the same time ascribing them meta-representational states, such as beliefs about beliefs or the higher-order concept of objectivity. As Chadha explains [Cha07:654]:

We may agree with Davidson that animals lack beliefs with fine-grained content, but disagree that they should therefore be deprived of a mental life altogether. I think we must respect the distinction between interpretive capacities of language users and nonlinguistic animals. When language users interpret, they assign relatively fine-grained content to mental states of language users. [...] to engage in linguistic interpretation, we must distinguish interpretations involving coreferring expressions, since linguistic interpretations are sensitive to opacity. Hence interpreters of language must have metalinguistic and metarepresentational

141

capacities. Nonlinguistic interpretations are based on causal generalizations that support behavioral predictions. Behavioral predictions can stand substitution salva veritate and thus they do not assign finegrained content to mental states of non-language-users. Hence, nonlinguistic interpreters do not require metalinguistic capacities.

All of this accords well with, and indeed, I think, is complemented by the previous account given of minimally interactive triangulation, along with the lessons learnt in respect to Davidson’s Argument from Surprise. The structural analysis of interactive triangulation explains how two overlapping intra-reactive triangles can allow creatures which are capable of correlating similar (and relevantly dissimilar) reactions in their conspecifics to come to engage in behaviour such as tactical deception, without thereby needing to attribute the capacity to ‘lie’ or the concept of objectivity.

Admittedly, these findings do not bode well for the role played by triangulation in Davidson’s Argument from Objectivity. Recall that Davidson claimed we could account for our concept of objectivity as arising out of intersubjectivity. Interactive triangulation, then, was then offered as a necessary condition for such intersubjectivity to occur. While it seems only commonsense to suppose that we must have been in contact with others to possibly know the contents of their thoughts, Davidson pushed further in claiming that the contact must not merely be intra-reactive, but mutually interactive. In our analysis of interactive triangulation thus far, however, mutual interactivity does not appear to require meta-representational states. And neither might we here call upon Davidson’s pessimism regarding the possibility of providing a linear account of the transition between non-semantic and semantic worlds. Rather, as Chadha’s account makes clear, we can readily account for complex social behavioural capacities of languageless creatures (such as acts of tactical deception) without recourse to ascribing meta-representational capacities to the creatures in question. That is to say, there is little to no question of having here hit upon the conjectured, ‘very deep conceptual difficulty or impossibility involved’ ‘in giving a full description of the emergence of thought’ [Dav97/01:128]. Rather, if the basic structure of interactive triangulation was hoped to locate (if not bridge) the point at which thought may emerge, then not only has it failed to do so, but, in so failing, it has not provided any theoretically substantive ground for Davidson’s claim that triangulation is a necessary condition either for intersubjectivity or for the supposed-complementary concept of objectivity.

142

VI. CONCLUSION

The shift from intra-reactive triangulation to interactive triangulation allows for far more complicated and purposeful reactions to occur between creatures. Interactive triangulation allows creatures to interact more sophisticatedly and efficiently than before, as was seen in the case of ‘touch-back’. It also allows creatures to coordinate their behaviour. To quote Davidson on this point again [Dav97/01:128]:

One sees triangulation in its simplest form in a school of fish, where each fish reacts almost instantaneously to the motions of the others. This is apparently a reaction that is wired in. A learned reaction can be observed in certain monkeys which make three distinguishable sounds depending on whether they see a snake, an eagle, or a lion approaching; the other monkeys, perhaps without seeing the threat themselves, react to the warning sounds in ways appropriate to the different dangers, by climbing trees, running, or hiding.

Although Davidson doesn’t emphasise the point (nor need he), it is interesting to note, that while vervet monkeys do learn specific reactions to the particular alarm calls of their conspecifics, they do not learn the alarm calls. As Tomasello explains: ‘In stark contrast to this picture of flexible comprehension, monkeys and apes do not learn to produce their vocal calls at all, and they have very little voluntary control over them.’ [Tom08:16] Moreover, the when and why of such alarm calls appears far more inflexible that one may perhaps assume. For instance, in most monkey species, the vocalisations are so closely tied to the emotional states of the creature that they will make the call even when none of their conspecifics (including their offspring) are present or in danger (see [Tom08:17–18] for an evolutionary account of this widespread phenomenon). Similarly, again as Tomasello reports [Tom08:19]: ‘chimpanzees give “pant-hoots” upon finding large amounts of food even if the whole group is already there and eating’. In light of these patterns, then, even when in certain species the need to call is more flexible than in others such that they may refrain from calling when they are alone, we have little reason to suppose that the particular members of that species (or any other) raise the alarm so as to communicate to their conspecifics, and not merely just because it is, as Tomasello puts it, ‘part of the genetically fixed adaptive specialization’ [Tom08:17]. (The argument here being that other animal

143

species, e.g. domestic chickens, equally refrain from alarm calling in such situations despite comparatively being unlikely candidates as communicators.)

Although Davidson himself offers no further explicit reason for dismissing learnt monkey calls as contenders for communicational behaviour, nonetheless, in light of the above, we may more readily sympathise with his following verdict regarding such phenomena [Dav97/01:128]:

On reflection we realize that the behavior of these primates, complex and purposeful as it is, cannot be due to propositional beliefs, desires, or intentions, nor does their mode of communication constitute a language.

Similarly, we can now get a better grasp on the significance of non-semantic communicational phenomena, such as the behavioural coordination that famously takes place between groups of chimpanzees while hunting monkeys. Tomasello introduces the example in what he argues to be the mistaken phraseology of shared goals and intentions [Tom08:173-4]:

In this account, one individual, called the driver, chases the prey in a certain direction, while others, so-called blockers, climb the trees and prevent the prey from changing direction – and an ambusher then moves in front of the prey, making an escape impossible.

Tomasello goes on to criticise [Tom08:174]:

...when the hunting event is described with this vocabulary of complementary roles, it appears to be a truly collaborative activity: complementary roles already imply that there is a joint goal, shared by the role-takers. But the question is whether this vocabulary is appropriate.

144

Tomasello instead proposes his own explanation of how such behavioural coordination could take place without the creatures sharing a goal in the sense of possessing a mutual awareness of one another’s mental states [Tom08:174]:

One chimpanzee begins by chasing the monkey, given that others are around (which he knows is necessary for success). Each other chimpanzee then goes to, in turn, the most opportune spatial position still available at any given moment in the emerging hunt. In this process, each participant is attempting to maximize its own chances at catching the prey, without any kind of prior joint plan or agreement on a joint goal or assignment of roles.

Tomasello also offers an extensive review of related empirical findings (see [Tom08:175–6]), the most striking of which, I feel, is the commonsense appeal that ‘it is almost unimaginable that two chimpanzees might spontaneously do something as simple as carry something heavy together or make a tool together’ [Tom08:176]. He concludes [Tom08:176]:

These results suggest that human-like collaborative activity—group activity with an intentional structure comprising both a joint goal and complementary roles—is something in which great apes do not participate.

All this fits well with our previous explication of interactive triangulation as structurally amounting to little more than two overlapping intra-reactive triangles, for, while interactive triangulation may well facilitate enhanced behavioural coordination between conspecifics, as it stands presently in its minimal form, we have little reason to suppose that it might occasion the full-blown mutually intersubjective awareness that Davidson maintains is necessary for possession of the concept of objectivity, as dictated by The Argument from Objectivity.

Finally, looking to the relevant literature on the capacities for deceptive behaviours among creatures without language, again, we find that our present analysis of minimally interactive triangulation both accords well with and sheds light on the empirically-researched

145

phenomena, but still fails to give us any account of our possession of the concept of objectivity.

146

PART THREE

* * *

Saying & Sharing

In Part One, Davidson’s a priori arguments for holding that language is necessary for possession of the concept of objectivity were found wanting. However, it was left as an open possibility that some story might still be told as to how language is nonetheless (and perhaps contingently) able to support and give rise to the concept of objectivity. In Part Two, we turned to Davidson’s more substantive arguments for triangulation as providing grounds for content determination sufficient to support the concept of objectivity. Again, however, not even properly interactive triangulation is seen to occasion full-blown intersubjective awareness, such as was demanded by The Argument from Objectivity. This perhaps, isn’t all that surprising, for even Davidson himself acknowledges that ‘triangulation does not answer these questions…’, and yet he continues, ‘but it helps fix some of the conditions under which propositional thought can emerge’ [Dav99b:731]. In this third part, then, we move beyond that which Davidson himself has laid out for us, and look to more recent engagement with, on the one hand, work directly inspired by Davidson, such as that of Ingar Brink and Robert Brandom, and, on the other hand, closely related empirical research by the likes of Michael Tomasello and his colleagues. The aim of this final part of the thesis,

147

then, is to move even closer to the empirical realm, without losing sight of Davidson’s initial starting conditions, in particular, his notion of triangulation.

In Chapter Five, I consider a proposal made by Ingar Brinck to the effect that joint attention is attributable only if the creatures in question can attend to non-salient aspects of their shared environment. Such a capacity, I argue, is explicable without recourse to meta-representation, entirely in terms of the parallel attention made possible via interactive triangulation. Moreover, empirical findings made by Tomasello reveal that, while both humans infants and chimpanzees are able to follow pointing gestures to non-salient aspects of their environment, chimpanzees are only able to do so in competitive contexts. Outside of competitive contexts, only humans are able to capitalise on altruistically informative gesturing. I briefly rehearse the evolutionary account Tomasello provides for the emergence of such cooperative capacities in humans, and suggest a modest amendment in keeping with Davidson’s triangulation picture.

In Chapter Six, I seek to explain human performance in non-competitive contexts of the object-choice task. To do so, I employ the same Carruthers-style analysis as was employed in Chapter One. Subsequently, I argue that the results of this analysis bear striking similarity with the general structure of Robert Brandom’s inferentialistic account of the concept of objectivity. I conclude by arguing that the former is a likely forerunner for the latter. That is, cooperation looks key to the both the emergence of language, intersubjectivity and possession of the concept of objectivity.

Taken together, Chapters Five and Six form a complementary pair. On the one hand, Tomasello’s amended evolutionary accounts for how humans came to be able to gesture altruistically. And, on the other, the Carruthers-style analysis explains how we are able to capitalise on such gesturing. Together, these form an account of how non-competitive communication is possible such that, eventually, language, intersubjective awareness and the concept of objectivity may emerge. As I have said throughout, this isn’t meant to be an account of the necessary conditions, but it is plausibly how we humans manage it, contingently or otherwise.

148

CHAPTER FIVE

* * *

Cooperation

Imitation of gesture is older than language, and goes on involuntarily even now, when the language of gesture is universally suppressed, and the educated are taught to control their muscles. The imitation of gesture is so strong that we cannot watch a face in movement without the innervation of our own face (one can observe that feigned yawning will evoke natural yawning in the man who observes it). The imitated gesture led the imitator back to the sensation expressed by the gesture in the body or face of the one being imitated. This is how we learned to understand one another…

Friedrich Nietzsche. Human, All Too Human §216

I. INTRODUCTION

In the last chapter, Davidson’s Argument from Objectivity was critiqued. There, we moved from the intra-reactive triangulation discussed in Chapter Three, to full-blown, mutually interactive triangulation. Interactive triangulation was seen to bring with it a number of likely advancements over that of intra-reactive triangulation, including finer content determination, potential for deceptive behaviour, and at least the structural potential for a creature to be able to gesture to objects not in the perceptual view of a conspecific.

149

However, it was concluded on the basis of discussion concerning the difference between deception and lying, that interactive triangulation by itself was unable to fulfil the minimal requirements demanded of Davidson’s original argument. Specifically, the structure of interactive triangulation was not sufficient to break into ‘the intentional circle’ of semantic content.

In Section II of this chapter, I will rehearse a suggestion made by Ingar Brinck [Bri04] that even if the basic structural framework of interactive triangulation cannot enable us to bridge the divide between non-semantic reactions and semantic communication, nevertheless, joint attention – by which two creatures attend both to a particular object in their shared environment and to one another also – might bring us that little bit closer to the precipice. Brinck’s claim is that joint attention, although far from constituting semantic communication proper, can nonetheless provide some reason for giving credence to the claim that properly interactive triangulation is necessary for communication (and so, language).

In Section III, I evaluate Brinck’s proposal in light of the work of Michael Tomasello. A number of empirical experiments undertaken by Tomasello et al. suggest the behavioural indicators that Brinck holds to be indicative of a creature’s capacity to know what its conspecific has in mind (such as sustained eye-contact) are undermined given slight changes in the experimental conditions. Specifically, if a gesture is made outside of the context of competitive interaction, then chimpanzees (at least) cannot make effective use of structurally similar gestures that they are otherwise able to in competitive contexts. Moreover, drawing on the analysis undertaken in the previous chapter concerning the structural capacity of non-semantic interactive triangulation as being capable of supporting a creature’s capacity to both gesture and make use of gestures to non-salient objects, the emphasis Brinck puts on joint attention as the tipping point of semantic content is seen to falter.

In light of the above, Section IV introduces Tomasello’s own suggestion that it is cooperation that is required for two creatures to make the leap from non-semantic interaction to at least the minimal beginnings of communication. Here, I point out the strengths of such an approach in, if not overcoming, at least partially circumventing, the potential conceptual obstacles of providing an account of semantic content as arising from a non-semantic interactions.

150

In Section V, I seek to bolster Tomasello’s evolutionary account by proposing a suggestion of how the picture provided by Davidsonian interactive triangulation could plausibly support Tomasello’s evolutionary story. Specifically, I argue that altruistic gesturing might have first emerged as a result of unintentional imitative behaviours existing between two or more conspecifics, such that one might then be able to capitalise on those of the other, eventually leading to cooperative social contexts actively encouraging individuals to act in such altruistically informative ways.

II. JOINT ATTENTION

In this section, I introduce and evaluate Ingar Brinck’s proposal that the phenomenon of joint attention shared between two creatures may serve to bring us closer to the point of breaking into the intentional circle. As Brinck puts it [Bri04:12]: ‘I suggest that we approach the problem of how to get interpretation started by thinking of triangulation in terms of joint attention’. The difficulty with Davidson’s original triangulation argument, as Brinck sees it, is that it is unable to account for how creatures might become able to identify a common cause on the basis of repeated observations given that, at this pre-intentional stage, the contextual features shared between each instance must be more coarse-grained than the object in question. Instead, Brinck proposes that the behavioural phenomenon of joint attention may provide a wedge into the intentional circle. Specifically, Brinck claims that to qualify for joint attention, creatures must demonstrate explicitly convergent interest in some particular object, firstly, by way of checking the continued, shared direction of the attention of its conspecific, and secondly, they must be capable of doing so in the context of attention made by its conspecific towards some non-salient entity. In response, I look to particular empirical findings of Tomasello et al., which show that, although chimpanzee behaviour satisfies Brinck’s criteria, nonetheless, they remarkably cannot make effective use of structurally identical gesturing when the experimental conditions are changed to a cooperative rather than competitive context. In this light, Brinck’s model is compared with the account developed in the previous chapter of non-intentional interactive triangulation. I conclude that insofar as the structure of interactive triangulation can readily account for the behavioural phenomena described by Brinck’s joint attention, her criteria falls short, such

151

that we are advised to seriously consider Tomasello’s alternative account, which we shall do in the subsequent sections.

1. Brinck’s Analysis

In her 2004 article, ‘Joint Attention, Triangulation and Radical Interpretation: A Problem and its Solution’, Ingar Brinck first criticises Davidson’s Arguments from Content Determination & Objectivity as being unable to break into the intentional circle without begging the question, before then turning to argue that the phenomenon of joint attention may provide a wedge, if not a bridge within. Although in Chapters Three and Four of this work, independent reasons were given for supposing that neither intra-reactive nor interactive triangulation were by themselves sufficient to suppose that either creature had broken into the intentional circle, nonetheless it is necessary that we here briefly outline Brinck’s criticism of Davidson’s argument (irrespective of whether it is taken to be compelling or not) so as to better appreciate the significance of her proposed solution and its subsequent limitations.

Brinck argues that triangulation between two creatures faces a dilemma. As she puts it [Bri04:189]: ‘Either triangulation is pre-cognitive and cannot explain how speakers converge on a common cause, or it requires higher-order thought, and cannot provide for language entry, nor individuate the content of individual beliefs and utterances’. Recall from Chapter 3, that essentially the same objection was raised by Peter Pagin [Pag01:199–212] and Kathryn Glüer [Glü01:53–75]. At that stage in our development of the structural hierarchy of the triangular relationship, we deemed it appropriate to concede Davidson’s insistence that non-intentional correlative expectations formed intra-reactively were all that were being posited (despite our eventually coming to conclude that inter-reactive triangulation could not by itself definitively solve the problem of content determination). Brinck, however, goes beyond this more general objection, and instead paves the way to a potential solution. Her account rests on the familiar two-pronged criticism that, on the one hand, Davidson’s account of Content Determination requires that ‘speaker and interpreter converge on a common cause by observing each other’s responses to external situations’ [Bri04:189], whereas, on the other hand, his Argument from Objectivity acknowledges that ‘the similarity judgements that the speakers as interpreters make about each other are about

152

their respective ways of conceiving of the world, and not just about behaviour’ [Bri04:189].15 But, Brinck continues, given that it would be question-begging to suppose the latter capacity constitutes judgements proper, then without such capacities the creature will be unable to recognise similarly repeated particular instances of the common cause as common. As Brinck puts it [Bri04:191]: ‘… to at all be able to make such repeated observations, the interpreter must hold something constant across contexts of observation. There must be some way of telling that the observations are of a similar kind of behaviour’. And yet, as Brinck submits [Bri04:190]: ‘Intentional actions cannot be identified by their behavioural properties alone, taken in isolation from the feature that triggered them’. The difficulty then is couched by Brinck in terms of the coarseness of the feature common to the context in contrast with that of the pertinent aspect that deems the cause as common [Bri04:141]:

The contextual feature that is held constant across contexts will be more coarsely individuated than an aspect of a cause is. The contextual feature is not conceptually individuated, and cannot be so, until interpretation is well on its way. During the initial phase of radical interpretation, the interpreter can access only less fine-grained elements of the context of utterance.

2. Brinck’s Solution

The problem thus laid out, Brinck suggests that the behavioural phenomena of joint attention, whereby ‘attention is jointly focused on objects that are identified both perceptually and intersubjectively’, may be called upon, both, to ‘show it is possible to break into the intentional circle’ [Bri04:195], and to ‘pave the way for language entry and get interpretation going’ [Bri04:195].

Brinck lists three conditions that are commonly taken as needed for joint attention [Bri04:196]:15 Also perhaps relevant to Brinck’s objection is Claudine Verheggen’s insistence that Davidson’s Arguments from Content Determination and Objectivity should be considered singularly as two sides of the same coin (see [Ver97]).

153

Joint attention will occur iff speaker and hearer:

(i) attend to each other’s states of attention;

(ii) make attention contact [(such as pointing)]; and

(iii) alternate [their] gaze between each other and the object.

Brinck argues, however, that, as it stands, this account of joint attention is unable to overcome the problems facing Davidson’s triangulation arguments. This is because, as she explains it, one of the reasons it is required that a creature be sensitive to the contextual features which make it possible to recognise repeated occurrences of a cause as common, is that the salient feature of the cause in question may not (or likely will not) be salient in each observed instance. As Brinck puts it [Bri04:193]:

… the interpreter may not know which difference or contrast to look for in order to pick out the triggering feature. For one thing, there may not be any particularly salient item per se. There may be several competing items in the context, or no especially salient items present at all.

In this light then, Brinck proposes that in order to qualify as joint attention sufficient to resolve the dilemma faced by Davidson’s triangulation arguments, the first condition stated above needs to be made more demanding so as to ‘explain how the process of joint attention can begin in the absence of salient entities’ [Bri04:14]. Brinck’s explication of the necessary condition that must hold for joint attention if it is to possibly lead to semantic communication is thus stated as follows [Bri04:197 (italics added)]:

Joint attention will occur iff speaker and hearer:

(i´) attend to each other’s states of attention as states that are directed at objects in the environment and by being so directed are about the objects, and, moreover, attend to these states also when they are directed at non-salient entities.

154

Brinck exemplifies what this modified condition should look like if it were to be met, by way of a hypothetical, yet, to her argument, canonical illustration [Bri04:197]:

Let me illustrate my point. Suppose that the speaker has seen a third person hide something in a box. She wants to indicate the box to the hearer. The box stands on the floor among several other boxes. All the boxes look alike. There is no evidence that one of them contains something of value. In this case, the speaker would only be able to make the hearer take part in the process of joint attention if the following were the case. The hearer should be sensitive to the speaker’s manifest interest in one of the boxes. She should moreover react to the speaker’s interest as if it were a motive for engaging in attention contact with her and for following the line of her gaze to the box.

The point that Brinck aims to make by this hypothetical scenario is to illustrate that, even in the absence of any salient entity (such as the desired object contained within the opaque box), we only have sufficient grounds to suppose that the ‘hearer’ is capable of joint attention (for it is already presupposed that at least the speaker is capable insofar as she ‘wants to indicate the box to the hearer’ [Bri04:197]) if they are capable of engaging with their conspecific’s pointing gesturing, which, in the case of Brinck’s hypothetical, amounts to following the line of the speaker’s gaze in a particular way. If the hearer is unable to appropriately attend to the speaker’s gestural invitation to engage its conspecific’s attention toward the otherwise indistinguishable box, then, we have insufficient grounds to suppose that joint attention in other cases where the salient is present. Put simply, Brinck claims that in order to attribute joint attention (and not mere ‘parallel attention’, where two or more creatures attend to the same object, yet without attending to the other’s shared attention), the creature must also be capable of attending to non-salient aspects of the environment. If they are unable, then instances where the entity is salient are best regarded as only moments of parallel attention.

To be clear: the issue cannot rightly be that in the presence of some clearly salient entity, we could not possibly determine whether the creature is responding to, either, the speaker’s own interest in (or indicating to) the desired object, or the object itself. As was established in Chapter Three, there is no reason to suppose that given sufficiently sensitive observations, we could not in principle determine what was the initial cause of the hearer’s subsequent parallel attention. Rather, the point must lie in the underlying explanation of the hearer’s

155

capacity or tendency to sustain their attention parallel to that of the speaker even though no salient entity is seen by the hearer to motivate the speaker’s attention. Brinck’s suggestion, then, is that only by attributing joint attention can such behaviour be sufficiently accounted for. This is because, as was outlined above, Brinck holds that for the creature to have the capacity to recognise similarities such that allow it to enter into the intentional circle, it must be sensitive to the broader contextual features of intentional behaviour so it may be able to come to recognise the cause of repeated observations as having a common cause (and so as properly, being the cause of said observations). If, by virtue of engaging in prolonged joint attention, a creature is able to recognise the intention of its conspecific in the absence of any salient entity, then we have good reason for attributing to it the requisite contextual sensitivity needed for subsequent membership in the intentional circle of semantic meaning and eventual language use.

III. CONTEXTS

In this section, I evaluate Brinck’s proposal that joint attention must be possible even in the absence of any salient entities if it is to provide a wedge into the intentional circle. In doing so, I again turn to the work of Michael Tomasello, not only to draw on the related findings of his and his various research partners’ empirical observations, but also to consider the theoretical interpretations made thereof as potentially being better candidates for finding a wedge that will allow us to non-question-beggingly enter from non-intentional behaviours into the intentional circle.

By way of introducing the evaluation of Brinck’s proposed criterion regarding the importance of joint attention in the absence of salient entities, it is worthwhile pausing to emphasise that her hypothetical scenario is just that – hypothetical. Of course, humans already well ensconced within the intentional circle can readily accomplish feats as are exemplified by Brinck’s archetypal scenario. However, it remains an open question at this stage of our inquiry as to whether any non-language users do actually behave in ways that accord with Brinck’s archetype. As I hope to show, while they do, they also display

156

remarkable divergences in behaviour in such a way that critically undermines the supposed force of Brinck’s proposal.

* * *

In answering the question of whether the supposed core of Brinck’s hypothetical scenario is ever enacted outside what is clearly the intentional circle of human language users, Tomasello provides what seems to be a straightforward answer. He writes, ‘Great apes follow the gaze direction of others, even to hidden locations behind barriers’ [Tom08:38–9]. This comes as no great surprise given the previously related phenomena that was drawn upon to conjecture that interactive triangulation might give us some reason to suppose creatures will thereby be capable of themselves acting so as to manipulate the behavior of their conspecific even in the absence of the correlated objects and events typically associated with the desired behaviour. (Recall from Chapter Four that chimpanzees can point to doors that they would like to be unlocked or to pockets where they presume treats may be contained.) But, at that point, the possibility that joint attention was thereby occurring was not a live contender in light of the hypothesis that the creature could straightforwardly draw upon the non-intentional correlations of mutually interactive behaviour over time. As it stood in our progressive development of triangulation, the capacity to manipulate the behaviour of a conspecific – even in the absence (observational or actual) of any salient entity – could be accounted for straightforwardly by expectations formed by correlative sensitivities to causal responses and reactions only. In this related context, there was no reason to posit sensitivities to intentional states, such as Brinck supposes is required for the possibility of joint attention in the absence of salient entities.

Given that we were correct in this analysis, might we not expect that Brinck’s hypothetical scenario also be explicable in straightforwardly non-intentional terms of causal correlative responses to previously experienced objects and events (i.e. the relevantly similar reactions of conspecifics)? The main difference between the two scenarios is that ours asked whether interactive triangulation is sufficient to support and account for creatures behaving so as to manipulate and bring about desired responses in their conspecifics even if the typical cause of either’s behaviour is absent. That is to say, we were interested in non-linguistic directives. The focus of Brinck’s scenario, on the other hand, is to establish the capacity for the hearer to recognise relevant intentional contextual features.

157

However, there seems to be little reason why, if the former scenario can be accounted for without recourse to the capacity for joint attention, the latter might not also be similarly accounted for. After all, the correlatively formed expectations might lead a creature to gesture even in the absence of any salient entity so as to bring about a desired behaviour in its conspecific – recall the tactical deception of the baboon pretending that its sibling had attacked it so as to invite the mother’s protection and, so, gain the sibling’s food. It is not obvious, then, what more might be required on the part of the hearer to similarly respond on the basis of its own correlatively formed expectation when its conspecific gestures in a particular yet typical way, even though no salient entity may be observable to the hearer. If this is correct, then there is little reason to suppose that joint attention – and not merely, parallel attention – has occurred simply by virtue of the hearer following and checking the directive gestures of the speaker.

Indeed, matters appear much worse for Brinck’s argument insofar as Tomasello reports that, while chimpanzees are capable of following the gestural directives of creatures that they view as competitors, they are seemingly unable to capitalise on the same gestures outside of competitive contexts. This suggests that there might be more to joint attention than the ability to attend to non-salient features of a shared environment. To quote Tomasello at length [Tom08:39–41]:

If a human points and looks toward some food that an ape currently does not see, and by following the pointing/looking the ape comes to see the food, she will go get it. In this sense, one could say that the ape understood the intention behind the human’s attention-directing gesture in this simple situation. But a seemingly minor change in this procedure leads to a drastically different result – which might lead us to reassess the simpler situation. Tomasello, Call, and Gluckman (1997) introduced apes to a game in which one human, the hider, hid food in one of three buckets and a second human, the helper, helped them find to it – what has been called the ‘object choice’ task. Apes knew from previous experience that there was only one piece of food hidden, and they would get only one choice. In the key experimental condition, the hider hid the food from the ape while the helper peeked, and then the helper simply pointed informatively for the ape to the bucket in which the food was hidden. Astoundingly, apes then chose buckets randomly, even though they were highly motivated to find the food on almost every trial.

158

Quite often an ape followed the helpers’ pointing and looking to the correct bucket but then did not choose it. This means that following the directionality of the point was not the problem; they just did not seem to understand its meaning, its relevance to their search for the food. It is as if the apes said to themselves ‘OK. There’s a bucket. So what? Now where’s the food?’ Human infants perform well in this seemingly trivial task by 14 months of age, mostly before language (Behne, Carpenter, and Tomasello 2005). Task failures may be explained in an unlimited number of ways. But a follow-up study constrains the possibilities considerably. Hare and Tomasello (2004) conducted a competitive version of the basic object-choice task. Chimpanzees participated in two experimental conditions. One condition, the cooperative condition, was identical to the basic task, and so, not surprisingly, the results were identical as well: despite following the point to the correct bucket the apes chose randomly. In the other, competitive condition, however, a human began in the warmup session by competing with the chimpanzee for food, and then in the experimental session attempted to continue competing. Specifically, without looking to the ape in any way, the human reached toward the correct bucket, but due to the physical constraints in the situation (her arm would not go very far because the hole in the Plexiglas was not large enough), was unable to reach it. When the buckets were now pushed to the ape (by another experimenter), she now knew where the food was! Even though the superficial behavior in the two experimental conditions was highly similar – arm extended toward correct bucket – the apes’ understanding of the humans’ behavior was seemingly very different. They were thus able to infer: she wants to get into that bucket for herself; therefore, there must be something good in there. But they still were not able to infer: she wants me to know that the food is in the bucket.

Although both humans and chimpanzees can both readily capitalise on pointing gestures to even the hidden location of objects (typically food, in the case of chimpanzees), slight changes in the contextual features of an experimental scenario can lead to dramatic and rather surprising results. Specifically, if the warm-up session that sets the context for the experiment is competitive in nature (i.e. the human participant is established by repeated example to be competing for the same food as the chimpanzee), then, even when the desired food is hidden and the chimpanzee is given no reason for supposing that the human competitor knows where it is located, if the human gestures towards some relevantly non-salient object (such as an upturned bucket), then the chimpanzee will unhesitatingly choose that object, presumably in the hope of securing the prize before the human. If, however, the

159

warm-up session does not take place, then, although the human makes physically identical gesturing signals towards where the food is hidden, the chimpanzee does not heed the human’s signalling and will instead search for the food by examining the buckets randomly.

Bringing these findings to bear on Brinck’s position, there are two main points to consider. The first is the extent to which Tomasello’s findings support the possibility that chimpanzees are capable of shared intentionality in general. Failing the first, the second is whether chimpanzees might nonetheless be capable of joint attention though only in competitive contexts. Each shall be treated in turn in the following two sub-sections.

1. Shared Intentionality?

In answer to our earlier question of whether Brinck’s hypothetical scenario is ever actually enacted by creatures not already presumed to belong within the intentional circle, the capacity of chimpanzees to be sensitive to the contextual features at least of being in competition with another creature seems to be borne out. Remarkably, chimpanzees are seemingly sensitive to the contextualising features of the warm-up session where the human counterpart is shown to be in direct competition for the desired food, such that in subsequent sessions the chimpanzee then expects the human’s hindered attempt to reach for the bucket containing the food (being structurally similar to a directive pointing gesture) likely indicates the location of the hidden food. This scenario is structurally equivalent to the hypothetical scenario of Brinck’s archetype, such that, if Brinck were right, we should have some reason to suppose chimpanzee performance promises to illuminate (if not manifest) the missing link between non-intentional causal interactive behaviours and semantic communication. However, the possibility raised above of explaining such parallel attentional behaviours without recourse to intentional states (let alone of shared intentionality) demands that we must consider whether Brinck’s account of the sufficient capacities needed for joint attention is thereby seen to be limited from what Tomasello has said regarding the chimpanzees’ inability to make use of structurally identical gestures in a context where the chimpanzee has no reason to suppose that they are in competition with their human counterpart.

160

The findings of Tomasello and his colleagues suggest chimpanzees are unable to make use of structurally identical gestures if the contextual features are changed from competitive to non-competitive interactions. On the face of it, this may actually appear to support Brinck’s suggestion that parallel attention in the absence of any salient entity counts as joint attention. It would seem that if indeed it is sensitivity to the contextual features of the competitive warm-up session that led the chimpanzees to being able to make use of the subsequent frustrated reaching of their human counterparts, then Brinck appears correct in claiming that joint attention in the context of non-salient entities suggests sensitivity to the contextual features requisite for shared intentional communication. The relevant contextual features in this case would be that the chimpanzees were initially led to understand that food was hidden under one of the buckets and that they were in competition with their human counterpart to find it first. On this account, without such initial contextualisation, features such as the structurally identical gesturing would mean nothing to the chimpanzees such as when performed in a non-competitive (or perhaps any other) context. Presumably, this is the account that Brinck would likely have us endorse, and, taken by itself, it is fairly compelling. However, the force of Brinck’s proposed archetypal hypothetical scenario is undermined insofar as Tomasello further argues that the competitive context is the only shared intentional, coarse-grained state that chimpanzees are capable of being sensitive to. As Tomasello explains [Tom08:41]:

One reasonable hypothesis [regarding the failure of chimpanzees to make use of directive gesturing outside of the competitive coarse-grained context], then, is that apes simply do not understand that the human is communicating altruistically in order to help them toward their goals. That is, they themselves communicate intentionally only to request things imperatively, and so they only understand others’ gestures when they are imperative requests as well – otherwise they are simply mystified as to what the gesticulating is all about.

In this light, a question arises of whether we might be able to provide a non-meta-representational account of the ability to follow a speaker’s directive gestures to non-salient entities. If such a low-level account were to be equally applicable in the case of Tomasello’s findings that chimpanzees are seemingly able to comprehend the competitive context in which a human’s hindered reach is seen to indicate the location of otherwise hidden food, then we only have to explain the chimpanzees’ capacity to shift from the explicitly competitive context of the warm-up sessions to the implicitly identical context of the

161

experiment. But this we can readily do by virtue of the capacity already presupposed that chimpanzees are capable of forming correlative expectations of behaviours based on past experiences. Admittedly, when we first invoked such capacities in the context of both intra-reactive and interactive triangulation, we had something far more accretional in mind, but the capacity seemingly demonstrated by the chimpanzees to form future expectations on the basis of warm-up sessions, although perhaps more distinct, are structurally no more sophisticated. All that is required is that the chimpanzee forms an expectation that the explicit competitive behaviour displayed by its human counterpart in the warm-up session still holds in the experimental session that follows afterwards.

It seems that we are unable to establish that such sensitivity to contextualising features confidently demonstrates a level of shared intentionality (or, for that matter, meta-representation) among chimpanzees. Indeed, the plausibility of being able to explain how chimpanzees can seemingly be sensitive to coarse-grained contextual features without needing to invoke any hint of the notion of shared intentionality (let alone meta-representation), coupled with the evidence suggesting that the competitive context is the only context in which they are able to make use of directive gestures, greatly undermines the confidence we might have in supposing that Brinck’s test is adequate. This, however, is not necessarily to show contrariwise that chimpanzees are necessarily incapable of shared intentionality either; only that Brinck’s proposed test is not without significant challenges or limitations.

2. Only Competitive Contexts?

Is it plausible that chimpanzees are capable of joint attention if only in competitive contexts? Tomasello’s identification of the non-competitive context that chimpanzees are seemingly incapable of comprehending, may, on the one hand, actually appear to support this notion. After all, chimpanzees are able to follow the direction of a pointing gesture, even if they are only able to make effective use of it in competitive contexts. I argue that the opposite interpretation is more likely. That is, chimpanzee performance in non-competitive contexts undermines any claim that joint attention proper occurs in competitive ones. Essentially, this is because the capacity of humans already acknowledged as capable of joint attention in both competitive and non-competitive contexts weighs against any claim that parallel

162

attention in an exclusively competitive context provides sufficient evidence to attribute joint attention.

Humans normally are readily able to participate in joint attention, and from a relatively young age. Tomasello suggests that this typically occurs as the child approaches the age of one (see [Tom08:144]). Humans are also readily able to demonstrate sensitivities to and competencies in a wide range of contexts, including competitive ones. Given, then, that adult humans belong to the full-blown intentional circle and are capable, even as infants, of capitalising on their conspecific’s gestures in both competitive and altruistic contexts, this seemingly undermines Brinck’s proposed criteria as sufficient for attributing joint attention. While chimpanzees certainly appear to ostensibly qualify in meeting Brinck’s proposed criteria for joint attention, their surprising inability to capitalise on gestures in non-competitive contexts suggest that perhaps more is involved. As Tomasello puts it [Tom08:177 (italics added)]:

Apes understand that others have goals and perceptions and how these relate to one another in intentional action, perhaps even rational action. So this is not the reason they do not collaborate in human-like ways. Rather, as might be expected, we believe that whereas apes understand what the other is doing as an individual intentional agent, they have neither the skills nor the motivations to form with others joint goals and joint attention or otherwise participate with others in shared intentionality.

Here Tomasello rejects an alternate hypothesis ­­­– aired by Povinelli and O’Neil [Pov00] – that the reason why chimpanzees do not (or cannot) make use of gestures outside of a competitive context is because they ‘they do not understand the goals and perceptions of their partner as an individual actor in the situation’ [Tom08:176]. Rather, Tomasello invokes a wide range of subsequent empirical research (see [Tom08:45–9]) to establish that ‘[a]pes understand that others have goals and perceptions and how these relate to one another in intentional action, perhaps even rational action’ [Tom08:177]. In this light, Tomasello asserts that the inability to cooperate is the precise reason why chimpanzees fail to make use of gestures made in a cooperative context. As he puts it himself [Tom14:52]:

163

They make the competitive inference, ‘He wants [what’s] in that bucket; therefore the food must be in there’, but they do not make the cooperative inference, ‘He wants me to know that the food is in that bucket’.

Tomasello’s point here is twofold. Firstly, he invokes this evidence to justify the theory that chimpanzees are precisely in this way limited. This is of particular interest to us presently as we are arguing that Brinck’s account of joint attention either fails to establish the minimal conditions required for joint attention (as suggested by the argument given in the previous sub-section), or that it doesn’t adequately provide a sought-after wedge into the intentional circle that was argued for. With respect to the latter, Tomasello’s point here is that, contra Davidson’s Argument from Objectivity, while chimpanzees are aware of others as intentional (and perhaps even rational) agents, and are also capable of parallel attention, nonetheless, the capacity for recognising (or being able to make use of) cooperatively shared intentionality is, it appears, a significant watermark for the subsequent emergence of full-blown intentionality. Moreover, as we shall see in the following section and beyond, this limitation in chimpanzees is not only of significance in pinpointing the precise limitations of the still-remarkable capacities of chimpanzees, but also Tomasello in this way seeks to subsequently lend support to the further claim that, by contrast, humans are distinctly capable in precisely this manner.

IV. COOPERATION

Tomasello argues that the capacity for cooperation is a significant point of distinction between humans and non-human animals (specifically great apes, but especially our closest evolutionary relative, the chimpanzee). Specifically, he claims that a plausible story can be given as to how, both, the capacity for altruistically informative gesturing and the ability to capitalise on such non-competitive pointing are essential for ‘objectivity’ – i.e. a ‘generic, agent-neutral perspective possible by any rational person’ [Tom14:121]). The far-ranging account that Tomasello provides in this regard moves from joint intentionality to collective intentionality and culture, including ‘conventions, norms, and institutions.’ [Tom14:5]. He outlines the evolutionary development from minimalistic cooperative interactions of kind evidenced in the object-choice task to epistemic objectivity as follows [Tom14:5]:

164

As part of this process, cooperative communication became conventionalized linguistic communication. In the context of cooperative argumentation in group decision making, linguistic conventions could be used to justify and make explicit one’s reasons for an assertion within the framework of the group’s norms of rationality. This meant that individuals now could reason “objectively” from the group’s agent-neutral point of view (“from nowhere”).

While such an ambitious account is certainly sympathetic to the general concerns of this thesis, our focus herein will remain far narrower and more constrained. This is because, both, the path laid out so far by Davidson (along with Brinck), and the specific task we have set ourselves (especially in light of the challenges outlined in Section II.3 of The Introduction), recommends that we persist in looking for an entering wedge into the intentional circle. Specifically, in the remainder of this chapter, we will be concerned to explain how altruistically informative gesturing arose, and, in the following Chapter Six, to account for how humans are able to capitalise on such gesturing such that, subsequently, fully semantic communication and possession of the concept of objectivity may emerge.

In this section, I will rehearse Tomasello’s account of how cooperative gesturing first emerges. In the subsequent Section IV, I propose a small, yet, I believe, significant amendment to Tomasello’s phylogenetic account of how the capacity for informative gesturing arose in humans.

* * *

Tomasello’s strategy for accounting for the emergence of altruistically informative gesturing is threefold. Firstly, he seeks to establish that cooperation underlies human communication. Secondly, he works to demonstrate that in human infants, the capacity for cooperation exists ontogenetically prior to language acquisition, such that we thereby have further reason to suppose that it underlies human communication. Finally, Tomasello seeks to show that, although they come very close, chimpanzees fall short of such cooperative capacities. I shall consider these in reverse order.

165

To his credit, Tomasello concerns himself as much with celebrating the remarkable capacities of our nearest evolutionary relations, as he does in marking their relative limitations compared to humans. Specifically, he draws on numerous empirical studies to make the following claims regarding the capacities of great apes [Tom08:48]:

… apes understand others in terms of their goals and perceptions and how these work to determine behavioral decisions, that is, they understand others as intentional, perhaps even rational, agents. Based on this understanding, they can engage in the kinds of practical reasoning that underlie flexible, strategic social interaction and communication – for example, determining what the other wants, the reason he wants it, and what he is likely to do next.

While the reader is invited to consider for themselves the numerous other empirical studies invoked by Tomasello in support of his conclusions, permit me to rehearse one example I find particularly striking: When a chimpanzee sees that a human’s efforts to give it food is frustrated by circumstances beyond their control, then they respond patiently, but if the human arbitrarily postpones the delivery of food, say, by adding in some non-rational step into the process, then the chimpanzee becomes frustrated (see [Tom08:45]). Prima facie, this finding appears to support much of the previous quoted summary of chimpanzee capacities; particularly, that they are able to understand others as having intentional goals underwritten by some degree of rational agency.

And yet, despite their undeniably remarkable capacities, chimpanzees do not appear to be capable of even capitalising on altruistically informative gesturing, let alone gesturing for themselves in any way but imperatively. As Tomasello puts it, ‘When apes communicate with one another in other contexts, it is always directive’ [Tom14:51]. By comparison, Tomasello points to research that suggests even 12-month-old human infants will point to inform others of the location of some sought after object with no apparent desire to possess the object themselves (c.f. [Lis06&08]). And, indeed, the difference can seemingly be traced back to more ontogenetically formative behavioural patterns. Specifically, Tomasello highlights that, in a comparative study involving 18-month-old human infants along with full-grown chimpanzees and bonobos, all interacting with an adult human and some particular object, although all species ‘monitored the adult human’s behavior reasonably frequently […] human infants spent far more time than apes looking back and forth from object to adult, and their looks to the face of the adult were, on average, almost twice as

166

long as those of the apes’ [Tom08:179-80]. Tomasello explains this difference as follows [Tom08:181]:

Joint goals also structure joint attention, since acting with a partner toward a joint goal, with mutual understanding that we are doing this, quite naturally leads to mutual attention monitoring.

However, for what it’s worth, I would like to suggest that, in light of the discussion of the previous chapters concerning Davidsonian triangulation, perhaps it is the other way around. That is to suggest that it is the comparatively longer time that humans spend attending to the behaviours of their conspecifics which in turn supports parallel attention and parallel goals, such that the intersubjective intentionality of full-blown joint attention and goals might later arise. Although nothing of any great significance hangs on this suggestion being taken up, if it were so, it would seemingly be supported by the role played by Davidsonian interactive triangulation. On this reading, the human tendency to attend for longer periods of time to their conspecific’s attention might be more parsimoniously explained (and thus be more in keeping with Morgan’s Canon), not by the prior possession of joint goals (as Tomasello would have it), but by a greater capacity for forming expectations on the basis of perceived relevantly similar reactions.

Whatever the outcome of this particular issue, however, Tomasello rightly concludes in much broader brush-strokes [Tom08:176]:

…human-like collaborative activity – group activity with an intentional structure comprising both a joint goal and complementary roles – is something in which great apes do not participate. In general, it is almost unimaginable that two chimpanzees might spontaneously do something as simple as carry something heavy together or make a tool together.

That being taken as given then, the final stage of Tomasello’s strategy is to lay bare the underlying cooperative nature of all distinctly human communication. As Tomasello enumerates at length [Tom08:106-7]:

167

…we call our model the cooperation model of human communication. In this model: (i) human communicators and recipients create the joint intention of successful communication, adjusting for one another as needed; (ii) human communicative acts are grounded in joint attention and shared understandings of the situation at hand; (iii) human communicative acts are performed for fundamentally prosocial motives such as informing others of things helpfully and sharing emotions and attitudes with them freely; (iv) human communicators operate in all of this with shared assumptions (and even norms) of cooperation between participants; and (v) human linguistic conventions, as the crowning pinnacle of human discourse, are fundamentally shared in the sense that we both know together that we are both using a convention in the same manner.

Unfortunately it goes far beyond the scope of this thesis to critique this cooperative model of human communication to the degree that it deserves (but see the replies in [Tom09] for a start). Instead, it will have to suffice to herein tacitly endorse Tomasello’s account as it stands in rough outline (with a few minor suggested amendments to be made in both the following section and chapter).

As I have sketched above, Tomasello’s approach is threefold. First, he looks to cooperation as a necessary non-semantic feature of all semantic communication. He then works to establish on the basis of empirical research that this very feature is a distinctly human capacity that is remarkably close to being realised in our closest, yet languageless, evolutionary relatives. It remains to be seen just how close this account might bring us to the tipping point from non-semantic interaction to semantic communication (and indeed, such is the aim of the remainder of this thesis), but, as a natural successor to Davidsonian triangulation, Tomasello’s account appears to be highly promising in at least pointing to a more precise location for where to search for and investigate this crucial moment.

168

V. IMITATION

Tomasello’s account of the cooperative communication between humans, as distinct from its absence in chimpanzee interactions, makes a compelling case that it is precisely this capacity for cooperation which is the decisive moment for entering into the intentional circle. In this section, I seek not only to bolster Tomasello’s own account, but also, to further articulate the link between Davidsonian triangulation and Tomasello’s cooperation. I do so by proposing to add a more fine-grained element to the phylogenetic account that Tomasello himself provides. Specifically, I borrow from Friedrich Nietzsche’s insight (quoted in the abstract to this chapter) in suggesting that for early humans to have come to cooperatively gesture so as to inform, it is likely that, within the context of interactive triangulation, the expectations of the observing creature must have lead the observing creature to behave in ways that would be otherwise appropriate for the observed conspecific. That is to say, the observing creature must behave (even if only partially) as though the circumstances of its conspecific were its own.

To motivate this proposal, let’s begin by all too quickly rehearsing the evolutionary story provided by Tomasello. (See [Tom08&14] for the full account.) (For alternative accounts, see, for instance [Bog99], [Car06], [Des07], [Dun96], [Hau96], [Hrd09], [Lie98], [Sky10], [Spe96].) Tomasello speculates that, in order for cooperative communication and shared intentionality to arise, firstly, ecological pressures would have afforded evolutionary preference for more socially tolerant traits [Tom08:193]:

… step one in the direction of human-like collaboration and cooperative communication is for individuals to become more tolerant, generous, and less competitive with one another, perhaps especially in feeding contexts.

That these more tolerant traits turned out to be evolutionarily advantageous, Tomasello explains, is because they afforded interdependencies that mutually increased the group’s chances of survival [Tom14:37]:

The first and most basic point is that humans began a lifestyle in which individuals could not procure their daily sustenance alone but instead were interdependent

169

with others in their foraging activities which meant that individuals needed to develop the skills and motivations to forage collaboratively or else starve.

Once these closer-knit groups came into being, they allowed for the individuals in the group to specifically select for behaviours and capacities that further promoted such relatively close-knit interdependencies. As Tomasello himself puts it [Tom08:194]:

The second step is that these individuals, who are coordinating actions with one another more regularly and tolerantly, would then be in a position for natural selection – given the appropriate ecological conditions – to specifically favour cognitive and motivational machinery supporting more complex collaborative interactions.

The particular traits that Tomasello has in mind here are those that would support cooperation and, so, shared intentionality [Tom08:194]: ‘What might be selected for in these tolerant, peacefully co-feeding individuals is the ability to create joint goals and joint attention.’ With respect to communication specifically, Tomasello writes [Tom08:195–6]:

… as we are working toward our joint goal in mutualistic collaboration, it is to each of our advantages that we help the other – and we are also likely to understand attempts to request and offer help communicatively as we are in the common ground of the collaborative activity. In this context, the communicator’s tendency to request help and the recipient’s tendency to simply help might naturally arise as a way of facilitating progress toward a joint goal.

Note here that there are two distinct communicative gestures being identified. They are, on the one hand, ‘attempts to request […] help communicatively’, and, on the other, ‘attempts to […] offer help communicatively’. To be precise, only the latter mode of communication can be taken to support properly joint goals and joint attention, whereas the former – in light of the previous discussion concerning the limitations of Brinck’s analysis – is more modestly understood in terms of parallel goals and parallel attention. At this point in the unfolding of Tomasello’s phylogenetic account, however, only parallel goals and attention can be posited, for it is properly joint goals and attention that we are hereby seeking to

170

explain. That being noted, Tomasello explicitly downplays the difference between the two modes of communication when he writes [Tom08:196]:

… in mutualistic collaborative activities the difference between requesting help and offering help by informing is minimal. That is, if we are moving a log together toward a joint goal, if there is an obstacle in the way I can request that you remove it to help us, or I can inform you of its presence which I assume will lead you to want to remove it to help us.

Tomasello stresses this similarity because the account he offers of how informative gesturing plausibly arises, rests on it being, in a sense, continuous with imperative gesturing. As he explains at length and with more nuance [Tom14:50]:

The emergence of this motive [to help the other by informing her of situations relevant to her] was aided by the fact that in the context of a joint collaborative activity, directive communication and informative communication are not clearly distinct – because the partners’ individual motives are so closely intertwined. Thus, if we are gathering honey together and you are struggling with your role, I can point to a stick, which I intend as a directive to you to use it, or, alternatively, I can point to the stick intending only to inform you of its presence – because I know that if you see it you will most likely want to use it. When we are working together toward a joint goal, both of these work because our interests are so closely aligned.

By this Tomasello is intending to support his analysis that imperative gestures as communicative requests for help readily arise and are supported by mutualistic contexts of parallel goals, simply because ‘helping you helps me’ [Tom08:198]. Tomasello also further speculates that ‘indirect reciprocity’ – whereby ‘individuals choose to help or cooperate with others who have good reputations for helping and cooperating in general’ [Tom08:200] – can plausibly explain how altruistically informative gesturing might also arise out of this context. Again, as Tomasello puts it with finer detail [Tom14:50-1]:

171

The evolutionary proposal is thus that early humans’ first acts of cooperative communication were pointing gestures in joint collaborative activities, and these were underlain by a communicative motive not yet differentiated between requestive and informative. But at some point early humans began to understand their interdependence with others not just while the collaboration was ongoing but also more generally […]. And outside of collaborative activities, the difference between me requesting help from you, for my benefit, and me informing you of things helpfully, for your benefit, becomes crystal clear. And so there arose for early humans two distinct motives for their deictic communication, requestive and informative, which everyone both comprehended and produced.

While I do not wish to deny Tomasello’s general picture, nonetheless, I do believe that by invoking the structural analysis developed in the earlier discussion of interactive triangulation, Tomasello’s account might thereby be supplemented so as to more clearly explain how this capacity to help by informing may have arisen, at least in the context of signalling.

The point I want to make is very basic, but, I think, significant nonetheless. It is this: initially, the capacity to gesture altruistically requires that the creature must have formed expectations concerning the likely behaviour of its conspecific given the state of affairs as perceived by the observing creature. That is, given what the creature itself sees, it must expect that its conspecific will behave in a certain way as result. This holds for at least two reasons.

i) Firstly and most obviously, the creature could not possibly come to understand that its conspecific might be in need of helpful information unless the creature had expectations concerning the conspecific’s likely reaction to the state of affairs as perceived by the creature, such that, when its expectations come to be frustrated, then, it is given impetus to help. This is straightforward enough: one has no reason to indicate to someone, say, the location of some desired object unless one believes that, had they known it was there, then they would have acted upon it themselves. As Tomasello puts this point [Tom08:196]: ‘I can inform you of [an obstructing object’s] presence [because] I assume [doing so] will lead you to want to remove it to help us’. And again [Tom14:55]: ‘If I am going to be helpful, I must point out situations that are new for you, or else why bother’.

172

ii) Secondly, the creature needs to have formed expectations about its conspecific’s likely behaviour given the state of affairs perceived by the creature. This is because, as I shall argue, the creature must behave in a relevantly similar way to how it expected that its conspecific would have behaved had the conspecific itself perceived the state of affairs as perceived by the observing creature. That is, the observing creature must behave (even if only partially) as though the circumstances of its conspecific were its own. This stipulation, as I argue presently, is required so as to not beg the question in our account of how the capacity for cooperative communication might arise. As Daniel Dennett puts it [Den10:60]: ‘The panglossian (or polyannian) presumption of cooperation must not be built in’.

* * *

To begin the argument, let’s recap: on Tomasello’s account, mutualistic parallel goals arise as a result of early humans becoming more tolerant in response to ecological pressures. These parallel goals, then supported communicative gestures where ‘communicative motive [was] not yet differentiated between requestive and informative’ [Tom14:50]. But to put it this way is misleading as it implies that the latter was latently present along with the former. As we have already seen, the requestive/imperative gesture precedes the informative, both historically and conceptually.

We know that the requestive/imperative gesture precedes the informative historically as our nearest evolutionary relatives can gesture imperatively, but not informatively. This, of course, makes perfect sense in light of the kind of analysis that was provided in the previous chapter concerning the touch-back gesture. That is to say, imperative gestures are a kind of short-hand manifestation of genuine efforts to respond differentially to one’s environment for oneself, say by reaching to grasp an object. Over time, given the creature is reliably aided by another such that their genuine effort is cut short, the imperative gesture may arise in expectation that the short-hand will suffice to achieve the creature’s parallel goal with the aid of its helper (conspecific or otherwise).

Nor can Tomasello’s invocation of ‘mutualism’ by itself explain the emergence of informative gesturing – i.e. ‘the information is offered because it helps us both’ [Tom08:192

173

(emphasis added)]. This is because one cannot here claim that the creature might appreciate either the meaning or the personal benefit of its own imperative gesturing so as to then straightforwardly transfer this capacity to informative gesturing. It would be question begging to presume that the creature might know that doing so would be beneficial to either its conspecific alone or both creatures mutually. Given we are here attempting to account for how joint intention might arise, we are constrained from invoking an appreciation of mutualism in the creature as a sufficient motivation for cooperative gesturing.

As I mentioned above, the solution I wish to make is small but, I believe, significant nonetheless. Hopefully the above considerations render its significance clear enough. Its ‘smallness’, however, consists only in this: it is plausible (and, in light of the above, permissible) that the early human capacity for imperative gesturing was able to shift to being informative such that non-competitive contexts and so, genuinely cooperative behaviours and shared intentions arose as a result of the observing creature having formed such strong expectations of its conspecific’s likely behaviours that the observing creature itself behaved (even if only partially) as though the observed circumstances of its conspecific were its own. For example, if a creature sees that its conspecific is, unwittingly, about to be struck, the observing creature may itself flinch as though it were to be struck itself. If ‘displaced’ reactions were to occur commonly – like a ‘tell’ in the card game of poker – the otherwise ignorant conspecific might then capitalise on such reactions so as to eventually select for and bring about fully cooperative altruistic informative gesturing.

To see how this might come about, imagine a creature seeing its conspecific looking for an object that is just out of the conspecific’s sight, and as a result, the observing creature’s own hand unthinkingly moves towards the object. In this case, all that is required for our account is that the creature has formed sufficiently strong expectations concerning the likely behaviour of its conspecific given the circumstances as perceived by the observing creature. It is not that the creature is taken already to possess joint goals, such that mutualism might be invoked to explain why it should point, for the very issue in question is to account for joint goals as manifest in altruistically informative gesturing. Admittedly, some further cognitive story needs also to be told about why the creature would (even partially) gesture as though it were the creature itself that was in the place of the observed conspecific. However, aside from pointing any curious reader in the direction of formative empirical studies concerning the role of mirror neurons and mirror systems that might one day provide such an account (see, for instance, [Cor11:60–2,70–1]), nothing hangs on such an

174

account being given here. It is enough that interactive triangulation might explain – without invoking cooperation or shared intentionality – how a creature might come to form sufficiently strong expectations regarding its conspecific’s likely behaviours given the circumstances perceived by the observing creature. In this way, then, we can articulate a plausible account of how imperative gesturing might arise without having to question-beggingly invoke shared intentionality, cooperation or even mutualism. In sum: at this early stage, the creature doesn’t need to engage in shared attention or share intentions; it is sufficient that they need only have relatively similar parallel attention, as is readily supported by interactive triangulation.

One potential worry with such an account is that it may seem to suggest that the conspecific could straightforwardly capitalise on the unintentional behaviour of the observing creature in ways incompatible with the scenario described by Tomasello wherein the chimpanzee could not comprehend or capitalise on the cooperatively informative gestures of the human, even though the gesture was structurally identical to the imperative gestures made in competitive contexts that the chimpanzee could capitalise on. Such concern, however, is ungrounded, for the proposed scenario is a genetic one. That is to say, all that is required at this very early stage in the origins of informative gesturing is that the conspecific’s attention may be relevantly (though, again, not yet intentionally) drawn to aspects in the environment that may be beneficial to being noticed. For example, a creature observing its conspecific about to be unwittingly struck by a falling branch may, under the account being proposed, itself flinch in anticipation. The sudden jerk of the observing creature might subsequently draw the conspecific’s attention to the impending collision so as to then be able to avoid it. The same story, mutatis mutandis, would then apply in the case of unintentional gesturing introduced above. Note, also, that in the empirical scenario described by Tomasello, the chimpanzees are readily able to follow the direction of human partner’s informative gesture; it’s just that they do not comprehend it as relevant to their own individual non-competitive search for the food.

Finally, why it is that humans and not any other of the great apes species came to be able to altruistically gesture so as to inform, might be the result of contingent historical differences in the relevant ecological pressures faced by each species – as Tomasello rightly emphasises – or it might have had something to do with subtle differences in the architecture of the brains of each respective species, such that early humans, for instance, not only formed stronger expectations concerning their conspecific’s likely behaviour to circumstances as

175

perceived by the observing creature, but also had a stronger tendency for the observing creature to itself react unintentionally – again, like the ‘tell’ in poker – such that the ecological pressures highlighted by Tomasello might then render the selection and so development of such interactions as beneficial for all concerned. Obviously, the answers to these questions are far-ranging and well beyond the scope of this thesis. Nonetheless, I hope to have hereby made, as I said, a small but significant contribution to the non-question-begging story of how the cooperation needed to support eventual intersubjectivity (and perhaps even language and the concept of objectivity) plausibly arose for creatures such as ourselves.

VI. CONCLUSION

While cooperation might well be the turning point for the emergence of language and perhaps also the concept of objectivity, Tomasello’s story thus far doesn’t provide much in the way of illuminating how the concept of objectivity itself arises. Recall that we were particularly interested in Davidson’s account of triangulation, not merely because it promised to establish a necessary condition for the emergence of thought, but also because Davidson’s Argument from Objectivity claimed it was the possession of the concept of objectivity that was the decisive moment in the triangular relationship between mere reactions and intentionality. Davidson’s interactive triangulation was, however, seen to amount to little more than two overlapping intra-reactive triangles, thus providing insufficient reason for supposing full-blown mutual intersubjective awareness might emerge in this way. In this chapter, we have argued that Tomasello’s notion of cooperation brings us one step closer to breaking into the intentional circle than even Brinck’s characterisation of joint attention. But, while Tomasello’s evolutionary account of how altruistically informative gesturing first arose goes some way to bringing us closer to explaining cooperative communication as an entering wedge into the intentional circle, still, only half of the story has been told so far. The complementary part is to explain how humans, but remarkably not chimpanzees, are capable of capitalising on altruistically informative gesturing in non-competitive contexts of the object-choice task. It is one thing to be able to point to inform, but it is another thing entirely to be informed by pointing. Accounting for this seemingly unique human capacity will be the focus of the final chapter, and it is there that we shall see the concept of objectivity re-enter the scene.

176

CHAPTER SIX

* * *

Inference

The fourth citizen had been wrong about what they knew, what they thought, what they wanted … and now they knew that? The implications rebounded between the symbols, models of minds mirroring models of minds, as the network hunted for sense and stability.

– Greg Egan, Diaspora.

I. INTRODUCTION

In previous chapters, an account of mutually interactive triangulation was developed and brought to bear on Davidson’s Argument from Objectivity. Although interactive triangulation was seen to provide greater sophistication than its intra-reactive predecessor, nonetheless, the brute causal structure of interactive triangulation alone was deemed insufficient to provide the creatures in question with the requisite intersubjective awareness demanded by Davidson’s own argument. We turned instead to a promising line advanced by Brinck that emphasised the phenomenon of joint attention as providing an ‘entering wedge’ into the intentional circle. Of this view, however, it was argued that the proposed criterion of being able to make use of gestures to non-salient entities was, on the one hand, explicable

177

in the wholly non-intentional framework we had earlier laid out with respect to interactive triangulation, and, on the other hand, was unable to account for an experimental phenomenon emphasised by Michael Tomasello. Specifically, the object-choice experiment reveals that chimpanzees and human infants are both able to satisfy Brinck’s model of joint attention when the context is competitive, but only humans can do so outside of competitive contexts. In such light, it was concluded that cooperative communication (both informing and being informed) is a promising candidate to be the distinguishing feature allowing Davidson’s triangular picture to eventually account for the emergence of the concept of objectivity. Tomasello’s evolutionary account of how altruistically informative gesturing first arose was modestly amended in accordance with the prior analysis of interactive triangulation, but did not shed any light on the relationship (if any) between language and the concept of objectivity. In this chapter, then, I will pursue the other half of the story – i.e. how humans are uniquely able to capitalise on informative gesturing in non-competitive contexts – and, in doing so, I will argue that the account given bears striking parallels with Robert Brandom’s Inferentialistic account of the concept of objectivity such that the two are likely continuous with one another. That is to say, our ability to communicate cooperatively is intimately related to our possession of the concept of objectivity.

In Section II, Brandom’s inferentialistic account of the concept of objectivity is brought to bear on the Davidsonian account of triangulation developed thus far. The general viability of such an undertaking is then considered in light of a number of similar undertakings made by others, including Fergestad & Ramberg, Swindal, and Brandom himself. While it will be seen that none of these particular accounts satisfactorily resolve the issue of whether Brandom’s account can be appropriated in a Davidsonian framework, related commentary by Dennett, Price and Lance & Kukla suggest that such a naturalistic undertaking should not only be possible, but moreover, desirable.

Brandom’s account of how the ‘game of giving and asking for reasons’ might give rise to the concept of objectivity is then introduced. Briefly, Brandom argues that possession of the concept of objectivity is explicable entirely by the inferentialistic moves we are able to make as linguistic creatures in the social game of giving and asking for reasons. More specifically, Brandom argues that such moves are made possible by the inferentialistic differences arising from the following two complementary pairings: undertakings & attributions and commitments & entitlements. As Brandom puts it [Bra94/98:595 (italics added)]: ‘…the required notion of

178

objective correctness is just what is expressed by de re specifications of the conceptual contents of ascribed commitments.’

In Section III, I undertake a Carruthers-style analysis of both human and chimpanzee performance in both competitive and non-competitive contexts of the object-choice task. In doing so I argue that Tomasello’s own original characterisation of what he calls the cooperative inference can be made more transparent and parsimonious.

In Section IV, argue that the basic inferentialistic structure of human performance in non-competitive contexts of the object-choice task not only parallels Brandom’s Inferentialistic account of the concept of objectivity, but that the former are likely non-normative forerunners for eventual emergence of the latter. In this way, I not only make some headway in naturalising Brandom’s account, but, moreover, show that cooperative communication is plausibly intrinsically related to our possession of the concept of objectivity.

II. BRANDOMIAN INFERENTIALISM

As was mentioned above, Brandom’s inferentialistic account of our possession of the concept of objectivity will form one of the two core aspects of this chapter (see sub-sections I.4&5). However, before introducing Brandom’s account, a question of compatibility needs to be addressed, for, as we shall see in the following sub-section, Brandom’s inferentialism is explicitly anti-naturalistic and so seemingly incompatible with the loosely naturalistic Davidsonian approach adopted herein. I argue that while Brandom’s inferentialistic program may or may not be broadly incommensurable with small-n naturalism, at least his account of how the concept of objectivity may arise as a result of our discursive practices may still be productively drawn upon in a naturalised account of human infant performance in non-competitive contexts of the object-choice task. Indeed, as I shall argue in the final section of this chapter, the plausibility of both Brandom’s and Tomasello’s accounts is augmented by their being paired naturalistically.

179

1. A Question of Compatibility

There are many complementary moments shared between Davidson and Brandom. For instance, Brandom explicitly endorses Davidson’s view that ‘[the] attributions of contentful intentional states and contentful speech go hand in hand, that neither sort of attribution is intelligible apart from its relation to the other’ [Bra94/98:150–1]. Or as Davidson himself puts it [Dav74/01:156]:

… neither language nor thinking can be fully explained in terms of the other, and neither has conceptual priority. The two are, indeed, linked, in the sense that each requires the other in order to be understood; but the linkage is not so complete that either suffices, even when reasonably reinforced, to explicate the other.

Furthermore, although the account herein diverges (see Chapter One), Brandom at least concurs with the Davidsonian view that beliefs are only possible if the creature possesses beliefs about belief [Bra94/98:152]:

… what a [creature] has is not recognisable as a belief unless the creature [...] acknowledges the applicability of a distinction between beliefs that are correct and those that are incorrect, in the sense of being true or false. We are not permitted to attribute the belief that p (a propositionally contentful state) unless somehow the putative believer acknowledges in practice the objective representational dimension of its content – that its being held is one thing, but its being correct is another, something to be settled by how it is with what it is about.

These Davidsonian commitments lead Brandom to describing the overarching aim of his magnum opus, Making it Explicit, in the following terms [Bra94/98:152 (italics added)]:

… this work focuses on the development of an account of linguistic social practices within which states attitudes, and performances have, and are acknowledged by the practitioners to have, pragmatic significance sufficient to confer on them objective representational propositional contents. The view

180

propounded is like Davidson’s in seeing intentional states and speech acts as fundamentally of coeval conceptual status…

The justification Brandom gives for undertaking this task on his own despite being in agreement with Davidson’s philosophical achievements, is that Brandom sees Davidson as having given more the form of an argument for such an approach, rather than the requisite argument as such. As Brandom continues [Bra94/98:153]:

Turning [Davidson’s proposed approach] into an actual argument requires filling in various notions of content, of objective representational correctness of content, of practical acknowledgement of the significance of assessments of correctness of content, and so on. That is the task of the rest of [Making it Explicit].

That Brandom explicitly acknowledges Davidson’s argumentative ‘form’ to be the inspiration for his own work may give us some ground for supposing that Brandom’s subsequent account may be congruent with the Davidsonian approach carved out over the course of this dissertation so far. James Swindal [Swi07:113–126] and Ingvald Fergestad & Bjørn Ramberg [Fer11:221–257], for instance, have all undertaken comparative work between Brandom and Davidson, specifically with respect to triangulation.16 All three commentators invoke Davidson’s triangulation argument as needed to support Brandom’s inferentialism. Yet, Fergestad & Ramberg prefer Brandom’s approach, whereas Swindal sides straightforwardly with Davidson.

Fergestad & Ramberg argue that the role Davidson ascribes to triangulation is better thought of in a Brandomian light as more ‘radical’ and ‘idealised’. By this they mean that the significance of triangulation should be seen as distinct from ‘the perhaps more pronounced, even dominant, naturalistic tendency [in Davidson], which seeks accounts of mind and meaning in naturalistically acceptable terms’ [Fer11:221]. Fergestad & Ramberg take Davidson’s naturalisistic tendencies to be a weakness such that they thereby seek a more ‘radical’ reading of Davidson as providing a transcendental argument for the necessary

16 Christian Barth also has argued (see [Bar11]) that Brandom’s Inferentialism can support what Barth calls Davidson’s ‘Basic Idea’, yet, on the one hand, as Barth’s own framework is decidedly non-naturalistic, and, on the other, he provides no explicit consideration of the viability of pairing the work of these two philosophers together.

181

material conditions of Brandom’s inferentialism. Swindal, on the other hand, champions Davidson’s triangulation thesis as a ‘weak naturalistic move, [...which] furnishes an adequate answer to the objectivity problem [facing Brandom] while relying upon fewer problematic assumptions about what constitutes rationality’ [Swi07:113]. Swindal thus sees Davidson’s weak natural as a strength and so, is willing to opt quite straightforwardly for Davidson’s account as it stands (seemingly without acknowledging any of the many objections raised against it by others). Either way, neither of these two critical engagements directly address the question of whether Brandom’s account itself might be appropriated into the Davidsonian framework in which our discussion has thus far been couched. They are concerned only whether Davidson might serve Brandom, but not vice versa. The difficulty with the issue faced here, however, would appear to stem from the different respective attitudes to naturalism. We have already seen Davidson’s avowal of a common sense, scientifically informed, liberal notion of naturalism (see [Dav95/04:5]). Brandom, however, explicitly says his project is ‘opposed to naturalism, at least as that term is usually understood’ [Bra00:26)]. As he explains [Bra00:26]:

Once concept use is on the scene, a distinction opens up between things that have natures and things that have histories. Physical things such as electrons and aromatic compounds would be paradigmatic of the first class, while cultural formations such as English Romantic poetry and uses of the terms ‘nature’ and ‘natural’ would be paradigmatic of the second.

According to Brandom’s understanding, the normative inferences that govern our social discursive practices belong to a non-causal ‘historical’ realm, whereas the kind of mere ‘natural’, non-intentional phenomena belonging to triangular interactions – intra-reactive, interactive, cooperative, or otherwise – are said to belong to the physically causal domain. As Brandom puts it [Bra94/96:626]:

Norms (in the normative sense of the word) are not objects in the causal order. Natural science, eschewing categories of social practice, will never run across commitments in its cataloguing of furniture of the world; they are not by themselves causally efficacious…

182

In this way, Brandom denies that norms can be systematically reduced to non-normative phenomena. As Barth’s explains it [Bar11:124]: ‘[Brandom’s] normativism implies that it is necessary for content-conferring practices to be governed by irreducible norms, i.e., to be governed by norms that cannot be reduced to naturalistically specifiable regularities’. And as Mark Lance & Rebecca Kukla reiterate [Lan10:115]: ‘For Brandom, inferentially articulated discourse forms an autonomous domain of normativity, while our bodily encounters with the world in perception and in action serve as language entry and exist points respectively’.

Although such an explicit anti-naturalism appears to stand at odds with Davidson’s broadly naturalistic commitments, it is perhaps plausible to see Davidson’s oft-repeated claim that the conceptual is at least discursively irreducible to the natural as in someway commensurate with Brandom’s Inferentialism. As Brandom puts it [Bra94/98:155-6]:

…no attempt will be made to show how the linguistic enterprise might have gotten off the ground in the first place. But it should be clear at each stage in the account that the abilities attributed to linguistic practitioners are not magical, mysterious, or extraordinary. They are compounded out of reliable dispositions to respond differentially to linguistic and non-linguistic stimuli. Nothing more is required to get into the game of giving and asking for reasons – though to say this is not to say that an interpretation of a community as engaged in such practices can be paraphrased in a vocabulary that is limited to descriptions of such dispositions.

That said, Brandom has elsewhere undergone heavy criticism for the decidedly anti-naturalistic tenor of his work. Rebecca Kukla, for instance, has argued that the success of the complementary work of Joseph Rouse (see [Rou02]) has revealed Brandom’s work as failing ‘to take sufficiently seriously the way that our discursive and inferential practices are, in the first instance, concrete bodily practices that always engage our perceptual and practical skills’ [Kuk04:218]. Kukla and Mark Lance together repeat this perceived limitation in Brandom’s writing when they write [Lan10:115]: ‘MIE does not offer a systematic theory of how the domain of bodily interaction and the discursive domain fit together…’. But the most general, yet explicit, criticism has come from Daniel Dennett [Den10:60]:

[Brandom] excuses himself from [the task of explaining how we humans are capable of expressing content that is either objectively true or false] by retracting

183

into a very modest goal: ‘explaining what the trick consists in, what would count as doing it, rather than how it is done by creatures wired up as we are’ (p. 155). In other words, he wants to give the ‘specs’ without a word on the implementation, the (reverse) engineering. This is a remarkable truncation, given not just the ambition but the gratifying success of his own reverse engineering of the game of asking and giving reasons. He blandly assures the reader that “the abilities attributed to linguistic practitioners are not magical, mysterious or extraordinary” (p. 156) but we are entitled to ask an emperor’s new clothes question: Why not?’

Dennett’s attack continues [Den10:60]:

This [approach undertaken by Brandom] is, I think, an unacceptable flight from naturalism, not just because it refuses to address an entirely appropriate question [(specifically, of whether the norms Brandom’s Inferentialism relies on co-evolve with the same norms that underwrite morality),] but because by doing so it actually distorts the analysis at the higher level. A deeper account of communication would be a better account.

Moreover, Dennett further assures us that the relevant empirical literature is out there waiting to be drawn on in providing a naturalistic account of the emergence of language compatible with Brandom’s own account [Den10:61]:

[There are a number of] quite uncontroversial themes in the considerable naturalistic literature on the evolution of communication. Adding them to [Brandom’s] perspective would not commit Brandom to any doctrinaire scientism or evolutionism that might seem to threaten the philosophical austerity of his project, and it would go some way to removing the suspicious residue of magic that clings to any account that purports to ground meaning in community and then, having done that, declares victory.

And again [Den10:55]:

184

My point is not a procedural criticism – he should have cited this work – but a substantive one: by ignoring it, he creates the illusion (for himself and his readers) that his community-based account of meaning and representation can be an autonomous alternative to a naturalistic theory of the same topics. The themes he discusses so insightfully are not just accessible to naturalism; they are (already) explicable, to a significant degree, in naturalistic terms.

In this, Huw Price also agrees that the key insights of Brandom’s Inferentialism can be naturalistically deflated without loss, and it is fair to read much of Price’s work on what he calls ‘Brandomian Expressivism’ as a broad attempt to demonstrate this to be the case (see [Pri11] and [Pri013] for extended engagement with Brandom’s alleged ‘expressivistic’ tendencies). Price writes [Pri08/11:318]:

[Brandom’s] account only looks non-naturalistic (to him) because he tries to conceive of it as metaphysics. If he had stayed on the virtuous (anthropological) side of the fence to begin with, there would have been no appearance of anything non-naturalistic.

Although, citing Dennett and Price in this way is clearly an appeal to authority, it is one, I believe, that may be seen to be justified in light of the following analyses. Indeed, although Brandom coyly pleads ignorance as to exactly what in his philosophy might be improved by welding it to a more naturalistic account of the emergence of linguistic communication and propositional thought, he himself allows, in response to Dennett, that ‘it might well be that my account would benefit from supplementation by attention to some of the “… naturalistic literature on the evolution of communication”’ [Bra10:308].

2. Brandom on Triangulation

In both Making it Explicit [Bra94/98] and Reason in Philosophy [Bra09], Brandom critiques triangulation as failing to provide a sufficient solution to the problem of content determination – or, as Brandom calls it, the ‘fundamental problem’ of ‘the stimulus-response model of empirical content’ [Bra09:188–9]. Although he quotes Davidson’s own

185

construal of the problem, when it comes to considering the supposed role played by triangulation in resolving the problem, Brandom focuses exclusively on Fred Dretske’s account and not on Davidson’s. One possible reason for this move is that Dretske’s explication of the importance of triangulation came before Davidson’s, such that it may be more properly considered the progenitor of the idea (i.e. [Dre81] and [Dav82/01]). It is also the case that Dretske’s account is more straightforwardly aimed at the problem of content determination alone, whereas Davidson’s recourse to triangulation is put to many uses, as was outlined in the introduction to Part Two. One might also suppose (albeit perhaps somewhat uncharitably), that Dretske’s own account better serves Brandom’s rather sweeping engagement with the idea insofar as the former is less nuanced than Davidson’s, as I hope to have evidenced in Chapters Three and Four via my reconstruction of the structural hierarchy to be found in Davidson’s account.

That said, as Brandom explains it, invoking the notion of triangulation is ‘the most popular approach to identifying distal stimulus as what is classified by the exercise of reliable differential responsive dispositions...’ [Bra09:191]. This is to say, it is hoped to explain how interactive triangulation is able to pick out distal sources of stimulus, and not just our proximal stimulations. Brandom puts it [Bra09:191]:

The proximal approaches to distinguishing stimulus are disastrous for the project of understanding the contents of fundamental empirical concepts from the way in which their application in observation depends on the exercise of reliable differential responsive dispositions. For the stimuli that are classified by responses do not extend out to the world, but remain in the organism, or at best at its functional surface. [...] Nothing that looks like one of our ordinary observable objects, is within reach of such an approach.

Triangulation promises to guarantee that the world does play the role we naively assume it to in providing distal content to our beliefs. Brandom illustrates how the introduction of triangulation is intended to work by using the analogy of a thermostat, the function of which is maintaining the temperature of a room within a certain range. Typically, a single thermostat, Brandom explains, is taken to reliably respond to the temperature of the room insofar as its properly functioning mechanism has been calibrated to provide a standardised output of turning a furnace either on or off within a set temperature range. Brandom claims (and notes that Dretske too acknowledges) that of such a simple model, it cannot be said

186

that the thermostat is responding to the temperature of the room per se, as opposed to the temperature of the thermometer itself, or the electrical circuit as it closes when it reaches its assigned limit, or any other arbitrarily chosen point along the causal chains either internal or external to the thermostat. Notice also that, if the thermostat happens to malfunction in some way such that it gives an incorrect output, there is and can be nothing within the simple thermostat itself that might recognise such error has occurred. (Similar concerns were aired in Chapter Three.)

The intended role played by triangulation, Brandom continues, then, is akin to introducing a further, different thermometric mechanism that is systematically integrated with the first, such that the thermostat functions only when the two mechanisms satisfactorily correlate with one another. As Brandom puts it [Bra09:192], ‘If the thermostat has a second sensor … then the system has two ways of responding to the temperature’. Someone who might endorse triangulation as straightforwardly resolving the problem of distance may be tempted here claim that the whole thermostat can be said to be responding to the temperature of the room insofar as the two routes from each sensor ‘intersect in just two places – upstream at the change of temperature in a room which is included in the “flow” or causal chain corresponding to each route, and downstream in the response of turning the furnace on or off’ [Bra09:192]. We can better illuminate the intended significance of this triangular relation if we again consider the possibility of malfunction in either one of the sensors and we add to the system a switch that both sounds an alarm and defaults to turning the furnace off when the two inputs into the switch mechanism sufficiently deviate from their designed focal convergence on the ‘upstream’ temperature of the room. Here, then, it might seem that a check had been placed on proper functioning of the thermostat such that it is thereby capable reacting if an error has occurred – i.e. when it is likely that one of the two sensors is on the blink, the furnace shuts down and the alarm sounds to alert the user.

As Brandom points out, however, the shortcoming in appealing to triangulation in this straightforward way to resolve the problem of distance is that it only pushes the original difficulty back one step. It still may be said that the system is not responding distally to the temperature of the room as such, but only proximally to the divergent outputs of the sensors themselves. Again, to illustrate what is at stake here, we might note that, while the dual subsystem mechanism is able to respond in such a way as to sound an alarm if either or both of the sensors deviate from one another, the system itself still cannot recognise that the alarm has sounded. It doesn’t recognise any ‘error’ as such; it only alerts a competent user.

187

In this light, the sensors still cannot be said to be measuring the distal temperature of the room, but, again, only that, proximally, the sensors themselves are giving different readings. Thus we once again have an interpretation that is proximal rather than distal.

Brandom diagnoses this shortcoming of the basic triangular relationship as a result of being too inferentially weak to establish concepts (i.e. of an objective room temperature). As Brandom puts it [Bra09:193]: ‘… mere differential responsiveness is not sufficient for identifying the responses in question as applications of concepts’. Inferentialism maintains that in order to possess concepts, a sufficiently complex, holistic network of inferential relations must hold so as to adequately differentiate the concepts in question [Bra00:162]:

… the conceptual should be distinguished precisely by its inferential articulation. […] What is the difference between a parrot or a thermostat that represents a light as being red or a room as being cold by exercising its reliable differential responsive disposition to utter the noise ‘That’s red’ or to turn on the furnace, on the one hand, and a knower who does so by applying the concepts red and cold, on the other? What is the knower able to do that the parrot and the thermostat cannot? After all, they may respond differentially to just the same range of stimuli. The knower is able to use the differentially elicited response in inference. The knower has the practical know-how to situate that response in a network of inferential relations – to tell what follows from something being red or cold, what would be evidence for it, what would be incompatible with it, and so on. For the knower, taking something to be red or cold is making a move in the game of giving and asking for reasons – a move that can justify other moves, be justified by still other moves, and that closes off or precludes still further moves. The parrot and the thermostat lack the concepts in spite of their mastery of the corresponding non-inferential differential responsive dispositions, precisely because they lack the practical mastery of the inferential articulation in which grasp of conceptual content consists.

Brandom’s answer to the problem of distance, then, lies in what we do with our linguistic exchanges – which is as much to say, there’s no point asking what the meaning of any particular word or sentence is outside of the discursive practices in which it is used. Rather, Brandom argues that the inferentialistic relations of our utterances, both upstream and

188

downstream, can account entirely for what we mean by what we say. Again, in Brandom’s words [Bra09:195]:

What picks out one kind of thing as what is being reported out of all those that are being differentially responded to is a matter of the inferential commitments that response is involved in. These inferential consequences of going into a state make it clear that what is being classified is something outside of the system.

So, while Brandom is willing to allow that ‘the rationalist view can incorporate Dretske-style triangulation’ [Bra09:195], what more is needed, Brandom holds, is the inferentialistically articulated concept of objectivity, for it is this that will allow for the classification that something is outside of the system.

3. The Inferentialistic Structure of The Concept of Objectivity

In light of the above, it is no great surprise that Brandom claims that our possession of the concept of objectivity can be accounted for in purely inferentialistic terms [Bra00:203]:

… the objectivity of propositional content […] is a feature we can make intelligible as a structure of the commitments and entitlements that articulate the use of sentences: of the norms, in a broad sense, that govern the practice of asserting, the game of giving and asking for reasons.

When Brandom here writes of ‘the objectivity of propositional content’ it is likely he has something stronger in mind than we are hoping to establish here. Recall: our concern is only to account for our possession of the concept of objectivity, not to establish the objectivity of our beliefs. We are not seeking to defeat the challenge of global scepticism or to establish the epistemic soundness of our thoughts and words. Here, we are not concerned to establish whether or not grounding the objectivity of our beliefs is Brandom’s intent, let alone to evaluate whether he succeeds. Instead, our interests might more accurately be captured by the following paraphrasing of Brandom (see [Bra09:195] quoted above): The inferential

189

consequences of our discursive commitments allow us to conceptualise what is being talked about as being something outside of the system (regardless of whether or not it actually is). That is to say, we are interested in whether Brandom’s Inferentialism can account for our possession of the concept of objectivity as a concept.

Brandom’s account of how our discursive practices might allow for the specific inferential relations needed to give rise to this concept of objectivity hinges on the interplay of two paired distinctions: undertakings & attributions, commitments & entitlements.

Firstly, Brandom aims to explain how we are able to distinguish between attitudes we undertake for ourselves as commitments, and commitments that we attribute to others. The key, he argues, lies in the inferentialistic differences that follow from de re and de dicto statements and/or interpretations. The second step, then, is to account for the ability of distinguishing between commitments themselves and entitlements to said commitments. That is to say, Brandom aims to show that differing inferences follow from, on the one hand, what one is committed to, and, on the other, what one is (or isn’t) entitled to. Once this latter capacity is in play, then the concept of objectivity may arise. As Brandom has it, the concept of objectivity inferentialistically amounts to being able to realise that one may not necessarily be entitled to the commitments one has nevertheless undertaken, such that one may subsequently understand one’s own undertaken commitments in the ‘third-person attributional mode’ (as will be explained in the next section). That is, Brandom seeks to explain [Bra00:221]:

… how it is possible (what one has to do in order) to adopt, as it were, a third-person perspective toward one’s own attitudes, and so take them to be subject in principle to the same sort of assessment to which one subjects the attitudes of others, in offering de re specifications of their contents.

Or, as one might say more colloquially: the concept of objectivity depends on being able to meaningfully wonder if one has adequate grounds for what one believes. In doing so, we come to appreciate that what we believe may not be the case (i.e. objectively).

190

4. Attributing and Undertaking

The first stage in Brandom’s inferentialistic account of the possession of the concept of objectivity involves explaining how one can attribute beliefs to another individual without necessarily undertaking the belief oneself. That is, he aims to show how A may say of B, that B believes of x that it has the property, y, even though A does not believe x has property y. This capacity is essential to possession of the concept of objectivity, Brandom suggests, because, as we shall see in the following subsection, when it is combined with the further capacity to distinguish between what one actually believes (i.e. ‘is committed to’) and what has good justification for believing (i.e. ‘is entitled to’), then, together, it enables the individual to inferentialistically consider whether their own beliefs may not also be unjustified – which is to say, they can thereby realise that what they believe may not necessarily be the case objectively.

Brandom’s account of the inferentialistic capacity to attribute beliefs to another but to not necessarily to undertake them for oneself is arrived at by way of pointing out the difference in inferences that follow from either de re or de dicto readings of statements, such as,

1. ‘The president of the United States will be black by the year 2020.’

Typically, (1) will be taken in the de dicto mode, such that it can be equally rephrased as,

2. ‘By the year 2020, the United States will have had a black president.’

Given that Articulating Reasons was first published in 2000, the subsequent presidency of Barack Obama from 2009–2017 has now rendered true both (2) and (1) when read in the de dicto mode made explicit by (2).

There is, however, also a possible de re reading of (1) that is compatible with the following rephrasing,

191

3. ‘The current president of the United States will be black by the year 2020.’

Given that Bill Clinton was the president of the United States at the time (1) was first published, both the de re reading of (1) and its explicit rephrasing, (3), are highly unlikely to come true by 2020 or any time after.

Brandom takes these kinds of inferentialistic differences that follow from de re and de dicto readings as key to accounting for how we are able to attribute beliefs to others without thereby undertaking them ourselves. To make his argument, Brandom asks us to consider the following two statements made respectively by a defence attorney and a prosecutor:

1. [Attorney:] ‘The man who just testified is a trustworthy witness.’

2. [Prosecutor:] ‘I have presented evidence that ought to convince anyone that the man who just testified is a pathological liar.’

In order to capture the inferentialistic bearing that (2) has when uttered in response to (1), it would not do to interpret (2) in the following de dicto mode:

3. [Prosecutor:] ‘The defense attorney believes that a pathological liar is a trustworthy witness.’

It doesn’t make much sense to believe that a pathological liar could also be a trustworthy witness. However, the inferentialistic explanation of why (3) doesn’t adequately capture the intended inferentialistic significance that (2) is supposed to have on (1), is that the defense attorney is likely to be committed only to (1) their own statement, and not to (2) the prosecutor’s. On the other hand, by adopting the de re mode, Brandom argues, the following rephrasing of (3) can make explicit the difference in beliefs with respect to the same object,

4. [Prosecutor:] ‘The defense attorney believes of pathological liar that he is a trustworthy witness.’

192

In this way, (4) makes explicit that the prosecutor attributes to the defense attorney the commitment that the witness is trustworthy, but the prosecutor does not undertake this commitment for themselves – rather, the prosecutor is committed to the witness being a pathological liar.

Typically our ready facility in drawing the correct inferential relations that statements such as (2) bear on statements such as (1) mean that the kind of formal phrasing seen in (4) – i.e. ‘S claims of t that (𝜙)t’ [Bra00:176] – is not needed in differentiating between commitments undertaken by the speaker and those which are attributed to others. However, Brandom argues that, by making it explicit, we thereby reveal the inferential compatibilities and incompatibilities that make such distinctions possible. As we shall see, Brandom further argues that when this capacity is coupled with the ability to inferentialistically distinguish between what one is committed to and what one is entitled to, then the potential arises for one to ask of one’s own beliefs whether one is or isn’t entitled to them. That is, we may realise that what we believe may not be what is objectively.

5. Commitments and Entitlements

The second stage of Brandom’s account of the possession of the concept of objectivity involves explaining the inferential difference between what one is committed to and what commitments one is entitled to. To illustrate his case, Brandom seeks to explain why it is that what he calls ‘naïve assertibility theories’ cannot solve the problem of content determination. The naïve assertibilist, Brandom claims, seeks to equate statements such as:

1. ‘The swatch is red.’

with:

2. ‘The claim that the swatch is red is properly assertible by me now.’

193

The problem, as Brandom explains, is that (1) and (2) are not straightforwardly equivalent with one another. As he puts it [Bra00:198–9]: ‘(1) and (2) would be incompatibility-equivalent […] just in case everything incompatible with (1) were incompatible with (2), and vice versa. But […] this is precisely not so’. Brandom makes his case by pointing out that (1) and (2) have asymmetric compatibility relations to other statements, such as,

3. ‘I do not exist.’

4. ‘Rational beings never existed.’

5. ‘In the absence of a swatch, but otherwise in circumstances that are perceptually quite standard, my optic nerve is being stimulated just as it would be if there were a red swatch in front of me.’

So, with respect to the above asymmetric relations, it may be said:

• (1) is compatible with (3) and (4), as the redness of the swatch itself is taken to be independent of whether ‘I’ or any other rational beings exist.

• (2), however, is not compatible with either (3) or (4). (3) implies that there is no person which might be picked out by ‘I’. (4) implies that that there are can be commitments or entitlements, since only rational beings have commitments or entitlements to commitments.

• (1) is incompatible with (5), as (5) explicitly denies the presence of a swatch. • (5), however, according to Brandom, describes a circumstance where I am prima facie

entitled to my commitment that (1), such that (2) would hold, but not (1). As Brandom puts it: ‘I would have extremely good evidence that (1) is true, so that it is appropriately assertible by me, even though (1) is not in fact true [which] is just to say there are claims that are incompatible with (1), but not with its being assertible by me’ [Bra00:199].

These asymmetric compatibility relations can be mapped as follows:

194

Brandom then further argues that the reasons why these varying compatibility relations hold are entirely explicable via the distinction between commitments and entitlements, which can be glossed, respectively, in terms of beliefs and justifications [Bra00:199]:

The additional normative resources made available by distinguishing the status of being assertionally committed from that of being entitled to such a commitment are sufficient to distinguish the contents of ordinary claims from those claims about what is assertible.

That is to say, the distinction between commitments and entitlements, reveals that statements such as (1) and (2) are not equivalent by way of the reasons for the asymmetric compatibility relations they hold to statements such as (3), (4) and (5).

To block any objection that Brandom’s analysis ‘depends on setting up the test case in terms of the undifferentiated notion of appropriate assertibility, while assessing it using the more specific normative notions of commitment and entitlement’ [Bra00:199], Brandom replaces (2) the naïve assertibilist statement, with the following explicit articulations of commitments and entitlements, to show that the same asymmetric pattern of compatibility relations hold:

195

2′. ‘I am now committed to the claim that the swatch is red.’

2″. ‘I am now entitled to the claim that the swatch is red.’

Again, these can be mapped as follows, and show the same pattern of compatibility relations as before:

From this we see that, as was the case of Figure 1, while (1) is not directly incompatible with either (2´) or (2´´), neither it is equivalent to them by way of the asymmetric compatibility relations they hold with respect to statements such as (3), (4) and (5). As Brandom puts it [Bra00:200]: ‘…this additional specificity makes no difference’, but rather reveals that the above explanations just are what ‘distinguish the contents of ordinary claims from those claims about what is assertible’.

With the distinction between commitments and entitlements made explicit, Brandom then goes on to argue that, in the case of ‘red swatch’, the difference between (2´) or (2´´) can be drawn on to bestow (1) with ‘the objectivity of propositional content’ [Bra00:203]. The aim here is to not presume that the non-equivalence of (1) and (2´) or (2´´) straightaway reveals that (1) must be about the objective swatch, but rather to show how the inferential interplay between (2´) and (2´´) can be given objective propositional content that is independent of anyone's commitments or entitlements to (1).

196

In this vein, Brandom points out that it is now explicable that ‘one could be committed to the claim that the swatch is red, that is, to (1), without thereby being committed to the claim that one is entitled to it’. [Bra00:200] For instance, if the swatch had been seen only for a fleeting second, it is conceivable that I might form a commitment that the swatch is red without quite knowing how I did so. In such a case, I might not be committed to my being entitled to the commitment that I hold nonetheless. This particular inferentialistic relationship, enabled by the distinction between commitments and entitlement, is of paramount significance to this thesis. As Brandom puts it [Bra00:200]:

In general, one ought to be entitled to one’s commitments, but the game of giving and asking for reasons has a point precisely insofar as we must distinguish between commitments to which one is entitled and those to which one is not. So one must at least allow that it is possible that one is in such a situation in any particular case.

When the inferentialistic capacity to distinguish between commitments and entitlements is coupled with the capacity to attribute commitments without necessarily undertaking them for oneself, one can then make inferentialistic sense of statements that imply possession of the concept of objectivity, such as,

6. ‘The swatch looks red to me.’

7. ‘I think/believe the swatch is red.’

Or, more formally, in Brandomian terminology:

8. ‘I am committed to the claim, of the swatch, that it is red (but I might not be properly entitled to that claim: i.e. despite my best evidence, it may not in fact be red).’

As Brandom explains at length [Bra00:203]:

197

The point of all this is that the objectivity of propositional content – the fact that in claiming that the swatch is red we are not saying anything about who could appropriately assert anything, or about who is committed or entitled to what, are indeed saying something that could be true even if there had never been rational beings – is a feature we can make intelligible as a structure of the commitments and entitlements that articulate the use of sentences: of the norms, in a broad sense, that govern the practice of asserting, the game of giving and asking for reasons. And we can make sense of practices having that structure even if we understand commitment and entitlement as themselves social statuses, instituted by the attitudes of linguistic practitioners. All that is required is that the commitments and entitlements they associate with ordinary empirical claims such as ‘The swatch is red’ generate incompatibilities for these claims that differ suitably from those associated with any claims about who is committed to, entitled to, or in a position to assert anything. The recognition of propositional contents that are objective in this sense is open to any community whose inferentially articulated practices acknowledge the different normative statuses of commitment and entitlement.

This is a bit tricky. Typically, we have a tendency to want to be able to say what it is about statements such as (1) in themselves that enable them to be understood objectively. But, according to Brandom, it is not like this at all. Rather than there being anything peculiar about statements like (1), instead, the objectivity of their propositional content is nothing more or less than their being understood inferentialistically as independent of anyone’s beliefs regarding the matter. That is to say, the concept of objectivity arises insofar as what one is committed to is understood entirely by the compatibility relations that follow from it, as not being a matter, either, of one’s commitments or entitlements to said commitments. Because we can understand, both, that we may not be entitled to our commitments, or that we may have entitlements to commitments that we nonetheless fail to undertake – precisely because either or both of these legs may be kicked out from under us – we thereby obtain the notion that what we are either committed or entitled to is a matter beyond our commitments and entitlements. Inferentialistically, we understand that claims like (1) might still be true even if, say, one didn’t exist or if no rational beings existed. That is, we obtain the concept of objectivity – of a way things are, independently of the way we take them to be.

198

It is important to note that in the above I am not attempting to defend or critique Brandom’s Inferentialistic account of the concept of objectivity. Rather, as shall become clear in the following sections, I intend only to re-appropriate the general structural insights articulated by Brandom so as to further explicate Tomasello’s analysis of human performance in non-competitive contexts of the object-choice task.

III. ANALYSIS OF THE OBJECT-CHOICE TASK

In this section, I undertake a systematic analysis of the reported differences between human and chimpanzee performance in the object-choice task, in both competitive and non-competitive contexts. In doing so, I will again adopt the framework given by Peter Carruthers, as was employed in Chapter One, Section III. The aim here is to show that the most plausible account of human performance involves a minimalist inferentialistic structure, which, as I shall argue in the following section, parallels Brandom’s account of the concept of objectivity as given in the previous section.

Tomasello argues that, while chimpanzees will follow both the gaze and pointing of humans to locations outside of their immediate perceptual field, their behaviour in the object-choice task suggests that they do not comprehend pointing as such – i.e. ‘the intention behind the human’s attention-directing gesture’ [Tom08:39]. The explanation he gives for why chimpanzees are unable to capitalise on pointing in non-competitive contexts is that ‘apes simply do not understand that the human is communicating altruistically in order to help them toward their goals’ [Tom09:41]. Here, I am interested in further investigating this claim, not so much for what it might tell us about limitations in chimpanzee comprehension, but rather, for what it might tell us about what humans can do and whether these capacities might have anything to do with possession of the concept of objectivity. Bearing this in mind, let us start with chimpanzee performance and build our way up to humans.

199

1. Chimpanzee Performance in Response to Pointing Gestures

Tomasello prefaces his discussion of the object-choice task by first acknowledging that chimpanzees do generally appear to be able to follow pointing gestures to non-salient aspects of their environment, even to locations outside of their immediate perceptual field (see [Tom08:38–9]). Such ability is in keeping with the analysis of interactive triangulation given in Chapter Four, Section IV.

By again employing the same conventions that were borrowed from Carruthers in Chapter One of this thesis, we can give the following account of this phenomenon, with ‘A’ designating the chimpanzee, and ‘B’ the human:

AnalysisofChimpanzeePerformanceinResponsetoPointingGestures

1.PERCEPTION[Aseesnothingofinterest]

2.PERCEPTION[Bgesturestoalocation(beyondA’scurrentperceptualfield)]

3.ACTION[A’sgazefollowsdirectionof(2)(totheindicatedlocation)]

4.PERCEPTION[Anowseestreat]

5.ACTION[Agetstreat]

It is important to note that the qualifications in (2) and (3) are bracketed in order to indicate that the location being pointed towards is not explicit in the chimpanzee’s perceptual experience. That is, A need not have noticed that something (e.g. a partition) had been obstructing their view. A need not have comprehended their action as discovering the indicated location. Nor is it the case that B’s gestures lead A to doubt the veridicality of (1) its perception that there’s nothing of interest to be seen. It is sufficient merely that A’s gaze follows B’s pointing gesture such that A sees a treat, which it then proceeds to get.

It is perhaps also important to reiterate that the content of the cognitive states contained within the square brackets are not to be ascribed to A explicitly or semantically. In particular, while instances of ‘A’ itself appear in some of these descriptions, this is for clarity’s sake, and does not suppose that A itself possess a meta-cognitive self-awareness or notion of personal identity. This caveat applies throughout.

200

2. Chimpanzee Performance in Competitive Contexts of The Object-Choice Task

Turning, then, to competitive contexts of the object-choice task: chimpanzees are also able to follow gestures that are structurally similar to pointing (i.e. a human reaching in the direction of a particular bucket) so as to successfully locate the hidden treat. According to Tomasello, it is as if the chimpanzee reasons [Tom08:40]: ‘…she wants to get into that bucket for herself; therefore, there must be something good in there’. And, again [Tom14:52]: ‘He wants in that bucket; therefore the food must be in there’. These cognitive processes can be accounted for as follows:

AnalysisofChimpanzeePerformanceinCompetitiveContextsofTheObject-ChoiceTask

1.BELIEF[BisincompetitionwithA]

2.BELIEF[thereisfoodhiddenunderoneofthebuckets,a,b,c]

3.DESIRE[Awantsthehiddenfood](from(2))

4.PERCEPTION[Bisgesturinginparticulardirection]

5.ACTION[A’sgazefollows(4)tobucketa](from(4))

6.BELIEF[Bwantstogetwhateverisinbucketa](from(1),(4)&(5))

7.BELIEF[thefoodishiddeninbucketa](from(2)and(6))

8.ACTION[Aexaminesbucketa](from(3)&(7))

Step (1) is established via prior activities that create in A the expectation that B is its competitor. Similarly, it is on the basis of previous similar activities that A believes (2) that the task is to find the treat under one of the three buckets. As Tomasello describes it [Tom08:39]: ‘Apes knew from previous experience that there was only one piece of food hidden, and they would get only one choice’. Obviously, neither the chimpanzee nor the human need think explicitly of the three buckets as being labelled ‘a’, ‘b’, or ‘c’.

B’s movement in Step (4) is described, not as a reaching or pointing, but only as a ‘gesturing towards’ because, as a perception, B’s movement is most minimally described as nothing

201

more than a gross movement. Importantly, we mustn’t beg the question in supposing that A perceives B’s movement as intentionally conveying information, altruistically or otherwise. It is enough that in (5), A’s gaze follows (4) B’s movement to bucket a.

Up to this point, the account of chimpanzee performance in competitive contexts of the object-choice task has directly paralleled the earlier account given of pointing to objects beyond the immediate perceptual field of the chimpanzee. However, unlike the previous case, where the tendency of A’s gaze to follow the direction of B’s pointing gesture was what revealed to A the location of a previously unseen treat, in the case of the object-choice task, A’s gaze is instead lead to a bucket that they can already see. No further perceptual access is given by A’s gaze following the direction of B’s gesturing.

Step (6) follows from (1), (4) & (5). Taken together, (4) & (5) facilitate A’s comprehending that B is gesturing specifically towards bucket a. This goes no further than the chimpanzee’s ability to generally follow pointing gestures, as analysed above. (1), however, is needed to explain A’s belief that B wants whatever is in the bucket she is reaching for. This further element is required for two reasons. Firstly, without it, it would render (1) redundant, leaving the analysis of the general chimpanzee ability to follow pointing gestures as also entailing that they universally see the pointer as wanting what they are pointing at. Secondly, in doing so, it would imply that chimpanzees should also make the same ‘competitive inference’ even in non-competitive contexts of the object-choice task, which is obviously not the case, for otherwise they should proceed the same as they do in non-competitive contexts, which they do not.

Regarding the proposed inclusion of (7), it might be that (6) and (1) alone are sufficient to motivate (8). That is, given that A is in competition with B, and that B wants bucket a, A too should want whatever is in bucket a. It might be possible to motivate a chimpanzee to want whatever their competitor wants. I don’t know. If it is, then Morgan’s Canon would deem the inclusion of (7) as technically redundant. Instead, something like the following would then be a better substitute:

7´.DESIRE[AwantswhateverBwantsinbucketa](from(1)and(6))

202

While (7´) is cognitively less demanding than (7), it loses track of Tomasello’s original gloss, ‘He wants in that bucket; therefore the food must be in there’ [Tom14:52]. Moreover, it sees the chimpanzee abandon its initial engagement with the task – i.e. (2) that there is hidden food that (3) the chimpanzee wants. This would perhaps be a little odd; after all, as we shall see momentarily, chimpanzees are able to engage in non-competitive contexts of the object-choice task (even if they perform differently to their human counterparts), so at least the chimpanzee’s initial holding of (2) and (3) are not in doubt. Either way, the choice between (7) and (7´), has little bearing on the overall analysis, and so shall be left here as a question mark.

3. Chimpanzee Performance in Non-Competitive Contexts of The Object-Choice Task

Turning then to analyse chimpanzee performance in non-competitive contexts of the object-choice task, step (1) from the previous analysis is removed. That is, A is given no reason to suppose that B is in competition with A. Nor, however, is A given any priming to suppose that B will act cooperatively, which is why the context is described here more neutrally as ‘non-competitive’. (N.B. The original numbering has nevertheless been retained for the remaining steps in order to facilitate easier points of comparison.)

AnalysisofChimpanzeePerformanceinNon-CompetitiveContextsofTheObject-ChoiceTask

1.BELIEF[BisincompetitionwithA]

2.BELIEF[thereisatreathiddenunderoneofthebuckets,a,b,c]

3.DESIRE[Awantsthehiddentreat](from(2))

4.PERCEPTION[Bisgesturinginparticulardirection]

5.ACTION[A’sgazefollows(4)tobucketa](from(4))

6´.ACTION[Aexaminesrandombucket](from(2)and(3))

Tomasello’s overall claim here is that A doesn’t comprehend the meaning of B’s pointing, but, rather, is merely capable of following the direction of pointing. As he puts it

203

[Tom08:39]: ‘…following the directionality of the point was not the problem; they just did not seem to understand its meaning, its relevance to their search for the food.’17 A’s attention being drawn to bucket a provides A with no further information to challenge A’s belief that (2) the treat is hidden under any one of the buckets. It seemingly doesn’t occur to A that what B is pointing to may lie beyond A’s perceptual field. Presumably, however, this is not because A does not realise that the hidden treat is hidden (as is presumed by (2)). It’s just that A’s tendency to follow the direction of the pointing doesn’t engage with A’s holding (2). As Tomasello glosses the chimpanzee’s inner monologue [Tom08:39], ‘OK. There’s a bucket. So what? Now where’s the food?’ Furthermore, in the absence of the understanding that (1) B is in competition with A, it appears also that A does not form the belief that (6) B wants whatever is in bucket a for herself (i.e. what I referred to above as ‘the competitive inference’).

4. Human Performance in Non-Competitive Contexts of The Object-Choice Task

Finally, we must account for how human infants around a year old are able to do what chimpanzees surprisingly cannot. Our primary challenge in this regard is to account for what Tomasello calls the cooperative inference in a manner that is consistent with our analysis thus far. This is because, while it is not my intention to call into question Tomasello’s extraordinary findings as such, nonetheless, Tomasello has different aims from those set here. Given that we are pursuing a non-trivial, non-question-begging, account of the emergence of semantic meaning as it coincides with the concept of objectivity, the burden of Morgan’s Canon falls more heavily on the account to be developed herein. In particular, we must be sure not to presume that the human infant understands the meaning of their conspecific’s gestures. The account to be developed must (ideally) be, on the one hand, explicable entirely in non-semantic terms, but, on the other, one that suffices to exhaustively determine Tomasello’s cooperative inference.

17 Incidentally, we have some reason to infer from his talk here of ‘relevance’ that, at least, by Tomasello’s reckoning, the question raised at the close of the previous sub-section, as to whether chimpanzees (7) see B’s gesturing as indicative of the location of the food, or (7´) they simply want whatever B wants, is answered in favour of (7).

204

In this light, my strategy will be to begin by (uncharitably) taking Tomasello at his word in employing his glossed characterisation of the cooperative inference, and then systematically critiquing this until we reach a minimalistic disaggregation that, on the one hand, meets the demands of Morgan’s Canon, but on the other, has explanatory substance sufficient to account for human performance. Ultimately, I aim to show that in order to explain human performance, a basic structure of inferences is required, and that it closely parallels that given by Brandom in his Inferentialistic account of the concept of objectivity introduced in the previous section. As I explain subsequently, my claim here is not that human infants possess the concept of objectivity; only that they are seemingly capable of minimalistic inferences that structurally parallel those demanded by Brandom’s full-blooded inferentialistic account. The plausibility of this claim and its significance for my thesis in general will be considered in the subsequent section.

* * *

According to Tomasello, human infants in non-competitive contexts of the object-choice task are readily able to follow the pointing gestures of their conspecifics to find the hidden food. As he explains [Tom14:52 (italics added)]:

…prelinguistic infants of only twelve months of age trust that the adult is pointing out to them something relevant to their current search – they comprehend the informative motive – and so they know immediately that the pointed-to bucket is the one containing the reward.

By contrast, chimpanzee performance is so remarkably different because, according to Tomasello [Tom14:52 (italics added)]: ‘…they do not make the cooperative inference, “He wants me to know that the food is in that bucket”’.

If we denote a human subject by ‘A*’, then as a first pass, it might be tempting to suppose that Tomasello’s cooperative inference could be captured by the following Carruthers-style analysis:

205

AnalysisofHumanPerformanceinNon-CompetitiveContextsofTheObject-ChoiceTask

1.BELIEF[BisincompetitionwithA*]

2.BELIEF[thereisfoodhiddenunderoneofthebuckets,a,b,c]

3.DESIRE[A*wantsthehiddenfood](from(2))

4.PERCEPTION[Bisgesturinginparticulardirection]

5.ACTION[A*’sgazefollows(4)tobucketa](from(4))

6*.BELIEF[BwantsA*toknowthatthefoodishiddenunderbucketa] (from(2), (4)&(5))

7.BELIEF[thefoodis(likely)hiddenunderbucketa](from(6*))

8.ACTION[A*examinesbucketa](from(3)&(7))

Steps (1) to (5) exactly mirror the account given of chimpanzee performance under similar circumstances. In the case of humans, however – rather than moving, like chimpanzees do, to examine the buckets randomly – instead (6*) prima facie captures Tomasello’s gloss on the cooperative inference, with A* inferring that B wants A* to know that the food is hidden under the bucket indicated. If we were to accept this characterisation (which I shall argue momentarily that we should not), then, following Tomasello’s gloss, A* subsequently accepts what B wants A* to know on the basis of (6*), and so believes for itself that (7) the food is hidden in bucket a, such that A* will then (8) examine the indicated bucket for itself.

Let’s start with the core of the cooperative inference, (6*) ‘B wants A* to know x’. Such a belief is highly cognitively demanding. Firstly, it supposes that A* possesses a Theory of Mind in attributing to B desires. While it was acknowledged in Chapter Four that even chimpanzees possess an albeit lower-order Theory of Mind, the content of this particular belief ascription is quite sophisticated. This is because, on the one hand, in order for A* to first be capable of believing that B wants A* to know something, it presupposes that A* possesses a metacognitive concept of its own mental states (i.e. what A* itself knows and doesn’t know). But then, in order to be capable of believing that this is what B wants A* to know, A* must also be capable of attributing to B this belief and desire. Such high-order metacognitive thinking is not only implausibly attributable to a year-old human infant, but, moreover, a lower-order alternative appears ready to hand, i.e.:

206

6*.BELIEF[B’sgesturingissuggestivethatthefoodishiddeninbucketa](from(2),(4)&(5))

Such a belief doesn’t attribute any Theory of Mind, nor is it metacognitive. On this reading, A* does not attribute beliefs to B, let alone metacognitively understand the content of B’s belief to pertain to A*’s own beliefs. Rather, it is the significance of B’s gesturing alone that carries the explanatory weight.

In fact, this version of 6* may be parsed further. After all, that A* sees B’s gesturing as attracting A*’s attention to bucket a does not necessarily entail that A* sees B’s gesturing as meaning ‘the food is hidden in bucket a’. If such meaning were somehow intrinsic to the gesture, then some further story would need to be told to explain how the meaning of similar such gesturing motions might equally mean others things, such as ‘Look out, a snake!’ or ‘What’s that?’. To make this more apparent, (6*) may be better characterised as:

6*.BELIEF[B’sgesturingissuggestivethatbucketahasanoteworthyquality](from(4)&(5))

In this way, A* is seen to draw on (4) & (5) in coming to believe (6*) that B’s gestures are suggestive, but ensures that any role played by (2) in affecting A*’s performance is not here conflated in such a way as to suggest that A* sees the gesture as meaningful. That is, A* takes B’s gesture only as suggestive that, in this case, bucket a has a quality about it that is worth pointing out.

There are two things to say about this lattermost version of (6*):

i) The first is that it is equally able to capture the shared chimpanzee capacity to follow pointing gestures as at least attention grabbing (though likely not informative). Recall that Tomasello glosses the chimpanzee’s reaction [Tom08:39]: ‘OK. There’s a bucket. So what? Now where’s the food?’. This is significant, in that the initial characterisation of the cooperative inference made it seem that the human capacity was completely distinct from the

207

chimpanzee pattern of inference. In the analysis of chimpanzee performance in non-competitive contexts of the object-choice task, A straightaway moved from (5) looking at the bucket indicated, to (6´), examining the buckets randomly. Although (6*) would be (and remains) redundant in accounting for A’s performance in arriving at (6´), it more plausibly places chimpanzee behaviour as falling on a continuum with human capacities. That is to say, both humans and chimpanzees are capable, regardless of whether the context is competitive or not, of inferring on the basis of B’s gesturing that it is suggestive that there is something noteworthy to what is being pointed at, whether or not they come to accept the suggestion for themselves.

Admittedly, Tomasello himself writes of the cooperative inference as not only distinct, but entirely novel [Tom14:52 (italics added)]: ‘…this new cooperative way of communicating […] created a new kind of inference…’. Rhetorical flair aside, however, the history of the philosophy of biology warns against such anthropocentric claims to uniqueness, and so, I think it is worth considering whether the cooperative inference might be better parsed as belonging to a continuum shared with chimpanzees and other animals.

ii) The second point, related to the first, is that the original inference to (7) is now lost. That is, there is no reason to suppose that B’s gesturing being suggestive that bucket a has a noteworthy quality is by itself sufficient for either A or A* to infer that it is the food that is hidden under the indicated bucket. A further inference is required. This is important because it helps explains why chimpanzees perform in the manner they do, examining random buckets, even though they are capable of not only following the direction of pointing gestures with their gaze, but seeming of being perplexed when their attention is drawn to a bucket that B’s gesturing suggests has a noteworthy quality, but no such quality is apparent to the chimpanzee.

It is also significant in that, without (2), it is clearer that humans are able to capitalise on the pointing gestures of their conspecifics outside of contexts where they are already engaged in goal-oriented activities, such as the object-choice task. To illustrate this capacity, Tomasello introduces a hypothetical scenario where a human conspecific makes an unsolicited, un-contextualised pointing gesture towards a nearby bush, which, unbeknownst to ‘you’, contains edible berries. Tomasello continues [Tom14:53]: ‘So you ask yourself: why does he think that this bush is relevant for me – and this makes you look harder for something that is indeed relevant – and thus you discover the berries’. From this, it is enough that a

208

conspecific merely make a pointing gesture for the human to subsequently make the cooperative inference (or it cognitive-procedural equivalent). No presumption that the human subject is already engaged in an goal-based activity is required, and as such, the human need not have any idea of what its conspecific’s pointing gestures might be suggestive of. Instead, it seems that if and when human infants see their conspecific pointing seemingly to nothing, they still will wonder what it is over there that is suggested to be noteworthy. In this way, (6*) appears to better capture human performance both in and outside of the object-choice task.

* * *

With the above considerations in hand, we must now explain how our new version of (6*) might lead humans (but not chimpanzees) to examine the indicated bucket. In this regard, perhaps an initially tempting suggestion might be that humans are more ‘trusting’ than their ancestral counterparts.

By this, I do not mean that, phylogenetically, humans exhibit higher levels of various forms of social cooperation, i.e., economic, sexual, etc. While it is likely true that such broader cooperative tendencies are necessary complements to the kind of inferential ‘trust’ that we are considering here (for without certain levels of social cooperation being generally shared in a community, a heightened tendency to follow one’s conspecific’s suggestive gestures would not be evolutionarily advantageous), our present concern is only to explain how the humans are capable of capitalising on pointing gestures to non-salient aspects of their environment in non-competitive contexts, such that joint attention proper might be attributed.

With this in mind, to explain human performance by recourse to the notion of ‘trust’, would be to say that A* will simply move from believing (6*) that B’s gesturing is suggestive that bucket a has a noteworthy quality, to (8) examining the indicated bucket. While such a tendency would certainly abide by Morgan’s Canon, it is, however, unlikely. Not merely because we presume the human is cognitively more sophisticated than the chimpanzee, but, relatedly, because humans are not so gullible. Even though humans’ (and chimpanzees’) gaze is inclined to follow pointing gestures, we are not passively lured by such pointing irrespective of circumstance. To illustrate, imagine seeing someone pointing suggestively at

209

the empty sky; while it is likely that we would look up at least momentarily, it is unlikely that we would remain staring blankly for long. Now, it might be that such overriding limits to our credulity only occur post hoc – i.e. we do always immediately ‘trust’ instances of (6*) insofar as we see them as suggestive of something, but we only override them if certain subsequent conditions are met. But consider, not only is it possible to pre-emptively not trust a conspecific if they are deemed untrustworthy, but moreover, such credulity conditions already look to be at play in the object-choice task and, indeed, were appealed to in explaining chimpanzee behaviour. That is, both A and A* perceive and, so, believe that none of the buckets have any noteworthy qualities. This can be better appreciated by parsing (2), as follows:

2´.BELIEF[buckets,a,b,c,havenonoteworthyqualities](from2)

2´´.BELIEF[thefoodcouldbehiddenunderanyoneofthethreebuckets](from2)

Recall, chimpanzees do (5) direct their gaze and (6*) attention in accordance with (4) the direction of perceived pointing gestures even in non-competitive contexts and even to non-salient aspects of their environment. However, in the case of non-competitive contexts of the object-choice task, it appears that (2´) their perceptually based belief that the bucket B is gesturing towards has no noteworthy qualities is stronger than (6*) B’s suggestion to the contrary. In this way, (2´) overrides (6*) such that the chimpanzee examines the buckets randomly.

In the case of humans, however, while it is reasonable to presume that they also experience the cognitive tension of (2´) and (6*), nonetheless, it looks like A*’s belief that (6*) B’s gesturing is suggestive that there is in fact something noteworthy about bucket a, one way or another, trumps (2´). The cognitive vacillation leading to such a resolution (one way or another) might be accounted for as:

6**.DISPOSITION[A*vacillatesbetween(2´)and(6*)](from(2´)&(6*))

Here, the cognitive state is described not as a belief, but rather as a disposition, whereby the cognitive tension created by (2´) and (6*) needs to be resolved in favour of one of the two beliefs.

210

Note that it is besides the point here to provide any account of the manner in which the cognitive vacillation takes places (e.g. whether it is an oscillation between (2´) and (6*), or whether the increased strength in one is proportional to the decrease in the other, etc.). Such matters are questions for psychologists and neuroscientists to determine empirically.

Again, it is plausible that chimpanzees also share some degree of (6**), but appear to resolve it in favour of their perceptually based belief that (2´) bucket a has no noteworthy qualities. To describe it as a vacillating disposition, then, allows for either of the two outcomes to emerge, without needing to appeal to any potentially metacognitive belief that (2´) and (6*) contradict one another.

* * *

Why it is that humans uniquely ‘trust’ in (6*) the suggestive nature of their conspecific’s gestures over (2´) what they can see for themselves, is a matter beyond the scope of this thesis, though it forms the overarching focus of investigation for Tomasello’s own work (see [Tom08] & [Tom14]). Nonetheless, a number of possible contenders present themselves specifically with relation to explaining human performance in non-competitive contexts of the object-choice task:

i) Firstly, it may be that humans possess a ubiquitous standing dispositional belief:

1´.BELIEF[BiscooperativewithA*]

Such a belief would help better explain why A* is able to see (6*) as relevant to A*’s engagement in (2) the object-choice task. The presence of (1´) might allow A* to readily resolve (6**) in favour of (6*) over (2´) insofar as believing B to be in cooperation with A* would better contextualise B’s gesturing as being relevant to A*’s current goal in the object-choice task.

211

I think it fair to say that this explanation fits best, not only with Tomasello’s original characterisation of the cooperative inference, but also with his extended evolutionary account of human cooperation in general (see [Tom08]). Indeed, as I acknowledged previously, such general cooperative tendencies would likely have to be common for the ability to capitalise on pointing gestures to be advantageous, for otherwise such a tendency would lead the individual to being duped.

That said, it is important to note that it would be question begging to mistakenly suppose that a standing dispositional belief that one’s conspecifics are generally cooperative could alone sufficiently explain phenomena such as human performance in non-competitive contexts of the object-choice task. This is because the phenomena that we are attempting to account for is – in conjunction with the tendency to gesture informatively (as explored in Chapter Five) – itself taken to constitute the specific form of cooperation that we are concerned with. It wouldn’t do to say that the cooperative inference is possible because humans cooperate. To illustrate this point, take, for instance, the fact that humans look at their counterpart more often and for longer durations than chimpanzees do (see [Tom08:179–80]). While, as has been acknowledged, such tendencies are suggestive and supportive of general cooperative phylogenetic tendencies, nonetheless, this tendency clearly cannot be called upon directly to explain the human capacity to capitalise on altruistically informative gesturing. Firstly, insofar as chimpanzees also pay attention to their counterparts’ gaze (albeit less frequently and for shorter durations) (see [Tom08:36–7]), a story would still need to be told in accounting for the threshold that allows humans but not chimpanzees to capitalise on non-competitive gestures. But, more significantly, one could not call on such evidence of heightened parallel attention as itself indicating to A* that B is ‘cooperative’ such that A* might then understand that B’s gesture is likely suggestive that the food is hidden under the bucket indicated. To do so would be to beg the question. A*’s ability to capitalise on B’s gestures is what we are taking as the cooperative complement of B’s cooperative gesturing, so this capacity cannot be itself explained by saying A* understands B’s gesturing as cooperative. On our account, A*’s cooperative ‘understanding’ lies solely in successfully obtaining the food. Similar such reasoning applies to other cooperative capacities. While they may support and even be necessary for human performance in non-competitive contexts of the object-choice task (and other related tasks), they cannot by themselves sufficiently explain this human capacity.

212

ii) It may be that humans possess greater inferential abilities in bringing (2) the context of the object-choice task itself to bear on the relevance of (6*) B’s gesturing, perhaps even without any need to appeal to (1´) A* believing B to be cooperative. That is, A* is just sufficiently able to infer that the noteworthy quality that B’s gesturing is most likely suggestive of, is that bucket a has the food hidden under it.

iii) Another possible factor is human cognitive architecture is more suited to abstraction. Where chimpanzees readily default to perceptual based beliefs, such as (2´), humans are capable of putting more weight on the suggestive nature of behaviours such as pointing gestures.

iv) Finally, it may be that the key lies in humans having a greater dispositional tendency to vacillate between standing beliefs given contrary beliefs. Obviously this doesn’t by itself explain why A* should resolve (6**) in favour of (6*) over (2´) – meaning that one or more of the above considerations will also likely play a part – but, perhaps what is distinctive of humans is, not so much trust in their conspecifics, but rather, the capacity to deliberate more deeply about their own beliefs.

My own suspicion is that it will be likely a combination of all of the above, but, as was said, such matters go well beyond the scope of this thesis, and, regardless, the demand for (6**) stands whichever way the answer lies. Indeed, it is important to note that this true also of Tomasello’s original characterisation of the cooperative inference, for even if A* were to believe that B wanted it to know that the treat was hidden under a particular bucket, the question still stands as to why A* should then come to believe that this is where treat is hidden. (Unsurprisingly, it can’t be because of the ostensibly factive state of B’s wanting A* to know that the food is hidden under bucket a.) In this light, (6**) proves its standing irrespective of my proposed version of (6*) as a more minimalist alternative to Tomasello’s original cooperative inference.

* * *

213

Once A*’s disposition (6**) to vacillate between (6*) and (2´) subsequently falls in favour of (6*) then, again (as was the case with chimpanzees also (See Section III.2), the account being developed seems to be able to go either of two different ways:

i) A* infers that (7) the food is (likely) hidden under bucket a, such that A* no longer believes (2´´) that the food could be hidden under any one of the three buckets.

ii) A* moves straight from (6*) to (8), examining the bucket that B’s gesturing suggests has a noteworthy quality, without forming any beliefs about the likely whereabouts of the food that it desires. Rather, it is just motivated to investigate the bucket indicated by B, because (6*) B’s gesturing being suggestive that it has a noteworthy quality is a stronger belief than A*’s earlier, now faded dispositional belief that (2´) none of the buckets have any noteworthy qualities about them (despite also believing (2) that there is food hidden in one of the buckets).

For the purposes of the present argument, nothing hangs on which inference A* makes. Nonetheless, it’s worth noting that, unlike the case of the chimpanzees where the evidence seemed more balanced between the two options, in the case of humans (i) seems to be clearly the more plausible option. For one, it is unlikely that humans could or should be so easily distracted from their goal-oriented activity. That is, A* is actively pursuing their belief that (2) there is food hidden under one of the buckets; it is unlikely that they should simply forget this goal, just because (6*) B’s gesturing is suggestive of something that A* cannot perceive or, in the case of (ii), does not even conceive. Overall, then, (i) seems to provide a better explanation of why A*, who is interested in finding food, will look under the bucket B is gesturing at. In this light, (i) also suggests an explanation of why and when A* will ‘trust’ instances of (6*) their conspecific’s suggestive gesturing. A*’s beliefs will have certain relative strengths or level of certainty. If (6*) is suitably stronger than (2´), then A*’s belief in (2´´) will be extinguished. At that point, A*’s belief in (6*) will combine with their background interest in food to generate (7) the belief about the food. So the explanation in (i) provides the resources for explaining why the (6*) will often – but not always – generate an inference to (7). In cases where A* has reason to doubt the reliability of B’s gestures, A*’s credence in (6*) may not be strong enough to defeat (2’’), and (7) and (8) will not automatically follow. In contrast, if the correct explanation of A*’s behaviour is (ii), then the strength of A*’s credence in (6*) will not be relevant to explaining A*’s

214

behaviour. Instead, whether or not A* moves to examine bucket a, b or c will be entirely determined by A*’s random dispositions about which bucket to check first.

It is worth pointing out that this same issue stands with regard to Tomasello’s original account also. That is, even with the cooperative inference, ‘B wants A* to know…’, in play, a story needs to be given as to at what point A* defers to B’s desires over A*’s own belief that (2´´) the food could be hidden in any one of the buckets.

Here, then, we have our complete, developed account:

AnalysisofHumanPerformanceinNon-CompetitiveContextsofTheObject-ChoiceTask

1.BELIEF[BisincompetitionwithA*]

2.BELIEF[thereisfoodhiddenunderoneofthebuckets,a,b,c]

2´.BELIEF[buckets,a,b,c,havenonoteworthyqualities](from2)

2´´.BELIEF[thefoodcouldbehiddenunderanyoneofthethreebuckets]

(from2)

3.DESIRE[A*wantsthehiddenfood](from(2))

4.PERCEPTION[Bisgesturinginparticulardirection]

5.ACTION[A*’sgazefollows(4)tobucketa](from(4))

6*.BELIEF[B’sgesturingissuggestivethatbucketahasanoteworthyquality](from(4)&(5))

6**.DISPOSITION[A*vacillatesbetween(2´)and(6*)](from(2´)&(6*))

7. BELIEF [the food is (likely) hidden under bucket a] (from (6**) if (6*) is sufficientlystrongerthan(2´)suchthatA*nolongerbelieves(2´´)butstillbelieves(2))

8.ACTION[A*examinesbucketa](from(3)&(7))

215

III. SYNTHESIS

In this final section, I argue that the Inferentialistic structure Brandom argues is necessary for our possession of the concept of objectivity bears a promising structural analogue with the Carruthers-style account developed in the last section. Specifically, I claim that the key cognitive inferences underlying human performance in non-competitive contexts of the object-choice task parallel the inferentialistic norms that Brandom argues underlie our possession of the concept of objectivity. I argue that the former can, in this way, be seen as forerunners for both possession of the concept of objectivity and semantic language use.

1. Parallels

The analysis developed in the previous section takes the following two stages as vital:

6*.BELIEF[B’sgesturingissuggestivethatbucketahasanoteworthyquality](from(4)&(5))

6**.DISPOSITION[A*vacillatesbetween(2´)and(6*)](from(2´)&(6*))

Again, (6*) and (6**) are plausibly shared with chimpanzees also, though perhaps to a weaker degree. Humans and chimpanzees both appear to experience some degree of vacillation between the beliefs: (2´) that none of the buckets have any noteworthy qualities about them; and, (6*) that B’s gesturing is suggestive that bucket a has a noteworthy quality. In the case of chimpanzees, (6**) resolves in favour of A’s perceptually based belief, (2´), such that A then examines the buckets randomly. In the case of humans, however, (6**) leads to A* resolving in favour of (6*) the suggestive nature of their conspecific’s gesturing, such that A* arguably comes to believe (7) that the food is (likely) hidden under the bucket that B is gesturing towards and, so, no longer believes (2´´) that the food might be hidden under any one of the buckets, even though it still generally believes (2) that it is engaged in the object-choice task.

216

What makes (6*) and (6**) distinctive for humans is, then, likely to be a matter of degree. As was speculated in the previous section, there are a range of potential factors that might contribute to explaining how humans are capable of resolving in favour of their conspecific’s suggestive gesturing over the evidence of their own eyes. These include: a standing dispositional belief that our conspecifics are cooperative with us; a greater capacity for inference (to the best explanation); a greater capacity for abstract symbolism in seeing the gesture as meaningful; and, a greater capacity for vacillation between standing beliefs. Although answering which one or combination of these factors determine characteristically human performance is a matter beyond the scope of this thesis, nevertheless, whichever way the answer lies, it remains that (6*) and (6**) form the core of the account.

Given, then, that (6*) and (6**) do the lion’s share of the work in accounting for human performance in non-competitive contexts of the object-choice task, I wish to propose that interesting parallels can be observed, respectively, with the two key inferentialistic distinctions Brandom argues to be necessary and sufficient for possession of the concept of objectivity. (I will subsequently suggest, in the third sub-section, that the former are likely forerunners for the latter.)

i) Firstly, I want to suggest that (6*) parallels Brandom’s account of attributing a commitment without undertaking it for oneself. Recall Brandom’s proposed generic formulation of such a statement:

‘S believes of t that (𝛟)t’.

Similarly, when (6*) is taken in conjunction with (2´), it can be reasonably said that A* believes that B’s gesturing is suggestive ‘of’ bucket a that is has a noteworthy quality. This is so insofar as A* itself does not see or (yet) believe that any of the buckets have any noteworthy qualities (from (2´)). Obviously, for reasons of parsimony outlined earlier, (6*) does not require that A* meta-representationally attribute a belief to B – i.e. ‘that B believes of bucket a that it has a noteworthy quality’. However, as I shall argue momentarily, it is plausible that the general capacity to make inferences to beliefs such as exemplified by (6*) is a forerunner to the capacity to later make the distinction that Brandom identifies between attributing and undertaking commitments.

217

ii) Secondly, I claim that the inference to (6**) can be seen to parallel Brandom’s distinction between commitments and entitlements. Again, recall the generic formulation of how the distinction between commitments and entitlements can come apart such that one may doubt one’s entitlements to the commitments that one currently holds, that is:

‘I am committed to the claim p, but I am not committed to the claim that I am entitled to my commitment that p’.

Similarly, the disposition (6**) in humans to vacillate between their beliefs is, as I shall argue momentarily, analogous to the capacity to doubt or question one’s entitlements to one’s commitments. Briefly, this is because, in resolving in favour of (6*), it suggests that (6**) amounts to an implicit awareness of tension between A*’s commitments, (2´) and (6*), and, as such, arguably provides the kind of cognitive friction needed for it to later question its entitlements to said commitments ‘inferentialistically’ (as Brandom would have it).

2. Disclaimers

Before turning to argue that (6*) and (6**) are likely forerunners to the inferential possession of the concept of objectivity as given by Brandom, let us first be clear what I am not claiming.

I am not claiming that either (6*) or (6**) are instances of their respective Inferentialistic counterparts. As was critically emphasized in Section II.1 of this chapter, the inferential distinctions central to Brandom’s Inferentialism are taken to be fully normative (even to the questionable extent that he further claims his account to be non-naturalistic), whereas what is exciting about Tomasello’s experimental findings regarding the remarkable differences between human and chimpanzee performance in the object-choice task is that the human subjects in question are pre-linguistic infants around only one year old. The significance of this is that their capacities are more plausibly taken to be revelatory of non-normative,

218

phylogenetic characteristics of the human species. That is, the infant capacity to be able to captialise on their conspecific’s altruistic gesturing is thought to be non-cultural and, so, non-normative. When they undertake the object-choice task, they are not following rules dictated by societal norms, including any arising from language-use, artefact or universal.

By contrast, with regard to Brandom’s Inferentialism, the various inferential compatibilities and incompatibilities that are taken to hold between statements such as (1) ‘The swatch is red’ and (2´) ‘I am committed to the claim that the swatch is red’, are indeed taken to be fully normative. That is, must belong to a rational community of language users such that one will be able to follow the normatively governed inferential rules that, according to Brandom, give our statements any meaning at all.

Indeed, this distinction is vital to the story being developed here because it is essential in providing a non-trivial account of the role played by language in our possession of the concept of objectivity. As has been stressed throughout, the criterion that the account should be non-trivial has recommended that we look to the emergence of language and the concept of objectivity as arising simultaneously out of non-semantic interactions. In particular we have been seeking the point (or at least, a point) where an entering wedge might be found into the intentional circle. This was the primary characteristic, both, of Davidson’s triangulation arguments and Brinck’s emphasis on joint attention with regard to non-salient features of a shared environment as indicating intersubjective awareness proper.

Here then, also, the claim is that the Carruthers-style analysis of the cognitive processes underlying human performance in non-competitive contexts of the object-choice task is non-normative, non-semantic, and non-intentional. It is not being claimed that human infants possess the concept of objectivity at this stage. Nor is it claimed that they thereby possess full intersubjective awareness – i.e. an explicitly awareness of the content of their conspecific’s mental states. The inferences from (6*) to (6**) and on to (7) were argued to be preferable to Tomasello’s original characterisation of the cooperative inference in part because they did not ascribe a high-order, meta-cognitive Theory of Mind to A*. Instead, the minimalist analysis, developed in accordance with Morgan’s Canon, was intended to reveal what basic cognitive processes underlie joint attention proper. It’s not that humans have magical mind-reading capacities. It’s only that we are more likely to ‘trust’ in deferring to the suggestive behaviour of our conspecifics, such that cooperative interactions, like informative gesturing may then emerge.

219

Indeed, there is a central difference between Brandom’s Inferentialistic account of the objectivity of propositional content and the Carruthers-style account developed of human performance in the object-choice task, and it is one that further highlights the distinction between the normative and non-normative nature of the respective inferences given in each account. Brandom takes it that, if the subject, A*, is successfully able to make the inference, ‘S claims of t that (𝛟)t’, then it will be so on the basis of recognising the inferences that

follow for S (but not A*) according to S’s attributed commitment that (𝛟)t. Given,

however, that the subject does not themselves undertake the commitment to (𝛟)t, the inferences that are taken to apply to S, will not necessarily be taken as applicable to the subject themselves. For instance, suppose I attribute to you a commitment to the claim that of bucket a it has the food hidden under it, but (for whatever reason) I do not undertake that same commitment myself. Then, if I take us to be in competition with one another for the food, I may not bother looking under bucket a as I can infer that you will likely look there yourself and I will increase the likelihood that I will find the food before you if I look under one of the other two buckets first. By attributing a commitment to you without undertaking it for myself, I thereby appreciate that you will make different inferences than me.

By contrast, the non-normative inferences that the human infant is credited with having made does not imply that it will be able to differentiate between the subsequent inferential relations attributable to its conspecific and the various inferential compatibilities open to itself on the basis of its own undertaken commitments. Rather, all that is required for (6*) is that the human infant take its conspecific’s gesturing as suggesting of bucket a that it has a noteworthy quality, even though the infant itself cannot see or understand what that quality might be.

Much the same considerations apply to the second inference and its parallels with Brandom’s distinction between entitlements and commitments. Recall that the important point for this latter distinction is, again, that different inferential compatibilities and incompatibilities follow from each. For instance, my being committed to the claim that ‘the swatch is red’ does not guarantee that I am entitled to the claim ‘the swatch is red’ (for instance, if I was red-green colour blind). By contrast, it is argued that, in (6**), the cognitive tension generated by the two beliefs – i.e. that A* believes a has no noteworthy quality, and that B’s gesturing suggests that bucket a does have a noteworthy quality – is sufficient only to lead

220

the human subject to vacillate between the two beliefs in such a manner that it comes to be resolved, for one reason or another, in favour of the suggestive nature of B’s gesturing. The key difference between the two, then, is that Brandom’s account requires meta-cognition insofar as subject is capable of self-attribution (i.e. ‘I am committed to the claim p, but I am not

committed to my entitlement to the claim p’, or, more colloquially, ‘I think/believe p’), whereas the disposition to vacillate between standing beliefs that a human infant is plausibly capable of amounts to little more than an ability to adjust the strength of one’s own beliefs in light of conflicting evidence so as to reach a resolution. As was emphasized previously, no Theory of Mind or meta-cognition is required. Equally, no overarching concept of mistake or error is required to explain the performance of human infants here. It is enough merely that they have the first-order disposition to vacillate between the two particular beliefs (2´) and (6*). That is to say, their ability is here only taken to be context specific and not necessarily generalisable.

3. Forerunners

With the above qualifications in mind, I do wish to argue that the kind of non-normative inferences that are the key to explaining human infant performance in non-competitive contexts of the object-choice task are also likely forerunners to the kind of fully-normative inferences that Brandom argues are necessary to our possession of the concept of objectivity. Specifically, inferences to beliefs such as (6*) are likely precursors to the inferentialistic distinction that Brandom makes between attributions and undertakings. Whereas, inferential dispositions like (6**) are likely precursors for the kind of self-critical evaluations subsequently made possible by the distinction between commitments and entitlements. Allow me to elaborate each pairing in turn. i) (6*) captures the human infant the ability to see a conspecific's altruistically informative gesturing as being suggestive of the environment that it possesses certain characteristics even though one may not be able to see it for oneself. Inferences to the kind of belief exemplified by (6*) are arguably forerunners for the subsequent capacity to attribute commitments to a subject without undertaking them for oneself. That is, inferences to the kind of beliefs exemplified by (6*) are precursors to the distinction between attributing and undertaking. As I shall argue presently, the content of (6*) can, in light of it being simultaneously held along with (2´), be equally rephrased as, 'B's gesturing is suggestive 'of' bucket a that it has a

221

noteworthy quality'. Admittedly, the demands of Morgan's Canon meant that there was no need to phrase it as such in our original analysis. However, if we can justify doing so here, we will demonstrate the parallel it shares with Brandom's formulation of attributing a commitment without undertaking it for oneself, i.e. ‘S believes of t that (𝛟)t’. The human infant capacity (shared to an extent with chimpanzees) to infer from B's pointing gestures that such gesturing is suggestive bucket a has a noteworthy quality despite A* itself believing that none of the three buckets have any noteworthy qualities, is analogous to the ability to attribute commitments to another person without undertaking them for oneself. This is because, by inferring (6*) despite also holding (2´), A* is thereby capable of simultaneously holding two beliefs that stand in peculiar tension with one another. This tension is different than most non-meta-cognitive changes in belief, even if said changes are: surprising or unsurprising (as discussed in Chapter One); amount to a recognition of mistake (as discussed in Chapter Two); or diverge from one's correlatively formed expectations regarding the likely reaction of one's conspecific to a shared environment (as discussed in Chapters Three and Four). These latter changes in belief resolve straightforwardly on the basis of the creature's current perceptions of salient aspects of their environment (or lack thereof, as in the case of the dog who chases the cat up the wrong tree). For instance, if I believe that I have a coin in my pocket, it is by then seeing that my pocket is empty that I come to be surprised by my unexpected change in belief that I (now) don't have a coin in my pocket. By contrast, to infer (6*) in the face of (2´) does not result in such a straightforward change in belief. We can see this insofar as (6**) the disposition to vacillate between the two arises. Chimpanzees look momentarily perplexed as to why B should be gesturing at a bucket that, to A, has no noteworthy qualities. Whereas human infants (in contrast to chimpanzees) are able to subsequently defer to the suggestive content of (6*) in choosing bucket a for themselves. Admittedly, for chimpanzees, the subsequent moment of confusion that results from the disposition (6**) to vacillate between the two beliefs, does resolve in favour of its own perceptually-based belief (2´) that there is no noteworthy quality to any of the buckets. But the point stands: insofar as (6*) results in the momentary confusion of (6**), we have evidence that even A sees B’s gesturing as at least somewhat suggestive of a non-salient aspect of the environment, even given this is insufficient for A to capitalise on so as to reliably obtain the hidden treat in non-competitive contexts.

222

In the case of human infants, however, the ability to capitalise on the altruistically informative gesturing of their conspecifics reveals that they see such pointing gestures as distinctly suggestive. By ‘suggestive’ I mean that A* is able to understand from B’s gesturing that it is at least potentially the case that bucket a does have a noteworthy quality despite A* not (yet) believing it for itself. We can see this insofar as (6**) leads to A* examining (8) the bucket indicated by B. This is particularly striking in that (6*) is not a direct contradiction of (2´), and yet A* subsequently defers to the content of (6*). The fact that (6*) is not a direct contradiction of (2´) speaks to the suggestive nature of the former, especially insofar as A* defers in its favour. It does so because prima facie it seems reasonable to expect that the explicit content of (2´) would settle the matter – i.e. ‘B’s gesturing may suggest that bucket a has a noteworthy quality, but A* plainly sees that it does not’. But what occurs in the case of human infants at least, is that B’s gesturing is taken to be so particularly suggestive to A* that it is drawn not only to (6**) vacillate between the two, but to go against its own perceptually salient belief (2´). The inference to (6*) in the face of (2´) such that A* then subsequently defers in favour of B’s gesturing over that of its own perception of the buckets having no salient noteworthy qualities, reveals that A* sees B’s gestures as being in this way suggestive. And in this light, I claim, it legitimates the Brandomian paraphrasing of (6*), i.e. ‘B’s gesturing is suggestive ‘of’ bucket a that it has a noteworthy quality’. The inference to (6*) in the face of (2´), can be seen to parallel Brandom’s distinction between attributing and undertaking insofar as A* understands B’s gesturing as being suggestive ‘of’ non-salient aspects of their presumed shared environment despite A* not yet believing for itself that bucket a has any noteworthy qualities. This is subtly, though importantly different from merely taking it that B’s gesturing suggests that bucket a has a noteworthy quality. Although this is strictly speaking the content of (6*), nonetheless, it is given a distinct significance in light of A* being able to infer (6*) in the face of (2´) and to do in such a manner that A* doesn’t straightforwardly defer to either. Indeed, part of what makes Tomasello’s findings so remarkable is that chimpanzees also seem capable of some degree of (6*) and (6**), but that only humans appear capable of capitalising on B’s altruistically informative gesturing. The fact chimpanzees also appear to recognise the suggestive nature of B’s gesturing but then default to their own perceptually-based beliefs that there is nothing saliently noteworthy about the three buckets, reveals, by contrast, the extent to which human infants do see their conspecifics’ gesturing as being

223

distinctly suggestive ‘of’ the environment that it is a certain way despite the human not (yet) believing it for itself. Obviously, it would transgress the parameters we have set for our undertaking to suppose that A* understands B’s gesturing to be intentional or agential in any meta-representational sense.18 This is why the content of (6*) is explicit in stating that it is B’s gesturing that is taken to be suggestive. By characterizing the content of (6*) as ‘B’s gesturing is suggestive that bucket a has a noteworthy quality’, we are careful to avoid implying that A* believes B herself to be suggesting (let alone intending to suggest) anything whatsoever.19 Despite this, it is, however, reasonable to allow that A* sees B’s gesturing as in some sense purposeful – i.e. not an anomaly or obvious albeit lower-order error. This is fair to presume because, although (2´) stands in tension with (6*), in the case humans at least, A* defers to (6*) in examining the indicated bucket. In this way, we see that A* doesn’t dismiss B’s behaviour out of hand even though A* believes (2´). That being said, technically speaking, at this stage in A* development, B’s gesturing and B herself are taken by A* as no different in kind to any other evidential source of information in the environment (for example, dark clouds being indicative of coming rain). (See Chapter Three, Section IV for Davidson’s own discussion of this matter.) But, on reflection, this is exactly the kind of story that we hoped for. The aim from the outset was to explain how humans come to captialise on altruistically informative gesturing without presupposing that they yet belonged to the intentional circle. So, while it is true that the account given does not have A* see B’s gesturing as in any significant way distinct from any other kind of evidentially informative aspects of its environment (e.g. rain clouds), nevertheless, it does account for the process of A*’s seeing certain behaviour in their conspecifics – altruistically informative gesturing, in particular – as also being in this way informative. Our aim never was to bridge the divide between the non-semantic and semantic, but only to more precisely locate the moment where an entering wedge may be found into the intentional circle. Again, as Tomasello puts it: [Tom14:151]:

[Cooperation] does not solve Davidson’s (1982) problem of a common theoretical vocabulary that spans from ‘no thought’ to ‘thought’, but it does narrow the distance to be traversed by any one step significantly.

18 This is so despite Tomasello’s assurances elsewhere that chimpanzees, at least, do see the behaviour of others as intentional (see [Tom08:177]) 19 Obviously, to not credit A* with such abilities is not to thereby deny them, either. It’s just that, following Morgan’s Canon, our analysis does not warrant such attributions.

224

By showing how humans not only came to altruistically gesture (as discussed in Chapter Five), but how we are able to capitalise on such gesturing in non-competitive contexts, and to have done so in such a way as to also reveal striking parallels with Brandom’s Inferentialistic account of our possession of the concept of objectivity, we will have achieved our aim of providing a non-trivial account of the role played by language (or, cooperative communication, as it turns out) in our possession of the concept of objectivity. Furthermore, Brinck’s proposal (discussed at length in Chapter Five) that in order to establish joint attention creatures must be able to attend to non-salient aspects of the environment, appears borne out so long as the context is also non-competitive. This is because it is only in non-competitive contexts that we can establish that the creature is not merely taking their conspecific’s gesturing, not merely as indicative of B’s own desire (i.e. that B wants whatever is the bucket they are reaching for), but rather that ‘of’ bucket a that is has a noteworthy quality. For A* to believe (6*) that B's gesturing is suggestive of bucket a that it has a noteworthy quality, despite A* holding (2´) that it does not, is not equivalent to Brandom's distinction between attributing and undertaking such that one may thereby attribute a commitment to another person yet without undertaking it for oneself. It is, I claim, nonetheless, a significant forerunner to such a capacity insofar as it allows A* to hold a belief (6*), which, although not a direct contradiction of (2´), is nonetheless in peculiar tension with (2´) owing to the suggestive nature of (6*) that bucket a does have a noteworthy quality, despite A* not believing this for itself. Again, let me stress that I am not suggesting human infants possess a meta-representational Theory of Mind whereby they attribute beliefs to their counterparts; (6*) is only that it is B's gesturing that is suggestive that bucket a has noteworthy quality. Nor is it the case that (6*) & (2´) directly contradict one another. Rather, (6*) is remarkable precisely in that it is seemingly able to generate cognitive tension in A* (as evidenced by the subsequent inference to (6**)) despite (6*) and (2´) not directly contradicting one another, and without attributing any metarepresentational capacities to A*. (6*) is not equivalent to the capacity to attribute a commitment to another person without undertaking said commitment for oneself, but it is a striking analogue to it in that A* is seemingly able to believe that B's gesturing is suggestive of something that A* doesn't (yet) believe for itself. Again, as has been said throughout, the human capacity to make inferences to beliefs like (6*) is not taken to be unique to humans alone – chimpanzees also appear to see B’s gestures

225

as suggestive. The difference, then, is not one of kind, but rather a difference in degree (and likely other supporting mechanisms, such as a standing dispositional belief in humans that their conspecifics are generally cooperative). The threshold, then, that differentiates human capacity from that of chimpanzee (and likely other creatures besides) is only that (6*) leads, via its interplay with (6**) – which is also shared to some degree with chimpanzees – to (7) & (8) in allowing the human infant, but not the chimpanzee, to obtain the altruistically indicated treat. Inferences to beliefs such as (6*) are, then, I claim, likely necessary forerunners to Brandom’s inferential distinction between attributing and undertaking insofar as they enable us to capitalise on the kind of altruistically informative behaviour accounted for in Chapter Five.

* * *

ii) Turning, then, to inferentially derived dispositions of the kind exemplified by (6**), I similarly claim that these are likely precursors to the distinction Brandom makes between commitments and entitlements. Although A*’s vacillation between (2´) and (6*) does not properly amount to being committed to (2´) while, at the same time, not being committed to one’s entitlement to one’s commitment to (2´), nonetheless, it does suggest that A* possesses at least an implicit subjective awareness that the two beliefs in question stand in tension with one another. This is particularly so insofar as, in the case of human infants, A* is capable of deferring to (6*) over that of (2´). My suggestion, then, is that such an implicit awareness is a forerunner to the explicit inferentialistic awareness that Brandom argues is necessary for possession of the concept of objectivity. Inference to the kind of disposition exemplified by (6**) such that A* vacillates between the conflicting beliefs, implies that the human infant has an implicit subjective awareness that its own perceptually salient belief (2´) that there is nothing noteworthy about any of the buckets, is in tension with (6*) B’s gesturing being suggestive that there is something noteworthy about the bucket indicated. Moreover, this implicit subjective awareness by itself does not straightforwardly determine which of the two beliefs is to be resolved in favour of. In fact, as was mentioned above, prima facie it would seem more plausible that that A* would resolve the tension, like chimpanzees do, in favour of its own perceptually salient beliefs. However, as Tomasello’s remarkable empirical findings show, instead, human infants are indeed capable of capitalizing on their conspecific’s altruistically informative gesturing, so as to defer to the suggestive nature of such gesturing to non-salient

226

aspects of their environment over their own standing perceptually salient beliefs. As was explored in the Section III, the possible factors involved in fully explaining this capacity to defer in favour of their conspecifics’ gestures is varied (including, perhaps, a standing dispositional belief that our conspecifics are generally cooperative in a broad sense). And yet, the key point stands: insofar as (6**) defers to (6*) in subsequently inferring (7) & (8) and, so, examining the indicated bucket for itself, the significance of (6**) in humans is such that we have at least an implicit subjective awareness that we may not be entitled to our commitment to (2´). In this light, we can better appreciate the implicit subjective awareness in vacillating dispositions, like (6**), that the creature’s beliefs are in tension with one another. Again, this is not meta-cognitive nor does it possession of the concept error. Rather, my suggestion is that this implicit subjective awareness that one’s beliefs are in tension such that one may have the disposition to vacillate between the two to the extent that one resolves in favour of the suggestive nature of one’s conspecific’s behaviour over that of one’s own perceptual evidence to the contrary, is, I claim, a necessary precursor to subsequently becoming capable of normatively doubting the entitlements one has to the commitments that one nonetheless holds. The implicit subjective awareness of conflicting beliefs that one possesses by virtue of a disposition to vacillate between said beliefs, may provide for the subsequent potential to normatively and meta–cognitively reflect on the entitlement one has for one’s commitments. Put loosely: the ability to doubt one’s own perceptually salient beliefs on the basis of what one’s conspecific has to say is surely key to cooperative communication if anything is – even more so that altruistic informing.

4. Summation

I have stressed the difference between non-normative inferences like (6*) and (6**), on the one hand, and the fully normative inferentialistic distinctions between, both, attributing & undertaking and commitments & entitlements, on the other. However, I do not wish to imply that they do not fall on a naturalistic continuum with one another. After all, it is not that, at some point, humans come to be able to telepathically read one another’s minds. It is rather that we likely move from seeing each others’ behaviour as potentially informative without either meta-representation or meta-cognition (as in non-competitive contexts of the object-choice task), to being able to quite sophisticatedly infer each other’s intentions and

227

meanings, along with our own (as per Brandom’s inferentialistic account of the concept of objectivity). In arguing that the former are necessary precursors for the latter, I thereby hope to have made some headway in responding the challenge put to Brandom’s anti-naturalism by the likes of Dennett, Price, Kukla and Lance, as was outlined in Section II.1. That is to say, I hope to have made some headway in revealing the parallels between Brandom’s account and the non-normative inferential structure of human performance in non-competitive contexts of the object-choice task, and making it plausible that the latter are likely precursors to the former.

That said, for the likes of Brandom’s distinctions between attributing & undertaking and commitments & entitlements to emerge as fully normative aspects of our linguistic rational communities, much more than the likes of (6*) and (6**) will be required. What exactly that may be, I cannot say. Instead I defer to the wisdom of Kim Sterelny, when he writes [Ste14:75-6]:

Many accounts of hominin evolution are structured around identifying a critical breakthough, one that explains the unique features of life. […] My approach is quite different. There is no key adaption or magic moment. […] We evolved via coevolutionary interaction among a suite of cognitive, behavioural, and social capacities. One of these is our capacity to cooperate.

With this in mind, however, I nonetheless believe I have made a compelling case in supposing that the two key inferences needed to account for human performance in non-competitive contexts of the object-choice task, as revealed and emphasized by Tomasello, are likely forerunners for the fully normative inferences required for possession of the concept of objectivity, as argued for by Brandom. As we saw in the previous chapter, Tomasello provides an evolutionary account of how altruistic gestures arise as a result of increased cooperative communities. In this chapter, I hope to have improved on Tomasello’s original account of how we are able to capitalise on such altruistic gestures. It is one thing to point, it is another to understand what that pointing means. These two findings taken together, then, not only plausibly provide a substantive, non-trivial account of the role played by language in our possession of the concept of objectivity (and vice versa), but do so as a result of following the direction laid out originally by Davidson, especially with reference to his triangulation arguments (as I shall explore further in the following Conclusion).

228

As I explained to my 6-year-old son: ‘It looks like a key reason why humans have language but non-human animals don’t, is because we are willing to listen to one another on the presumption that the person talking might be trying to tell us something we don’t already know.’

229

CONCLUSION

* * *

Davidson’s Legacy

The externalist … proposes that we begin with the world that we find ourselves in, and with what either common sense or our best scientific theories tell us about it. Among the things we find are human being – ourselves – who are things that (it seems) can know about the world, can experience it, have a point of view on it. Our problem is to explain how our objective conception of the world can be a conception of a world that contains things like us who are able to think about it and experience it in the way that we do.

– Robert Stalnaker, Our Knowledge of the Internal World

I. SUMMARY

The inspiration for this thesis was found in the problem of objectivity as posed by Donald Davidson, namely: How is it that creatures such as ourselves possess the concept of objectivity – of a reality, be it ontological, epistemic, moral or aesthetic, or otherwise, that is independent of us? Following Davidson’s lead, the capacity for language was tipped as a most promising line of investigation, such that the thesis aim was refined to providing a non-trivial account of the role played by language in our possession of the concept of objectivity.

230

Part One found (perhaps unsurprisingly) that language cannot be shown to be a priori necessary for the concept of objectivity. But, neither, I argued, can it be proven a priori that there isn’t some interesting, substantive story to be told that might nonetheless explain how language by itself is still capable of supporting and even perhaps giving rise to at least a higher-order concept of objectivity.

With such hopes, we turned, in Part Two, to consider Davidson’s arguments from triangulation, systematically reconstructing the increasingly complex relations from an unobserved creature alone, to intra-reactive triangulation and, finally, interactive triangulation proper. There is good reason to suppose that such interrelations do plausibly sharpen the content of a creature’s beliefs, as interactive triangulation effectively amounts to two overlaid intra-reactive triangles. However, despite allowing for the capacity to manipulate the behaviour of conspecifics via tactical deception, interactive triangulation was by itself unable to establish mutual intersubjective awareness, and so no entering wedge into the intentional circle was to be found there.

Moving beyond the work of Davidson himself, then, in Part Three we turned our attention to more recent, subsequent work of Brinck, Tomasello and Brandom. Brink’s proposal is that, in order to establish mutual intersubjective awareness, creatures should be capable of attending to non-salient aspects of their shared environment. However, while non-salience is significant in the story of joint attention, the remarkable empirical findings of Tomasello et al., suggest that social context – as either competitive or non-competitive – can make all the difference when it comes to whether chimpanzees and human infants are able to pass Brinck’s proposed criteria. Tomasello’s evolutionary account of how altruistically informative gesturing arose, was further developed by drawing on the prior analysis of interactive triangulation. Specifically, it was proposed that, if a creature’s expectations regarding its conspecific’s likely responses to the environment grew strong enough to illicit a sympathetic response from the observing creature itself, such ‘informative’ tendencies might be selected so as to support eventual altruistic gesturing. The thesis proper was brought to a close by undertaking a Carruthers-style analysis of the non-normative inferences required to explain human performance in non-competitive contexts of the object-choice task. The analysis was seen to bear a striking parallel to the inferentialistic structure Brandom holds gives rise to the concept of objectivity proper, such that the former is arguably a forerunner for the latter. This finding not only gained some ground in naturalising Brandom’s erstwhile anti-naturalistic account, but suggests that cooperatively

231

interactive triangulation serves as the foundation of, both, the human capacity for language and our possession of the concept of objectivity.

The main findings of this thesis can be summarised as follows:

• While Davidson’s Argument from Surprise aims to establish that meta-cognitive beliefs about beliefs are necessary for any beliefs whatsoever, it was conceded that the phenomenon of surprise can (following Carruthers) be adequately accounted for by entirely first-order cognition.

• Carruthers’ account was further developed to better explain why some changes in beliefs, but not others, cause us to be surprised. This is achieved by allowing beliefs to be accompanied by differing degrees of expectation with regard to the duration of the belief.

• Davidson’s Argument from Error aims to establish that language is necessary for the concept of error, but the various scenarios proposed by Lepore & Ludwig also make it appear equally plausible that the concept of error can have ready application outside the context of language.

• The three novel scenarios I offer in response, shift the dialectic by suggesting that certain applications of the concept of objectivity can be entirely language dependent. This, in turn, emphasised the need for a more substantive account of the non-trivial role that is played (if at all) by language in our possession of the concept of objectivity.

• Davidson’s Argument from Content Determination was seen to rely too heavily on untenable scepticism which flies in the face of our best evolutionary and ethological commitments regarding the question of whether there are any grounds for our attributing determinate distal content to the perceptions of creatures either in isolation or that are unobserved.

• Although Davidson’s Argument from Content Determination is unable to establish that triangulation is necessary to attain a degree of precision sufficient to support the concept of objectivity, my reconstruction of the various levels of complexity leading to intra-reactive triangulation demonstrates how intra-reactive triangulation plausibly gives rise to greater precision in a creature’s determination of content than merely intrapersonal triangulation on the basis of, say, memory and various sensory modalities.

• Davidson’s Argument from Objectivity cannot be established, since interactive triangulation by itself is equivalent to two intra-reactive triangles overlaid one another.

232

• That said, the account of interactive triangulation I developed beyond that of Davidson’s own writing, does account for a variety of ethological phenomena, including tactical deception (following Chadha) and group hunting (following Tomasello) without recourse to meta-representational intersubjective awareness or the concept of objectivity.

• I argued against Ingar Brinck’s proposal that joint attention to non-salient features of a shared environment might indicate intersubjective awareness, citing not only the preceding account of interactive triangulation, but also Tomasello’s findings that it is only competitive contexts that chimpanzees can capitalise on pointing gestures to non-salient features of the environment.

• Tomasello’s evolutionary account of cooperative communication is not only supported by Davidsonian triangulation, but our earlier analysis suggests that interactive triangulation might give rise to altrustically informative gesturing if the expectations of each conspecific were strong enough to lead the creature to behave in response to the perceived circumstances of their conspecifics as though they were its own.

• Finally, I developed a Carruthers-style analysis of human performance in non-competitive contexts of the object-choice task which was better able to abide by Morgan’s Canon than was Tomasello’s original cooperative inference.

• Moreover, I argued that the two key, non-normative inferences of the account not only interestingly paralleled the two fully normative inferentialistic distinctions Brandom argues to be necessary for the concept of objectivity – i.e. attributions & undertakings and commitments & entitlements – but further proposed that the former are likely forerunners of the latter.

• In doing so, I not only made some headway in showing how Brandom’s anti-naturalism might be naturalised to the benefit of both accounts, but advanced a non-trivial account of the role played by language in our possession of the concept of objectivity.

233

II. SIGNIFICANCE & FUTURE RESEARCH

1. Reflection

Davidson’s early engagement with the significance of the concept of objectivity now looks (if it didn’t always) to have severely overstated the case. As Jason Bridges compellingly argues in support of the general consensus of ethological sciences (not to mention pet lovers everywhere), it is undoubtable that many (if not most or all) languageless animals think and have beliefs (see [Bri:06]). Davidson’s subsequent claim that the term ‘concepts’ should be specially reserved for cognitions where the concept of truth is not only relevant, but actually meta-cognitively applied by the creature in question (see [Dav95/04:8-9]) now looks to be hyper-intellectualised, as is convincingly argued by Camp (see [Cam09]). Let’s be sure, then, not to fall into the same trap with respect to the significance of the findings made herein. On the one hand, I do not mean to imply that possession of the concept of objectivity is essential to any and all thought or even concept use. And, on the other, I do not wish to overstate the significance of possession of the concept of objectivity in explaining human cooperation. On this latter point, there are a few things to say:

i) The first is that, while it is fitting that Tomasello characterises the difference between human and chimpanzee performance in non-competitive contexts of the object-choice task as ‘cooperative’, I am certainly not claiming that the inferences underlying this phenomenon are the key to explaining cooperative phenomena more generally. This is pertinent given that philosophers of biology (and Tomasello himself) use the term to refer to a much wider range of social phenomena, including ecological cooperation, economic cooperation, reproductive cooperation, etc, etc.. It is also likely the case that the latter provide the general conditions necessary for the former to emerge (as emphasised in Chapter Six).

ii) The second is that, while I hope to have gone some way in providing a non-trivial account of the role played by language in our possession of the concept of objectivity (and vice versa), I do not intend to imply that this account explains the key characteristic of humankind, either existentially or in terms of our evolution. Again, I defer to the sagacity of Sterelny, as quoted in the previous chapter (see [Ste14:75-6]), when he writes that the remarkable capacities of humans are likely the result of a suite of evolutionary features, cooperation being only one among many.

234

iii) Thirdly, and in keeping with the above, the account given in the previous chapter made it clear that the non-normative inferences made by human infants in capitalising on the pointing gestures of their conspecifics to non-salient features of their shared environment is what distinguishes us from chimpanzees (and likely a huge swathe of other languageless creatures) only by a matter of degree. On the one hand, chimpanzees are similarly able to follow the direction of pointing gestures such that, if, in doing so, nothing of relevance comes into view, they still appear to wonder what the point was. And on the other hand, nothing ‘magical’ happens when humans give greater epistemic weight to the suggestions of their conspecifics; they are just more likely, for one reason or another, to investigate the indicated location further. Human thought and behaviour – as truly remarkable as it is – is naturalistically akin to the cognitions and behaviours of other animals, such as our closest evolutionary relatives, chimpanzees. Of course, when complemented by the suite of attendant capacities hinted at by Sterelny, the difference may be remarkable as to almost appear completely and utterly unique in the animal kingdom, but accounts such as this one hope to show that each and every step of the way can be accounted for without recourse to anything non-naturalistic. As entertainer, Tim Minchin, put it in his polemically naturalistic poem, Storm [Min14]: ‘… throughout history, every mystery ever solved has turned out to be not magic.’

2. Joining the dots

Obviously much, much more work needs to be done in order to provide a complete account of how the basic non-normative inferentialistic structure eventually gives rise to the emergence of the concept of objectivity. As was implied above, I doubt that this story can be told without recourse to ‘a [whole] suite of cognitive, behavioural, and social capacities’ [Ste14:76]. That said, one question in particular arises from the heavy focus placed herein on Tomasello’s empirical findings, namely: what further instances are there where similarly structured non-normative inferential patterns are observed and, so, can be analysed? For instance, can we find more experimental paradigms that elicit similarly structured inferential patterns, which seem to require implicit sensitivity to one’s own conflicting ommitments and to the possibility of error? Although the object-choice task is the clearest and most remarkable instance that I know of, Tomasello, for one, provides a huge range of other, related empirical observation of seemingly distinctive human performance capacities

235

(See [Tom08:62-82, 114-15 & 148]) for an extensive list of examples.) What is needed in light of the findings of this thesis, then, is a systematic analysis of the literature regarding such empirical experiments and observations, with a view to determining the extent to which our loosely Brandomian analysis may also further illuminate the phenomena observed. Unfortunately, it was beyond the scope of this thesis to undertake such an extensive review. Rather, it would be my most far-reaching hope that at least some of the findings of this thesis might one day help contribute to ethologists and philosophers of biology coming to better understand how it is that we came to think and speak and act as we do.

3. Davidson’s Goal

I’ve argued that none of Davidson’s original arguments appear able to achieve his aim of solving the problem of objectivity in their own right. Nonetheless, he is still to be given due credit for having articulated the problem naturalistically, and to have pointed – by way of his notion of intersubjective triangulation – towards the general social approach to the development of the concept of objectivity that I’ve sought to articulate in the final chapters of the thesis. Rather than rehash what the findings of this thesis are once again, let us close by considering whether the progress made herein on the problem of objectivity have any bearing on the traditional problem of knowledge and the challenge of global scepticism, as Davidson suggested such work might also accomplish.

Recall from the Introduction to this thesis that Richard Rorty characterises Davidson’s approach as effectively telling the sceptic to ‘get lost’. By this he means that Davidson refuses to meet the sceptic on their own terms. And as has been emphasised time and time again throughout this thesis, neither have we been concerned herein with whether the concept of objectivity we have been pursuing provides us with either epistemic or ontological veridicality. That is, our aim was not to establish the objectivity of human, language-based thought – let alone the mind-independent existence of the world. It was only ever to attempt to explain possession of the concept of objectivity, irrespective of whether or not that concept is applied legitimately or effectively. (Indeed, as was mentioned in the above sub-section regarding possible corollaries of the findings of this thesis, there are many areas of philosophical investigation where it is largely held that the concept of objectivity is widely misapplied.) But not engaging the sceptic on their own terms is only

236

half of Rorty’s point. The other half is that Davidson’s presumed naturalism – i.e. that the external world does exist and that science and common sense provide reliable access to it – might be invoked to provide a substantive account of how the sceptic came to be able to entertain the very question that they now challenge the naturalist with. That is to say, the sceptical challenge cannot be asked if one does not already possess the concept of objectivity. When Davidson seeks to answer the problem of objectivity, then, by presuming the very conditions to be subsequently thrown into question by the problem knowledge, he takes the wind out of the sails of the global sceptic. If the story to be told about how we come to possess the concept of objectivity depends on presuming the world exists such that we might then ask the question of the world’s existence, it makes the question look, if not foolish, then already substantially resolved in favour of the naturalist. Of course, as was said at the outset, this does not answer the sceptical worry; rather it puts it to rest. Such is the intended significance of Davidson’s appeal to triangulation. Such was the central effort of this thesis in taking the story that one step further. To finish with Davidson’s own characteristically wise and well-formed form [Dav91/01:219-20]:

If I am right, our propositional knowledge has its basis not in the impersonal but in the interpersonal. Thus, when we look at the natural world we share with others, we do not lose contact with ourselves, but rather acknowledge membership in a society of minds. If I did not know what others think, I would have no thoughts of my own and so would not know what I think. If I did not know what I think, I would lack the ability to gauge the thoughts of others. Gauging the thoughts of others requires that I live in the same world with them, sharing many reactions to its major features, including its values. So there is no danger that in viewing the world objectively we will lose touch with ourselves. The three sorts of knowledge form a tripod: if any leg were lost, no part would stand.

237

BIBLIOGRAPHY

* * *

1. Works Cited

[Ada10] Adams, Fred & Aizawa, Ken. (04.02.2010) ‘Causal Theories of Mental Content’ in Stanford Encyclopedia of Philosophy (accessed 26.12.2017 https://plato.stanford.edu/entries/content-causal/)

[Aes03] Aesop. The Complete Fables, trans. O. Temple (Penguin, 2003)

[Bar11] Barth, Christian. Objectivity and the Language-Dependence of Thought (Routledge, 2011)

[Ber1710/1988] Berkeley, George. Principles of Human Knowledge/Three Dialogues (Penguin, 1988)

[Bog99] Bogucki, P. The Origins of Human Society (Blackwell, 1999)

[Bra94/98] Brandom, Robert. Making it Explicit (Harvard University Press, 1994)

[Bra00] Brandom, Robert. Articulating Reasons (Harvard University Press, 2000)

238

[Bra09] Brandom, Robert. Reason in Philosophy: Animating Ideas (Belknap Press of Harvard University Press, 2009)

[Bri06] Bridges, Jason. (2006) ‘Davidson’s Transcendental Externalism’, Philosophy and Phenomenological Research Vol. LXXIII, No. 2, pp. 290-315

[Bri04] Brink, Ingar. (2004) ‘Joint Attention, Triangulation and Radical Interpretation: A Problem and its Solution’, Dialectica, Vol. 58, No 2, pp. 179-206

[Brä05] Bräuer, J., Call, J., and Tomasello, M. (2005). ‘All four great ape species follow gaze around barriers.’ Journal of Comparative Psychology, Vol.119, pp. 145–154.

[Bur10] Burge, Tyler. The Origins of Objectivity (Oxford University Press; Oxford, 2010)

[Byr91] Byrne, R. W., & Whiten, A. ‘Computation and mindreading in primate tactical deception’ in A. Whiten (Ed.), Natural theories of mind. (Oxford, Basil Blackwell, 1991) pp. 127–141.

[Byr92] Byrne, R. W., & Whiten, A. (1992) ‘Cognitive evolution in primates: Evidence from tactical deception.’ Man, Vol. 27, No. 3, pp. 609–627.

[Cam04] Camp, Elisabeth. (2004) ‘The Generality Constraint and Categorical Restrictions’, The Philosophical Quarterly, Vol. 54, No. 215, pp. 209-231

[Cam09] Camp, Elisabeth. (2009) ‘Putting Thoughts to Work: Concepts, Systematicity, and Stimulus-Independence’ in Philosophy and Phenomenological Research, Vol. LXXVIII, No. 2, March, pp. 275-311

239

[Car08] Eds. De Caro, Mario; Macarthur, David. Naturalism in Question (Harvard University Press, 2008)

[Car06] Carruthers, Peter. The Architecture of the Mind. (Oxford University Press, 2006)

[Car08] Carruthers, Peter. (2008) ‘Meta-cognition in Animals: A Skeptical Look’, Mind and Language, Vol. 23, pp. 58-89

[Cha07] Chadha, Monima. (2007) ‘No Speech, Never Mind!’, Philosophical Psychology, Vol. 20, No. 5, pp. 641–657

[Cor11] Corballis, Michael. The Recursive Mind: The Origins of Human Language, Thought, and Civilization (Princeton University Press, 2011)

[Cro17] Crockford, Catherine; Wittig, Roman; Zuberbühler, Klaus. (2017) ‘Vocalizing in chimpanzees is influenced by social-cognitive processes’, Science Advances, Vol. 3, no. 11

[Das07] Daston, L & Galison, P. Objectivity (Zone Books, 2007)

[Dav70/01] Davidson, Donald. ‘Mental Events’, Essays and Actions and Events (Oxford University Press, 2001)

[Dav74/01] Davidson, Donald. ‘Psychology as Philosophy’, Essays and Actions and Events (Oxford University Press, 2001)

[Dav75/84] Davidson, Donald. ‘Thought and Talk’, Inquiries into Truth and Interpretation (Oxford University Press, 2001)

240

[Dav82/01] Davidson, Donald. ‘Rational Animals’, Essays and Actions and Events (Oxford University Press, 2001)

[Dav87/01] Davidson, Donald. ‘Knowing One’s Own Mind’, Subjective, Intersubjective, Objective (Oxford University Press, 2001)

[Dav87/04] Davidson, Donald. ‘Problems in the Explanation of Action’, Problems of Rationality (Oxford University Press, 2004)

[Dav90/01] Davidson, Donald. ‘Epistemology Externalised’, Subjective, Intersubjective, Objective (Oxford University Press, 2001)

[Dav91/01] Davidson, Donald. ‘Three Varieties of Knowledge’, Subjective, Intersubjective, Objective (Oxford University Press, 2001)

[Dav92/01] Davidson, Donald. ‘The Second Person’, Subjective, Intersubjective, Objective (Oxford University Press, 2001)

[Dav94/05] Davidson, Donald. ‘The Social Aspect of Language’, Truth, Language and History (Oxford University Press, 2005)

[Dav95/04] Davidson, Donald. ‘The Problem of Objectivity’, Problems of Rationality (Oxford University Press, 2004)

[Dav97/01] Davidson, Donald. ‘The Emergence of Thought’, Subjective, Intersubjective, Objective (Oxford University Press, 2001)

[Dav97b/01] Davidson, Donald. ‘Indeterminism and Antirealism’, Subjective, Intersubjective, Objective (Oxford University Press, 2001)

241

[Dav97/05] Davidson, Donald. ‘Seeing Through Language’, Truth, Language and History (Oxford University Press, 2005)

[Dav99a] ‘Interpretation: Hard in Theory, Easy in Practice’, in Mario De Caro (ed.), Interpretations and Causes: New Perspectives on Donald Davidson’s Philosophy (Kluwer Academic Publishers, 1999).

[Dav99b] (ed.) HAHN, Lewis Edwin. The Philosophy of Donald Davidson. (Open Court: Illinois, 1999)

[Dav00] Davidson, Donald. ‘An Interview with Donald Davidson, May 2000, University of California at Berkeley’ (http://www-lehre.inf.uos.de/~milic/Coxi/InterviewDDavidson.pdf accessed 15.12.2015)

[Dav01] Davidson, Donald. ‘Externalisms’ in Interpreting Davidson, Eds. Kotatko, Pagin, Segal (Center for the Study of Language and Information, 2001)

[Dav03] Davidson, Donald. (2003) ‘Responses to Barry Stroud, John Mcdowell, and Tyler Burge’ Philosophy and Phenomenological Research. Vol. 67, No. 3, pp. 691-699

[Den10] Dennett, Daniel. ‘The Evolution of “Why?”’ in Reading Brandom: On Making it Explicit, Eds. B. Weiss & J. Wanderer (Routledge, 2010)

[Den13] Dennett, Daniel. Intuition Pumps and Other Tools For Thinking (Penguin, 2013)

[Der76/07] Derrida, Jacques, Who’s Afraid of Philosophy?: Right To Philosophy 1 (Stanford University Press, 2002)

242

[Des1641/1996] Descartes, René. Meditations on First Philosophy (trans. J. Cottingham) (Cambridge University Press, 1996)

[Des07] Dessalles, J-L. Why We Talk: The Evolution of Language (Oxford University Press, 2007)

[Dre81] Dretske, Fred. Knowledge and the Flow of Information (MIT Press, 1981)

[Dun96] Dunbar, R. Grooming, Gossip, and the Origin of Language (Faber, 1996)

[Ega97] Egan, Greg. Diaspora (Orion, 1997)

[Ega99] Egan, Greg. Teranesia (Victor Gollancz, 1999)

[Eva82] Evans, Gareth. The Varieties of Reference (Clarendon Press, 1982)

[Fer11] Ferestad, Ingvald & Ramberg, Bjørn. ‘Triangulation in Action: A Rationalizing Proposal’ in Triangulation: From an Epistemological Point of View, eds. M. C. Amoretti & G Preyer (Ontos Verlag, 2011)

[Fig06/10] Figal, Günter. Objectivity: The Hermeneutical and Philosophy, trans. T. George (State University of New York Press, 2010)

[Gar11] Garavaso, Pieranna. ‘Review of Naomi Scheman’s Shifting Ground: Knowledge and Reality, Transgression and Trustworthiness’, Notre Dame Philosophical Reviews (https://ndpr.nd.edu/news/24786-shifting-ground-knowledge-and-reality-transgression-and-trustworthiness/ accessed 15.12.2015)

243

[Glo03] Glock, Hans-Johann. Quine and Davidson on Language, Thought and Reality (Cambridge University Press, 2003)

[Glü06a] Glüer, Karthrin. ‘Triangulation’, The Oxford Handbook of Philosophy of Language, Eds. E. Lepore & B. Smith (Oxford University Press, 2006) pp.1006-1019

[Gop98] Gopnik, A., and Astington, J. W. ‘Children’s Understanding of Representational Change and Its Relation to the Understanding of False Belief and the Appearance–Reality Distinction.’ Child Development, Vol. 59: 26–37. (1988)

[Góm90] Gómez, J. C. ‘The emergence of intentional communication as aproblem-solving strategy in the gorilla. In “Language” and Intelligence in Monkeys and Apes, ed. S. T. Parker & K. R. Gibson (Cambridge University Press, 1990), pp. 333–355

[Hau96] Hauser, M. The Evolution of Communication. (MIT Press, 1996)

[Hey98] Heyes, C. M. (1998) ‘Theory of mind in non-human primates.’, Behavioral and Brain Sciences, 21, pp. 101–48.

[Hrd09] Hrdy, S. B. Mothers and Others: The Evolutionary Origins of Mutual Understanding (Harvard University Press, 2009)

[Hur03] Hurley, S. (2003) ‘Animal action in the space of reasons.’ Mind & Language, 18, pp. 231–256.

[Kuk04] Kukla, Rebecca. ‘Review of How Scientific Practices Matter: Reclaiming Philosophical Naturalism by Joseph Rouse’ in The Philosophy of Science, Vol. 71, No. 2, pp. 216-219 (The University of Chicago Press, April 2004)

244

[Lan10] Lance, Mark & Kukla, Rebecca. ‘Perception, Language, and the First Person’ in Reading Brandom: On Making it Explicit, Eds. B. Weiss & J. Wanderer (Routledge, 2010)

[Las04] LASONEN, Maria and MARVAN, Tomass. (2004) ‘Davidson’s Triangulation: Content-Endowing Causes and Circularity’, International Journal of Philosophical Studies, Vol. 12, No. 2, pp. 177-195

[Lep05] Lepore, Ernie and Ludwig, Kirk. Davidson: Meaning, Truth, Language and Reality. (Clarendon Press, 2005)

[Lep07] Lepore, Ernie & Ludwig, Kirk. (2007) ‘Radical Misinterpretation: A Reply to Stoutland’, International Journal of Philosophical Studies (Vol. 15, Issue 4, pp. 557-585)

[Lie98] Lieberman, P. Eve Spoke: Human Language and Human Evolution (W. W. Norton, 1998)

[Lis06] Liszkowski, U., M. Carpenter, T. Striano, and M. Tomasello. (2006).’12- and 18-month-olds point to provide information for others. Journal of Cognition and Development, 3 173-187.

[Lis08] Liszkowski, U., M. Carpenter, and M. Tomasello. (2008). ‘Twelve-month-olds communicate helpfully and,appropriately for knowledgeable and ignorant partners.’ Cognition, 108, 732-739.

[Lur15] Lurz, Robert. (2015) ‘Animal Minds’ in Internet Encyclopedia of Philosophy (http://www.iep.utm.edu/ani-mind/#SSH1ciii accessed 26.12.2017)

[Mal02] Malpas, Jeff. From Kant to Davidson: Philosophy and the Idea of the Transcendental. (Routledge; New York, 2002)

245

[Mer16] Mercer, Christia. (2016) ‘Descartes’ debt to Teresa of Avila, or why we should work on women in the history of philosophy’, Philosophical Studies Vol.174, No. 10 pp.2539-2555

[Mid11] Middleton, Tim. ‘History: The Great Trigonometrical Survey’ (ßluesci: The Cambridge University Science Magazine, 29.01.2011) (https://bluesci.wordpress.com/2011/01/29/history-the-great-trigonometrical-survey/accessed 05.11.2017)

[Mil12] Millikan, Ruth. (2012) “What’s Inside a Thinking Animal”, Do animals live in the space of reason?; Action and decision in non-human animals (XXII Deutscher Kongress fár Philosophy, Proceedings Welt der Gründe, Kolloquium 19), pp. 889-893

[Min14] Minchin, Tim. Storm (Orion Publishing Group, 2014)

[Mon03] Montminy, Martin. (2003) ‘Triangulation and the Ambiguity Problem’, Crítica, Vol. 35, No. 105, pp. 25-48

[Nag79] Nagel, Thomas. Mortal Questions (Cambridge University Press, 1979)

[Nie1879/1996] Nietzsche, Friedrich. Human, All too Human (Penguin, 1879,1996)

[Noz03] Nozick, Robert. Invariances: The Structure of the Objectivity World (Harvard University Press, 2001)

[Pag01] Pagin, Peter. (2001) ‘Semantic Triangulation’ in Interpreting Davidson (CSLI Publications: Stanford, 2000)

246

[Per87] Perner, J., Leekam, S. R., and Wimmer, H. (1987) ‘Three-Year-Olds’ Difficulty with False Belief: The Case for Conceptual Deficit.’ British Journal of Developmental Psychology, Vol. 5: pp.125–37.

[Pla94] Plato. Gorgias. Trans. Robin Waterfield (Oxford, 1994)

[Pla14] Plato. The Republic trans. B. Jowett(https://ebooks.adelaide.edu.au/p/plato/p71r/index.html accessed 16.12.2015)

[Pop72] Popper, Karl. Objective Knowledge: An Evolutionary Approach (Oxford University Press, 1972)

[Pri08/11] Price, Huw. ‘One Cheer for Representationalism’ in Naturalism without Mirrors (Oxford, 2011)

[Pri08/13] Price, Huw. ‘Naturalism without Representationalism’ in Expressivism, Pragmatism and Representationalism (Cambridge University Press, 2013)

[Qui81] Quine, W. V. O. Theories and Things (Harvard University Press, 1981)

[Ror80] Rorty, Richard. Philosophy and the Mirror of Nature (Princeton University Press, 1980)

[Rou02] Rouse, Joseph. How Scientific Practices Matter: Reclaiming Philosophical Naturalism (University of Chicago Press, 2002)

[Rou15] Rouse, Joseph. Articulating the World: Conceptual Understanding and the Scientific Image (University of Chicago Press, 2015)

247

[Sad09] Sadedin, S. & Paperin, G. (2009) ‘Implications of the social brain hypothesis for evolving human-like cognition in digital organisms’ – conference paper for 10th European Conference on Artificial Life.

[Sel62/91] [SPR] Sellars, Wilfrid. Science, Perception and Reality, (Routledge & Kegan Paul Ltd; London, and The Humanities Press: New York, 1963; reissued in 1991 by Ridgeview Publishing Co., Atascadero, CA.)

[Ser95] Searle, John. The Construction of Social Reality (Free Press; New York, 1995)

[Sky10] Skyrms, B. Signals: Evolution, Learning, and Information (Oxford University Press, 2010)

[Spe96] Sperber, D. Explaining Culture: A Naturalistic Approach (Blackwell, 1996)

[Sta08] Stalnaker, Robert. Our Knowledge of the Internal World (Oxford University Press, 2008)

[Ste03] Sterelny, K. Thought in a hostile world: The evolution of human cognition. (Blackwell, 2003)

[Sto73] Stoppard, Tom. Rosencrantz and Guildenstern Are Dead (Faber & Faber, 1972)

[Sto06] Stoutland, Fredrick. (2006) ‘A Mistaken View of Davidson’s Legacy’, International Journal of Philosophical Studies (Vol. 14, Issue 14, pp. 579-596)

[Sto07] Stoutland, Fredrick. (2007) ‘Radical Misinterpretation Indeed: Response to Lepore & Ludwig’, International Journal of Philosophical Studies (Vol. 15, Issue 4, pp. 587-597)

248

[Swi07] Swindal, James. (2007) ‘Can a discursive pragmatism guarantee objectivity: Habermas and Brando on the correctness of norms’ in Philosophy & Social Criticism, (SAGE Publications), Vol. 33, No. 1, pp. 113-126

[Tol68/91] Tolkien, J. R. R., Lord of the Rings (Harper Collins, 1991)

[Tom99] Tomasello, M., Hare, B., and Agnetta, B. (1999). ‘Chimpanzees, Pantroglodytes, follow gaze direction geometrically.’ Animal Behaviour, Vol. 58, No. 4, pp. 769–777.

[Tom08] Tomasello, Michael. Origins of Human Communication (The MIT Press, 2008)

[Tom09] Tomasello, Michael. Why We Cooperate (The MIT Press, 2009)

[Tom14] Tomasello, Michael. A Natural History of Human Thinking (Harvard University Press, 2014)

[Ver97] Verheggen, Claudine. ‘Davidson’s Second Person’, Philosophical Quarterly. Vol. 47, pp, 361-69 (1997)

[Wie72] Wiesel, Elie. Souls on Fire: Portraits and Legends of Hasidic Masters (Simon & Schuster, 1972)

[Witt53/99] Wittgenstein, Ludwig. Philosophical Investigations, trans. G.E.M. Anscombe (Blackwell, 1999)

[Wri92] Wright, Crispin. Truth and Objectivity (Harvard University Press, 1992)

2. Further Reading

249

[All99] Allen, Colin. (1999) ‘Animal Concepts Revisited: The Use of Self-Monitoring as an Empirical Approach’, Erkenntnis, Vol. 55, pp. 33-40

[Arc04] Eds. Archer, M. & Outhwaite, W. Defending Objectivity: Essays in Honour of Andrew Collier (Routledge; London, 2004)

[Aud94] Audi, Robert. (1994) ‘Dispositional Beliefs and Dispositions to Believe’, Nous, Vol. 28, No. 4, pp. 419-434

[Bar13] Bar-On, Dorit & Sias, James. (2013) ‘Varieties of Expressivism’ in Philosophy Compass, Vol. 8, No. 8, pp. 699-713

[Bei02] Beisecker, David. (2002) ‘Some More Thoughts About Thought and Talk: Davidson and Fellows on Animal Belief’, Philosophy, Vol. 77, No. 299, pp. 115-124

[Beh05] Behne, T., Carpenter, M., and Tomasello, M. (2005). ‘One-year-olds comprehend the communicative intentions behind gestures in a hiding game.’ Developmental Science, Vol. 8, 492–499.

[Ben64] Bennett, J. Rationality (Routledge & Kegan Paul; New York, 1964)

[Ber83] Berka, K. Measurement (D. Reidel Publishing; Dordrecht, 1983)

[Ber03] Bermudez, José Luis (2003). ‘Ascribing Thoughts to Non-Linguistic Creatures’, Facta Philosophica, Vol. 5, pp. 313-335

[Ber03a] Bermudez, José Luis. Thinking Without Words (Oxford University Press, 2003)

250

[Bor09] Borgoni, C & Souza, H. ‘Davidson’s Externalisms’, Universitas Philosophica, Vol. 53, No. 26, pp. 65-87

[Bra00] Ed. Brandom, Robert. Rorty and His Critics (Blackwell; Massachusetts, 2000)

[Bri13] Brink, I. & Liljenfors, R. ‘The Developmental Origin of Metacognition’, Infant and Child Development, Vol. 22, No. 1, pp. 85-101

[Bri07] Briscoe, Robert Eamon. (2007) ‘Communication and Rational Responsiveness to the World’, Pacific Philosophical Quarterly, Vol. 88, No. 2, pp. 135-159

[Bro84] Ed. Brown, S. C. Objectivity and Cultural Divergence (Cambridge University Press; Cambridge, 1984)

[Bur86] Burge, Tyler. (1986) ‘Cartesian Error and the Objectivity of Perception’ in Subject, Thought, and Context Eds. P. Pettit & J. McDowell (Claredon Press, 1986)

[Bur92] Burge. Tyler. (1992) ‘Philosophy of Language and Mind 1950-1990’ in The Philosophical Review, Vol. 101, No. 1, pp. 3-51

[Bur03] Burge, Tyler. (2003) ‘Social Anti-Individualism, Objective Reference’ in Philosophy and Phenomenological Research, Vol. 67, No. 3, pp. 682-690

[Cal09] Calvert-Minor, Chris. ‘Commonsense Realism and Triangulation’, Philosophia, Vol 37, pp. 67-86

251

[Cam09] Camp, Elisabeth. ‘A Language of Baboon Thought?’ in The Philosophy of Animal Minds, ed. R. Lurz (Cambridge University Press, 2009) pp. 108-126

[Cap07] Capone, Alessandro. (2007) Book Review: ‘Donald Davidson: Meaning, Truth, Language, and Reality’, Journal of Pragmatics, Vol. 39, pp. 1039-1046

[Car96] Carruthers, Peter. Language, thought and consciousness. (Cambridge University Press, 1996)

[Car07] Carruthers, Peter. (2007) ‘Invertebrate concepts confront the Generality Constraint (and win)’, Journal of Ethics, Vol. 11, 2007, pp. 275-297

[Car08] Carruthers, Peter. (2008) Meta-cognition in Animals: A Skeptical Look’, Mind and Language, Vol. 23, pp. 58-89

[Chi01] Child, William, (2001) ‘Triangulation: Davidson, Realism and Natural Kinds’, Dialectica, Vol. 55, No. 1, pp. 29-49

[Chi06] Child, William. (2006) ‘Interpreting People and Interpreting Texts’, International Journal of Philosophical Studies, Vol. 14(3), pp. 423-441

[Col03] Collier, Andrew. In Defense of Objectivity and other Essays (Routledge; London, 2003)

[Das92] Daston, Lorraine. (1992) ‘Objectivity and the Escape from Perspective’, Social Studies of Science, Sciences (D. Reildel; Dordrecht, 1987)

[Das07] Daston, L & Galison, P. Objectivity (Zone Books; New York, 2007)

252

[Dav89] Davidson, Donald. (1989) ‘The Conditions of Thought’ in Le Cahier, Vol. 7, pp. 167-171

[Dav90] Davidson, Donald. (1990) ‘The Structure and Content of Truth’ in The Journal of Philosophy, Vol. 6, pp. 279-328

[Dav94] Davidson, Donald. (1994) ‘Radical Interpretation Interpreted’ in Philosophical Perspectives. Vol. 8, pp. 121-128

[Dav08] Davidson, Donald. Truth & Predication (Harvard University Press, 2008)

[Den95] Dennet, Daniel. ‘Do Animals Have Beliefs?’ in Comparative Approaches to Cognitive Sciences, ed. Roitblat, H. (MIT Press, 1995)

[Dre99] Dretske, Fred. I. (1999) ‘Machines, Plants and Animals: The Origins of Agency’, Erkenntnis, Vol. 55, pp. 19-31

[Eng01] Engel, Pascal. (2001) ‘The Norms of Thought: Are They Social?’, Mind and Society, Vol. 2, pp. 129-148

[Evn07] Evnine, Simon. (2007) ‘Book Review: Donald Davidson: Meaning, Truth, Language, and Reality’, Mind, Vol. 116, No. 462, pp. 446-453

[Fel00] Fellows, Roger. (2000) ‘Animal Belief’, Philosophy, Vol. 75, No. 294, pp, 587-598

[Fod80] Fodor, Jerry. The Language of Thought. (Harvard University Press, 1980)

[Fod03] Fodor, Jerry. (2003) ‘Review of Bermudez’ Thinking Without Words’, The Guardian

253

[Gil87] Gillet, G. R. (1987) ‘The Generality Constraint and Conscious Thought’, Analysis, Vol. 47, No. 1, Jan., pp 20-4

[Gil60] Gilliespie, Charles Coulston. The Edge of Objectivity (Princeton University Press; New Jersey, 1960)

[Glo00] Glock, Hans-Johann. (2000) ‘Animals, Thoughts and Concepts’, Sythese, 123, pp. 35-64

[Glo03] Glock, Hans-Johann (2003) ‘Critical Notices: Donald Davidson, Subjective, Intersubjective, Objective’, Philosophical Investigations, Vol. 26, No. 4, pp. 348-360

[Glü06b] Glüer, Katherin, (2006b) ‘The Status of Charity I: Conceptual Truth or A Posteriori Necessity?’, International Journal of Philosophical Studies, Vol. 14(3), pp. 337-359

[Gol08] Goldberg, Nathaniel. (2008) ‘Tension within Triangulation’, The Southern Journal of Philosophy. Vol. 96, pp. 368-383

[Gol09] Goldberg, Nathaniel. (2009) ‘Triangulation, Untranslatability, and Reconciliation, Philosophia, Vol. 37, pp. 261-280

[Hau02] Hauser, M. & Chomsky, N. & Fitch, W. (2002) ‘The Faculty of Language: What Is It, Who Has It, and How Did It Evolve?’ Science, Vol. 298, pp. 1569-1579

[Har04] Hare, B., and Tomasello, M. (2004). ‘Chimpanzees are more skillful in competitive than in co-operative cognitive tasks.’ Animal Behaviour, Vol. 68, 571–581.

254

[Hur01] Hurley, Susan. (2001) ‘Overintellectualizing the Mind’, Philosophy and Phenomenological Research, Vol. LXIII, No. 2, pp. 423-431

[Hur06] Hurley, Susan. (2006) ‘Making Sense of Animals’ in Rational Animals Eds. S. Hurley & M. Nudds (Oxford University Press; Oxford)

[Joh07] Johnston, Mark. ‘Objective Mind and the Objectivity of Our Minds’, Philosophy and Phenomenological Research. Vol. 75, No. 2, pp. 233-268

[Jos04] Joseph, Marc. Donald Davidson (Acumen: Chesham, 2004)

[Kor16] Koren, Ladislav. ‘Joint Intentionality: From Thick to Thin’, Journal of Social Ontology, Vol. 2, No. 1, pp. 75-85

[Kuk06] Kukla, Rebecca. (2006) ‘Objectivity and Perspective in Empirical Knowledge’, Episteme, pp. 80-95

[Kuk08] Kukla, Rebecca. (2008) ‘Naturalizing Objectivity’, Perspective on Science, Vol. 16, No. 3, pp. 285-302

[Kyb84] Kyburg, Henry J. Jr. Theory and Measurement (Cambridge University Press; Cambridge, 1984)

[Lan07] Lance, M. & White, W. (2007) ‘Stereoscopic Vision: Persons, Freedom, and Two Spaces of Material Inference’ in Philosophers’ Imprint, Vol. 7, No. 4, pp. 1-21

[Las04] Lasonen, Maria and Marvan, Tomass. (2004) ‘Davidson’s Triangulation: Content-Endowing Causes and Circularity’, International Journal of Philosophical Studies, Vol. 12, No. 2, pp. 177-195

255

[Lep86] Ed. Lepore, Ernest. Truth and Interpretation: Perspectives in The Philosophy of Donald Davidson (Basil Blackwell, 1986)

[Lep04] Lepore, Ernest & Ludwig, Kirk. (2004) ‘Donald Davidson’ in Midwest Studies in Philosophy, Vol. 28, pp. 309-333

[Lep05] Lepore, Ernest and Ludwig, Kirk. Davidson: Meaning, Truth, Language and Reality. (Clarendon Press: Oxford, 2005)

[Lep06] Lepore, Ernest and Ludwig, Kirk. (2006) ‘Ontology in the Theory of Meaning’, International Journal of Philosophical Studies, Vol. 14, pp. 325-335

[Lep07] Lepore, Ernest & Ludwig, Kirk. (2007) ‘Radical Misinterpretation: A Reply to Stoutland’, International Journal of Philosophical Studies, Vol. 15, Issue 4, pp. 557-585

[Lep07] Lepore, Ernest & Ludwig, Kirk. Donald Davidson’s Truth-Theoretic Semantics (Oxford University Press, 2007)

[Lep11] Lepore, Ernest & Ludwig, Kirk. (2011) ‘Truth and Meaning Redux’ in Philosophical Studies, Vol. 154, pp. 251-277

[Lep13] Lepore, Ernest & Ludwig, Kirk. A Companion to Donald Davidson (Wiley-Blackwell, 2013)

[Lud03] (ed.) Ludwig, Kirk. Donald Davidson. (Cambridge University Press: Cambridge, 2003)

256

[Lil07] Lillehammer, Hallvard. (2007) ‘Davidson on Value and Objectivity’, Dialectica, Vol. 61, No. 2, pp. 203-217

[McD03] McDOWELL, John. (2003) ‘Subjective, Intersubjective, Objective’, in Philosophy and Phenomenological Research. Vol. 67, No. 3, pp. 675-81

[Mac04] Machan, T. Objectivity: Recovering Determinate Reality in Philosophy, Science, and Everyday Life (Ashgate; Aldershot, 2004)

[Mal11] Malpas, Jeff. (ed.), Dialogues with Davidson: On the Contemporary Significance of his Thought (Cambridge, Mass.: MIT Press, 2011).

[Mal92] Malpas, Jeff. Davidson’s Holism: Epistemology and the Mirror of Meaning, (Cambridge: Cambridge University Press, 1992)

[Mat07] Matthews, Robert. The Measure of Mind (Oxford University Press, 2007)

[Müh88] Mühlhölzer, Felix. (1988) ‘On Objectivity’, Erkenntnis, pp. 185-230

[Mye16] Myers, Robert H. & Verheggen, Claudine. Donald Davidson’s Triangulation Argument: A Philosophical Inquiry (Routledge, 2016)

[Nag86] Nagel, Thomas. The View From Nowhere (Oxford University Press; Oxford, 1986)

[Nat95] Eds. Natter, W. & Schatzki, T. R. & Jones, J. P. Objectivity and Its Other (Guilford Press; New York, 1995)

257

[New86] Newell, R. W. Objectivity, Empiricism and Truth (Routledge & Kegan Paul; London, 1986)

[Nul06] Nulty, Timothy. (2006) ‘Davidsonian Triangulation and Heideggerian Comportment’, International Journal of Philosophical Studies, Vol. 14, No. 3, pp. 443-453

[Pag06] Pagin, Peter. (2006) ‘The Status of Charity I: Conceptual Truth or A Posteriori Necessity?, International Journal of Philosophical Studies, Vol. 14(3), pp. 361-383

[Pér05] Pérez, Diana. (2005) ‘Is Thought Without Language Possible?’, Principia, Vol. 9, No. 1-2, pp. 177-191

[Pet11] Petrenko, Anton. (2011) ‘Christian Barth: Objectivity and the Language-Dependence of Thought’ in Philosophy in Review, No. 6, pp. 396-399

[Pin94] Pinker, Steven. The Language Instinct. (Penguin: London, 1994)

[Pro06] Proust, Joëlle. (2006) ‘Rationality and Metacognition in Non-human Animals’ in Rational Animals? (Oxford University Press; Oxford) pp. 247-274

[Pro99] Proust, Joëlle. (1999) ‘Mind, Space and Objectivity in Non-Human Animals’, Erkenntnis, Vol. 55, pp. 41-58

[Put92a] Putnam, Hilary. ‘The Craving for Objectivity’ in James Conant (ed.) Realism with a Human Face (Harvard University Press, 1992)

[Put92b] Putnam, Hilary. ‘Objectivity and the Science/Ethics Distinction’ in James Conant (ed.) Realism with a Human Face (Harvard University Press, 1992)

258

[Put81] Putnam, Hilary. Reason, Truth and History (Cambridge University Press; Cambridge, 1981)

[Ram15] Ramberg, Bjørn. ‘Davidson and Rorty: Triangulation and Anti-Foundationalism’ in Malpas & Helmuth (Eds), The Routledge Companion to Hermeneutics (Routledge, 2015)

[Res97] Rescher, Nicholas. Objectivity: The Obligations of Impersonal Reason (University of Notre Dame Press; Notre Dame, 1997)

[Ror91] Rorty, Richard. Objectivity, Relativism, and Truth (Cambridge University Press; Cambridge, 1991)

[Sac00] Sacks, Mark. Objectivity and Insight (Clarendon Press; Oxford, 2000)

[Sch67] Scheffler, Israel. Science and Subjectivity (Bobbs-Merrill Company; Indianapolis, 1967)

[She06] Shettleworth, S & Sutton, J. ‘Do Animals Know What They Know?’ in Rational Animals Eds. S. Hurley & M. Nudds (Oxford University Press, 2006 ) pp. 235-246

[Sin02] Sinclair, Robert. (2002) ‘What is Radical Interpretation? Davidson, Fodor, and the Naturalization of Philosophy’, Inquiry, Vol. 45, pp. 161-184

[Sin05] Sinclair, Robert. (2005) ‘The Philosophical Significance of Triangulation: Locating Davidson’s Non-Reductive Naturalism’, Metaphilosophy, Vol. 36, No. 5, pp. 708-728

259

[Sma59] Smart, J. J. C. ‘Measurement’, The Australasian Journal of Philosophy, Vol. 37, No. 1, 1959, pp. 1-13

[Smi06] Smith, Barry C. 2006) ‘Davidson, Interpretation and First-Person Constraints on Meaning’, International Journal of Philosophical Studies, Vol. 14(3), pp. 385-406

[Ste99] Stephan, Achim. (1999) ‘Are Animals Capable of Concepts?’, Erkenntnis, Vol. 55, pp. 79-92

[Sto93] (ed.) Stoeker, Ralf. Reflecting Davidson (Walter de Gruyter: New York, 1993)

[Str03] Stroud, Barry. (2003) ‘Ostentation and the Social Character of Thought’, Philosophy and Phenomenological Research, Vol. 67, No. 3, pp. 667-674

[Str59] Strawson, P. F. Individuals (University Paperbacks; London, 1959)

[Tal97] Talmage, Catherine, J. L. (1997) ‘Meaning and Triangulation’, Linguistics and Philosophy, Vol. 20, pp. 139-145

[Tye97] Tye, Michael. (1997) ‘The Problem of Simple Minds: Is there anything it is to be like a honey bee?’, Philosophical Studies, Vol. 88, pp. 289-317

[Tye06] Tye, Michael. ‘Another look at Representation about pain’ in Pain: New Essays on its Nature and the Methodology of its Study. Ed. M. Aydede (MIT Press; Cambridge, 2006)

[Ver06] Verheggen, Claudine. (2006) ‘How Social Must Language Be?’, Journal for the Theory of Social Behaviour, Vol 36, No. 2, pp. 203-219

260

[Ver07] Verheggen, Claudine. (2007) ‘Triangulating with Davidson’, Philosophical Quarterly.

[War07] Warneken, F. & Tomasello, M. ‘Helping and Cooperation at 14 Months of Age’, Infancy, Vol. 11, No. 3, pp. 271-294

[Wil00] Williams, Meredith. (2000) ‘Wittgenstein and Davidson on the Sociality of Language’, The Journal for the Theory of Social Behaviour

[Wil11] Williams, Meredith. (2011) ‘Master and Novice in Wittgenstein’s Later Work’, American Philosophical Quarterly, Vol. 48

[Yal99] Yalowitz, Steven. ‘Davidson’s Social Externalism’, Philosophia, Vol. 27, pp. 99-136

[Zeg99] (ed.) Zeglen, Urszula M. Donald Davidson: Truth, meaning and knowledge. (Routledge: London, 1999)

Minerva Access is the Institutional Repository of The University of Melbourne

Author/s:

Barham, Ross Campbell

Title:

Davidson's Objective: language & the concept of objectivity

Date:

2018

Persistent Link:

http://hdl.handle.net/11343/213938

File Description:

Davidson's objective: language & the concept of objectivity

Terms and Conditions:

Terms and Conditions: Copyright in works deposited in Minerva Access is retained by the

copyright owner. The work may not be altered without permission from the copyright owner.

Readers may only download, print and save electronic copies of whole works for their own

personal non-commercial use. Any use that exceeds these limits requires permission from

the copyright owner. Attribution is essential when quoting or paraphrasing from these works.