the anti-naturalism of some language centered accounts of belief

The Anti-Naturalism of Some Language Centered Accounts of Belief

Ruth BARCAN MARCUS’

Summary Common sense explanations of human action are often framed in terms of an agent’s beliefs

and desires. Recent widely received views also take believing and desiring (as well wishing, intending and even thinking) as attitudes of an agent to linguistic or quasi-linguistic entities. It is here claimed that such a narrow view of cognitive attitudes is not supportable, since even among linguals non-verbal responses are often overriding evidence for belief and desire, even where they run counter to sincere verbal assents. The view is also curiously non naturalistic in that it disallows ascribing beliefs and desires altogether to non-linguals and pre-linguals. In the present paper a “common sense” explanation of action in accordance with the triad <Desire, Belief, Action, is seen as a useful phenomenological “theory” provided that language centrality is not taken as essential.

Language centered accounts of belief and believing have been a focus of philosophy of mind and philosophy of action in recent decades. Some trace its origins to Plato’s silent forum. It is such accounts of belief which have been viewed as central to what has been called, often pejoratively, “folk psychology,” or more neutrally “common-sense” psychology, and lately come under attack.

F.P. Ramsey ’ at one point was a proponent of the language centered view but with some ambivalence. He allows a sense of “belief” in which, for example, a belief may be attributed to a chicken who has an aversion to eating a species of caterpillar due to unpleasant prior experiences. But he suggests that in attributing to the chicken a belief that the caterpillar is inedible is an “ambiguous” use, and he goes on to say that “without wishing to depreciate the importance of this kind of belief, . . . I prefer to deal with those beliefs which are expressed in words. . . consciously asserted, or denied. . . The mental factors [my emphasis] of such a belief I take to be words spoken aloud or to

Halleck Professor of Philosophy, Yale. National Humanities Center. The Foundations of Mathematics (Routledge, London, 1931): 144.

Dialectics Vol. 49, No 2-4 (1995)

114 Ruth Barcan Marcus

oneself or merely imagined, connected together and accompanied by a feeling or feelings of belief.” On Ramsey’s language centered view an agent (pres- umably a competent language user) believes that ‘P’ where the agent is conscious of those “mental factors” designated as utterances “spoken aloud or to oneself,” and toward which the agent has an affirmative feeling. “Accompa- nied by a feeling,” allows prying about, at least conceptually, the feeling and the understood sentence. [In his later account of degrees of belief, he defines such degrees in terms of betting behavior.]

Ramsey pries apart, as is common to language centered views, the attitude from the object of the attitude - the understood sentence. He may be seen as endorsing what has come to be called a disquotation principle. On the as- sumption that an agent is a sincere competent language user there is a condi- tional and biconditional version.

Disquot.: If A assents to ‘P’ then A believes that P.

The converse

Disquot,: If A believes that P then A assents to ‘P’

requires amplification. Assenting is an occurrent action, and is evoked in special circumstances as, for example, when an agent is asked whether he or she believes that P. If believing is always a relation of an agent to a linguistic entity, then the converse of Disquot, would have to suppose that agent wouldsincer- ely assent to ‘P’ if he or she believes that P. The conjunction of Disquot and Disquot is the biconditional version.

What should be noticed on the Ramseystyle account is that (1) it clearly excludes the attribution of belief to non-linguals and (2) it disallows attribution of unconscious believing where unconscious supposes that a sentential correlate, a correlative assent cannot be evoked. It excludes the possibility that the agent believes but cannot generate the appropriate sentence - doesn’t have the words. “If you can’t say what you believe you can’t whistle it either.” Unconscious beliefs whether or not of psychoanalytic interest are especially problematic since they are attributions which are initially and often persistently disavowed by the agent. Topics to which I will return.

Frege3 pries apart the sentence from its content and makes the latter the object of believing yet he must be counted as having a language focussed ac-

See S. Kripke. “A Puzzle About Belief,” Meaning and Use, ed. A. Margalit (Dordrecht,

“On sense and Nomenatum” and other essays in G. Frege. Translations from The Phil- 1979): 234-83.

osophical Writings, trans. P. Geach and M. Black (Oxford, Blackwell, 1952).

The Anti-Naturalism of Some Language Centered Accounts of Belief 115

count. Believing for Frege relates an attitude to a content under a mode of presentation. That content is a Fregean proposition - which he postulates as an abstract non-psychological entity. It is a curious fact that he calls them “thoughts” of which he believed there was a common stock. We “have them in mind when we entertain them or believe them. I will not discuss here the many difficulties with this notion of proposition. Suffice it to say they are quasi-linguistic intensional entities that mimic those properties which on a Tarskian semantics are more properly attributes of interpreted sentences such as true, false, logically valid, contradictory, consistent, etc. Frege’s contents are parasitic on sentences in their formal features. They are not actual or non- actual states of nor are they structures with non-intensional real world constituents such as Russell’s propositions. Recall Frege’s astonishment at Russell’s5 bland agreement that on his (Russell’s) use Mt. Blanc itself was a constituent in the proposition that Mt. Blanc is more than 4000 feet high. Nor are they sets of worlds or functions from worlds to truth values. They are abstract intensional entities which mimic sentences, share formal and semantical features of sentences and can arguably be understood as quasi-linguistic entities which are the objects of believing.

A minimalist version of a language centered view is (or was) that of J. Fodor.7 On the model of some computational theories of artifical intel- ligence, propositions are sentences in the language of thought, independent formulae stored in the brain. Much simplified the subject has an internal regis- ter of sentence constituents and basic sentences which are correlated with natural language sentences. Syntax and rules generate complex sentences. Mental analogies of ‘yes’ or ‘no’ responses are elicited on appropriate cues. An agent believes ‘P’ when the formula which is a correlate of ‘P’ in the language of thought elicits an assenting response.

Fodor is narrowly focussed on questions about cognitive attributions to language using agents. But a more global language-centered view of mind is that of Davidson, since he denies that any cognitive attributions to non-language users are warranted. Non-language users do not have thoughts, beliefs; not even desires or intentions. For Davidson, beliefs of agents are the objects

4 See for example, A. Plantinga, The Nature of Necessity (New York: Oxford Press,

G. Frege, Philosophicaland Mathematical Correspondence, ed. G. Gabriel, et al. (Chi- 1974) and R. Chisholm, “Events and Propositions,” Now, IV, 1 (1970).

Modularity of Mind (Cambridge, Mass.: MIT Press, 1983).

ford Press 1975). “Thought and Talk,” in Mind and Language, ed. Samuel Gattenplan (New York: Ox-


of attitudes of members of a linguistic community. Those objects are sentences in an interpreted language. The attitude is that of holding the sentence true. ‘True’ as a predicate of beliefs is parasitic on ‘true’ as attached to sentences and unlike truth as a sentential property it is redundant and dispensable. To believe that P is true is simply to believe that P. Davidson says, “Such an account of true belief depends on the notion of true sentence,” and the semantical account of truth requires a sharedxlanguage and a large measure of competence. The “sincere assent of competent language users” of the disquotation principle is a manifestation of the “holding true” of Davidson. Since for Davidson it is only language users who have beliefs, he makes adjustments, especially for the converse. Since assenting is episodic, an agent must be dis- posed to assent. Where some language users lack vocabulary or are claimed to have tacit beliefs whether or not they are of the psychoanalytic kind, such species of belief, if there are such, can be accommodated by Davidsonian views, but not without some highly speculative stretching. Given the agent is in a language community of interpreters, if it is unavailability of vocabulary, the vocabulary can be ferreted out. Some tacit beliefs, given the known absence of closure for believers over logical consequence, might be those which a deductively rational agent would assent to up to a point, which does not require logical omniscience but is a minimal criterion of logicality. Psychoana- lytic or motivated unconscious beliefs, require an account of unconscious assent to sentences which would not readily be consciously assented to. Some theory about divided mentality is required, one submerged, but with its own complement of sentences unconsciously assented to, a topic to which I will return. But Davidson’s language centered views are not confined to an account of cognitive verbs of attributions. It is global in that it is extended to a large range of mental phenomena.

In Thought and Talk he9 says, “the attribution [emphasis mine] of desires and beliefs (and other thoughts) must go hand in hand with the interpretation of speech.” Such a claim is ambiguous. In one sense it is a conceptual truth. At- tributions are made by language users. If I attribute to a dog Fido a desire for food or the belief that what is in his dish is edible, I am using language to make the attribution. Fido makes no attributions. But Davidson goes on to say that “the attribution of thought depends on interpretation of speech” by those of whom the attribution is made. One reason given is that otherwise “ we cannot make the fine distinctions between thoughts that are essential to the explanations we can sometimes confidently supply” in making attributions. He notes the familiar examples of belief claims of human agents where substitution of

“Throught and Talk,” especially pp. 19-23.


coreferential terms converts true belief claims into false ones. Language is a means of “fine tuning.” We sometimes mistakenly or ambiguously report the thoughts of others or even of ourselves. We do not always “know our own mind,” but non-language users make no attributions at all.

Granted that faulty attributions to language users can often, be corrected by proper interpretation - often by an agent’s first-person honest testimony about what he or she thinks, believes, desires, although recent research seems to support the conclusion that even sincere first person verbal accounts of rea- sons for behavior are not as reliable as we supposed. But confirming assents are not possible in the case of non-linguals.

Yet is that, on the face of it, sufficient grounds for denying a non-lingual thoughts, beliefs, or desires altogether? That is a large step. Davidson claims that any attribution of thoughts of any kind to “dumb creatures,” to non-language users, is ruled out altogether. Not because of the obvious difficulties of correct attribution and fine tuning but because all thought he says is essentially linked to language. Davidson outlines some considerations by virtue of which ”to have a thought would be to have a disposition to utter certain sentences with appropriate force under given circumstances.” He sets out to show, “that only creatures with speech have thoughts” and that believing, doubting, desiring and other “sentential” attitudes, are not detachable from thinking or conversely.

The argument runs something like this. The basis for interpretation is the sentential attitude - the attitude of holding true. A feeling about sentences. That feeling is in turn a function of what the sentence means and what else is believed. The meaning of the sentence is given by agreed upon truth conditions in a linguistic community on an extension of a Tarski-style semantics. It is also taken as true that most beliefs are correct since too many false beliefs frustrate the possibility of identifying true beliefs. A large set of sharedstable beliefs is required to identify further beliefs whether true or false. On this account we are supposed to see how essential common agreed upon assentings are to a theory of belief. Not only must there be a large measure of agreement but “a theory of interpretation cannot be correct that makes a man assent to very many false sentences. It must generally be the case that a sentence is true when a speaker holds it to be. . . A method of interpretation. . . puts the interpretation in general agreement with the speaker.” On Davidson’s account, which he calls a theory of radical interpretation, concepts of truth and error are language bound and “Belief takes up the slack between objective truth and sentences held true.’’ But, and here is the giant step, belief as an attitude outside of the context of a shared language is therefore not intelligible. More generally, only a speech interpreter can have the concept of a thought or the


concept of a belief or the concept of a mistake. Without the exercise of a capacity for linguistic behavior an agent, he says, has no such concepts and hence has no thoughts or beliefs or even desires. Where language is absent, other behavioral markers of believing, desiring, thinking are ruled out altogether. It is in this sense that such language-centered views are in my view anthropocentric and non-naturalistic and in many features just a small step from Cartesianism. One is reminded that Descartes denied that non-humans had pains since they had no pain thoughts. Never mind the physical trauma and overt behavior so similar to ours. That “behavior” was just a failure of mechanism. Our empathy is misplaced anthropomorphism. I’m supposing that Davidson, unlike Descartes and Le Mettrie, allows animals their pains. But is it not equally Cartesian to deny that organisms engaged in such seemingly intentional behavior, often similar to our behavior in similar circumstances, have no intentions or desires or beliefs. In a previous paper lo I described a man and his dog in a desert, long deprived of water with observable physiological manifestations of thirst, racing toward a mirage which appears as water. The mirage recedes, they both behave as one does when expectations are unfulfilled. If the disappearance of the mirage was sudden, they may act startled. But on Davidson’s account, only the language user desires to drink, believes that there is something potable ahead, and then is mistaken.

In “Rational Animals,” Davidson l1 reviews his arguments for the insepar- ability of belief (and indeed all intentionality and intentional action) from language. To have a belief it is necessary to have the concept of belief, and to have the concept requires language. He bolsters his argument with the element of surprise. Surprise he says is “a necessary and sufficient condition for thought in general.’’ “Surprise” he takes to be what would be evoked in an agent when a sentence held true is shown not to be so. But it does not seem to me that first- order surprise of an agent requires that there be a sentence held true which is shown to be false. It can also be manifested by distinguishable non-linguistic responses to the unexpected and can be noted in linguals as well as pre-lingual children - and for all I know, in animals.

We noted above that Descartes, viewing non-human animals as robots, denied them pain. There is a curiously resonating analogy in “Rational Ani- mals.” Davidson notes that a self-guided missile might be described as want- ing to destroy its target and believing that the end would be realized by taking

lo “Some Revisionary Proposals about Belief and Believing,” Philosophy and Phenom-

l1 “Rational Animals” in Actions and Events, ed. E. Le Pore and B. I. McLaughlin enological Research I (Supplement 1990): 133-53. Also Modalities Oxford 1993.

(Blackwell, 1985): 473-80.


the path it takes. But of course, he says, we know that those are the beliefs and desires of the design engineer, and to attribute them to the missile is clearly misattribution. The missile lacks the mental, physiological, or behavioral complexity of a thinking being, as well as language. In the case of similar attributions to some non-lingual animals we are misled by the fact that many non-lingual animals have a repertoire of a wide range of non-longual behavior, as well as physiological make-up, similar to humans. It may be useful, Da- vidson says, as a manner of speaking, to deploy a belief, desire, action framework and vocabulary in the absence of further explanations about the origins of the seemingly intentional actions as in the case of the missile. We can, he says, “continue to explain the behavior of speechless creatures by attributing propositional attitudes to them while at the same time recognizing that such creatures do not actually have propositional attitudes and hence have no intentions, thoughts, beliefs, desires.” Such creatures ore autometa.

Richard Jeffrey’s paper, “Animal Interpretation” is a succinct response. He describes undertakings of animals which are continually adjusted to ends and have all the features of intentional actions. He notes that rats have pat- terns of preferences which have the clear structure of probabilistic judgments, absent the concept of subjective probability. He claims, as have I, that we can try to deploy a minimalist language in belief attribution to non-linguals which more accurately describe non-lingual motives or intentions or beliefs. He re- jects the implausible global, holistic, account of all belief - where ascribing a belief also requires ascribing an awesomely complex (possibly total) network of beliefs. We can, and I agree, determine that Fido is thirsty, desires to drink, believes that there is something drinkable nearby, and would satisfy that desire in the act of drinking as in the case of humans, without getting tangled in the total web of objects of propositional, i.e., sentential, attitudes. We can for example, empathize from our own experience. As Jeffery puts it, in the case of animals, the nodes in the account are non-propositional and sub-doxastic.

The range of believings of non-linguals are clearly vastly narrower than our believings. We have beliefs about language, for example, about remote pasts and futures, and also second order beliefs, second order desires and intentions, and the like which figure in Davidson’s narrow account of rationality. Such second order thoughts and intentions may in fact entail propositional or sentential attitudes. But recent research in animal psychology and psychology of pre-linguals would seem to suggest that the apparent range of intentional,

l2 “Animal Interpretation,” ibid.: 481-87.


planned, experientially grounded, motivated behavior of many non-lingual animals and pre-lingual children is not so narrow as previously supposed. l3

What makes me read Davidson’s program as non-naturalistic and Carte- sian is not that he believes in disembodied minds, or thoughts. He does not. Tokens of believing (i.e., the state of holding a sentence true) in an agent are, for him, correlated with tokens of brain events. It is rather the radical discon- tinuity, the great divide he sees between all intentionality of post-lingual human behavior and seemingly intentional behavior of pre-lingual and non- lingual animals. In the latter case, such attribution, as in the case of rockets, is also only a manner of speaking.

Consider by contrast the radical continuity argued by Hume. l4 In “Of the Reason of Animals,” he says:

Next to the ridicule of deying an evident truth, is that of taking much pains to defend it; and no truth appears to me more evident, than that beasts are endow’d with thought and reason as well as men. The arguments are in this case so obvious, that they never escape the most stupid and ignorant.

We are conscious, that we ourselves, in adapting means to ends, are guided by reason and design, and that ‘tis not ignorantly nor casually we perform those actions, which tend to self-preservation, to the obtaining pleasure, and avoiding pain. When therefore we see other creatures, in millions of instances, perform like actions, and direct them to like ends, all our principles of reason and probability carry us with an invincible force to believe the existence of a like cause. ‘Tis needless in my opinion to illustrate this argument by the enumeration of particulars. The smallest attention will supply us with more than are requisite. The resemblance betwixt the action of animals and those of men is so entire in this respect, that the very first action of the first animal we shall please to pitch on, will afford us an incontestable argument for the present doctrine.

This doctrine is as useful as it is obvious, and furnishes us with a kind of touchstone, by which we may try every system in this species of philosophy. ‘Tis from the resemblance of the external actions of animals to those we ourselves perform, that we judge their internal actions likewise to resemble ours; and the same principle of reasoning, carry’d one step farther, will make us conclude that since our internal actions resemble each other, the causes, from which they are deriv’d, must also be resembling. When any hypothesis, therefore, is advanc’d to explain a mental operation, which is common to men and beasts, we must apply the same hypothesis to both; and as every true hypothesis will abide this trial, so I may venture to affirm, that no false one will ever be able to endure it. The common defect of those systems, which philosophers have employ’d to account for the actions of the mind, is, that they suppose such a subtility and refinement of thought, as not only exceed the capacity of mere animals, but even of children and the common people in our own species; who are notwithstanding susceptible of the same emotions and affections as persons of the most accomplish’d genius and understanding.

l3 Donald Griffin, Animal Thinking (Cambridge, Mass: Harvard Press, 1984). l4 David Hume, A Treatise of Human Nature, ed. L. A. Selby-Bigge, second edition

(New York: Oxford University Press, 1978): 176-79.


Such a subtility is a clear proof of the falshood, as the contrary simplicity of the truth, of any system.

Let us therefore put our present system concerning the nature of the understanding to this decisive trial, and see whether it will equally account for the reasonings of beasts as for those of the human species.

Here we must make a distinction betwixt those actions of animals, which are of a vulgar nature, and seem to be on a level with their common capacities, and those more extraordinary instances of sagacity, which they sometimes discover for their own preservation, and the propagation of their species. A dog, that avoids fire and precipices, that shuns strangers, and caresses his master, affords us an instance of the first kind. A bud, that chooses with such care and nicety the place and materials of her nest, and sits upon her eggs for a due time, and in a suitable season, with all the precaution that a chymist is capable of in the most delicate projection, furnishes us with a lively instance of the second.

As to the former actions, I assert they proceed from a reasoning, that is not in itself different, nor founded on different principles, from that which appears in human nature. ‘Tis necessary in the first place, that there be some impression immediately present to their memory or senses, in order to be the foundation of their judgment. From the tone of voice the dog infers his master’s anger, and foresees his own punishment. From a certain sensation affecting his smell, he judges his game not to be far from him.

Sesondly, The inference he draws from the present impression is built on experience, and on his observation of the conjunction of objects in past instances. As you vary this experience, he varies his reasoning. Make a beating follow upon one sign or motion for some time, and afterwards upon another; and he will successively draw different conclusions, according to his most recent experience.

Now let any philosopher make a trial, and endeavour to explain that act of the mind, which we call belief, and give an account of the principles, from which it is deriv’d, independent of the influence of custom on the imagination, and let his hypothesis be equally applicable to beasts as to the human species; and after he has done this, I promise to em- brace his opinion. But at the same time I demand as an equitable condition, that if my system be the only one, which can answer to all these terms, it may be receiv’d as entirely satisfactory and convincing. And that ’tis the only one, is evident almost without any reasoning.

Paul Churchland l5 said that what is called by him and others “folk psychology,” a “theory” which takes believing and intending as sentential or quasi- sentential attitudes “has changed little or none since ancient times.” That is an inaccurate and a - historical view. Hume’s Treatise is explicity a theory of psy chology. Granted that Plato described thinking as an internal dialogue, but for Plato proper thinking was not about particulars. It was not about practical reason. “Opinions” about particular persons or events were not for Plato of doxastic interest. For Aristotle, who addressed practical as well theoretical reasoning, not all reasoning is propositional. Aristotle’s account of practical

l5 “eliminative Materialism and Porpositional Attitudes,” Journal of Philosophy 78 (1981).


reasoning, if he is taken literally, does not take language or propositions as essential to an explanation of intentional actions.

It has been plausibly claimed that the practice of srafting propositional or sentential argument forms onto Aristotle’s theory of practical inference of agents is mistaken. That such a grafting is spurious except as a third person attribution, has been persuasively defended by M. Nussbaum, l6 nor is her reading idiosyncratic. Action or refraining from action is the conclusion of a practical syllogism, not a proposition.

Actions must be motions. The triple <Desire, Belief, Action> is, for A r i s - totle, explanatory but neither desires nor beliefs are necessarily dependent on language. His theory is after all deployed to explain intentional or goal directed behavior of all animals, including non-linguals.

As Nussbaum puts it 17, “Desires and Beliefs [in Aristotle] are in the practical syllogism itself not [isolatable] Humean causes: They cannot be inde- pendently picked out, and yet, under another description, Aristotle believed they are causes. Each of them can also be described as some physiological change, and these changes cause the motion.. . Of course in most particular cases the appropriate physiological description for a complicated desire or belief will not be available to us. Aristotle never (I believe), ever claims that the two descriptions are related in any constant or predictable fashion. Token-re- lations like these are not going to be of much use to the scientist.. . In some cases, e.g. desire, it seems to be an article of faith with Aristotle for which he has no evidential support that there is such a physiological description. But that he wants to assert both a causal and conceptual relationship among desire, belief, and action and along what lines he tries to do this should be evident from MA‘s (De Mom Animalium’s) double account.’’ More recently, G.H. von Wright, l8 in a paper on practical inference specifically developed along Aristotelian lines, says, “It‘is of the essence of propositions that they are expressed by sentences.. I’ He notes that theoretical reasoning is propositional and language bound - but goes on to say that in practical inference, “Wants, states of knowing or believing, and acts have no analogous essential, COMedion with language. The relation of language to practical inference in the first person is in principle different from the relation to language of practical inference in the third person? Language, on von Wright‘s reading, may be

l6 Aristotle’s de Motu Animalium (Harvard, 1978), esp. Essay 1 and Essay 4. Ibid., p. 188. G. H. von Wright, “Practical Inference,” Philosophical Review 72 (1963): 168-69. It

was Nussbaum’s reference to von Wright which directed me to his interesting discussion of the practical syllogism.


It occurred to me in reading De Motu Animalium and Nussbaum’s and von Wright’s commentaries, that Aristotle might consistently have taken the position that theoretical reasoning is a species of practical inference peculiar to language users who, in addition to having a range of speculative, represen- tational, logical abilities, have higher order desires to which linguistic abilities may be essentially connected.

Nussbaum says of Aristotle, that in the literal reading of De Motu “the conclusion is in the practical case, not a proposition but an action. . . of course a human agent might verbalize his conclusion but this will not in any way alter our account of his behavior.” But here again, there is the persistent tendency to ignore the fact that assenting to a sentence which describes a state of affairs, or to a sentence which is a logical consequence of prior assents, is an action. Such assenting may sometimes be instrumental in practical reasoning which culminates in actions that are not speech acts. In purely theoretical reasoning, viewed as a species of practical reasoning, we might take the operative desirelg as the desire to know for its own sake, which Aristotle attributed only to humans, and the speech act of assenting or asserting is the outcome of theoretical reasoning. Here too, the conclusion is an action, a speech act.

The model of Plato’s internal dialogue may be, despite Churchland, plausible for much of theoretical reason but even here it seems inadequate to account for idiot savants or, for example, the mathematical insights of someone like Ramanujan 2o some of whose astonishing claims about numbers seemed often, to Hardy’s despair, not to be bolstered by arguments or proof or prior premises. It is more plausible to suppose that the mode of thought even in such an abstract context, was not to be explained as wholly linguistic although jus-g those insights, given the norms of mathematics, required finally, set- ting out a proof. In my example of the thirsty desert wanderer and his dog, there is no reason to suppose that either the desires, beliefs, and actions of the desert wanderer or his dog are accompanied by a sententidobligato of mental assentings.

Aristotle, on literal interpretation, Hume, and, more recently, versions of behaviorism, Ryle, versions of pragmatism, some recent modal accounts of cognition, do not fit Churchland’s description of folk psychology. Nor is what they call folk psychology very folksy or commonsensical. Recall that Hume in

l9 “All men desire knowledge,” The Metaphysics, Book I, 1 ( h e b Library, Harvard

2o Robert Kanigel, The Man Who Knew Infinity: A Life ofthe Genius, Ramanujan(New Press, 1980).

York: Scribner’s, 1991).


the quoted passage argues that attributing thought and reason to non-lingual animals is so obvious it does not escape even the most stupid or ignorant folk.

Representing in language is one mode of representation which is a dra- matic enlargement of our cognitive abilities but it is a natural ability and far from the whole story about thought. Imaging, for example, is a mode of thinking. It is closer to direct perception but it is a mode of thinking. Some language-centered accounts of thought, though they acknowledge mental “action” like silent affirmations, have resisted the notion that there actually are mental picturings. They claim that the notion of a mental-picture is mistaken psychology, and a conceptual mistake at that. The status of mental images in reasoning has been a matter of continuing debate among psychologists, *l although its cognitive role is once again acquiring respectability.

There are very likely, even among language users, a wide range of cognitive states and styles which ground belief and which have a wide range of neuro-physiological underpinnings. The extraordinary thing about linguistic capacity is that we can communicate about what we believe through speech acts in a public language despite the variant causal histories. The speech act of assenting to a sentence ‘P’ which assenting we share with others in our speech community may in each case terminate a process which emerges from quite divergent mental or neurophysiological causal histories. One can say impress- ionistically, that we might view Hume as focussing on a narrow range of representations; sensory or pictorial modes or representation. Language focussed theories of mind are preoccupied with linguistic representation which are at some causal distance from objects represented. The causal link between pictorial representation, perceptual and sensory experience, and the world, is close. But as we Know, a picture theory of linguistic representation is not sus- tainable. In a complex organism there is a chain of learning grounded initially in encounters with the environment, and undergirded by our neuro-physio- logy - which may show considerable individual variation given our variant histories and cognitive styles. A genuine folk psychologist is more likely to offer highly condemned platitudes such as that seeing and feeling is believing, or that non-verbal behavior of even language users is a reliable guide to what they believe, than that believing is having a pro-attitude to an understood sentence by a member of a speech community. The Cartesian analogue of language centered views is to see believing, thinking, intending, in all their possible manifestations as necessarily restricted to language users. Replace “having a soul” with “having language”. Nor am I motivated to question this languagecentered non-naturalistic account because I am one of those who

See Michael Tye, The Imagery Debate (Bradford, M.I.T., 1991).


Davidson refers to as an “animal lover,” to which he might have added, baby lover. It is because such a restricted theory generates problems, not just with respect to the believings of non-language users but in conjunction with believings of language users.

I1

The legitimacy of attack on what is called folk psychology should not be addressed to the desire-belief-explanatory model of action, writ large, but with the centrality of sentential or quasi-sentential entities or states as the only proper objects of verbs such as think, intend, believe, desire, expect, and the like. No language, then no thought, no intentions, no desires, etc. On an inter- nalist folk-psychological view, the mind is in its essential features a complex system of evocable attitudes to internal linguistic entities or semantical states isolable and determinable. Hence the seeming plausibility of the mind as a souped up Turing machine or conceivable as a brain in a vat. But it is not plausible, although there are surely features of human mentality which can be so described.

Nor is the implausibility simply that the objects of propositional attitudes are on this “narrow” view, all in the mind (or head). Adding the “externalist rider” which requires that the sentential object of attitudes must be interpreted sentences within the framework of a public language does not appreci- ably ameliorate the implausibility. Meanings on an externalist account are world bound and are not wholly in the agent’s head. But that does not make the objects of thought, desire, belief, intention less linguistic, but only less so- lipsistic.

It seems to me that the Desire-Belief model of action explanation is a plausible, rough, phenomenological theory provided that the role of language is assigned its proper place. A kind of Copernican rearrangement is warranted with language no longer at center stage.

Complex organisms vary enormously with respect to their needs and capacities - and those capacities will naturally play a role in action explanation. The special capacities for language and for the wide range of communication, speculation, and inference it affords will, of course, play a role in human action explanation, but it is hardly the whole story.

We began the paper by characterizing the disquotation principle as central to language-focussed accounts of belief. Sincere assent entailed belief and arguably conversely, among competent non-confused language users. On a more general account, the objects of believing are not necessarily linguistic entities although such entities will sometimes be the objects of believing as in


the belief that a certain sentence is grammatical. The disquotation principle will fall into place as a claim which takes speech acts of assent by competent, sincere, non-conceptually confused linguals as a privileged mark of believing. A claim which recent research suggests may not be wholly warranted. 22

Consider the following more general dispositional account. 23

D. An agent believes that S just in case (1) under agent-centered circumstances such as desires, need, an other psychological states including other believings and (2) external circumstances (3) the agent wil l act as if S obtained, i.e., will act in ways appropriate to S being the case, where S is a state of affairs, actual or non-actual.

Paraphrasing an example of Braithwaite, 24 elaborating on what it is to believe that strawberries give him indigestion, he says that to believe that strawberries give him indigestion means that under relevant external circumstances such as being offered strawberries and internal needs and desires such as the need to preserve health and to desire to avoid pain, he will behave in a manner appropriate to their indigestibility. For example, he will refuse them. Under other circumstances where he has a need to have indigestion to avoid an unpleasant obligation, he will accept the strawberries.

Who, under D are the agents? They are organisms of sufficient complexity and apparently goal directed behavior such that viewing them, as closer to hard wired rockets, the design of which we happen to be ignorant, is unwar- ranted - unless of course we should also so view ourselves.

The hungry chimpanzee who improvises “tools” and tries out various strategies for obtaining what looks like bananas out of reach could surely be counted as an agent who believes that the objects within view (we call them ‘bananas’) are edible. The chimp is related to a possible state of affairs, the edibility of that object and acts accordingly.

It is equally plausible that Norman Malcolm’s now celebrated dog believes that whatever it is it pursued (we call it a ‘cat’) is in an object where it disap- peared from view (we call it a ‘tree’).

** See T. Wilson, “Strangers to Ourselves: The Origins and Accuracy of Beliefs About Ones Own Mental States,” ed. J. Harvey and G. Weary, Attribution in Contempomry Psycho- logy (New York: Academic Press, 1988). Also, notes 26 and 28.

23 Principle D is adapted from R. B. Braithwaite, “The Nature of Believing” in Knowl- edge und Belief, ed. A. P. Griffiths (Oxford Press, 1967).

Braithwaite required the entertaining of proposiaons and that the agent act as S were true. Since Braithwaite’s propositions do seem to be quasi-linguistic entities, I have modified his Bcco\ult so that the principle is not essentially language bound. Braithwaite also required that S bs contingent (or in my account, a possible state of affairs).

See my “Some Revisionary Proposals About Belief and Believing” and ”Rationality and Believing the Impossible,” Journal of Philosophy LXXX (1983).

a Bmithwaite, p. 31.


Of course as we descend the hierarchy of animal organisms, the behavior may become more analogous to hard wired rocketry. But it does not seem to me that a cut-off point is clear, or necessary, or possible. Sorites are every- where.

What counts as “acts as if“ under D? Perhaps “responds as if“ would be preferable to “acts as if“ which in turn is preferable to “behaves as if.’’ The latter has been identified with the rigid constraints of Watsonian or Skinnerian Behaviorism which insisted on publicly displayable responses - thereby allowing uttering out loud or vocal chord motion as behavior, but excluding mental utterances in accounts of thought. But mental events such as framing intentions, making decisions, picturing, humming to onself will count as actions - whose connections with explanation of further action are part of common sense psychology divested of language centricity. If the “acts as if“ is broadened to “responds as if“ then that leaves room for further neuro-physiological responses in some causal hierarchy. But so far, and despite the promises of cognitive theorists, such deep theory is unavailable, and we “theorize” or explain on the phenomenological level.

It is a curious fact that the obsession with linguistic behavior for defining belief, intention, thought, rationality seems to motivate animal psychologists to try to locate languages in animals25 who are otherwise so patently intentional in their behavior. As if that were required if their behavior is to be en- dorsed as intentional or rational or even cognitive. But the putative language ascribed for example by Primack to chimps is abjectly primitive compared to the complexity of some of the chimps’ other goal directed behavior. Indeed, some linguists are reluctant to characterize the behavior as genuinely linguistic.

Robert Stalnaker is among those who has proposed an account of the objects of propositional attitudes which are not the states of &airs I have in mind, but they are also clearly non-linguistic entities. For Stalnaker the inten- sions of sentences are not linguistic nor do they closely mimic language. They are rather functions from worlds to truth values. I have difficulties with those objects as the objects of attitudes although there is a powerful rationale behind it which must be weighed but I will not consider here. But what k stressed by Stalnaker and which I have been urging, is that “If desires and beliefs are to be understood in terms of their role in the determination of action, then their objects have nothing essential to do with language.” “It is conceivable (whether or not true) that there are rational creatures who do not use language, who

25 D. Premack, The Mind of An Ape (New York, 1983). Also Gavagai: Or the Future History of the Animal Language Contmversy (Cambridge, Mass.: MIT Press, 1986).


have no internal representations of their attitudes which have a linguistic form. I think this is true of many animals - even rather stupid ones - but there might be clearer cases.”

Even among language users, an account of the attitudes which does not take the objects of such attitudes to be necessarily linguistic or quasi-linguistic, permits a better solution to problems not easily addressed by language bound uses. Consider the related phenomena of tacit, unconscious and conflicted belief.

In accordance with our account, the propositional attitude of believing is not confined to a disposition to verbal behavior. A range of actions other than speech acts may count as evidence for an agent’s being in a believing relation to a state of affairs, not necessarily actual. Long before some recent psychological research, 26 which seems to support the claim that “cognitive” causes of non-verbal behaviors are distinct from cognitive causes of verbal behavior, it was, among real common-sensical folk, recognized that a language-endowed agent’s non-verbal behavior might run counter to his sincere assents to sentences understood and held true. There is the familiar case of someone who sincerely avows he loves another yet through non-verbal action and seemingly unaware of the dissonance, persistently inflicts harm. Such examples of conflicted believing are echoed in examples of conflicted desiring.

If believing is a verbal disposition then how to account for the occasions of failure of fit of non-verbal actions (broadly conceived) with verbal actions of sincere assent. This is not a matter of inconsistency of sentences assented to but of a lack of coherence between sincere verbal responses and other manifestations of belief. Those for whom believing is, of necessity, language bound could say that there are sentences unconsciously held true to which there is unconscious assent. Those sentences causally and selectively are claimed to influence our non-verbal actions but not our conscious verbal actions which manifest our conscious believing. It is as if there is within us a hidden alter- ego, assenting to sentences which do contradict those consciously assented to.

But that is an hypothesis difficult to sustain. 27 More plausible is the range of explanations recently advanced for the fairly frequent divergence of verbal from other behavioral indicators of believing, not all of which can be ac- counted for by a repressive mechanism. Non-linguistic behavioral indicators of believing sometimes have no linguistic behavioral correlate. Sometimes they are in parallel with verbal behavioral indicators and their failure to mesh

See R. Nisbett and T. Wilson, “Telling More Than We Can Know,” Psychological Re-

27 S e e A. Collins, “Unconscious Belief,” Journal of Philosophy LXVI, No. 20 (1969). view 84, No. 3 (1977).


seems not always to be motivated in a way consistent with psychoanalyhc accounts. Indeed, Timothy Wilson says 28 “people can often make accurate re- ports about their attitudes, moods, motives and evaluations. Often they cannot and it is under these conditions that behavioralmeasures of internal states are useful.”

A non-language centered account of cognitive attitudes allows a more plausible account of rationality. Rationality as common-sensical folk know, is not merely a matter of being a reliable deducer of logical consequences or a good computer of probabilities or a non-assentor to contradictory sentences ans so on. Rationality is a feature of behavior writ large. It is not misplaced rocket-talk to describe some behavior of a non-lingual as rational, or some behavior of a logical lingual as irrational where in the latter case norms of logic are preserved but other behavior is at odds with sincere verbal assent. I ex- clude here such cases as those given in the puzzle about belief. (See footnote

A purpose in this paper was to challenge what philosophers have recently called “folk psychology”, not for its recalcitrance to reduction or for the difficulties in pinpointing reference, but for its being a highly non-commonsensical, non-naturalistic account of intention, cognition, thought and action. It is an account of thought, of intentional states of cognition, of many seemingly affective attitudes which is grounded exclusively in attitudes to linguistic or quasi-linguistic entities. That is already questionable for verbs like “believe” and far more questionable for verbs like desire, need, want, pursue, seek, im- agine, and a large family of cognates. For genuine common sense psychology, sentences or quasi-linguistic propositions do not play an exclusive role. Speech acts are among the many acts which are manifestations of cognitive states and crucial as language is to theoretical reasoning, it is only a part of the story in the explanation of action of linguals. It plays no role in explanations of action of non-linguals. But the absence of language does not primafaciecon- vert non-linguals into thoughtless automata.

2)

za T. Wilson, “Self Deception Without Repression,” p. 113.

Dialectica Vol. 49, No 2-4 (1995)

the anti-naturalism of some language centered accounts of belief

Documents