Consciousness, Intentionality and Function. What is the right order of explanation?


(published in Philosophy and Phenomenological Research, vol. LV, n1, March 1995, 195-200)



Pierre Jacob





Searle has written a provocative book. What is most controversial about it is not his solution to the mind-body problem, namely that "mental phenomena are caused by neurophysiological processes in the brain and are themselves features of the brain" (p. 1); nor that 'the brain causes certain "mental" phenomena, such as conscious mental states, and these conscious states are simply higher-level features of the brain' (p. 14). In spite of Searle's own disclaimers, this is all fairly standard physicalism.

Nor is his dichotomy between intrinsic and mere as-if intentionality controversial, at least not in the sense in which, as most other intentional realists, Searle assumes that, in the proper order of explanation, words derive their meanings from the contents of thoughts. For Searle, however, "intrinsic" does not merely mean primitive or underived. It also means internal or non-relational: "the moon intrinsically has a mass, but is not intrinsically a satellite" (p. 80). A fortiori it means non-relative to a conscious agent with beliefs and desires: having a mass, unlike being a chair, is thus said to be intrinsic (p. 211). Finally, if a property is not observer-dependent, then it is intrinsic in the sense of being "the real thing and not just something more or less like the real thing (as-if), or something that is the result of somebody else's uses of or attitudes toward the thing (derived)" (p. 80). Like other intentional realists, and unlike Dennett, Searle then assumes that "my having [a given] belief is not a matter of what anybody else chooses to say about me or how I behave or what sort of stance someone might adopt toward me" (p. 155).

Before I turn to the two really controversial theses in Searle's book, let me note one puzzling consequence of his distinction between intrinsic and non-intrinsic properties. Syntactic properties are not intrinsic properties of a symbol: they are assigned or stipulated by conscious agents (pp. 207-10). Now, it is uncontroversial both that syntactic properties of a symbol are not basic physical properties of the symbol and that syntactic properties of symbols of formal artificial languages are assigned by conscious agents. But what of the syntactic properties of symbols of natural languages? They might be relational in the sense in which being a satellite of the earth is relational. But it does not follow that they are in any obvious sense observer-dependent, i.e., that they are relative to a conscious linguist with beliefs and desires. If they were, then, not only (as Searle emphasizes in ch. 9) would the idea of deeply unconscious rules of syntax be wrong (because of the primacy of consciousness thesis about which more in a minute), but, as Searle does not make explicit, the goal, widely shared in linguistics, of trying to discover by empirical investigation the syntactic properties of natural languages, would be incoherent.

Now, the two theses in Searle's book which are really controversial are his thesis of the primacy of intentionality over function and his thesis of the primacy of consciousness over intentionality. According to the former, all function-attributions are relative to systems having intrinsic intentionality. According to the latter (the so-called Connection Principle), a system's intrinsic intentionality depends upon the system's being conscious. I take both to be deeply puzzling.




It follows from Searle's thesis of the primacy of intentionality over function that function-ascriptions are not straightforwardly factual; they are normative and value-dependent. Their truth is observer-dependent or relative to a conscious agent with intrinsic intentionality.

One might object to the claim that artifacts are always designed by conscious agents with beliefs and desires: arguably, dams and nests are artifacts built by beavers and birds; and it is at least controversial whether beavers and birds are conscious agents with beliefs and desires. I will, however, waive this objection. Presumably, Searle would agree that, unlike artifacts, biological organs do not owe their existence to a conscious designer with intentions, beliefs and desires. If you wonder about the function of an artifact, just ask the designer. He will tell you what the function is. If a conscious agent decides to change the function of an artifact, he too can tell you what the new function is. Unlike artifacts, however, biological organs have not been created by conscious agents. Now, the beliefs of a biologist investigating the function of an organ are theoretical, fallible, and open to empirical refutation in a way that the propositional attitudes of the conscious designer who created an artifact are not. The latter are constitutive of the function of the artifact. It follows, on Searle's view, that the function of an artifact is less indeterminate (more objective) than the function of a biological organ which has been shaped by natural selection, not by a conscious designer with beliefs and desires.

Consider a human heart. On Searle's view, the facts "the hard, brute, blind, physical facts" make it true that the heart pumps blood, makes a thumping noise and exerts gravitational attraction on the moon. No such fact, however, can make it true that the function of the heart is to pump blood, not to make a thumping noise, nor to exert gravitational attraction on the moon. "In addition to its various causal relations, the heart does not have any function. When we speak of its functions, we are talking about those of its causal relations to which we attach some normative importance" (p. 238). Biological function, then, on this view, is entirely in the eye of the intrinsically intentional beholder.

In spite of the sharp contrast between Searle's strong intentional realism and Dennett's interpretive anti-realist conception of the intentional stance (which Searle so adamantly rejects), it would be really ironic if they turned out to agree that the ascription of biological function is somehow indeterminate, if not observer-dependent. I am not saying that this consequence has been demonstrably shown to be wrong although I do think it is. But if Searle is right, then we have all been wrong to think that Harvey discovered that the function of the heart is to pump blood. Given the enormously revisionistic implications of Searle's view for the ascription of biological functions in biology proper, it seems to me that the onus of proof is clearly on Searle: he must convince practitioner biologists that his philosophical interpretation of their function-talk is the right one. But in the book, he does not argue for his view of biological function; he merely asserts it.

If one thinks of the biological function of an organ as a subset of the set of effects the organ may have, then I see no reason why there may not be facts bearing on the truth of an ascription of function, independently of any intrinsically intentional observer. Along the lines of Wright (1973) and Millikan (1984), I would argue that a biological organ may derive its function - through natural selection - from the fact that its presence and/or persistence across time depends on its having this function, i.e., its producing a particular kind of effect.




I now turn to the thesis of the primacy of consciousness over intentionality. The Connection Principle (CP) asserts that intrinsically intentional unconscious mental states must be in principle potentially conscious or available to consciousness. If the CP is correct, how wrong, then, is the picture of cognitive science according to which a great deal of information processing occurs in an agent's mind (or brain), which is both intentional and inaccessible to the agent's consciousness?

Given the obvious loophole opened by the notion of accessibility in principle, one might argue - as Block (1990) has - that the CP is in fact trivially satisfied by standard cognitive science. By producing the very kind of linguistic theories or computational theories of early vision which Searle so vehemently objects to (ch. 9), cognitive scientists, such as Chomsky and Marr, have brought to the conscious awareness of some minds rules and representations, which would otherwise have remained deeply unconscious. Isn't this evidence that such unconscious rules and computations are potentially conscious in the required sense? Alternatively, representations and computations, which are inaccessible to our consciousness, might turn out to be consciously accessible to the minds of other creatures, with a different (perhaps more powerful) cognitive architecture.

Presumably, Searle's intended version of the CP is that, for an agent's inner state to be genuinely intentional (as opposed to being merely neurophysiological), the content of the state must be potentially conscious to the agent at the very moment it is causally efficacious in interacting with other intentional states or in contributing to the agent's intentional behavior. But again, what are the constraints on potential availability?

Consider the semantic facilitation obtained by Marcel (1983) in the subliminal perception of words, where semantic information about a word seems to be extracted unconsciously by a subject. Or consider blindsight patients studied by Weiskrantz (1986). In both cases, a person's intentional behavior is being influenced by the detection of a stimulus of which the person remains unaware. In such cases, the agent's behavior would seem to be caused by an intentional state whose content is inaccessible to his or her consciousness. And such states clearly are causally efficacious in the production of the agent's intentional behavior.

Of blindsight patients, Searle seems tempted to say (p. 163) that the contents of their perceptual states are potentially conscious, on the grounds that, in the same circumstances, the perceptual experience of a normal person would be available to his or her consciousness. By the same token, he might want to say of the subliminal perception of words that, were words presented slowly enough to the subject, the content of his or her experience would be consciously available to him or her. By making this move, however, Searle, it seems to me, turns the CP into a vacuous or irrefutable principle as noted above. Nor is it, I think, open to Searle to deny that, in the subliminal perception of a word, the state of the person who detects the word subliminally is intentional. The reason why not is that the person's behavior (caused by the state under discussion) is intentional and the relevant property of the detected stimulus is semantic, not physical.

Let us now consider Searle's argument for the CP. In steps 1 and 2 (p. 156), Searle assumes that conscious and unconscious intentional mental states must have intrinsic intentionality. In step 3 (pp. 156-7), conscious and unconscious intentional states are said to have "aspectual shape": "where there is no aspectual shape", says Searle (p. 164), "there is no intentionality". What is aspectual shape? As far as I can see, it is what most philosophers call intensionality: namely, the property of intentional mental states whereby a referential thought-constituent cannot always be substituted salva veritate by a coreferential one. So I can believe that the Evening Star is above the tree without believing that the Morning Star is. Or I can think that the bottle in front of me contains water without thinking that the bottle in front of me contains H2O. Notice that Fodor's language of thought hypothesis is precisely designed to explain what Searle calls aspectual shape and it has nothing to do with consciousness.

Now, according to step 4, the aspectual shape of a person's thought cannot be fixed either by the person's behavior or by neurophysiological facts about the person's brain. According to step 5, the only facts about an unconscious person ("in a sound dreamless sleep") that can make it true that he or she has unconscious beliefs (with aspectual shape) are neurophysiological states and processes in the person's brain (p. 159). Such neurophysiological facts lack aspectual shape. However, an unconscious person may still be ascribed beliefs with aspectual shape. The conclusion is that what preserves the aspectual shape of a person's beliefs when he or she is unconscious is their potential availability to the person's consciousness.

I now want to share with Searle my deep puzzlement about his account of aspectual shape. I fail to see how the appeal to conscious experience could solve Quine's puzzles about the indeterminacy of translation (referred to by Searle in his discussion of why there could not be an intentional zombie, pp. 163-64). Both a rabbit and its undetached (rabbit) parts will cause me to have the same rabbitish visual experience. When I scrutinize my rabbitish visual experiences, I simply cannot find any basis upon which to erect a distinction between my thinking about rabbits and my thinking about rabbit parts.

Both water and H2O will cause me to have the same waterish conscious experience, because water is H2O and everything having the former property will have the latter as well. What makes my water-thoughts differ from my H2O-thoughts - what gives them different aspectual shapes - is their respective structure, not my visual experience of a transparent or murky (as the case may be) looking liquid, nor my conscious "gustatory" experience of the "tasteless" liquid dripping on my taste buds and then flowing down my throat. What makes my water-thoughts differ from my H2O-thoughts is that the latter, unlike the former, involves my thinking about hydrogen, the number 2 and oxygen.

How on earth could my conscious gustatory or visual experience of a glass of water account for the above difference in aspectual shape? This is a mystery one is left pondering over after one has read Searle's book. One cannot but feel that consciousness is being assigned an impossible mission. Given my previous reservations about Searle's priority of intentionality over function and my present puzzlement over his view that somehow conscious experiences must have the power to account for aspectual shape, I think it would be worth considering the alternative strategy of deriving intentionality from function and consciousness from intentionality.[1]






Block, N. (1990) "Consciousness and accessibility", in Behavioral and Brain Sciences, 13, 4, 596-98.


Marcel, A.J. (1983) "Conscious and unconscious perception: experiments on visual masking and word recognition", Cognitive Psychology, 15, 197-237.


Millikan, R.G. (1984) Language, Thought, and Other Biological Categories, Cambridge, Mass.: MIT Press.


Weiskrantz, L. (1986) Blindsight: A Case Study and Implications, Oxford: Oxford University Press.


Wright, L. (1973) "Functions", in E. Sober (ed.) Conceptual Issues in Evolutionary Biology, Cambridge, Mass.: MIT Press, 1984.



[1] Many thanks to Karen Neander, Dan Sperber, and especially Georges Rey.