Portal de Periódicos da CAPES

Intension and representation: Quine's indeterminacy thesis revisited

2005; Taylor & Francis; Volume: 18; Issue: 4 Linguagem: Inglês

10.1080/09515080500229878

1465-394X

Itay Shani,

Embodied and Extended Cognition

Abstract This paper re-addresses Quine's indeterminacy of translation/inscrutability of reference thesis, as a problem for cognitive theories of content. In contradistinction with Quine's behavioristic semantics, theories of meaning, or content, in the cognitivist tradition endorse intentional realism, and are prone to be unsympathetic to Quine's thesis. Yet, despite this fundamental difference, I argue that they are just as vulnerable to the indeterminacy. I then argue that the vulnerability is rooted in a theoretical commitment tacitly shared with Quine, namely, the commitment to the view that the perceptual input to the cognitive system is extensional—differentiating objects, but not the aspects (or, properties) they manifest. Thus, input extensionalism, and not behaviorism, is what forces the indeterminacy. I conclude by suggesting that the solution to Quine's indeterminacy problem hinges on the elaboration of an intensional theory of perceptual input, and of content in general. Notes I use quotation marks to emphasize the distinction between the position Quine identifies as mentalism, and traditional idealism or Cartesian dualism. The point is that mentalism, as criticized by Quine and defended by cognitivist theories, is not tantamount to an outright rejection of the materiality of the mind but, rather, to an affirmation of cognitive realism—realism about inner mental states. Quine's disapproval of cognitive realism is primarily based on empiricist constrains on knowledge justification, i.e. on epistemological considerations, but the thesis itself is compatible with an ontologically naturalistic world view, a compatibility that is presupposed at the very core of cognitive science. In his classical address of the issue, Quine (1969 Quine WVO 1969 Ontological relativity Ontological relativity and other essays pp. 26–68 New York Columbia University Press [Google Scholar]) talks, quite generally, about ‘the inscrutability of reference’ and ‘the indeterminacy of meaning and reference’. Quine's indeterminacy is now conceived as only one amongst several problems of content indeterminacy—the ‘disjunction’ problem (Fodor, 1987 Fodor JA 1987 Psychosemantics Cambridge MA MIT Press [Crossref] , [Google Scholar], 1990 Fodor JA 1990 A theory of content and other essays Cambridge MA MIT Press [Google Scholar]) and the ‘depth’ or ‘distality’ problem (Dretske, 1986 Dretske F 1986 Misrepresentation In R. Bogdan (Ed.) Belief: Form, content, and function pp. 17–36 New York Oxford University Press [Google Scholar]; Godfrey-Smith, 1989 Godfrey-Smith, P. 1989. Misinformation. Canadian Journal of Philosophy, 19: 533–550. [Taylor & Francis Online], [Web of Science ®] , [Google Scholar]) are two familiar examples of alternative problems of content indeterminacy. Chomsky (1975 Chomsky N 1975 Quine's empirical assumptions In D. Davidson & J. Hintikka (Eds.) Words and objections: Essays on the work of W.V.O. Quine pp. 53–68 Dordrecht The Netherlands Reidel [Google Scholar]) has argued that Quine's problem is, essentially, no more than the problem of the underdetermination of theory by the data. However, as Searle (1987 Searle, JR. 1987. Indeterminacy, empiricism and the first person. Journal of Philosophy, 84: 123–146. [Crossref], [Web of Science ®] , [Google Scholar]), Gibson (1988 Gibson R 1988 Enlightened empiricism: An examination of W.V.O. Quine's theory of knowledge Tampa FL University of South Florida Press [Google Scholar]) and Quine himself (1987) show, there is an important difference. On Quine's view there are simply no semantic facts beyond dispositions to respond to verbal stimuli, and since such dispositions are indeterminate vis-à-vis competing translation manuals and referential assignments there are no determinate semantic facts either. The problem, then, is not that the facts are not determined by the data but rather, that there are no facts of the matter at all. One may argue, of course, that once Quine's behaviorism is rejected and our confidence in the existence of hidden semantic facts is restored the problem disappears; but this is misleading and mistaken. It is misleading because even if it is agreed that there are underlying semantic facts Quine's indeterminacy problem is not solved as long as our theories cannot explain why ‘rabbit’ means rabbits and not rabbit parts, stages, etc. As for the claim that it's mistaken, the rest of the paper is devised as an existential proof to that effect. In order to prevent possible misunderstanding, I would like to emphasize that in the remaining parts of the discussion I shall address QI as a problem for cognitive theory rather than as a linguistic puzzle and moreover, as a problem of individual meaning rather than a problem of communication on a social scale. My assumption is that if there's a solution to the problem, that solution necessarily involves facts about inner mental states. I don'st take an explicit stance on precisely what other conditions must obtain and how, nor do I argue here in favor of my assumption. The assumption, I believe, is in accord with the accepted wisdom in cognitive science, and since I am primarily concerned here with cognitivist theories of content I shall simply proceed and call some of them to task on account of their apparent lack of a satisfactory solution to Quine's riddle. Reference to linguistic matters will only be made, if at all, while mentioning Quine's original exposition of the subject. For a more exhaustive critical discussion of IS in connection with Quine's indeterminacy argument, see Gates (1996 Gates, G. 1996. The price of information. Synthese, 107: 325–347. [Google Scholar]). The selection may be on an evolutionary scale, as when the heart is ascribed the function of pumping blood, or on a developmental scale as in behavioral learning (cf. Dretske, 1988 Dretske F 1988 Explaining behavior: Reasons in a world of causes Cambridge MA MIT Press [Google Scholar]). In any case, the normative criterion for functional ascription is historical; an appropriate functional explanation is one that anchors the function of a trait in the effect for which it was selected. Millikan (1991 Millikan RG 1991 Speaking up for Darwin In B. Loewer & G. Rey (Eds.) Meaning in mind: Fodor and his critics pp. 151–164 Cambridge MA Blackwell [Google Scholar]) claims that her teleological account of content individuation is as fine-grained as causal explanations can be, and that the theory only fails to differentiate between a pair of properties F and G when they are indistinguishable in their causal powers. We can now see why this is wrong. The constraints that a teleological account such as Millikan's impose on content individuation, are simply not as fine-grained as she imagines. For any property F that Millikan identifies as the one that has been selected ‘for’ there is an equivalence class of alternative properties that could have been selected with equal ease because their adaptive value in the context of selection would have been similar enough to F. But this does not imply that the properties need to be causally identical. In fact, they could differ from each other in many significant respects—intentional as well as non-intentional. Indeed, Millikan's normalizing explanations are simply too abstract to manifest sensitivity to causal details, and they also run into serious problems of mental causation. Narrow content in the functional role sense of CRS is only one way of portraying the idea of a narrow content. Fodor (1987 Fodor JA 1987 Psychosemantics Cambridge MA MIT Press [Crossref] , [Google Scholar]), for example, once held that the narrow content of a thought is simply whatever it is that determines its truth conditions (and thus its broad content) relative to a context (see Marras, 1989 Marras A 1989 Psychological content Unpublished maniscript [Google Scholar]). However, this difference in characterizing narrow content makes no difference to the point I wish to make since the pairing problem (which, I argue below, is symptomatic to two-factor functional role theories) is replicated, mutatis mutandis, in Fodor's ‘truth-functional’ version of narrow content. I present the problem as a practical problem for representational systems, but the problem is also, a fortiori, a theoretical problem for the theory of representation. Note that it is unhelpful to assume that the arbitrariness could be blocked by reference to connections with other representations in the system since the (standard) content of these other representations is no less arbitrary. Interestingly, Fodor (1990 Fodor JA 1990 A theory of content and other essays Cambridge MA MIT Press [Google Scholar]) describes the similarities between Dretske's information semantics and Skinner's account of stimulus meaning. Fodor emphasizes the fact that both theories are physicalist and atomistic, and that both identify semantic relations with relations of lawful covariance between symbols and their denotata. Fodor opposition to Quine's holism has the consequence that he emphasizes the differences between Skinner and Quine, and between Quine and information semantics and de-emphasizes the similarities. But Quine's stimulus meaning is Skinnerian and the inability of SMT to generate semantic information is invariant between the two. If I am correct, then information semantics, which Fodor describes as ‘updated Skinnerian semantics’ (1990, p. 57), is in a similar condition and for similar reasons—covariance relations are extensional and any attempt to reduce content to such relations makes the problem of indeterminacy insolvable. Despite his eliminativism, even Quine himself realized that the existence of this semantic gap constitutes a problem. In particular, he was troubled by the absurdity of what follows from the reduction of meaning to stimulus meaning—the conclusion that there is no difference whatsoever between referring to rabbits and referring to rabbit stages, or between referring to formulas and referring to their Gödel numbers (1969, p. 48). Quine's way of dealing with the absurdity is to introduce his celebrated theory of ontological relativity. He argues that all that the indeterminacy shows is that it is an absurdity to think of meaning in absolute, non-relativistic, terms and that linguistic expressions are, indeed, meaningful only relative to a linguistic network, which network is analogous to a coordinate system in relativistic physics. Only with such background network as a frame of reference, Quine argues, meaningful speech acts are possible (‘relative to it we can and do talk meaningfully and distinctively of rabbits and parts, numbers and formulas’ (1969, p. 48; for a detailed criticism of Quine's relativity argument, see Searle, 1987 Searle, JR. 1987. Indeterminacy, empiricism and the first person. Journal of Philosophy, 84: 123–146. [Crossref], [Web of Science ®] , [Google Scholar]). ‘Enrichment’ is the terms chosen by the Gibsons (Gibson & Gibson, 1955 Gibson, EJ and Gibson, JJ. 1955. Perceptual learning: Differentiation or enrichement?. Psychological Review, 62: 32–41. [Crossref], [PubMed], [Web of Science ®] , [Google Scholar]) to characterize the view that perceptual input must be constantly supplemented by additional epistemic factors (be it ideas, or—as in the more recent vocabulary of cognitive science—inferences). Bickhard and Richie (1983 Bickhard MH Richie DM 1983 On the nature of representation: A case study of james gibson's theory of perception New York Praeger [Google Scholar]) use the term ‘enhancement’ to express the same idea. Fodor and Pylyshyn don'st use the term ‘intensionality’. Instead, they speak of intentionality and of intentional contexts in opposition to extensionality and extensional contexts. However, nothing in this terminological difference need bother us. Although ‘intentionality’ and ‘intensionality’ are not identical in meaning (see Cornman, 1972 Cornman JW 1972 Intentionality and intensionality In A. Marras (Ed.) Intentionality, mind, and language pp. 52–75 Chicago University of Illinois Press [Google Scholar]; Searle, 1983 Searle JR 1983 Intentionality: An essay in the philosophy of mind Cambridge England Cambridge University Press [Crossref] , [Google Scholar]), nothing in Fodor and Pylyshyn's discussion hinges on the difference. Indeed, the properties they identify with intentional entities and intentional discourse—opacity, representation under a mode of presentation, failure of substitutivity, etc.—are all intensional. For the gist of this argument against the appeal to nativism, applied to the problem of representation but without special emphasis on the question of intensionality, see Thelen and Smith (1994 Thelen E Smith L 1994 A dynamic systems approach to the development of cognition and action Cambridge MA MIT Press [Google Scholar], p. 31), Bickhard (1991 Bickhard MH 1991 The import of Fodor's anti-constructivist argument In L. P. Steffe (Ed.) Epistemological foundations of mathematical experience pp. 14–25 New York Springer-Verlag [Google Scholar]) and Gibson (1979 Gibson JJ 1979 The ecological approach to visual perception Hillsdale NJ Lawrence Earlbaum [Google Scholar], p. 253). Another significant feature of the perception of affordances, and, indeed, part of what makes this perception aspectual, is the fact that it is made from a perspective—the perspective of an active, embodied, psychological agent situated in an environment. Gibson has always insisted that exteroception, the perception of one's external environment, cannot be separated from proprioception, the perception of one own's self and its relations to the environment. As various authors, from Gibson to Searle (1980 Searle, JR. 1980. Mind, brains, and programs. Behavioral and Brain Sciences, 3: 417–424. [Crossref], [Web of Science ®] , [Google Scholar], 1992 Searle JR 1992 The rediscovery of the mind Cambridge MA MIT Press [Crossref] , [Google Scholar]), have argued, one of the major problems of the cognitivist approach is that it neglects this perspectival dimension of intentional phenomena. But the problem of perspectival representation deserves a separate treatment. Intensional contexts also result from the use of modal notions such as ‘necessarily’ and ‘possibly’. For reasons of focus and manageability however, I don't pursue the investigation of the modal aspects of intensionality in the context of the present discussion.