Abstract

Naturalists about intentionality seek to find sufficient conditions, couched in a purely physicalistic, non-semantic vocabulary, for one thing to be about another. The motivation behind this enterprise is the thought that, given the materialist thesis that our world is entirely physical, it is possible for something to be about another thing only if aboutness turns out to be a perfectly ordinary species of physical relation. ;Thus the naturalist faces a serious challenge. On the one hand, the account offered must not make use of any semantic notions in the explanans; on the other, it must be sufficiently robust to do a great deal of explanatory work: It must show how representation is possible. ;I argue that it is unlikely that the naturalist will meet this challenge. Paying particular attention to the most recent theories proposed by Fodor and Dretske, I argue that naturalistic theories lose their apparent explanatory power when they are given the appropriate non-semantic reading: When the conditions the naturalist proposes for one thing to be about another are understood in the required--non-semantic--way, then the fact that a thing meets those conditions cannot be what explains how it comes to represent just what it does. The theories have an intuitive plausibility only insofar as we allow semantic notions to creep, illicitly, into our understanding of the conditions. ;The failure of naturalism may seem to force us to choose between materialism and realism about intentionality. I argue, on the contrary, that even if naturalism cannot be made to work it is still possible to believe in the reality of intentionality without giving up the idea that everything is in some sense physical. ;The argument for this claim proceeds in two stages. First, I argue that the naturalist criterion for what is to count as physical is too restrictive. The naturalist idea is that the way to show that intentionality is physical is to show that intentional facts reduce, in some sense, to sets of facts about causal relations. But if this is the criterion we use to decide whether something is physical, many paradigmatically physical things will fail to count as physical. This suggests that the criteria need to be relaxed in an appropriate way, and that, when they are, intentional facts--even unnaturalized intentional facts--will count as physical, too. This would mean that we can be both materialists and realists about intentionality, even if we give up on the naturalist goal of finding non-semantic conditions for representation. ;In the second stage of the argument, I sketch a positive account of the representational relation, an account that is plausibly consistent with materialism yet sensitive to the earlier conclusion that intentionality cannot be naturalized. The central challenge here is to find an appropriate conception of the propositional content of mental states. I defend an account according to which talk of propositions functions to specify relations in which a person stands to the world rather than to identify a special set of objects of belief.