1 I am indebted to William Lycan for letting me read a helpful unpublished manuscript on functionalism, and to Morton O. Beckner, Virgil Aldrich, Ronald Rubin, and Alise Carleton for conversation on this subject. This paper was delivered to Philosophy Collquia at the University of Calgary, University of Lethbridge, & University of Utah.

2 Lewis, David K. “An Argument for the Identity Thesis”, Journal of Philosophy, 52 (1966) 17–25.CrossRefGoogle Scholar Armstrong, David A Materialist Theory of Mind (Routledge & Kegan Paul, London, 1968).Google Scholar

3 In doing this, I will not try to argue for any general equation of mental with physical states. There are two reasons for this. The first is that there are striking differences between the items which identity theorists and their opponents classify as ‘mental’. For example, intentionality, which is commonly called the mark of the mental, does not hold in any obvious way for pains and other feelings, which are surely mental states. At the same time, privacy and privileged access, which are (rightly or wrongly) said by some to characterize mental phenomena and which (under some modest and careful considerations) are plausibly ascribed to pains, do not hold obviously for beliefs, desires, and some other mental states. Such differences and the attendant philosophical disagreements over what constitutes the mental make it hard to see Just what the general question whether the mental is to be identified with the physical amounts to. If the unclarity is as severe as I think, it is the better part of valor to attack the question of mental-physical identity piecemeal, considering interesting mental phenomena one at a time. Pain is interesting. The second reason is that in the absence of scientific knowledge philosophers discuss the identity question at their peril. It won't do to pretend that scientific findings and theories are irrelevant, and a prior predictions of how the science will turn out tend to be unreliable. More is known about the neurophysiology of pain than about many other mental phenomena. What is known seems compatible with functionalist arguments for the identification of pains with CNS states. Whether the neurophysiology of such phenomena as remembering will also turn out to be a comfort to the identity theorist is another question altogether. In view of these considerations, I am not optimistic about the prospects of generalizing from anything this paper may say that is true.

4 From lectures given by Rubin at Pitzer College.

5 Rubin's point does not depend, as it may seem to, on the theory (espoused by Ryle and many others) that the identification of physical with mental items involves a ‘category mistake’. Even if one thought there are no such things as ‘category mistakes’ or a coherent theory of them, one could still ask the functionalist why he thinks the variables in the argument take the same values, and complain that the premises of the argument are consistent with their ranging over different values.

6 Whether this detail is relevant depends upon how complicated a flow chart is given for pain. The simple chart I gave could be instantiated by a school with less fish with less tail flicks. But if the functional state were sufficiently complex (if for example, pain were held to be a complex, higher order function of states which were themselves functionally characterized) an instantiation might have to have the multiplicity of Ant.

7 If it sounds odd to say that At is in State F Just because N is in F, consider men who are said to be paralyzed in virtue of the paralysis of their limbs, and teenagers who are acned in virtue of the acne on their faces.

8 This example comes from Morton O. Beckner, who treats it differently than I do in an unpublished paper. It is related to Ned Block's example in which the population of China is made to model the CNS of a man. Block's example appears in Perception and Cognition, Minnesota Studies Philosophy of Science Series, Vol. IX, Ed. Savage, C. Wade (U. of Minnesota Press, Minneapolis, 1978), 261–326.Google Scholar

9 A man may welcome pain as a sign of restoration of feeling in his frozen toes, as an aid to his malingering, etc. These are the sorts of cases I want to ignore, and they may be ignored on the grounds that even here, the pain per se is unpleasant. In special, and well marked out cases, for purposes of certain sorts of psychological theorizing, it may be reasonable to speak of unfelt pains. (See, for example, R. Melzack's The Puzzle of Pain.). I shall ignore these as well. (Basic Books, New York, 1973).

10 On some theories, pains are neurological messages which may or may notregister consciously, and what is noxious is the registering or the resulting awareness, not the pains. Someone who holds such a theory may read the term ‘pain’ in this paper as signifying the message together with the registering or the awareness, or as signifying the registering or awareness itself. I believe that some philosophers hold that to suffer is to perform a mental act of somehow grasping a mental object called a pain. I do not subscribe to any such theory, and nothing in this paper requires one to think that what I call ‘introspectibility’ involves acts of introspecting or grasping. On other theories, to be in pain is to experience in a certain way, or to experience a state or part of one's body in a certain way. According to such theories, ‘pain’ does not stand for anything — mental or physical. This too is compatible with the version of functionalism I want to defend, and to the claim that pains are introspectibly noxious. What the functionalist identifies with a physical state, and what I claim to be noxious are states: having-pains, having-toothaches, etc., regardless of whether ‘having’ is believed to signify a relation the sufferer and something called a pain.

11 As a result of the work of Beckner, M. (in The Biological Way of Thought (U. of California Press, Berkeley, 1968))Google Scholar, Wimsat, W. in Studies in History and Philosophy of Science, 2 (1972)Google Scholar) and others, it is no longer necessary to think that the assignment of a goal to a natural system or its treatment as a teleological system commits a theorist to the belief that that system has consciousness or intentions, or that it was designed for a purpose by an intelligent designer — let alone a divine one. Nor need it be thought that because goals are assigned to natural systems from the standpoint of evolutionary or other sorts of theories that the assignment is arbitrary — that one can supply a theory under which any old goal may be assigned to Just anything, or any function can be assigned to any old part of a natural system. While theories which assign goals are invented by theorists, the facts and data they must explain are not, and these place severe constraints both upon the sorts of theories which can be reasonably maintained, and the sorts of goals and functions which can be reasonably assigned under the theories.

12 I am indebted to Joseph E. Bogen for this information.

13 That a transplant which involves no physiological, etc. changes is possible here is due to the fact that Ant was deliberately imagined to be the sort of system which allows for such transplants. It should not be inferred from this that if having a pain is identical to being in a certain CNS state, the relevant parts of the CNS could be excised without physiological changes, etc., that would make them incapable of being in the crucial physical state. Whether this is possible is an empirical question, and I know of no good reason to think its answer is ‘Yes’.

14 David K. Lewis, “An Argument for the Identity Thesis”.

15 Someone will probably want to know what is meant by claiming that I and P are identical properties. I do not think it means merely that they are coextensional, or even necessarily coextensional — that whatever has the property signified by ‘I’ does or must have the property signified by ‘P’, and vice versa. Consider the trivial properties signified by the predicates ‘flies if it flies’ and ‘swims if it swims’. Since nothing can lack these properties, whatever has one must have the other. But it does not seem to me to follow from that that they are the same property. Related examples of coextensional, but apparently nonidentical properties can be invented and multiplied until boredom and fatigue set in. Nor, does it sound promising to say that what is signified by ‘I’ and by ‘P’ is the same property Just in case the two expressions are synonymous. Aside from the difficulties with determining and explaining synonymity, this view converts an identity claim into a linguistic claim, and it would be surprising if a purely linguistic maneuver is enough to save the functionalist from quasi parallelism, epiphenomenalism, and from the brain tissue in a vat objections. I have no positive account to offer, but the following consideration may suffice for present purposes. Instead of asking how property identity is discovered, or what evidence establishes that ‘I’ and ‘P’ signify the same property, I think we should view the identity claim as a stipulation made by the functionalist as part of a commitment to a strategy for explaining the phenomenon of pain. The situation which leads to the stipulation would be roughly the following. Suppose we have a theory, T, under which P is an observable or empirically detectable property, by appeal to which, the materialist can explain how pains are caused, how they cause pain behavior, etc. Suppose that the explanations are based upon well confirmed causal principles, and that T is reasonably elegant, economical, and otherwise aesthetically satisfying. Suppose also that there is no satisfactory way of explaining the same things in terms of I as long as ‘I’ is thought to signify anything other than what is signified by ‘P’. It would then be reasonable for the functionalist to treat the predicates ‘I’ and ‘P’ as if they signified the same thing. To do this would be to commit oneself to the policy of trying to explain whatever needs to be explained concerning I in terms of theory T, and treating ‘I’ as signifying whatever T takes P to be. If this strategy fails — if some mental phenomena cannot be explained by T in terms of P, the commitment may be questioned or abandoned. If an alternative theory develops which allows for equally good or better explanations on the assumption that ‘I’ signified what is different from what T takes ‘P’ to signify, the commitment may also be abandoned. Failing this, it may reasonably be upheld and pursued. I believe that this sort of pragmatic account according to which property identity claims are to be assessed on the basis of the fruitfulness of an explanatory strategy is the best that can be hoped for at present.