Identity theorists claim that mental states are identical with brain states. Identity theorists usually espouse one of two versions of this claim. Either 1) tokens of types of mental states are identical with tokens of types of brain states, or 2) tokens of mental states are identical with tokens of brain states. In either case, the particular feeling I get when I stub my toe is said to be identical with some particular thing going on in my brain usually C-fiber firings. Saul Kripke argues that any such identity statement (where a mental state is said to be identical to a physical state) is false based solely on the structure of the statements. His argument shows that any statement that strictly equates mental events with physical events will be false.
In order to examine the structure of identity statements, Kripke introduces the concept of a rigid designator. A rigid designator is a term that picks out an entity based on some essential property. Since we cannot imagine an entity without its essential properties, a rigid designator will pick out the same thing regardless of how we imagine the world to have turned out (that is to say in all possible worlds). Kripke holds that names are rigid designators. An essential property for the entity called Aristotle is being-the-person-that-Aristotle-was. The name “Aristotle” is a rigid designator for an entity with that property. Aristotle could have had a different name, but he could not have been a different person. Rigid designators are contrasted by nonrigid designators, which are terms or descriptions that pick out an entity based on a nonessential property. The point is that when we say “Aristotle could have not written the Metaphysics,” we intend to describe a world in which Aristotle was still the person that he was without having been the writer of the Metaphysics (138-9, “Naming”).
Within the lectures in which Kripke introduced rigid designators, he applied the idea to the identity statements proposed by identity theorists. To be clear, Kripke meant for his argument to apply to claims made by both kinds of identity theorists (148, “Naming”). Token identity theorists feel their claims are unaffected by Kripke’s argument, but as I recount the argument, we will see that both kinds are refuted.
In the following argument, R stands for any rigid designator and S stands for the statement, “The particular feeling I get when I stub my toe is identical with some particular thing going on in my brain.” The argument goes like this (148-51, “Naming”):
1) If R1 = R2 is true, R1 = R2 will be necessarily true.
2) S is an identity statement of that form.
3) S is not necessarily true, it could be false.
Line 1 means if R1 = R2, we can never imagine a world in which R1 ≠ R2. Line 2 says if S, then we can’t imagine S to be false. Line 3 states that we can, in fact, imagine S to be false. We can imagine my particular brain state occurring without being accompanied by the particular feeling I get. We can also imagine the feeling without the brain state. Now, since S is not necessarily true, by lines 1 and 2:
4) S is not true.
The real force of the argument is found in line 1. R1 = R2 will have to be necessarily true, that is to say, true in all possible worlds, if it is true at all, because R1 and R2 designate the same thing in all possible worlds. If it is true in one world it will have to be true in all worlds. If this is at all difficult to understand, it will be because the logic is so tight; it seems like a tautology.
Token identity theorists sometimes try to dodge Kripke’s argument. John Searle attempts to show that they are not spared. He writes, “If you grant me that there really are two features to this experience, the feeling of pain and the firing of C-fibers, then it will look like Kripke’s argument will go through” (69, Mind). Before Searle writes any more, his argument has already failed; he has assumed that which he is trying to prove. The original claim was that the feeling of pain was identical to some firing of C-fibers. They claimed that there are not “two features to this experience” but one under two names. Searle should have been trying to show why Kripke’s argument still applies to token identity statements; instead, he asked us to assume with him that the statement wasn’t one of identity.
Identity theorists will not simply give up. They usually attack the third premise by saying we actually cannot imagine the sensation without the brain state. They say it is just an illusion akin to the illusion that it is possible to imagine heat not being identical to molecular movement. They say we cannot actually imagine p (the pain at time t) without b (brain state at time t).
Kripke himself was successful in maintaining his attack on token identity statements. When people think they can imagine heat without molecular movement they have only imagined something by which we identify heat (a contingent property of heat) existing without heat itself. They imagine the sensation of heat without any molecular movement. But in the case of p and b, that by which we identify p (a pain) is pain. We do not pick out p by some contingent property. Being a pain is an essential property of p (145-6, “Identity”).
Another way of identifying a pain is by its causal relations. Such is claimed by a subset of identity theorists called functionalists. They say pain is that brain state which is caused by bodily damage and which causes avoidance of bodily damage. So when people say they can imagine bodily damage without the feeling of pain, the functionalist cannot say they are only imagining a contingent property. It would be like saying that pain might not feel like pain (Kripke, 146, “Identity”). This points out the fundamental flaw of functionalism. The functionalist account leaves out the feeling of pain. If functionalists say being such-and-such-a-mental state is a contingent property of the brain state, they have simply conceded that identity statements are contingent (line 3 of the above argument).
David Lewis adopts a different strategy to avoid Kripke’s conclusion. He rejects line 2 of the above argument. He claims that “pain” is not a rigid designator. So, he must say that “pain” picks out an entity by a property it might not have had. He gives two cases where pain doesn’t have the properties traditionally thought essential (behavioristic causal relations and types of brain states). He does, in the end, give the common feature he believes makes all of these experiences to be instances of pain. He says “If it feels like pain, then it is pain” (221). As a materialist, he should give a physical account of what a feeling is. But, for the purpose of this debate, he has given that essential property by which the term “pain” picks out instances of pain. If faced with this description, he should admit, “pain” is a rigid designator.
Kripke’s argument simply establishes that mental states are not identical to brain states. It still is possible that they be correlated, maybe even concomitant phenomena. Materialists do not like this because they want to explain the mind with only reference to physical facts. As Searle points out, “[this argument] is essentially the commonsense objection in a sophisticated guise” (39, Rediscovery). The commonsense objection is that pains and brain processes are simply two different kinds of things.
Kripke, Saul. “Identity & Individuation” Identity & Individuation. Munitz, Milton, ed. New York: NYUP, 1971.
—. “Naming & Necessity” Nature of Mind. New York: Oxford University Press, 1991, 138-158.
Lewis, David. “Mad Pain and Martian Pain” Readings in Philosophy of Psychology. Block, Ned, ed. 1980.
McGinn, Colin. “Anomalous Monism and Kripke’s Cartesian Intuitions” Analysis 37, 2 (1977): 78-80.
Searle, John. Mind: A Brief Introduction. New York: Oxford University Press, 2004.
—. Rediscovery of the Mind. Cambridge: MIT Press, 1994.