Originally published in Philosophia Philosophia Vol. 18, No. 2-3, July 1988


Steven R. Bayne

Saul Kripke has devised a subtle argument against mind-body identity. Although it has been subject to considerable discussion, many of Kripke's critics, I believe, fall short in their appraisal. After stating Kripke's argument in an attenuated form I will proceed to offer another criticism; one free from the introduction of new theories or additional ontological commitments. It simply turns Kripke's argument against itself.

In his Meditations Descartes argued that the mind might exist in the absence of the body.2 He maintained that the ideas of the mind and the body are different, and that my ability to conceive the mind apart from the body shows that they are distinct substances.3 Being different substances, minds and bodies differ essentially.4 Being essentially different, it is impossible that minds should be bodies.

Modern materialists argue argue that mind and body are identical. They maintain identifying the mental with the physical requires only an identity of what is referred to. Consider 'water=H2O'. Just because the ideas of water and H2O are different is no reason, it is argued, to deny that what the ideas refer to are identical.

Kripke's argument against the materialists is directed against making such use of identity. The argument itself comes out of Kripke's theory of rigid designation - a theory our criticism will leave untouched. According to Kripke5, a rigid designator is a term which refers to the same thing in all possible worlds in which it refers at all. If 'A' and 'B' are rigid designators, then, if 'A=B' is true it is necessarily true. Consider the rigid designators, 'pain' and 'C-fiber stimulation'. An identity theorist will maintain that what they designate is one and the same thing. If the identity theorist is right, then on the theory of rigid designation it is necessary that what they designate is one and the same thing. But, it does not appear to be necessary. It appears that they might have designated different things. The identity appears tgo be contingent. If the identity theorist's argument is to work, he must explain away the appearance of contingency. In the case of 'water=H2O' this appearance can be explained. I can conceive of an "epistemic counterpart" of water, a substance possessing all the properties consistent with what I know to obtain in the case of water, which is neither water nor H2O. However, I can conceive of no such epistemic counterpart of pain. As Kripke puts it5:

It is impossible to explain the apparent possibility of C-fiber stimulation not having been pain in the same way. Here too, we would have to suppose that we could have been in the same epistemological situation, and identify something in the same way we identify pain, without its corresponding to C-fiber stimulation. But the way we identify pain is by feeling it, and if a C-fiber stimulation could have occurred without our feeling any pain, then the C-fiber stimulation would have occurred without there being any pain, contrary to the necessity of the identity. The trouble is that although 'heat' is a rigid designator, heat is picked out by the contingent property of its being felt in a certain way; pain, on the other hand, is picked out by an essential (and indeed necessary and sufficient) property. For a sensation to be felt as pain is for it to be pain.
Therefore, the identity theorist cannot answer Descartes by introducing even the appearance of a contingent identity. However, the identity theorist may have a reply.

Kripke's argument depends in an important way upon the fact that if two things are identical they are necessarily identical. Given Leibniz's Law this can be shown to be a theorem of standard logic.6 Although there is little reason to be surprised that

(A) (x)(y)(x=y L(x=y))
there is, I think, some reason to be surprised that something like
(B)(x)(y)(~(x=y) L~(x=y)
might be true. B is derivable from A, but only in S5.7 For the purpose at hand, however, this fact is not important. To see why, consider the role A plays in Kripke's argument. In an early paper on the subject by David Wiggins, the view was expressed that (A) ruled out contingent identity statements altogether.8 Kripke replied that this was merely appearance, and an extended discussion followed in an effort to explain the appearance.9 What was less controversial was the necessity of statement of the form
(A') ((A=B L(A=B))
where 'A' and 'B' are rigid designators. According to Kripke
If names are rigid designators, then there can be no question about identities being necessary...10
Similarly one might ask what doubts could arise over the truth of
(B') ((~A=B) L~(A=B))
as long as 'A' and 'B' are rigid designators. Even though B' is provable only in modal system S5, given the theory of rigid designation, it is difficult to imagine any doubt arising as to its truth. The force of B' in the argument which follows derives from the inconsistency of affirming the theory of rigid designation, while denying B'. If '~(A=B)' is true, and 'A' and 'B' rigidly designate different objects different objects, then there is no possible world in which what they designate are one and the same. If there were such a world (A) would guarantee that the contradiction '~(A=B)' and 'A=B' would follow.

When confronted with the prospect that the mental and physical are not only diverse, but necessarily diverse, the identity theorist need only imitate Kripke's approach to the idea that the mental and the physical are necessarily identical, if they are identical at all. After all, it appears that they might have been identical, even if they are not identical. The non-identity appears contingent. The identity theorist need only ask what accounts for the apparent contingency of '~(A=B)'. Kripke cannot say that we can conceive of an epistemic counterpart to pain. Kripke, then, has difficulty explaining in what sense the apparent contingency of '~(pain = C-fiber stimulation)' is illusory. We must account for the appearance of contingency just in case it turns out that pain and C-fiber stimulation are diverse. If Kripke cannot explain the appearance of contingency, then the identity theorist need only say that it is not necessary that pain differs from C-fiber stimulation. But, if it is not necessary that pain differs from C-fiber stimulation, it cannot be the case that pain differs from C-fiber stimulation.

So, by using Kripke's own line of argument, we can nullify its force. Indeed what we have resembles a paradox. If 'pain = C-fiber stimulation' then 'L(pain = C-fiber stimulation)' but the apparent contingent of 'pain = C-fiber stimulation' can't be explained, so '~(pain = C-fiber stimulation)', but, if so, 'L~(pain = C-fiber stimulation)', but since we can't account for the apparent contingency of '~(pain = C-fiber stimulation)', 'pain = C-fiber stimulation'. So if 'pain = C-fiber stimulation', '~(pain = C-fiber stimulation)', and if ~(pain = C-fiber stimulation)', 'pain = C-fiber stimulation'. I conclude Kripke's subtle argument agains mind-body identity leads to paradox, and appears, therefore, without force.

One could deny that '~(pain = C-fiber stimulation)' appears contingent but this appeal to intuition is no more convincing than the denial of apparent contingency in the case of '~(pain = C-fiber stimulation)'. Another solution leaves untouched both (B') and the theory of rigid designation, but it requires not accepting Kripke's argument in its current fomulation. The weakest premise employed in the argument is the one which says that "for a sensation to be felt as pain is for it to be pain." I see no alternative but to relinquish this plausible but troublesome notion. But in doing so pain like heat, and water, can only be picked out by a contingent property. Allowing this reintroduces contingent identity and places the mind-body dispute back in the court of scientific, rather than logical argument. Perhaps there is more than a little irony in the fact that Kripke's approach collapsed the sensation of pain into pain itself, making it more difficult to utilize the standard arguments for dualism based on intentionality. Our investigation may further encourage dualists to examine the logical features of these traditional arguments.



1. I greatly appreciate the comments of Jerome Shaffer,     John Troyer, and Samuel Wheeler, III.

2. The Philosophical Works of Descartes, trans. Haldane     and Ross, V.I. (Cambridge, 1982), p. 152.

3. Ibid. p. 190.

4. Ibid, V. II. p. 99.

5. "Identity and Necessity" by Saul Kripke, in Naming,     Necessity, and Natural Kinds, ed. Stephen P. Schwartz     (Cornell University Press, 1977), pp. 66-101.