Kripke's Argument

Kripke's Modal Argument is a distant and far more sophisticated cousin of Descartes' arguments for Cartesian Dualism (Interactionism).

Recall that we rejected this argument on the grounds that it was an illegitimate application of the corollary to the weak form of Leibniz' Law,

For any X and any Y, IF, for some property F, X has F but Y does not have F, THEN X is not identical to Y,

because my beliefs about Spiderman's superheroism and Peter Parker's cowardice surely do not warrant the conclusion that Spiderman is not Peter Parker. The attitudes I might have towards objects don't seem to be the right kinds of properties for application of Leibniz' Law or any of its corollaries.

Kripke's Modal Argument escapes this criticism. It is important to recognize that the text from which this argument has been drawn is a small portion of a much larger and altogether different project. That is, Naming and Necessity is not fundamentally a text in the philosophy of mind or artificial intelligence. Rather, it is a text in the philosophy of language, and as such we are missing much of the context that would make this a more accessible reading than it is.

In Naming and Necessity, Kripke articulates and defends the view a term refers directly to an object without any mediation by the term's connotation, associated ideas, or Fregean sense. According to the theory of Direct Reference, terms directly pick out or refer to the objects they denote solely in virtue of their causal history in a community of language users.

An important consequence of this view is that any true identity is necessary. We can state this more formally as the principle that,

For any X and any Y, if X = Y, then necessarily X = Y.

If we understand modal terms like 'necessarily' and 'possibly' using Possible Worlds Semantics, then the principle becomes

For any X and any Y, if X = Y at the actual world, then at every possible world, X = Y.

On the face of it this seems absurd, for what it says is that since it turns out at the actual world that Hesperus = Phosphorus, then there is no possible world at which Hesperus ≠ Phosphorus. Yet it seems conceivable, and thus perfectly possible, that Hesperus ≠ Phosphorus. After all, the term 'Hesperus' refers to the Morning Star and the term 'Phosphorus' refers to the Evening Star, yet we were surprised to find that the Morning Star just is the Evening Star. Surely, we want to say, it could have turned out otherwise?

Notice, however, that if Hesperus directly refers to the planet Venus and Phosphorus likewise directly refers to the planet Venus and do so at every possible world, then our suspician that, possibly, Hesperus ≠ Phosphorus is itself nothing more than the clearly absurd suspician that Venus ≠ Venus, a claim no one would, or could, make.

Consider another example. It turns out that heat = the motion of molecules at the actual world. By the principle of the necessity of true identities, it follows that heat = the motion of molecules at every possible world. Why couldn't heat have been anything else, though? Our strong suspician that it could have been, Kripke notes, is a confusion on our part: We only make this claim because we confuse heat with the appearance of heat or the feeling of heat. How heat feels to us surely could have been caused by something other than the motion of molecules, but heat itself just is the motion of molecules.

One more example. According to either type or token physicalists, every mental state of a certain kind is a brain state of some kind (type-identity), or every mental state is a brain state (token-identity). Revisit our old staple,

1. Pain = C-fibre nerve firings.

By the principle of the necessity of true identities, it follows that

2. Necessarily, pain = c-fibre nerve firings (or, pain = c-fibre nerve firings at every possible world).

if (1) is true. Spelled out more fully,

3. If pain = c-fibre nerve firings at the actual world, then at every possible world pain = c-fibre nerve firings.

Yet, Kripke points out,

4. Conceivably, pain ≠ c-fibre nerve firings.

Nor need we go very far for justification of (4)--viz., phantom limb syndrome. Yet if (4) is true, then

5. Possibly, pain ≠ c-fibre nerve firings (or, there is a possible world at which pain ≠ c-fibre nerve firings).

Why isn't pain like heat, though? Couldn't it just be that we're fooling ourselves like we did in the case of heat, when we confused heat itself with the feeling of heat? Here is the lynchpin of Kripke's Argument:

6. The feeling of pain = pain.

Unlike heat and the feeling of heat, pain just is the feeling of pain. There is no difference between having a pain and having the feeling of having a pain. So (6) must be true: we can't confuse pain with the feeling of pain in the same way we can confuse heath with the feeling of heat.

The resulting argument is stunning. For if there is a possible world at which pain ≠ c-fibre nerve firings, then it follows that

7. It is not the case that at every possible world pain = c-fibre nerve firings.

So by (3) above we conclude,

8. It is not the case that pain = c-fibre nerve firings at the actual world,

or, put more simply,

8'. Pain ≠ c-fibre nerve firings.

Hence physicalism, whether type or token, is false: Pain states (and, presumably, conscious states in general) are not physical states.