Some Considerations Regarding Identity

To frame what the identity theory is up against and, and prime for the Kripke thing, I’ll look in a little more detail here at what it would be for any one thing to be identical to another. Here is a plausible and more or less typical account (let’s call it L-identity, because I like nicknames and I think it’s Leibniz’s) of what it is for two things to be identical (tip your 40 to Leibniz):

(x)(y)(x = y → (Fx iff Fy))

That is, take any thing x and anything y: x is y only if, for any property F, if x is F then y is F (and vice versa). In other words, finally, two things are identical if and only if they share all properties in common. For our purposes, L-identity dictates that:

Consciousness is a brain process only if, for any property F, if consciousness is/has property F then some brain process is/has property F (and vice versa).

This seems to be enough to support the claim that strict identity clearly fails on this widely accepted criterion for identity. The non-identity theorist needs only to point out that conscious states like pain, pleasure, joy, and sorrow apparently have properties that whatever brain process(es) are correlated with them do not. Presumably these would be the so-called phenomenal properties. Neuro-cellular events are not, in and of themselves, joyous or painful unless a subject is consciously experiencing the states.

Of course, whether or not types or tokens of such states always underwrite the qualitative states is not especially important, because correlation is no criterion for identity. Smoking is correlated with lung-cancer, but they aren’t the same thing. Even perfect correlation doesn’t necessarily implicate an identity relation or even a causal relation. The point generalizes to cases of perceptual states like visual representations; for instance, there are no black or hot pink neural states, though particular neural states presumably underwrite the experience of color. By all counts, strict material identity fails if L-identity is our criterion.

But L-identity is not what Place appears to be talking about, and the ‘is’ of composition seems like it has the potential to do a lot of work here: ‘lightning is the motion of electric charges’ is an identity statement, but one with merely contingently identical relata, both of which are incapable of being simulaneously apprehended by a subject. We might ask whether the two relata share all properties in common, as is necessary if we accept L-identity, but this would miss the point that some phenomena are apparently identical in a different sense, though they may have radically different descriptions and radically disparate verification conditions.

So, Place leans on the following: (1) that whether or not the application of LIGHTNING and MOTION OF ELECTRIC CHARGES are appropriate in a particular case is an empirical question, depending on independently verifiable objects, and (2) to doubt that identical objects can have different verification conditions is to fall victim to the Phenomenological Fallacy (PF) which I’ll repeat here:

When a subject s describes his ‘introspective’ observations of so-called qualitative states (i.e. his experience of looks, smells, sounds, feels, and other seemings), s describes literal properties of objects and events experienced in the ‘phenomenal field’.

He explains PF in the context of after-images in a phenomenal field, but we can generalize to the experience of lightning. The lightning case, too, seems to fail L-identity, but this only seems important if we place too much stock in the phenomenological ‘data’. It may be that the point here is just the relatively simple one that one and the same event/object can have different descriptions. Do the motion of electric charges and lightning share all properties in common? Not phenomenologically, but if it is a true identity statement by composition, then yes, they do; we just can't from the perspective of a single subject, verify the identity simultaneously. All this is of course consistent with the fact that a hallucination of a bolt of lightning, or a dreamt experience of lightning is not identical with some instance of electrical charges in motion, since it does not pick out something so composed. It’s just that we’re in a brain state similar to that which are in when we experience real lightning. But to make the inferential leap from the conceptual independence of LIGHTNING and MOTION OF ELECTRIC CHARGES, together with their phenomenological independence, to non-identity is a species of PF, and is not clearly warranted at all. Moreover, it appears to be a contingent fact that the relata are identical, and that Place was right to insist that the claim ‘Consciousness is a Brain Process’ is a scientific hypothesis on par with ‘Lightning is the Motion of Electric Charges’.

What we might try to do is to qualify L-identity with a supervenience clause or something, so as to accommodate the ‘is’ of composition. This'll be messy, but, take anything x and anything y:

x is y only if, for any property F, if x is F then y is F (and vice versa), and bona fide occurrences of x are perfectly correlated with y such that if F is a directly observable phenomenal property, then F supervenes on G (a predicate describing the composition of x).

It’s sloppy and I won’t even bother trying to formalize it. But you can see what I’m thinking, F might have multiple descriptions, each depending on the level of material composition one is addressing. As Place notes, an accurate description of composition should not be ruled out as a candidate for an identity statement just because different properties are involved. Of course, this is not super-satisfactory with regard to other intuitions about identity. We still might be wont to say, motivated by L-identity, that we need more than descriptions that appear to be related, we need something like L-identity to guide us, otherwise we’re just begging the question as to whether two relata pick out the same phenomena.