22. The So-Called Identity Theory

I always try to avoid discussing the meaning of words, and as a consequence, I also try to avoid criticizing a theory for using the wrong words or words with a wrong meaning or with no meaning at all. My standard policy in these cases is to sec whether the theory under discussion cannot be so reformulated or interpreted that objections based on the meaning of words disappear.

This applies to the identity theory. (The identity theory often occurs linked with panpsychism, for example in its founder, Spinoza, or in our own time in the work of Rensch. However, it needs to be clearly distinguished from panpsychism as a theory.) I very much doubt whether a formulation like "mental processes are identical with a certain kind of (physico-chemical) brain processes" can be taken at its face value, in view of the fact that we understand, since Leibniz, "a is identical with b" to imply that any property of the object a is also a property of the object b. Some identity theorists certainly seem to assert identity in this sense; but it seems to me more than doubtful whether they really can mean it. (Two very powerful though very different criticisms of the identity claim in this sense can be found in Judith Jarvis Thomson [1969] and in Saul A. Kripke [1971]; each of them appears to me pretty conclusive.) In view of this situation I will adopt here the following policy. I will criticize the identity theory by criticizing a logically weaker consequence of it, the Spinozistic theory that mental processes are physical processes experienced "from the inside"; that is, I shall criticize a form of parallelism. (A parallelist theory is weaker than an identity theory because the identity of two lines or two surfaces is a limiting case of their being parallel: they are parallel with the distance zero.) In this way I can avoid criticizing the identity claim and still criticize the identity theory, and some weaker theories at the same time. Moreover, in adopting this policy, I am not prevented from presenting the identity theory in as reasonable and convincing a form as I can.

The identity theory in some of its versions is very old. It is reformulated in Diogenes of Appolonia (DK B5). Democritus no doubt regarded psychical processes as identical with atomic processes, and Epicurus (Letter I, to Herodotus, 63ff.) indicates clearly that he regards sensations and passions (or feelings) as mental or psychical, and the soul or mind as a body of fine particles; and these ideas are no doubt even older. Descartes stresses the different character of the mental (unextended; intensive) and the physical (extended); but the Cartesian Spinoza stresses that "the order and the connection of [mental] ideas is the same as [or identical with] the order and the connection of [physical] things" (Ethics Part II, proposition VII; Part V, proposition I, demonstration); and he explains this by the theory that mind and matter are two different ways of comprehending, or aspects of, one and the same substance (or thing in itself) which he also called "Nature" or "God". This theory — a parallelism between mind and matter, explained by their being two aspects of a thing in itself — is, I suppose, the beginning of the modern physicalistic identity theory which replaces "Nature" by either "mental process" or by "physical process" and which restricts the identity thesis to a small subclass of material processes: to a subclass of the brain processes, which it identifies with mental processes.

It is interesting that the Spinozistic theory of two aspects was often described as an identity theory. Thus the great nineteenth century neurologist John Hughlings Jackson ([1887] = [1931], volume ii, p. 84) distinguished the following three doctrines of the relation between consciousness and "the highest nervous centres" of the nervous system. [The remarks in square brackets have been added by me.]

(1) The "mind acts through the nervous system". [Interactionism.]

(2) The "activities of the highest centres and mental states are one and the same thing, or are different sides of one and the same thing". [Identity theory; and Spinozism.]

(3) The two things, though "utterly different", "occur together ... in parallelism", there being "no interference of one with the other". [Parallelism.]

It is clear that (2) comprises the identity theory and Spinozism, while (3) distinguishes a non-Spinozistic (Leibnizean?) parallelism from (2). (Jackson himself opted for (3).) I too shall regard here the identity theory as a more radical form of Spinozistic parallelism.

The theory called by Herbert Feigl "the psychophysical identity theory" aims at avoiding the implausibilities and the difficulties of epiphenomenalism. It does so by stressing that the mental phenomena — or the mental processes — are real. (Feigl, following Schlick, goes so far as to say that they are the real things or, in Kantian terminology, the things in themselves.²²) Thus mental processes do not play here the objectionable role of redundant epiphenomena. They are conjectured, however, to be "identical" with a certain subclass of the physical processes that occur in our brains. This is the central conjecture of the theory. It does not mean that the mental experiences or processes as known by acquaintance must be logically identical with physical processes as described by physical theory. On the contrary, as Schlick emphasizes, the mental processes of which we have a knowledge by acquaintance are conjectured, according to his theory, to be "identical" with a kind of physical process of which we may obtain only a knowledge by description; "identical" in the sense that the objects the brain physiologist tries to describe in theoretical terms turn out empirically to be, in part, our subjective experiences. This knowledge is theoretical knowledge (and thus, incidentally, conjectural knowledge). Or as Feigl likes to put it; the mental processes of which we have knowledge by acquaintance turn out, if we want to obtain of them knowledge by description, to be physical brain processes. Thus according to Feigl's theory a type of mental process may, regarded as a type of brain process, consist of the fact that a sufficiently large number of neurons are all doing microchemically the same thing — say, synthesizing certain transmitter molecules in a peculiar rhythm.

The identity theory (or the "central state theory") can be formulated thus. Let us call "World 1" the class of processes in the physical world. Then let World 1 (or the class of objects belonging to it) be divided (as in section 16) into two exclusive subworlds or subclasses, in such a manner that World 1_m (m for mental) consists of the description in physical terms of the class of all the mental or psychological processes that will ever be known by acquaintance, while the vastly larger class, World 1_p (p for purely physical) consists of all those physical processes (described in physical terms) which are not mental processes as well.

In other words, we have

(1) World 1 = World 1_p + World 1_m

(2) World 1_p . World 1_m = 0 (that is, the two classes are exclusive of each other)

(3) World 1_m = World 2

The identity theory stresses the following points:

(4) Since World 1_p and World l_m are parts of the same World 1, there is no problem raised by their interacting. They can clearly interact according to the laws of physics.

(5) Since World 1_m = World 2, mental processes are real. They interact with World processes, exactly as interactionism asserts. So we have interactionism (without tears).

(6) Accordingly, World 2 is not epiphenomenal, but real (also in the sense of section 4, above). Therefore the clash between the Darwinian point of view and the epiphenomenal view of World 2 described in section 20 does not occur (or so it might seem - but see the next section).

(7) The "identity" of World 1_m and World 2 can be made intuitively acceptable by considering a cloud. It consists, physically speaking, of an accumulation of water vapour, that is, a region of physical space in which water drops of a certain average size are distributed with a certain density. This is a physical structure. It looks from the outside like a white reflecting surface; it is experienced, from the inside, as a dull, only partially translucent, fog. The thing as experienced is, in theoretical or physical description, identical with a structure of water drops.

According to U.T. Place [1956], we can compare the inside view and the outside view of the cloud with the inside or subjective experience of a brain process and the outside observation of the brain. Moreover, the theoretical description in terms of water vapour, or of a structure of water drops, can be compared to the not yet fully known theoretical physical description of the relevant physico-chemical brain processes involved.

(8) If we say that fog was the cause of a car accident, then this can be analysed, in physical terms, by pointing out how the water drops absorbed light, so that light quanta which otherwise would have stimulated the driver's retina never reached the retina.

(9) The upholders of the central state theory or identity theory point out that the fate of the theory will depend on empirical corroboration which can be expected to come from the progress of brain research.

I have presented what I regard as the essentials of the theory. The following points I regard as inessential.

(a) Herbert Feigl's suggestion ([1967], p. 22) that the theory does not assume the hypothesis of emergent evolution. (This is even a crucially important point for Smart.) I think that the theory does assume it: there was no World 1_m before it emerged from World l_p. Nor could its peculiar mental properties be predicted. However, I regard this emergent character of World 1_m as perfectly in order, and not as a point of weakness of the theory.²³

(b) It will be remembered that in my presentation of the theory I have tried to avoid any purely verbal argument connected with the term "identical" or with the question what it may mean to say that mental or experienced processes (World 2 = World l_m) are "identical" with objects of our physical descriptions. This "identity" certainly has its difficulties. But in my view it need not be taken as crucial for the theory or for some version of it; just as it may not be essential for our metaphor of the cloud to decide in which sense the three aspects — the outside view, the inside view, and the description in physical terms — are all aspects of one and the same object. What I regard as crucial is that the identity theory adheres to the physicalist principle of the closedness of World 1. Thus in my view a theory which gives up the term "identity" and replaces it by "very close association" (say) would be just as mistaken if it adhered to this physicalist principle.

(c) Feigl rightly stresses the "reality" of the mental processes, and this point seems to me essential. But he also stresses the character of the mental processes as things in themselves. This seems to me to make him a spiritualist rather than a physicalist, but it invites discussions which may easily become quite verbal. Take again our cloud metaphor. It seems to me (but I should be loth to argue the point) that the physical description — that of the cloud as a space in which water drops of a certain kind are distributed — comes in a way perhaps nearer to describing the thing in itself than either the outside description as a cloud or as a bulky surface reflecting light, or the inside experience as a fog. But does it matter? What does matter is that all the descriptions are descriptions of the same real thing — a thing that can interact with a physical body (for example, by condensing on it and thus making it wet).

I think that we may assume that there is no problem here; we can still criticize the theory on non-verbal grounds. In the next section I will criticize the identity theory as a physicalist theory. In a later section (54) I will point out that as a spiritualist theory, bordering on panpsychism, it accords badly with modern cosmology.

²²  Feigl ([1967], pp. 84, 86, 90.) In a footnote on p. 84 Feigl refers to Kant's Critique of Pure Reason (see Kritik der reinen Vernunft, first edition, p. 361; transcendentale Dialektik, zweites Buch, dritter Paralogism = Kants Werke, Akademieausgabe, Band 4, 1911, p. 227; Cassirer's edition, Band 3, 1913, p. 643), where indeed the theory is mentioned that the thing in itself may be of a mindlike character. Thus we get the following parentage of this form of the identity theory: Kant — Schopenhauer (the thing in itself = will) — Clifford (whose identity theory is a kind of parallelism) — Schlick — Feigl — Russell (Russell's "Mind and Matter" [1956] is discussed in H. Feigl and A. E. Blumberg [1974], pp. XXII ff., and Feigl [1975]). For Clifford, see note 4 to section 16, above; for some additional remarks on the history of the identity theory, see section 54 below.

²³  The point is far from crucial; but it is not merely verbal. Smart, more especially, has a different attitude towards scientific knowledge from mine: while I am impressed by our immense ignorance on all levels, he holds that we can assert that our knowledge of physics will one day suffice to explain everything — even (to quote Peter Medawar) our foreign exchange deficit; see section 7 above.