Against the Identity Theory of Physical Objects
The identity theory of physical objects says that a perceived object is identical to an object described in physics (or perhaps in physiology). For example, the table before me is identical to an object described in physics as consisting of a collection of atoms. Logically, it is like the identity theory of mental objects: a pain, say, is identical to a brain state as described in neurophysiology. In other terminology, both are reducible to objects described in physical science—reducible via identity. The identity might be of types or tokens or both. A dualist theory will maintain that there is no such identity; rather, the objects in question are numerically distinct. I am about to argue for a dualist theory of (what we call) physical objects—tables, rocks, animal bodies, etc. I will put this for simplicity as the theory that perceived objects are not identical to physical objects (thus restricting the term “physical” to objects-as-described-in-physics). Physical objects (vernacular) are not identical to physical objects (scientific).
I will start with a thought experiment to get the intuitive juices flowing. Suppose I come into possession of a medium-sized dog; call him Jack. Jack is a perfectly ordinary dog. However, due to some quirk in my visual system, I don’t see Jack as a dog at all; I see him as a cat. I think I have a cat as a pet; I call it Wendy. Suppose I never discover my error: I think my visual cat-representing percept corresponds to an actual cat. The same goes for my other dog-caused cat-like percepts, tactile and auditory. Actually, Jack has none of the properties my percepts attribute to him (or just a few of a general nature). Are Jack and Wendy identical? Wendy is a cat, but Jack is not; and Wendy looks exactly like a cat, which Jack doesn’t. In philosophical jargon, the intentional object of my percepts is completely different qualitatively from the object occupying my house. I think the intuitive answer is that they are not numerically identical. Jack is simply the causal trigger of my Wendy percepts. We would no more say that Wendy is identical to Jack than that she is identical to the brain states that trigger my percepts of her. Does Wendy exist? It doesn’t matter to the thought experiment whether she does or not, but I am inclined to say she does; in any case, I will occasionally talk this way to facilitate discussion. So, we can say that the perceived object is not identical to the physical object, though there is a certain correspondence between them. Leibniz’s law would appear to back me up, because Jack doesn’t have cat properties and Wendy doesn’t have dog properties. An identity theory here would be false. Intuitively, my mind has created Wendy and she is not identical to the dog that elicits cat impressions in my deranged nervous system.
This case may remind you of two well-known philosophical examples: the Muller-Lyer illusion and Eddington’s two tables. On the face of it, I see two lines of unequal length and a solid object, respectively. But the two lines in the external stimulus are equal in length and the table is really not solid (according to Eddington). We are under an illusion in both cases. The perceived object and the physical object do not coincide in their properties; therefore, they are not identical. Thus, Eddington speaks of two tables (he is a table dualist) and a perceptual psychologist might follow suit with regard to the visual illusion (he is a geometry dualist). Visual objects are not physical objects: the objects of perception are not identical to the objects of physics. The same kind of reasoning can be applied more generally: visual objects are perceived as colored and are colored, but physical objects are not; therefore, they are not identical. Ditto for other secondary qualities. The two things do not have all properties in common, so they cannot be identical. This conclusion is neutral on the question of whether we see physical objects, directly or indirectly; the point is that even if we do the two are not identical. I see colored objects, but the objects of physics are not colored; so, the two are not identical.
We can pursue the parallel between the two sorts of identity theory further. First, the type-token distinction: we can ask whether being a red object is identical to being an object and such and such a physical type, and we can ask whether the identity only extends as far as token red objects (so that redness may have a different basis in different tokens). Second, we can imagine cases of red objects in a possible world that fail to be correlated with the physical objects they are correlated with in the actual world, thus undermining identity (Kripke cases). And we can conceive of zombie-type cases that involve the same physical objects but no corresponding perceptual objects (non-supervenience). Also, knowledge arguments to the effect that knowledge of the physical object never adds up to knowledge of the perceptual object (in respect of color). The same dialectic applies in the two cases. You can pull apart the perceptual object and the physical object, so no necessity binds them, as required by identity. Granted, then, that both identity claims are demonstrably false, it turns out that neither mental states nor perceptual objects are physical entities. This means that the paradigms for the classic identity theory are incorrectly described: the perceptual object we call “heat” is not identical to molecular motion, because the same such perceptual object can be elicited by different physical phenomena. In the case of a hallucination, impressions of heat are present but no physical correlate is—we have heat, the perceptual object, but no molecular motion to go with it. Similarly, for water: same perceptual object but correlated either with H2O or XYZ. Or the same physical thing could cause perceptions of different intentional objects. The objects of perception are never identical with the objects described by physics. So, physicalism fails for both perceptual objects and perceptual states themselves. There is obviously a pattern here—a systematic failure of physicalist reduction.
This is particularly clear for sounds. Suppose I hear the sound of a bell as a piano sound: can the heard sound be identified with the physical stimulus? No, because the heard piano sound is not necessarily correlated with the air perturbations caused by the bell being struck. That perceptual object cannot be identical with the physical stimulus, though the two are conjoined in this instance. The connection is contingent, so not the relation of identity. It is even possible to have the perceptual object in the complete absence of a physical stimulus. There are two things here: the external physical stimulus and the internally generated intentional object. It is the same with the body: the body you perceive is not the same as the body studied by physical science. You could hallucinate a body, in which case there would be no physical body to be identical with the body hallucinated. There are really two bodies here—the perceptual body and the physical body.
This solves a problem that has long puzzled me—the puzzle of the location of pain. We are often told that a pain apparently in the toe is not really in the toe but in the brain, that being the place where pain is processed. But now we can see that the pain might be in the perceptual body’s toe but not in the physical body’s toe. You perceive pain in your toe qua part of the perceived body, but in the physical body it occurs in the brain and only seems to occur in the toe. Seeming rules in the case of the perceived body, but not the physical body. You are not really under any illusion about the pain’s location; we just need to distinguish the two bodies. In the Muller-Lyer illusion the two lines are of different lengths in the perceptual object, though not in the physical object. The mistake is to confuse the two objects; visual illusion occurs when the two objects diverge in their properties. There is nothing visually false in the case of so-called visual illusion, though you may make a false judgment regarding the physical stimulus. Certainly, these cases are conceptually intricate, but it isn’t that the physical stimulus is being seen as otherwise than it is; what is seen is the intentional object, and it is as it is seen. We make inferences from one to the other, which may lead us to error, but the visual system itself is not committing any errors—it is correctly representing the perceptual object. You saw two unequal lines, and those lines compose the immediate perceptual object. Likewise, the pain is where it seems in the perceived body, though not in the physical body. In any case, the two bodies are not identical.[1]
[1] I have avoided the term “sense-datum” here, because of its ambiguity and general unclarity. I prefer to use “intentional object”, but this is a technical term and carries unwanted connotations. I have therefore used “perceptual object” throughout. The whole topic is plagued with terminological pitfalls. It really is amazing how hard it is to describe ordinary visual experience, given its immediacy. You feel like you are walking on eggshells.
