The Inadequacy of Materialistic Explanation A Review of Joseph Levine's Purple Haze

Purple Haze: The Puzzle of Consciousness, by Joseph Levine, is reviewed. The position that Levine takes in the current philosophical debate about consciousness is identified and the general approach of the essay outlined. I focus on two of the more important issues in the book the conceivability argument against materialism, and the explanatory gap argument against dualism and argue that Levine's argument against the former is unconvincing and his diagnosis of the source of the latter leads him into problems. I suggest a more promising route. One of the many problems which the existence of phenomenal consciousness poses is the task of explaining just how it arises (if it in fact does) from the biological and thus physical/functional systems from which we are composed. The idea that there is an epistemic divide between the two kinds of phenomena was first discussed by Levine in the early eighties (Levine, 1983) and is now one of the central issues in the philosophy of mind. Here, in his first monograph, he presents, develops, and defends this impervious puzzle in a succinct, thoroughgoing, yet comprehensive survey of the field of which it is part, illustrating convincingly why the mind-body problem remains just that. Levine takes a middle ground position between materialism and dualism, holding that mental properties are realised by physical properties, whilst acknowledging the fact that we have no clear idea of how this is so. The aim of the essay is not to provide a positive solution to the problem, but to show just why it is a problem, why materialist theories of phenomenal consciousness are inadequate, while defending materialism from antimaterialist arguments. He covers a lot of ground, discussing such wide ranging views as functionalism, panpsychism, representationalism, higher order theory, property dualism, and eliminativism (each chapter taking its title from the lyric of a Jimi Hendrix song). The first task is to give an account of materialism, which he defines negatively as the thesis that only nonmental properties are instantiated in a basic way, all mental properties being realised by the instantiation of these basic ones. This would, ofcourse, be vacuous without an account of what it is for a property to be mental, so Levine offers the following fairly standard categories: those properties which are intentional or are directed towards the world (and which can be subject to intelligent manipulation), and those which are phenomenally conscious (are "bits of awareness" or qualia). The motivation for this materialist approach is the commonplace view that mental properties interact causally with physical properties and vice versa, and so if mental properties are part of the causal order, they must themselves be or be realised by physical properties. This vague ontological picture doesn't really set Levine apart from most of his contemporaries, it is only when the question of explanatory adequacy is considered that his views become interesting. He claims that "while we seem to have some idea how physical objects, or systems, obeying physical laws, could instantiate rational and intentional properties, we have no idea. how a physical object could instantiate a subject of experience, enjoying, not merely instantiating, states with all sorts of qualitative character" (p.76). To demonstrate his point, he devotes a chapter to exposing the inadequacies of most of the more recent reductive accounts of conscious experience, and another to defending the reality of qualia from the ostrich-head-in-the-sand option of denying that we actually enjoy any phenomenal experiences at all. Though some of the more problematic scenarios for a qualia realist which have been outlined by Dennett (1990) are not discussed, both chapters forcefully make out the case for the idea that there is something missing from our physicalistic worldview. The meat of the book though, is his defence of materialism from the conceivability argument, and his general argument for the existence of the explanatory gap. The main obstacle to a purely materialistic metaphysic, he tells us, is the logical possibility of zombies. From a complete physical/functional description of a creature that is conscious (which picks out the lower order properties), we cannot derive by a priori means only that that creature has conscious experiences (the higher order properties), so it is conceivable (conceptually possible) that the physical/functional mechanisms responsible for the nature and existence of conscious experiences and the experiences themselves could exist independently of one another. If this is conceptually possible, then there is a possible world in which this situation is a metaphysical reality, but then materialism must be false for materialism is the thesis that all properties are or are realised by non-mental properties. An obvious response to this is to point out that standard cases of necessary a posteriori identities such as "water = H20", fall foul of the same argument, since it is conceptually possible that H20 could have none of the higher order properties that it in fact has (such as liquidity, transparency, etc.) Because this is a metaphysical impossibility (for H20 cannot fail to have the higher order properties that it in fact has) there must be something wrong with the argument, and so what must really be going on in this case and the mental-physical case is that there are two different modes of presentation two different concepts which pick out the same situation. However, this response cannot account for the fact that in the mental-physical case, we don't seem to have two different concepts picking out the same situation, but, rather, two distinct concepts picking out two distinct properties a physical one and an experiential one and no matter how we try to analyse an identification of the two, we will always be left with what Smart had called "an irreducibly psychical property" (Levine also considers arguments from Kripke and Chalmers but the main point is the same). The issue can be resolved, thinks Levine, by focusing on a general question in the theory of meaning whether or not we have a priori access to enough information to determine the referent of a term in a possible world considered as actual. If we think that we have, then we are what Levine calls "ascriptivists" about that mode of presentation. When we use a term to refer to something, we have in mind, either implicitly or explicitly, some description that enables us to pick out that something in a given possible world. The term "water", for example, always picks out the substance that has the higher order properties of actual water, since it is part of the meaning of "water" that it is liquid, transparent, etc. In other words, "water" always refers to "watery" stuff. However, if we think that we have very little or no conceptual content in mind when using such referring terms, we are "non-ascriptivists" about the mode of presentation. All there is to the correct application of a term is the appropriate relation of the symbol to its referent in the actual world, so "water" always refers to H20 in any given possible world. Using this latter approach, Levine can explain why we can't derive a priori the higher order properties of something from its lower order properties (or microproperties) why we can't derive liquidity and transparency from the microphysics of H20. We can conceive of H20 with few or even none of the higher order properties that it in fact has, since it is conceptually possible that H20 be opaque in normal conditions, or that it not be capable of being in a liquid state. This is conceivable because when we refer to H20 with the term "water", it is not essential that we have any of the properties that we usually associate with this term in mind. But though such a situation is conceptually possible, it does not follow that it is metaphysically possible, for, as already noted, H20 cannot fail to have the higher order properties that it in fact has (given that the rest of physics and chemistry remains the same). The same considerations apply to the mental-physical case. Even if we had a complete physical/functional theory of consciousness, we could not infer the nature or existence of the experiential properties which the physical/functional mechanisms involved give rise to because of the non-ascriptive character of that mode of presentation. But to infer from this that the physical and phenomenal are not identical is unwarranted, and so the conceivability argument is no threat to materialism. An epistemic problem remains though. Nothing we can say about the physical/functional basis of consciousness makes it fully intelligible why conscious states have the particular nature that they have, or why, indeed, there should be any at all. Though the higher order properties of water can't be derived from the microphysical properties a priori, once all the relevant empirical information is complete (physics and chemistry), there is no sense left in wondering how H20 could have the higher order properties that it has. This is not so in the mental-physical case, for even when we have all the relevant empirical information about the physical/functional basis of phenomenal consciousness, there still seems to be "genuine, substantive cognitive significance" (p.83) left to the question of just how such properties could give rise to phenomenal experiences or qualia. This is due to the fact that our concepts of qualia are "presentationally thick" they serve as their own modes of presentation (are "substantive") and present themselves as having a specific quality (are "determinate"). This is in contrast to the "presentationally thin" conceptions we have of other properties or substances such as water, in which the referring term involved acts as no more than a label for its referent. But now this motivates a second conceivability argument. The existence of zombies now seems possible due to the very fact that there is this kind of substantive and determinate qualitative residue left which is unexplained by our physical/functional theories, and so we can conceive of the existence of a creature which is physically/functionally identical to a conscious one, but which does not instantiate this qualitative residue. Levine's response is a desperate one. He simply denies that this "gappiness" must be explained by a distinction in properties, telling us that the assumption that it must is based on a kind of Cartesian model of access to the facts, and that the possibility that distinct concepts can refer to the same thing must always remain a live option. I agree, but then Levine's whole argument seems to be, in the end, little more than a straight denial that distinct concepts must pick out distinct properties, but that is the very issue which the distinct property