Here, in Chalmers' words, we can see his claim to a form of dualism--that consciousness is in some sense "beyond" the physical:
"Consciousness is a feature of the world over and above the physical features of the world. This is not to say it is a separate 'substance'; the issue of what it would take to constitute a dualism of substances seems quite unclear to me. All we know is that there are properties of individuals in this world--the phenomenal properties--that are ontologically independent of physical properties" (p.125, Chalmers' italics).This next quote illustrates how Chalmers' view implies that consciousness is causally irrelevant to the physical world--e.g. that consciousness is causally irrelevant to human behavior:
"A problem with the view I have advocated is that if consciousness is merely naturally supervenient on the physical, then it seems to lack causal efficacy. The physical world is more or less causally closed, in that for any given physical event, it seems that there is a physical explanation.... This implies that there is no room for a nonphysical consciousness to do any independent causal work (p.150).My interest in this essay is to argue against Chalmers' version of epiphenomenalism. (Note, Chalmers agrees that his view is very similar to epiphenomenalism, but does not outright endorse the term; however, for the sake of brevity, I will loosely refer to his position as epiphenomenalism.) Although Chalmers spends a good deal of time arguing against materialist views of consciousness, like logical supervenience and reductivist approaches, that part of the book strikes me as decisive so I have no criticisms to offer on that front: I agree with Chalmers that materialist accounts of consciousness are fatally flawed. My disagreement with Chalmers centers on his rejection of substance dualism in favor of epiphenomenalistic dualism. Specifically, I will argue that epiphenomenalism is problematic for two reasons: 1) it is conceptually flawed, and 2) it denies consciousness causal relevance.
Objection 1: Conceptual Flaws with Epiphenomenalism
Chalmers argues that consciousness is a non-physical property or feature of the world, not a kind of substance. Chalmers describes his view as "a kind of property dualism: conscious experience involves properties of an individual that are not entailed by the physical properties of that individual, although they may depend lawfully on those properties" (p.125, Chalmers' italics). He goes on to claim, "The position we are left with is that consciousness arises from a physical substrate in virtue of certain contingent laws of nature, which are not themselves implied by physical laws" (p.125).
We can see the sense of dualism implied by Chalmers' view by the fact that he says consciousness "arises" from a physical substrate: consciousness is something "extra" or "different from" physical stuff. It is not that consciousness is merely identical or reducible to matter (like the brain), but rather that consciousness is created, generated, or produced by matter. My best interpretation of Chalmers' view suggests that he calls this property dualism rather than substance dualism because he holds that consciousness depends on a physical substrate; consciousness cannot exist independently or in the absence of some kind of physical substrate. This is to say, Chalmers holds that consciousness can only exist insofar as it is causally supported by some physical event or other. I consider this analogous to a light bulb that is only lit insofar as there is an electrical current causing its illumination; the moment the electrical current stops flowing, the light bulb ceases to glow. So, just as the illumination only exists insofar as it is sustained by electricity, so consciousness only exists insofar as it is sustained by appropriate physical substrates (like neurons, say).
Chalmers further suggests that the mental and the physical are related by some kind of psycho-physical laws that are not derivable from physical laws alone. This is to say, there is a law-like connection between the mental and the physical such that identical physical substrates will produce identical mental effects; it's just that one cannot derive these law-like connections merely from examining the physical substrates alone. For example, if physical state p1 is instantiated at time t1, then mental state m1 is necessarily also instantiated at t1. Furthermore, if p1 goes out of existence at time t2 (and p1 is not replaced by some homologous physical state at t2), then m1 also goes out of existence at t2. Chalmers' view aspires to preserve a deterministic system of causes and effects. This deterministic system is simply broader than the physical world alone.
Now, with this much of Chalmers' view on the table, I want to explore some of the implications and complications of his view. Since Chalmers grants that experiences (also called "phenomenal properties") exist, he needs some kind of account of "subjectivity," "subjective properties," or simply "the subject of experience." Chalmers agrees early-on in the book (page 6) that he means "consciousness" or "experience" to include "having subjective experience." The issue I wish to raise is the problem of integrating subjectivity into an otherwise objective world.
As I see it, Chalmers has two options: either (a) physical objects are intrinsically objective and they create subjectivity entirely out of objective "ingredients," or (b) physical objects are intrinsically both objective and subjective and they just breathe life into an already existant subjective substrate. I argue that both options are problematic.
Option (a) is problematic because it ascribes rather special powers to physical objects. To explain my objection, let me clarify what I mean by "objective" and "subjective." I take these terms to describe what I'll call "modes of existence." If something has an "objective" mode of existence, it is open to observation from the "outside." This is to say, it exists in such a way that different observers can observe the same thing in the same way. Physical objects are characteristically taken to be objective in this way. For example, the same chair can be observed by many different people--they can each see its shape, color, size, etc. By contrast, if something has a "subjective" mode of existence, it is closed to observation from the "outside." This is to say, it exists in such a way that different observers cannot observe the same thing in the same way. Mental states are characteristically taken to be subjective in this way. For example, an individual's visual experience is only open to the one who is the subject of the visual perception. I, for instance, cannot see exactly what you see when you have a visual experience. I can infer what you might be seeing, but I cannot have your exact visual experience directly. Your visual experience is closed to me, and mine to you.
Furthermore, I claim that objective and subjective describe "modes of existence" because it is only entities or substances (i.e. ontological particulars) that can be open or closed to observation from the outside. Objective and subjective are categories or "ways of existing" that entities can be classified under.
Chalmers does not discuss objectivity and subjectivity at much length in his book, so it is possible he could dispute my definitions of these terms. Nevertheless, I will proceed on the assumption that these definitions are reasonable. What follows from my definition is that objectivity and subjectivity are mutually exclusive categories. In effect, something cannot be simultaneously "open" and "closed" to observation from the outside, so something cannot simultaneously have an objective and a subjective mode of existence. Analogously, circles and squares are also mutually exclusive categories: the same figure cannot simultaneously be a circle and a square.
Now, here is what I think is wrong with option (a) above--that physical objects are intrinsically objective and they create subjectivity out of purely objective ingredients: In order for (a) to work, it must be possible for objective entities to possess the necessary and sufficient conditions for bringing subjective entities into existence; but, I argue, this is impossible. I say this is impossible because it seems incoherent to assert that anything subjective could come into existence from purely objective antecedents. For illustration, let's consider the typical example of objectivity and subjectivity--the human brain and human consciousness. Presumably, Chalmers would accept that some minimal number of physical molecules (e.g. brain molecules) are sufficient for bringing states of human consciousness into existence. For simplicity, let's just say that that minimal number is 100 molecules. Each of these molecules, as physical objects, has an objective mode of existence. In order for option (a) to work, it would have to be possible that these 100 objective physical molecules--and only these 100 objective physical molecules--could be combined in such a way that a subjective entity comes into existence. The problem, though, is that subjectivity is in no way conceptually derivable from exclusively objective antecedents. How is it that 100 molecules, each of which is objective, can pool their objective resources together to get something subjective--something fundamentally different in kind? This strikes me as analogous to asserting that one can create a log cabin out of nothing but bricks. No matter how one arranges a bunch of bricks, he or she would not magically create wood. However, epiphenomenalism under option (a) would have to do something like that. Even worse, though, option (a) not only posits that the purely objective can cause the subjective, but that the subjective is actually an ontologically new entity in addition to the physical entities. Those 100 brain molecules we suppose create consciousness do not go out of existence once the mental states come into existence--i.e., the molecules do not transform into, or get replaced by, mental states; rather, the brain molecules remain intact in their objective physical forms as a new subjective mental state gets added to the universe. Analogously, this is like constructing a building out of bricks, then, in addition to having a brick building, a wooden building also comes into existence right next to it. It is not that the brick building turns into a wooden building; but, rather, that the brick building remains intact and an entirely new wooden building comes into existence from the bricks. I claim that there is no conceptual foundation for this kind of assertion.
I do not know of any other examples of the kind required for objective-to-subjective causation, and Chalmers himself does not provide any. Further, I cannot even conceive of possible examples of this kind of causation. Thus, I claim the possibility of creating subjective from objective is simply incoherent. This sort of objective-to-subjective problem seems to plague all "matter-first" theories of mind, which is what Chalmers' epiphenomenalist view seems to be.
Now, in a section in his book entitled "The Intrinsic Nature of the Physical," Chalmers did somewhat grapple with the objection I just raised, though he did not frame it in terms of "subjective" and "objective" but rather in terms of "physical" and "phenomenal." He offered a solution that is similar to option (b) I listed earlier--that physical objects are intrinsically both objective and subjective. Chalmers writes:
"A less extreme case in which intrinsic properties are protophenomenal, or in which some are neither phenomenal nor protophenomenal, is perhaps best regarded as a version of Russell's neutral monism. The basic properties of the world are neither physical nor phenomenal, but the physical and the phenomenal are constructed out of them. From their intrinsic natures in combination, the phenomenal is constructed; and from their extrinsic relations, the physical is constructed" (p.155).Here we can see Chamlers proposing a way of getting the objective and the subjective into the same entity. (Note, I think his term "physical" includes my concept of "objective," and his term "phenomenal" includes my concept of "subjective." I will continue to use the objective/subjective terminology because I find it more precise.) On this neutral monist view, no entity is entirely objective nor entirely subjective; instead, all entities are primarily "neutral," and the objective and subjective would be different expressions of that same neutral "stuff."
The advantage of this view is that it bypasses the objection I raised against option (a), because the objective no longer has to bear the burden of creating the subjective. That burden is instead shifted to something more fundamental than the physical--something "neutral." However, the disadvantage, as I see it, is that we have no conceptual basis with which to evaluate this "neutral stuff," so I have no way of knowing whether something neutral causing both subjective and objective is any more plausible than something objective causing subjective. Somewhat along these lines, Bertrand Russell wrote in his 1927 book, The Analysis of Matter that neutral monism is not itself "demonstrable":
"To show that the traditional separation between physics and psychology, mind and matter, is not metaphysically defensible, will be one of the purposes of this work; but the two will be brought together, not by subordinating either to the other, but by displaying each as a logical structure composed of what, following Dr. H. M. Sheffer, we shall call 'neutral stuff.' We shall not contend that there are demonstrative grounds in favour of this construction, but only that it is recommended by the usual scientific grounds of economy and comprehensiveness of theoretical explanation" (p.10).Russell admits here that he does not have "demonstrative grounds" for invoking neutral stuff, but that it is instead preferable on theoretical grounds. I grant that the view does have theoretical advantages because it is an attempt to get the objective and subjective into the same universe, but this theoretical victory is at the expense of any kind of conceptual understanding. As such, I am not inclined to outright reject the view, but I also feel no rational compulsion to embrace it either. Furthermore, if I am right that subjective and objective are mutually exclusive categories, and if I understand Chalmers correctly that he thinks the objective and subjective properties of the brain molecules exist simultaneously in those molecules wherever consciousness exists, then this neutral view runs into the problem of getting the same entity into both objective and subjective categories. This is again like trying to get the same figure to simultaneously satisfy the categories "circle" and "square." Since the neutral view is conceptually vague enough, I am not sure that this charge is fatal; I will simply record it as a worry.
So, I argue that Chalmers' epiphenomenalism is conceptually problematic because it has trouble reconciling the subjective with the objective. If Chalmers opts to claim that the objective physical objects cause the subjective mental states, there seems to be no way to get anything subjective from purely objective antecedents, just like one cannot make a log cabin out of nothing but bricks. If Chalmers opts to claim that the objective and subjective are both constructed out of some more fundamental neutral stuff, his view gains some theoretical appeal, but it seems to do so at the expense of a clear conceptual foundation. I happen to prefer conceptual clarity over theoretical appeal, so I find this last option unsatisfying.
Objection 2: Causal Flaw with Epiphenomenalism
As I mentioned earlier, Chalmers admits that his position leaves no room for the mind to uniquely cause anything. He accepts that the physical world is "causally closed," which means that for any physical event, there is a complete causal story that explains that physical event by reference to only physical entities and forces. In Chalmers' words:
"The best evidence of contemporary science tells us that the physical world is more or less causally closed: for every physical event, there is a physical sufficient cause. If so, there is no room for a mental 'ghost in the machine' to do any extra causal work" (p.125).Even though Chalmers is aware of this causal problem, he oddly does not discuss the subject of free will at all in his book. I scanned the index for any mention of "free will," "determinism," "agency," and the like, but found nothing. He does say, "A problem with the view I have advocated is that if consciousness is merely naturally supervenient on the physical, then it seems to lack causal efficacy" (p.150). I infer from this statement and others in the book that Chalmers would have to hold that free will (as commonly understood) does not exist--whatever behaviors human bodies carry out, none of them are the result of standard notions of "choice," "volition," "intention," "will," "decision," etc. Instead, all human behavior is a consequence of determinate laws of physics, just as the gears of a clock move according to determinate laws of physics. So, Chalmers accepts that our experiences are real subjective, first-person states of awareness, but he denies that any of these subjective, first-person states of awareness have unique causal effects on the physical world, including our physical bodies. For example, according to Chalmers' view, if I experience thirst, that sensation of thirst itself has no causal effect whatever on whether my hand reaches for a glass of water. The motions of my hand are ruled exclusively by the determinate laws of physics, so my mental experience has nothing to do with how my hand moves.
Although Chalmers does not discuss free will directly, he does discuss causal concerns of his view in a section he titles "Strategies for Avoiding Epiphenomenalism" (pp.151-6). One of these strategies (and the one he says he prefers) is to opt for a version of Russell's neutral monism discussed above. This strategy incorporates phenomenal (or protophenomenal) properties into the basic structure of reality, which includes its causal structure. On this view, the mental (phenomenal) will inherit causal powers by virtue of its being woven into the fabric of the fundamental--but physically/mentally neutral--ontological stuff. As I mentioned earlier, the idea of neutral ontological stuff is so abstract that I cannot get a conceptual grip on it in order to evaluate it, but my point here will be to emphasize that the causal powers Chalmers retains for the mental by this maneuver are very unlike the causal powers we normally attribute to free will. Chalmers recognizes as much when closing his section about strategies for avoiding epiphenomenalism:
"It remains the case that natural supervenience [Chalmers' view] feels epiphenomenalistic. We might say that the view is epiphenomenalistic to a first approximation: if it allows some causal relevance for experience, it does so in a subtle way. I think we can capture this first-approximation sense by noting that the view makes experience explanatorily irrelevant" (p.156, Chalmers' italics).If we preserve our common-sense notion of free will causation--that our decisions are causally relevant to the physical world--Chalmers' view essentially rejects this kind of causation. The kind of causation Chalmers preserves is much more like "regular" physical causation in physics: there are basic entities that move and interact according to definite laws and forces; but, on Chalmers' view, the basic entities are not exclusively physical. To whatever extent nonphysical basic entities exist, they "obey" determinate laws just as much as physical basic entities "obey" determinate laws. I get the impression that Chalmers aspires to tie his view as closely as possible to what we might call "regular physics" as possible while still including non-physical phenomena in the picture.
Now, the most interesting consequence of Chalmers' view of causation is what he has to say in response to interactionist dualism. Interactionist dualism is basically the Cartesian view which says that mind and body are separate substances that interact with each other. For example, dualism holds that physical states can affect mental states, like when a stubbed toe causes pain. Likewise, dualism holds that mental states can affect physical states, like when a decision to paint a picture causes one's hand to wield a paintbrush.
Chalmers argues against interactionist dualism as follows:
[A]ll versions of interactionist dualism have a conceptual problem that suggests that they are less successful in avoiding epiphenomenalism than they might seem; or at least that they are no better off than the view I have advocated. Even on these views, there is a sense in which the phenomenal is irrelevant. We can always subtract the phenomenal component from any explanatory account, yielding a purely causal component. Imagine (with Eccles) that 'psychons' in the nonphysical mind push around physical processes in the brain, and that psychons are the seat of experience. We can tell a story about the causal relations between psychons and physical processes, and a story about the causal dynamics among psychons, without ever invoking the fact that psychons have phenomenal properties. Just as with physical processes, we can imagine subtracting the phenomenal properties of psychons, yielding a situation in which the causal dynamics are isomorphic. It follows that the fact that psychons are the seat of experience plays no essential role in a causal explanation, and that even in this picture experience is explanatorily irrelevant....The denial of the causal closure of the physical therefore makes no significant difference in the avoidance of epiphenomenalism" (156-8).I find this argument an incredibly novel and interesting objection to interactionist dualism. Here, Chalmers grants the dualist the rejection of physical causal closure (something the dualist needs for his position to work), yet Chalmers goes on to argue that dualism is still problematic. Chalmers does this by conceptualizing phenomenal entities (i.e. experiential entities) as similar to physical particles (like electrons). These so-called "psychons" obey determinate laws just like electrons obey determinate laws, it's just that psychons happen to be non-physical and have experiential properties. Chalmers' suggestion is that the psychons would have whatever causal powers they have independently of the experiential properties they happen to have. Analogously, I think Chalmers would agree that, in a similar way, a blue billiard ball would have the causal powers it has independently of its blueness property--i.e., if we subtract the blueness off of the billiard ball, it would still bounce around the table and play the same causal role it usually does. Thus, Chalmers says that if we subtract off the phenomenal from the psychons, the same causal system would remain intact--there just wouldn't be any experiences. The thrust of Chalmers' argument here is that the psychons would be obeying determinate causal laws in exactly the same way whether or not experiences exist, thus leaving no role for free will or volition to cause anything. Thus, Chalmers claims that interactionist dualism is ultimately no better at preserving free will than epiphenomenalism, so epiphenomenalism is not actually as bad an option as typically believed.
Although I find Chalmers' argument here novel and interesting, I do think it is flawed for two reasons: (a) it leaves out the role of a "subject" in mental causation, and (b) his argument is potentially circular.
With respect to (a) Chalmers' argument leaves out the role of the "subject" in the interactionist dualist's picture. Chalmers seems to present interactionist dualism as a kind of extension or broadened version of regular physics. That is, he seems to take our usual understanding of basic physical particles (like electrons, protons, etc.) and then adds a few more basic--but non-physical--particles (e.g. the idea of a "psychon") to the picture, and gives these new particles the properties of experience and non-physicality. However, he ascribes to these psychons basically the same causal roles he would ascribe to electrons and protons--namely, that they deterministically follow whatever laws of nature happen to exist.
My objection is that interactionist dualism need not be conceptualized in this way. Interactionist dualists can plausibly hold that a non-physical subject can exist that is not like a regular physical particle or a psychon--nor is the subject a "collection" of individual particles or psychons; instead, interactionist dualists could say that a non-physical subject is more "soul-like." That is, it is an enduring non-physical substance that possesses a single, unified point of view. This idea is perhaps best captured by thinking about what it means to be a "subject" of experience. A subject is generally the entity of which an experience is a property. For example, let's take the experience of visiting the Eiffel Tower in France. Specifically, let's take the visual experience. There is "something it is like" to see the Eiffel Tower in person. Now, this experience is not sort of "free floating" out there in the world the way electrons are sort of free floating around. Instead, this experience is anchored to some subject--someone to whom this experience is attached. So, for Chalmers' position to work, he needs to give some sort of account of a subject of experience; however, I do not think he successfully does this.
The best passage I could find in Chalmers' book discussing the role of a subject of experience is one I found in a section entitled, "What is it like to be a thermostat?" This section is part of a larger discussion where Chalmers speculates that consciousness might be linked to information processing systems. This is to say, Chalmers suggests that conscious states are somehow created or instantiated by information states. Chalmers introduces the relevant section this way:
"To focus the picture, let us consider an information-processing system that is almost maximally simple: a thermostat. Considered as an information-processing device, a thermostat has just three information states (one state leads to cooling, another to heating, and another to no action). So the claim is that to each of these information states, there corresponds a phenomenal state. These three phenomenal states will all be different, and changing the information state will change the phenomenal state" (p.293).Since Chalmers suggests in this part of the book that information is possibly the basis of consciousness, he must grapple with the implication that any information-processing system is conscious--like a thermostat. As the above quote mentions, there would be a different experiential state associated with each of the three information states of the thermostat. In other words, there is "something it is like" to be a thermostat. In this discussion, Chalmers goes on to comment on the issue of a subject of experience:
"Some intuitive resistance may come from the fact that there does not seem to be room in a thermostat for someone or something to have the experiences: where in the thermostat can a subject fit? But we should not be looking for a homunculus in physical systems to serve as a subject. The subject is the whole system, or better, is associated with the system in the way that a subject is associated with a brain. The right way to speak about this is tricky. We would not say that my brain has experiences, strictly speaking, but that I have experiences. However we make sense of this relation, the same will apply to thermostats: strictly speaking it is probably best not to say that the thermostat has the experiences (although I will continue to say this when talking loosely), but that the experiences are associated with the thermostat. We will not find a subject 'inside' the thermostat any more than we will find a subject inside a brain" (296-7).Chalmers' comments about the subject of experience here strike me as severely deficient. On the one hand, he seems to accept that there does exist something we could call a "subject of experience." On the other hand, though, he provides vague and somewhat contradictory statements about what the subject is. He asserts that the subject is not found "inside" a system, like a brain or a thermostat; this implies, perhaps, that the subject could be "outside" the system, but I'm not sure he would agree. Furthermore, he says "the subject is the whole system." If we take a brain or a thermostat to be a "whole system," then, presumably, the subject would be coextensive with the molecules of that system. As far as I can tell, this is equivalent to saying the brain or thermostat just is the subject. But, in the same sentence, Chalmers goes on to say something apparently in contradiction to this: he says the subject is "associated with the system." The word "associated" implies the existence of two separate entities: in this case the subject and the system. If this is right, then the subject cannot be equal to the system. Thus, both Chalmers' statements that "the subject is associated with the system" and that "the subject is the whole system" conflict with each other. The former implies the existence of something beyond the system, whereas the latter denies the existence of something beyond the system. Thus, I claim Chalmers' account of the subject is deficient.
I want to make a big deal out of this because I think the concept of a subject strongly pushes one into a substance dualist position. The substance dualist position provides the ontological foundation for the existence of a unified, enduring, self-aware seat of experience. If we only took Chalmers' statement that "the subject is associated with the system," we should find that this statement is entirely compatible with the substance dualist position: the non-physical, "soul-like" subject is associated with a system like a brain. That is, the non-physical subject interacts with the physical brain that it is associated with.
Now, as far as Chalmers' causal objection to interactionist dualism is concerned, I claim that his argument is unsuccessful against a specifically substance dualist position. I argue that the causal powers of mind are, under substance dualism, situated in the subject rather than in something like a psychon. As Chalmers seems to accept, experiences already belong to subjects: e.g., in the quote above he says, "We would not say that my brain has experiences, strictly speaking, but that I have experiences." It seems to me, then, that the substance dualist can go a step further and say that the subject--the "I"--also has causal powers. After all, Chalmers does ascribe both experiential and causal properties to psychons. All the substance dualist does is ground both experiential and causal properties in a non-physical substance instead of a non-physical psychon. With this much in place, the substance dualist can now assert that the subject's causal powers include our common notions of "volition," "will," or "choice." That is, the non-physical subject--as an autonomous ontological substance--can initiate its own causal chains independently of whatever universal conditions preceded it.
Granting all of this for the sake of argument, I think Chalmers' argument against interactionist dualism fails to hold. Recall that his argument suggests subtracting off the experiential component from something like a psychon such that all that remain are the causal components of psychons. The thrust of this argument is that, even though the physical world is no longer causally closed, experience is still causally inert. The causal components of psychons would still causally interact with each other and the physical world just the same as they would if experiences were present. That is, the causal components of psychons would continue to obey the determinate laws of nature whether or not experiences went along for the ride. Now, suppose we run the same sort of argument against the substance dualist position I've described. We subtract off the experiential components of non-physical subjects (assuming this is possible); the causal components of the non-physical subjects will still remain. Just as with the psychons, we still have a world of physical and non-physical causation happening; however, the sort of non-physical causation going on would no longer be "determinate." Instead, the subjects would still have their same causal powers of volition, will, and choice that they had when experience was present (just as the psychons had the same causal powers whether or not experience was present). Although Chalmers' argument would still succeed at making experience causally irrelevant in this case, it would not make volition, will, or choice causally irrelevant. This is because the whole point of the argument is that the same causal system remains intact independently of experiential properties. Thus, if volition, will, and choice exist as part of the causal system when experience is present, then volition, will, and choice should continue to exist as part of the causal system when experience is absent. As far as I'm concerned, volition, will, and choice are the crucial motivating components of the interactionist dualist position. Insofar as these can be salvaged against Chalmers' argument, I think his argument loses its bite. In short, Chalmers' argument does not show that interactionist dualism necessarily collapses into something analogous to epiphenomenalism.
Furthermore, my second objection (b) claims that Chalmers' argument runs the risk of being circular in the sense that he's presupposing that experience is in fact separable from causation. In the previous paragraph, one might have thought it odd to think that something like volition, will, or choice could exist in the absence of experience. It seems natural to me that one would deny that he or she could conceptually separate experience from something like choice. How could one make a choice if he or she were not aware of the choices or the experience of choosing, for example? If Chalmers were to run with this idea and respond to me by saying that the deletion of experience from the picture also entails the deletion of volition, will, and choice, then his argument would beg the question at issue: he would simply presuppose that experience is independent of mental causation, which (on our present assumption) is identical to just presupposing that free will does not exist (because free will requires experience, but experience cannot cause anything, therefore free will cannot cause anything). I say this is potentially circular because his argument is supposed to show that experience is causally irrelevant, not presuppose it. So, to the extent Chalmers equates experience and free will, his argument begs the question. However, if Chalmers grants that experience and free will are separable, then he is vulnerable to objection (a) described previously: namely, by deleting experience, we would still have free will causation operating in the world.
So, I claim that Chalmers' version of epiphenomenalism is problematic because it denies any unique causal role for the mind. His discussion of this causal problem is deficient because he neither addresses free will directly nor produces an adequate account of the subject of experience. Further, his claim that interactionist dualism does no better than epiphenomenalism at preserving a causal role for experience falls short for two reasons: first, substance dualism can plausibly preserve causal powers of volition, will, and choice in the absence of experience; second, Chalmers' argument is potentially circular if he makes experience necessary for the existence of free will because this presupposes, rather than demonstrates, that epiphenomenalism is true.
David Chalmers' The Conscious Mind is a great work of philosophy. Even though my review sought to argue against his epiphenomenalistic thesis, this is not to say the book is poorly done. In fact, I take Chalmers' arguments here to be some of the strongest, and most thorough, I have yet seen in the philosophy of mind. It is precisely for this reason that I had to expend great effort to build arguments against him. I do not know if my objections here are successful, but I am certainly grateful to Chalmers for putting his book together with as much care as he did. This book is easily a must-read for any serious student of the philosophy of mind.