ZOMBIES AND THE CASE OF THE PHENOMENAL PICKPOCKET

Transcription

1 M.P. LYNCH ZOMBIES AND THE CASE OF THE PHENOMENAL PICKPOCKET ABSTRACT. A prevailing view in contemporary philosophy of mind is that zombies are logically possible. I argue, via a thought experiment, that if this prevailing view is correct, then I could be transformed into a zombie. If I could be transformed into a zombie, then surprisingly, I am not certain that I am conscious. Regrettably, this is not just an idiosyncratic fact about my psychology; I think you are in the same position. This means that we must revise or replace some important positions in the philosophy of mind. We could embrace radical skepticism about our own consciousness, or maintain the complete and total infallibility of our beliefs about our own phenomenal experiences. I argue that we should actually reject the logical possibility of zombies. A prevailing view in contemporary philosophy of mind is that zombies are logically possible. I will argue, via a thought experiment, that if this prevailing view is correct, then I could be transformed into a zombie. If I could be transformed into a zombie, then surprisingly, I am not certain that I am conscious. Regrettably, this is not just an idiosyncratic fact about my psychology; I think you are in the same position. This means that we must revise or replace some important positions in the philosophy of mind. We could embrace radical skepticism about our own consciousness. But I suggest that we should actually reject the logical possibility of zombies. Others have worried that the logical possibility of zombies implies that you or I might be one (e.g., Dennett, 1991, see also Bayne 2001). The case of the phenomenal pickpocket shows us why we must take this worry seriously. 1. There are several different senses of the word consciousness. I am concerned with the phenomenal sense. I am conscious in this way in that there is something that it is like to be me. Accordingly, a mental state is conscious in this way when there is something it is like to be in that mental state, when it has a particular qualitative feel or aspect. If we assume for the sake of argument that I am conscious in this sense when I have some qualitatively conscious mental states, then Synthese (2006) 149: Ó Springer 2006 DOI /s

2 38 M.P. LYNCH (1) I am certain that I have qualitatively conscious mental states or experiences. In saying that I am certain that I have at least some qualitatively conscious mental states, I mean that I am able to justifiably rule out all possibilities that I do not have conscious mental states. 1 Alternatively, we might say that I know that I have such states in the strongest sense of know, or that I am conclusively justified in believing that I have such states. Certainty is a much stronger epistemic category than knowledge in the usual sense of knowledge the sense in which we say that we know that we have enough money to pay the bill, or that we know where our car is parked. Obviously, the same can be said for the relationship between certainty and mere justification; I have many justified beliefs that are not certain. So we should modify (1) a little to account for some small misgivings. For one thing, no one is certain that they will be conscious in the future. You might even doubt that you were conscious yesterday (perhaps your memory has been altered). Further, you probably aren t certain that every state of your mind is conscious. So I ll stipulate that (1) is equivalent to (1a) I am certain that I have some qualitatively conscious mental states now. Few things in life are certain. But for most people, (1) will be among them, and requires no argument. But if arguments are needed, here are three. First, try to seriously consider the thought that (C-): I am having no qualitatively conscious thoughts right now. I would be surprised if you can do so. To consider (C-) requires a conscious thought with a distinctive qualitative feel (some of my students describe it as making their head hurt ). Thus it is impossible to consciously entertain that (C-) is true. To even consciously consider the thought that I am not conscious I must have at least one conscious thought. Therefore, it seems as if I can rule out the possibility that I am not conscious now simply by consciously wondering whether it is true. And that strongly suggests that (1) is true. Not everyone may be convinced by this argument. Some, for example, may not find (C-) to have a qualitative feel; if so, then perhaps I might unconsciously consider whether (C-) is true. Therefore, here is another, and perhaps more intuitive route to the same conclusion. One reason to think that a proposition is certain is that you can t imagine that it is false. I can t imagine that I am having no conscious, phenomenal experience now. This is not only because imagining is a form of conscious thought, it is because imagining myself without any

3 ZOMBIES AND THE CASE OF THE PHENOMENAL PICKPOCKET 39 conscious experience is imagining nothing. It is similar to trying to imagine what it would be like to be dead (right now). This provides me with another good reason for thinking (1) is true. Finally, some might give a transcendental argument of sorts of (1). It seems to many that certainty of my own consciousness is a logical presupposition of any discussion of the subject (Chalmers, 1996, p. 193; see also Strawson, 1994, p ). These philosophers believe that an appeal to my own first-person experiences is the conceptual starting point of any investigation of consciousness. Such states of mind act as the necessary ground on which to build third-person explanations of consciousness. Without them, the problem of consciousness wouldn t even make sense. From these facts we can reason in reverse, or transcendentally. Since the problem of consciousness does make sense, we must have at least some conscious experience. This suggests, we might think, that (1) is true. Appreciated individually, these arguments are perhaps not conclusive. But taken together, they make (1) very plausible. Yet, together with some additional premises, I shall argue that if (2) Zombies are logically possible is true, then (1) must be false. In the sense of the term I am concerned with, a zombie is an imaginary being which is a perfect physical duplicate of a human being but which lacks phenomenal experience. This type of zombie has received quite a bit of attention over the last decade. 2 Except for the fact that it completely lacks any qualia, my zombie twin would be identical to me in every conceivable respect e.g., functionally, psychologically, biologically and chemically (ibid., 94 95). We would both react in the same way to stimuli, process information identically and even share all the same beliefs at least in a minimal or functional sense of belief according to which beliefs are not essentially qualitative. A minimal belief (as I ll call it) is a functional state, understood in terms of its causal connections to the environment, behavior and other mental states. Many philosophers obviously take minimal beliefs to be beliefs proper. Others demur. But following Chalmers, (1996,174) I ll lay this issue aside, and use minimal beliefs as a term for whatever what is left of a belief after any associated phenomenal quality is subtracted. Presumably, then, my zombie twin would share not only my minimal beliefs about the external environment; it would also share my minimal phenomenal beliefs, or the judgments I make about my

4 40 M.P. LYNCH conscious experience. Phenomenal beliefs, like beliefs generally, can be sorted into many different types. For the moment, we can distinguish between basic and non-basic phenomenal beliefs. 3 Basic phenomenal beliefs are beliefs about particular conscious experiences and about particular kinds of conscious experiences. Examples include my belief that the wine tastes fine, that the ache in my back is much sharper today than yesterday, and that hunger pangs can vary in their intensity. My zombie twin will have the same basic phenomenal beliefs that I have, but since he is not conscious, his will be false, while mine (most likely) will be true. Non-basic phenomenal beliefs, on the other hand, are beliefs about conscious experiences in general. This includes the belief that I have some conscious mental states, but also peculiarly philosophical beliefs about consciousness, such as understanding the nature of consciousness is difficult. My zombie twin will also have these beliefs. Some of his non-basic phenomenal beliefs such as his belief that he is conscious will clearly be false, for their truth depends on the truth of his more basic phenomenal beliefs. Some others may not be true (such as his belief that there are conscious beings, or that consciousness is difficult to explain). Zombies of this sort seem naturally impossible, not to mention implausible. Claim (2) is that zombies are merely logically possible. We can say that a situation is logically possible just when it is not contradictory. Since there appears to be no contradiction in something that is otherwise identical to a normal adult human being not having any qualia, it would appear that zombies are logically possible. 4 This does not prove that zombies are logically possible. But it does make that claim very plausible, barring further argument. Of course it is controversial what if anything the logical possibility of zombies implies. Some philosophers see the possibility of zombies as a threat to materialism; others do not. David Chalmers, for example, uses the logical possibility of zombies to argue that consciousness is not logically supervenient on the physical, and hence the facts about consciousness are not reducible to physical facts any physical facts. Materialist philosophers such as Owen Flanagan and Tom Polger, on the other hand, find zombies useful fictions; they help us test our intuitions about whether consciousness is essential for beings with our behavioral and functional organization (Flanagan and Polger, 1995, p. 313). In any event, many (although certainly not all) philosophers of mind agree that zombies are logically possible in the sense I ve described.

5 ZOMBIES AND THE CASE OF THE PHENOMENAL PICKPOCKET 41 Considered just by themselves, claims (1) and (2) certainly appear consistent. The fact that I could have a zombie twin, after all, does not itself imply that I could be a zombie, nor that I am not certain that I am conscious. For that, one needs an argument. The strategy appropriate for undermining claims to certainty is familiar from traditional skeptical arguments. The skeptic attempts to make some possibility epistemically real for us by asking us to engage in a thought experiment, one designed to show that we might be radically mistaken in some of our beliefs in ways that we had not previously imagined. Once we appreciate that fact, the skeptic claims, we can no longer rule out the possibility that such mistakes are happening now. In what follows, I will adopt this strategy in order to show that if zombies are logically possible, then contrary to what I believe, I can t be certain that I have any phenomenally conscious mental states. Specifically, I will argue that (3) It is logically possible that any particular basic phenomenal belief of mine is not true. And moreover, I will claim that if (2) and (3) are true, then we should also grant (4) It is logically possible that all of my basic phenomenal beliefs are not true. But if (4) is true, then so, I ll argue, is: (5) I cannot rule out the possibility that all of my basic phenomenal beliefs are not true. If (5), then (1) must be false. For if I cannot rule out the possibility that I am mistaken in all my basic phenomenal beliefs right now, then it is possible that, despite what I believe about my conscious experiences, I have none, and thus am a zombie. And if I cannot rule out the possibility that I am a zombie, then (1) is false; I cannot be certain that I am conscious right now Most of my beliefs about my conscious experiences are true; but it is now widely held, even by the most ardent qualiaphiles, that they can be mistaken. This is opposed to traditional Cartesian conceptions of epistemology, according to which my experiences provide conclusive support for phenomenal beliefs which are directly about those experiences. On the Cartesian view, while I might be mistaken in believing that there is an apple on the table, I can t be mistaken in believing that I seem to see an apple on the table that I am having an apple on a table experience. 6 My having that experience conclusively

6 42 M.P. LYNCH justifies my believing that I am having it it makes that belief certain. Strong Cartesian foundationalism of this sort is not much in favor these days; contemporary epistemology finds it difficult to accept that the mind has infallible access to its own contents. It seems much more plausible that inner experience provides defeasible evidence for a belief that has that experience as its content (Chalmers, 1996, p ). As Chalmers remarks, beliefs about conscious experiences are not infallible, because beliefs about experiences lie at a distance from experiences, they can be formed for all sorts of reasons, and sometimes unjustified beliefs will be formed. If one is distracted, for example, one may make judgments about one s experiences that are quite false (ibid. p. 197). Thus, if I am distracted by our conversation, I may mistakenly believe that the wine tastes fine when in fact it tastes terrible. And of course, distraction is not the only source of possible mistakes. Tasting several sweet things may lead me to form a false expectation that the next taste will be sweet as well, and therefore judge it to be so before realizing my mistake. An experience of something extremely cold can be mistaken as an experience of something hot. Lust can be mistaken for love and so on. And misclassification is not the only sort of error I can make in my basic phenomenal believings. I can also falsely believe that I am having a qualitative conscious experience of a certain type when I am not in fact having any conscious experience of that type at all. People with blindness denial, for instance, insist that they are having visual experiences when it is overwhelmingly likely that they are not. Or consider someone who has chronic pain in some part of her body who forms a habitual belief that she has such pain. Suppose that she is given a shot of what she believes is a placebo but which in fact totally numbs the part of the body where the pain is located. It seems perfectly plausible that for a short time, she may continue to believe that she is experiencing pain in that part of her body even after she is not in fact experiencing anything in that area. 7 So while the reasons may vary, the facts suggest that our basic phenomenal judgments can and often are mistaken. These reflections strongly suggest that (3) is true that any basic phenomenal belief of mine might not be true. We can bolster the case even further by considering what I ll call the case of the phenomenal pickpocket. As some of us unfortunately know better than others, pickpockets first distract their victims and then lift their possessions while the victim s attention is elsewhere. Now suppose that there is an all-powerful demon with tendencies toward kleptomania. This demon

7 ZOMBIES AND THE CASE OF THE PHENOMENAL PICKPOCKET 43 might pick our phenomenal pockets. Specifically, we can imagine that the demon can distract me from any particular conscious experience that I believe I am having, e.g., when tasting wine. The demon might arrange for loud noises, or false expectations, or simply ensure that I become wrapped up in philosophical argument. At the very instant I am distracted, it removes that particular conscious experience (the taste of the wine) from my mind but leaves the belief that I am having that experience intact. Further, it does not replace that experience with any other. This also could be done in a variety of ways. For example, the demon could prevent me from having an experience of x by fiddling with conditions very far downstream from consciousness. It might change the physical structure of the relevant objects, or fiddle with my neural processing, therefore removing my ability to have the experience in question. More directly still, the demon may simply eliminate or remove the qualia themselves. At the end of the day the exact method the demon employs is irrelevant. As with Cartesian skepticism about the external world, we don t need to know how the demon accomplishes its task, just that it can. So postpickpocketing, I will still believe that I am having the experience in question. Only now that belief is false, for I won t be having that experience at all. Intuitively, my belief hasn t changed; but its truthvalue has. I just don t notice that fact. Beliefs about experiences that are formed while I am attending to those experiences are less likely to be mistaken. Yet although this is true, it in no way rules out the possibility of our phenomenal pickpocket. For I can be wrong about my experiences even when I am attentively examining them. Consider the case of believing (based on false expectations) that you are experiencing heat when in truth you are experiencing extreme coldness. Or more damning, consider that you can have all sorts of false beliefs about your emotional experiences even when you are concentrating intently upon them. Reflecting sincerely on my own state of mind, I might believe that I am not angry at Jim or not in love with Sue even when I am indeed experiencing both emotions (as Sue or John might attest). I can confuse emotional states for each other no matter how attentive I m being. (Indeed, when it comes to emotional experiences, being too attentive sometimes seems to add to the confusion). I might confuse jealousy over a lover with concern for her welfare; I might confuse resentment with righteous indignation or I might confuse being appropriately detached with feeling nothing at all.

8 44 M.P. LYNCH Second, my attention is hardly perfect. It is difficult to know when I am truly attending to my experience and when I am not. I can sometimes drift off without noticing. Therefore, I can t rule out the possibility that I have drifted off or been led off by our demon even when I believe I am attending to what I believe are red experiences. In short, attentiveness can only guarantee the truth of phenomenal beliefs to the degree to which it is guaranteed itself. But we can be misattentive, and therefore can t rule out the possibility that we are misattentive whenever we are examining our phenomenal beliefs. 8 The case phenomenal pickpocket shows that (3) is extremely plausible. It is logically possible that any one of my phenomenal beliefs could be mistaken. Does it also show (4) that all of my phenomenal beliefs could be mistaken? Some may point out that they can no more imagine the demon removing all of my qualia and therefore becoming a zombie than they can imagine being a zombie. To convince them, we need to provide a more detailed case. I call this more detailed case the spectrum of experience-removal. Here, we imagine not that the demon steals a single phenomenal experience, but that he deftly removes them one at a time, until gradually all are gone. In this way the spectrum helps us to correctly conceptualize turning into a zombie via phenomenal pickpocketing. At each point on that spectrum, as my experiences are gradually removed, I will continue to believe as I would have believed had those experiences remained untouched (see Figure 1 below). So throughout the process, I will continue to not only retain my belief that e.g., the wine tastes fine, but I will also have the second-order belief that my experience is roughly the same as it was a moment ago. From that second-order standpoint, I simply won t recognize any moment along the spectrum when my experience becomes suddenly different. Therefore, I will never detect a difference between any one point on the spectrum of experience-removal from any other point; so I will not detect a difference between being fully conscious and being without any conscious experience whatsoever. Figure 1. The spectrum of experience removal.

9 ZOMBIES AND THE CASE OF THE PHENOMENAL PICKPOCKET 45 Compare the case of an ordinary mistake in phenomenal belief something that we ve already assumed (for good reason) is possible. At time t 1 I believe truly that I am experiencing heat. At t 2, for plain old ordinary reasons, I am distracted. At t 3 I judge that I am still experiencing heat but am wrong, and do not notice the mistake. If this is possible (as we are assuming) then even if there is a difference in what seems to be the case from one stage to the next, there can be no difference in my over-all or second-order point of view between the three temporal stages. If there were a difference in my second-order reflective stance I wouldn t be making the mistake. My over-all point of view must remain the same even and especially in the first-personal sense or the very idea of making a mistake loses all sense whatsoever. Yet if we concede as much in the ordinary case, then we must concede the same in the case of the phenomenal pickpocket. This is why it is clearly logically possible that all of my qualia are pickpocketed in succession until none are left. For whether t 1 and t 2 are moments along the spectrum of experience-removal or simply the moments at which an ordinary misjudgment occurs, my belief remains the same even as my experience changes. This is not to deny that there are differences between most ordinary cases of phenomenal misjudgment and what happens during the spectrum. In most ordinary cases, we are having some kind of qualitative experience (e.g., a particular visual experience) but misdescribing it as another visual experience. In the case of the phenomenal pickpocket, however, the evil demon is removing my conscious visual qualia but not putting anything not even other visual qualia in their place. While this is true, we need to keep two important facts in mind. First, as I pointed above, mis-classification is not the only form of error for phenomenal beliefs; it is possible for someone to believe that she is having visual conscious experiences when she is having no such experiences at all. Second, even if this weren t the case, for our present philosophical purposes, the (alleged) dissimilarity between ordinary cases of phenomenal misjudgment and the pick-pocket case is entirely irrelevant. Consider mistakes in perceptual beliefs about the external world. It is true that most mistaken beliefs about the external world also involve confusing one sort of object for another. But in the traditional Cartesian demon scenario, this needn t be the sort of mistake being made, since in that scenario, there might not be any external objects at all you might simply be a mind floating in the void. 9 The fact that this is not

10 46 M.P. LYNCH the ordinary type of mistake we make in forming our beliefs about the world is no objection to it s being logically possible. The same point holds for phenomenal belief as well. It doesn t matter whether this happens in ordinary mistakes over phenomenal beliefs or not. It only matters that it could happen. The case of the phenomenal pickpocket shows that it is logically possible for any one of my phenomenal beliefs to be mistaken. The spectrum of experience-removal shows that it is logically possible that all of my phenomenal beliefs are mistaken. It helps us to understand, in short, how I might gradually become a zombie. It shows that if it is possible to make an unnoticed mistake in one s phenomenal beliefs, then it is possible to imagine a continual series of such mistakes due to demonic intervention. If I can imagine the phenomenal pickpocket slowly stealing all of my qualia, then, prima facie, I can t rule out the possibility that this has already happened. If zombies are logically possible, then the above possibilities apparently imply that I am not certain that I have some qualitatively conscious mental states after all So far, we ve established at least a prima facie case for (5). But some will not be convinced. Accordingly, I ll consider some reasons for thinking that we can rule out the possibility that all of my basic phenomenal beliefs are not true right now. The first objection I ll consider goes like this. The thought experiment given in the last section is meant to show that (3) and (4) are true. But in order for (3) and (4) to be true, not only must zombies be logically possible; it must be logically possible that any particular phenomenal belief of mine be mistaken. 11 Therefore, the objection continues, we should reject this second assumption and accept that at least some of my phenomenal beliefs are incorrigible or incapable of being false. If so, then I can rule out as impossible any scenario that implies otherwise. Consequently, I can still believe I am certain that I am conscious without rejecting the logical possibility of zombies. The problem with this line of objection is that it heads back in the direction of a Cartesian epistemology of mind, according to which while I can be wrong about what is the case, I can t be wrong about what seems to be the case. As I noted above, however, there is considerable evidence, both empirical and otherwise, to think that the Cartesian view is deeply flawed. It is far more plausible to believe that the connection between our basic phenomenal beliefs and our conscious experiences is contingent; the fact that I believe that I am having some experience does not logically entail that I am having that

11 ZOMBIES AND THE CASE OF THE PHENOMENAL PICKPOCKET 47 experience. Accordingly, if this objection is going to fly, there must be a good argument for accepting the incorrigibility of our phenomenal beliefs. I ll discuss two. First, one might claim that I am essentially such that I have at least one qualitative mental state, and therefore, at least the belief that I have one phenomenal belief cannot be mistaken. This is unpersuasive. Even my initial argument for the plausibility of (1) does not show that I am essentially conscious. Its conclusion is not de re but de dicto: namely, that it is impossible that I consciously think that (C-) is true. Even if this is the case, it does not show that I am essentially conscious. In fact, it could not do so, since it is false that I am essentially conscious. I am capable of going into a deep state of unconsciousness during which I exist but have no qualitatively conscious experiences at all. Arguably, this is just what happens during periods of deep sleep. Therefore, I am not essentially conscious in the phenomenal sense. Perhaps I am essentially capable of having at least some qualia. Yet my having this property would not prevent the demon from pickpocketing all of my qualia now but returning them later. For being capable of having at least some qualia is a property that could be had by my zombie twin as well. Second, one might claim that while most phenomenal beliefs can be false, some very special phenomenal beliefs can t be. Chalmers has made this suggestion recently. These very special beliefs he calls direct phenomenal beliefs. A direct phenomenal belief is formed out of direct phenomenal concepts, the content of which is partly constituted by an underlying phenomenal quality (2003, p. 235). Thus a direct phenomenal belief emerges when a subject predicates the concept of the very experience responsible for constituting its content (2003, p. 236). If I have a phenomenally red experience and attend to it, forming the direct phenomenal concept R, the belief this experience is R is a direct phenomenal belief. According to Chalmers, this way of understanding phenomenal beliefs underwrites a limited incorrigibility thesis. Namely, no direct phenomenal belief can be false (2003, p. 242). This is a limited incorrigibility thesis because the vast majority of phenomenal beliefs (such as the belief that I am in pain, for example) remain completely corrigible in Chalmers view. Nonetheless, the incorrigibility of direct phenomenal beliefs is true by definition, Chalmers suggests. Alleged counterexamples cannot truly be counterexamples since the truth of the incorrigibility thesis is guaranteed by the definition of direct phenomenal beliefs (2003, p. 243). One can t have a direct

12 48 M.P. LYNCH phenomenal belief, in other words, unless one has successfully demonstrated a phenomenal experience. I will not quibble over the definition of what is, after all, a bit of technical vocabulary. As Chalmers knows, the more important matter is whether there are such things as direct phenomenal beliefs. This to me seems a difficult question; I am not at all convinced that there are such things. Indeed, I cannot see how there could be, if Wittgenstein s private language argument is at all persuasive. Luckily, however, we can put that thorny issue aside; for even if there are direct phenomenal beliefs, and they are incorrigible, this would in no way undermine the case for either (3) or (4). Here s the point. Direct phenomenal concepts, Chalmers notes, are based in acts of attention to instances of phenomenal qualities. That is, they require attention to a quality for its formation (2003, p. 237). Such acts of attention he calls demonstrations. Yet as Chalmers admits, demonstrations are not always successful. Indeed, there may be cases where I form a concept of an instance of a phenomenal quality but where that particular phenomenal quality is not in fact instantiated. Chalmers notes that one might attempt to demonstrate a particular phenomenal quality to your experience, say a highly specific shade of phenomenal red, but be unsuccessful because you fail to notice that the visual patch you are attending to is actually a slightly dissimilar shade (ibid.). Our phenomenal pickpocket of course, presents another possibility: that in a momentary lapse of attention, you fail to note that the phenomenal quality you were attempting to demonstrate is not instantiated because it has been stolen by our demon friend. Significantly, Chalmers handles such cases by saying that your corresponding belief is not a direct phenomenal belief but a pseudo-direct phenomenal belief. Pseudo-direct phenomenal beliefs are formed by a pseudo-direct phenomenal concept, where a pseudodirect phenomenal concept is grounded in an unsuccessful act of demonstration. Accordingly, they are either false or neither true nor false that is, they aren t true. Thus on Chalmers view, a victim of our pickpocket has no direct phenomenal beliefs, only pseudo-direct phenomenal beliefs. But here a natural worry arises: even if direct phenomenal beliefs can t be false, how do I know I have any? How, in other words, can I rule out the possibility that while I have phenomenal beliefs, I have only pseudo-direct phenomenal beliefs? Given the pickpocket possibility, it seems likely that I cannot. For it may be that whenever I form an

13 ZOMBIES AND THE CASE OF THE PHENOMENAL PICKPOCKET 49 intention to demonstrate a phenomenal quality, the demon distracts me and then removes the relevant instantiation of that quality, nonetheless leaving my pseudo-direct phenomenal belief about it intact. That belief would then be false, but I would be none the wiser. And if that can happen once, surely our powerful demon can ensure it happens over and over again, even while he is also robbing me of all my other experiences. If so, then I would eventually be left with only false indirect phenomenal beliefs and false pseudo-direct phenomenal beliefs. Therefore, even if Chalmers is correct that I can have direct phenomenal beliefs, I can t rule out the possibility that I don t have them. So even if some sorts of direct phenomenal beliefs are incorrigible, it does not follow that I have any of those beliefs. Accordingly, I can t rule out the possibility that all of my phenomenal beliefs are not true. If so, then I can t be certain that I am conscious. The next line of objection is related to the first. As we ve seen, one can argue that in order to have phenomenal beliefs, I need phenomenal concepts, and in order to have phenomenal concepts, I need to have phenomenal experiences, that is, experiences with a particular phenomenal quality. But if so, then one might argue that ML at the start of the spectrum of experience-removal and ML at the end of that spectrum don t share the same concepts, and therefore, don t share corresponding beliefs with the same content. Accordingly, I can t believe that my experience is the same as it was a moment ago. I cannot do this, because in order to have a belief with that content I would have to have an experience I do not have. Yet if I can t have that sort of belief, then the pickpocket can t steal my qualia without my noticing. This objection is flawed in several ways. To begin with, even if I can have the full-blooded phenomenal concept of a red experience only if I have at some point had a red experience, it doesn t follow that I need to be having a red experience every time I employ that concept. Recently, blinded people can obviously continue to use the concept red experience and entertain all sorts of propositions about that sort of experience. And what holds for recently blinded persons holds for recently pick-pocketed persons. For a short time anyway, they will be able to form all the same beliefs they enjoyed prior to being robbed by our demon friend including, e.g., the belief that everything seems roughly the same now as it did a second ago and the belief that they are conscious. And for all you know, you are a member of this group yourself. You may have been recently pick-pocketed; and therefore might recently have become a zombie, and therefore, you cannot be certain that you are conscious. 12

14 50 M.P. LYNCH One might think that the problem is more pressing when it comes to direct phenomenal beliefs supposing, again, that there are such things. Direct phenomenal beliefs, Chalmers tell us, are short in lifespan, since the direct phenomenal concepts that compose them are limited to the lifetime of the experience that constitutes those concepts (2003, p. 240). Consequently, even if I were pickpocketed just a moment ago, I could no longer have direct phenomenal beliefs, but only, as we noted above, what Chalmers calls pseudo-direct phenomenal beliefs. Yet as we ve already seen, this fact (if it is a fact) has no bearing on the pickpocket: for the pickpocket can as easily rob us of our direct phenomenal beliefs as he can steal the qualia that compose their content. And he can do so without our believing that he has. For as Chalmers admits, (ibid. p. 245) one can make a mistake about whether one is having a direct phenomenal belief or a pseudo direct phenomenal belief. Moreover, beliefs of the sort singled out in the above objection (like, my experience is the same as it was a moment ago) are not direct phenomenal beliefs at all, but second-order, indirect phenomenal beliefs employing standing or pre-existing phenomenal concepts like experience. And I will be able to employ such concepts, as we just noted, for at least some time after being pickpocketed. These reflections remind us that the case of the phenomenal pickpocket does not compare you to your zombie twin but asks you to consider what your own situation might be like were you to be the unwitting victim of phenomenal pickpocketing. The question: Could I be a zombie? is ambiguous. If it means, Could I have always been a zombie? then the answer perhaps is no. But if it means Could I have recently become a zombie via a quick but gradual removal of my phenomenal properties? the answer seems to be yes granting, that is, that zombies are indeed logically possible. Notice, by the way, that the same line of argument counters what might appear to be a simple refutation of the pickpocket argument. One might be inclined to argue that without at least some conscious experience, I cannot even form a thought with the content that I am conscious. Given that I can form that thought, it follows that I have some conscious experience and therefore that I am not a zombie. But again, this line of argument only works if we forget that one can retain most phenomenal concepts long after losing the ability to have the phenomenal experience which helped constitute that concept. Just so long as I ve had some conscious experience, I can form the belief that I am conscious even when I am not. Furthermore, even if this was not so, this objection would still fail. For this line of reasoning assumes

15 ZOMBIES AND THE CASE OF THE PHENOMENAL PICKPOCKET 51 that we can know with certainty the contents of our own beliefs and other mental states. But of course, just as we can be skeptical about whether my phenomenal beliefs are accurate, we can also be skeptical about whether they have the content that we believe them to have. I now turn to a third sort of objection. So far, we ve seen that even if a few very special phenomenal beliefs are incorrigible, this by itself does not entail that my phenomenal beliefs can t all be false or at least not true. Significantly, Chalmers himself would seemingly grant this point, for in his view, the incorrigibility of my direct phenomenal beliefs is separate from their epistemic status (2003, p. 252). He accepts that certainty requires more and that certainty about p involves ruling out all skeptical possibilities that not p (Ibid.). Nonetheless, he objects that the hypothesis that I could be a zombie because, he thinks, it assumes that my beliefs are the only or primary determinants of my epistemic situation: From the first-person point of view, my zombie twin and I are very different: I have experiences, and he does not. Because of that, I have evidence for my belief and he does not. Despite the fact that he says the same things I do, I know that I am not him (though you might not be sure) because of my direct first-person acquaintance with my experiences. This may sound somewhat paradoxical at first, but really it is simply saying the obvious: our experience of consciousness enables us to know that we are conscious (1996, p. 199). So in Chalmers view, my phenomenal beliefs are justified by the very fact of having the experiences those beliefs are about (1996, p. 196), that is, by acquaintance (2003, p. 250). Acquaintance, according to its advocates, is a primitive epistemic relation between a subject and the properties of a phenomenal experience: whenever a subject has a phenomenal property, the subject is acquainted with that phenomenal property (ibid.). If so, then we might argue that even if it is an abstract logical possibility that the phenomenal pickpocket steals my qualia, from within my actual epistemic situation I can still rule out the possibility that this has already happened. What is the argument here? Well, the first part is clear enough. It goes like this: My phenomenal beliefs are justified by acquaintance with phenomenal experiences (or properties). I am acquainted with some phenomenal experiences (or properties). Therefore, the beliefs formed on the basis of such acquaintance are justified. Many will quarrel with the first premise, since it is controversial whether acquaintance can justify beliefs. 13 I will not. For the pickpocket scenario is completely consistent with the hypothesis that my

16 52 M.P. LYNCH basic phenomenal beliefs about my conscious experience are justified by the mere fact that I am acquainted with those experiences. That is, there is nothing in the pickpocket scenario which prevents us from adopting the following principle: Were I to be acquainted with my conscious experiences that acquaintance would justify or contribute to the justification of my beliefs about my experiences. The issue at hand is not whether acquaintance would justify my basic phenomenal beliefs. It is whether I can be certain that I have conscious experiences. So let us turn to the second premise, which asserts that I am acquainted with some conscious experiences. It implies that I have some conscious experiences. Since it so explicitly concerns the very issue at hand, how can it be used in an argument for the conclusion that I know that I am conscious? Here is one possibility. According to a strong externalist approach to epistemology, I do not have to justifiedly believe that a belief is justified in order for that belief to be justified. 14 If so, then I don t have to justifiedly believe that I am acquainted with my conscious experiences in order for my phenomenal beliefs to be justified by acquaintance with conscious experiences. All that is required for the justification of my belief that I am having some conscious experience is that I be acquainted with that experience. If I am, then I am justified in believing that I am having that experience. And if I am justified in believing that I have conscious experiences then, Chalmers reasons, I am justified in believing that I am conscious. Yet as noted, justification is one thing, certainty is another. In order to defeat the pickpocket argument, Chalmers needs an additional claim: Namely, the justification by acquaintance my phenomenal beliefs enjoy is strong enough to make some of those same phenomenal beliefs certain. As he says,...acquaintance with a property enables one to eliminate all (a priori) epistemic possibilities in which the property is absent. If so, then in the right cognitive background (with sufficient attention, concept formation, lack of confusion and so on), the justification of a direct phenomenal belief P by acquaintance with a property will sometimes enable a subject not just to know that P by the usual standards of knowledge, but to eliminate all skeptical counterpossibilities in which P is false (2003, 252-3). Therefore, I can be certain that I have some conscious experiences. I want to stress again that I have not argued that the pickpocket scenario proves that we are completely unjustified in believing that we have at least some conscious mental states. Indeed, if knowledge is

17 ZOMBIES AND THE CASE OF THE PHENOMENAL PICKPOCKET 53 meant in a very weak sense, then we may even grant that, if I am in fact acquainted with conscious experiences, then I know that I have conscious experiences. But if we take knowledge to imply certainty which we must do if it is meant to counter the problem raised by the phenomenal pickpocket then it is clear that the above chain of argument does not show that I am certain that I am conscious. Here s why. As Chalmers himself notes, in order for my acquaintance with an experience to be sufficiently robust to allow me to be certain in my direct phenomenal belief about that experience, the cognitive background conditions would have to be perfect. In his view, apparently, this is sufficient. Crucially, they just have to be perfect; I don t even have to believe that they are. But while this externalist stance, while controversial, is perhaps sufficient when it comes to prima facie justification, the standards for certainty are clearly much higher. As Bayne (2001, p. 415) has pointed out, one needs to engage in doxastic ascent to meet the skeptic s challenges. Even a tough-minded externalist like Alston holds that in seeking certainty, as opposed to less demanding epistemic goals, one is intuitively seeking what Alston has called full reflective assurance for one s beliefs (1989, p. 334). This is why we say that when the issue is the certainty of p, one is obligated to rule out all the possibilities that not-p. And intuitively, the process of ruling a possibility out involves appealing to reasons and conclusive reasons for thinking those possibilities don t obtain. It is not enough to note that if some of these possibilities don t obtain, (e.g., if we have conscious experiences and are acquainted with them), then one may (justifiably) believe that p. Certainty is not a matter of luck; one can only be certain that p on the basis of premises that are themselves certain. That is the point of calling it certainty. Prior to encountering the pickpocket scenario, I thought that it was certain that I am now having conscious experiences. But now that the pickpocket scenario is raised, I cannot rule out the possibility that the demon has stolen all of my qualia. And since I cannot rule that possibility out, (1) is false. That is, it is no longer certain that I am having conscious mental experiences whether or not I am merely justified in believing that I have them. 4. The case of the phenomenal pickpocket places us in an uncomfortable position. It does so because while we have good reason to believe (1) and (2), the above two sections show that I cannot completely rule out the possibility that I have recently been turned into a zombie. If so, then if (2) is true, (1) must be false. One of my initial claims must be given up.

18 54 M.P. LYNCH So perhaps we should go ahead and just give up (1) and admit that we are not certain that we are conscious. Maybe it really is all dark inside. Some might argue that indubitable certainty about anything may be impossible, even about this. Besides, logical possibility is hardly real possibility, let alone plausibility. So to admit the logical possibility of being a zombie is hardly worrying. Or at least it is no more worrying than admitting the logical possibility of everyone else being a zombie. Doubtless there will be philosophers inclined in this direction. Yet I am reluctant to join them myself. I think that giving up certainty about one s own consciousness is quite a bit more worrying than giving up certainty about someone else s consciousness. I am inclined to agree with those philosophers I mentioned at the outset who take certainty of our own consciousness as the necessary bedrock which allows us to even understand the subject. Another option, if you can call it that, would be to embrace a truly radical Cartesian epistemology and hold that none of our phenomenal judgments, basic or otherwise, can be mistaken. This would be to claim, in effect, that any belief I have about my phenomenal experience including the belief that I have phenomenal beliefs cannot be false. If a suitably strong version of this position were true, then perhaps we could be certain that we had not been pickpocketed. Yet such a view is not very plausible, to say the least. Thus we reach our final, and I suspect, best option; namely that zombies are logically impossible. If we reject this assumption, then it is not logically possible for someone to be functionally and physically identical to me and to have all my minimal phenomenal beliefs and not be conscious. Therefore, since this description obviously applies to myself, I can t be as I am and not be conscious. Yet some will wish for a more positive reason to believe that zombies are logically impossible. They will hold that if one is going to deny that zombies are possible even with the support of the phenomenal pickpocket argument, you need to say something positive and independent of the present considerations about why they are. Otherwise, one has simply a prima facie case for the impossibility of zombies. One such explanation for why zombies are impossible is this. 15 Suppose we take it that for a state to be conscious is for it to be the object of a higher-order mental state. This is the core of the so-called HOT (higher-order thought) theory. 16 Such theories come in different

19 ZOMBIES AND THE CASE OF THE PHENOMENAL PICKPOCKET 55 varieties and admit of numerous complexities, but for our purposes, I only want to consider the implications of taking the core idea of the higher-order thought theory as a necessary truth. That is, assume that in every possible world, any mental states, whether belief or experience, is qualitatively conscious whenever it is the object of a higher-order thought. If so, then necessarily, any being that has higher-order thoughts about its mental life, true or false, has conscious mental states. My zombie is physically and functionally identical to me. As a result he has my basic phenomenal beliefs. Yet crucially, he also has my thoughts about those basic phenomenal beliefs. His basic phenomenal beliefs will be false, as will many of the beliefs he has about those basic phenomenal beliefs. Nonetheless, on the higher-order thought theory I am imagining, some of his basic phenomenal beliefs would nonetheless be conscious, since they are targets of higher-order thoughts (which needn t be conscious of course). It follows that my zombie would, impossibly, have some qualitatively conscious mental states; and so zombies would seem logically impossible. 17 Of course, the HOT theory is open to numerous objections even without interpreting its core thesis so strongly. Yet, if zombies are truly logically impossible, (not just nomologically impossible or simply crazy-sounding) then something at least vaguely like it must be true. This is the important point. There must be a logical connection between our functional, physical or behavioral states of mind and our conscious states of mind. In short, if zombies are logically impossible, then what Owen Flanagan calls conscious inessentialism, or the view that intelligent activities can be done without conscious accompaniments, is false (Flanagan, 1992, p. 5). Consciousness would be logically essential for the sorts of cognitive functioning we typically perform with conscious accompaniment. This is a surprising consequence. The case of the phenomenal pickpocket confronts us with three evils: Either admit that we are not certain that we are conscious, accept a truly radical Cartesian epistemology of consciousness, or admit that zombies are logically impossible after all (and thus agree that conscious inessentialism is ruled out). No matter which option we take to solve the problem, the case of the phenomenal pickpocket requires us to revise or reject a widely held position in the philosophy of mind. Forced to choose, I reluctantly give up the possibility of zombies. To me, this seems the least costly option; but I leave you to your own accounting. 18