Frankish Not disillusioned eprint


1  Not Disillusioned: Reply to Commentators Keith Frankish There is nothing more deceptive than an obvious fac t. (Arthur Conan Doyle, 1892, p. 80) I am grateful to the commentary authors for their c ontributions. The aim of this special issue is to give the reader a sense of the potentia l of illusionism as an approach to consciousness, and the commentators do an excellent job of this, both those who defend the approach and those who challenge it. Each comme ntary deserves a far more detailed reply than there is space for here, so I shall conc entrate on the most salient issues for points of disagreement rather than the overall evaluation of illusionism and focus on a piece, this should not be taken to mean agreement. (Thus, if I say relatively little about s reply a smoother read, I shall group that I dismiss it; quite the opposite.) To make thi advocates, explorers, sceptics, and similar commentators together, classifying them as opponents . 1. Advocates I begin with a group of commentators who offer furt her arguments in support of illusionism. Daniel Dennett provides a characteristically robust statement of the case for illusionism as the default theory of consciousness, arguing that we should thoroughly explore the mundane possibility of illusion before turning to exotic theoretical positions, especially when the latter offer few, if any, empirical predictions. I could not ng the case for illusionism, it is only agree more, and if I have been less robust in stati that the methodological principle to for tactical reasons. Of course, opponents will say which Dennett appeals is not applicable in this cas e, since illusionism denies the existence of the very thing to be explained. But th is is begging the question, which is . The explanandum is the thing we precisely whether phenomenality is real or illusory estion whether this involves call ‘conscious experience’, where it is an open qu phenomenality or the illusion of it (compare the in clusive sense of ‘consciousness’ defined in Section 1.6 of the target article). Illu sionists agree that we have a potent intuition that phenomenality is real, but they hold that the rational policy (at least given our current, rudimentary understanding of the neuro science of consciousness) is not to trust it and to pursue an illusionist research prog ramme. If the programme proves  This is the author’s eprint of an article publish ed in Journal of Consciousness Studies , 23 (11-12), 2016, pp. 256-89, and later reprinted in K. Frankis h (ed.), Illusionism as a Theory of Consciousness , Imprint Academic, 2017. It may differ in minor ways from the print version. The definitive text is 16/00000023/f0020011/art00020 . available at [v.17/4/18]

2 fruitful, then the realist intuition may loosen its grip on us (or we may loosen our grip on it). Of course, at present illusionists can do little t o make it seem plausible that this will happen. They can at best offer vague sketches of ho w the illusion of phenomenality might be generated, which are easily dismissed. But if truth is our aim, then we should t. The widespread reluctance to do this be prepared to put our realist intuition to the tes urking in the background. Perhaps suggests that there may be non-epistemic concerns l people worry that ceasing to trust the intuition wo uld erode our sense of self or our sympathy for the suffering of others, and feel that we should hold onto it regardless of its truth. Such worries are, I think, misconceived, but they deserve detailed articulation and assessment. (I shall make some brief remarks la ter, in responding to Katalin Balog.) Dennett also urges caution in framing illusionist hypotheses and warns against supposing that there is a clear-cut range of questi ons and theoretical options that can be identified in advance of detailed empirical work (as some passages in my target article may have suggested). I think these points a re wholly salutary. In his commentary, Jay Garfield attacks phenomenal realism, arguing that phenomenal properties would be unknowable, that int rospection affords no good evidence for their existence, and that belief in th em arises from mistaking properties of external objects for properties of the sensory syst ems by which we perceive them. In making these arguments he draws on Sellars and Witt genstein, but he goes on to show that similar ideas have long been present in Buddhi st philosophy. In particular, he outlines Vasubandhu’s view that it is a misconcepti on to think of experience as having dual subjective and objective aspects — a misconcep tion that yields a doubly distorted view of the causal processes involved. I am, of course, sympathetic to Garfield’s argumen ts. There are points at which phenomenal realists will want to object (arguing, f or example, that zombies do not share the same phenomenal beliefs as us and that we have a special kind of epistemic access to our phenomenal properties), but I shall not disc uss these objections here (some relevant points are made in Section 3 of the target article). Instead, I shall offer a couple of general observations. First, a comment on the nature of Garfield’s illus ionism. If I read him right, Garfield sees phenomenal realism as a purely cognitive illusion, which consists in the mistaken belief that our experiences have phenomenal propert ies. This may be correct, but there might also be a quasi-perceptual element to the ill usion. It is possible that we have sensory systems that target aspects of our brain ac tivity, and that these systems play a role in generating the illusion of phenomenality (a s proposed in Humphrey, 2011, for example). I take it that this possibility is compat ible with Garfield’s arguments, and indeed with rejection of subject/object duality. Su ch neuro-senses would be on a par with the other senses, including other body-directe d ones such as proprioception. The properties they detect would be inner only in a spa tial sense, would not be immediately and infallibly known (and would be known by zombies ), and would not be phenomenal in any substantive sense (though they might be represented as phenomenal). I don’t think we should rule out the possibility that such neuro-senses play a role in consciousness. 2

3 Second, a comment on Garfield’s discussion of Budd hist philosophy. Garfield has ntroducing them to Buddhist done Western philosophers a tremendous service in i in much of great contemporary philosophical traditions, which, as he shows, conta interest and importance (see in particular Garfield , 2015). I am not qualified to comment in detail on the points he makes, but the f act that illusionist ideas can be found cant. Illusionists are sometimes accused in ancient Buddhist philosophy is in itself signifi scientism — as if only blind science worship could prompt so meone to deny the of existence of phenomenal properties. I think this is unfair, and the fact that similar views emerged in a quite different intellectual culture l ong before the development of modern science helps to rebut it. Vasubandhu’s illusionism was the product of a long tradition of metaphysical reflection on the nature of the wor ld and our place in it, and the fact that many Western philosophers find illusionism utt erly implausible may say more about their cultural horizons than about the nature of consciousness itself. Georges Rey devotes his commentary to exploring the nature and origin of the intuition that underlies phenomenal realism. He dis tinguishes w(eak)-consciousness, which involves implementing various computational p rocesses of attention and internal awareness, and s(trong)-consciousness, whi ch involves meeting some additional, non-computational condition (these noti ons correspond to the two senses of what-it’s-like-ness distinguished in Section 1.7 of the target article). We have a powerful intuition that we have s-consciousness, bu t we have no idea what the extra presence. It is often assumed that condition might be nor any independent test for its we have rationally compelling introspective grounds for believing in s-consciousness, e condition and the known fallibility but Rey questions this. Given the elusiveness of th eally have s-conscious states, as of introspection, there is scope to doubt that we r opposed to merely having the attitudes and reaction s we associate with them. Moreover, Rey notes that we have an equally strong conviction that other people possess s- sensitive to behavioural factors consciousness, suggesting that our concept of it is l as to introspective ones. All these (perhaps including marks of biological life) as wel points are, I think, very well taken. Rey goes on to offer a Wittgensteinian diagnosis, according to which talk of ‘consciousness’ (in the strong sense), like that of ‘the sky’, has a role within a particular everyday linguistic practice, or ‘language game’, w hich cannot be smoothly integrated with science. The concept plays a useful role, refl ecting everyday needs, interests, and moral concerns, but we cannot specify its condition s of application, and it does not appear to pick out a well-defined natural phenomeno n. (I would add that the fact that we cannot specify its conditions of application mea ns that we cannot be sure that they are not wholly functional and behavioural.) I think Rey’s diagnosis is useful and that underst anding the nature and function of the concept of s-consciousness will be crucial to d eveloping the illusionist case. Our intuitions here may be put to the test in the not t oo distant future, as we create humanoid robots that have w-consciousness and exhib it a rich variety of human-like behaviour (enabling us to interact with and control them using our existing social skills and knowledge). I suspect such machines will provok e conflicting intuitions. When we concept of s-consciousness, but when interact with them, they will strongly activate our 3

4 we reflect on how they were made and how they work, we shall have a strong intuition despread scrutiny of the concept that they lack s-consciousness. This may lead to wi itself, and perhaps to its revision or replacement. Rey concludes with some cautionary remarks: some a spects of experience (especially of colour experience) seem deeply resis tant to illusionist explanation, and tenacious. It is important that the intuition that s-consciousness is real remains have explanations for specific features illusionists say these things. They do not claim to of conscious experience, or even to see how such ex planations will go. They simply claim that the illusionist programme is the most pr omising one, and that our current intuitions about what can and cannot be explained i n illusionist terms may not be reliable. Detailed empirical work may open new theo retical and conceptual options. We may never fully dispel the illusion of s-consciousn ess; it may be hardwired into our mechanisms of introspection and social perception, just as some visual illusions are hardwired into our visual systems. But recognizing that that is the case will be a major step forward. One final point: Rey notes that there is a passage in the target article where I speak of phenomenal properties, conceived as non-existent intentional objects, as being causally potent (Rey, this issue, p. 201, fn. 9). R ey dissociates himself from this view: it is the representations of non-existent intentional objects that are causally efficacious, and talk of the objects themselves having certain e ffects is merely a convenient shorthand. In fact, I agree with Rey on this; the p assage in question was loosely phrased. Amber Ross menal realism. highlights some epistemological problems for pheno Real properties are independent of our beliefs abou t them (reality, in Philip K. Dick’s t, doesn’t go away’ (Dick, 1995, p. words, ‘is that which, when you stop believing in i 261) — and, we might add, real properties are ones that don’t come to be there just because you think they are). If phenomenal properti es do not exhibit this sort of belief- independence, then the natural conclusion is that t hey are not real but merely e content of a fiction or hallucination. intentional objects of our representations, like th As Ross puts it: Any view according to which the subject’s beliefs a bout the character of her conscious experience do play a role in determining the facts of the matter about her conscious experience is a non-realist, illusion ist type of view. (Ross, this issue, p. 221) to describe a plausible scenario in which Yet, as Ross shows in some detail, it is very hard a subject has a false belief about the phenomenal c haracter of an experience they are currently attending to. Of course, it wouldn’t exac tly help the realist if we could construct such a scenario; for, as Ross notes, we h ave a strong intuition that we cannot make this kind of mistake (this issue, p. 219). In this respect, then, illusionism is better placed to account for the common-sense view of cons ciousness than phenomenal realism. I think the line of attack Ross pursues — question ing the coherence of phenomenal and one that was perhaps insufficiently realism — is an important one for the illusionist, 4

5 stressed in the target article. It is, of course, a line that Dennett has pressed with ses is that of change blindness considerable force over the years. One example he u ce repeated shifts of colour in an (Dennett, 2005, chapter 4). People can fail to noti image, provided each presentation of the image is s eparated by a brief masking stimulus. In such cases, the colour shifts must be registered at some level by the subject’s visual system, but do they show up in their visual phenome nology? Dennett argues that realists face a dilemma: if they were to experience change blindness themselves, what 1 If they would say that their would they say about their own phenomenology? phenomenal properties changed without their noticin g it, then they must accept that we are not authoritative about our phenomenal prope rties and that, for all we know, they may change all the time without our noticing. If they say that their phenomenal properties did not shift until they noticed the cha nge in the image (that is, registered it cognitively), then it looks as if our phenomenal pr operties are simply constructions out of our judgments, as illusionists claim (and if the y say they don’t know if their phenomenal properties shifted, then it is unclear w hat could possibly settle the matter). The upshot, Dennett concludes, is that the notion o f a phenomenal property is simply a mess, a source of nothing but confusion. Ross’s c ontribution illustrates the force of considerations like this, which are, I think, still widely underestimated. James Tartaglia advocates a surprising position: non-physicalist i llusionism. I did not consider the possibility of such a position in the target article, since I was concerned tegy, but of course illusionism does with illusionism as a conservative explanatory stra not entail sciousness while holding physicalism. One could be an illusionist about con Tartaglia argues that illusionism that reality is fundamentally non-physical. Indeed, actually provides grounds for holding that. In the first part of his commentary, Tartaglia att acks ‘intermediate’ positions, which attempt to combine physicalism with phenomenal real ism. Such positions typically rely argues, this reliance is unwise, since on the phenomenal concept strategy, but, Tartaglia phenomenal concepts aren’t simply neutral ones, whi ch do not present their objects as physical, but substantive ones, which present their objects as having a qualitative, subjective nature that no physical property could h ave. If physicalism is true, then phenomenal concepts must misrepresent their objects : So if the phenomenal concept is a concept of a brai n state, it must be a radical misconception of it; we must be misconceiving the b rain state beyond all recognition, in fact. We are thinking of a brain st ate as a subjective experiential array, but that is not what it is at all. Consequen tly, the array must be an illusion, even if thinking about it somehow allows us to think about real brain states. (Tartaglia, this issue, p. 238) Tartaglia has no sympathy with dualist or panpsychi st explanations of consciousness, and he accordingly adopts an illusionist position. This line of argument is, of course, one that I endorse, and Tartaglia’s presentation of it is elegant and compelling. 1 Dennett uses the term ‘qualia’, but for consistenc y I’ll put the point in terms of phenomenal properties. 5

6 In the rest of his commentary Tartaglia turns to t he metaphysical implications of uld be compatible with our manifest illusionism. He points out that our metaphysics sho cally with the fact that we seem to situation — with how things seem to us, and specifi be confronted with arrays of phenomenal properties. Physicalist illusionists explain our manifest situation by appealing to our judgments an d representations: things seem that em to be. Tartaglia thinks this has way because that is how we are inclined to judge th argues that it means we cannot be profound epistemic consequences. In particular, he confident in our physical conception of reality, si nce that conception was created on the basis of illusory experience. He is not suggesting that we should doubt our science. By taking experience as a guide to reality, we have bu ilt up a coherent and detailed picture of an objective world — a picture that has eventual ly led us to the hypothesis that experiences themselves are illusory. But, he argues , we should not forget our starting point: experience itself has a reality that must be accounted for. Hence, we must supplement our scientific picture of reality with a distinctively philosophical account. There must be an independent reality behind our exp erience, which transcends the objective world in the same way that the objective world transcends the world of a dream (however, Tartaglia denies that this independ ent reality is mental; his view is not a form of idealism or phenomenal realism — this iss ue, p. 251, fn. 10). What should we make of this? Does illusionism requ ire us to reject physicalism? I am unpersuaded. I place myself in the tradition of Quinean naturalism, which (as Tartaglia notes) denies a sharp distinction between philosophy and science and holds picture of reality. At any rate, I do that science, broadly construed, provides our best not see how positing a transcendent reality could s hed any further light on the nature lf says that we can say ‘nothing of consciousness and subjectivity. (Tartaglia himse p. 250.) Tartaglia worries that substantive’ about independent reality; this issue, physicalist illusionism renders many other aspects of our manifest situation illusory too, including our sense of being spatio-temporally loca ted. Even if this were so (and I’m ect physicalism. If certain illusions are not sure it is), I do not see it as a reason to rej reating them as enabling fictions, important to us, we can continue to live by them, t which do not need metaphysical underpinning. Moreover, I think Tartaglia overestimates the nega tive epistemic implications of illusionism. He suggests that it renders our judgme nts about our experiences ‘completely unreliable’ (this issue, p. 244). But t his is too swift. Illusionism does not claim that our conscious experiences are wholly ill usory, only that their apparent phenomenal aspect seeing a red postbox is. I may be correct to judge that I am currently (where red is a reflectance property of surfaces), even though I’d be wrong to judge that the experience has a reddish phenomenal feel. Tarta glia asks how we can correlate experiences with worldly properties, but I fail to see the problem. Evolution has set up the correlations, designing our perceptual systems to reliably track worldly properties (perhaps disjunctive, gerrymandered ones), and by d oing science we can get a better understanding of the nature of those properties. Ta rtaglia doubts that we can ‘cherry- pick’ experience for veridical elements, but that i s just what the scientific method has enabled us to do. Even if it is not the fundamental reality, the objective world has a g the scientific method we have complex structure independent of us, and by applyin 6

7 acquired a powerful grip on that structure. In the process, we have come to question endorse some form of some of the beliefs we started with, but unless we foundationalism, this is not a problem. In the same vein, illusionists need not deny that phenomenal concepts represent real properties, albeit under distorted guises. Tar taglia finds it implausible that brain states or of distal objects, phenomenal concepts represent properties either of e too radical (this issue, pp. 244– arguing that the misrepresentation involved would b 5). Again, I fail to see the worry. A concept may t rack a certain property even if it radically misrepresents it. Recall Humphrey’s examp le of the Penrose triangle, discussed in the target article (Frankish, this iss ue, p. 17). Deployed in visual experience, the concept of a Penrose triangle tracks a certain sort of three-dimensional structure (a ‘Gregundrum’), which it represents as a physically impossible object. The misrepresentation involved is radical yet perfectly possible. Moreover, it may be useful if we need to distinguish Gregundra from non-Gregun dra. Gregundra are simply objects that create the illusion of a Penrose trian gle, and the best way to tell if an object is a Gregundrum is to see if it creates the illusio n — if our visual system misrepresents it as a Penrose triangle. It would be impossible to develop a veridical perceptual concept that reliably distinguishes Gregundra from all the other similar structures that do not create the illusion. Something similar may be the c ase with phenomenal concepts. They may pick out highly disjunctive physical properties (either of our cognitive systems or of distal objects) which it is useful for us to tra ck but which are unified only by the fact that they trigger the concept. The fact that they r adically misrepresent their objects is no bar to their performing this function. Perhaps my Quinean sympathies — or my lack of what Tartaglia calls ‘historically- s issue, p. 252) — are blinding me informed metaphilosophical self-consciousness’ (thi the science of consciousness goes, he to Tartaglia’s deeper point. At any rate, as far as and I are in agreement: we should adopt an illusion ist view. 2. Explorers I turn now to four commentators I have dubbed . These use their explorers commentaries to explore ways of developing illusion ism — either building theories, responding to objections, or reviewing experimental evidence. Their papers illustrate how illusionism can form the core of a research pro gramme, which can be supplemented and developed in different ways. François Kammerer plaining addresses the illusion problem — the problem of ex how the illusion of phenomenality arises. As he not es, the problem has a particularly hard aspect. It is not just that we are strongly di sposed to think that phenomenality is not an illusion; we find it hard to understand how it could be an illusion: [W]hat makes us reluctant to accept illusionism is not only that we are disposed to believe that we are conscious, it is also that w e have difficulties making sense of the hypothesis that we are not conscious while i t seems to us that we are . (Kammerer, this issue, p. 127) 7

8 enality apart from all other illusions This, Kammerer argues, sets this illusion of phenom em. and means that it cannot be usefully modelled on th 2 Kammerer proposes a solution. Simplified somewhat, it runs as follows. Introspection is informed by an innate and modular theory of mind and epistemology, which states that (a) we acquire perceptual informa tion via mental states — experiences — whose properties determine how the world appears to us, and (b) experiences can be fallacious, a fallacious experience of A being one in which we are mentally affected in the same way as when we have a veridical experience of A, except that A is not present. Given this theory, Kammerer notes, it is incoherent to suppose that we could have a fallacious experience of an experience, E. For that would involve being mentally affected in the same way as when we have a veridical experie nce of E, without E being present. But when we are having a veridical experience of E, we are having E (otherwise the experience wouldn’t be veridical). So, if we are me ntally affected in the same way as when we are having a veridical experience of E, the n we are having E. So E is both present and not present, which is contradictory. (K ammerer couches the argument in terms of experiences, but it could easily be recast in terms of the phenomenal properties of experience. Having a fallacious experience of a phenomenal property involves being mentally affected in the way one would be if the pr operty were present, which involves it being present. Generalized further, this argumen t might explain our sense that introspection is infallible.) Kammerer proposes tha t this explains the peculiar hardness of the illusion problem. The illusionist thesis can not be coherently articulated using our everyday concept of illusion, which is rooted in ou r naïve concept of fallacious sketches does inform our experience. Moreover, if the naïve theory Kammerer introspective activity, then we shall not be able t o form any imaginative conception of what it would be like for illusionism to be true. H ence the common claim that, where consciousness is concerned, appearance is reality. As Kammerer stresses, this does not mply means that in order to state it we mean that illusionism actually is incoherent. It si ay, a cognitively impenetrable, non- must employ a technical concept of illusion — as, s veridical mental representation that is systematica lly generated in certain circumstances. Kammerer’s approach to the illusion problem is, I think, a promising one, and the idea that introspection is theoretically informed i s likely to figure prominently in any developed illusionist theory. Of course, even if Ka mmerer is right about the source of our intuitive resistance to illusionism, this would not show that illusionism is true, though it would help to dispel one common objection to it. Realists will say that phenomenality is not an illusion even in a technica l sense: our relation to our phenomenal properties is one of direct acquaintance , which does not depend on potentially fallible representational processes. Pe rhaps Kammerer could employ the strategy again here, arguing that our concept of in trospective acquaintance is also a theoretical one. At any rate, considerations like t his should help to move the debate forward, beyond the simple assertion that illusioni sm is unintelligible. 2 ope I have captured the core of his argument. I have omitted a lot of Kammerer’s detail, but I h 8

9 Derk Pereboom table approach has done much to establish illusionism as a respec ed illusionist theory (the ‘qualitative to consciousness, setting out a carefully articulat sed to rebut standard anti- inaccuracy hypothesis’) and showing how it can be u physicalist arguments (see Pereboom, 2011). In his commentary, he discusses the form an illusionist theory should take, challenging the functionalist view I suggested in the erns what illusionists should say about target article. He makes two points. The first conc l object that illusionists are still illusions of phenomenality themselves. Realists wil committed to phenomenal realism, since there is som ething it is like to have the illusion of a phenomenal property. In the target article, I suggested that illusionists should deny that phenomenal illusions themselves seem to have p henomenal properties. Pereboom thinks this won’t do, and argues that we should ins tead explain their apparent phenomenality as a further illusion. If introspecti on misrepresents quasi-phenomenal states as phenomenal, then it can misrepresent our modes of presentation of those states as phenomenal too. This need not create a regress, Pereboom argues, since there is no reason to think that we also represent those higher -order modes of presentation, or at least that we do so under phenomenal modes of prese ntation. It is good to have this proposal on the table. I t hink it is a coherent position, and I agree with Pereboom that the regress objection is n ot serious (the point to stress is that mental states seem to possess phenomenal properties only when introspected, and psychological limitations on the introspection proc ess will naturally block the regress). It is in some ways a puzzling proposal, however. Wh y should introspection represent modes of presentation as having the same properties as the states they represent? Why should the representation of an experience feel lik e the experience itself? Moreover, we may not need to posit higher-order introspective pr ocesses in order to account for our ves have phenomenal properties. sense that illusions of phenomenality would themsel Recall Kammerer’s proposal about the theory-laden n ature of introspection. If Kammerer is right, then when we try to conceive of an introspective illusion we shall iginal experience, with all its conceive of a mental state that incorporates the or (apparent) phenomenal properties. The apparent high er-order feel may simply be an ction. artefact of the innate theory that informs introspe si-phenomenal properties. Pereboom’s second point concerns the nature of qua Introspection represents these as intrinsic propert ies rather than functional ones. Illusionism removes the pressure to think of them i n this way, allowing us to incorporate them smoothly into a functionalist acco unt of the mind. But, Pereboom argues, illusionism doesn’t require us to adopt a functionalist view. We could regard quasi-phenomenal properties as consisting, at least absolutely intrinsic partially, of aptnesses , which form the categorical bases for the causal p owers of physical entities. Pereboom argues that this view not only vindicates our common-sense intuition that phenomenal properties are intrinsic causal powers b ut also gives the illusionist a stronger response to standard anti-physicalist argu ments. Again, it is good to have this view on the table, though personally I find it unpersuasive. Of course, if one thinks that all cau sal powers are ultimately grounded in absolutely intrinsic aptnesses, then one will think that the powers of quasi-phenomenal reasons for thinking of properties are too. But I don’t think there are specific 9

10 consciousness in this way. As Pereboom acknowledges , we do not need to posit he anti-physicalist arguments, and our absolutely intrinsic properties in order to rebut t can be explained as a sense that phenomenal properties are intrinsic ones misrepresentation. As illusionists, we do not need the heavy metaphysical machinery of absolutely intrinsic aptnesses in order to explain why conscious experiences seem to be my mind, bring illusionism intrinsic causal powers, and employing it would, to uncomfortably close to a form of Russellian monism. I turn now to two contributions from scientists. M uch scientific work on consciousness has been conducted, wittingly or not, in a quasi-dualistic spirit. Theorists seek to identify the neural processes that produce consciousness, without offering any explanation of how they produce it. This isn’t surprising if consciou sness is conceived in a realist way: it is impossible to gain any expl anatory purchase on such a nebulous phenomenon. But illusionism provides a much more tr actable target for scientific investigation. To explain consciousness we need to identify and explain the (broadly representational) processes that collectively const itute the illusion of phenomenality. The two commentaries considered next adopt this per spective. Michael Graziano provides a clear introduction to his attention schema theory , according to which consciousness depends on possess ion of an internal model of one’s attentional processes. Graziano’s conception of the explanandum for a theory of consciousness is thoroughly illusionist. As he expl ains: processing information, Here by ‘consciousness’ I mean that, in addition to people report that they have a conscious, subjectiv e experience of at least some a specific explanation for of that information. The attention schema theory is uman machine claims to how we make that claim... It is a theory of how the h have consciousness and assigns a high degree of cer tainty to that conclusion. (Graziano, this issue, p. 98) The aim is to explain our sense that we are conscio us, rather than consciousness itself as a distinct property. This sense arises, Graziano argues, from the fact that, in addition to representing features of the world and of oursel ves, we represent our mental relation to things via attention. We have an ‘attention sche ma’, which models covert attention (the deep processing of selected information), allo wing us to monitor and control it. This model does not provide a detailed representati on of the mechanisms involved; rather, it represents attention in an abstract, sch ematic way, as a sort of private mental possession of something. As a result, when we introspect our attentional processes we seem to find an inner world where a subjective self has an immediate grasp of the properties of things, leading us to issue reports l ike this (which Graziano puts into the mouth of a robot equipped with an attention schema) : ‘my mental possession of the apple, the mental poss ession in-and-of-itself, has no physically describable properties. It’s an essen ce located inside me... It’s my mind taking hold of things — the colour, the shape, the location. My subjective self seizes those things.’ (Graziano, this issue, p p. 102–3) 10

11 When we talk of consciousness and its features, we are reporting the deliverances of our imental evidence that attention schema. Graziano goes on to outline exper ons of the attention schema, as the consciousness is associated with the control functi theory would predict. I shall not attempt to assess attention schema the ory here but simply comment on its relation to the illusionist programme. Graziano notes the affinity (especially with Dennett’s views) but argues that the term ‘illusion ism’ has misleading connections. To describe consciousness as an illusion suggests that it is nothing at all and that introspection is simply in error. But, he points ou t, covert attention itself is real, and our internal model of it, though schematic and abst ract, is well adapted for its function of tracking and controlling attention. As he puts i t, ‘consciousness is not an illusion but a useful caricature of something real and mechanist ic’ (this issue, p. 112). I think these are excellent points, and they indic ate the need for an important thin illusionist theorizing. On the one clarification. Talk of illusion does double duty wi hand, it may refer to quasi-perceptual introspective representations generated by self- monitoring processes, such as the attention schema. These representations may be highly abstract and distorted, and in that sense il lusory, but they may also carry valuable information for the system and facilitate important tasks of control and self- manipulation. An illusion need not be a fault and m ay have been carefully designed (compare Dennett’s analogy with the ‘user illusions ’ produced by the icons and pointers on a computer desktop — Dennett, 1991). On the othe r hand, illusion talk may refer to the cognitive illusion involved in judging that we are acquainte d with an internal world of intrinsic phenomenal properties. Here it is appr opriate to talk of error (certainly in theoretical contexts), though perhaps still not of a fault: belief in the metaphysical ying an important role in human specialness of our inner lives may be adaptive, pla psychology and social interaction (Humphrey, 2011). These illusions, quasi-perceptual and cognitive, a re of course closely related; we judge that we are acquainted with phenomenal proper ties because introspection gives us such a partial view of internal reality (indeed, natural selection may have sculpted our neural processes in order to create the cogniti ve illusion; ibid .). Phenomenal consciousness, we might say, is a theoretical illus ion built on an introspective caricature. In their commentary, Nicole Marinsek and Michael Gazzaniga look at illusionism from the perspective of split-brain research. Patie nts who have undergone surgical severing of the corpus callosum display various beh avioural dissociations, which suggest that each hemisphere is operating as a sepa rate mind. This presents a challenge for illusionism. Both hemispheres show signs of bei ng phenomenally conscious (in the everyday sense), so if phenomenality is an introspe ctive illusion, then both must possess a capacity for introspection and be susceptible to illusions. Marinsek and Gazzaniga review relevant experimental evidence and tentative ly conclude that this is indeed the case. One moral of this, they suggest, is that, eve n without callosotomy, phenomenal consciousness may be fragmented, comprising numerou s ‘modular illusions’ with different characteristics. I think these points are well taken, and, as Marin sek and Gazzaniga note, the split- nd for detailed illusionist proposals (it brain literature will provide a useful testing grou 11

12 would be interesting to explore its implications fo r attention schema theory and for stion that consciousness may be Humphrey’s ‘sentition’ theory). Moreover, the sugge the split-brain literature has shown fragmented is, I think, an important one. One thing is that our sense of psychological unity can be ill usory: despite the dissociations in their behaviour, split-brain patients continue to feel un ified, and they unconsciously we conceive of subjecthood in non- confabulate to preserve that feeling. Of course, if ce with phenomenal properties, then psychological terms, as involving direct acquaintan it is hard to see how we can establish any objectiv e criteria for identifying conscious subjects. But illusionism provides a much more trac table approach. To be a conscious subject is (putting it very sketchily) to be a syst em that produces appropriate introspective representations of its own mental act ivity and uses them to modulate its activity in appropriate ways. In this sense, we may each incorporate multiple conscious or semi-conscious subjects, either modular or parti ally integrated with each other. 3. Sceptics This section looks at contributions from four comme ntators who, although not full- blown opponents of illusionism, express reservation s about the position or feel that it is in some way misguided. Susan Blackmore distinguishes illusionism from a more cautious vie w, which she calls delusionism . Whereas illusionists deny the existence of phenom enal consciousness outright, delusionists hold that we have many mista ken theories about it. Blackmore expresses reservations about illusionism, but she e ndorses delusionism, arguing that we are wrong to think that there is a stream of consci ousness, with rich, unified, and observes it. Whenever we introspect, determinate contents, and a persisting self, which we always find some conscious experience, and this leads us to think that there is a continuous inner stream of such experiences and an inner self waiting to observe them. But these claims, she argues, are baseless — neithe r neuroscience nor careful conscious experiences are present introspection offers any way of determining whether ther, there are just moments of at times when we are not actively introspecting. Ra consciousness, temporary constructions bonding thou ghts and perceptions to a representation of the self. This is a valuable piece, which usefully summarize s ideas that Blackmore has defended at length in earlier work. There are many important issues here, but I shall confine myself to commenting on the relation betwee n delusionism and illusionism. Illusionism clearly entails delusionism: if there a re no phenomenally conscious experiences, then there is no continuous stream of them either. Could there be a continuous illusion of consciousness? It depends on what kind of illus ion we are thinking of. If it is a personal-level cognitive on e, which occurs when we actively introspect and judge that we are currently having a n experience with such-and-such phenomenal properties, then the answer is obviously no. But illusionists might want to say that there is a continuous subpersonal illusion , or something like it, consisting in the production of abstract, quasi-perceptual repres entations of neural processes, which rm the basis for our phenomenal are used for internal control purposes and which fo 12

13 judgments when they occur (perhaps Graziano’s atten tion schema theory supposes something like this). What about the converse? Does delusionism entail i llusionism? Blackmore thinks it does not. She does not endorse illusionism and seem s to accept the reality of immediate conscious sensations. Prima facie this seems right — there could be moments of m of it. However, I am not sure phenomenal consciousness without a continuous strea then there are properties of one’s brain this position is stable. If there are such moments, state at those moments that make it phenomenally co nscious — physical properties, let us assume. But then it should be possible, at least in principle, to determine whether our brain states have these properties at times when we are not introspecting, and thus to determine whether or not there is a stream of consc iousness. If delusionists deny that this is possible, then, it seems, they should deny that there are such properties and 3 accept that phenomenal consciousness does not exist . Blackmore closes her commentary by suggesting that our delusions of consciousness are malign memes, which we can, with effort, rid ourselves of. I am unsure about this. It may be true that our concepti ons of consciousness and the self are culturally shaped, though rooted in the deliverance s of real introspective processes. However, they may not be malign — they may play val uable social and psychological roles, as Humphrey has argued (Humphrey, 2011). As we understand more about why we conceptualize our inner lives in the way we do, we should gain more purchase on these questions, perhaps with beneficial practical consequences. cterizing Nicholas Humphrey uses his commentary to question the value of chara consciousness in terms of illusion. In the past, Hu mphrey has proposed an explicitly periences reflect internalized illusionist theory, according to which conscious ex h incoming sensory signals to expressive responses to stimuli, which interact wit generate complex feedback loops. When these loops a re internally monitored, Humphrey argued, they appear to possess strange qua litative and temporal properties, ibid. ). creating the illusion of a magical inner world ( e label ‘illusionist’ and insists In his commentary, however, Humphrey repudiates th that his view is better characterized as a realist or ‘surrealist’ one (though not, he stresses, in any anti-physicalist sense). He offers two reasons for this. One is tactical: to characterize one’s view as the claim that phenomena l consciousness is an illusion is to invite people to ignore or ridicule it; it’s ‘bad p olitics’ (this issue, p. 122). I shall discuss this worry in a moment. Humphrey’s other, more subs tantive, point is that sensations represent something real and important — namely our evaluative responses to stimuli: [W]hen considering whether sensations are or are no t ‘real’, we must never let go of the fact that sensations do indeed represent our take on stimuli impinging on the body. In doing so they represent some of the objective facts about what’s happening: the what, where, and when, for example. But, crucially, they also 3 There are passages in Blackmore’s commentary which suggest that her sympathies are more illusionist than she admits. She writes, for exampl e, that neuroscientists ‘will never find the neural correlates of an extra added ingredient — “consciou sness itself” — for there is no such thing’ (this i ssue, p. 61). 13

14 represent how we what’s happening, how we feel about it. And this is evaluate nsations represent how where phenomenal properties come into their own. Se box of phenomenal concepts we relate to stimulation using, as it were, a paint to depict what it’s like for us. (Humphrey, this is sue, p. 118) mulation: how we are being stimulated Sensations, he argues, represent two aspects of sti (the objective side) and how we respond to the stim ulation (the subjective side). Their phenomenal aspect corresponds to the latter — it re presents our subjective take on stimulation. And this aspect, Humphrey argues, cann ot be illusory or nonveridical: How could you... be experiencing a feel that ‘doesn’t exist’? To be blunt, I think the very notion of this is absurd. When the sensati on represents you as feeling a certain way about the stimulation, that is all there is to it. The phenomenal thereby its existence becomes a fact . feel arises with the representation, and (Humphrey, this issue, p. 119) There are two ways of reading this. On one, Humphre y is making a point similar to Graziano’s: sensations are not mere illusions but r epresentations of something real and important — our evaluative responses to stimuli, wh at they mean for us. (It is interesting that both Graziano and Humphrey hold th at consciousness is based in a dynamic relation rather than passive awareness. Den nett makes a similar point in his commentary.) This reading is compatible with illusi onism in my sense. For phenomenal properties may still be illusory. It may be that sensation mi srepresents our evaluative s of efferent neural activity) as responses (which are constituted by complex pattern simple intrinsic phenomenal feels — that it tells u s how we feel in the language of fault in sensation. The distortion phenomenal fictions. Again, this need not imply any may be necessary to achieve the effect; the represe ntation of a huge swathe of neural ntation of phenomenal pain, just as activity wouldn’t have the same impact as a represe a pile of sociological reports on parent–child rela tions wouldn’t have the same impact King Lear . The analogy is with a skilled magician producing as a performance of nating an oasis. (I would add that, pace astonishing effects, not a desert traveller halluci Humphrey, genuine error may be possible on the subjective side as we ll as the objective one. Introspection may sometimes go awry, represent ing the presence of an internal response that has not in fact been triggered.) On the other reading, Humphrey is claiming that re presentations of our evaluative create or constitute responses to stimuli phenomenal feels as distinct properties. This, of course, is not an illusionist position but a rea list one. There are places in the commentary where Humphrey appears to endorse this v iew, describing phenomenal properties as an ‘inherent feature’ of brain activi ty (this issue, p. 120). On the whole, however, I think the illusionist rea ding is more accurate. When Humphrey defends his realism, it is the reality of our relation to stimuli that he stresses, not the reality of phenomenal feels themselves — wh ich are, after all, usually characterized as nonrelational properties (this iss ue, pp. 119–20). And when Humphrey resentation’ he may mean that they talks of phenomenal feels ‘aris[ing] with their rep 14

15 are intentional objects, which are real — part of their represented inner for the subject llusionist at heart: introspection uses a world. I suspect, then, that Humphrey is still an i ain internalized responses that are ‘paintbox of phenomenal concepts’ to represent cert not really phenomenal. This chimes well with his ad option of the term ‘surrealism’. Surrealist paintings distort reality, albeit in a c reative, expressive way. s adopt a less provocative label for Back to the bad politics worry. Should illusionist rable connotations (any single-word their position? ‘Illusionism’ does have some undesi label would have — ‘magicism’ would be even worse!) . And, as I have stressed, we can employ an inclusive concept of consciousness that d oes not carry a commitment to phenomenal realism and allows us to affirm the real ity and significance of consciousness in a natural way. But as a term for a theoretical approach to consciousness, I prefer to stick with ‘illusionism’ — at least at this stage in the debate, where the ghost of phenomenality has yet to be exor cised from cognitive science. It is all too easy to adopt a conception of what needs to be explained that encourages scientists and philosophers to ask bad questions an d to ignore good ones. In my view, we need to challenge this misconception head on. Th ere’s no point mincing words: we don’t have phenomenal properties, only representati ons of them. Pete Mandik sm, agreeing that expresses a sympathetic scepticism about illusioni we do not have phenomenal properties but denying th at we are under the illusion of having them. He makes two points. First, the term ‘ phenomenal’ (as used in this context) has no clear content. Attempts to define i t cycle uninformatively through a series of synonyms, and illusionists won’t want to rely on private introspective ostension to explicate the concept. Mandik’s second worry concerns the notion of illusion. Illusions are systematic: it is appropria te to talk of illusion only when a certain erception or fallacious judgment in stimulus or scenario reliably evokes a certain misp is supposed to be that when we people. In the case of consciousness, the illusion introspect our experiences they seem to possess ano malous, inexplicable properties. But, Mandik notes, introspection elicits this judgm ent in few people outside philosophy departments — most would say their experiences seem perfectly mundane and natural. Mandik is tempted to call his position but settles on qualia quietism : meta-illusionism questions about qualia, or phenomenal properties, a re simply not well enough defined to be worth pursuing. I am sympathetic to Mandik’s quietism (which echoe s points Daniel Dennett has made), and to some extent I think the difference be tween us is one of emphasis. However, I think Mandik overstates his case. Take ‘ phenomenal’ first. The concept is typically introduced via a sort of language game (c all it ‘the phenomenality language game’), which involves a combination of inner osten sion (think of how pain feels, coffee 4 reflection on the appearance/reality distinction ( where is the colour of an smells, etc.), after-image located?), thought experiments (imagine inverts and zombies), and scientific knowledge (science tells us that colours are really ‘in’ us), supplemented with theoretical claims (phenomenal properties are ineff able, intrinsic, radically private, and 4 Illusionists can engage in inner ostension, though they will take the objects identified to be merely intentional. 15

16 so on). This game isn’t played only by philosophers ; many of the moves are widely em for themselves. I don’t claim known, and children spontaneously invent some of th t is not contentless either, and I think it that the resulting concept is fully coherent, but i 5 Moreover, I is meaningful (and important) to deny that it picks out something real. concept. It is sometimes argued that think there is a genuine introspective basis to the experience is wholly transparent — that we are awar e only of aspects of the external world (including our bodies). But it is plausible t o think that experience tells us more than this. As several commentators have argued, con scious experience also tells us about our relation to worldly properties — how we feel about them (Hu mphrey, this issue), or how we are attending to them (Graziano, this issue), or what expectations and reactions they evoke in us (Dennett, 2013). Introsp ection represents these relations under highly abstract, caricatured guises (creating what Dennett calls the user illusion ), but in doing so it provides a substantive, though m isleading, content to the notion of phenomenality. Second, what about illusion? As I’ve mentioned, il lusion talk does double duty — referring both to quasi-perceptual introspective re presentations of the sort just mentioned and to the cognitive illusion involved in judging that introspection acquaints us with phenomenal properties (and that t hese properties are anomalous or magical). The former might be better described as a caricature rather than an illusion, but I think the latter deserves the title. It is tr ue, as Mandik observes, that mere introspection does not elicit these judgments; one needs to have been inducted into the relevant language game. But with that induction (wh ich is common), and a little ealism and judge that it presents a reflection, people do reliably endorse phenomenal r nty Hall problem, which Mandik hard problem. The case is similar to that of the Mo cites as an example of a cognitive illusion. In ord er to fall for the illusion, one needs some prior (albeit imperfect) grasp of the concept of probability. Finally, a point about quietism. From a tactical p oint of view, I think, quietism is not the best approach for the qualia irrealist. Peo ple easily fall into dualist ways of thinking, which lead them to ask bad questions abou t consciousness. To counter this, the irrealist needs to provide a robust explanation of why we are susceptible to dualist intuitions and why they seem so compelling. Simply telling them not to talk about qualia won’t do. There is a Bob Newhart sketch in w hich he plays a psychiatrist whose only advice to a neurotic patient is ‘Stop it!’, re peated over and over. Quietism is a bit like that. It may be sound advice, but it’s not ver y helpful therapeutically. Eric Schwitzgebel also considers the problem of defining phenomenal consciousness, responding to the target article’s c hallenge to identify a notion of phenomenal consciousness that is substantive yet fr ee of dubious theoretical commitments. He proceeds by offering a definition b y example, describing a range of uncontentious positive and negative cases and ident ifying phenomenal consciousness as ‘the most folk-psychologically obvious thing or feature that the positive examples possess and that the negative examples lack’ (this issue, p. 229). Because it relies on 5 I do, however, doubt that there is any content to the weaker, ‘diet’ conception of qualia sometimes enomenality language game — see Frankish (2012). proposed, which is supposedly independent of the ph 16

17 examples and does not adjudicate on contentious cas es, this definition is a theoretically room for puzzlement about the nature innocent one, yet it is not deflationary and leaves of consciousness. I think Schwitzgebel succeeds in identifying an im portant folk-psychological kind — indeed the very one that should be our focus in t heorizing about consciousness. the target article. For, precisely However, I don’t think he has met the challenge of compatible with illusionism. As I because his definition is so innocent, it is not in stressed in the target article, illusionists do not deny the existence of the mental states we describe as phenomenally conscious, nor do they deny that w e can introspectively recognize these states when they occur in us. Moreo ver, they can accept that these states share some unifying feature. But they add that this feature is not possession of phenomenal properties (qualia, what-it’s-like-ness, etc.) in the substantive sense created by the phenomenality language game. Rather, it is p ossession of introspectable properties that dispose us to judge that the states possess phenomenal properties in that substantive sense (of course, we could call this fe ature ‘phenomenality’ if we want, but I take it that phenomenal realists will not want to do that). Now, the challenge of the target article was to articulate a concept of pheno menality that is recognizably substantive (and so not compatible with illusionism ) yet stripped of all commitments incompatible with physicalism. Schwitzgebel hasn’t done this, since his conception is not substantive. Nevertheless, Schwitzgebel has succeeded in someth ing perhaps more important. He has defined a neutral explanandum for theories o f consciousness, which both realists and illusionists can adopt. (I have referred to thi s as consciousness in an inclusive sense. consciousness , We might call it simply , or, if we need to distinguish it from other forms putative phenomenal consciousness .) In doing this, Schwitzgebel has performed a valuable service. 4. Opponents f illusionism. Each makes This section responds to four papers by opponents o important points, which deserve more discussion tha n there is space for here, but I shall indicate the general lines of reply that I favour. As I stressed in the target article, my aim is not to refute alternative positions but simply t o establish the attractions of illusionism. I begin with two commentators who write from a cons ervative realist perspective. Katalin Balog enal realism. mounts a forthright defence of common-sense phenom In particular, she argues that explanatory gap cons iderations do not give physicalists reason to prefer illusionism, since they can be exp lained by a version of the phenomenal concept strategy. Specifically, Balog proposes that direct phenomenal concepts are partly constituted by the experiences they refer to and refer to them in virtue of this fact: phenomenal states serve as their own modes of prese ntation. This, she argues, gives us a direct and substantial access to our phenomenal s tates, which is very different from the access science gives us and creates the impress ion of an explanatory gap. I indicated my misgivings about this strategy in the target art icle (see also Tartaglia’s commentary here. in this issue), but I shall add a few more comments 17

18 First, it is not clear how constitutive self-refer ence is supposed to work. A concept erring to it, and Balog does not explain may be partially constituted by a state without ref enomenal concepts. More what further factors are involved in the case of ph importantly, at best the account explains why we th ink phenomenal states could be non - be physical. To see this, it is enough physical; it does nothing to explain how they could or a non-phenomenal state, such as to note that we might have a concept of this kind f a propositional attitude. The concept might represe nt the state in a substantial and direct way, opening a potential gap between facts a bout it and the neural facts (indeed illusionists might accept that phenomenal concepts represent sensory states in this way). Yet we might feel no resistance to the idea t hat the state represented is physical — nothing in our grasp of it need its physicality or make it difficult to conceive o f. rule out Yet that is just what we feel about phenomenal stat es — we can’t understand how they could be physical, and the phenomenal concept strat egy sheds no light on the matter. Moreover, Balog makes it clear that she thinks that experiences possess introspectable phenomenal character when they are not being target ed by phenomenal concepts: Thinking about [an experience] and simply having th e experience will then share something very substantial, very spectacular: namely the phenomenal character of the experience. (Balog, this issue, p. 45) is independent of our phenomenal But if we have a grasp of phenomenal character that concepts, then we cannot explain away its puzzling features by reference to those raise all the familiar questions. What concepts. Simple acquaintance will be sufficient to is this ‘very substantial, very spectacular propert y’? How do brain processes generate it? Why is it not detectable from other perspectives? A nd how can we be aware of it when we are not representing it to ourselves? Balog raises other objections to illusionism. It i s, she says, ‘utterly implausible’ (p. mental ways the world presents itself to 42). It ‘flies in the face of one of the most funda dangerous — scientism (p. 42). I us’ (p. 47) and manifests a misguided — and perhaps am not unsympathetic to Balog’s worries, but I thin k they are unfounded. As we have seen, illusionists do not deny the existence of con sciousness in the innocent sense defined by Schwitzgebel; they merely offer a differ ent account of its nature. And here, as Balog puts it, ‘the question comes down to the e pistemic authority accorded to introspective awareness vs. scientific theorizing’ (p. 47). In developing a theory of consciousness, I plump for the latter. We have abun dant evidence of the unreliability of introspection, and there is no reason why an evolve d cognitive system should represent its internal states to itself in a transparent way, as opposed to an adaptively useful one. I do not accept that this view manifests a scientis tic attitude. My motive for adopting it is the same as that for relying on scientific theor izing to explain any other aspect of the natural world — namely, a desire to have an account of the phenomenon that is as far as possible undistorted by human interests and bias es. But seeking such an account need not involve dismissing or devaluing other ways of d escribing the world, including folk theories, the humanities, the creative arts, and sp iritual traditions, all of which may pick tably expressible in the language of out patterns and capture insights that are not trac 18

19 science. Nor does endorsing illusionism require us to give up the language of eful way of characterizing our phenomenality: we may continue to employ it as a us s phrase, an essentially dramatic idiom inner life, while recognizing that it is, in Quine’ (Quine, 1960, p. 219). (In comparing qualia to fict ions, I am not expressing a negative 6 .) view of qualia so much as a positive one of fiction tative properties in presenting the case Balog also claims that I illicitly appeal to quali for illusionism. One of her worries concerns intros pective phenomenal concepts. If illusionism is true, she points out, these will be either universally misapplied or meaningless. In the former case, it will be miracul ous that we have them, and in the latter they will be merely ‘mental junk’. It is, sh e suggests, only because we are already acquainted with phenomenal properties that we can m ake sense of our having introspective representations that refer to them. I accept that providing a theory of content for ph enomenal concepts will be a major challenge for illusionism, but I don’t think we hav e reason at this stage to write it off as unsurmountable. Balog argues that the problem is pa rticularly hard because phenomenal concepts, unlike other non-referring one s, are simple and direct ones, with no compositional structure. I suspect this is wrong and that the apparent simplicity of phenomenal concepts belies a lot of structure, whic h we shall tease out as we learn more about the mechanisms involved. (For example, it is plausible that phenomenal concepts contain a distinct affective component. Consider su fferers from pain asymbolia, who recognize pains but no longer find them unpleasant or distressing. Is their introspective which lacks an affective dimension?) concept of pain the same as ours, or a thinner one, As I indicated in the target article, illusionists can appeal to a wide range of factors to nks, links to nonconceptual explain phenomenal content, including conceptual li gnitional capacities for neural states, sensory and introspective representations, and reco and they can explain our acquisition of phenomenal concepts as due to developmental processes, individual theorizing, cultural transmis sion, or a combination of all three. pick out metaphysically real It may be the case that phenomenal concepts do not embody inconsistent theoretical (albeit uninstantiated) properties (I suspect they commitments, in which case they presumably do not). But even if so, it does not follow that they are simply ‘mental junk’ as Balog puts it . They may still play an important role in expressing our relation to stimuli and orienting us with respect to our own sensory processes. I grant that it is not easy to see how this approa ch can account for the apparent richness of our phenomenal worlds, but it is not cl ear that realists are much better placed to do this. Would the mere existence of a ca usal connection to a phenomenal property account for the richness of our conception of it? Balog may say that it is our direct acquaintance with phenomenal properties that gives content to our phenomenal concepts. But this is not a genuinely explanatory m ove, since the acquaintance relation itself is wholly unexplained (if anything, it is th e assumption of acquaintance that is 6 Balog suggests that I take a ‘negative view’ of qu alia, as indicated by my use of the term ‘embarrass ed’ in reference to them (Balog, this issue, p. 47). In fact, I used the term to refer to the difficulty realists face difficulty illusionists avoid. in accounting for the potency of consciousness — a 19

20 illicit in this context). My sketch of an illusioni st theory of phenomenal content may be, and-waving in the direction of a as Balog puts it, hand-waving, but it is at least h coherent research programme. Balog’s other worry concerns the functional sense of ‘what it is like’, which I introduced in contrast to the phenomenal one. To sa y that one’s experiences are like in this functional sense is to say that one has in something formation about them, provided by functionally defined representational m echanisms. We have a strong intuition that such informational processes would n ot be sufficient to give us inner lives of the kind we have, but I suggested that if we had a richer and more detailed account of the representations involved, then we might lose this intuition. Balog objects that in suggesting this I am proposing a functional-represe ntational analysis of what-it’s-like- ness, which is not only highly implausible but a fo rm of conservative realism rather than illusionism. This mistakes the suggestion, however (which was perhaps not clearly expressed). The idea was not that our current notion of what-it’s-like-ness is a functional one. I think it is not. Rather, it was t hat as we develop a richer understanding of the representational processes involved in intro spection, we may reconceptualize our inner lives in terms of the functional notion of ‘w hat it is like’, coming to see them as constituted by representational processes that crea te the illusion of phenomenality. This wouldn’t vindicate the reality of what-it’s-like-ne ss in the phenomenal sense, any more than coming to see a magician’s performance as a se ries of deceptive manipulations would vindicate the reality of the apparent effect. In his commentary, Jesse Prinz argues that illusionism, while not absurd, is less term ‘conservative’), which identifies attractive than reductive realism (he dislikes the phenomenal properties with functional or physical o nes. He suggests that illusionists, heory of consciousness: identities are like dualists, expect too much from a physicalist t not deducible but inferred from correlations and pa rtial explanations, and if a reductive ousness, then we are justified in theory can explain enough of the features of consci adopting it. I take this challenge seriously. Prinz has done tr emendously important work in identifying the psychological and neural correlates of consciousness and in providing the sort of partial explanations that he thinks are the best we can hope for in this area (see, in particular, Prinz, 2012). But while I do n ot deny that we may be able to provide reductive accounts of many aspects of conscious exp erience, I doubt that these will be sufficient to justify realism about phenomenal prop erties in anything like the traditional sense. My worries centre on explanatory gaps. While ident ities may be initially inferred on the basis of partial explanations, we expect to be able to render them intelligible, giving reductive explanations of higher-level properties i n terms of more basic ones. Why should consciousness be an exception, especially wh en the feature that resists explanation is such a central one? A partial explan ationism that doesn’t explain phenomenality itself is too partial. Prinz suggests that gaps arise in the case of cons ciousness because we have direct acquaintance with phenomenal properties. We know ou r experiences by having them, al sort of knowledge, phenomenal not by representing them, and this gives us a speci 20

21 knowledge, which science cannot provide. In the tar get article I dismissed the notion to us, but Prinz argues that we can that physical states can directly reveal themselves hat constitute conscious experiences explain it in processing terms. The neural states t directly reveal themselves to us in virtue of being accessible to higher cognition. They do not need to be represented; they are themselves representations and constitute our l us something about themselves and awareness of external properties. Yet they also tel the subjective way they present the world to us. This is an ingenious move, which deserves detailed consideration, but my first response is sceptical. I see how the accessibility of a representational state would give us a grasp of the worldly properties it represents, bu t I fail to see how it could give us any knowledge of intrinsic properties of the state itse lf. (Of course, those who adopt a representational theory of consciousness will say t hat grasping the content of an experience just is grasping its phenomenal characte r. But then phenomenal knowledge would have no distinctive content. I assume Prinz does not want to take this line: indeed, he thinks that phenomenal properties are intrinsic ones; this issue, p. 191.) Prinz argues that conscious states inform us about themselves be cause they present the world to us in a subjective way, reflecting categories and divi sions imposed by our minds, such as categorical colour boundaries. But, arguably, this is just to say that experience misrepresents the world in certain ways, and it is unclear how a ccess to a misrepresentation of the world can afford us any kn owledge of the representing state’s intrinsic nature. Moreover, even if acquaintance di d give us knowledge of intrinsic phenomenal properties of neural states, I do not se e how this knowledge could have any cognitive significance for us. Neural states affect cognitive processing in virtue of having causal properties that correlate with their represe ntational content. Other causal ly effects. So, if properties they may possess can have no distinctive cognitive phenomenal features are not themselves represented, then they cannot have cognitive effects, even if they have effects of other kinds. This is all very brief of course, but there is a strong case for thinking that all knowledge of the physical world, including those parts of it that constitute our own minds, is repre sentationally mediated. Prinz also makes some points against illusionism, which he thinks is prone to collapse into reductive realism. His first point co ncerns the reference of phenomenal terms. We learn these terms, he argues, by pointing to examples, not by description, so if the exemplified states have physical correlates, then it is to these that the terms refer, and realism is true. My response is that a term may be acquired by pointing yet also have a descriptive component. We may learn phenomen al terms from examples but conceptualize their referents as phenomenal in the sense created by the phenomenality language game (this may involve a later theoretical accretion to concepts originally defined ostensively). If, as illusionists claim, th is conception radically misrepresents the states referred to, then it is misleading to use ph enomenal terms for them. Of course, if Prinz is right, then the phenomenal conception does not misrepresent those states; but this just takes us back to the issue of whether red uctive realism is true. Prinz’s second point concerns the nature of illusi ons. Illusions require seemings — representations of the illusory situation. What sho uld illusionists say about the hey say they are phenomenal states, seemings that constitute phenomenal illusions? If t 21

22 then they have not eliminated phenomenality, but if they say they are beliefs, then they since experience is more fine- cannot explain the apparent richness of experience, issue in various places, so I shall be grained than belief. I have already discussed this brief. I suspect that beliefs may do a lot more wor k here than we imagine, but the illusionist need not claim that they do it all. Phe nomenal illusions may depend on ot phenomenal. They may, for representational states that are fine-grained but n emselves (functionally defined), example, be parasitic on sensory representations th arising when these representations are targeted by conceptual mechanisms. (The thought is that we might introspectively represent a phenomenal property via a sensory content, as the phenomenal property with this kind of content.) Or they might depend on quasi-perceptual introspective mechanisms, which give us the sort of caricatured access to our own neural processes that I discussed earlier. These fine-grained representations might be bound up with various beli efs, creating multi-faceted introspective states. This again is hand-waving, bu t I think it is enough to divert a priori objections to the illusionist project. Prinz concludes his comment with some thoughts on the illusion problem. Explaining the illusion of phenomenal consciousness , he suggests, may be no easier than explaining phenomenal consciousness itself and may require very similar resources, in which case the attractions of illusio nism diminish. I don’t wish to play down the hardness of the illusion problem, but I th ink this overstates the case. As a rule, the more magical and inexplicable something seems, the easier it is to create the illusion hard problem should give us reason to of it than the reality, so the very hardness of the think that the illusion problem will be easier to s olve. If we resist this conclusion, it may no appearance/reality distinction for be because of the familiar intuition that there is gical as the reality. I have already consciousness — the illusion would have to be as ma argued that we should not trust this intuition, but I want to add a further point. Coming to see consciousness as an illusion may inv olve not only theorizing about the mechanisms involved but also reconceptualizing our inner lives. Compare stage magic again. Working out how a magic trick is done involves resisting our natural interpretation of what we see — our intuitive sense of what forces are at work, what causal sequences occur, what properties items have, and so on. In doing this, we can achieve a sort of aspect switch, reconceptualizing the events we see as a sequence of clever manipulations rather than a simple miraculou s effect. In order to understand consciousness, we may need to achieve a similar asp ect switch in introspection, reconceptualizing our inner lives as constituted by complex multi-dimensional representational processes rather than simple pheno menal effects. The test will be what happens as we learn more about the psychology and n europhysiology of consciousness and try to apply its findings introspectively. I su spect we shall find that the illusion problem seems increasingly tractable. Of course, we may still default to the intuitive phenomenal aspect, and when we do we may still feel the pull of the hard problem; but this psychological fact will seem increasingly irre levant. I turn finally to two commentators who adopt a rad ical, non-physicalist view of consciousness. Philip Goff does not attack illusionism directly but challenge s one is incompatible with our scientific motivation for it — the claim that radical realism 22

23 worldview. The target article offered some reasons for thinking that there is a tension . Radical realism, he argues, is not here, but Goff argues that these are not compelling as realists can adopt a Russellian inconsistent with third-person science, especially monist view, which identifies phenomenal properties with absolutely intrinsic properties of physical entities and systems (this v iew is sometimes characterized as a and it is a radical position in my form of physicalism, but it is a non-standard one, additional metaphysical commitments, sense). And although radical realism involves beyond those of third-person science, Goff argues t hat these are not unacceptable. We have introspective grounds for making them, and the ir disadvantages are not overwhelming. In reply, I concede that radical realism need not be strictly inconsistent with science, at least if it takes a Russellian monist form (inte ractionism is a different matter, though there is doubtless a lot more to be said on the mat ter). If phenomenal properties are absolutely intrinsic ones, then they are simply inv isible to third-person science. Moreover, I do not think it is incoherent to suppos e that we might supplement our scientific picture of reality in the way Goff propo ses, though it might involve some 7 (This isn’t to say that I think Russellian rather extravagant metaphysical commitments. monism is without internal problems. In particular, it faces serious problems in explaining how and when subjects of consciousness c ombine and how subjects correspond to physical organisms.) The case for ill usionism is not that there are no alternatives, but simply that it is much better than the alternatives — more economical, more elegant, and, most importantly, more explanato ry. This is not the place for an assessment of Russellian monism, but I shall say a few words about the last point — explanatory power. Goff writes as if Russellian monism offers a new f ramework for research on consciousness, freed from the constraints of what h e calls radical naturalism . But it is difficult to see how it could do this. Since the th eory treats phenomenal properties as intrinsic ones, it offers no predictions as to the behaviour of physical systems (this is what ensures its consistency with third-person scie nce), and its data, which are introspective episodes, cannot be intersubjectively compared and checked. There is no basis here for a collective science of consciousnes s, but, at best, for multiple individual sciences. Indeed, even this is too optimistic. The data for a radical realist science are immediate introspective episodes, and we have no way of comp aring such episodes over time or checking that our beliefs about past episod es are accurate. These are not merely ‘methodological difficulties’, as Goff puts it (thi s issue, p. 91), but in-principle obstacles to a radical realist science of consciousness. Russ ellian monism gives us no new explanatory purchase on the world but merely adds a fifth wheel, which serves no function other than to underwrite our conviction th at we have direct introspective access to phenomenal properties. Perhaps if all att empts to explain consciousness within the standard scientific framework were to fa il, we might fall back on this view. 7 Goff himself argues for a cosmopsychist form of Russellian monism, according to which the universe itself is conscious (Goff, forthcoming). 23

24 But it would be premature to adopt it now. Radical realism may not be incompatible ute for it. with third-person science, but it is a poor substit why I assume that consciousness I shall add one further, minor comment. Goff asks must be known with certainty if it is to be a datum ; in general, we needn’t be certain of our data (Goff, this issue, p. 92). This is true, b ut in other contexts we are prepared to ts wish to insist that the reality of question our data in the light of theory. If realis phenomenal consciousness is a datum, which cannot be questioned (which was bedrock the suggestion under consideration in the context o f my original remark), then I think they do need to claim that it is known with certain ty. Of course, if realists accept that the introspective data are open to question, then i t is a different matter. But illusionists will be happy to fight on this ground. Martine Nida-Rümelin begins her commentary by rejecting the widely held view that phenomenal consciousness consists in having ex periences with phenomenal properties. Talk of experiences having phenomenal p roperties is, she argues, a confused way of talking about subjects having experiential properties, where these are properties that it is like something to undergo. If this is co rrect, then it immediately undercuts illusionism as originally presented. If it is confu sed to think that being phenomenally conscious involves having experiences with phenomen al properties, then it is equally confused to think that it involves having experienc es that are misrepresented as having phenomenal properties. However, as Nida-Rümelin not es, illusionists may simply recast their view as the claim that we misrepresent ourselves as having experiential properties, and she goes on to argue against this c laim. (In fact, she argues that the claim is necessarily oes not fit into our false. She accepts that phenomenal consciousness d considerations would favour standard scientific worldview and that theoretical s the crux of the matter, I shall focus on illusionism, were it a possible view.) Since this i her arguments, granting her earlier move for the sa ke of argument. Nida-Rümelin’s main argument appeals to facts abou t reference fixing. She argues introduced in the following, two-step that reference to experiential properties should be suffering pain, feeling sad, or being way. First, we point to paradigm examples, such as visually presented with blueness. Second, we establ ish reference to a feature all the examples share, using metaphors and provisional des criptions (perhaps talking of ‘what it is like’ to have the properties), but without ma king any theoretical commitments as to the nature of the feature. This shared feature i s what marks out experiential properties and thus phenomenal consciousness. Since illusionists deny the existence of experiential properties, Nida-Rümelin argues, they must either deny that this procedure picks out a common feature of experiential properti es or say that experiential properties are never instantiated. Neither option, she argues, is attractive. This argument is similar to Schwitzgebel’s, and my response is similar. I grant that the first step picks out real properties — the pers onal-level properties we call ‘being in 8 and so on. Illusionists do not deny that pain’, ‘feeling sad’, ‘seeing a blue colour’, something is going on when we are in pain or feeling sad. An d I grant, too, that the 8 I’m not sure about ‘being visually presented with blueness’, which is how Nida-Rümelin puts it (this issue, p. 165). 24

25 second step establishes reference to a common featu re of these properties. However, I it is. It is not some intrinsic quality, akin deny that it is the sort of feature realists think language game. Rather, it is to the property characterized by the phenomenality (roughly) the property of having a cluster of intro spective representational states and cquainted with some intrinsic quality. I dispositions that create the illusion that one is a am sure that this is not what Nida-Rümelin thinks t he procedure picks out, but I don’t see how she can rule out the possibility. She makes it clear that in the second step reference is to be fixed by ostension, not descript ion (she says that any descriptions used are merely an aid to identification and may not sur vive later theorizing — this issue, p. 168). So I am happy to concede the truth of realism about experiential properties in this sense ch is compatible with the . However, this is a very weak kind of realism, whi ontology of illusionism. Nida-Rümelin briefly outlines another argument aga inst illusionism, which turns on the claim that our awareness of experiential pro perties is direct and unmediated: If to have a property P and to be aware of having P is one and the same thing, then the awareness of having P cannot possibly ‘mis represent’ oneself as having P. On that view, being aware of having an experient ial property by having that experiential property does not involve any further step (no reflection, no introspection, no conceptualization) and therefore leaves no room for any kind of illusion. (Nida-Rümelin, this issue, p. 167) The appeal here is, of course, to direct acquaintan ce: experiential properties reveal rror. This notion has cropped up themselves to us immediately, leaving no room for e one. Pace Prinz, I don’t believe it is frequently in this discussion, and it is a central possible to give a physicalist account of direct ac quaintance in any robust sense. I cannot n this way to a physical cognitive see how physical properties can reveal themselves i and dispositional properties of the system (by ‘physical properties’ I mean structural sort described by third-person science, not absolut ely intrinsic ones). So, as a physicalist, I maintain that our sense of being dir ectly acquainted with experiential properties is itself an illusion, an artefact of th e way our sensory and introspective systems are structured. This will not convince Nida -Rümelin, of course, whose metaphysical commitments are different from mine. B ut I think we are close to bedrock here, and that’s a good place to stop. 5. Conclusion I conclude by thanking the commentators once again. They have forced me to think hard about my position and its commitments, but the y haven’t shaken my belief in it. If it’s an illusion to think that phenomenal consci ousness is an illusion, then I’m not 9 disillusioned. 9 I am grateful to Daniel Dennett, Eileen Frankish, Nicholas Humphrey, Maria Kasmirli, and Miloš Tomin for comments, advice, and assistance. 25

26 References Dennett, D.C. (1991) Consciousness Explained , New York: Little, Brown. Dennett, D.C. (2005) Sweet Dreams: Philosophical Obstacles to a Science of Consciousness , Cambridge, MA: MIT Press. Dennett, D.C. (2013) Expecting ourselves to expect: The Bayesian brain as a projector, Behavioral and Brain Sciences , 36 (3), pp. 209–210. Dick, P.K. (1995) The Shifting Realities of Philip K. Dick: Selected Literary and Philosophical Writings, New York: Vintage Books. ew York: Harper and Brothers. Doyle, A.C. (1892) Adventures of Sherlock Holmes, N 21 (2), pp. 667– ess and Cognition, Frankish, K. (2012) Quining diet qualia, Consciousn 676. Garfield, J.L. (2015) Engaging Buddhism: Why it Mat ters to Philosophy, Oxford: Oxford University Press. l Reality, New York: Oxford Goff, P. (forthcoming) Consciousness and Fundamenta University Press. Humphrey, N. (2011) Soul Dust: The Magic of Conscio usness, Princeton, NJ: Princeton University Press. Pereboom, D. (2011) Consciousness and the Prospects of Physicalism, New York: Oxford University Press. Prinz, J.J. (2012) The Conscious Brain, New York: O xford University Press. : MIT Press. Quine, W.V.O. (1960) Word and Object, Cambridge, MA 26

Related documents