P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

CHAPTER 3 Philosophical Theories of : Contemporary Western Perspectives

Uriah Kriegel

Abstract With each approach, I will present in order (i) the leading account of consciousness This chapter surveys current approaches to along its line, (ii) the case for the approach, 1 consciousness in Anglo-American analytic and (iii) the case against the approach. I will philosophy. It focuses on five approaches, not issue a final verdict on any approach, to which I will refer as mysterianism, dual- though by the end of the chapter it should ism, representationalism, higher-order mon- be evident where my own sympathies lie. itoring theory, and self-representationalism. Before starting, let us draw certain dis- With each approach, I will present in order tinctions that may help fix our for the (i) the leading account of consciousness discussion to follow. The term “conscious- along its line, (ii) the case for the approach, ness” is applied in different senses to dif- and (iii) the case against the approach. I will ferent sorts of things. It is applied, in one not issue a final verdict on any approach, sense, to biological species, as when we say though by the end of the chapter it should something like “Gorillas are conscious, but be evident where my own sympathies lie. snails are not”; in a different sense, to individ- ual organisms or creatures, as when we say “Jim is conscious, but Jill is comatose”; and in a third sense, to particular mental states, Introduction: The events, and processes, as when we say “My of Consciousness thought about Vienna is conscious, but Jim’s belief that there are birds in China is not.” This chapter surveys current approaches to To distinguish these different senses, we may consciousness in Anglo-American analytic call the first species consciousness, the second philosophy. It focuses on five approaches, creature consciousness, and the third state con- to which I will refer as mysterianism, dual- sciousness.2 ism, representationalism, higher-order mon- There appear to be certain concep- itoring theory, and self-representationalism. tual connections among these three senses, 35 P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

36 the cambridge handbook of consciousness

such that they may be analyzable in functions.5 The sense is that an insight of a terms of one another. Plausibly, species completely different order would be needed consciousness is analyzable in terms of to make scientific theories, and indeed sci- creature consciousness: a species S is species- ence itself, at all relevant to our understand- conscious just in case a prototypical spec- ing of phenomenal consciousness. Some sort imen of S is creature-conscious. Creature of conceptual breakthrough, which would consciousness may in turn be analyzable in enable us to conceive of the problem of con- terms of state consciousness: a creature C sciousness in new ways, is required. This is is creature-conscious just in case C has (or where philosophical theories of conscious- is capable of having) mental states that are ness come into the picture.6,7 state-conscious. If so, state consciousness is the most fundamental notion of the three. State consciousness is itself ambiguous as Mysterianism between several senses. If Jim tacitly believes that there are birds in China, but never con- Some philosophers hold that science cannot sciously entertained this belief, whereas Jill and will not, in fact, help us understand con- often contemplates consciously the fact that sciousness. So-called mysterianists hold that there are birds in China, but is not doing so the problem of consciousness – the prob- right now, there is a sense of “conscious” in lem of how there could be something like which we may want to say that Jim’s belief is phenomenal consciousness in a purely nat- unconscious whereas Jill’s is conscious. Let ural world – is not a problem we are capa- 3 us call this availability consciousness. By ble (even in principle) of solving. Thus con- contrast, there is a sense of “conscious” in sciousness is a genuine mystery, not merely which a is conscious when and a prima facie mystery that we may one day only when there is something it is like for the demystify. 4 subject – from the inside – to have it. Thus, We may introduce a conceptual distinc- when I take in a spoonful of honey, there is tion between two kinds of mysterianism – an a very specific – sweet, smooth, honey-ish, ontological one and an epistemological one. if you will – way it is like for me to have the According to ontological mysterianism, con- resulting conscious experience. Let us call sciousness cannot be demystified because it this phenomenal consciousness. is an inherently mysterious (perhaps super- Some of the leading scientific theories natural) phenomenon.8 According to episte- of consciousness – such as Baars’ (1988, mological mysterianism, consciousness is in 1997) and Crick no way inherently mysterious, and a greater and Koch’s (1990, 2003) synchrony-based could in principle demystify it – but “neurobiological theory” – shed much light it just so happens that we humans lack the on availability consciousness and neighbor- cognitive capacities that would be required. ing notions. But there is a persistent feel- Epistemological mysterianism has actu- ing that they do not do much to explain ally been pursued by contemporary West- phenomenal consciousness. Moreover, there ern philosophers. The most comprehensive is a widespread sense that there is some- development of the view is offered in Colin thing principled about the way in which McGinn’s (1989, 1995, 1999, 2004) writ- they fail to do so. One way to bring out this ings. We now turn to an examination of his feeling is through such philosophers’ con- account. cepts as the explanatory gap (Levine, 1983)or the hard problem (Chalmers, 1995). Accord- McGinn’s Mysterianism ing to Chalmers, for instance, the problems of explaining the various cognitive func- McGinn’s theory of consciousness has two tions of conscious experiences are the “easy central tenets. First, the phenomenon of con- problems” of consciousness; the “hard prob- sciousness is in itself perfectly natural and lem” is that of why there in no way mysterious. Second, the human should be something it is like to execute these mind’s conceptual capacities are too poor P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

philosophical theories of consciousness: contemporary western perspectives 37

to demystify consciousness. That is, McGinn induction. From the fact that all the theories is an epistemological mysterianist: he does of consciousness we have come up with to not claim that the world contains, in and date are hopelessly unsatisfactory, it should of itself, insoluble mysteries, but he does not be concluded that our future theories contend that we will never understand con- will be the same. It may well be that a thou- sciousness. sand years hence we will look back with At the center of McGinn’s theory is the amusement at the days of our ignorance and concept of cognitive closure. McGinn (1989, self-doubt. p. 529) defines cognitive closure as follows: However, McGinn’s main argument for “A type of mind M is cognitively closed with his position is not the inductive argument respect to a property P (or a theory T) if and just sketched. Rather, it is a deductive argu- only if the concept-forming procedures at ment based on consideration of our cog- M’s disposal cannot extend to a grasp of P (or nitive constitution. The argument revolves an understanding of T).”9 To be cognitively around the claim that we do not have a sin- closed to X is thus to lack the procedure for gle mechanism, or faculty, that can access concept formation that would allow one to both consciousness and the brain. Our access form the concept of X. to consciousness is through the faculty of To illustrate the soundness and appli- . Our access to the brain is cability of the notion of cognitive clo- through the use of our senses, mainly vision. sure, McGinn adduces the case of animal But unfortunately, the senses do not give us and their constitutive limitations. access to consciousness proper, and intro- As James Joyce writes in A Portrait of spection does not give us access to the brain the Artist as a Young Man, rats’ minds proper. Thus, we cannot see with our eyes do not understand trigonometry. Likewise, what it is like to taste chocolate. Nor can we snails do not understand quantum physics, taste with our buds what it is like to taste and cats do not understand market eco- chocolate. We can, of course, taste choco- nomics. Why should humans be spared this late. But we cannot taste the feeling of tast- predicament? As a natural, evolved mecha- ing chocolate. The feeling of tasting choco- nism, the human mind must have its own late is something we encounter only through limitations. One such limitation, McGinn introspection. But alas, introspection fails to suggests, may be presented by the phe- give us access to the brain. We cannot intro- nomenon of consciousness. spect neurons, and so could never introspect Interestingly, McGinn does not claim that the neural correlates of consciousness. we are cognitively closed to consciousness Using the term “extrospective” to denote itself. Rather, his claim is that we are cog- the access our senses give us to the world, nitively closed to that property of the brain McGinn’s argument may be formulated as responsible for the production of conscious- follows: ness. As someone who does not wish to por- 1 tray consciousness as inherently mysterious, ) We can have introspective access to con- McGinn is happy to admit that the brain sciousness but not to the brain; has the capacity to somehow produce con- 2) We can have extrospective access to the scious awareness. But how the brain does so is brain but not to consciousness; something he claims we cannot understand. 3) We have no accessing method that is both Our concept-forming procedures do extend introspective and extrospective; there- to a grasp of consciousness, but they do not fore, extend to a grasp of the causal basis of con- 4) We have no method that can give us sciousness in the brain. access to both consciousness and the brain. The Master Argument for Mysterianism As we can see, the argument is based on con- A natural reaction to McGinn’s view is that siderations that are much more principled it may be based upon an overly pessimistic than a simple pessimistic induction from P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

38 the cambridge handbook of consciousness

past theories. Dismayed as we may be by kinds of states a person or organism may be the prospects of mysterianism, we must not in: brain states on the one hand and con- confuse McGinn’s position for sheer despair. scious states on the other. Instead, we must contend with the argument Recall that McGinn’s mysterianism is of just formulated. the epistemological variety. The epistemo- Some materialists would contest the first logical claim now appears to be conditional premise. (1985) has repeat- upon an ontological claim, namely dualism. edly argued that we will one day be able So at the end of the day, as far as the ontology to directly introspect the neurophysiolog- of consciousness is concerned, McGinn is a ical states of our brains. Perception and straightforward dualist. The plausibility of introspection are theory-laden, according to his (epistemological) mysterianism depends, Churchland, and can therefore be funda- to that extent, on the plausibility of (onto- mentally changed when the theory they are logical) dualism. In the next section, we con- laden with is changed.10 Currently, our intro- sider the plausibility of dualism. spective practice is laden with a broadly Before doing so, let us raise one more diffi- Cartesian . But when we culty for mysterianism, and in particular the mature enough scientifically, and when the notion of cognitive closure. It is, of course, right neuroscientific theory of consciousness undeniable that rats do not understand makes its way to our classroom and living trigonometry. But observe that trigonomet- room, this will change and we (or rather ric problems do not pose themselves to rats our distant offspring) will start thinking (Dennett, 1995,pp.381–383). Indeed, it about ourselves in purely neurophysiological is precisely because rats do not understand categories. trigonometry that trigonometric problems Other materialists may deny the second do not pose themselves to rats. For rats to premise of the argument. As long as brain grapple with trigonometric problems, they states are considered to be merely correlates would have to understand quite a bit of of conscious states, the claim that the con- trigonometry. Arguably, it is a mark of gen- scious states cannot be perceived extrospec- uine cognitive closure that certain questions tively is plausible. But according to material- do not even pose themselves to the cogni- ists, conscious states will turn out to be iden- tively closed. The fact that certain questions tical with the brain states in question, rather about consciousness do pose themselves to than merely correlated therewith. If so, per- humans may therefore indicate that humans ceiving those brain states would just be per- are not cognitively closed to consciousness ceiving the conscious states.11 To assume that (or more accurately to the link between con- we cannot perceive the conscious states is to sciousness and the brain).14 beg the question against the materialist.

Dualism The Case Against Mysterianism To repeat the last point, McGinn appears Traditionally, approaches to the ontology of to assume that conscious states are caused mind and consciousness have been divided by brain states. His argument does not go into two main groups: and dual- through if conscious states are simply iden- ism. The former holds that there is one kind tical to brain states. In other words, the of stuff in the world; the latter that there argument does not go through unless any are two.15 Within monism, there is a further of conscious states with brain states distinction between views that construe the is rejected.12 But such rejection amounts single existing stuff as material and views to dualism. McGinn is thus committed to that construe it as immaterial; the former dualism.13 On the view he presupposes, the are materialist views, the latter idealist.16 conscious cannot be simply identified with Descartes framed his dualism in terms of the physical. Rather, there are two different two different kinds of substance (where a P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

philosophical theories of consciousness: contemporary western perspectives 39

substance is something that can in principle At the same time, Chalmers does not exist all by itself). One is the extended sub- take phenomenal properties to be accidental stance, or matter; the other is the thinking or random superpositions onto the physical substance, or mind. A person, on this view, world. On the contrary, he takes them to be is a combination of two different objects:a causally grounded in physical laws. That is, body and a soul. A body and its correspond- instantiations of phenomenal properties are ing soul “go together” for some stretch of caused by instantiations of physical proper- time, but being two separate objects, their ties, and they are so caused in accordance existence is independent and can therefore with strict laws of nature.21 come apart.17 This means that phenomenal conscious- Modern dualism is usually of a more sub- ness can be explained in physical terms. It is tle sort, framed not in terms of substances just that the explanation will not be a reduc- (or stuffs), but rather in terms of properties. tive explanation, but rather a causal expla- The is that even though there is only nation. To explain an event or phenomenon one kind of stuff or substance, there are two causally is to cite its cause, that is, to say kinds of properties, mental and physical, and what brought it about or gave rise to it.22 neither can be reduced to the other.18 This According to Chalmers, one could in princi- is known as . A particularly ple explain the instantiation of phenomenal cautious version of property dualism claims properties by citing their physical causes. that although most mental properties are A full theory of consciousness would reducible to physical ones, conscious or phe- uncover and list all the causal laws that gov- nomenal properties are irreducible. ern the emergence of phenomenal proper- ties from the physical realm. And a full description of nature and its behavior would Chalmers’ Naturalistic Dualism have to include these causal laws on top For many decades, dualistic arguments were of the causal laws obtained by “ultimate treated mainly as a challenge to a physical- physics.”23 ist worldview, not so much as a basis for Chalmers himself does not attempt to a non-physicalist alternative. Thus dualism detail many of these laws. But he does pro- was not so much an explanation or account pose a pair of principles to which we should of consciousness, but rather the avoidance expect such laws to conform. These are of one. This state of affairs has been recti- the “structural coherence” principle and the fied in the past decade or so, mainly through “organizational invariance” principle. The the work of (1995, 1996, former concerns the sort of direct avail- 2002a). ability for global control that conscious Chalmers’ theory of consciousness, which states appear to exhibit, the latter the sys- he calls naturalistic dualism, is stronger than tematic correspondence between a system’s ordinary dualism, in that it claims not only functional organization and its phenomenal that phenomenal properties are not identical properties.24 to physical properties, but also that they fail to supervene – at least with metaphysical or The Case for Dualism logical necessity19 – on physical properties.20 We tend to think, for instance, that biological The best-known arguments in favor of prop- properties necessarily supervene on physical erty dualism about consciousness are so- properties, in the sense that two systems can- called epistemic arguments. The two main not possibly differ in their biological proper- ones are Frank Jackson’s (1984) “Knowl- ties if all their physical properties are exactly edge Argument” and ’s (1974) similar. But according to Chalmers, phenom- “what is it like” argument. Both follow a sim- enal properties are different: two systems ilar pattern. After describing a situation in can be exactly the same physically, but have which all physical facts about something are different phenomenal properties. known, it is shown that some knowledge is P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

40 the cambridge handbook of consciousness

still missing. It is then inferred that the miss- tion of a new piece of knowledge one cannot ing knowledge must be knowledge of non- infer the existence of a new fact – and that physical facts. is precisely the inference made in the above The Knowledge Argument proceeds as dualist arguments.27,28 follows. Suppose a baby is kept in a black- A different argument for dualism that is and-white environment, so that she never widely discussed today is Chalmers’ (1996) has color experiences. But she grows to argument from the conceivability of zom- become an expert on color and color vision. bies. Zombies are imaginary creatures that Eventually, she knows all the physical facts are physically indistinguishable from us but about color and color vision. But when she lack consciousness. We seem to be able to sees red for the first time, she learns some- conceive of such creatures, and Chalmers thing new: she learns what it is like to see red. wants to infer from this that is That is, she acquires a new piece of knowl- false. The argument is often caricatured as edge. Since she already knew all the physical follows: facts, this new piece of knowledge cannot be 1 knowledge of a physical fact. It is therefore ) Zombies are conceivable; knowledge of a non-physical fact. So, the fact 2) If As are conceivable, then As are (meta- 29 thereby known (what it is like to see red) is physically) possible; therefore, a non-physical fact. 3) Zombies are possible; but, Nagel’s argument, although more obscure 4) Materialism entails that zombies are not in its original presentation, can be “format- possible; therefore, ted” along similar lines. We can know all the 5) Materialism is false. physical facts about bats without knowing what it is like to be a bat. It follows that the Or, more explicitly formulated: knowledge we are missing is not knowledge 1 of a physical fact. Therefore, what it is like ) For any physical property P, it is conceiv- to be a bat is not a physical fact. able that P is instantiated but conscious- These arguments have struck many mate- ness is not; rialists as suspicious. After all, they infer an 2) For any pair of properties F and G, if it ontological conclusion from epistemological is conceivable that F is instantiated when premises. This move is generally suspicious, G is not, then it is (metaphysically) pos- but it is also vulnerable to a response that sible that F is instantiated when G is not; emphasizes what philosophers call the inten- therefore, sionality of epistemic contexts.25 This has 3) For any physical property P, it is possible been the main response among materialists that P is instantiated and consciousness is (Loar, 1990; Tye, 1986). The claim is that not; but, the Knowledge Argument’s protagonist does 4) If a property F can be instantiated when not learn a new fact when she learns what it property G is not, then F does not super- is like to see red, but rather learns an old vene on G;30 therefore, fact in a new way; and similarly for the bat 5 26 ) For any physical property P,consciousness student. does not supervene on P. Consider knowledge that the evening star glows and knowledge that the morning star To this argument it is objected that the sec- glows. These are clearly two different pieces ond premise is false, and the conceivability of knowledge. But the fact thereby known of something does not entail its possibility. is one and the same – the fact that Venus Thus, we can conceive of water not being glows. Knowledge that this is what it is like H2 O, but this is in fact impossible; Escher to see red and knowledge that this is the neu- triangles are conceivable, but not possible.31 ral assembly stimulated by the right wave- The zombie argument is more subtle length may similarly constitute two separate than this, however. One way to get at the pieces of knowledge that correspond to only real argument is this.32 Let us distinguish one fact being known. So from the acquisi- between the property of being water and P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

philosophical theories of consciousness: contemporary western perspectives 41

the property of appearing to be water, or water or being conscious do exist, “unnatu- being apparent water.33 For a certain quan- ral” properties do not, and appearance prop- 36 tity of stuff to be water, it must be H2 O. erties are unnatural in the relevant sense. But for it to appear to be water, it need To avoid this latter objection, a dualist only be clear, drinkable, liquid, and so on – may proceed to develop the argument dif- or perhaps only strike normal subjects as ferently, claiming that in the case of con- clear, drinkable, liquid, etc. Now, although sciousness, there is no distinction between the unrestricted principle that conceivabil- appearance and reality (Kripke, 1980). This ity entails possibility is implausible, a version would amount to the claim that the property of the principle restricted to what we may of being conscious is identical to the prop- call appearance properties is quite plausible. erty of appearing to be conscious. The con- Thus, if we can conceive of apparent water ceivability argument then goes like this: not being H2 O, then it is indeed possible that 1) For any physical property P, it is conceiv- apparent water should not be H2 O. Once the restricted principle is accepted, able that P is instantiated but apparent there are two ways a dualist may proceed. consciousness is not; The zombie argument seems to be captured 2) For any pair of properties F and G, such more accurately as follows:34 that F is an appearance property, if it is conceivable that F is instantiated when G 1) For any physical property P, it is conceiv- is not, then it is (metaphysically) possi- able that P is instantiated but apparent ble that F is instantiated when G is not; consciousness is not; therefore, 2) For any pair of properties F and G, such 3) For any physical property P, it is possible that F is an appearance property, if it is that P is instantiated when apparent con- conceivable that F is instantiated when G sciousness is not; but, is not, then it is (metaphysically) possi- 4) If property F can be instantiated when ble that F is instantiated when G is not; property G is not, then F does not super- therefore, vene on G; therefore, 3 ) For any physical property P, it is possible 5) For any physical property P,apparent con- that P is instantiated when apparent con- sciousness does not supervene on P; but, sciousness is not; but, 6) Consciousness = apparent consciousness; 4 ) If a property F can be instantiated when therefore, property G is not, then F does not (meta- 7) For any physical property P,consciousness physically) supervene on G; therefore, does not supervene on P. 5) For any physical property P, apparent consciousness does not (metaphysically) Materialists may reject this argument by supervene on P. denying that there is no distinction between appearance and reality when it comes to con- A materialist might want to reject this sciousness (the sixth premise). 2 argument by denying Premise (the res- The debate over the plausibility of the tricted conceivability-possibility principle). various versions of the zombie argument Whether the restricted principle is true is continues. A full critical examination is something we cannot settle here. Note, how- impossible here. Let us move on, then, ever, that it is surely much more plausible to consideration of the independent case than the corresponding unrestricted princi- against dualism. ple, and it is the only principle that the argu- ment for dualism really needs. The Case against Dualism Another way the argument could be rejected is by denying the existence of such The main motivation to avoid dualism con- properties as apparent water and appar- tinues to be the one succinctly worded ent consciousness.35 More generally, perhaps, by Smart (1959,p.143) almost a half- while “natural” properties such as being century ago: “It seems to me that science is P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

42 the cambridge handbook of consciousness

increasingly giving us a viewpoint whereby that is merely counter- organisms are able to be seen as physico- intuitive, but does not face serious argumen- chemical mechanisms: it seems that even tative challenges. This is not particularly sat- the behavior of man himself will one day be isfying, however: all arguments must come explicable in mechanistic terms.” It would be to an end, and in most of philosophy, the curious if consciousness stood out in nature end is bound to be a certain or as the only property that defied reduc- intuitively compelling claim. As tive explanation in microphysical terms. go, the intuition that consciousness is not More principled arguments aside, this simple epiphenomenal is very strong. observation seems to be the chief motivat- The second prong is more interesting. ing force behind naturalization projects that Chalmers notes that physics characterizes attempt to reductively explain conscious- the properties to which it adverts in purely ness and other recalcitrant phenomena. relational terms – essentially, in terms of As I noted above, against traditional dual- the laws of nature into which they enter. ists it was common to present the more The resulting picture is a network of inter- methodological argument that they do not related nodes, but the intrinsic character of in fact propose any positive theory of con- the thus interrelated nodes remains opaque. sciousness, but instead rest content with It is a picture that gives us what Bertrand arguing against existing materialist theories, Russell once wittily called “the causal skele- and that this could not lead to real progress in ton of the world.” Chalmers’ suggestion is the understanding of consciousness. Yet, this that phenomenal properties may constitute charge cannot be made against Chalmers, the intrinsic properties of the entities whose who does propose a positive theory of relational properties are mapped out by consciousness. physics. At least this is the case with intrin- Chalmers’ own theory is open to more sic properties of obviously conscious entities. substantial criticisms, however. In particu- As for apparently inanimate entities, their lar, it is arguably committed to epiphenom- intrinsic properties may be crucially similar enalism about consciousness, the thesis that to the phenomenal properties of conscious conscious states and events are causally inert. entities. They may be, as Chalmers puts it, As Kim (1989a,b, 1992) has pointed out, “protophenomenal” properties. it is difficult to find causal work for non- Although intriguing, this suggestion has supervenient properties. Assuming that the its problems. It is not clear that physics physical realm is causally closed (i.e., that indeed gives us only the causal skeleton of every instantiation of a physical property the world. It is true that physics characterizes has as its cause the instantiation of another mass in terms of its causal relations to other physical property), non-supervenient prop- properties. But it does not follow that the erties must either (i) have no causal effect property thus characterized is nothing but a on the physical realm or (ii) causally overde- bundle of causal relations. More likely, the termine the instantiation of certain physical relational characterization of mass is what properties.37 But because pervasive overde- fixes the reference of the term “mass,” but termination can be ruled out as implausi- the referent itself is nonetheless an intrinsic ble, non-supervenient properties must be property. The bundle of causal relations is causally inert vis-a-vis` the physical world. the reference-fixer, not the referent. On this However, the notion that consciousness is view of things, although physics character- causally inert, or epiphenomenal, is extremely izes mass in causal terms, it construes mass counter-intuitive: we seem to ourselves to not as the causing of effects E, but rather as act on our conscious decisions all the time the causer (or just the cause) of E. It con- and at will. strues mass as the relatum, not the relation. In response to the threat of epiphenom- Furthermore, if physics did present us enalism, Chalmers pursues a two-pronged with the causal skeleton of the world, then approach.38 The first prong is to claim physical properties would turn out to be P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

philosophical theories of consciousness: contemporary western perspectives 43

epiphenomenal (or nearly so). As Block have my conscious experience of the sky is (1990b) argued, functional properties – just a matter of my experience’s representa- properties of having certain causes and tion of the blue sky. The phenomenal char- effects – are ultimately inert, because an acter of my experience can be identified with effect is always caused by its cause, not by (an aspect of) its representational content.42 its causing. So if mass was the causing of E, This would be a theoretically happy rather than the cause of E, then E would result, since we have a fairly good notion as not be caused by mass. It would be caused, to how may be itself rather, by the protophenomenal property reductively explained in terms of informa- that satisfies the relational characterization tional and/or teleological relations between attached to mass in physics.39 The upshot neurophysiological states of the brain and is that if mass was the causing of E, rather physical states of the environment.43 The than the cause of E, mass would not have the reductive strategy here is two-stepped, then: causal powers we normally take it to have. first reduce phenomenal properties to repre- More generally, if physical properties were sentational properties, then reduce represen- nothing but bundles of causal relations, they tational properties to informational and/or would be themselves causally inert.40 other physical properties of the brain. Chalmers faces a dilemma, then: either he violates our strongly held intuitions Tye’s PANIC Theory regarding the causal efficacy of phenomenal properties, or he violates our strongly held Not every mental representation is con- intuitions regarding the causal efficacy of scious. For this reason, a representational physical properties. Either way, half his account of consciousness must pin down world is epiphenomenal, as it were. In any more specifically the kind of representation event, as we saw above, the claim that phys- that would make a mental state conscious. ical properties are merely bundles of causal The most worked-out story in this genre is relations – which therefore call for the postu- probably Michael Tye’s (1992, 1995, 2000, lation of phenomenal and protophenomenal 2002) “PANIC Theory.”44 properties as the putative causal relata – is The acronym “PANIC” stands for Poised, implausible. Abstract, Non-conceptual, Intentional Con- Problems concerning the causal efficacy tent. So for Tye, a mental representation of phenomenal properties will attach to qualifies as conscious when, and only when, any account that portrays them as non- its representational content is (a) inten- supervenient upon, or even as non-reducible tional, (b) non-conceptual, (c) abstract, and to, physical properties. These problems are (d) poised. What all these qualifiers mean is less likely to rear their heads for reduc- not particularly important, though the prop- tive accounts of consciousness. Let us erties of non-conceptuality and poise are turn, then, to an examination of the main worth pausing to explicate.45 reductive accounts discussed in the current The content of a conscious experience is literature. non-conceptual in that the experience can represent properties for which the subject lacks the concept. My conscious experience Representationalism of the sky represents the sky not simply as being blue, but as being a very specific shade According to the representational theory of blue, say blue17. And yet if I am presented of consciousness – or for short, representa- a day later with two samples of very simi- tionalism – the phenomenal properties of lar shades of blue, blue17 and blue18, I will conscious experiences can be reductively be unable to recognize which shade of blue explained in terms of the experiences’ rep- was the sky’s. This suggests that I lack the 41 resentational properties. Thus, when I look concept of blue17. If so, my experience’s rep- 46 up at the blue sky, what it is like for me to resentation of blue17 is non-conceptual. P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

44 the cambridge handbook of consciousness

The property of poise is basically a func- It appears, then, that when you pay atten- tional role property: a content is poised tion to your experience, the only thing you when it is ready and available to make a become aware of is which features of the direct impact on the formation of beliefs and external sky your experience represents.In desires. Importantly, Tye takes this to dis- other words, the only introspectively acces- tinguish conscious representation from, say, sible properties of conscious experience are blindsighted representations. A square can its representational properties. be represented both consciously and blind- The transparency of experience provides sightedly. But only the conscious representa- a straightforward argument for representa- tion is poised to make a direct impact on the tionalism. The argument may be laid out as beliefs that the subject subsequently forms. follows: PANIC theory is supposed to cover not only conscious perceptual experiences but 1) The only introspectively accessible prop- also all manners of phenomenal experi- erties of conscious experience are its rep- ence: somatic, emotional, and so on. Thus, a resentational properties; toothache experience represents tissue dam- 2) The phenomenal character of conscious age in the relevant tooth, and does so inten- experience is given by its introspectively tionally, non-conceptually, abstractly, and accessible properties; therefore, with poise.47 3) The phenomenal character of conscious experience is given by its representational The Master Argument for properties. Representationalism The first premise is the thesis of The main motivation for representational- transparency; the second one is intended ism may seem purely theoretical: it holds the as a conceptual truth (about what we promise of a reductive explanation of con- mean by “phenomenal”). The conclusion is sciousness in well-understood informational representationalism. and/or teleological terms. Perhaps because Another version of the argument from of this, however, the argument that has been transparency, one that Tye employs, centers most influential in making representational- on the idea that rejecting representational- ism popular is a non-theoretical argument, ism in the face of transparency would require one that basically rests on a phenomenolog- one to commit to an “error theory.”48 This ical observation. This is the observation of version may be formulated as follows: the so-called transparency of experience. It has been articulated in a particularly influential 1) The phenomenal properties of conscious manner by Harman (1990), but goes back at experience seem to be representational least to Moore (1903). properties; Suppose you have a conscious experience 2 of the blue sky. Your attention is focused ) It is unlikely that the phenomenal prop- on the sky. You then decide to turn your erties of conscious experience are radi- attention away from the sky and onto your cally different from what they seem to be; experience of the sky. Now your attention therefore, is no longer focused on the sky, but rather 3) It is likely that the phenomenal proper- on the experience thereof. What are you ties of conscious experience are represen- aware of? It seems that you are still aware tational properties. of the blueness of the sky. Certainly you are not aware of some second blueness, which Here the transparency thesis is again the attaches to your experience rather than to first premise. The second premise is the the sky. You are not aware of any intermedi- claim that convicting experience of massive ary blue quality interposed between yourself error is to be avoided. And the conclusion is and the sky. representationalism. P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

philosophical theories of consciousness: contemporary western perspectives 45

The Case against Representationalism Tye, 2000). To be sure, you do not necessar- ily possess the concept of subtending a visual Most of the arguments that have been mar- angle. But recall that the content of expe- shaled against representationalism are argu- rience can be construed as non-conceptual. ments by counter-example. Scenarios of So your experience can represent the two varying degrees of fancifulness are adduced, trees to subtend different visual angles with- in which allegedly (i) a conscious experi- out employing the concept of subtend- ence has no representational properties, or ing a visual angle. Thus a representational (ii) two possible experiences with different difference is matched to the phenomenal phenomenal properties have the same repre- difference. sentational properties, or (iii) inversely, two Perhaps the most prominent alleged possible experiences with the same phenom- counter-example is Block’s (1990a) Inverted enal properties have different representa- Earth case. Inverted Earth is an imaginary tional properties. For want of space, I present planet just like Earth, except that every only one representative scenario from each object there has the color complementary to category. the one it has here. We are to imagine that a Block (1996) argues that phosphene expe- subject is clothed with color-inverting lenses riences are non-representational. These can and shipped to Inverted Earth unbeknownst be obtained by rubbing one’s eyes long to her. The color inversions due to the lenses enough so that when one opens them again, and to the world cancel each other out, one “sees” various light bits floating about. so that her phenomenal experiences remain Such experiences do not represent any exter- the same. But about represen- nal objects or features, according to Block. tational contents ensures that the represen- In response, Tye (2000) claims that such tational content of her experiences eventu- experiences do represent – it is just that they ally change.49 Her bluish experiences now misrepresent. They misrepresent there to be represent a yellow sky. When her sky expe- small objects with phosphorescent surfaces riences on Inverted Earth are compared to floating around the subject’s head. her earthly sky experience, it appears that A long-debated case in which phenome- the two groups are phenomenally the same nal difference is accompanied by represen- but representationally different. tational sameness is due to Peacocke (1983). This case is still being debated in the Suppose you stand in the middle of a mostly literature, but there are two representa- empty road. All you can see in front of you tionalist strategies for accommodating it. are two trees. The two trees, A and B, have One is to argue that the phenomenal char- the same size and shape, but A is twice as far acter also changes over time on Inverted from you as B. Peacocke claims that, being Earth (Harman, 1990); the other is to devise aware that the two trees are the same size, accounts of representational content that you represent to yourself that they have the make the representational content of the same properties. And yet B “takes up more subject’s experiences remain the same on of your visual field” than A, in a way that Inverted Earth, externalism notwithstanding makes you experience the two trees differ- (Tye, 2000).50 ently. There is phenomenal difference with- There may be, however, a more princi- out representational difference. pled difficulty for representationalism than Various responses to this argument have the myriad counter-examples it faces.51 Rep- been offered by representationalists. Perhaps resentationalism seems to construe the phe- the most popular is that although you rep- nomenal character of conscious experiences resent the two trees to have the same size purely in terms of the sensuous qualities they properties, you also represent them to have involve. But arguably there is more to phe- certain different properties – for example, nomenal character than sensuous quality. In B is represented to subtend a larger visual particular, there seems to be a certain mine- angle than A (DeBellis, 1991; Harman, 1990; ness, or for-me-ness, to them. P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

46 the cambridge handbook of consciousness

One way to put it is as follows (Kriegel, Higher-Order Monitoring Theory 2005a; Levine, 2001; Smith, 1986). When I have my conscious experience of the blue One theory of consciousness from analytic sky, there is a bluish way it is like for me philosophy that can be interpreted as target- to have my experience. A distinction can ing subjective character is the higher-order be drawn between two components of this monitoring theory (HOMT). According to “bluish way it is like for me”: the bluish com- HOMT, what makes a mental state con- ponent, which we may call qualitative char- scious is the fact that the subject is aware acter, and the for-me component, which we of it in the right way. It is only when the may call subjective character. We may con- subject is aware (in that way) of a mental strue phenomenal character as the comp- state that the state becomes conscious.55 resence of qualitative and subjective charac- HOMT tends to anchor consciousness in ter. This subjective character, or for-me-ness, the operation of a monitoring device. This is certainly an elusive phenomenon, but it device monitors and scans internal states and is present in every conscious experience. events and produces higher-order represen- Indeed, its presence seems to be a condi- tations of some of them.56 When a mental tion of any phenomenality: it is hard to state is represented by such a higher-order make sense of the idea of a conscious expe- representation, it is conscious. So a mental rience that does not have this for-me-ness state M of a subject S is conscious when, to it. If it did not have this for-me-ness, it and only when, S has another mental state, would be a mere subpersonal state, a state M∗, such that M∗ is an appropriate represen- that takes place in me but is not for me tation of M. The fact that M∗ represents M in the relevant sense. Such a subpersonal guarantees that there is something it is like state seems not to qualify as a conscious for S to have M.57 experience. Observe that, on this view, what con- The centrality of subjective character (as fers conscious status on M is something out- construed here) to consciousness is some- side M, namely, M∗. This is HOMT’s reduc- thing that has been belabored in the phe- tive strategy. Neither M nor M∗ is conscious nomenological tradition (see Chapter 4; in and of itself, independently of the other Zahavi, 1999). The concept of prereflec- state. It is their coming together in the right tive self-consciousness – or a form of self- way that yields consciousness.58 awareness that does not require focused Versions of the HOMT differ mainly in and explicit awareness of oneself and one’s how they construe the monitoring device current experience, but is rather built into and/or the representations it produces. The that very experience – is one that figures most seriously worked out version is prob- centrally in almost all phenomenological ably David Rosenthal’s (1986, 1990, 2002a, 52 accounts of consciousness. But it has been b). Let us take a closer look at his “higher- somewhat neglected in analytic philosophy order thought” theory. of mind.53 The relative popularity of representation- Rosenthal’s Higher-Order Thought alism attests to this neglect. While a repre- Theory sentationalist account of sensuous qualities – what we have called qualitative character – According to Rosenthal, a mental state is may turn out to win the day (if the all- conscious when its subject has a suitable eged counter-examples can be overcome), it higher-order thought about it.59 The higher- would not provide us with any perspective order state’s being a thought is supposed on subjective character.54 Therefore, even if to rule out, primarily, its being a quasi- representationalism turns out to be a satis- perceptual state. factory account of qualitative character, it is There is a long tradition, hailing from unlikely to be a satisfactory account of phe- Locke, of construing the monitoring device nomenal consciousness proper. as analogous in essential respects to a sense P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

philosophical theories of consciousness: contemporary western perspectives 47

organ (hence as being a sort of “inner sense”) It is worth noting that according to Rosen- and accordingly as producing mental states thal the second-order representation is nor- that are crucially similar to perceptual rep- mally an unconscious state. To be sure, it need resentations and that may to that extent not necessarily be: in the more introspective, be called “quasi-perceptual.” This sort of or reflective, episodes of our conscious life, “higher-order perception theory” is cham- the second-order state becomes itself con- pioned today by Armstrong (1968, 1981) scious. It is then accompanied by a third- and Lycan (1987, 1996). Rosenthal believes order state, one that represents its occur- that this is a mistake and that the higher- rence in a suitable way. When I explicitly order states that confer consciousness are not introspect and dwell on my conscious expe- analogous to perceptual representations.60 rience of the sky, there are three separate Rather, they are intellectual, or cognitive, states I am in: the (first-order) experience, a states – that is, thoughts. (second-order) awareness of the experience, Another characteristic of thoughts – in and a (third-order) representation of that addition to being non-perceptual – is their awareness. When I stop introspecting and being assertoric. An assertoric state is one turn my attention back to the sky, however, that has a thetic, or mind-to-world, direc- the third-order state evaporates, and con- tion of fit.61 This is to be contrasted with sequently the second-order state becomes states (such as wanting, hoping, disapprov- unconscious again. In any event, at any one ing, etc.) that have primarily a telic, or world- time the subject’s highest-order state, the to-mind, direction of fit.62 A third character- one that confers consciousness on the chain istic of thoughts – at least the kind suitable of lower-order states “below” it, is uncon- for conferring consciousness – is that they scious.68 are occurrent mental states.63 In summary, Rosenthal’s central thesis Crucially, a suitable higher-order thought is that a mental state is conscious just in would also have to be non-inferential, in that case the subject has a non-perceptual, non- it could not be the result of a conscious infer- inferential, assertoric, de se, occurrent repre- ence from the lower-order state (or from any sentation of it. This account of consciousness other state, for that matter).64 To be sure, is not intended as an account of introspec- the thought is formed through some process tive or reflective consciousness, but of regu- of information processing, but that process lar, everyday consciousness. must be automatic and unconscious. This is intended to reflect the immediacy, or at least The Master Argument for Higher-Order felt immediacy, of our awareness of our con- Monitoring Theory scious states.65 The fact that my experience The master argument for the higher-order of the sky has for-me-ness entails that I am monitoring approach to consciousness has somehow aware of its occurrence; but not been succinctly stated by Lycan (2001): any sort of awareness would do – very medi- ated forms of awareness cannot confer con- 1) A mental state M of subject S is conscious scious status on their objects. when, and only when, S is aware of M in One last characteristic Rosenthal ascribes the appropriate way; to the “suitable” higher-order representation 2) Awareness of X requires mental represen- is that it represents the lower-order state tation of X; therefore, as a state of oneself. Its content must be, as 66 3) M is conscious when, and only when, S this is sometimes put, de se content. So ∗ ∗ has a mental state M , such that M rep- the content of the higher-order represen- resents M in the appropriate way. tation of my conscious experience of the sky is not simply something like “this bluish Although the second premise is by no means experience is occurring,” but rather some- trivial, it is the first premise that has been thing like “I myself am having this bluish the bone of contention in the philosophical experience.”67 literature (see, e.g., Dretske, 1993). P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

48 the cambridge handbook of consciousness

One can defend the claim that conscious side world. The claim is rather that we are states are states we are aware of having sim- at least peripherally aware of every conscious ply as a piece of conceptual analysis – as a state we are in.69 As long as M is conscious, platitude reflecting the very meaning of the S is aware, however dimly and inattentively, word “conscious” (Lycan, 1996). To my ear, of M. Once S’s awareness of M is extin- this sounds right: a mental state of which guished altogether, M drops into the realm of the subject is completely unaware is a sub- the unconscious. This seems highly plausible personal, and therefore unconscious, state. on both conceptual and phenomenological To some, however, this seems plainly grounds.70 false. When I have an experience of the sky, I am attending to the sky, they stress, not to The Case against Higher-Order myself and my internal goings-on. By con- Monitoring Theory sequence, I am aware of the sky, not of my experience of the sky. I am aware through my Several problems for the monitoring the- experience, not of my experience. ory have been continuously debated in the This objection seems to rely, however, philosophical literature. I focus here on what on an unwarranted assimilation of awareness I take to be the main three.71 and attention. There is a distinction to be The first is the problem of animal and made between attentive awareness and inat- infant consciousness. It is intuitively plausi- tentive awareness. If S attends to X and not ble to suppose that cats, dogs, and human to Y, it follows that S is not attentively aware neonates are conscious, that is, they have of Y, but it does not follow that S is com- conscious states; but it appears empirically pletely unaware of Y. For S may still be inat- implausible that they should have second- tentively aware of Y. order representations (Lurz, 1999). The Consider straightforward visual aware- problem is particularly acute for Rosenthal’s ness. The distinction between foveal vision account, since it is unlikely that these crea- and peripheral vision means that our visual tures can have thoughts, and moreover of the awareness at any one time has a periphery complex form, “I myself am enjoying this as well as a focal center. Right now, I am milk.” (visually) focally aware of my laptop, but There are two ways to respond to this also (visually) peripherally aware of an ash- objection. One is to deny that having such tray at the far corner of my desk. A similar higher-order representations requires a level distinction applies to perceptual awareness of sophistication of an order unlikely to in other modalities: I am now (auditorily) be found in (say) cats. Thus, Rosenthal focally aware of Duke Ellington’s voice and (2002b) claims that whereas adult human (auditorily) peripherally aware of the air higher-order thoughts tend to be conceptu- conditioner’s hum overhead. ally structured and employ a rich concept There is no reason to think that a similar of self, these are not necessary features of distinction would not apply to higher-order such thoughts. There could be higher-order awareness. In reflective moods I may be thoughts that are conceptually simple and focally aware of my concurrent experiences employ a rudimentary concept of self, one and feelings, but on other occasions I am just that consists merely in the ability to distin- peripherally aware of them. The former is guish oneself from anything that is not one- an attentive form of second-order awareness, self. It may well turn out that worms, wood- the latter an inattentive one. Again, from the peckers, or even day-old humans lack even fact that it is inattentive it would be falla- this level of conceptual sophistication – in cious to infer that it is no awareness at all. which case we would be required to deny When it is claimed that conscious states them consciousness – but it is unlikely that are states we are aware of, the claim is not cats, dogs, and year-old humans lack them. that we are focally aware of every conscious The second possible line of response is state we are in. That is manifestly false: the to dismiss the intuition that animals, such focus of our attention is mostly on the out- as cats, dogs, and even monkeys, do in P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

philosophical theories of consciousness: contemporary western perspectives 49

fact have conscious states. Thus, Carruthers There are several ways higher-order mon- (1998, 1999) claims that there is a signifi- itoring theorists may respond to this objec- cant amount of projection that takes place tion. Let us briefly consider three possible when we ascribe conscious states to, say, responses. our pets. In reality there is very little evi- First, they may claim that when M∗ is dence to suggest that they have not only targetless, the property of being conscious, perceptual and cognitive states but also although not instantiated by M, is instanti- conscious ones. ated by M∗. But as we saw above, according Both lines of response offer some hope to their view, M∗ is normally unconscious. to the defender of higher-order monitor- So to say that M∗ instantiates the property ing, but also implicate the theory in certain of being conscious would be to say that it is, counter-intuitive and prima facie implausi- in the normal case, both conscious and not ble claims. Whether these could somehow conscious – which is incoherent.75 be neutralized, or accepted as outweighed by Second, they may claim that the prop- the theoretical benefits of HOMT, is some- erty of being conscious is, in reality, not thing that is very much under debate. a property of the discrete state M, but Perhaps more disturbing is the problem rather attaches itself to the compound of of so-called targetless higher-order thoughts M and M∗.76 But this will not work either, (or more generally, representations). When because HOMT would then face the fol- someone falsely believes that the almond lowing dilemma. Either the compound state tree in the backyard is blooming again, there M + M∗ is a state we are aware of having, are two ways he or she may get things wrong: or it is not. If it is not, then HOMT is false, (i) it may be that the backyard almond tree since it claims that conscious states are states is not blooming, or (ii) it may be that there we are aware of having. If it is, then accord- is no almond tree in the backyard (blooming ing to the theory it must be represented by a or not). Let us call a false belief of type (ii) third-order mental state, M∗∗, in which case a targetless thought. HOMT gets into trou- the same problem would recur when M∗∗ is ble when a subject has a targetless higher- targetless. order thought (Byrne, 1997).72 Suppose at a Third, they may claim that there are time t subject S thinks (in the suitable way) no targetless higher-order representations. that she has a throbbing toothache, when in But even if this can be shown to be the reality she has no toothache at all (throb- actual case (and it is hard to imagine how bing or not). According to HOMT, what it this would be done), we can surely con- is like for S at t is the way it is like to have a ceive of counterfactual situations in which throbbing toothache, even though S has no targetless higher-order representations toothache at t. In other words, if S has an do occur.77 M∗ that represents M when in reality there A third problem for the HOMT is its is no M,73 S will be under the impression treatment of the epistemology of conscious- that she is in a conscious state when in real- ness (Goldman, 1993b; Kriegel, 2003b). Our ity she is not. (She is not in a conscious state knowledge that we are in a conscious state because M does not exist, and it is M that is is first-person knowledge, knowledge that is supposed to bear the property of being con- not based on inference from experimental, scious.) Moreover, on the assumption that a or theoretical, or third-personal evidence. person is conscious at a time t only if she But if HOMT were correct, what would has at least one conscious state at t,74 this make our conscious states conscious is (nor- would entail that when a subject harbors mally) the occurrence of some unconscious a targetless higher-order misrepresentation, state (i.e., the higher-order representation), she is not conscious, even though it feels to so in order to know that we are in a conscious her as though she is. This is a highly counter- state we would need to know of the occur- intuitive consequence: we want to say that a rence of that unconscious state. But knowl- person cannot be under the impression that edge of unconscious states is necessarily the- she is conscious when she is not. oretical and third-personal, since we have P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

50 the cambridge handbook of consciousness

no direct acquaintance with our unconscious popularity, is what we may call the “self- states. representational theory.” According to this Another way to put the argument is this. view, mental states are conscious when, and How does the defender of HOMT know only when, they represent their own occur- that conscious states are states of which we rence (in the right way). Thus, my conscious are aware? It does not seem to be some- experience of the blue sky represents both thing she knows on the basis of experimenta- the sky and itself – and it is in virtue of repre- tion and theorization. Rather, it seems to be senting itself that it is a conscious experience. intuitively compelling, something that she Historically, the most thorough devel- knows on the basis of first-person acquain- opment and elucidation of the self-rep- tance with her conscious states. But if resentational theory is Brentano’s (1874). HOMT were correct, it would seem that Through his work, the view has had a sig- that knowledge would have to be purely nificant influence in the phenomenologi- theoretical and third-personal. So construed, cal tradition. But apart from a couple of this “epistemic argument” against HOMT exceptions – Lehrer (1996, 1997) and Smith may be formulated as follows: (1986, 1989) come to mind – the view had enjoyed virtually no traction in Anglo- 1) If HOMT were correct, our awareness of American philosophy. Recently, however, our conscious states would normally be versions of the view, and close variations an unconscious state; that is, on it, have been defended by a number of 2) We do not have non-theoretical, first- philosophers.78 person knowledge of our unconscious Rather than focus on any one particular states; therefore, account of consciousness along these lines, I 3) If HOMT were correct, we would not now survey the central contributions to the have non-theoretical, first-person knowl- understanding of consciousness in terms of edge of the fact that we are aware of our self-representation. conscious states; but, 4) We do have non-theoretical, first-person Varieties of Self-Representational Theory knowledge of the fact that we are aware of our conscious states; therefore, Brentano held that every conscious state is intentionally directed at two things. It is 5) HOMT is incorrect. primarily directed at whatever object it is The upshot of the argument is that the about, and it is secondarily directed at itself. awareness of our conscious states must in the My bluish sky experience is directed pri- normal case be itself a conscious state. This is marily at the sky and secondarily at itself. something that HOMT cannot allow, how- In more modern terminology, a conscious ever, since within its framework it would state has two representational contents: an lead to infinite regress. The problem is to other-directed (primary) content and a self- reconcile the claim that conscious states are directed (secondary) content. Thus, if S states we are aware of having with the notion consciously fears that p, S’s fear has two that we have non-theoretical knowledge of contents: the primary content is p, the sec- this fact. ondary content is itself, the fear that p. The distinction between primary intention- ality and secondary is pre- sumably intended to capture the differ- The Self-Representational Theory ence (discussed above) between attentive or of Consciousness focal awareness and inattentive or peripheral awareness.79 One approach to consciousness that has a Caston (2002) offers an interesting gloss venerable tradition behind it, but has only on this idea in terms of the type/token dis- very recently regained a modest degree of tinction. For Caston, S’s conscious fear that p P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

philosophical theories of consciousness: contemporary western perspectives 51

is a single token state that falls under two sep- thesis – and this seems to amount to the arate state types: the fear-that-p type and the claim that M is conscious just in case it rep- awareness-of-fear-that-p type. The state has resents itself (constitutes a representation of two contents, arguably, precisely in virtue of itself). But the point is that there are other, falling under two types. weaker constitutive relations that fall short Brook and Raymont (2006) stress that of full identity. the self-representational content of the con- One such relation is the part-whole rela- scious state is not simply that the state tion. Accordingly, one version of the view, occurs, but rather that it occurs within one- the one defended by Gennaro (1996, 2006), self – that it is one’s own state. Just as Rosen- holds that M∗ is a part of M; another version, thal construed the content of higher-order apparently put forth by Kobes (1995), holds states as “I myself am having that state,” so that M is part of M∗; and yet another ver- Brook and Raymont suggest that the full self- sion, Van Gulick’s (2001, 2006), holds that representational content of conscious states M is conscious when it has two parts, one of is something like “I myself am herewith hav- which represents the other. ing this very state.”80 In Van Gulick’s “higher-order global For Brentano and his followers, the self- states theory,” S’s fear that p becomes con- directed element in conscious states is an scious when the fear and S’s awareness of the aspect of their intentionality, or content. fear are somehow integrated into a single, In David Woodruff Smith’s (1986, 2004) unified state. This new state supersedes its “modal account,” by contrast, the self- original components, though, in a way that directed element is construed not as an makes it a genuine unity, rather than a sum aspect of the representational content, but of two parts, one of which happens to rep- rather as an aspect of the representational resent the other. The result is a state that, if attitude (or mode). When S consciously it does not represent itself, does something fears that p, it is not in virtue of figuring very close to representing itself.84 in its own secondary content that the fear is conscious. Indeed, S’s fear does not have a secondary content. Its only content is p. The Master Argument for the The “reflexive character” of the fear, as Smith Self-Representational Theory puts it, is rather part of the attitude S takes The basic argument for the self-represen- toward p. Just as the attitudes toward p can tational approach to consciousness is that vary from fear, hope, expectation, and so it is the only way to accommodate the on, so they can vary between self-directed notion that conscious states are states we are or “reflexive” fear and un-self-directed or aware of without falling into the pitfalls of “irreflexive” fear. S’s fear that p is conscious, HOMT. on this view, because S takes the attitude of The argument can be organized, then, as 81,82 self-directed fear toward p. a disjunctive syllogism that starts from the One way in which the self-represen- master argument for HOMT, but then goes tational thesis can be relaxed to make a sub- beyond it: tler claim is the following. Instead of claim- ing that a mental state M of a subject S is 1) A mental state M of subject S is con- conscious just in case M represents itself, the scious when, and only when, S is aware thesis could be that M is conscious just in of M; case S has an M∗ that is a representation of 2 M and there is a constitutive, non-contingent ) Awareness of X requires mental represen- relation between M and M∗.83 One consti- tation of X; therefore, tutive relation is of course identity. So one 3) M is conscious when, and only when, S ∗ ∗ version of this view would be that M is con- has a mental state M , such that M rep- scious just in case M is identical with M∗ – resents M. this is how Hossack (2002) formulates his 4) Either M∗ = MorM∗ = M; P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

52 the cambridge handbook of consciousness

5) There are good reasons to think that it is at all. The same problem does not arise for not the case that M∗ = M; therefore, self-representing states, however: although 6) There are good reasons to think that it is M could in principle misrepresent itself to be the case that M∗ = M; therefore, F when in reality it is not F, it could not possi- 7) Plausibly, M is conscious when, and only bly misrepresent itself to be F when in reality when, M is self-representing. it does not exist at all. For if it did not exist it could not represent anything, itself included. The fourth premise could also be formu- Thus the problem of targetless higher-order lated as “either M∗ and M do not enter- representations has no bite against the self- tain a constitutive, non-contingent relation, representational theory. or they do,” with appropriate modifications These are already two major problems in Premises 5 and 6 to suit. The conclu- that affect gravely the plausibility of HOMT, sion of the relevantly modified argument but do not apply to the self-representational would then be the thesis that M is conscious theory. They make a strong prima facie case when, and only when, S has a mental state for the fifth premise above. The fourth M∗, such that (i) M∗ represents M and (ii) premise is a logical truism, and the first and there is a constitutive, non-contingent rela- second ones were defended above. So the tion between M and M∗. argument appears to go through. The fallacy in the master argument for HOMT is the supposition that if S is aware Problems for the Self-Representational of M, then S must be so aware in virtue of Theory being in a mental state that is numerically different from M. This supposition is brought One problem that does persist for the self- to the fore and rejected in the argument just representational theory is the problem of sketched. . The ability to have The case for the fifth premise consists in self-representing states presumably requires all the reasons to be suspicious of HOMT, as all the conceptual sophistication that the elaborated in the previous section, although ability to have higher-order monitoring it must also be shown that the same prob- states does (since the self-representational lems do not bedevil the self-representational content of a conscious state is the same as theory as well. the representational content that a higher- Consider first the epistemic argument. order state would have), and perhaps even We noted that HOMT fails to account for greater sophistication.85 the non-theoretical, first-person knowledge Another problem is the elucidation and we have of the fact that we are aware viability of the notion of self-representation. of our conscious states. This is because it What does it mean for a mental state to construes this awareness as (normally) an represent itself, and what sort of mecha- unconscious state. The self-representational nism could subserve the production of self- theory, by contrast, construes this awareness representing states? There is something at as a conscious state, since it construes the least initially mysterious about the notion awareness as the same state, or part of the of a self-representing state that needs to be state, of which one is thereby aware. So the confronted. self-representational theory, unlike HOMT, In fact, one might worry that there are can provide for the right epistemology of principled reasons why self-representation consciousness. is incompatible with any known natural- Consider next the problem of target- ist account of mental representation. These less higher-order representations. Recall, the accounts construe mental representation as problem ensues from the fact that M∗ could some sort of natural relation between brain in principle misrepresent not only that M states and world states. Natural relations, as is F when in reality M is not F, but also opposed to conceptual or logical ones, are that M is F when in reality there is no M based on causality and causal processes. But P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

philosophical theories of consciousness: contemporary western perspectives 53

causality is an anti-reflexive relation, that ture with no conscious awareness whatso- is, a relation that nothing can bear to itself. ever who harbors mental states that rep- Thus no state can bring about its own occur- resent themselves. Thus Chalmers’ zombie rence or give rise to itself. The argument can argument can be run in a particularized ver- be formulated as follows: sion directed specifically against the self- representational theory.90 1) Mental representation involves a causal relation between the representation and the represented; Conclusion: Directions for 2) The causal relation is anti-reflexive; Future Research therefore, 3) No mental state can cause itself; and Much of the philosophical discourse on therefore, consciousness is focused on the issue of 4) No mental state can represent itself. reducibility. As we just saw, the zombie argu- ment and other dualist arguments can be The basic idea is that there is no naturalist tailored to target any particular reductive account of mental representation that could account of consciousness. This debate holds allow for self-representing mental represen- great intrinsic importance, but it is impor- tations. tant to see that progress toward a scientific Even more fundamentally, one may worry explanation of consciousness can be made whether the appeal to self-representation without attending to it. really explains consciousness. Perhaps self- All three reductive approaches to con- representation is a necessary condition for sciousness we considered – the represen- consciousness, but why think it is also a suf- tational, higher-order monitoring, and self- ficient condition? A sentence such as “this representational theories – can readily be very sentence contains six words” is self- refashioned as accounts not of consciousness representing, but surely there is nothing it itself, but of the emergence base (or causal is like to be that sentence.86 basis) of consciousness. Instead of claiming One may respond to this last point that that consciousness is (or is reducible to) phys- what is required for consciousness is intrinsic ical structure P, the claim would be that or original self-representation, not derivative consciousness emerges from (or is brought self-representation.87 Sentences and linguis- about by) P. To make progress toward the tic expressions do not have any representa- scientific explanation of consciousness, we tional content in and of themselves, indepen- should focus mainly on what the right phys- dently of being interpreted. But plausibly, ical structure is – what P is. Whether P is mental states do.88 The same goes for self- consciousness itself or only the emergence representational content: sentences and lin- base of consciousness is something we can guistic expressions may be derivatively self- set aside for the purposes of scientific expla- representing, but only mental states can be nation. If it turns out that P is conscious- non-derivatively self-representing. A more ness itself (as the reductivist holds), then we accurate statement of the self-representation will have obtained a reductive explanation of theory is therefore this: A mental state M of a consciousness; if it turns out that P is only subject S is conscious when, and only when, the emergence base of consciousness (as the M is non-derivatively self-representing. dualist holds), then we will have obtained Still, self-representing zombies are read- a causal explanation of consciousness. But ily conceivable. It is quite easy to imagine both kinds of explanation are bona fide sci- unconscious mental states in our own cog- entific explanations. nitive system – say, states formed early on In other words, philosophers could use- in visual processing – that represent them- fully reorganize their work on conscious- selves without thereby being conscious.89 ness around a distinction between two sep- Furthermore, it is easy to imagine a crea- arate issues or tasks. The first task is to P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

54 the cambridge handbook of consciousness

devise a positive account of the physical case against. Obviously, space and other lim- (or more broadly, natural) correlate of con- itations do not allow me to present the full sciousness, without prejudging whether it story on each of these approaches. will constitute a reduction base or merely 2. The distinction between creature conscious- an emergence base. Work along these lines ness and state consciousness is due to Rosen- will involve modifying and refining the rep- thal (1986). resentational, higher-order monitoring, and 3. Availability consciousness as construed here is self-representational theories and/or devis- very similar to the notion of access conscious- ing altogether novel positive accounts. The ness as defined by Block (1995). There are second task is to examine the a priori and certain differences, however. Block defines a posteriori cases for reducibility. Work here access consciousness as the property a men- tal state has when it is poised for free use will probably focus on the issue of how much by the subject in her reasoning and action can be read off of conceivability claims, control. It may well be that a mental state as well as periodic reconsideration of the is availability-conscious if and only if it is intuitive plausibility of such claims in light access-conscious. For a detailed discussion of of newer and subtler positive accounts of the relation between phenomenal conscious- 91 consciousness. ness and access consciousness, see Kriegel Another front along which progress can (2006b). certainly be made is tightening the connec- 4. It is debatable whether thoughts, beliefs, tion between the theoretical and experi- desires, and other cognitive states can at all mental perspectives on consciousness. Ulti- be conscious in this sense. I remain silent on mately, one hopes that experiments could be this issue here. For arguments that they can 1993 designed that would test well-defined empir- be conscious, see Goldman ( a), Horgan 2002 1998 ical consequences of philosophical (or more and Tienson ( ), and Siewert ( ). generally, purely theoretical) models of con- 5. The terms “easy problems” and “hard prob- sciousness. This would require philosophers lem” are intended as mere labels, not as to be willing to put forth certain empiri- descriptive. Thus it is not suggested here that understanding any of the functions of con- cal speculations, as wild as these may seem, sciousness is at all easy in any significant sense. based on their theories of consciousness, and Any scientist who has devoted time to the experimental scientists to take interest in study of consciousness knows how outstand- the intricacies of philosophical theories in ing the problems in this field are. These terms an attempt to think up possible ways to test are just a terminological device designed to them. bring out the fact that the problem of why All in all, progress in our understand- there is something it feels like to undergo a ing of consciousness and the outstanding conscious experience appears to be of a dif- methodological and substantive challenges it ferent order than the problems of mapping presents has been quite impressive over the out the cognitive functions of consciousness. past two decades. The central philosophical 6. This is so even if phenomenal consciousness issues are today framed with a clarity and does not turn out to have much of a func- precision that allow a corresponding level of tional significance in the ordinary cognitive life of a normal subject – as some (Libet, 1985; clarity and precision in our thinking about Velmans, 1992; Wegner, 2002) have indeed consciousness. Even more happily, there is argued. no reason to suppose that this progress will 7. In the course of the discussion I avail myself come to a halt or slow down in the near 92 of philosophical terminology that may not future. be familiar to the non-philosophically trained reader. However, I have tried to recognize all the relevant instances and such and include an Notes endnote that provides a standard explication of the terminology in question. 1. More accurately, I present central aspects of 8. No major philosopher holds this view, to my the main account, the case in favor, and the knowledge. P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

philosophical theories of consciousness: contemporary western perspectives 55

9. Many of the key texts discussed in this chap- is that if a conceptual scheme is power- ter are conveniently collected in Block et al. ful enough to frame a problem it should (1997). Here, and in the rest of the chapter, I be powerful enough to frame the solution. refer to the reprint in that volume. Whether the correct solution will actually 10. This is what Churchland often discusses be framed is of course anyone’s guess. But under the heading of the “plasticity of mind” the problem cannot be a constitutive limita- (see especially Churchland, 1979). tion on concept formation mechanisms. (For a more detailed development of this line of cri- 11 . It may not be perceiving those brain states tique, see Kriegel, 2004a.) There is a counter- as brain states. But it will nonetheless be a example of this sort of claim, however. Cer- matter of perceiving the brain states. tain problems that can be framed within the 12. The view – sometimes referred to as emer- theory of rational numbers cannot be solved gentism – that consciousness is caused by the within it; the conceptual machinery of irra- brain, or causally emerges from brain activ- tional numbers must be brought in to solve ity, is often taken by scientists to be material- these problems. It might be claimed, how- ist enough. But philosophers, being interested ever, that this sort of exception is limited to in the ontology rather than genealogy of con- formal systems and does not apply to theo- sciousness, commonly take it to be a form of ries of the natural world. Whether this claim dualism. If consciousness cannot be shown to is plausible is something I do not adjudicate be itself material, but only caused by matter, here. then consciousness is itself immaterial, as the 15. Monism divides into two subgroups: mate- dualist claims. At the same time, the position rialist monism, according to which the only implicit in scientists’ work is often that what kind of stuff there is is matter, and idealist is caused by physical causes in accordance monism, according to which the stuff in ques- with already known physical laws should be tion is some sort of mindstuff. immediately considered physical. This posi- tion, which I have called elsewhere inclusive 16. is not really considered a live option materialism (Kriegel, 2005b), is not unreason- in current philosophical discussions, although able. But the present chapter is dedicated to it is defended by Foster (1982). I do not discuss philosophers’ theories of consciousness, so I set it here. it aside. 17. Such coming-apart happens, for Descartes, 13. It should be noted that McGinn himself has upon death of the physical body. We should repeatedly claimed that his position is not note that Cartesian substance dualism drew dualist. Nonetheless others have accused him much of its motivation from religious con- of being committed to dualism (e.g., Brueck- siderations, partly because it provided for the ner and Berukhim, 2003). There is no doubt survival of the soul. The main difficulty his- that McGinn does not intend to commit to torically associated with it is whether it can dualism. In a way, his position is precisely account for the causal interaction between that, because of our cognitive closure we can- the mind and the body. not even know whether materialism or dual- 18. So property dualism is compatible with sub- ism is true. Yet it is a fair criticism to suggest stance monism. Unlike Descartes and other that McGinn is committed to dualism despite old-school dualists, modern dualists for the himself because his argument for mysterian- most part hold that there is only one kind of ism would not go through unless dualism was stuff, or substance, in the world – matter. But true. matter has two different kinds of properties – 14. More generally, it is curious to hold, as material and immaterial. McGinn does, that an organism’s concept- 19. A kind of property F supervenes on a kind forming procedures are powerful enough to of property G with logical necessity – or for frame a problem, without being powerful short logically supervenes on them – just in enough to frame the solution. To be sure, case two objects differing with respect to their the wrong solution may be framed, but this F properties without differing with respect would suggest not that the conceptual capa- to their G properties would be in contraven- bilities of the organism are at fault, but rather tion of the laws of logic. A kind of prop- that the organism made the wrong turn some- erty F supervenes on a kind of property G where in its reasoning. The natural thought with metaphysical necessity – or for short P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

56 the cambridge handbook of consciousness

metaphysically supervenes on them – just in tial generalization (the inference from “a is F” case it is impossible for two objects to dif- to “there is an x, such that x is F”) and sub- fer with respect to their F properties with- stitution of co-referential terms salva veritate out differing with respect to their G prop- (the inference from “a is F” and “a = b” to “b is erties. Philosophers debate whether there is F”). Epistemic contexts – contexts involving a difference between the two (logical and the ascription of knowledge – are intensional metaphysical supervenience). That debate in this sense. does not concern us here. 26. Another popular materialist response to these 20. This stronger claim will require a stronger arguments is that what is being gained is argument. The claim that phenomenal prop- not new knowledge, but rather new abil- erties are not identical to physical proper- ities (Lewis, 1990; Nemirow, 1990). Upon ties could be established through the now being released from her room, the Knowledge familiar argument from multiple realizability Argument’s protagonist does not acquire new (Putnam, 1967). But multiple realizability knowledge, but rather a new set of abilities. does not entail failure of supervenience. To And likewise what we lack with respect to obtain the latter, Chalmers will have to appeal what it is like to be a bat is not any particular to a different argument, as we will see in the knowledge, but a certain ability – the ability next subsection. to imagine what it is like to be a bat. But from 21. As a consequence, phenomenal properties do the acquisition of a new ability one can surely supervene on physical properties with nomo- not infer the existence of a new fact. logical necessity, even though they do not 27. Materialists reason that because what it is supervene with metaphysical or logical neces- like to see red is identical to a neurophysi- sity. A kind of property F supervenes on a kind ological fact about the brain, and ex hypoth- of property G with nomological (or natural) esi the Knowledge Argument’s protagonist necessity – or for short nomologically super- knows the latter fact, she already knows the venes on them – just in case two objects differ- former. So she knows the fact of what it is like ing with respect to their F properties without to see red, but not as a fact about what it is differing with respect to their G properties like to see red. Instead, she knows the fact of would be in contravention of laws of nature. what it is like to see red as a fact about the 22. So causal explanation is the sort of explana- neurophysiology of the brain. What happens tion one obtains by citing the cause of the when she comes out of her room is that she explanandum. For discussions of the nature comes to know the fact of what it is like to of causal explanation, see (e.g., Lewis, 1993). see red as a fact about what it is like to see red. That is, she learns in a new way a fact 23. The latter will govern only the causal inter- she already knew in another way. The same action among physical events. They will not applies to knowledge of what it is like to be a cover causal interaction between physical and bat: we may know all the facts about what it phenomenal, non-physical events. These will is like to see a bat, and still gain new knowl- have to be covered by a special and new set edge about bats, but this new knowledge will of laws. present to us a fact we already know in a way 24. In Baars’ (1988, 1997) Global Workspace The- we do not know it yet. ory, consciousness is reductively explained in 28. It could be responded by the dualist that terms of global availability. In a functionalist some pieces of knowledge are so different that 1981 1991 theory such as Dennett’s ( , ), con- the fact known thereby could not possibly sciousness is reductively explained in terms of turn out to be the same. Knowledge that the functional organization. Chalmers’ position is evening star is glowing and knowledge that that neither theory can explain consciousness the morning star is glowing are not such. But reductively, though both may figure as part consider knowledge that justice is good and of the causal explanation of it. These the- knowledge that banana is good. The dual- ories are not discussed in the present chap- ist could argue that these are such different ter, because they are fundamentally psycho- pieces of knowledge that it is impossible that logical (rather than philosophical) theories of the facts thereby known should turn out to consciousness. be one and the same. The of evening 25. A linguistic context is intensional if it disal- star and morning star are not different enough lows certain inferences, in particular existen- to exclude the possibility that they pick out P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

philosophical theories of consciousness: contemporary western perspectives 57

the same thing, but the concepts of justice not mean, however, that the property is actu- and banana are such that it cannot possibly ally instantiated by any actual object.) But on be the case that justice should turn out to be a sparse conception of property – one that the same thing as bananas. rejects the latitudinous assumption – there 29. The kind of possibility we are concerned with may not be appearance properties at all. here, and in the following presentation of 36. The notion of a natural property is hard to variations on this argument, is not practical pin down and is the subject of philosophi- possibility, or even a matter of consistency cal debate. The most straightforward way of with the laws of nature. Rather it is possibil- understanding natural properties is as proper- ity in the widest possible sense – that of con- ties that figure in the ultimate laws of nature sistency with the laws of logic and the very (Armstrong, 1978; Fodor, 1974). essence of things. This is what philosophers 37. That is, they would have their causal effi- refer to as metaphysical possibility. cacy restricted to bringing about physi- 30. The modal force of this supervenience claim cal events and property-instantiations that is concordant with that of the claim in already have independent sufficient causes Premise 2; that is, that of metaphysical (and that would therefore take place any- necessity. way, regardless of the non-supervenient prop- 31. The reason it is impossible is that there is erties. (This is the second option of the no such thing as contingent identity, according dilemma.) to the official doctrine hailing from Kripke. 38. This is the strategy in Chalmers (1996). Since all identity is necessary, and necessity is Later on, Chalmers (2002a) embraces a three- cashed out as truth in all possible worlds, it pronged approach, the third prong consisting follows that when a = b in the actual world, in accepting causal overdetermination. a = b in all possible worlds, that is, a is neces- 39. When a cause C causes an effect E, C’s caus- sarily identical to b. ing of E may have its own (mostly acciden- 32 . The interpretation I provide is based on cer- tal) effects (e.g., it may surprise an observer 1996 131 tain key passages in Chalmers ( ,pp. – who did not expect the causing to take place), 134 ), but I cast the argument in terms that are but E is not one of them. This is because mine, not Chalmers’. E is caused by C, not by C’s causing of E. 33. I mean the property of apparent water to be Dretske (1988) distinguished between trig- more or less the same as the property philoso- gering causes and structuring causes, the lat- phers often refer to as “watery stuff” (i.e., ter being causes of certain causal relations the property of being superficially (or to the (such as C’s causing of E), and offers an naked eye) the same as water – clear, drink- account of structuring causes. But this is an able, liquid, etc.). account of the causes of causal relations, not 34. Chalmers (1996, 132) writes, “...theprimary of their effects. To my knowledge, there is intension [of “consciousness”] determines a no account of the effects of causal relations, perfectly good property of objects in possible mainly because these seem to be chiefly acci- worlds. The property of being watery stuff dental. [or apparent water] is a perfectly reasonable 40. Or at least they would be nearly epiphenom- property, even though it is not the same as the enal, having no causal powers except perhaps property of being H2 O. If we can show that to bring about some accidental effects of the there are possible worlds that are physically sort pointed out in the previous endnote. identical to ours but in which the properly 41. By “representational properties” it is meant introduced by the primary intension is lack- properties that the experience has in virtue ing, then dualism will follow [italics added].” of what it represents – not, it is important to 35. Our discussion so far has presupposed a “lat- stress, properties the experience has in virtue itudinous” approach to properties, according of what does the representing. In terms of the to which there is a property that corresponds distinction between vehicle and content, rep- to every predicate we can come up with. resentational properties are to be understood (Thus, if we can come up with the predi- as content properties rather than vehicular cate, “is a six-headed space lizard or a flying properties. We can also make a distinction cow,” then there is the property of being a six- between two kinds of vehicular properties: headed space lizard or a flying cow. This does those that are essential to the vehicling of P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

58 the cambridge handbook of consciousness

the content and those that are not. (Block’s Shoemaker (1994a, b, 1996, 2002) and Thau (1996) distinction between mental paint and (2002). Some of these versions are impor- mental latex (later, “mental oil”) is supposed tantly different from Tye’s, not only in detail to capture this distinction.) There is a sense in but also in spirit. This is particularly so which a view according to which phenomenal with regard to Shoemaker’s view (as well properties are reductively accountable for in as Lurz’s). For a limited defense and elab- terms of vehicular properties essential to the oration of Shoemaker’s view, see Kriegel vehicling is representational, but the way the (2002a, b). In what way this defense is term “representationalism” is used in current limited will become evident at the end of discussions of consciousness, it does not qual- this section. ify as representationalism. A view of this sort 45. The properties of intentionality and abstract- 1989 is defended, for instance, by Maloney ( ), ness are fairly straightforward. The for- but otherwise lacks a vast following. I do not mer is a matter of intensionality; that is, discuss it here. the disallowing of existential generalizations 42. By the “phenomenal character” of a mental and truth-preserving substitutions of co- state at a time t I mean the set of all phenom- referential terms. The second is a matter of enal properties the state in question instanti- the features represented by experience not ates at t. By “representational content” I mean being concrete entities (this is intended to whatever the experience represents. (Experi- make sense of misrepresentation of the same ences represent things, in that they have cer- features, in which case no concrete entity is tain accuracy or veridicality conditions: con- being represented). ditions under which an experience would be 46. This line of thought can be resisted on a said to get things right.) number of scores. First, it could be argued 43. See Dretske (1981, 1988) for the most thor- that I do have a short-lived concept of blue17 , oughly worked out reductive account of men- which I possess more or less for the dura- tal representation in informational and tele- tion of my experience. Second, it could be 1981 ological terms. According to Dretske ( ), claimed that although I do not possess the every event in the world generates a certain descriptive concept “blue17 ,” I do possess the amount of information (in virtue of exclud- indexical concept “this shade of blue,” and ing the possibility that an incompatible event that it is the latter concept that is deployed can take place). Some events also take place in my experience’s representational content. only when other events take place as well, Be that as it may, the fact that conscious and this is sometimes dictated by the laws of experiences can represent properties that nature. Thus it may be a law of nature that the subject cannot recognize across rela- dictates that an event type E1 is betokened tively short stretches of time is significant only when event type E2 is betokened. When enough. Even if we do not wish to treat them this is the case, E1 is said to be nomically depen- as non-conceptual, we must treat them at dent upon E2 , and the tokening of E1 carries least as “sub-recognitional.” Tye’s modified the information that E2 has been betokened. claim would be that the representational con- Or more accurately, the tokening of E1 carries tent of experience is poised, abstract, sub- the information generated by the tokening of recognitional, intentional content. E2 . Some brain states bear this sort of rela- 47 tion to world states: the former come into . To be sure, it does not represent the tissue being, as a matter of law, only when the latter damage as tissue damage, but it does repre- do (i.e., the former are nomically dependent sent the tissue damage. Since the represen- upon the latter). Thus, a certain type of brain tation is non-conceptual, it certainly cannot state may be tokened only when it rains. This employ the concept of “tissue damage.” brain state type would thus carry the informa- 48. An error theory is a theory that ascribes tion that it rains. An informational account of a widespread error in commonsense beliefs. mental representation is based on this idea: The term was coined by J. L. Mackie (1977). that a brain state can represent the fact that it Mackie argued that values and value judg- rains by carrying information about it, which ment are subjective. Oversimplifying the it does in virtue of nomically depending on it. dialectic, a problem for this view is that such a 44. Other representational theories can be found judgment as “murder is wrong” appears to be, in Byrne (2001), Dretske (1995), Lurz (2003), and is commonly taken to be, objectively true. P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

philosophical theories of consciousness: contemporary western perspectives 59

In response Mackie embraced what he termed 54. There are versions of representationalism that an error theory: that the common view of may be better equipped to deal with the sub- moral and value judgments is simply one huge jective character of experience. Thus, accord- mistake. ing to Shoemaker’s (2002) version, a men- 49. Externalism about representational content, tal state is conscious when it represents a or “content externalism” for short, is the subject-relative feature, such as the disposi- thesis that the representational content of tion to bring about certain internal states in experiences, thoughts, and even spoken state- the subject. It is possible that some kind of ments is partially determined by objects out- for-me-ness can be accounted for in this man- side the subject’s head. Thus, if a person’s ner. It should be noted, however, that this interactions with watery stuff happen to is not one of the considerations that moti- be interactions with H2 O, and another per- vate Shoemaker to develop his theory the way son’s interactions with watery stuff happen he does. to be interactions with a superficially simi- 55. Rosenthal prefers to put this idea as follows: lar stuff that is not composed of H2 O, then conscious states are states we are conscious even if the two persons cannot tell apart of. He then draws a distinction between con- H2 O and the other stuff and are unaware sciousness and consciousness of – intransi- of the differences in the molecular com- tive and transitive consciousness (Rosenthal, position of the watery stuff in their envi- 1986, 1990). To avoid unnecessary confusion, ronment, the representational contents of I state the same idea in terms of awareness- their respective water thoughts (as well as of, rather than consciousness-of. But the idea water pronouncements and water experi- is the same. It is what Rosenthal calls some- ences) are different (Putnam, 1975). Or so times the “transitivity principle” (e.g., Rosen- externalists claim. thal, 2000): a mental state is intransitively 50. Another option is to go internalist with conscious only if we are transitively conscious respect to the representational content that of it. determines the phenomenal properties of 56. The representation is “higher-order” in the conscious experiences. With the recent sense that it is a representation of a repre- advent of credible account of narrow con- sentation. In this sense, a first-order represen- tent (Chalmers, 2002b, Segal, 2000), it is now tation is a representation of something that is a real option to claim that the phenomenal not itself a representation. Any other repre- properties of experience are determined by sentation is higher-order. experience’s narrow content (Kriegel, 2002a; 57. More than that, according to Rosenthal Rey, 1998). However, it may turn out that (1990), for instance, the particular way it is this version of representationalism will not like for S to have M is determined by the par- be as well supported by the transparency of ticular way M∗ represents M. Suppose S tastes experience. an identical wine in 1980 and in 1990. During 51. For one such line of criticism, on which I do the 1980s, however, S had become a wine con- not elaborate here, see Kriegel (2002c). noisseur. Consequently, wines she could not 1980 1990 52. Elsewhere, I construe this form of pre- distinguish at all in strike her in as reflective self-consciousness as what I call worlds apart. That is, during the eighties she intransitive self-consciousness. Intransitive self- acquired a myriad of concepts for very spe- consciousness is to be contrasted with transi- cific and subtle wine tastes. It is plausible to tive self-consciousness. The latter is ascribed claim that what it is like for S to taste the 1990 in reports of the form “I am self-conscious wine in is different from what it was 1980 of my thinking that p,” whereas the former like for her to taste it in – even though is ascribed in reports of the form “I am self- the wines’ own flavors are identical. Arguably, consciously thinking that p.” For details see the reason for the difference in what it is like Kriegel (2003b, 2004b). to taste the wine is that the two wine-tasting experiences are accompanied by radically dif- 53. Part of this neglect is justified by the thesis ferent higher-order representations of them. that the for-me-ness of conscious experiences This suggests, then, that the higher-order rep- is an illusory phenomenon. For an argument resentation not only determines that there is for the psychological reality of it, see Kriegel something it is like for S to have M, but also (2004b). what it is like for S to have M. P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

60 the cambridge handbook of consciousness

58. I do not mean the term “yield” in a causal 64. Earlier on, Rosenthal (1990) required that sense here. The higher-order monitoring the- the higher-order thought be not only ory does not claim that M∗’s represent- non-inferential but also non-observational. ing of M somehow produces,orgives rise This latter requirement was later dropped to, M’s being conscious. Rather, the claim (Rosenthal, 1993). is conceptual: M’s being conscious consists 65. A person may come to believe that she is in,oris constituted by,M∗’s representing ashamed about something on the strength of of M. her therapist’s evidence. And yet the shame 59. Other versions of the higher-order thought state is not conscious. In terms of the ter- view can be found in Carruthers (1989, 1996), minology introduced in the introduction, the Dennett (1969, 1991), and Mellor (1978). state may become availability-conscious, but not phenomenally conscious. This is why 60 1990 739 40 . Rosenthal ( ,pp. – ) claims that it is the immediacy of awareness is so crucial. essential to a perceptual state that it has a sen- Although the person’s second-order belief sory quality, but the second-order represen- constitutes an awareness of the shame state, it tations do not have sensory qualities and are is not a non-inferential awareness, and there- 2001 therefore non-perceptual. Van Gulick ( ) fore not immediate awareness. details a longer and more thorough list of 66. De se content is content that is of one- features that are characteristic of perceptual self, or more precisely, of oneself as oneself. states and considers which of them is likely Castaneda (1966), who introduced this term, to be shared by the higher-order represen- ˜ also claimed that de se content is irreducible tations. His conclusion is that some are and to any other kind of content. This latter claim some are not. is debatable and is not part of the official 61 . The notion of direction of fit has its origins in higher-order thought theory. the work of Anscombe (1957), but has been 67. Rosenthal’s (1990,p.742) argument for this developed in some detail and put to extensive requirement is the following. My awareness work mainly by Searle (1983). The idea is that of my bluish experience is an awareness of mental states divide into two main groups, the that particular experience, not of the general cognitive ones (paradigmatically, belief) and type of experience it is. But it is impossi- the conative ones (paradigmatically, desire). ble to represent a mental state as particu- The former are such that they are supposed to lar without representing in which subject it make the mind fit the way the world is (thus occurs. Therefore, the only way the higher- “getting the facts right”), whereas the latter order thought could represent my experience are such that they are supposed to make the in its particularity is if it represented it as world fit the way the mind is (a change in the occurring in me. world is what would satisfy them). 68. This is necessary to avert infinite regress. If 62. Kobes (1995) suggests a version of higher- the higher-order state was itself conscious, order monitoring theory in which the higher- it would have to be itself represented by a order representation has essentially a telic yet higher-order state (according to the the- direction of fit. But Rosenthal construes it as ory) and so the hierarchy of states would having only a thetic one. go to infinity. This is problematic on two 63. Carruthers (1989, 1996, 2000), and probably scores. Firstly, it is empirically implausible, also Dennett (1969, 1991), attempt to account and perhaps impossible, that a subject should for consciousness in terms of merely tacit entertain an infinity of mental states when- or dispositional higher-order representations. ever conscious. Secondly, if a mental state’s But these would not do, according to Rosen- being conscious is explained in terms of thal. The reason for this is that a merely dis- another conscious states, the explanation is positional representation would not make the “empty,” inasmuch as it does not explain con- subject aware of her conscious state, but only sciousness in terms of something other than disposed to being aware of it, whereas the cen- consciousness. tral motivation behind the higher-order mon- 69. This claim can be made on phenomenological itoring view is the fact that conscious states grounds, instead of on the basis of conceptual are states we are aware of having (Rosenthal analysis. For details, see Kriegel (2004b). 1990,p.742). 70. To repeat, the conceptual grounds are the fact that it seems to be a conceptual truth that P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

philosophical theories of consciousness: contemporary western perspectives 61

conscious states are states we are aware of 2002, 2006), Kobes (1995), Kriegel (2003a, having. This seems to be somehow inherent 2005, 2006a), and Van Gulick (2001, 2004). in the very concept of consciousness. 79. For fuller discussion of Brentano’s account, 71. There are other arguments that have been see Caston (2002), Kriegel (2003a), Smith leveled against the higher-order monitoring (1986, 1989) Thomasson (2000), and Zahavi theory, or specific versions thereof, which I (1998, 2004). do not have the space to examine. For argu- 80. So the self-representational content of con- ments not discussed here, see Block (1995), scious states is de se content. There are places Caston (2002), Dretske (1995), Guzeldere where Brentano seems to hold something like (1995), Kriegel (2006a), Levine (2001), Nat- this as well. See also Kriegel (2003a). 1993 1988 1999 soulas ( ), Rey ( ), Seager ( ), and 81. For more on the distinction between content 1998 Zahavi and Parnas ( ). and attitude (or mode), see Searle (1983). 72. The argument has also been made by Caston For a critique of Smith’s view, see Kriegel (2002), Levine (2001), and Seager (1999). For (2005a). a version of the argument directed at higher- 82. A similar account would be that conscious order perception theory (and appealing to states are not conscious in virtue of stand- higher-order misperceptions), see Neander ing in a certain relation to themselves, but 1998 ( ). this is because their secondary intentionality ∗ 73. Note that M does not merely misrepresent should be given an adverbial analysis. This M to be F when in reality M is not F, but is not to say that all intentionality must be misrepresents M to be F when in reality there treated adverbially. It may well be that the is no M at all. primary intentionality of conscious states is a 74. This would be a particular version of the sup- matter of their standing in a certain informa- position we made at the very beginning of this tional or teleological relation to their primary chapter, by way of analyzing creature con- objects. Thus, it need not be the case that S’s sciousness in terms of state consciousness. conscious fear that p involves S’s fearing p- 75. Furthermore, if M∗ were normally conscious, ly rather than S’s standing in a fear relation the same problem would arise with the third- to the fact that p. But it is the case that S’s order representation of it (and if the third- awareness of her fear that p involves being order representation were normally con- aware fear-that-p-ly rather than standing in scious, the problem would arise with the an awareness relation to the fear that p.To fourth-order state). To avert infinite regress, my knowledge, nobody holds this view. the higher-order monitoring theorist must 83. A constitutive, non-contingent relation is a somewhere posit an unconscious state, and relation that two things do not just happen to when she does, she will be unable to claim entertain, but rather they would not be the that that state instantiates the property of things they are if they did not entertain those being conscious when it misrepresents. relations. Thus A’s relation to B is constitutive 76. This appears to be Rosenthal’s latest stance if bearing it to B is part of what constitutes A’s on the issue (in conversation). being what it is. Such a relation is necessary 77. There are surely other ways the higher-order rather than contingent, since there is no pos- monitoring theorist may try to handle the sible world in which A does not bear it to B – problem of targetless higher-order represen- for in such a world it would no longer be A. tations. But many of them are implausible, 84. Elsewhere, I have defended a view similar in and all of them complicate the theory con- key respects to Van Gulick’s – see Kriegel siderably. One of the initial attractions of the (2003a, 2005, 2006a). theory is its clarity and relative simplicity. 85. Indeed, the problem may be even more press- Once it is modified along any of the lines ing for a view such as the higher-order global sketched above, it becomes significantly less states theory. For the latter requires not only clear and simple. To that extent, it is consid- the ability to generate higher-order contents, erably less attractive than it initially appears. but also the ability to integrate those with the 78. See Brook and Raymont (2006), Caston right lower-order contents. (2002), Hossack (2002), Kriegel (2003b), and 86. For a more elaborate argument that self- Williford (2006). For the close variation, see representation may not be a sufficient condi- Carruthers (2000, 2006), Gennaro (1996, tion for consciousness, one that could provide P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

62 the cambridge handbook of consciousness

a reductive explanation of it, see Levine Armstrong, D. M. (1978). A theory of univer- (2001, Ch. 6). sals, Vol. 2. Cambridge: Cambridge University 87. I am appealing here to a distinction defended, Press. e.g., by Cummins (1979), Dretske (1988), Armstrong, D. M. (1981). What is consciousness? and Searle (1992). Grice noted that some In D. M. Armstrong (Ed.), The nature of mind things that exhibit aboutness of meaningful- (pp. 55–67). Ithaca, NY: Cornell University ness, such as words, traffic signs, and arrows, Press. do so only on the assumption that someone Anscombe, G. E. M. (1957). Intention. Oxford: interprets them to have the sort of meaning Blackwell. they have. But these acts of interpretation Baars, B. (1988). A cognitive theory of consciousness. are themselves contentful, or meaningful. So New York: Cambridge University Press. their own meaning must be either derived by Baars, B. (1997). In the theater of consciousness: The further interpretative acts or be intrinsic to workspace of the mind. Oxford: Oxford Univer- them and non-derivative. Grice’s claim was sity Press. that thoughts and other mental states have 1990 an aboutness all their own, independently of Block, N. J. ( a). Inverted earth. Philosophical 52 79 any interpretation. Perspective, 4, – . 1990 88. This is denied by Dennett (1987), who claims Block, N. J. ( b). Can the mind change that all intentionality is derivative. the world? In G. Boolos (Ed.), Meaning and method: Essays in honor of .New 89. One might claim that such states are York: Cambridge University Press. less clearly conceivable when their self- 1995 representational content is fully specified. Block, N. J. ( ). On a confusion about the Thus, if the content is of the form, “I myself function of consciousness. Behavioral and Brain 227 247 am herewith having this very bluish experi- Sciences, 18, – . ence,” it is less clearly the case that one can Block, N. J. (1996). Mental paint and mental latex. conceive of an unconscious state having this Philosophical Issues, 7, 19–50. content. Block, N. J., Flanagan, O., & Guzeldere, G. (Eds.). 90. The conceivability of unconscious self- (1997). The nature of consciousness: Philosophi- representing states may not be proof of their cal debates. Cambridge, MA: MIT Press. possibility, but it is evidence of their possibil- Brentano, F. (1973). Psychology from empirical ity. It is therefore evidence against the self- standpoint (O. Kraus, Ed. & A. C. Rancurello, representational theory. D. B. Terrell, & L. L. McAlister, Trans.). 91. The reductivist may claim that zombies with London: Routledge and Kegan Paul. (Original the same physical properties we have are con- work published 1874) ceivable only because we are not yet in a posi- Brook, A., & Raymont, P. (2006). A unified tion to focus our mind on the right physical theory of consciousness. Cambridge, MA: MIT structure. As progress is made toward identi- Press. fication of the right physical structure, it will Brueckner, A., & Berukhim, E. (2003). McGinn become harder and harder to conceive of a on consciousness and the mind-body problem.” zombie exhibiting this structure but lacking In Q. Smith & D. Jokic (Eds.), Consciousness: all consciousness. New philosophical perspectives. Oxford: Oxford 92. For comments on an earlier draft of this University Press. chapter, I would like to thank George Byrne, D. (1997). Some like it HOT: Conscious- Graham, David Jehle, Christopher Mal- ness and higher order thoughts. Philosophical oney, Amie Thomasson, and especially David Studies, 86, 103–129. Chalmers. Byrne, A. (2001). Intentionalism defended. Philo- sophical Review, 110, 199–240. Carruthers, P. (1989). Brute experience. Journal References of Philosophy, 85, 258–269. Carruthers, P. (1996). Language, thought, and con- Armstrong, D. M. (1968). A materialist theory of sciousness. Cambridge: Cambridge University the mind. New York: Humanities Press. Press. P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

philosophical theories of consciousness: contemporary western perspectives 63

Carruthers, P. (1998). Natural theories of con- Dennett, D. C. (1991). . sciousness. European Journal of Philosophy, 6, Cambridge, MA: MIT Press. 203–222. Dennett, D. C. (1995). Darwin’s dangerous idea. Carruthers, P. (1999). Sympathy and subjectiv- New York: Simon and Schuster. ity. Australasian Journal of Philosophy, 77, 465– Dretske, F. I. (1981). Knowledge and the flow of 482. information. Oxford: Clarendon. Carruthers, P. (2000). Phenomenal consciousness. Dretske, F. I. (1988). Explaining behavior. Cam- New York: Cambridge University Press. bridge MA: MIT Press. Carruthers, P. (2006). Conscious experience ver- Dretske, F. I. (1993). Conscious experience. sus conscious thought. In U. Kriegel & K. Willi- Mind, 102, 263–283. ford (Eds.), Consciousness and self-reference. Dretske, F. I. (1995). Naturalizing the mind. Cam- Cambridge, MA: MIT Press. bridge MA: MIT Press. 1966 Castaneda,˜ H.-N. ( ). ‘He’: A study in the Fodor, J. A. (1974). Special sciences. Synthese, 28, 130 157 logic of self-consciousness. Ratio, 8, – . 97–115. 2002 Caston, V. ( ). Aristotle on consciousness. Foster, J. (1982). The case for idealism. London: 751 815 Mind, 111, – . Routledge. 1995 Chalmers, D. J. ( ). Facing up to the prob- Gennaro, R. J. (1996). Consciousness and self- lem of consciousness. Journal of Consciousness consciousness. Philadelphia: John Benjamin. Studies, 2 , 200–219. Gennaro, R. J. (2002). Jean-Paul Sartre and the 1996 Chalmers, D. J. ( ). The conscious mind. HOT theory of consciousness. Canadian Jour- Oxford: Oxford University Press. nal of Philosophy, 32, 293–330. 2002 Chalmers, D. J. ( a). Consciousness and its Gennaro, R. J. (2006). Between pure self- place in nature. In D. J. Chalmers (Ed.), Philos- referentialism and (extrinsic) HOT theory. In ophy of mind. Oxford: Oxford University Press. U. Kriegel & K. Williford (Eds.), Conscious- Chalmers, D. J. (2002b). The components of con- ness and self-reference. Cambridge, MA: MIT tent. In D. J. Chalmers (Ed.), Philosophy of Press. mind. Oxford: Oxford University Press. Goldman, A. (1993a). The psychology of folk psy- Churchland, P. M. (1979). Scientific realism and chology. Behavioral and Brain Sciences, 16, 15– the plasticity of mind. New York: Cambridge 28. University Press. Goldman, A. (1993b). Consciousness, folk psy- Churchland, P. M. (1985). Reduction, , and chology, and cognitive science. Consciousness the direct introspection of brain states. Journal and , 2 , 364–383. 8 28 of Philosophy, 82, – . Grice, P. (1957). Meaning. Philosophical Review, Crick, F., & Koch. C. (1990). Towards a neurobio- 66, 377–388. logical theory of consciousness. Seminars in the Guzeldere, G. (1995). Is consciousness the per- 263 275 Neurosciences, 2 , – . ception of what passes in one’s own mind? Crick, F., & Koch. C. (2003). A framework for In T. Metzinger (Ed.), Conscious experience. consciousness. Nature Neuroscience, 6, 119–126. Padborn: Schoeningh-Verlag. Cummins, R. (1979). Intention, meaning, and Harman, G. (1990). The intrinsic quality of truth conditions. Philosophical Studies, 35, experience. Philosophical Perspectives, 4, 31– 345–360. 52. DeBellis, M. (1991). The representational con- Horgan, T., & Tienson, J. (2002). The intention- tent of musical experience. Philosophy and Phe- ality of phenomenology and the phenomenol- nomenological Research, 51, 303–324. ogy of intentionality. In D. J. Chalmers (Ed.), Dennett, D. C. (1969). Consciousness and content. . Oxford: Oxford University London: Routledge. Press. Dennett, D. C. (1981). Towards a cognitive the- Hossack, K. (2002). Self-knowledge and con- ory of consciousness. In D. C. Dennett (Ed.), sciousness. Proceedings of the Aristotelian Soci- Brainstorms. Brighton: Harvester. ety, 163–181. Dennett, D. C. (1987). The intentional stance. Jackson, F. (1984). Epiphenomenal qualia. Philo- Cambridge, MA: MIT Press. sophical Quarterly, 34, 147–152. P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

64 the cambridge handbook of consciousness

Kim, J. (1989a). The myth of nonreductive mate- Kripke, S. (1980). The identity thesis. Reprinted rialism. Proceedings and Addresses of the Amer- in N. J. Block, O. Flanagan, & G. Guzeldere ican Philosophical Association, 63, 31–47. (Eds.), (1997). The nature of consciousness: Kim, J. (1989b). Mechanism, purpose, and Philosophical debates. Cambridge, MA: MIT explanatory exclusion. Philosophical Perspec- Press. tives, 3, 77–108. Lehrer, K. (1996). Skepticism, lucid content, and Kim, J. (1992). Multiple realization and the the metamental loop. In A. Clark, J. Ezquerro, of reduction. Philosophy and Phe- & J. M. Larrazabal (Eds.), Philosophy and cogni- nomenological Research, 52, 1–26. tive science. Dordrecht: Kluwer. Kobes, B. W. (1995). Telic higher-order thoughts Lehrer, K. (1997). Self-trust: A study of rea- and Moore’s paradox. Philosophical Perspec- son, knowledge, and autonomy. Oxford: Oxford tives, 9, 291–312. University Press. Kriegel, U. (2002a). Phenomenal content. Erken- Levine, J. (1983). Materialism and qualia: The ntnis, 57, 175–198. explanatory gap. Pacific Philosophical Quar- terly, 64, 354–361. Kriegel, U. (2002b). Emotional content. Con- 2001 sciousness and Emotion, 3, 213–230. Levine, J. ( ). Purple haze: The puzzle of con- sciousness. Oxford: Oxford University Press. Kriegel, U. (2002c). PANIC theory and the prospects for a representational theory of phe- Lewis, D. K. (1990). What experience teaches. In nomenal consciousness. Philosophical Psychol- W. G. Lycan (Ed.), Mind and cognition. Oxford: ogy, 15, 55–64. Blackwell. 1993 Kriegel U. (2003a). Consciousness, higher-order Lewis, D. K. ( ). Causal explanation. In content, and the individuation of vehicles. Syn- D.-H. Ruben (Ed.), Explanation. Oxford: these, 134, 477–504. Oxford University Press. Kriegel, U. (2003b). Consciousness as intransi- Libet, B. (1985). Unconscious cerebral initiative tive self-consciousness: Two views and an argu- and the role of conscious will in voluntary ment. Canadian Journal of Philosophy, 33, 103– action. Behavioral and Brain Sciences, 8, 529– 132. 566. Kriegel, U. (2004a). The new mysterianism and Loar, B. (1990). Phenomenal states. Philosophical the thesis of cognitive closure. Acta Analytica, Perspectives, 4, 81–108. 18, 177–191. Lurz, R. (1999). Animal consciousness. Journal of Kriegel, U. (2004b). Consciousness and self- Philosophical Research, 24, 149–168. consciousness. The Monist 87, 185–209. Lurz, R. (2003). Neither HOT nor COLD: An Kriegel, U. (2005a). Naturalizing subjective alternative account of consciousness. Psyche, 9. character. Philosophy and Phenomenological Lycan, W. G. (1987). Consciousness. Cambridge, Research, 71, 23–57. MA: MIT Press. 2005 Kriegel, U. ( b). Review of Jeffrey Gray, Lycan, W. (1996). Consciousness and experience. Consciousness: creeping up on the hard problem. Cambridge, MA: MIT Press. Mind, 114, 417–421. Lycan, W. G. (2001). A simple argument for 2006 Kriegel, U. ( a). The same-order mon- a higher-order representation theory of con- itoring theory of consciousness In U. sciousness. Analysis, 61, 3–4. Kriegel & K. Williford (Eds.), Conscious- McGinn, C. (1989). Can we solve the mind-body ness and self-reference. Cambridge, MA: MIT problem? Mind, 98, 349–366. Press. McGinn, C. (1995). Consciousness and space. Kriegel, U. (2006b). The concept of conscious- Journal of Consciousness Studies, 2 , 220– ness in the cognitive sciences: Phenomenal con- 30. sciousness, access consciousness, and scientific 1999 practice. In P. Thagard (Ed.), Handbook of the McGinn, C. ( ). The mysterious flame. Cam- philosophy of psychology and cognitive science. bridge, MA: MIT Press. Amsterdam: North-Holland. McGinn, C. (2004). Consciousness and its objects. Kriegel, U., & Williford, K. (eds.). (2006). Oxford: Oxford University Press. Self-representational approaches to consciousness. Mackie, J. L. (1977). Ethics: Inventing right and Cambridge, MA: MIT Press. wrong. New York: Penguin. P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

philosophical theories of consciousness: contemporary western perspectives 65

Maloney, J. C. (1989). The mundane matter of the Macmillan encyclopedia of cognitive science. New mental language.New York: Cambridge Univer- York: Macmillan. sity Press. Seager, W. (1999). Theories of consciousness. Mellor, D. H. (1978). Conscious belief. Proceed- London: Routledge. ings of the Aristotelian Society, 78, 87–101. Searle, J. R. (1983). Intentionality: An essay in the Moore, G. E. (1903). The refutation of ideal- philosophy of mind. New York: Cambridge Uni- ism. In G. E. Moore (Ed.), Philosophical papers. versity Press. London: Routledge and Kegan Paul. Searle, J. R. (1992). The rediscovery of mind. Nagel, T. (1974). What is it like to be a bat? Philo- Cambridge, MA: MIT Press. sophical Review, 83, 435–450. Segal, G. (2000). A slim book on narrow content. Natsoulas, T. (1993). What is wrong with Cambridge, MA: MIT Press. appendage theory of consciousness? Philosoph- Shoemaker, S. (1994a). Phenomenal character. 137 154 ical Psychology, 6, – . Nous, 28, 21–38. 1998 Neander, K. ( ). The division of phenomenal Shoemaker, S. (1994b). Self-knowledge and labor: A problem for representational theories ‘inner sense.’ Lecture III: The phenomenal of consciousness. Philosophical Perspectives, 12, character of experience. Philosophy and Phe- 411 434 – . nomenological Research, 54, 291–314. 1990 Nemirow, L. ( ). and the cogni- Shoemaker, S. (1996). Colors, subjective reac- tive role of acquaintance. In W. G. Lycan (Ed.), tions, and qualia. Philosophical Issues, 7, 55–66. Mind and cognition. Oxford: Blackwell. Shoemaker, S. (2002). Introspection and phe- Peacocke, C. (1983). Sense and content. Oxford: nomenal character. In D. J. Chalmers (Ed.), Clarendon. Philosophy of mind. Oxford: Oxford University 1967 Putnam, H. ( ). The nature of mental states. Press. In D. M. Rosenthal (Ed.), The nature of mind. Siewert, C. P. (1998). The significance of conscious- Oxford: Oxford University Press. ness. Princeton, NJ: Princeton University Press. Putnam, H. (1975). The meaning of ‘meaning.’ In Smart, J. J. C. (1959). Sensations and brain pro- H. Putnam (Ed.), Mind, language, and reality, cesses. Philosophical Review, 68, 141–156. New York: Cambridge University Press. Smith, D. W. (1986). The structure of (self-) con- Rey,G.(1988). A question about consciousness. sciousness. Topoi, 5, 149–156. In H. Otto & J. Tueidio (Eds.), Perspectives on mind. Norwell: Kluwer Academic Publishers. Smith, D. W. (1989). The circle of acquaintance. Rey, G. (1998). A narrow representationalist Dordrecht: Kluwer Academic Publishers. account of qualitative experience. Philosophi- Smith, D. W. (2004). Return to consciousness. cal Perspectives, 12, 435–457. In D. W. Smith (Ed.), Mind world. New York: Rosenthal, D. M. (1986). Two concepts of Cambridge University Press. consciousness. Philosophical Studies, 94, 329– Thau, M. (2002). Consciousness and cognition. 359. Oxford: Oxford University Press. Rosenthal, D. M. (1990). A theory of con- Thomasson, A. L. (2000). After Brentano: A one- sciousness. ZiF Technical Report 40, Bielfield, level theory of consciousness. European Journal Germany. of Philosophy, 8, 190–209. Rosenthal, D. M. (1993). Thinking that one Tye, M. (1986). The subjective qualities of expe- thinks. In M. Davies & G. W. Humphreys rience. Mind, 95, 1–17. (Eds.), Consciousness: Psychological and philo- Tye, M. (1992). Visual qualia and visual content. sophical essays. Oxford: Blackwell. In T. Crane (ed.), The contents of experience. Rosenthal, D. M. (2000). Consciousness and New York: Cambridge University Press. metacognition. In D. Sperber (Ed.), Metarep- Tye, M. (1995). Ten problems of consciousness. resentation. Oxford: Oxford University Press. Cambridge, MA: MIT Press. Rosenthal, D. M. (2002a). Explaining conscious- Tye, M. (2000). Consciousness, color, and content. ness. In D. J. Chalmers (Ed.), Philosophy of Cambridge, MA: MIT Press. mind. Oxford: Oxford University Press. Tye, M. (2002). Visual qualia and visual content Rosenthal, D. M. (2002b). Consciousness and revisited. In D. J. Chalmers (Ed.), Philosophy of higher-order thoughts. In L. Nadel (Ed.), mind. Oxford: Oxford University Press. P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

66 the cambridge handbook of consciousness

Van Gulick, R. (1993). Understanding the phe- Wegner, D. M. (2002). The illusion of conscious nomenal mind: Are we all just armadillos? will. Cambridge, MA: MIT Press. Reprinted in N. J. Block, O. Flanagan, & Williford, K. W.(2006). The self-representational G. Guzeldere (Eds.), (1997). The nature structure of consciousness. In U. Kriegel & of consciousness: Philosophical debates. Cam- K. Williford (Eds.), Consciousness and self- bridge, MA: MIT Press. reference. Cambridge, MA: MIT Press. Van Gulick, R. (2001). Inward and upward – Zahavi, D. (1998). Brentano and Husserl on self- reflection, introspection needed and self- awareness. Etudesˆ Ph´enom`enologiques, 27(8), awareness. Philosophical Topics, 28, 275– 127–169. 305. Zahavi, D. (1999). Self-awareness and alterity. Van Gulick, R. (2006). Mirror mirror – is that Evanston, IL: Northwestern University Press. all? In U. Kriegel & K. Williford (Eds.), Con- Zahavi, D. (2004). Back to Brentano? Journal of sciousness and self-reference. Cambridge, MA: Consciousness Studies, 11, 66–87. MIT Press. Zahavi, D., & Parnas, J. (1998). Phenomenal con- Velmans, M. (1992). Is human information pro- sciousness and self-awareness: A phenomeno- cessing conscious? Behavioral and Brain Sci- logical critique of representational theory. Jour- ences, 14, 651–669. nal of Consciousness Studies, 5, 687–705.