Anyone who is looking for a clear, concise and accurate lay of the land with respect to contemporary, analytic, theories of consciousness would do well to get hold of this book. Its first half contains a handy survey and critical assessment of current theories of (1) qualia, and (2) what awareness of qualia involves. Yet it is not a textbook. For its second half, beginning at Chapter five, develops a new, representationalist theory of consciousness. Building on the insightful, but underdeveloped, ideas of Gilbert Harman, Hill's main ambition is to defend a thorough-going representationalism about consciousness, while, along the way, refuting dualism and establishing that materialism, specifically a central state materialism sort, is still in good nick. He explores the implications of his new theory not just to the central cases of visual awareness and qualia, but also for pain, emotional experience and introspection.

Achieving all of that in a short work requires Hill to move through some rather complex, but also rather familiar material, at a good clip. Readers are required to keep up. Doubtless they will find that some topics are not treated in as much depth as they would like to make a fully compelling case; some possible replies have been overlooked. Nevertheless, Hill's analyses, which are almost always careful and even-handed, are well worth the attention, even if one is not entirely persuaded by the force of his objections or the truth of his positive proposal.

On the whole, Hill does a fine job of exposing the limitations and lacuna in the arguments of his opponents and in exposing the shaky assumptions upon which their theoretical edifices rest. He is less successful, however, in securing his main objective – i.e. of convincing the reader of the truth of representationalism. A major shortcoming, which he himself recognizes at the close of the book, is that:

It is impossible to comprehensively examine all the items on Hill's list of ‘reasons for believing in representationalism about consciousness’ in a short review. Still, consideration of at least some of these provides a useful way of highlighting important concerns about the overall success of his project.

The first item on this list is peculiar in light of the book's arguments. Far from being intuitively plausible Hill stresses that representationalism is a ‘highly theoretical’ position that is neither built into folk psychology nor immediately endorsed by introspection (112, see also 17). Indeed, much of his work takes the form of explaining away our intuitions i.e. in showing us that we are not obliged to take them at face value (112). Apart from this, it is hard to see why defenders of representationalism (or anyone) ought to place any great value on intuitions anyway. When it comes to deciding important matters in philosophy they are hardly authoritative; more often they are downright wrong and misleading. And Hill is not shy of noting this when criticizing the views of others. Thus in reviewing Moore's introspection-based conclusion that consciousness is ‘featureless’ and ‘metaphysically simple’ because it is diaphanous, Hill points out that Moore took this view because he was ‘presupposing the claim that what is transparent to introspection is metaphysically transparent’ (78). He immediately points out that this ‘has very little to recommend it, but Moore seems to have regarded it as obvious’ (79). This shows something more than that Moore's judgement was poorly based in this case, it underscores a general methodological problem with a tendency of philosophers to found arguments on what they find ‘intuitive’, ‘obvious’ or otherwise ‘natural to think’. Unless there is a special link between intuitions and truth such appeals fail to offer secure support for one's favoured positions. Intuitions look incapable of playing that sort of role – they are neither indefeasible evidence nor final arbiters – at best they indicate which claims some of us may find initially attractive. This leaves out the work of making a case for the truth of such claims that is logically and scientifically compelling.

The second item on Hill's list is, by contrast, quite pivotal. The book's motto might have been: Awareness doesn't come free. Hill takes this a step further and promotes Principle P: If an agent is aware of X, then the agent is in a mental state that represents X. Hill cites the successes of cognitive science as the key motivation for believing in this principle. We are told that those successes rest on making the assumption that ‘the mind is essentially representational in character’ (70). Hill not only bills this as an argument for his position, he identifies it as his primary argument (88). But it only goes through if his explanation of the successes of cognitive science is correct. There is no doubt that many traditional cognitive scientists are committed to representationalism. But our understanding of cognitive science and what makes it successful is still developing, thus knowing what really does the work in its best theories, and how to characterize it, is a matter of controversy in the philosophy of cognitive science. New developments in embodied and embedded cognition suggest that what early theories took to be essentially and paradigmatically representational phenomena – including basic forms of perception – are, in fact, nothing of the sort. As Ramsey (2007) observes, ‘something very interesting is taking place in cognitive science… cognitive science has taken a dramatic anti-representational turn’ (xiv–xv). This topic is too large to say more about it here; suffice to say that it would be bad news for Hill's representationalism and his argument for it if Ramsey's remark turns out to be true.

What of the third item? Hill attempts to provide additional backing for Principle P by casting doubt on competing hypotheses about experiential awareness, thus arguing for P by elimination. In line with this strategy he reviews Russell's account of acquaintance, one that is ‘innocent of all forms of representation’ (73); Moore's view that experience is featureless because it is diaphanous or transparent; and Ducasse's adverbalism (which is regarded as an ‘astonishing’ view). All of this is neatly brought up to date with a discussion of Campbell's relational view. Hill's fundamental complaint about all such approaches is that ‘It seems wrong to attribute all of the distinguishing features of every fact of perceptual consciousness to entities that count as objects of consciousness. How an object of consciousness appears to us sometimes depends, at least in part, on factors that lie on the subject side of the subject/object divide’ (83). This is a powerful criticism, yet the result of this treatment of rivals is less than conclusive because Hill's examination of possible competitors is not wholly exhaustive. He overlooks the possibility that both the relational and representational views might be false; i.e. that, which we might call, the responsive view (the idea that experiencing is to be understood in terms of how organisms systematically respond where such responding is non-representational) might be true.

The possibility of a non-representationalist understanding of experience becomes a useful foil in thinking about Hill's fourth reason. He tells us that representationalism is to be preferred over disjunctivism because the former, while not logically defective, forces us to accept a complex and inelegant explanation of veridical experiences and their hallucinatory counterparts. Representationalism is simpler in that it says ‘hallucinating p and seeing that p are fundamentally akin – they both involve a representation of the state of affairs p’ (81). But this hardly follows. Hallucinating p and seeing p might be fundamentally akin in the way things appear to a subject and even, and for that reason, in provoking the same sort of reactions and actions without assuming that any representations are involved in this process. This would be the case, for example, if there were good independent reasons to embrace a non-representational understanding of perception in the cognitive sciences more generally.

Moreover, by considering such cases it is possible to point to an uncomfortable tension in the sort of representationalist account that Hill offers. On the one hand, he says ‘Suppose that a perceptual representation R represents a property P. It is always possible for P to be instantiated without being represented by a corresponding token of R, and for, R to be tokened without there being a corresponding instance of P’ (175). This is a salient reminder about what representationalism requires. It is this principle that drives Hill to accept that there is always an appearance/reality gap (possibly, even a large and systematic one) between the intrinsic nature of what is represented and the features that are captured by representations of it (176). Here it helps to bear in mind his useful distinction between two conceptions of qualia at large in the literature. According to the first, qualia are ‘the intrinsic features of perceptual experience that are given in introspection’, according to the second they denote ‘the way things look, feel, smell and so on’ (58). Presented with these options Hill proposes that qualia should be identified with the latter, following Harman. This is consistent with his claim that we should think of representational contents as endowing perceptual experiences with their qualitative characters.

At the same time, he also tells us that qualia are wholly physical; they reduce to objective, physical properties (25). But here Hill seems to be using qualia to refer to what is represented – i.e. pain qualia as bodily disturbances, visual qualia as appearances, etc. – as opposed to how what is represented is represented. But it is not at all clear what entitles him to do so or why we should be inclined to follow suit.

It is easy enough to understand how both the veridical and hallucinatory might share representational content, if we buy into representationalism. Crudely, tokens of R will occur in both cases, but only in the former cases does P co-occur. As far as it goes, that seems fine. But once one accepts the necessity of the appearance/reality gap with respect to appearances it is not at all clear why we should believe that qualitative properties are also what is represented in veridical cases of experience (unless some kind of implausible resemblance theory of representation is also endorsed). Yet this seems to be what Hill supposes. Consider his example of what the veridical experience of pains involves. He says:

But, pace Hill, the claim of the last sentence is not strictly entailed by what comes before it. For one can surely accept that representational contents fix the qualitative characters of pain experiences without accepting that pains exist as such (understood as essentially qualitative properties – albeit physical ones – that are detectable in cases of veridical experience). Why should we believe in a more robust kind of physicalist qualia-realism? It is surely logically compatible with the story that Hill wants to tell that, for example, the experience of pain is all there is in existence that has qualitative properties. That would be so if the function of having experiences of such-and-such qualitative character is to lawfully covary or correlate with the presence of certain kinds of bodily disturbances (Mutatis mutandis for all those worldly things to be identified with Hill's so-called visual and emotional qualia). Underwriting such correlations may be all there is to the proper biological functioning of the perceptual systems that Hill appears to think lies at the basis of experience. If so, there is no reason to suppose that what it represented and how it is represented need share any properties in common.

Despite these concerns this is a valuable, provocative and carefully argued book. It is well worth wrestling with, even though – as I believe – it fails to show that representationalism about consciousness is a foregone conclusion.