Reading Tye's new book reminded me of slowly sipping a good specimen of a dry vodka Martini. In both cases much is accomplished by the skilful assembly of only a few key ingredients. I don't really like dry vodka Martinis, though, and similarly I found many of the thoughts offered by Consciousness Revisited to be too bitter to swallow. A sophisticated piece of work, however, it certainly is.

The main argumentative thread is as follows: Tye works to develop an account of perception from which he is able to give a particular diagnosis of what the phenomenal character of a perceptual experience comes to; that qualitative character we seem easily able to detect on introspection, and about which certain sorts of error appear, at least prima facie, to be impossible for the introspecting subject. Bringing in one or two more tools, notably a Russellian distinction between factual and acquaintance knowledge, Tye takes himself to be able to offer a new defence against the by now traditional thorns in the physicalist side: Jackson's Mary, the explanatory gap/‘hard’ problem, and zombies. The availability of this new defence obviates the need for the physicalist to rely on the popular ‘phenomenal concept strategy’, of which Tye has hitherto been a major proponent. Indeed Tye in this new work even offers a considerable amount of argumentation against the notion that there exist phenomenal concepts at all, in any distinctive and important sense. In the form of the slimline conception of conscious perceptual experience and its content that he puts forward, Tye considers that he has all the materials needed to rebut the anti-physicalist considerations.

I have objections to Tye's story at each stage. I will critique his account of perception and the content of consciousness, since I think this wrongheaded in its own right. A more serious problem for Tye's project, however, is that even if his views on perception and phenomenal content are granted, it is less obvious than Tye thinks how this serves to take the sting out of the anti-physicalist thought experiments. Given Tye's rather convincing case against the phenomenal concept strategy, therefore, the main upshot of this book is apparently to leave Tye's ‘thoroughly modern materialist’ in something of a sticky situation.

So far as I understand it, and I cannot claim to have found the book an easy read,Footnote ¹ Tye's view of what goes on in perception is the outcome of trying to accommodate two intuitive and popular theses: first that experience is transparent, and second that we are in some direct way aware of the phenomenal character of our experiences just by introspecting them. Transparency is the idea that when we introspectively examine the qualities of an experience we are undergoing, for example the visual experience as of a red fire engine before us, we do not become aware of the qualities of the experience independently of becoming aware of the qualities that the experience represents as belonging to the world outside: if you try to isolate the phenomenal qualities of your experience as of the red fire engine, what you end up doing is just attending to the redness of the fire engine. There do not appear floating phenomenal instantiations existing apart from the way the experience tells you the world is. More snappily put, perceptual experience is transparent, one ‘sees’ straight through it to the world. But the thesis that we are directly aware of the phenomenal character of our experiences might seem to conflict with the transparency thesis: if what we're in contact with when we introspect is the phenomenal character of our experience, then how can transparency also hold? For doesn't the phenomenal character of the experience precisely get in the way, by being in between us and the way the world is?

To reconcile these apparently clashing theses, Tye adopts an elegant and surprising solution; in this respect it reminds me, in fact, of the way that Davidson deduces the identity of the mental with the physical from the anomalousness of the mental plus the claim that all cause and effect is law-governed. Tye's way out of his dilemma is to identify the phenomenal character we are directly aware of in introspection with the qualities of external objects. As he says ‘The phenomenal character of the experience of red just is red [the property of our fire engine]. In being aware of red [the property of our fire engine], I am aware of what it is like to experience red’ (120) Thus what Tye is endorsing is a form of direct realism: perceptual experiences include their objects and the qualities these have. More precisely, the view on offer is a disjunctivism: in the veridical case perceptual experiences are partially composed of the objects/qualities perceived, and in the hallucinatory case the experience appears to contain a suitable object, but in reality does not. For Tye veridical perceptual states have singular contents into which the perceived object literally enters (it fills a ‘slot’ in the experience). In the case of hallucinations the relevant slot is unfilled, although it appears filled: the content is therefore a ‘gappy’ one.

I find two features of this account puzzling, one mildly puzzling but the second seriously so. My first puzzlement perhaps just expresses an insufficiently supple philosophical mind, but I don't quite grasp the sense in which Tye considers the present theory a variety of intentionalism, in addition to being disjunctivism. Intentionalism, at a minimum, holds that the phenomenal content of an experience supervenes on its representational content. So Tye will say that two relevantly similar perceptual experiences, into which the same worldly object enters, will be phenomenally identical. This seems correct; what I can't make proper sense of is Tye's talk of perceptual experiences, on his conception, having representational content at all. If I follow, then the content of my veridical visual perception of the fire engine is the fire engine itself, at least in part. The experience actually contains the fire engine. But whence the representational element? It doesn't seem correct to say that the fire engine present in my experience represents itself there to me. For there is no need: the fire engine is simply there, present in my experience. Being present, it doesn't need to represent itself to me. Now, it is true that I may, for example, represent myself in court, certainly. But in this case what I do is to partition myself into two aspects: I am present qua defendant (say) and present also qua counsel for the defence. Here it is easy enough to understand that one of my modes of being present in the court represents the other mode, in the relevant sense. But this model doesn't seem transferable to the fire engine perception. It is not the case that we capture the idea of the fire engine being present in the street and being present in my experience of it in terms of distinct aspects or ways of being of the fire engine, one of which stands in for the other, so to speak. For the whole point of saying that an object itself figures in one's experience is to say that the object itself, i.e. just as it is, figures in the experience. There is no duality here to be invoked, and correspondingly I find the talk of the representational content of the experience hard to fathom.

The second puzzle is this. Hallucinatory experiences may have the same phenomenal character as veridical ones, says disjunctivist Tye, despite differing in their contents. The idea is fine until spelled out using the mechanics of Tye's account. In the veridical case of an experience as of a red fire engine, the relevant slot in the experience is filled by the red fire engine. This makes it a simple matter to account for the phenomenal redness of the experience, its being of a fire engine and so on. But now what of the hallucinatory case? Here the fire engine-sized slot is unfilled: that's what makes it an hallucination. But I still have an experience as of a red fire engine; in fact I have an experience indistinguishable from an experience where there is a whopping great red fire-engine filling my perceptual slot. How can this be? What is it that gives me the hallucinatory experience as of a red fire engine, and, much more seriously, how does having an experiential state just like that present in the veridical case except that there is a fire-engine shaped hole in it give me an experience indistinguishable from the case where that hole is filled? This is a metaphysically bizarre picture. To see clearly why, note that it seems well suited to an argument to the effect that hallucinations cannot be indistinguishable from veridical perception. It seems to me that Tye has serious further work to do towards a satisfying account of the possibility of shared phenomenal character between veridical experience and hallucination.

Now on to consciousness and physicalism. For ease I will stick to Tye's treatment of Jackson's Mary, but the points I make can be extended to his treatments of the other anti-physicalist devices. Mary the monochrome scientist apparently poses a problem for physicalism by knowing all the physical facts about seeing red without knowing what it is like to see red. When she finds out what it's like to see red she seems to acquire a new bit of factual knowledge, which involves, ex hypothesi, a non-physical fact.Footnote ² According to the phenomenal concept strategy, Mary learns nothing new about the world when she sees red, she merely acquires a new way of conceptualising some physical facts she already knew about – albeit in a more boring way – from within her black and white laboratory. Tye, previously a notable advocate of the phenomenal concept strategy, is (I'm glad to say) now impressed by the idea that when Mary sees red she ‘comes to know new worldly things – that is, things she did not know before’ (129). Physicalism must accommodate this fact or perish, opines Tye.

The solution offered comprises two elements: first, as a consequence of the account already given of perception and the content of experience, the property that Mary does not know, the ‘redness’, is pushed from the mind into the world: the red Mary learns about is not a property of her experience as such, but a property of the facing surfaces of things like red fire engines. This move, Tye thinks, is going to make the identification of redness with a physical property less problematic. The second element of Tye's response to Mary is to invoke, following Russell, a distinction between factual-knowledge and acquaintance- (or thing-) knowledge. We arguably make such a distinction in speech quite routinely: we talk about knowing lots (of facts) about Seville (from guidebooks and TV, say) as opposed to knowing Seville itself, in the sense of being acquainted with it – having directly encountered it. We talk in a similar way about individual people. The gap in Mary's knowledge within her room then, according to Tye, concerns not facts about seeing red. Rather, Mary is simply yet to be acquainted with red, due to not seeing it yet: what she lacks is acquaintance knowledge of the property red. This move promises to benefit the physicalist cause by accommodating the intuition that Mary really confronts some novel feature of the world, without granting that she learns any new, and so non-physical, fact.

The first thing to note about Tye's answer to Mary, I think, is that the first of the two elements that comprise it appears entirely useless. To see this imagine that the first element was not present, and that Tye's thesis about Mary was simply that she comes to be acquainted with a new property of phenomenal experience through seeing red, but does not learn any new facts about the world by doing so. It seems to me that this response, essentially that proposed by Connee some time ago, is every bit as forceful as Tye's two-component update. One might think, along with Tye, that making redness a property of surfaces rather than a property of experiences makes a physical identification of the property easier to broach: we might now hypothesise that red is identical to some surface reflectance property, for example, whereas we were previously at a loss for a candidate brain property with which to identify red qualia, when we thought there were such things. And won't the proposed identity now be easier to swallow, whereas identifying phenomenal property with physical property just seemed so, well, implausible? But this line of thought is delusory. As Tye notes, there will be no perspicuous and satisfying account of why a given surface reflectance property is identical with red as opposed to, say, blue: there will not be any neat a priori or conceptual transition from a story about reflectance properties to a story about colours, of the sort that Chalmers and Jackson claim physicalists are committed to. We will have to make do with correlations to build identity statements, says Tye, just as scientists normally do. This being so – not that I agree that it is so – there is nothing gained by shifting redness from the mind into the world: the physicalism-preserving identity to be proposed remains every bit as unsatisfying and brute as it ever did before.

The real work in defusing Mary is being done, it follows, by the second element of Tye's account. But this element also faces its problems. In particular, I think, there is this problem: normally new acquaintance knowledge comes with some new factual knowledge; in fact, more strongly, it is hard to imagine gaining acquaintance knowledge without thereby also gaining factual knowledge. Consider Seville again. It is not just because a certain metaphysical relation has failed to hold between you and the city that I say to you that cribbing from guidebooks and watching holiday programs is really no substitute for going to Seville itself. What I have in mind is that you'll know the place better if you go there, and this doesn't just mean that you'll become acquainted with it in the sense of meeting with it: the thought is that there are some facts about Seville that you won't get short of stepping out in Seville. Similarly, if acquaintance with red would add to Mary's knowledge at all, it seems irresistible to think that this will in part consist in Mary learning some new fact concerning red. In this case the glaring fact there to be learned through her meeting with red is what it is like to see red. Let me sum up this objection: How can Tye persuade us that Mary can become acquainted with red without learning any new fact about what it is to see red? He makes no effort to do this in Consciousness Revisited. He simply accepts that Mary can know all the physical facts about seeing red, and then add pure acquaintance knowledge to her epistemic stock by seeing red. But this does not conform, far from it, to the normal model of acquaintance knowledge, which Tye invokes. Something further of a justificatory sort therefore needs to be said. Without the phenomenal concept strategy, the most plausible thing to say on Tye's position seems to be that Mary's acquaintance with red comes with some new factual knowledge about the nature of seeing red, a result that falsifies physicalism.