In her book Locating Consciousness (1995), Hardcastle argues that an empirical, materialistic theory of consciousness can and will emerge from an interdisciplinary Cognitive Science, reviews a broad range of empirical and philosophical literature on consciousness, and defends a "multiple memory model" of consciousness. Hardcastle's defence of the general possibility of an empirical theory of consciousness is convincing, as is her argument for the necessity of a multilevel interdisciplinary approach. Her defence of the multiple memory model is suggestive but not entirely compelling.
2. Hardcastle sets three principal objectives in this volume: First, she argues that an empirical, scientific, broadly materialist theory of consciousness is possible. Secondly, she attempts to draw together and to integrate information processing and neuroscientific research programmes and results pertaining to consciousness. Third, she defends a particular theory: the multiple memory model of consciousness, according to which consciousness is instantiated in the semantic memory system, and is present when and only when that system is activated. She is extremely successful in the first project, richly suggestive with regard to the second and ultimately unconvincing with regard to the third, despite calling our attention to a plethora of relevant information and offering a number of interesting and original integrative insights and critical discussions of other unsuccessful attempts to solve this conundrum at the heart of cognitive science.
3. Hardcastle clears the ground by arguing in the first chapter for what she calls "naturalism about subjective experience." This is the doctrine that natural science can provide a theory and explanation of subjective experience with all of the descriptive and explanatory power of any other theory or explanation in any other domain of science--that there is nothing special about the domain of experience excepting it from the purview of science and ordinary scientific methodology. Most of the argument in this chapter is quite compelling. The bulk of it is directed at undermining arguments against this position, as Hardcastle correctly observes that the burden of proof here is squarely on the shoulders of those who would posit an exception to naturalism. But there is also some solid positive argument for the applicability of science to the inner.
4. I note, however, a few small difficulties with Hardcastle's defence. On p. 7 she says, "I maintain that when we introspect, we perceive some appearance and then make a judgment about that experience." This move is central to her argument for the conclusion that the deliverances of common sense introspection are not the last word regarding the nature of consciousness or its contents, and hence that science may displace rather than simply explain or reduce our commonsense notions in this domain. Hardcastle borrows this Lockean two-stage view of introspection--and indeed this argument--from Churchland (1979) and Rorty (1979). I have criticised this view at length elsewhere (Garfield, 1989). Suffice it to say here that, as Hardcastle herself implicitly acknowledges elsewhere, particularly in her discussion of perception in chapter 5, the idea of a conceptually uncontaminated perceptual experience, accomplished prior to the operation of judgment is as bankrupt in the domain of inner experience as it is in the domain of outer perception. Hardcastle is at pains both to assimilate introspection to perception and to abandon the notion of a pre-conceptual perceptual given in perception. These are both laudable epistemological goals. (p. 8) But then she should abandon the commitment to such a given in introspection as well. In fact, of course, this would only strengthen her hand, in virtue of pulling up the last anchor to privilege on which her mysterian opponent might depend in order to preserve the deliverances of naive introspection against the tide of science.
5. A second part of this discussion that remains unsatisfactory is that in which Hardcastle replies to the mysterian claim that no amount of scientific theory concerning a subjective state could tell us what it is like to be in that state (pp. 14-18). She replies in several perplexing ways. First, she argues that a sufficiently robust scientific description could enable us to imagine what it would be like to be in such a state, much as a robust description of the American Civil War could allow us to imagine what it would be like to be a soldier in that war (p 15). Second, she argues that a robust enough scientific description would allow us to put our own brains into a similar enough state to the one we are describing that we would then be in a position from the first person perspective to know what it feels like to be in that state (p 16). Third, she argues that all languages have descriptive lacunae, and so the fact that the language of science might not give us a description of subjective experience should not count against it any more than the inability to represent quantificational inferences should count against the propositional calculus (pp 17-18).
6. All three of these arguments miss the opponent's point badly. With regard to the first, the mysterian opponent will simply deny that a description of the American Civil War alone could ever allow us the imaginative knowledge Hardcastle claims it could engender. Rather, he will reply, what allows us that knowledge is the fact that we ourselves have had relevantly similar conscious experiences. Without that similar subjective constitution, he would argue, the external description could never provide an entree into the subjective realm. And since it is just this possibility that is at issue between Hardcastle and her opponent, she appears to beg the question.
7. The second argument appears to concede the point completely to the opponent. For the opponent argues that no scientific description or explanation could provide insight into subjective experience, simply because in order to know what any subjective experience is like you have to have had it. The fact that science can tell you how to produce it is, even if true, nothing to the point.
8. Likewise, the third argument must fail to impress. The mysterian opponent does not deny that all languages are expressively incomplete; nor does he think that scientific language is abominable on grounds of its incompleteness with respect to the subjective, but only that it IS incomplete in this domain. Hardcastle's attempt at a reply merely concedes all that is at issue.
9. None of this is necessary, however. Hardcastle has ample resources to reply to this mysterian challenge. All she needs to do is to point out that the opponent begs an important question: The opponent simply presumes that there is no possible third-person scientific description of any inner experience on the grounds that he has never seen one. No current description of any event in computational or neuropsychological terms, he says, tells him what it is like to see a ripe tomato. Fair enough. In order for such a description to do that, we would have to have enough background theory for that "what it is like" to be understood in such terms. A medieval alchemist saying that no description of a relation between atoms of two gases could ever explain what it is to be the element water would be right. Without enough theory, the "what it is to be water" is not explained by saying that it is H20. With enough theory, there is no problem. Mutatis mutandis, probably, with conscious experience. This is not, of course a positive argument to the conclusion that such a theoretical success IS forthcoming, but only a disarming of the mysterian argument to the conclusion that none CAN BE forthcoming. But that is all that needs to be accomplished at this stage, and indeed this is all that Hardcastle attempts with her three failed arguments.
10. Hardcastle offers another needlessly problematic argument for a true conclusion that can be (and is elsewhere in this chapter) defended using successful arguments in chapter 2. Here (pp. 23-24) she argues against privileging current conceptions of consciousness on the grounds that "other languages, including ancient Greek, Chinese, and Croatian, and English before the Seventeenth Century) have no word for 'conscious' or 'consciousness'. Presumably consciousness itself remains unchanged; people from these linguistic communities... cannot talk about [conscious] experiences since they don't regard the mind the same way we do." [pp. 23-24]
11. There is a lot wrong with this argument. The first premise is a good place to start. Classical Chinese, for instance, does have terms for "consciousness" and for "conscious." In fact quite a few, and it needs them, since the understanding of consciousness is a central preoccupation for many classical Chinese Buddhist philosophical schools. While there may be no direct synonyms in classical Greek, the distinction between conscious and unconscious states was important to the stoics and sceptics. I cannot speak for the Croatians.
12. The second premise is also problematic. Given Hardcastle's own position regarding the penetration of introspective consciousness by theory it must be at least an open question whether "consciousness itself remains unchanged" across the kind of conceptual variation she suggests. My own guess would be that it does not, on very much the kinds of grounds Hardcastle adduces in the first chapters of the book for the plasticity of introspection. Third, the inference from the absence of a particular lexical item in a language to the absence of a corresponding concept is dangerous. Very similar concepts may be present and expressed with circumlocution in one language, and a single word in another.
13. Hardcastle also argues in this chapter that no scientific theory of qualitative experience has any obligation to tell us about the nature of any particular experience of phenomenon of consciousness, just as "Theoretical physicists only want to explain the general pattern of behaviour of moving objects; hence they can rightly ignore the colour of the body, the date of the falling... etc." [p. 24] Here she sets the descriptive/explanatory burden for cognitive science a bit too light, even according to her own bill of lading. The colour of the falling body is irrelevant to the physicist's task at hand, which is a trajectory explanation. But if the task of a naturalistic theory of consciousness is to explain the character of consciousness, then what yellow looks like is hardly irrelevant to that enterprise, even though the yellowness of a falling ball may be irrelevant to explaining the velocity of the ball.
14. A better analogy is this: A theory of consciousness may not have to say what I am experiencing now. But just as a physics of colour needs to specify what wavelengths of light are yellow, a theory of colour experience needs to say what it is to see yellow even if it doesn't need to say when I experience it. There is no reason for cognitive science as Hardcastle characterises and envisions it to shirk this burden. And if Hardcastle is right, there is no reason to think that this descriptive/explanatory burden cannot in principle be born.
15. In the fourth chapter Hardcastle articulates and defends the multiple memory system framework. Here she distinguishes a memory system dedicated merely to stimulus-specific habits and responses; a more complex but automatic structural memory recruited in perception and automatic recognition, and controlled, highly articulated conscious semantic memory involved in explicit recall, episodic and factual memory. She makes deft use of a wide variety of psychological and neurological evidence, weaving a large number of different experimental paradigms to create a compelling argument from convergence of evidence.
16. At the end of the account, however, one is left with some questions concerning the clarity of what are presented as sharp principled distinctions. In particular, the distinction between habitual and structural memory is not adequately defended. Why are motor habits so different from perceptual habits? Isn't is plausible that many motor skills of the kind represented in habitual memory are utilised by experts in a wide range of contexts, and that they are semantically sensitive as well in expert performance? (See for instance Jordan and Rosenbaum, 1993) These questions are not addressed by Hardcastle.
17. Hardcastle argues in the fifth chapter that consciousness is associated with the activation of semantic memory. This is the central section of the book. Hardcastle again is extremely effective and persuasive in her integration of neuroscience and computational cognitive psychology in the service of a philosophical argument. Difficulties, however, remain for the attempt to align consciousness so closely with semantic memory as Hardcastle characterises it. It would appear that a lot of cognition draws on semantic memory proceeds unconsciously. The psycholinguistic processes necessary to assign semantic representations to sentences; selection of chess moves for conscious consideration; driving through a city while carrying on a conversation to which one is attending provide plausible examples. It might be that these prima facie semantic memory-dependent tasks are in fact handled by structural memory. But that would require a good deal of evidence that is absent in the book. Otherwise we would need a distinction between conscious and unconscious use of this system, and we would be back to square one with respect to providing an account of the nature of consciousness in general. Nor is it at all obvious that none of the tasks ascribed to structural memory are unconscious. Consider recognising faces in a lineup, or identifying a piece of music.
18. The sixth chapter of the volume is devoted to what has come to be known as "the hard problem"--the problem of explaining "why it should be that THAT is conscious." [p. 104]. This is a central chapter in the book, as it is here that Hardcastle really accomplishes her principal aim--to show that and how there could be a theory of cognition that incorporates neuroscientific and computational levels of description yielding insight into cognitive mechanisms that could lead us to say, "Yes, understanding how that works allows us to see how we could be conscious, and what it is to be conscious." I do not pretend to evaluate the details of the story. In fact, I would be astonished if anything like the story Hardcastle tells were to turn out to be the final version of a theory of consciousness. There is as yet just too much that is unknown in the relevant domains. What is important though is the fact that she provides a plausible story which, were it true, would yield the requisite understanding, and which hence demonstrates, insofar as this is demonstrable, that we could gain insight into consciousness from the right kind of cognitive science.
19. Particularly noteworthy in this chapter is the fine critique of Dennett's replacement of the "Cartesian theatre" with what Hardcastle calls a "travelling side show" [p. 105] (and what called in another place "the Cartesian American Multi Cinema [Garfield, 1992]): her attack on Dennett's odd claim that we do not re-present in cognition [p. 106] and her attack on his equally odd claim that any mental process can become conscious depending on the timing and nature of introspective probes [p. 105]. This critique is sharp, well-argued, and demonstrates the strength of a suitably realistic materialist approach to consciousness research as a reply to Dennett's instrumentalism. There is, however, one difficulty with this chapter as it stands: The various pieces of the empirical story Hardcastle tells, while each told in great detail and with considerable precision, are never as clearly assembled by the author into a unified whole. That task is left to the reader as an exercise, and the empirically unsophisticated philosophical reader will find it a challenging exercise indeed.
20. The seventh chapter could have been omitted without loss. Here Hardcastle recapitulates the absent qualia debate, defending the only plausible position in that debate (defended long ago by Shoemaker (1981)--that beings that sincerely believe themselves to have conscious experience and that behave like and are organised internally like beings that have conscious experience in fact have conscious experience. This is by now well-trodden terrain, and the chapter could have been replaced with a footnote.
21. Hardcastle argues that her account of semantic memory as the mechanism for the implementation of consciousness is radically distinct from executive processing models. I worry that it is not as different as it needs to be. For, as Hardcastle concedes, the difficulty with executive system models is not that it is called "executive," but that it supposes a central location, or as Dennett (1991) derisively calls it, a "Cartesian theatre" in which a thought must appear in order to be conscious. Now, Hardcastle (and I) agree with Dennett regarding the plethora of difficulties attending to such a view. But notice that Hardcastle's multiple memory model, with its association of consciousness with semantic memory recapitulates the crucial feature of these models: There IS, in her view, a central location where thoughts appear in order to be conscious: semantic memory. If the account of consciousness at the end boils down to this (granting the marvellous detail in which Hardcastle spells this out in computational and neurological terms) Hardcastle has not really abandoned the executive model she castigates, but has rather substituted one theory of the executive for another with all of the difficulties she adduces for that model left in place. If, on the other hand, semantic memory isn't the mediator of consciousness, we don't yet have an account at all, despite Hardcastle's having put a lot at our disposal that might constitute the raw material for such an account.
22. For this reason I have reservations about Hardcastle's otherwise persuasive reply to Dennett regarding identifying the moment of consciousness. Dennett, of course, argues that this makes no sense, because Dennett argues that there is no central location where consciousness emerges. If Hardcastle is right about the role of semantic memory, then she is right when she argues in the final chapter of the book that there is no principled difficulty in determining the moment when a thought becomes conscious. The experiment may be difficult to design, and delicate to perform, but there is an empirical question to be posed and there are clear kinds of evidence to which one could point in answering it. But if Dennett is right, and if my worries about the identification of semantic memory with conscious experience are warranted, the antecedent of this conditional is false.
23. There are a few minor features of this book about which one could complain: There are too many typographical errors, too many unnecessary and non-mnemonic acronyms, and a few moments at which Hardcastle faced with a choice between humour and clarity opts for the former. But these are minor blemishes on an otherwise excellent book. If Hardcastle's principal aim is to show how explanations at the level of neuroscience, computational psychology and philosophy of mind can be brought together in a single account, she has succeeded admirably, even if that account remains in need of further development.
Churchland, P. (1979). Scientific Realism and the Plasticity of Mind. Cambridge: Cambridge University Press.
Dennett, D. (1991). Consciousness Explained. Boston: Brown and Little.
Garfield, J. (1989). "The Myth of Jones and the Mirror of Nature: Reflections on Introspection." Philosophy and Phenomenological Research, Vol 50, no 1, pp 1-32.
Garfield, J. (1992). "Reply to Dennett." "Has Consciousness Been Explained?", Amherst College.
Hardcastle, V.G. (1995). Locating Consciousness. Amsterdam & Philadelphia: John Benjamins Press.
Hardcastle, V.G. (1996). Precis of: Locating Consciousness. PSYCOLOQUY 7(33) locating.consciousness.1.hardcastle.
Jordan, M. and D. Rosenbaum (1993) "Action," in M. Posner, ed., Foundations of Cognitive Science. Cambridge: MIT Press.
Rorty, R. (1979). Philosophy and the Mirror of Nature. Princeton: Princeton University Press.
Shoemaker, S. (1981). "Absent Qualia are Impossible--Reply to Block," Philosophical Review 90, pp 581-599.