Boxing with qualia

10 April 2005

The sub-title of Jeffrey Gray's book 'Consciousness - creeping up on the hard problem'  signals an attractive modesty about his achievement. Elsewhere he describes himself as ducking and weaving round the problem of qualia like a boxer, putting in a jab here and there and dancing away again, picking up points where he can without looking for a knockout. His approach is distinctive, using well-known elements in slightly unexpected ways, so that at times it's almost as though he's using judo moves in his boxing ring. I'm not sure his approach always works, but I think his strategy entitles him to a large degree of latitude over some of the issues, especially the philosophical ones, since he declares that though his  book recognises the need to deal with all the issues, it is not a philosophical treatise.

One case where his philosophical stance seems dubious to me appears very near the beginning. We do not, he declares trenchantly, see the real world; only an illusion constructed in our heads. Far from having direct experience of an external world, we only perceive a mental model, full of things like colour and music which don't really exist in the world (other than as complex and rather arbitrary properties based on the differential reflectance of surfaces or vibrations in the air). Gray recognises the danger of this argument - if we never perceive the external world, what reason can we have to believe in it at all? He carefully affirms on several occasions that there really is an objective external world which has something to do with our perceptions. But I don't think he really realises the gravity of the problem. When I think about a cow, I think about a real cow, not a model cow in my head. One of the properties of thought is that I just have the power to do this. If Gray agrees with this, I think he ultimately has to abandon his view; if he disagrees, then the status of his assertions about the reality of the external world becomes problematic - can he even talk about the real extrnal world?

A second pillar of Gray's view also seems debatable, if not to quite the same extent. This is the idea that all conscious experience suffers from a kind of time-lag. Gray takes Libet's conclusions (that our decisions are made before we even become aware of them) pretty much at face value, and he also quotes the case of tennis, where the stroke, it seems, often has to be played before there has been sufficient time for conscious perception of the ball. The idea of a systematic delay is undoubtedly tenable, but personally I suspect that the appearance of a delay arises merely from a failure to distinguish between decisions consciously made, and decisions which have had time to become the objects of conscious examination themselves.

Anyway, these two dubious propositions go to support a theory which in itself is not at all unreasonable. Gray has unconscious processes putting together a model for the examination of the conscious faculty while they (the unconscious systems) get on with actually playing the tennis, driving the car, or whatever. These unconscious systems seem so proficient at looking after all the detailed business of life that one wonders why they bother keeping consciousness in the picture at all - a question Gray goes on to address. There must, he says, be some Darwinian justification for consciousness; and in fact, he believes it serves the purpose of 'late error detection'. Normally, our behaviour is under unconscious control, but the model is used to extrapolate future situations: where unexpected elements appear or a process takes an unexpected route, the matter is referred to consciousness where there is the opportunity for a second look at the circumstances.

The idea that something along these lines does happen is quite persuasive. Just as our eyes automatically jump to look at anything surprising, our attention focuses on the unexpected and the irregular. Gray also makes the claim that, rather paradoxically, consciousness also tends to identify features of the environment which remain constant: the unconscious processes operate on a moment-to-moment basis, but things tend to hang around in consciousness for an appreciable period. As a result, consciousness permits the identification of those common referents in the real world which we need in order to communicate with each other.

Any system which has a mental faculty perceiving a model is vulnerable to a charge of homuncularism or infinite regress: if perceiving the outside world requires a model which is then perceived by consciousness, doesn't the perception by consciousness require a second model perceived by an inner consciousness, and so on? Gray, justifiably, I think, rebuts this accusation as it applies to his own theory. The kinds of perception he is talking about are sufficiently different, and the helpful role of the model is clear enough to avoid any further regress.

I think the general scheme is plausible, but whether it accounts for more than one limited, attentional aspect of consciousness seems more dubious. On Gray's model, consciousness is in thrall to unconsciousness, constantly having its attention directed to potential problems. In reality, most of the time consciousness directs attention wherever it wishes, and often goes on operating in complete detachment from current unconscious perceptions, whether problematic or routine.

Worse, while such error-handling processes may well occur, the theory does not explain why they are conscious, let alone why or even how they involve qualia. What has phenomenal experience got to do with it? Much of Gray's discussion of qualia after this has the air of shadow-boxing: the real qualia have escaped somehow. He argues, for example, that qualia can have real causal effects; but on his account it seems hard to understand why anyone should ever have supposed there was any difficulty about it: instead of the ineffable phenomenal qualia of other discussions, we seem to be talking about fairly straighforward pieces of data.

It must not be thought, however, that Gray sees it this way: in fact he presents a case against any functionalist analysis of qualia. One reason to reject functionalism is that Searle's anti-computationalist arguments strike Gray as convincing; but his main argument is a novel one. According to functionalism, he says, qualia are identical with appropriate brain functions. It follows that qualia and functions are in a one-to-one relationship: one function, one quale and vice versa. But synaesthetes get the same qualia from two quite different mental functions - from seeing a colour and from hearing a number, for example. This is no voluntary process, but one which the synaesthetes are powerless to dispense with even when it becomes a nuisance.

I think the functionalists have at least two escape routes from this argument. I think moderate functionalists need not actually identify qualia with functions: they might simply believe that qualia arise from functions (or indeed, they might just deny the existence of qualia altogether). Alternatively, it could be that the qualia are functions, and that seeing colour and reading a number (for the synaesthete) are two distinct functions which nevertheless both trigger the same qualic function. I don't see why that should be impossible.

Gray himself ultimately takes the respectable but limited view that qualia must indeed be reducible to some (non-functional) properties of brains: we just don't know at the moment which or how. He has some kind words to say about what he calls the Penroff (Penrose-Hameroff) theory, though stopping well short of endorsing it.