Dennett is the great demystifier of consciousness. According to him there is, in the final analysis, nothing fundamentally inexplicable about the way we attribute intentions and conscious feelings to people. We often attribute feelings or intentions metaphorically to non-human things, after all. We might say our car is a bit tired today, or that our pot plant is thirsty. At the end of the day, our attitude to other human beings is just a version – a much more sophisticated version – of the same strategy. Attributing intentions to human animals makes it much easier to work out what their behaviour is likely to be. It pays us, in short, to adopt the intentional stance when trying to understand human beings. This isn’t the only example of such a stance, of course. A slightly simpler example is the special ‘design stance’ we adopt towards machines when we try to understand how they work (that is, by assuming that they do something useful which can be guessed from their design and construction). An axe is just a lump of wood and iron, but we naturally ask ourselves what it could be for, and the answer (chopping) is evident. A third stance is the basic physical one we adopt when we try to predict how something will behave just by regarding it as a physical object and applying the laws of physics to it. It’s instructive to notice that when we adopt the design stance towards an axe, we don’t assume that the axe is magically imbued with spiritual axehood: but at the same time its axehood is uncontroversially a fact. If we only understood things this way all the time, we should find the real nature of people and thoughts no more worrying than the real nature of axes. One day there could well be machines which fully justify our adopting the intentional stance towards them and hence treating them like human beings. With some machines, some of the time, and up to a point, we do this already, (think of computer chess) but Dennett would not predict the arrival of a robot with full human-style consciousness for a while yet. So it’s all a matter of explanatory stances. But doesn’t that mean that people are not ‘real’, just imaginary constructions? Well, are centres of gravity real? We know that forces really act on every part of a given body, but it makes it much easier, and no less accurate, if our calculations focus on a single average point. People are a bit like that. There are a whole range of separate processes going on in the relevant areas of your brain at any one time – producing a lot of competing ‘multiple drafts’ of what you might think, or say. Your actual thoughts or speech emerge from this competition between rival versions – a kind of survival of the fittest, if you like. The intentional stance helps us work out what the overall result will be.

The ‘overall result’? But it’s not as if the different versions get averaged out, is it? I thought with the multiple drafts idea one draft always won at the expense of all the others. That’s one of the weaknesses of the idea – if one ‘agent’ can do the drafting on its own, why would you have several?

It’s just more effective to have several competing drafts on the go, and then pick the best. It’s a selective process, comparable in some respects to evolution – or a form of parallel processing, if you like.

‘Pick the best’? I don’t see how it can be the best in the sense of being the most cogent or useful thought or utterance – it’s just the one that grabs control. The only way you could guarantee it was the best would be to have some function judging the candidates. But that would be the kind of central control which the theory of multiple drafts is supposed to do away with. Moreover, if there is a way of judging good results, there surely ought to be a way of generating only good ones to begin with – hence again no need for the wasteful multiple process. I’m always suspicious when somebody invokes ‘parallel processing’ . At the end of the day, I think you’re forced to assume some kind of unified controlling process.

Absolutely not- and this is a key point of Dennett’s theory. None of this means there’s a fixed point in the brain where the drafts are adjudicated and the thinking gets done. One of the most seductive delusions about consciousness is that somewhere there is a place where a picture of the world is displayed for a ‘control centre’ to deal with – the myth of the ‘Cartesian Theatre’. There is no such privileged place; no magic homunculus who turns inputs into outputs. I realise that thinking in terms of a control centre is a habit it’s hard to break, but it’s an error you have to put aside if we’re ever going to get anywhere with consciousness. Another pervasive error, while we’re on the subject, is the doctrine of ‘qualia’ – the private, incommunicable redness of red or indescribable taste of a particular wine. Qualia are meant to be the part of an experience which is left over if you subtract all the objective bits. When you look at something blue, for example, you acquire the information that it is blue: but you also, say the qualophiles, see blue. That blue you really see is an example of qualia, and who knows, they ask, whether the blue qualia you personally experience are the same as those which impinge on someone else? Now qualia cannot have any causal effects (otherwise we should be able to find objective ways of signalling to each other which quale we meant). This has the absurd consequence that any words written or spoken about them were not, in fact, caused by the qualia themselves. There has been a long and wearisome series of philosophical papers about inverted spectra, zombies, hypothetical twin worlds and the like which purport to prove the existence of qualia. For many people, this first person, subjective, qualia-ridden experience is what consciousness is all about; the mysterious reason why computers can never deserve to be regarded as conscious. But, Dennett says, let’s be clear: there are no such things as qualia. There’s nothing in the process of perception which is ultimately mysterious or outside the normal causal system. When I stand in front of a display of apples, every last little scintilla of subtle redness is capable of influencing my choice of which one to pick up.

It’s easy to deny qualia if you want to. In effect you just refuse to talk about them. But it’s a bit sad. Qualia are the really interesting, essential part of consciousness: the bit that really matters. Dennett says we’ll be alright if we stick to the third-person point of view (talking about how other people’s minds work, rather than talking about our own); but it’s our own, first-person sensations and experiences that hold the real mystery, and it’s a shame that Dennett should deny himself the challenge of working on them.

I grant you qualia are grist to the mill of academic philosophers – but that’s never been any sign that an issue was actually real, valid, or even interesting. But in any case, Dennett hasn’t excluded himself from anything. He proposes that instead of mystifying ourselves with phenomenology we adopt a third-person version – heterophenomenology. In other words, instead of trying to talk about our ineffable inner experiences, we should talk about what people report as being their ineffable inner experiences. When you think about it, this is really all we can do in any case. That’s Dennett in a nutshell. Actually, it isn’t possible to summarise him that compactly: one of his great virtues is his wide range. He covers more aspects of these problems than most and manages to say interesting things about all of them. Take the frame problem – the difficulty computer programs have in dealing with teeming reality and the ‘combinatorial explosion’ which results. This is a strong argument against Dennett’s computation-friendly views: yet the best philosophical exposition of the problem is actually by Dennett himself.

Mm. If you ask me, he’s a bit too eager to cover lots of different ideas. In ‘Consciousness Explained’ he can’t resist bringing in memes as well as the intentional stance, though it’s far from clear to me that the two are compatible. Surely one theory at a time is enough, isn’t it? Even Putnam disavows his old theory when he adopts a new one.

It seems to me that a complete account of consciousness is going to need more than one theoretical insight. Denett’s broad range means he’s said useful things on a broader range of topics than anyone else. Even if you don’t agree with him, you must admit that that sceptical view about qualia, for example, desperately needed articulating. And it typifies the other thing I like about Dennett. He’s readable, clear, and original, but above all he really seems as if he wants to know the truth, whereas most of the philosophers seem to enjoy elaborating the discussion far more than they enjoy resolving it. His theory may seem strange at first, but after a while I think it starts to seem like common sense. Take the analogy with centres of gravity. People must be something like this in the final analysis, mustn’t they? On the one hand we’re told the self is a mysterious spiritual entity which will always be beyond our understanding: on the other side, some people tell us paradoxically that the self is an illusion. I don’t think either of these positions is easy to believe: by contrast, the idea of the self as a centre of narrative gravity just seems so sensible, once you’ve got used to it.

The problem is, it’s blindingly obvious that whether something is conscious or not doesn’t depend on our stance towards it. Dennett realises, of course, that we can’t make a bookshelf conscious just by giving it a funny look, but the required theory of what makes something a suitable target for the stance (which is really the whole point) never gets satisfactorily resolved in my view, in spite of some talk about ‘optimality’. And that business about centres of gravity. A centre of gravity acts as a kind of average for forces which actually act on millions of different points. Well there really are people like that – legal ‘persons’, the contractual entitities who provide a vehicle for the corporate will of partnerships, companies, groups of hundreds of shareholders and the like. But surely it’s obvious that these legal fictions, which we can create or dispell arbitrarily whenever we like, are entirely different to the real people who invented them, and on whom, of course, they absolutely depend. The fact is, Dennett’s view remains covertly dependent on the very same intuitive understanding of consciousness it’s meant to have superseded. You can imagine a disciple running into problems like this…

Disciple: Dan, I’ve absorbed and internalised your theory and at last I really understand and believe it fully. But recently I’ve been having a difficulty.

Dennett: What’s that?

Disciple: Well, I can’t seem to adopt the intentional stance any more.

Dennett: Wow. It’s really very simple. Deep breaths now. Look at the target (use me if you like). Now just attribute to me some plausible conscious states and intentions.

Disciple: But… What would that be like? What are conscious states? For you to have conscious states just means I can usefully deal with you as if you had … conscious states. I seem to be caught in a kind of vicious circle unless I just somehow know what conscious states are…

Dennett: Steady now. Just think, what would I be likely to do if I had the kind of real, original intentions which people talk about? How would things with intentions behave?

Disciple: I have no idea. There are no things with real intentions. I’m not even sure any more what ‘real intentions’ means…

Yes, very amusing I’m sure. I suppose I can sympathise with you to some extent. Grasping Dennett’s ideas involves giving up a lot of cherished and ingrained notions, and I’m afraid you’re just not ready (or perhaps able) to make the effort. But the suggestion that Dennett doesn’t tell us what makes something a good target for the intentional stance is a shocking misrepresentation. It could hardly be more explicit. Anything which implements a ‘Joycean machine’ is conscious. This Joycean machine is the thing, the program if you like, which produces the multiple drafts. The idea is that consciousness arises when we turn on ourselves the mechanisms and processes we use to recognise and understand other people. Crudely put, consciousness is a process of talking to ourselves about ourselves: and it’s that that makes us susceptible to explanation through the intentional stance. It’s all perfectly clear. You obviously haven’t grasped the point about optimality, either. Suppose you’re playing chess. How do you guess what the other player is likely to do? The only safe thing to do is to assume he will make the best possible move, the optimal move. In effect, you attribute to him the desire to win and the intention of out-playing you, and that helps dramatically in the task of deciding which pieces he is likely to move. Intentional systems, entities which display this kind of complex optimality, deserve to be regarded as conscious to that extent.

Yes, yes, I understand. But how do you know what behaviour is optimal? Things can’t just be inherently optimal: they’re only optimal in the light of a given desire or plan. In the case of a game of chess, we take it for granted that someone just wants to win (though it ain’t necessarily so): but in real-life contexts it’s much more difficult. Attributing desires and beliefs to people arbitrarily won’t help us predict their behaviour. Our ability to get the right ones depends on an in-built understanding of consciousness which Dennett does not explain. In fact it springs from empathy: we imagine the beliefs and desires we would have in their place. If we hadn’t got real beliefs and desires ourselves, the whole stance business wouldn’t work.

It isn’t empathy we rely on – at least, not what you mean by empathy. The process of evolution has fitted out human beings with similar basic sets of desires (primarily, to survive and reproduce) which can be taken for granted and used as the basis for deductions about behaviour. I don’t by any means suggest the process is simple or foolproof (predicting human behaviour is often virtually impossible) just that treating people as having conscious desires and beliefs is a good predictive strategy. As a matter of fact, even attributing incorrect desires and beliefs would help us falsify some hypotheses more efficiently than trying to predict behaviour from brute physical calculation. Speaking of evolution, it ocurs to me that a wider perspective might help you see the point. Dennett’s views can be seen as carrying on a long-term project of which the theory of evolution formed an important part. This is the gradual elimination of teleology from science. In primitive science, almost everything was explained by attributing consciousness or purpose to things: the sun rose because it wanted to, plants grew in order to provide shade and food, and so on. Gradually these explanations have been replaced by better, more mechanical ones. Evolution was a huge step forward in this process, since it meant we could explain how animals had developed without the need to assume that conscious design was part of the process. Dennett’s work takes that kind of thinking into the mind itself.

Yes, but absurdly! It was fine to eliminate conscious purposes from places where they had no business, but to eliminate them from the one place where they certainly do exist, the mind, is perverse. It’s as though someone were to say, well, you know, we used to believe the planets moved because they were gods; then we came to realise they weren’t themselves conscious beings, but we still believed they were moved by angels. After a while, we learnt how to do without the angels: now it’s time to take the final step and admit that, actually, the planets don’t move. That would be no more absurd that Dennett’s view that, as he put it, ‘we are all zombies’.

A palpably false analogy: and as for the remark about zombies, it is an act of desperate intellectual dishonesty to quote that assertion out of context!