Message to Stan Klein From A.Sloman@cs.bham.ac.uk Sun May 25 10:06 BST 1997 Date: Sun, 25 May 1997 10:06:29 +0100 Subject: Re: subjective awareness and machinery thanks Stan, > ...He also questions the > distinction between access and phenomenal consciousness. I have now (yesterday) read Block's BBS article and a lot of the commentary articles, and found the usual mixture of confusion and cross-purposes that always pervades discussions of consciousness. There are several problems with Block's distinction, including one which the objectors did point out and one which nobody seemed to notice: (a) the extraordinary vagueness of his specification of P-consciousness, which amounts to little more than a definition by ostension, and (b) the unconving presuppositions behind his specification of A-consciousness, namely that the uses of A consciousness have something to do with rationality, planning, etc. Of course, there may be ways of introducing an architecture-based distinction which are immune from these criticisms. But if you take the architecture seriously there will be more than just two varieties of consciousness. I got the impression that, like most philosophers who had never actually taken part in any engineering activity, Block has no idea how to design a working system that could function like a mind. Would you put him in charge of a team of robot designers? It's a very common assumption of philosophers that somehow rationality has to be involved in high level specifications of important mental states and processes, which I think is just false. There's a pre-rational virtual machine level of information processing (i.e. information-based control, as opposed to purely physical control) which provides the engine in which cognitive processes run, and which is in itself neither rational nor irrational, but which can sometimes be used for rational purposes and sometimes generates irrational behaviour. Most of the time works in the way that evolution (plus individual adaptation) designed it to work which is neither rational nor irrational. (We can interpret evolution as an implicit rational designer of the whole system. But that's another long story. Likewise aspects of a robot's mind and brain that it would be rational for a designer to incorporate need not be rational from the point of view of the robot.) My guess is that if we had a way of counting conscious human mental events and processes, we'd find that the vast majority could not be described as rational or irrational: they just happen. That's also what's wrong with Dennett's intentional stance: it presupposes rationality as a basis for attributing intentionality (i.e. the ability to make use of semantic content), and therefore fails to account for the vast majority of types of intentionality. (Curiously Alan Newell, who really should have known better, made a similar mistake in defining "the knowledge level" and suggesting that there's only one level below it, the "physical symbol level"). In criticising Block's distinction I don't claim that there are no partly analogous distinctions between varieties of types of consciousness. E.g. there's obviously a distinction between passively experiencing some state and actively making use of that experience in decision making, for example. I'll try later on to get the article you referred to. From your summary it sounds as if it is consistent with the sort of architecture I've been working on. I explain qualia in terms of intermediate levels of perceptual processing together with the ability of a meta-management mechanism to "attend to" those levels (i.e. select their contents for processing), as a painter has to do when focusing on how things are experienced as opposed to how they are. Drawing a table and using a table to put things on generate quite different perceptual requirements. > I don't think that developing an improved conceptual framework > will change people's minds. I agree. One of the reasons for this has to do with how compelling the various delusions gained by introspection are. (eg the belief that we really do know what we mean...) Another is that many of the people who discuss consciousness nowadays are not involved in a search for the truth but a search for some kind of metaphysical security. For both groups, what is needed is philosophical therapy, not argument. Moreover, the two groups need different sorts of therapy! And even long term therapy won't work for all of them. I've already managed to change the mind of one person through close face to face discussion, but I don't expect to be able to convince many (especially as all this is a sideline for me and I don't have time to do the job properly!) I gave a talk to Sussex university philosophers a couple of weeks ago. In the discussion afterwards I got the impression that ONE of them was beginning to shift his position. But usually it takes much, much longer. At Elsinore in August I'll no doubt be faced by a collection of workshop attendees who will prove you completely right! Cheers. Aaron PS The latest New Scientist (24th May) inside back cover, page 97, has a lovely example. Count the number of occurrences of "F" in this sentence: FINISHED FILES ARE THE RESULT OF YEARS OF SCIENTIFIC STUDY COMBINED WITH THE EXPERIENCE OF YEARS. It seems that most people (including my wife and me) see only three Fs at first, and have to work quite hard to see all of them (six Fs). Question: before they have found them all are they having six F qualia or only three?