X-Message-Number: 11749 From: Date: Fri, 14 May 1999 13:09:47 EDT Subject: Once more, with feeling--zombies etc. Reiterating his case for consciousness in computers, Daniel Crevier writes (in part): >We said that we could replace a piece of the >brain by a simulation with the same input-output >properties. It's pretty hard to argue against that, since what happens in >any part of the brain are physical processes, and a computer can simulate >any physical process. Now if we can do it for a piece of the brain, we can >do it for the whole brain: just do it for all the pieces, connect the >resulting simulations together, and voilą. >The resulting simulated brain would therefore have the same input-output >properties as the real brain. For a brain, the input is what the senses >tell it, and the output is motor signals determining the person's words >and deeds. The simulated person would thus react in exactly the same way >as the real person. In particular, if you ask her whether she is >conscious, she will answer yes. We have been over this countless times, going back years, but let's try yet again, since apparently I still fail to convey my points clearly. Let's separate it into two parts. 1. Could there be "zombies"--systems that, seen from the outside, behave just like people would, and in particular converse just like people would, but nevertheless have no consciousness and are not people? If there could be zombies, would any "paradoxes" or philosophical problems arise? Certainly there could be zombies. This is just the Turing Test revisited. A sufficiently sophisticated scanner-cum-computer could analyze a person and predict his actions, and then program a robot to follow that script. To a very limited extent, THIS HAS ALREADY BEEN DONE. Conversational programs already exist, and fool some of the people some of the time, yet no one claims that these programs (or the computers in which they run) are conscious--and making them more elaborate and more accurate will change nothing in principle. No unique philosophical problems arise, only practical ones. Certainly you might worry, and could err, if you suspect that someone is a zombie--so what? Either you find some way to decide (see below), or you live with the uncertainty. Mr. Crevier's position reminds us of the Behaviorist school in psychology--remember B.F. Skinner? They thought we had to regard people as black boxes, with only inputs and outputs as the allowable variables, internal structure off limits. This again reminds us of the Positivist school of philosophy of science, which (roughly) held that, if you can't get an answer to a question by performing an experiment, then the question is meaningless. This school also is in decline; remember that, at a particular time in a particular context, what you "can" or cannot do in principle is generally not at all clear. (Right now, for example, some people take the view that the mathematical postulates of quantum mechanics represent the last word, and it is meaningless to ask whether those postulates can be explained by, or might result from, deeper underlying realities, such as hidden variables. And I might add that the "many worlds" or "multiverse" interpretation is really just a form of hidden variables, or a metaphor for one possible system of hidden variables.) 2. What is proven by thought experiments involving gradual replacement of brain parts by computer simulations? First, I repeat that thought experiments are useful only if the premises are sound. Mr. Crevier assumes, as a premise, that if brain parts are gradually removed, replaced by electromagnetic and chemical inputs/outputs generated by a computer, the subject will notice nothing at any stage. But this is not obvious at all. If indeed--as I surmise--consciousness requires time binding and space binding activity in the brain, then the computer could not supply that, and subjectivity would be lacking. If the person's "self circuit" were removed or shut off, then the person would lose consciousness. The key question then becomes, would or could the subject nevertheless REPORT that he was still conscious, and if so are we obliged to consider only that report and nothing else? If the computer can simulate the details of the self circuit, and if it is so fast that it doesn't need to be parallel or time-binding or space-binding, but can produce the same input/output signals as the self circuit, then we would indeed have a zombie, with consequences discussed above. The subject would lose consciousness, but would report that he was still conscious. But we are NOT reduced to considering the report and nothing else. If a decoy looks like a duck and walks like a duck and quacks like a duck (these actually exist), we can still demand to feel its feathers, x-ray its innards, etc., before admitting that maybe it really is a duck. We cannot yet do the equivalent for consciousness in a person, because we do not yet know the anatomy/physiology of consciousness. But when we do learn the innards of consciousness in mammals, we will be justified in skepticism about the consciousness of an artifact that lacks those features. Another reminder: If a computer can be conscious, so can a book. If isomorphism is everything--and this is indeed the real, basic premise of Mr. Crevier's school of thought--then the principle of isomorphism should apply to time as well as space and matter. If nothing is essential except eventual generation of the right sets of numbers, then the pages of a book can correspond to the successive states of a Turing computer, and therefore to a living person. I think this is a valid reductio-ad-absurdum. Finally, I said there are no "unique" philosophical problems arising from the "zombie" question. But there are still the usual problems of duplication and continuity, which also exist for example with a beam-me-up machine; these are still open questions, and anyone who thinks he knows the answers is kidding himself. And finally-finally, in the face of the uncertainties, the common-sense course of action is to try to have ourselves cryopreserved if we die, with minimal change, and not be side-tracked by the much more remote possibilities of uploading. Robert Ettinger Cryonics Institute Immortalist Society http://www.cryonics.org Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=11749