X-Message-Number: 8682 Date: Wed, 15 Oct 97 18:56:05 From: Mike Perry <> Subject: Turing's thesis, computers Thomas Donaldson, #8677, wrote: >As for Turing's thesis as distinct from Turing machines, exactly what >was his thesis? Quoting from the book I'm writing: "It was a conjecture of Turing, now known as Tur- ing's Thesis, that any effective procedure of the sym- bol-manipulation variety could be performed by a Turing machine--and thus by a UTM [universal Turing machine] or a computer. (Sometimes this is known as the Church or Church- Turing thesis or hypothesis, in honor of American logician Alonzo Church, who independently ad- vanced an equivalent formalism and conjecture, slightly earlier than Turing. Turing's approach is more accessible to the non-specialist and is more widely cited.) This would include any task of the computational variety that a human might do, such as numerical calculations, devising moves to a game such as chess (since these moves are describable in strings of symbols), and in short, any well-defined operation that produces a description of something from another description." [end of quote] "Symbol manipulation" though has a very broad applicability, if you want to go far enough with it, like describing what is happening in any finite chunk of reality at the quantum level. To me it is important to be able to argue in the book that every finite process whatever could be described in all its significant details (i.e. at the quantum level) by a Turing computation. Here I am not trying to claim this would lead to any sort of practical application. But the claim does have significance philosophically speaking--quite a bit I think. For instance it sheds light on the problem of resurrecting the dead, if we also consider a few other principles. It is interesting that Tipler, in *The Physics of Immortality* argues, for certain theological reasons connected with "free will," that reality in the large is (probably) not Turing-computable. And he may well be right, which would seem to mean that life as a whole might be more interesting through being less "computable," i.e. predictable. (I don't accept his "theological" conclusions however.) If he is right, then a slight generalization--the Turing machine with infinitely inscribed tape, seems to fit the bill, and still allows the "digital" argument I need to make in my own book. > It seems to me, given that computing has already seen >lots of developments which make Turing machines at best only abstract >models, that if we want to refute or argue against any spiritualistic >approach to explaining how we work, then tying our argument to Turing >machines is unwise. On the contrary, a Turing machine is a simple device that is easy to describe and very convenient for "in priniciple" type arguments, so long as you don't insist on using this model inappropriately for any claims about making a practical device (and are reasonably careful in other ways). [snip] >I'll also say, for myself, that it would help understanding both on >cryonet and in general if somebody came up with a definition of just >what was a computer. As I see it, a machine of the symbol-manipulation variety. It's task is to transform strings of symbols, which we could regard as descriptions of things, into other strings of symbols. At least this seems to be the main thing "computers" do. The process of "writing" symbols in certain locations can have physical consequences too, e.g. to activate motor responses in a device. But here you are, ostensibly, going beyond "computing." On the other hand, "symbols" can be broadly interpreted. You may have to go clear down to the quantum level to argue that brains operate with "symbols"--but your reward, philosophically speaking, may be worth it. Mike Perry Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=8682