X-Message-Number: 21678 Date: Sun, 27 Apr 2003 00:55:01 -0700 From: Mike Perry <> Subject: Computations and Symbolics A reply to Thomas Donaldson, #21663 >First, you might do a bit of reading about quantum computing and >qubits. None of the points you bring up actually discuss Turing >machines as compared to quantum computers. A Turing machine is a >machine which works on an infinitely long tape divided into small >lengths, for brevity containing 1's or 0's. It can move forwards >or backwards on that tape, and mark it with its own 1's and 0's >to wipe out those on the tape. (I hardly believe you did not >know this!). So just how does this relate to a quantum computer? The way it relates is--I think--through isomorphism. My understanding of quantum computing is far from comprehensive and I do hope to do more reading. However, from what I've read so far it seems that there are certain rather strong claims being made (such as that the halting problem can be solved through quantum computing) versus a much more modest success with actual quantum computers to date. This suggests that we should approach such claims with caution at least, yet I didn't want to ignore quantum computing when I was doing my book; it does seemingly have great potential. But when I researched quantum computing for the book the main sources I found claimed, in effect, that the functions effectively computable via quantum computing were identical to those computable by Turing machines, only that quantum computing was vastly more efficient in some cases. This is basically what I mean by saying that a Turing machine's computation relates to that of a quantum computer "through isomorphism." In other words, for any computation you could perform with one system you set up a corresponding computation using the other. And there exists algorithms for translating the results of one computation into the results of the other, and vice versa. So the two systems, though very different in how they work, do equivalent things. At least this is my understanding. Admittedly it might be faulty. Maybe quantum computers really will be able to solve the halting problem that Turing machines cannot (nor can classical computers, which do compute exactly the same functions as Turing machines). I think that remains to be seen however. If it turns out that quantum computers really are more powerful and not just more efficient, then, according to Deutsch in *The Fabric of Reality*, you still have the same sort of computational "universe"--only with different operations allowed. In particular you will have a universal quantum computer, analogous to the universal Turing machine. And you can still regard all processing that goes on, all events and all history that is, as as a form of computation. >First of all, quantum computers work with qubits, which simply >aren't the same as marks on a tape and cannot be assimulated to >them. Whoever said that a computer had to be able to solve the >halting problem in order to not be considered a Turing machine? Well, a classical computer is not a Turing machine, true, but computationally it is equivalent to one, and once again, a quantum computer appears to be too, despite some claims--though granted it remains to be seen. >Second, you end up actually agreeing with me about the problem >of making a completely unattached set of symbolic statements >which could be interpreted in only one way. Besides, what about >"right" and "left"? Well, I suppose you could invoke particle physics. There are certain particles that decay differently in their southern than northern hemispheres (as defined by spin direction, I believe), which could distinguish right from left. I won't say there could not be unresolvable ambiguities, but I'm not sure they would be significant. For instance, you could imagine a communication channel occurring between humans and aliens. In the event of ambiguity, you just ask what they mean (or they ask you), providing you already have enough basic understanding to communicate. >If such a set of symbolic statements cannot >be made, then it is not enough for thinking that it be >symbolic, and simulated creatures which basically are our own >symbolic constructions cannot think or even be alive. I don't think I follow you here. To get to concrete cases, do you think a robot with a fancy computer as its brain could have feeling? From some of your postings I do think you allow that something other than protoplasm could have feeling and be alive. But I'm not sure how you draw the lines between what can and cannot, or what I've referred to as "success" versus "failure." Some clarification would be helpful. Best wishes to all and a very long life, Mike Perry Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=21678