Re: [math-fun] quantum theory foundational issues, my theory of how they should be resolved
David Makin wrote:
I disagree strongly., why are "Turing machines" obviously not consdcious ?
Turing machines are mathematical concepts, not materialistic entities. When programmed correctly, they are capable of miraculously sophisticated behaviour -- probably greater than that of human minds, potentially, due to the unlimited storage -- but they can't possibly be conscious. Here is a thought experiment: Suppose a Turing machine X with initial tape T_0 is `conscious'. Then, by definition of consciousness, it must be aware of its existence. Now, I suppose that cannot be ruled out in the following situation: 1. An electronic computer simulates X with the initial tape T_0, and is therefore conscious. But the Platonic abstraction of `X with the initial tape T_0' has always existed as a mathematical abstraction, so even before anyone built an electronic computer it must have been conscious. Even before the universe existed, `X with the initial tape T_0' has always had a predetermined set of future states. Going in the other direction, what about the following? 2. Two electronic computers each simulate X with the initial tape T_0. Are you suggesting that the `consciousness' therefore inhabits both electronic computers simultaneously? Finally, if Turing machines are conscious, then so are (by emulation) Diophantine equations.
Turing machines *with no method of sensing the outside world* may well be "obviously not conscious" but give the computer the programmed ability to self-evolve and sense the rest of existence - what then ?
Suppose you give the computer a method of sensing the rest of existence. Without loss of generality, it may as well receive a binary input stream through a USB port (or whatever). But, equivalently, we could write the same binary input stream on its initial tape, so this self-contained Turing machine behaves exactly the same as your `computer with a method of sensing the rest of existence'. Sincerely, Adam P. Goucher
There are many definitions of consciousness, such as being able to think or being aware of one's own existence (which seems to demand further clarification). But the one that most interests me by far is what might be called conscious awareness, or simply the having of experiences, or sentience. In any case, to the extent our discussion involves consciousness, maybe we should make sure we are talking about the same thing -- or at least devise terminology to refer to the various things we might want to mean by the word. --Dan On 2013-08-04, at 1:43 PM, Adam P. Goucher wrote:
David Makin wrote:
I disagree strongly., why are "Turing machines" obviously not consdcious ?
Turing machines are mathematical concepts, not materialistic entities. When programmed correctly, they are capable of miraculously sophisticated behaviour -- probably greater than that of human minds, potentially, due to the unlimited storage -- but they can't possibly be conscious. Here is a thought experiment:
Suppose a Turing machine X with initial tape T_0 is `conscious'. Then, by definition of consciousness, it must be aware of its existence. Now, I suppose that cannot be ruled out in the following situation:
1. An electronic computer simulates X with the initial tape T_0, and is therefore conscious.
But the Platonic abstraction of `X with the initial tape T_0' has always existed as a mathematical abstraction, so even before anyone built an electronic computer it must have been conscious. Even before the universe existed, `X with the initial tape T_0' has always had a predetermined set of future states. Going in the other direction, what about the following?
2. Two electronic computers each simulate X with the initial tape T_0. Are you suggesting that the `consciousness' therefore inhabits both electronic computers simultaneously?
Finally, if Turing machines are conscious, then so are (by emulation) Diophantine equations.
Turing machines *with no method of sensing the outside world* may well be "obviously not conscious" but give the computer the programmed ability to self-evolve and sense the rest of existence - what then ?
Suppose you give the computer a method of sensing the rest of existence. Without loss of generality, it may as well receive a binary input stream through a USB port (or whatever).
But, equivalently, we could write the same binary input stream on its initial tape, so this self-contained Turing machine behaves exactly the same as your `computer with a method of sensing the rest of existence'.
Sincerely,
Adam P. Goucher
_______________________________________________ math-fun mailing list math-fun@mailman.xmission.com http://mailman.xmission.com/cgi-bin/mailman/listinfo/math-fun
On 8/4/2013 1:43 PM, Adam P. Goucher wrote:
David Makin wrote:
I disagree strongly., why are "Turing machines" obviously not consdcious ? Turing machines are mathematical concepts, not materialistic entities. When programmed correctly, they are capable of miraculously sophisticated behaviour -- probably greater than that of human minds, potentially, due to the unlimited storage -- but they can't possibly be conscious. Here is a thought experiment:
Suppose a Turing machine X with initial tape T_0 is `conscious'. Then, by definition of consciousness, it must be aware of its existence. Now, I suppose that cannot be ruled out in the following situation:
1. An electronic computer simulates X with the initial tape T_0, and is therefore conscious.
But the Platonic abstraction of `X with the initial tape T_0' has always existed as a mathematical abstraction, so even before anyone built an electronic computer it must have been conscious. Even before the universe existed, `X with the initial tape T_0' has always had a predetermined set of future states. Going in the other direction, what about the following?
2. Two electronic computers each simulate X with the initial tape T_0. Are you suggesting that the `consciousness' therefore inhabits both electronic computers simultaneously?
Finally, if Turing machines are conscious, then so are (by emulation) Diophantine equations.
Turing machines *with no method of sensing the outside world* may well be "obviously not conscious" but give the computer the programmed ability to self-evolve and sense the rest of existence - what then ? Suppose you give the computer a method of sensing the rest of existence. Without loss of generality, it may as well receive a binary input stream through a USB port (or whatever).
But, equivalently, we could write the same binary input stream on its initial tape, so this self-contained Turing machine behaves exactly the same as your `computer with a method of sensing the rest of existence'.
I think the correct inference is that "to be conscious" requires a an evnironment, something to interact with. Then for a Turing machine to be conscious it either has to be implemented physically in a world with some input-output, as in a robot. OR it could be simulating a whole world including as part of that conscious beings - and in that case it would not necessarily have to be physically realized. This is the theory of Bruno Marchal who claims to prove that the physical (and the mental) are both implicit in an abstract computation by a Universal Dovetailer. Brent
But, equivalently, we could write the same binary input stream on its initial tape, so this self-contained Turing machine behaves exactly the same as your `computer with a method of sensing the rest of existence'.
No because neither you, nor anyone (or anything) else could predict what that binary stream should be *until the time when that state exists* (true based on either uncertainty in the current physics model or based on my own belief that existence is fractal in nature). On 4 Aug 2013, at 21:43, Adam P. Goucher wrote:
David Makin wrote:
I disagree strongly., why are "Turing machines" obviously not consdcious ?
Turing machines are mathematical concepts, not materialistic entities. When programmed correctly, they are capable of miraculously sophisticated behaviour -- probably greater than that of human minds, potentially, due to the unlimited storage -- but they can't possibly be conscious. Here is a thought experiment:
Suppose a Turing machine X with initial tape T_0 is `conscious'. Then, by definition of consciousness, it must be aware of its existence. Now, I suppose that cannot be ruled out in the following situation:
1. An electronic computer simulates X with the initial tape T_0, and is therefore conscious.
But the Platonic abstraction of `X with the initial tape T_0' has always existed as a mathematical abstraction, so even before anyone built an electronic computer it must have been conscious. Even before the universe existed, `X with the initial tape T_0' has always had a predetermined set of future states. Going in the other direction, what about the following?
2. Two electronic computers each simulate X with the initial tape T_0. Are you suggesting that the `consciousness' therefore inhabits both electronic computers simultaneously?
Finally, if Turing machines are conscious, then so are (by emulation) Diophantine equations.
Turing machines *with no method of sensing the outside world* may well be "obviously not conscious" but give the computer the programmed ability to self-evolve and sense the rest of existence - what then ?
Suppose you give the computer a method of sensing the rest of existence. Without loss of generality, it may as well receive a binary input stream through a USB port (or whatever).
But, equivalently, we could write the same binary input stream on its initial tape, so this self-contained Turing machine behaves exactly the same as your `computer with a method of sensing the rest of existence'.
Sincerely,
Adam P. Goucher
_______________________________________________ math-fun mailing list math-fun@mailman.xmission.com http://mailman.xmission.com/cgi-bin/mailman/listinfo/math-fun
The meaning and purpose of life is to give life purpose and meaning. The instigation of violence indicates a lack of spirituality.
participants (4)
-
Adam P. Goucher -
Dan Asimov -
David Makin -
meekerdb