Re: Indefinite Extensibility and Computationalism

On Jun 20, 4:55 pm, Aatu Koskensilta <aatu.koskensi...@xxxxxxxxx>
wrote:
stevendaryl3...@xxxxxxxxx (Daryl McCullough) writes:
Not quite. I don't believe that concepts *have* unique extensions,
but if they are coherent they can be given a set-theoretic interpretation.

Can the concept of being a juicy hamburger, which on the face of it is
perfectly coherent, be given a "set-theoretic interpretation"?

I don't agree, but I don't want to argue about it right now.
The point is this:

(1) A computer program can very well make the argument that
you've outlined.
(2) In that case, the conclusion is false.
(3) Therefore, the argument is invalid.

Maybe you think that the argument is valid if it is truly
*asserted* but is invalid if it is merely uttered. I don't
think that makes sense, but assuming it does, what reason
is there for believing that *humans* can assert your argument
(as opposed to merely uttering it)?

This is indeed a most telling objection to the various "purely
logical" arguments for the non-mechanicality of this or that piece of
human competence; it is invariably the case that there is nothing to
stop a robot from making precisely the same argument.

There is a trivial sense in which no robot can make an argument, as
you would probably say. But the words that a robot can output can
indeed reproduce the argument when they are read and this causes no
trouble to the argument. There is a Turing machine able to 'prove'
Turing's theorem on the limit of Turing machines, and this says
nothing against the theorem.



As a more general observation we can also note that all these
arguments about "computationalism" and what not share the same defect:
it is without any justification claimed that if, in some unspecified
sense, human cognitive behaviour is "computational", such things as
"believing that ...", "knowing that ...", "wondering whether ...",
must be definite, "physical" and "computational". In a trivial sense
an algorithm can no more believe or know anything than it can eat a
hamburger. So to support these assertions some explanation, some
theory, of what it would mean for an algorithm to believe or know
something must be provided -- or it must be argued that the assertions
are consequences of any such explanation. Nothing of that kind has
been put forth, and there is indeed no reason to think that "Robbie
believes P", "Robbie knows P", "Robbie thinks about P", are any more
definite when Robbie happens to be a robot than when Robbie refers to
Robbie Lindauer.


I have set forth a couple of arguments resting on the fact that a non
propositional sentence would have to be propositional if
computationalism held. The point is not that something is more or less
definite for robots than for humans.; it is not about vagueness,
fuzziness or definiteness.The point is just that algorithms are
perfectly objectified so that it is impossible to produce a non well
founded (or ungrounded) sentence about them, as it is impossible to
produce one about natural numbers.

I think you miss the point.

Best regards

.

Relevant Pages

  • Re: Indefinite Extensibility and Computationalism
    ... There is a trivial sense in which no robot can make an argument, ... of what it would mean for an algorithm to believe or know ... definite when Robbie happens to be a robot than when Robbie refers to ... definite for robots than for humans.; ...
    (sci.logic)
  • Re: consciousness
    ... exact same way the location of a wire in a camcorder gives rise to the ... signal processing circuits that break the visual information into signals ... Why don't you try to explain to me how a robot is able to see color without ... That is, both humans, and robots, can produce ...
    (comp.ai.philosophy)
  • Re: Can artificial intelligence be a danger to life?
    ... want to build will ever become a threat to humans, ... created by hard wired systems which define pain and pleasure for it. ... reward, and punished, based on how much money it makes. ... Another simple way to motivate a robot would be to get a remote control ...
    (comp.ai.philosophy)
  • Bots on The Ground
    ... Bots on The Ground ... In the Field of Battle, Robots Are a Soldier's Best Friend ... At the Yuma Test Grounds in Arizona, the autonomous robot, 5 feet long and modeled on a stick-insect, strutted out for a live-fire ... Humans have long displayed an uncanny ability to make emotional connections with their manufactured helpmates. ...
    (soc.retirement)
  • Re: Forged vs rolled steel?
    ... solvability of Chess or Go really isn't relevant. ... "think outside the box" like humans do. ... your definition of "robot" which excludes whatever it is humans do to ... sort of creativity that does not come from strict logic. ...
    (rec.arts.sf.science)