Re: SUO: What the hell was CSP talking about?
Jon and Doug,
Peirce insisted that his "non-psychological" theory of logic
(and signs in general) applies to "any scientific intelligence",
by which he meant "any intelligence capable of learning from
experience". In his writings, he gave frequent examples of
intelligence in dogs and parrots, and he was very interested in
the primitive computer-like things of his day, including Babbage's
machines, and various mechanical things for solving logic problems.
In fact, he was the first person to observe that electrical circuits
could be used to represent logical AND and OR, and he suggested
electricity as a better mechanism for supporting logic machines.
I agree with Jon's points:
DM> > My (limited) reading of Peirce has led me
> > to the conclusion that he is talking about
> > the mechanism of cognition.
JA> This is a possible application of pragmatic sign theory
> but it does not constitute a necessary limitation on it.
> If you read the material that I have cited on this score,
> I think that you will notice the characterization of logic
> as "formal semiotic" and the emphasis that Peirce lays on
> what he calls his "non-psychological" conception of logic.
DM> > The interpretant is an inside-the-mind thing, ...
JA> This is a "can be", but not a "must be" of the theory.
DM> > The sign is a representation of some outside-the-mind thing,
>
JA> Again, maybe, maybe not.
DM> > Jon seems to be talking about triads where all three
> > parts of a triad are external representations. Sowa
> > seems to support the conclusion that Peirce is talking
> > about thought, where Awbrey seems to be saying that he
> > is talking about representation.
I agree with Jon on this point. Signs can be in the mind, but
the triadic representation relation can support "mind-like"
relationships in a wide range of media. In another of Peirce's
examples, he suggested "crystals and bees".
JA> One of the features of this formal approach, or what we
> would more often today refer to as an "abstract" theory,
> is that all of these worries about the moment to moment
> locations of objects, signs, interpretants are entirely
> incidental to the "form", id est, the isomorphism class,
> pattern, shape, or structure of the whole sign relation.
> The in-or-out of your mind question is an "accidental",
> a secondary characteristic that gets "abstracted out"
> of the "formal" question, which is all that is really
> and truly of the "essence" here.
In Section 6.6 of my KR book, I show how Peirce's non-psychological
semeiotic can be used to analyze and resolve several of the troubling
mind-body issues that plague discussions of "artificial intelligence":
John Searle's Chinese room example, the talk about symbolic vs.
imagelike reasoning, subsymbolic processes in neural nets, and the
discussions of phenomenology by Hubert Dreyfus and Terry Winograd.
In fact, I say that instead of "artififical intelligence", a more
accurate term for what we are trying to do with computers would be
"computational semiotic". Using that term would immediately dispense
with the tons of verbiage about whether computers can "really" think.
Instead, it would focus the discussion on what kinds of signs are
used at various stages of the process and how they are related to one
another.
John Sowa