[ExI] Wernicke's aphasia and the CRA.

[Stathis Papaioannou]

2009/12/9 Gordon Swobe <gts_2000 at yahoo.com>:
> --- On Tue, 12/8/09, Stathis Papaioannou <stathisp at gmail.com> wrote:
>
>>> Premise A1: Programs are formal (syntactic).
>> Premise A2: Minds have mental contents (semantics).
>> Premise A3: Syntax is neither constitutive of nor
>> sufficient for semantics.
>
>
>> The A1/A2 dichotomy is deceptive. A human child learns that
>> if it utters a particular word it gets a particular response, and
>> a computer program can learn the same thing. Why would you say the
>> child "understands" the word but the program doesn't?
>
> Presumably the child has mental contents, i.e., semantics, that accompany if not also cause its behaviors, whereas the machine intelligence has only syntactical rules (even if self-written) that define its behaviors.

The child learns to make a particular noise when it is hungry and that
*becomes* semantics. The syntax is more difficult and comes later.

> Keep in mind, and this is in reply also to one of your other messages, that Searle has no problem whatsoever with weak AI. He believes Software/Hardware systems will eventually mimic the behaviors of humans, pass the Turing test and in the eyes of behaviorists exceed the intelligence of humans.
>
> But will such S/H systems have semantic understanding? More generally, will any S/H system running a formal program have what philosophers call intentionality? Never, says Searle. The Turing test will give false positives.

Searle's insistence that weak AI is possible but not strong AI is his
other big failing. He has not addressed David Chalmer's argument in
his 1995 paper (http://consc.net/papers/qualia.html) showing that IF
it is possible to replicate the behaviour of neurons with electronic
replacements THEN any subjective experiences associated with the
original neurons will also be replicated. In fact, I've never seen an
attempt at a rebuttal of this argument by anyone who has understood
it.


-- 
Stathis Papaioannou