From newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!rpi!think.com!wupost!uunet!tdatirv!sarima Mon Mar  9 18:34:44 EST 1992
Article 4215 of comp.ai.philosophy:
Xref: newshub.ccs.yorku.ca comp.ai.philosophy:4215 sci.philosophy.tech:2228
Path: newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!rpi!think.com!wupost!uunet!tdatirv!sarima
>From: sarima@tdatirv.UUCP (Stanley Friesen)
Newsgroups: comp.ai.philosophy,sci.philosophy.tech
Subject: Re: Definition of understanding
Message-ID: <465@tdatirv.UUCP>
Date: 2 Mar 92 23:42:22 GMT
References: <1992Feb23.044200.29383@mp.cs.niu.edu> <1992Feb23.015634.9079@husc3.harvard.edu> <447@tdatirv.UUCP> <1992Feb29.133816.9316@husc3.harvard.edu>
Reply-To: sarima@tdatirv.UUCP (Stanley Friesen)
Followup-To: comp.ai.philosophy
Organization: Teradata Corp., Irvine
Lines: 89

In article <1992Feb29.133816.9316@husc3.harvard.edu> zeleny@zariski.harvard.edu (Mikhail Zeleny) writes:
|SF:
|>or iii) we prefer to leave the question open until actual verifiable evidence
|>is available as to what is true, leaving vacuous logical argumentation to
|>philosophers and reactionaries.
|
|For a number of reasons, you are in no position to leave the question open.
Good grief, i can leave any blasted question open I want to unless it is
a matter of immediate life or death to answer it.

|First, Searle's argument constitutes a legitimate attempt to formulate an
|a priori claim that your enterprise is flawed in principle;

Which has failed because it *is* a priori, and in general I do not accept
a priori arguments unless they are *fully* based on verifiable facts.

| while the
|attempts to refute Searle on philosophical grounds have all failed so far.

But I *am* addressing it on a 'philosophical' ground.  I am saying that his
premises lack observational validity, and must be verified before being used
in any argument, at least if ti is to be conclusive.

|Second, the available empirical evidence, not least of which is the abject
|request for verifiable evidence doesn't address the issue at all,failure of
|AI research to deliver on its promises, is all against you.

O, really?  Given our *abysmal* ignorance of how minds work, I thonk there
is a tremendous *lack* or evidence, on either side.

||Third, if the AI side were really "prefer[red] to leave the question open",
|there wouldn't be so many hacks of the Dennett--Hofstadter type jumping up
|and down screaming at it; the fact is, your side is bothered by it and
|wishes it would go away.

Personally, my main problem is the fear that the money-counters who fund
the research will listen to Searle and prevent us from ever finding the
answer (or at least in my lifetime).

| Finally, the premisses and the methodology of
|Searle's argument are well estabilished both in empirical disciplines, like
|linguistics, and in analytic ones, like logic and mathematics;

The only well-established premisses he uses are generic ones, like
propositional logic.  The ones I doubt, and demand demonstrations of,
are *not* so established.  Or if they are, they should not be, since they
are highly questionable.
|
|SF:
|>I do *not* claim to have any final answer, I just find that Searle's
|>arguments rely too much on unverified, and currently unverifiable,
|>opinions about what constitutes 'understanding'.
|
|I normally abstain from responding to comments of this sort, just as I
|abstain from commenting on the foundations of any other religious belief.
|This time, I'll make an exception.  The whole point of Searle's argument is
|that, if he is right, there is an a priori reason as to why the question of
|computer understanding may not be decidable by empirical evidence like the
|Turing test.

My whole point is that this is true only *if* he is right.  How do we tell
if he is right?  Since his premises are different than mine, whose are
better?  In all honesty, I cannot say with certainty, I may believe mine
to be better, but I, like all humans, have been mistaken at one time or
another.  So, since he is *not* universally held to be right, and his
premisses are, at least in theory, capable of being verified by observation,
*that* is the only reasonable means of determining who is *actually* right.

So, *if* he is right, then further research will make that obvious, if
he is wrong, further research will also make that obvious.  But unless
*every* reasonable person accepts all of his premisses, then a priori
reasoning is *not* conclusive, since at least some reasonable people
disagree with the premisses.  (Actually it is not conclusive even then,
but it at least has claim to being convincing).

|SF:
|>Thus, nothing proven, I keep an open mind.
|
|This has to be the funniest thing I've ever seen on the net!

Why?  That is the best attitude in cases of major disagreement about
matters of fact.

Or are you claiming that whether or not a computer can have understanding
is *not* a matter of fact, but a matter of opinion?
-- 
---------------
uunet!tdatirv!sarima				(Stanley Friesen)



