From newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.ecf!utgpu!cs.utexas.edu!qt.cs.utexas.edu!zaphod.mps.ohio-state.edu!pacific.mps.ohio-state.edu!linac!att!emory!gwinnett!depsych!rc Tue Jan 21 09:26:59 EST 1992
Article 2869 of comp.ai.philosophy:
Path: newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.ecf!utgpu!cs.utexas.edu!qt.cs.utexas.edu!zaphod.mps.ohio-state.edu!pacific.mps.ohio-state.edu!linac!att!emory!gwinnett!depsych!rc
>From: rc@depsych.Gwinnett.COM (Richard Carlson)
Newsgroups: comp.ai.philosophy
Subject: Searle and the Curse of the Mediocre Philosophers
Message-ID: <FN6qeB2w164w@depsych.Gwinnett.COM>
Date: 18 Jan 92 15:46:02 GMT
Lines: 155

It has seemed to me for some time that philosophers, especially
mediocre philosophers, have the very bad habits of generalizing
from inadequate data and of using very shaky and fuzzy concepts as
if they were rock-solid, empirically grounded and unproblematic
commonplaces.  A good example of these philosophical felonies is
John Searle's Chinese Room.  After several weeks of heated
discussion and tortuously analytic reasoning, everybody seems to
be coming to the conclusion that the Chinese Room proves nothing
at all.

Let me first state in my own words what I think is wrong with it
and then look at what some others are saying.  Primarily it is an
argument which uses the philosophical notion of "in principle."
(The dark side of the moon was knowable "in principle" even before
space flight but whether animals think is not knowable in
principle since we can't think of any procedure that would tell us
-- never mind that the example of animals thinking immediately
calls to mind anecdotes of animals like Clever Hans, the counting
horse, and the Turing-Test-like procedures used to assess his
arithmetic abilities -- and whether there is conscious life after
death is not knowable in principle -- unless we are willing to
consider the Turing-Test-like procedure by which Houdini unmasked
spiritualists or the Turing-like "Houdini test" he proposed,
namely communicating with his wife after his own death -- hmm, odd
that these "unprovable in principle" paradigms should suggest a
Turing-like test, but who knows what that means.)

Searle argues that even though it may be possible "in principle"
to simulate (mimic, fake, model, whatever) human language
behavior, there  would still be no "real" understanding.  He
places a real, live human brain (his own) in the circuit, and
"shows" that there is no understanding of the semantic processes
modeled.  Actually, I think in an attempt to make his argument
stronger -- and just why he thought it would in fact make his
argument stronger is something to consider -- he makes it the most
"human" and least "machine-like" form of language, namely
narrative language -- it was Roger Schanks' indexing scheme that
was proposed as the program that was simulating the
"understanding," a not unimportant point, although no one in the
thread has made much of it.  But the point is that that human
brain is itself merely used to mimic or simulate a computer
processor, so the fact that it is organic is really irrelevant to
his argument.  It is just there to be a brain which is out of the
loop of understanding (but in a rote computational loop, like a
clerk -- say Bob Cratchet, who copies down Ebenezer Scrooge's
calculations and figures but doesn't understand the thought behind
them that leads Mr. Scrooge to buy this week's corn at this price
but not last week' corn at last week's prices), although a brain
should "in principle" understand what it is doing.

It took a month for everybody to realize that Searle has _assumed_
the passing of the Turing Test.  (Or even if they did realize it,
some of the early posters did not seem to realize the significance
of the assumption.)  All he's saying is that the Turing Test
wouldn't per se prove that the processes in the brain were
identical or isomorphic to the processes in the computer which
simulated them.  But everybody already knew this!  Or almost
everybody.

Why did he do this?  Why did John Searle "jump ahead" to the
"next" step before the current one was fairly taken?  My answer is
that it is because he is a mediocre philosopher.  That adjectival
qualifier is important.  Not just that John Searle is a
philosopher, but that he is a *MEDIOCRE* philosopher.  All
philosophers use crude and shaky concepts as the basis for their
reasoning -- they have to, otherwise they'd be scientists, not
philosophers.  These crude constructs are made up of chains of
signifieds, some of which may be inappropriate or contradictory.
The creative philosopher has a mind which enables him to process
this semantic information in such a way that it gives him clues
and hunches about where to look for the next meaningful facts and
he points the way.  The mediocre philosopher, operating with the
tokens in his own head, with their less richly articulated chains
of signifieds, zeroes in on irrelevant or misconceived aspects of
the construct, which might look "crystal clear" to him, but are in
reality false starts and dead ends.

By eliding the question of whether a computer constructed on the
lines of today's computers could in fact pass the Truing Test, and
how they might go about doing it, and jumping ahead to an imagined
day when they have done so (with no real or philosophically
relevant consideration of the way they might have done so except
the guess that it would be something like Roger Schanks's
narrative indexing procedure -- which guess, as I have mentioned,
no one has even picked up on), he refocuses attention from the
interesting scientific and philosophical questions of the
mechanisms of "semantics" and "understanding" and leaves us with
pseudo-questions like, can a "room"  (i.e., four walls and some
furniture and a person inside) "understand." The potential for
fruitless semantic bickering is endless.

My point is, what we call "philosophical" thought is not "logical"
thought, it is "semantic" or phenomenological thought.  That's
what we mean by calling "philosophical" thought "reflective."  As
such it is inherently non-algorithmic (even when it is being used
to create algorithms!), so it can lead either to brilliant
insights or to garbage and there is not algorithmic or
deterministic decision procedure to judge any sample of
philosophical thought (any philosophical "text," since reflective
or phenomenological thought is inherently and irredeemably
_textual_, i.e., "semantic"), leading to the conclusion that
philosophical thought must be judged pragmatically and
heuristically by the conclusions and findings to which it lights
the way.

Can I "prove" that?  Not in general, but let us return to a
"intellectual historical" consideration of this Chinese Room
thread in this comp.ai.philosophy Newsgroup.

Jeff Dalton writes:
>I suspect that what you're getting at is that if I think conversation
>without understanding is impossible, then I should accept the Turing
>Test, because whenever there was conversation there would (in my view)
>have to be understanding.  Well, if I could _show_ that conversation
was impossible without understanding, then I should indeed accept
>the Turing Test.  But I can't show it's impossible, and neither can
>the people who want us to accept the TT right now.
>
>The arguements for acepting the TT right now do look rather like
>residual operationalism and behaviorism.  They often involve saying
>(or implying) that there's no way to test for "real understanding",
>that the question of "real understanding" is meaningless or
>unscientific, and so on.
>
>Another point I think you were making before was that if Searle can
>show that computers can't understand by using "syntax isn't enough for
>semantics", then what does the Chinese Room add?  Well, you can think
>of it as Searle having two arguments, or an argument and an example
>(or, as Dennett says, an intuition pump).  Since different people may
>find different arguments convincing, why not use both?  Note that
>using both does not mean putting them together in one argument as
>you did, I think, in <1992Jan14.151104.16978@aifh.ed.ac.uk>.


It is, of course, the second argument, the "in principle" argument
that calls focus to itself precisely because the Turing Test is
presently inconclusive.  It hasn't been passed yet!  It is
Searle's act of jumping over that fact and "granting" that it
_might_ "in principle" be passed which creates all the semantic
confusion.

Kristoffer Ericksson writes:
>Apparently some claim that the Chinese Room would still not have "real"
>understanding, even it it were to pass this test (which of course IS the
>famous Turing test), since Searle has "showed" that it can't have that.


This is wonderfully clear and succinct.  But it took a month to
get here!

--
Richard Carlson        |    rc@depsych.gwinnett.COM
Midtown Medical Center |    {rutgers,ogicse,gatech}!emory!gwinnett!depsych!rc
Atlanta, Georgia       |
(404) 881-6877         |


