From newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.ecf!utgpu!cs.utexas.edu!usc!sol.ctr.columbia.edu!destroyer!gumby!yale!cs.yale.edu!mcdermott-drew Tue May 12 15:49:41 EDT 1992
Article 5483 of comp.ai.philosophy:
Newsgroups: comp.ai.philosophy
Path: newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.ecf!utgpu!cs.utexas.edu!usc!sol.ctr.columbia.edu!destroyer!gumby!yale!cs.yale.edu!mcdermott-drew
>From: mcdermott-drew@CS.YALE.EDU (Drew McDermott)
Subject: Re: Functionalist Theory of Qualia
Message-ID: <1992May8.145326.12358@cs.yale.edu>
Keywords: qualia,functionalism
Sender: news@cs.yale.edu (Usenet News)
Nntp-Posting-Host: aden.ai.cs.yale.edu
Organization: Yale University Computer Science Dept., New Haven, CT 06520-2158
Date: Fri, 8 May 1992 14:53:26 GMT
Lines: 125

   At some remote time in the past, David Chalmers wrote:

   [Sorry about the slow reply on this one.  It slipped past me somehow.]

I've been busy, too, but here I am.  The following reply does not
address topics in the original order.

   I originally wrote:

   >I'd like this argument to show that qualia are totally vacuous.  But
   >I'm afraid all it really shows is that qualia are probably
   >supervenient on the physical world somehow, which everyone but
   >Chalmers probably already accepts.

   ... 

   Incidentally, I certainly believe that qualia are supervenient on the
   physical.  The supervenience is contingent, that's all.

Let me try to strengthen my argument, to prove the following claim: If
there is a computational (functionalist) account of behavior and
intentionality, then conscious sensations are computational or even
physical events.  I take it that this is a stronger claim than
"contingent supervenience."  That is, if they're purely computational,
then any system realizing the same computational structure would have
the same qualia.  If they're physical, then there can't exist a world
with the same physical structure without the same qualia in the same
places.  In either case, the qualia-free Earth is impossible.

The argument goes as follows: A key aspect of a conscious sensation (a
mental event with "qualitative" content) is its similarity to other
conscious sensations.  When we say that an experience has a certain
"red quale," we mean that its the same quale (or very close to being
the same) as previous experiences that we've come to label "red."
But, by hypothesis, there is a computational account of all such
labelings and similarity judgings, because they're in the causal
stream that ultimately accounts for behavior.  If qualia are not
computational or physical, then there could exist a system with the
same computational or physical states but quite different qualia.  But
"difference of qualia" is impossible without "difference of judgement
of similarity."  There's no way to put in a tap and siphon off some of
the qualia to see if their owner is correct in his judgements of their
similarity.

You might think we could get a counterexample of this sort: Suppose
Fred sees a red car on Monday, but by Wednesday misremembers its color
as yellow.  He then sees a yellow bird, and claims it's the same color
as the car.  Is he not mistaken about his own qualia?  No.  He's
mistaken about the color of the car.  Presumably, the quale of his
experience on Monday has been lost, replaced by a bogus memory.  

If qualia have any reality at all, then we have to separate two
judgements: (a) the judgement of which sensations occurred at which
times; (b) the judgement of what a particular sensation was like.
Judgements of type (b) are the ones that I am claiming are impossible
without comparisons of sensations at different times.  These
comparisons are computational events, and so the qualia are
computational events, too.

Now back to Chalmers's particular objections:

   I'm not sure, however, that one has a
   "subjective experience that the two qualia are the same"; one certainly
   has a subjective experience of two qualia, and forms a judgment that
   they're the same, but it's not obvious that that judgment is itself a
   subjective experience in the sense that qualia are.

It needn't be.

   [me:]

     >I could imagine an evil neurosurgeon reaching in and fiddling with the
     >output of the comparator entirely independent of the actual contents
     >of the color signals from the retina, so that I might say, "This red
     >has hue H, saturation S, etc. [assuming I'm properly trained]," and
     >also say with complete sincerity, "It's quite different from that
     >other red with, even though I know I reported that it also had hue H
     >and saturation S; I must have been mistaken."

   [chalmers:]

   ... But even if ... the two experiences really are qualitatively identical,
   but are judged different -- I don't think that's a real problem.

What could it mean to say that they're identical?  Where can they be
brought together to be compared if not in the mind of their owner?

   [me:]
     >Okay, let's suppose that the universe is only like this from 9 to 5
     >(Eastern Standard Time) every day.  In other words, we only really
     >have qualia after work.  The rest of the time we don't really exist.
     >Of course, we don't notice this because all our memories are intact at
     >5 o'clock; there's no contradiction in supposing that we can have
     >memories of qualia we never really experienced.  It's enough that when
     >we retrieve a memory there's a certain qualitative pang.

   [chalmers:]

   I don't see that this shows anything.  It's conceivable, consistent,
   and almost certainly false.  It seems to be precisely of a kind with
   Descartes' evil demon hypothesis, and the hypothesis that the world was
   created five minutes ago.  One can't *disprove* such hypotheses (well,
   not unless it's between 9 and 5, but unfortunately it's 7 pm now),
   but one shouldn't believe in them either.  Is the very fact that this
   hypothesis is conceivable meant to be a reductio ad absurdum of qualia,
   or of dualism?  That doesn't seem to follow, any more than the
   five-minutes-ago hypothesis is a reductio ad absurdum of the concept
   of the past.

You're right.  I withdraw that argument.  But the reason it's so
tempting is that the notion of quale allows so little evidence to
bear on what qualia we actually have.  If qualia are not properties of
information-processing events, the the qualia fan has to believe that
we could be completely certain in our claims that two qualia were the
same when "really" they could be routinely, wildly different.  For
instance, God could have created the world with just one "generic
sensation."  All experiences actually have this single quale, caused
by everything from pizza to poison ivy.  Of course, we *judge* them to
be different, but that's just an information-processing event.  Okay,
okay, this is another overly-skeptical argument.  It comes from there
not being enough to qualia for us to be able to get ahold of them.

                                             -- Drew McDermott




