🚀 go-pugleaf

In article <Bw2zzB.IsL@usenet.ucs.indiana.edu> chalmers@bronze.ucs.indiana.edu (David Chalmers) writes:
>In article <7694@skye.ed.ac.uk> jeff@aiai.ed.ac.uk (Jeff Dalton) writes:

>>But the views I'm complaining about are in fact widespread.
>>Why just the other day I read someone comparing your views to
>>vitalism and arguing against vitalism (and by implication you)
>>on the verificationist grounds that there were no consequences. 
>>And even someone like Steven Harnad, who thinks the Chinese Room
>>works against the TT, and who must surely count as sophisticated,
>>will argue in favor of the TT as follows (from Comp.ai.philosophy
>>and SIGART):
>>
>>  So if the pen-pal TT is enough to allow us to correctly infer that a
>>  real person has a mind, then it must by the same token be enough to
>>  allow us to make the same inference about a computer,
>
>Certainly there are some verificationists out there.  I think you're
>misreading Harnad, however, who is far from a verificationist; 

I'm not saying he is.  I'm saying that even someone who agrees
with Searle on certain points against the TT may defend the Test
anyway.  There are many other sophisticated defenders of AI who
don't even go that far towards agreeing with Searle.  I'm sorry
if mentioning verificationism earlier in the same paragraph obscured
this point.

In any case, I'm sure I clarified in earlier messages that I am
not saying everyone is a behaviorist or verificationist.  If people
have forgotten this, let me repost it:

  [...] you will have noticed that I said the ideology is _something
  like_ a verificationist (or falsificationist) philosophy of science.
  I also mentioned a multi-headed beast, I seem to recall.

  I am trying to diagnose something I have seen again and again in
  these debates, over several years.  People demand a definition of
  consciousness that can be tested by a third-person observer or
  else, they say, the concept is meaningless.  People defend the
  Turing Test and claim that we judge that other humans have minds
  using such behavioral evidence.  And so on.

  Now, these people do not all agree with each other.  There are a
  number of different views involved, and they can exploit this to
  attack any attempt to characterize them.  Some of them aren't
  behaviorists, some aren't verificationists, etc.  But there is a
  persistent bias towards things that can be assessed from the
  third-person point of view, and I think this needs some explanation.

Let me also say that I think the people involved are, in general,
reasonable people, even when they employ somewhat dubious rhetorical
tricks such as trying to link someone to vitalism.  (Indeed, some of
the things I do are at least that bad as rhetoric.)  I don't mean
any of this as a personal attack.

Instead, my point is that such views are widespread on the net, that
they're the wrong way to approach questions that touch on conscious
experience, and that this sort of thinking can be dangerous if applied
more generally wihtout, shall we say, adequate care.

>he's here making only an obvious point about justification (if X is *enough*
>to conclude Y, then X should cause one to conclude Y in other cases.

Well, that isn't the only sentence he's ever written on the subject.

>Of course, maybe you don't think the TT alone is *enough*, and that
>one is also using e.g. knowledge of internals in the human case).

What I think is worth noting here is that such arguments are offered
in support of the TT and not only by Daryl McCullough.  Indeed, there
are a number of reasonable people who argue in support of the TT.
For instance O'Rourke has argued that our ability to test "understands
geometry" by looking at answers to questions might be extended to
the Chinese Room notion of understanding.

Now, it's entirely reasonable for someone to respond to the "obvious
point about justification" as roughly follows: "sure, behavior used to
seem sufficient, and I've relied on it in the past [I happen to think
this is false, as an account of what most people actually did, but
that's a different issue], but that was before I started to seriously
consider the implications of modern computers and the philosophical
issues around the Chinese Room and AI generally.  So I've had to
revise my view that the pen-pal TT is enough to allow us to correctly
infer that a real person has a mind."

Quite a few people will want to argue against this.  It's a line
of reasoning that didn't get me anywhere with McCullough or Harnad.
But I think the arguments have to go beyond obvious points about
justification.

-- jd