Newsgroups: comp.ai.philosophy
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!howland.reston.ans.net!cs.utexas.edu!utnut!utgpu!pindor
From: pindor@gpu.utcc.utoronto.ca (Andrzej Pindor)
Subject: Re: Bag the Turing test (was: Penrose and Searle)
Message-ID: <D0KAJw.E8M@gpu.utcc.utoronto.ca>
Organization: UTCC Public Access
References: <CzFr3J.990@cogsci.ed.ac.uk> <3c5nml$370@news1.shell> <3c68og$ql8@agate.berkeley.edu> <3c7dli$a5m@news1.shell>
Date: Fri, 9 Dec 1994 21:04:44 GMT
Lines: 44

In article <3c7dli$a5m@news1.shell>, Hal <hfinney@shell.portal.com> wrote:
>...........
>The HLT seems to have no representation of the richness of our internal
>mental life.  It is little more than a tape recording of all possible
>conversations.
>
>These internal mental states that are apparently missing from an HLT can
>be observed to some extent in the brain.  With electrical probes we can
>observe states of arousal, moments of decision, and other correlates of
>the subjective aspects of consciousness.  But I maintain that there is no
>way even in principle to observe these phenomena in the HLT, because they
>are not there.
>
Sorry if I am repeating myself, but I have an impression that my posting in 
which I have addressed this issue (and Jim Balter also talked about) has not
somewhow made it to your site.
Point is that you are treating HLT based program as basically a static table
which is a mistake. The output has to include past history of the conversa-
tion and has to have a mechanism for making choices between various replies
which would make sense in any given situation. 

>These missing internal mental states are what justify denying that
>the HLT has a mind.  The fundamental structural difference between the
>HLT and biological minds (which are the only things we really know to
>be conscious) give reason to hesitate in extrapolating from our
>personal experiences of consciousness to the assumption that a
>recording of conversations could be conscious.
>
In the process of converstaion program would have to create links between its
database and the conversation history. If every time you started the 
conversation with the same opening the program replied the same way and 
effectively you could make the converstaion run indentically, would you
pass the program? So the program would have to have dynamic internal states to
pass TT convincingly.

>Hal Finney
>hfinney@shell.portal.com

Andrzej
-- 
Andrzej Pindor                        The foolish reject what they see and 
University of Toronto                 not what they think; the wise reject
Instructional and Research Computing  what they think and not what they see.
pindor@gpu.utcc.utoronto.ca                           Huang Po
