From newshub.ccs.yorku.ca!ists!torn.onet.on.ca!utgpu!news-server.csri.toronto.edu!rutgers!ub!zaphod.mps.ohio-state.edu!mips!darwin.sura.net!europa.asd.contel.com!uunet!mcsun!uknet!edcastle!edcogsci!sharder Mon Jun 15 16:04:59 EDT 1992
Article 6238 of comp.ai.philosophy:
Path: newshub.ccs.yorku.ca!ists!torn.onet.on.ca!utgpu!news-server.csri.toronto.edu!rutgers!ub!zaphod.mps.ohio-state.edu!mips!darwin.sura.net!europa.asd.contel.com!uunet!mcsun!uknet!edcastle!edcogsci!sharder
>From: sharder@cogsci.ed.ac.uk (Soren Harder)
Newsgroups: comp.ai.philosophy
Subject: Re: 'home' is where you find it: Re: Transducers
Message-ID: <9744@scott.ed.ac.uk>
Date: 12 Jun 92 11:35:36 GMT
References: <1992Jun10.203412.19158@news.Hawaii.Edu> <4138.708217481@mp.cs.niu.edu> <1992Jun11.055038.9628@Princeton.EDU> <l3f2h3INNarn@exodus.Eng.Sun.COM>
Organization: Centre for Cognitive Science, Edinburgh, UK
Lines: 55

silber@orfeo.Eng.Sun.COM (Eric Silber) writes:

> It may require a robot to pass the 'TTT', but it is not at all clear that
> 'mind', PER SE, requires 'transduction'.  Once I have a 'mind', you
> can lop off all my transduction, and I can still think, so a computer-mind
> could be loaded with the transduction-rich-'ESSENTIALS', and then think,
> (perhaps only about 'abstractions' or 'memories') without further
> 'transduction'.  Yes, I have already submitted this observation, but I
> haven't seen a convincing refutation, so I submit it again!

In that case you have lost the ability to learn, which many people
consider an important part of intelligence. But worse than that, I
believe intelligence is so unstable that it will disappear if not
sustained by a connection to the real world.

[I would have said 'grounding' and not 'connection', but even though I
to a large extent agree with Harnad, I don't want to limit what this
connection can be].

To elaborate on this: the human mind is an eternally transforming. We
get new ideas, change our theories about the world. If we cannot get
feed-back from the world we will slowly drift into a virtual world,
and intelligence will disappear, because it is not needed,
consciousness will also disappear, because there is nothing to be
conscious about. (True, consciousness will linger on for a long while,
while we process our previous experiences, but at some point of time
it will peter out).

throop@aurs01.UUCP (Wayne Throop) writes

>And I'm certainly not convinced that it matters where these
>physical squiggles and squoggles came from, any more than it
>matters what's inside a successful TTT testee.

It does matter. One thing important for intelligence (I believe) is
originality; the ability to make your own mind up on things. (You
might take other peoples _opinions_ in part or whole, but these are
still only part of your intelligent behaviour, and you are still
making the _choice_). 

If the mentality of the AI is just dependent on what the programmar
has foreseen, it is not intelligent. It needs its own intelligence,
and not (like chess games, word processors and all other current
programs that seem intelligent) be dependent on the intelligence of
its constructor and users.

(Sorry if this is a bit condensed. I have to run).

Soren

^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Soren Harder, (MSc student)
Centre for Cognitive Science, 2 Buccleuch Place, Edinburgh
E-mail: sharder@cogsci.ed.ac.uk
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^


