Newsgroups: comp.ai,comp.ai.philosophy,sci.cognitive
Path: cantaloupe.srv.cs.cmu.edu!rochester!hood.cc.rochester.edu!news.acsu.buffalo.edu!news.uoregon.edu!leto!hammer.uoregon.edu!arclight.uoregon.edu!enews.sgi.com!ix.netcom.com!jqb
From: jqb@netcom.com (Jim Balter)
Subject: Re: Emotional computers
Message-ID: <jqbE451JK.1uz@netcom.com>
Organization: Netcom
References: <32DBBF95.64C0@esumail.emporia.edu> <staring-1501970853160001@smisch.tiac.net> <jqbE43Gtx.FFA@netcom.com> <staring-1601970830000001@smisch.tiac.net>
Date: Fri, 17 Jan 1997 05:50:08 GMT
Lines: 63
Sender: jqb@netcom.netcom.com
Xref: glinda.oz.cs.cmu.edu comp.ai:43544 comp.ai.philosophy:51147 sci.cognitive:14490

In article <staring-1601970830000001@smisch.tiac.net>,
Samson <staring@my.screen> wrote:
>In article <jqbE43Gtx.FFA@netcom.com>, jqb@netcom.com (Jim Balter) wrote:
>
>> In article <staring-1501970853160001@smisch.tiac.net>,
>> Samson <staring@my.screen> wrote:
>> >In article <32DCCC23.41C6@cs.bham.ac.uk>, Ian P Wright <ipw@cs.bham.ac.uk>
>> >wrote:
>
>> >As anyone who has nearly had a collision or neary walked off a cliff
>> >knows, the emotional response comes after -- not before -- the system
>> >interrupts its resource-bound processing. How does MINDER1 behave _after_
>> >it avoids destruction?
>> >
>> >Of course, I can't see why anyone would _want_ a robot that starts shaking
>> >and sweating after all danger has passed.
>> 
>> And I can't see why anyone would want to present an argumentum ad ignorantiam.
>> The fright/fight/flight mechanisms resulting from evolution are rather
>> effective; mere failure to understand them does not make them inappropriate.
>> But some insight might be derived from the fact that a race car does not
>> stop short the moment it crosses the finish line, nor does it cease to
>> "uselessly" dissipate heat.
>
>Ah, so this is an emotional response on the part of the car? Does the car
>also rejoice in victory or sulk in defeat?

That wasn't the point of the analogy; the point was that many effects are
"incidental".

>I don't deny the "usefulness" of a f/f/f reaction to an organism -- or in
>some cases, to a computer, provided that once the task of fleeing or
>fighting is completed, "fear" does not incapacitate the agent in
>completing further tasks. 

A description of an ineffective biological mechanism is likely to be
inaccurate.  I suggest that yours is here.

>By the way, you might notice that I made no argument in my post. I asked a

Oh come now.  You argued ("anyone knows" isn't a very *good* argument, but it
is an argument nonetheless) that "the emotional response comes after --
 not before -- the system interrupts its resource-bound processing."  I
suggest that this is inaccurate; there is emotional response throughout the
episode, but subjective memory and post-editing skews the picture.  Secondly,
you argued ("I can't see why" is again not a good argument, but it is an
argument) that one should not "_want_ a robot that starts shaking and sweating
after all danger has passed.".  I suggest that such an emotional response
plays a vital part in the learning process.  There also may be refractory
effects, as I tried to suggest with the race car, that are an aftereffect of
an extremely potent *systemic response* that involves dumping of adrenalin
that very rapidly raises reflex times and energy utilization capabilities.
That costs; there are good reasons why we don't go around all the time in
such an aroused state.

>question: namely, what does MINDER1 do _after_ it has instantaneously
>reacted to adversity in a so-called "emotional" way? How "emotional" is
>MINDER1?

You do well to put those scare quotes there.
-- 
<J Q B>

