Newsgroups: comp.ai.alife
Path: cantaloupe.srv.cs.cmu.edu!nntp.club.cc.cmu.edu!godot.cc.duq.edu!ddsw1!news.kei.com!travelers.mail.cornell.edu!newstand.syr.edu!galileo.cc.rochester.edu!prodigal.psych.rochester.edu!stevens
From: stevens@prodigal.psych.rochester.edu (Greg Stevens)
Subject: Re: Simulating emotion
Message-ID: <1994Dec25.062241.27925@galileo.cc.rochester.edu>
Sender: news@galileo.cc.rochester.edu
Nntp-Posting-Host: prodigal.psych.rochester.edu
Organization: University of Rochester - Rochester, New York
References: <3dhh69$tt5@news.imssys.com> <1994Dec24.181433.6081@galileo.cc.rochester.edu> <3di74q$8g2@news.imssys.com>
Date: Sun, 25 Dec 94 06:22:41 GMT
Lines: 56

In <3di74q$8g2@news.imssys.com> jimk01@imssys.imssys.com (Jim Kennedy) writes:

>I was thinking more of a social-psych kind of simulation,
>where you model social dynamics.  From that point of view, I see emotion
>as the local event, and social order as the emergent phenomenon.  

That is how I'm looking at it, except I feel that under certain parameter
conditions there is a causal link between innate neurophysiology and
emotions and social behavior that social STRUCTURE can be seen as DIRECTLY
emergent from neurphysiological activity.  For the basic premises of 
what the parameter conditions are when this happens, I have co-authored a
paper that has been printed in the Santa Fe Institute working-paper
series, (T. Smith and G. Stevens, "Emergence, Self-Organization and Social
Interaction: Arousal-Dependent Phenomena in Social Structure") and can
probably email some version of it to you.

>If
>I understand you, your model sees neuronal/hormone/neurotransmitter-level
>events as local, and response to another person as the emergent event.
>Of course both views are appropriate!  

In fact, my view is two-tiered: psychological emergent from neurophsiological
and social emergent from psychological -- the catch being that certain
conditions exist when the neurological is not "screened off" from the socal,
so that formal models can capture emergent social structure from studying
properties of neurochemical dynamics (and their links to social attachment).

>I am curious now-- are you considering "attachment" in the sense that 
>Bowlby & Ainsworth etc use it?  

Yes, expanded on the premise that such attachment is not necessarily
restricted to the mother-infant domain and can reappear throughout life
(see Reite and Field, 1985, _The Psychobiology of Attachment and Separation_,
and Smith, 1992, _Strong Interaction_).

>I was thinking of the old "avoidance-
>approach gradient" as a useful way of looking at emotion.  (By the way,
>it seems that laboratory psychology is surprisingly lacking in 
>describing emotional phenomena).  

It depends on where you look.  Laboratory stuff is lacking, yes.  I have had
to look at clinical models and pclinical interaction for data.  Here in
Rochester Highland Hospital has a bio-psycho-social research group which
collects some data which is relevant to the emergent relationship between
psychological emotions and social and biological outcomes.

>As I re-read your comment, I think I should point out that I wasn't
>thinking of simulating emotion itself, but of using it as a parameter in
>a model of another kind of process.

That is, in fact, how I've been using it.

Greg Stevens

stevens@prodigal.psych.rochester.edu

