Path: utzoo!attcan!uunet!know!cs.utexas.edu!sun-barr!rutgers!news-server.csri.toronto.edu!dgp.toronto.edu!mccool
From: mccool@dgp.toronto.edu (Michael McCool)
Newsgroups: comp.ai.philosophy
Subject: Re: Bird flight as an emergent property
Message-ID: <1990Oct26.133824.8111@jarvis.csri.toronto.edu>
Date: 26 Oct 90 17:38:24 GMT
References: <1990Oct23.170118.27104@ecn.purdue.edu> 	<1990Oct24.114805.3306@idayton.field.intel.com> 	<1990Oct25.100748.2501@jarvis.csri.toronto.edu> <N025FC.90Oct25194726@tamuts.tamu.edu>
Lines: 54

n025fc@tamuts.tamu.edu (Kevin Weller) writes:

>Hmmm ... if my memory of zoology class is accurate (it's been a
>while), birds flying in an "inverted V" pattern do so for aerodynamic
>reasons.  I don't know the physics behind it (yet), but the leader
>supposedly makes the flying easier on the rest of the flock if they
>remain in formation.  The periodic changeoffs take place to give the
>original leader a (relative) rest while the new leader takes over the
>burden.  A bird flying by itself must work harder to stay up,
>significantly harder on long-term (i.e., migration) flights.  I will
>try to find a reference (in my zo book) if you want some real physics.

>-- Kev

Well, when we talk about emergent properties I thought we were talking about
"flocking" behaviour, like masses of sparrows or crows wheeling as a "unit".
This type of behaviour has been well simulated in computer graphics without
taking air drag into account.

The V shape IS more dependent on the physics of the situation, I think,
than typical flocking behaviour.  I feel that the V migration pattern
falls under a different, more specialized and stereotypical type of
behaviour.  The most interesting collective behaviours to me are where
the group acts like an "organism" and responds immediately and appropriately
to its enviroment with seemingly no communication between members, and
where the responses are varied to fit the environment; they can almost
seem to be "intelligent" responses.

Focussing on this type of behaviour also seems more in tune with the
discussion in this group.

Any sociologists out there?  In what respect are related behaviours, i.e.
voting patterns, emergent?  Are there any socially emergent properties 
of groups of PEOPLE?  Especially interesting behaviours are those that
are unplanned, and even uncommunicated, but arise as the result of individual
descisions and adaptations to what others are doing (or are expected to
be doing).  What springs to mind are pricing strategies in ogliopolies,
a la Galbraith's "The New Industrial State": nobody communicates, but
everyone has such a good model of everyone else that prices end up
being stable.  Each company predicts what the other is going to do, and
trys to avoid a price war which would be detrimental to all parties.
The actual price chosen ACTS, and can be treated as, the result of
collusion, even though that is not how it is arrived at.

How does this relate to AI?  Well, the question is, can a society of
autonomous units be treated as an organism even if no *explicit* communication
channels exist?  What are the types of communication between subunits that
can lead to goal-directed behaviour?  Is only interaction through actions 
taken in the common environment enough?  And how do the goals of the subunits
become translated into the goals of the collective body?

Comments? Flames?

Michael McCool@dgp.toronto.edu