From newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!bonnie.concordia.ca!uunet!wupost!micro-heart-of-gold.mit.edu!news.media.mit.edu!minsky Wed Apr 22 12:04:06 EDT 1992
Article 5151 of comp.ai.philosophy:
Path: newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!bonnie.concordia.ca!uunet!wupost!micro-heart-of-gold.mit.edu!news.media.mit.edu!minsky
>From: minsky@media.mit.edu (Marvin Minsky)
Newsgroups: comp.ai.philosophy
Subject: Re: O-I ( other intelligence )
Message-ID: <1992Apr20.010947.23359@news.media.mit.edu>
Date: 20 Apr 92 01:09:47 GMT
References: <kuu171INNqpp@exodus.Eng.Sun.COM> <1992Apr20.001759.10684@cbnewsj.cb.att.com>
Sender: news@news.media.mit.edu (USENET News System)
Organization: MIT Media Laboratory
Lines: 41
Cc: minsky

In article <1992Apr20.001759.10684@cbnewsj.cb.att.com> nea@cbnewsj.cb.att.com (norman.e.andrews) writes:
>The question came up concerning how "numerically aware" insects might be,
>given that they can build hives and spider webs.  I confess I never met
>an insect who was explicitly conceptually aware of the minimum concept "unit",
>conceptual awareness of which is required for the formation of (and also the
>formalization of) of all other concepts, numerical or not.  So I'd say
>based on the evidence (there isn't any I'm aware of) that insects haven't
>any concepts, in spite of their "geometric" abilities.
>
>What would you have insects do, to say they are numerically aware?
>What measurements would they make, geo-metric or otherwise?
>Would they quantify anything, or rank things according to "more" or "less"?
>
>What evidence is there that insects
>have percepts?
>perceive objects?
>recognize objects as entities?
>have minds of any kind (name the kind...)?

Good questions!  I often wonder how to find out, when my cat jumps up
onto a table, if it has some sort of representation for anticipating
finding a surface when it gets there.  In a couple of decades or
centuries, we'll have scanners to show us the wiring diagrams of such
representations.

In the meantime, I think in Tinbergen's "A study of Instinct", he
shows a map of five different locations in which a wasp has deposited
some eggs (in accomodating caterpillers).  The wasp visits each site 3
times, to deposit more supplies there, and then seals off the place.

Presumably some local feature of the state of each site triggers the
appropriate response in the wasp, so it needn't be able to "count to
three".  I have no idea how it remembers where it placed those sites
-- they're many meters apart, as I recall.  And Tinbergen doesn't say
how often the wasps lose track of those locations.  

My favorite instance -- I can't remember if it is in an essay by
Tinbergen or Lorenz -- is an illustration showing what happened when a
bird made the mistake of building a nest on a ladder rung.  As I
recall, there are fully 7 partly finished nests on various rungs,
forming a pretty good Gaussian around the central most-finished one.


