From newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!rpi!usc!wupost!uunet!psgrain!qiclab!nosun!techbook!szabo Wed Apr 22 12:04:05 EDT 1992
Article 5148 of comp.ai.philosophy:
Xref: newshub.ccs.yorku.ca sci.environment:8019 comp.ai.philosophy:5148
Newsgroups: sci.environment,comp.ai.philosophy
Path: newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!rpi!usc!wupost!uunet!psgrain!qiclab!nosun!techbook!szabo
>From: szabo@techbook.com (Nick Szabo)
Subject: The productivity of models of intelligence
Message-ID: <1992Apr19.080246.21406@techbook.com>
Organization: TECHbooks of Beaverton Oregon - Public Access Unix
References: <40911839@hpindda.cup.hp.com> <1466601338@igc.org> <JMC.92Apr17211816@SAIL.Stanford.EDU>
Date: Sun, 19 Apr 1992 08:02:46 GMT
Lines: 49

In which Alan McGowen again regales us with the overwhelming value of
biology and the evil of everything else.  In this case, it is alleged
that cognitive science and AI have failed; that biology provides
much more productive models for intelligence.  McGowen claims a
dearth of commercialization for cognitive-based AI, while conveniently
not mentioning commercialization of biological-based AI.

John McCarthy addressed recent AI research well, including the market in 
expert systems.  Going back a bit further in intellectual history, 
the very circuitry in our computers and calculators is organized around 
Boole's "laws of thought".  Programming languages and the basic architecture 
of modern computers are based on "an automata theory describing the 
organization and dynamics of cognitive processes", inspired by Leibniz and 
brought to life by Von Neumann and Turing.  

The modern computer's inventors were also AI researchers; their inspiration
was the human mind and their goal to create intelligent machines.  They 
succeeded to a quite signficant extent that we now take for granted.  For 
example, calculating speed was once considered a sine quo non of human 
intelligence, as well as being quite important to science and industry.  
"Computer" was a human job description up to the 1960's.

Similarly, the relational databases McGowen raves about originated in 
research on logic and sets, not in biology.  Image processing, while not 
"AI" by the narrow definition, is one of the most recent advances based on 
logical rather than biological models.  I could go on to describe the
plethora of software -- spreadsheets, word processors, desktop publishers,
compilers, operating systems, GUIs, CAD, this very Usenet, etc. layered over 
Von Neumann's logical model of intelligence.  Using philosophical and 
mathematical constructs to describe and mimic cognition has a much longer 
and more sucessful track record than using biological models.  This isn't to 
say that the synergism between philosophical and biologcial models might not 
be even more productive, but McGowen's dismissal of decades of spectacularly 
successful intellectual work and $hundreds of billions worth of industry is 
bizarre.

Perhaps the inspiration might be working more the other way around.
Note how often McGowen uses ideas derived from Von Neumann's game theory 
and Weiner's cybernetics to describe his beloved ecology.  These theories 
are so entrenched into the ecologist's mindset, he might not even be aware of 
these debts.  Game theory originated as a model of human economic choices 
-- a cognitive theory.  Cybernetics started as an abstract model of 
goal-directed behavior.  McGowen looks at the very thing he would like 
to use as a model through the lenses of a philosophy he claims to be sterile.


-- 
szabo@techbook.COM  Public Access User --- Not affiliated with TECHbooks
Public Access UNIX and Internet at (503) 644-8135 (1200/2400, N81) 


