Newsgroups: comp.ai.alife,comp.ai.philosophy,comp.ai,alt.consciousness
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!howland.reston.ans.net!gatech!newsfeed.pitt.edu!dsinc!ub!galileo.cc.rochester.edu!prodigal.psych.rochester.edu!stevens
From: stevens@prodigal.psych.rochester.edu (Greg Stevens)
Subject: Re: Thought Question
Message-ID: <1995Jan20.204411.19044@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: <3fbdcb$44t@ixnews2.ix.netcom.com> <1995Jan15.225423.23577@galileo.cc.rochester.edu> <3fd097$di1@mp.cs.niu.edu> <1995Jan19.061707.19578@galileo.cc.rochester.edu> <3fm2km$odg@mp.cs.niu.edu>
Date: Fri, 20 Jan 95 20:44:11 GMT
Lines: 62
Xref: glinda.oz.cs.cmu.edu comp.ai.alife:1883 comp.ai.philosophy:24834 comp.ai:26672

In <3fm2km$odg@mp.cs.niu.edu> rickert@cs.niu.edu (Neil Rickert) writes:
>In <1995Jan19.061707.19578@galileo.cc.rochester.edu> stevens@prodigal.psych.rochester.edu (Greg Stevens) writes:
>>In <3fd097$di1@mp.cs.niu.edu> rickert@cs.niu.edu (Neil Rickert) writes:

>>>Is there any good evidence for *innate* heuristic knowledge?

>>Universal grammar?  A heuristic used to cut down the processing time of
>>finding the correct grammar from the input set of words?

>Since I have been arguing against UG on another thread, I am not
>going to be very impressed with UG as an example of innate heuristic
>knowledge.

>Let me comment on why I questionned the idea of innate heuristic
>knowledge.  There seems to be a common assumption (folk psychology)
>according to which people make their decisions by applying rational
>methods to their beliefs.  There is significant evidence contrary to
>this, such as that developed by Kahneman and Tversky.  As I see it,
>heuristics have been introduced as a fudge, so as to account for the
>deficiencies in folk psychology.  I would prefer to say that folk
>psychology is an incorrect causal theory, and that most attributed
>beliefs don't actually exist except as attributions.  Having
>dispensed with folk psychology, I do not need heuristics to prop it
>up.

I can see how you would take this stance, considering the etiology of the
term as a prop for rational choice theory and folk psychology.  However,
I guess I was trying to appropriate the term away from its origins to
simply refer to any process which is not deductive, "rational" or
otherwise normally considered to be propositionally structured.

As a result, I can make a distinction between rules of thumb that we
arrive at either through association or innate predispositions and types
of knowledge we arrive at through deductive means.

Neural networks would account for learned (non-innate) heuristics by
associative mechanisms that are not propositional.  "Most of the time
X is linked with Y, so the next time X happens, let's link it with Y!"
Neural network theory seems inadequate to describe certain kinds of
structured knowledge and deductive knowledge.

Thus, I would term a general rule of thumb as "Phrases must have a head
and can have optional modifiers" as innate heuristic knowlege (assuming
you buy into Chomsky).

But reflexes in general could be considered "innate heuristics" under the
way I am using the term -- blinking when there is sharp movement near the
eye is reflecting heuristic knowledge that things going into your eye can
be bad.

In this sense, if someone has a "gut feeling" or "intuition" about creating,
what is usually associated with "talent," that too would be an innate 
heuristic.

From here, it seems one could disagree with me either by claiming that these
innate mechanisms are not "rules of thumb" i.e. non-deductive, or by arguing
that I should not construe the word "heuristic" in this way.

Greg Stevens

stevens@prodigal.psych.rochester.edu

