Re: One of the posts from the Blog Experiment on Jack Park's site

[Howard Pattee <***>]

Judith and Tim,

I would like to make a comment on Rosen?s statement in Judith?s last post about 
artificial intelligence.
Tim, if you do not consider such comments appropriate for your list, please let me know.

Rosen wrote (Essays on LI):
?An early attempt to pursue biological correlates of technology, was pursued under the 
general (though diffuse and ill-defined) rubric of bionics. Chapter 19 is a review of the 
history of this endeavor; it flourished for less than a brief decade (roughly 1960 to 
1970). As we note, all that exists of it today is the field of artificial intelligence-- 
and that in a vastly mutated form based entirely on a concoction of software, very 
different from what was initially envisioned.?

HP: Bionics is still a flourishing field (e.g. 
http://www.bionik.tu-berlin.de/institut/xstart and links, or Google ?Bionics?). Also, the 
field of artificial intelligence long predated bionics and had little relation to it. AI 
was stimulated by Turing?s question, ?Can a machine think?? (See Turing, ?Computing 
machinery and intelligence,? Mind, 59, 433-46, 1950, which few people have actually 
read.) Early work on problem-solving was done by Newell  (e.g., Newell, Shaw, and Simon, 
?Empirical exploration with the logic machine: a case study in heuristics? in 1957 and 
reprinted in Feldman, ed., Computers and Thought, McGraw Hill, 1963, pp. 109-133). Today 
the field is too vast to characterize.

Newell?s concept of AI, and AI studies today, are certainly not ?based entirely on a 
concoction of software? and it is not mutated from bionics. In the same sense that Rosen 
was a relational biologist, Newell was a relational cognitavist. Newell, like Rosen, 
believed that  ?thought itself? was best understood in terms of its functions and did not 
depend on a particular material embodiment. Furthermore, he understood that thought was 
generally not algorithmic in the formal sense, but heuristic, with the result that his 
programs did not depend on formal axioms or follow strict rules. Trial and error searches 
are necessary and solutions are not guaranteed. Heuristics was coined as a 
problem-solving technique by the mathematician George Polya, one of Newell?s teachers. 
One should understand why impredicativity and non-formal programs are essential for 
heuristic reasoning.

If Bob had known Newell I believe they would have many principled ideas in common about 
biological models.

Howard