Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!ames!ucsd!ucsbcsl!silber@sbphy.ucsb.edu
From: silber@sbphy.ucsb.edu
Newsgroups: comp.ai
Subject: Question re: conditions of turing/other "understanding" tests
Message-ID: <1321@hub.ucsb.edu>
Date: 9 Mar 89 17:05:35 GMT
Sender: news@hub.ucsb.edu
Organization: UC, Santa Barbara. Physics Computer Services
Lines: 7

We cannot, practically speaking, exhaustively test the understanding of
a given system.  
Can "optimal" statistical criteria + knowledge domain selections
for an "understanding" test be characterized?   (a highly domain
constrained test with a limited number of interchanges can coincide
with "understanding" (over the small set of interchanges) without
actually being understanding.