00100	COMMENT ⊗   VALID 00002 PAGES
00200	C REC  PAGE   DESCRIPTION
00300	C00001 00001
00400	C00002 00002	YORICK
00500	C00006 ENDMK
00600	C⊗;
     

00100	YORICK
00200	
00300	I'd  like  to  hear  some  final  observations  from  people  on  the
00400	following:  the last discussion  about models has  shown me something
00500	I'd rather not have seen  and also answered Bruce's earlier  question
00600	about the  relation between  Phil-man and  Robo-man.I find  myself in
00700	more  or less complete agreement  with Tom about  the sorts of active
00800	nondeductive models we'd like to build or see built.At  the same time
00900	I  disagree strongly  with him  about the  METAPHYSICS of  models and
01000	reference (some may read "how those  words ought to be used" if  they
01100	wish).  That all suggests that the metaphysics are independent of the
01200	constructive activity and I dont really want to believe that. Further
01300	confirmation of that nasty conclusion  is to be found in our  varying
01400	metaphysical versus constructive alliances  here.I suspect Bruce ,Tom
01500	and I agree on what we'd like to see (something non-deductive,active,
01600	without  strong  theory),   whereas  Arthur  disagrees  strongly  and
01700	Richard holds an intermediate position .But metaphysically Arthur and
01800	Tom at least agree,and I and I think Bruce ,oppose them and what they
01900	agree on.Hence the agreements in the two domains are conflicting,thus
02000	supporting Bruce's independece of philosophy and AI thesis.
02100	
02200	
02300	ARTHUR
02400	
02500	I'm  sure your  analysis of our  respective positions  is accurate,
02600	Yorick, but I don't see  that your conclusion about the  independence
02700	of AI and philosophy follows at all from it. If I wanted to be harsh,
02800	I  might  simply suggest  that  all of  you other  people  are simply
02900	inconsistent in  your views  on metaphysics  and activity,   while  I
03000	simply hold  a strongly  consistent view.   That's an unlikely,   but
03100	surely plausible, conclusion, isn't  it? I must  add that I  consider
03200	holding such a strongly consistent view to be important simply on the
03300	grounds that  it makes me much more  vulnerable to definitive contra-
03400	diction, which is an altogether good thing from a  general scientific
03500	point of view.
03600	
03700	YORICK
03800	
03900	No,Arthur,there's  straightforward  misunderstanding   here.I'm   not
04000	advocating that independence( of philosophy and AI----on the contrary
04100	OUR point of agreement is that  were  both  advocating  consistency).
04200	Bruce  is  advocating  it,  and  I'm raising again for discussion the
04300	possibility that it might be true, and adducing as evidence the cross
04400	agreements  we  have  established.  You correctly point out that your
04500	metaphysics of reference is consistent with your taste in models, and
04600	I  agree.  May  I  point out that that consistency of yours in no way
04700	contradicts  my  also  being   consistent   in   my   non-referential
04800	metaphysics   and  MY  taste  in  models.  You  and  I  can  disagree
04900	fundamentally and still both be consistent.Of course,  what  I  infer
05000	but  didnt  say,is  that its those who CROSS AGREE WITH YOU AND ME ON
05100	MODELS AND METAPHYSICS who are inconsistent. [TOM,this is for you  to
05200	reply to,I suspect] .
05300	
05400	TOM
05500	
05600	I disagree with Yorick's summary of the putative agreement
05700	among Bruce, Tom and Yorick.  I do feel that a formal theory 
05800	independent of semantic domains is not enough; that we must have
05900	detailed structures for individual semantic domains, and that this
06000	knowledge is our dominant interest.  I am not opposed to formal
06100	systems; on the contrary, I favor formal systems.  I infer a
06200	statement that there is an inconsistency between the metaphysical
06300	position (meaning is model-referential) and the constructive 
06400	(build those models as representations of knowledge) position.
06500	I fail to see the inconsistency.  Nor do I really see clearly
06600	what our differences are; they do not seem so clear-cut to me.
06700	When I have tried to explore them, the differences have largely
06800	escaped like steam, and I remain with a feeling that we differ
06900	strongly in our estimate of the pragmatic values of formal systems,
07000	philosophy, and detailed knowledge of individual domains.  Further,
07100	we differ in the problem areas in which we want to make progress,
07200	ie formal systems, language translation, perception.  I sense
07300	among us an intolerance that I expect among the most delightful
07400	people.