00001	TOM Feb 20 volleys in flight at the cease-fire of the pop-philosophy war
00005	It seems that much discussion turns on points taken for granted at
00010	first exposure long ago: impossibility of direct reference.  That
00015	is such a trivial thing it is useful for brilliant argumentation,
00020	but is best forgotten lest it impede conversation. 
00021	Only philosophy has the time to make profundities of such trivia.
00022	 Of course, we
00025	cannot point to a thing.  We can suggest a set of experiments
00030	(look, touch, listen) which have reproducible results and we
00035	can store those results in a coherent way.  To a certain extent
00040	we can communicate the results of experiments to other people,
00045	but  not very accurately.  At best we say "Look, now, there.
00050	Forever after, I mean something like that when I say orange."
00055	There are enormous difficulties in saying "similar", but we share
00060	the same meaning of similar, so saying "forever after" works.
00065	Let me repeat "a model is a computational structure which mimics
00070	the world", ie a model is intended to allow thought experiments.
00075	Thought experiments are safer, quicker, and more economical than
00080	blundering along.
00100	I do not understand what Yorick means when he says explanation.
00200	Does it mean substituting word strings for word strings?  If so,
00300	how does the poor soul who receives this treatment know when to shut up,
00400	ie be satisfied with the explanation (if he perceives that is happening,
00500	he should shut up immediately, of course, and seek better company).
00600	And do you deny entirely reference in language?  Do I or do I not
00700	have a structure "that orange" which refers to the particular orange
00800	which I have brought for lunch?  There are two argumentative copouts
00900	to avoid:
00905		not all things are referential;
01000		reference cannot be infallible.
01005	
01100	Clearly, many words or phrases do not refer to objects; they refer
01200	to models, which refer to other models.  Some models have reference to
01300	objects.
01400	A robot must have a belief in objects in the world (in the above sense).  This might seem
01500	just a useful self-delusion.  It seems more fundamental in that it
01600	corresponds to a discrimination about validity of types of knowledge;
01700	touch/pain and manipulation seem primal-they do not seem equivalent
01800	to vision, hearing, etc.  Solipsism does not seem very important; for
01900	a solipsist to function in a world like mine, he must use descriptions
02000	rather like mine, with all that entails.
02200	So reference is not infallible, but it is consistent and predictable
02300	(I actually can find that orange).