perm filename CONTEX[F84,JMC]2 blob sn#782642 filedate 1985-01-16 generic text, type C, neo UTF8
COMMENT āŠ—   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	contex[f84,jmc]		Ideas about context
C00004 ENDMK
CāŠ—;
contex[f84,jmc]		Ideas about context

	Consider the following first order treatment of context.
We have  value(concept,context)  as a function and  true(proposition,context)
as a predicate.  We have modifiers of both concepts and contexts
and ``trade-off axioms'' like

	true(future(prop),context) ≔ true(prop,future'(context)).

Contexts are regarded as rich entities, like situations.

contexts can be values.

Suppose we consider a system that uses sentences.  The system includes
both the program and the language in which the sentences are written.
The situation is simplest if the language has a semantics that
can be described without reference to the program that uses the
language.  However, these matters are sufficiently mixed up in
humans that we have no guarantee that the separation is always
possible.  Are  the RUM continuations relevant to this?

1985 Jan 16

Another trade-off axiom

true(on(x,y,s),c) ≔ true(on(x,y),assign(situation,s, c))

Actually it isn't quite an assignment; c is a context without
an assigned situation.  However, on(x,y) can be true only in
a context in which there is some situation specificied.