In the definition of logical entailment, all interpretations are taken into account;
there is no constraint. In certain situations, it is desirable to restrict the
possible interpretations to those in which certain constants are assigned values
having to do with the set of sentences itself. In this case, the constants are said
to be *indexical*. An interpretation then is *indexical* if and only if it
assigns these indexical constants correctly.

In KIF, there is a single indexical constant, viz. the object constant `knowedge-base`. An indexical interpretation of a knowledge base is one in which this constant is assigned as value. This one indexical makes it possible for the user to write sentences that depend on the knowledge base within which the sentences are contained.

Finally, we say that a set of sentences *indexically entails* a conclusion if and
only if every indexical interpretation and variable assignment that satisfies the set
of sentences also satisfies the conclusion.

Wed Dec 7 13:23:42 PST 1994