Meaning postulates are used in Natural Language Processing
(NLP) to allow restrictions to be placed on the parsing
of sentences based on the meaning
of the word being considered.
Meaning postulates usually take the form of a list of implications, rather like a set of production rules, which can then be compared and combined with the meaning postulates of the other words in the sentence.
The set of all the consistent combinations of meaning postulates is then (supposedly) the set of all the "sensible" readings of the sentence.
Needless to say, this rarely happens.