Helmreich [53] uses Montague semantics in an elaboration of the viewpoint that most of what is currently treated as ``lexical semantics'' should actually be modeled as ``lexical pragmatics''. As an illustration, some words have two senses that can both be considered as basic. Such cases are claimed to be indeterminate as to which sense speakers use to derive the other. Therefore, if lexical semantics is construed in a normative framework (as with syntax), then this knowledge would be outside its scope. Pragmatic referring functions provide a solution to this and related problems by allowing for the incorporation of context and speaker belief. A Montague semantics framework is used for the formal specification of these referring functions.
Helmreich's work can be construed as making a strong statement about representing word-sense distinctions, namely that in general such knowledge is best handled by pragmatic means rather than semantic ones (e.g., perhaps these distinctions are really just inferences people make when using the words in question). This is one way of looking at the later work by Helmreich and Farwell [52] on the use of pragmatic referring functions as a replacement to lexical functions. They argue that lexical rules apply at the pragmatic level and not at the level of semantics as commonly assumed. They present several arguments in support of this. For example, they point out that the ``semantic relatedness'' of certain uses is marginal at times and thus pragmatic knowledge must be brought to bear in the lexical functions that would derive one use from the other.