Gerald Penn from the University of Toronto came back to Frankfurt this summer for an entire month of intense implementation work on CLLRS, the constraint language that gives a computational interpretation to Lexical Resource Semantics (LRS). LRS is the semantic framework that we use in teaching introductory semantics, and it is developed actively in various ongoing research projects.
This summer we focused on extensions to the graphical interface for parse output, and on new syntactic constructs in the constraint language. The interface extensions aim at making the concept of semantic contribution that is central to LRS more salient in the graphical output. At the same time, contribution information can be hidden from the output in situations in which it distracts developers.
The new syntactic constructs in CLLRS provide the means to underspecify the structure of logical functors, giving the implementation the same flexibility that semanticians working in LRS are used to. It is now also possible to characterize logical terms by only stating their semantic type, another important construct of LRS for the specification of lexical or phrasal restrictions on semantic composition.