Dot-to-Dot Semantic Representation

This paper introduces dot-to-dot semantic representation as an encoding for abstract dependency interpretations of natural language data. The approach is illustrated with content created initially as syntactic tree annotations. This is transformed into construction information for Discourse Representation Structures. This gives a basis for resolving event and entity dependencies across discourse for transformation to Conjunctive Normal Forms with skolemization from which governor to dependent relations arise for graph structures that are dot-to-dot semantic representations. Use of the representation approach is demonstrated with two sample applications: database querying, and analysis feedback for parsed corpus annotation of Old Japanese language data.