Domes as a Prodigal Shape in Synthesis-Enhanced Parsers

Research on logic based bottom-up parsing in particular, around Constraint Handling Rule Grammars [3]is uncovering shape as an untapped fertile ground for natural language processing in general, and for bottom-up parsing and grammar induction in particular [1]. For instance, commonalities between shapes implicit in molecular biology strings and in natural language strings motivated an award winning paper that, based on this commonality of shapes, uncovered a dual processing scheme for both human and biological languages [5] illustrated around CHRG. In this article we investigate more generally useful ways of visually enhancing spoken human language analysis with static or interactive ways of expressing some of the parsing processes in terms of shape. In particular we examine a) shape replication as a visual aid for determining implicit elements in bottom-up parsing, and b) visual reformulation of parsing results as an economical approach to user-interactive disambiguation and correction. Methodologically, we use CHRG as the computational backbone.