Intensionalizing-AMR

Accompanying code for Williamson et al. (2021). Intensionalizing Abstract Meaning Representations: Non-Veridicality and Scope. In Proceedings of The Joint 15th Linguistic Annotation Workshop and 3rd Designing Meaning Representations Workshop.

Contents translation-scripts

• main.py : converts AMR penman.Graph objects into nltk.sem.logic.Expression objects.
• node_depth.py : uses penman.Graph.epidata to create tree information stored as a dictionary of triple-depth key-value pairs.
• translation_functions.py : contains translation functions.
• World.py : defines the World class.
• Operators.py : defines Existential class, Universal class, and HigherOrder class for representing quantificational determiners. Defines Negation, And, Or classes for propositional negation and connectives.

Contents conversion-script

• convert_content_role.py : converts intensional AMR :ARGs to :content role.
  • Uses list of verbs extracted from the MegaVeridicality Project v1 data set (http://megaattitude.io/projects/mega-veridicality/)
  • Reference: White, Aaron Steven, and Kyle Rawlins. 2018. The Role of Veridicality and Factivity in Clause Selection. In Proceedings of the 48th Annual Meeting of the North East Linguistic Society, edited by Sherry Hucklebridge and Max Nelson, 221–234. Amherst, MA: GLSA Publications.