I developed a context-free grammar for the chart parser in NLTK, each covering an interesting and not-entirely-trivial subset of English, along with a lexicon and a set of test sentences. My grammar encompasses verb agreement and subcategorization. Subcategorizaiton refers to the specification that a particular word places on the possible complements (objects) that it can occur with.
My report contains what my grammar can and can’t do. It also includes how I tested my grammar and the limitations my grammar is bounded by.
This grammar was created for the purposes of University of Toronto's CSC485 - Computational Linguistics and redefined by me after the course was over.