Appendix G
Combinatory Categorial Grammar
Appendix: Combinatory Categorial Grammar, in Daniel Jurafsky and James H. Martin, Speech and Language Processing, 3rd edition draft, August 19, 2026.
Topics to learn
Page numbers refer to the linked chapter PDF.
- The formalism p. 1
- Categories p. 1
- The lexicon p. 1
- Combinatory rules p. 2
- Data and ambiguity p. 5
- CCGbank p. 5
- Ambiguity in CCG p. 6
- Parsing p. 7
- Supertagging p. 7
- A* parsing p. 8
- Heuristic functions p. 9
- A worked example p. 10
FAQ
Where does most of the grammar live in CCG?
In the lexical categories. A word can have a complex category that says what arguments it takes and in which direction. The combinatory rules are a smaller shared set.
What is a supertag?
A lexical category assignment. Supertagging chooses likely categories before the full parse, which cuts the search. A wrong supertag can remove the correct derivation.
What is A* searching for?
A derivation whose cost is optimal under the chosen score, using a heuristic to avoid expanding every partial result. The heuristic has to be admissible for the optimality claim to hold.
Does one CCG derivation mean one constituent tree?
Not automatically. CCG has its own derivations, and CCGbank relates them to treebank structure. Do not treat a CCG step and a context-free rule as the same object.