3

I am working on theoretical NLP things, in particular to do with combinatory categorial grammar (CCG). I don't have much knowledge of CCG, or of grammar in general.

I was wondering how much of English has been formalised in CCG? For instance, I've seen the following tables in a presentation on CCG. Is there a resource I can consult that has a big list/summary of different parts of speech of English with their corresponding CCG categories? And as a follow up question, how much of English hasn't been formalised in CCG?

Thanks

enter image description here

enter image description here

1
2

Are you aware of CCGbank (Hockenmaier, 2003)? This is the largest-scale corpus of English text annotated with CCG categories, consisting of ~1 million tokens of text, and is derived semi-automatically from the Penn Treebank.

This was used by Clark & Curran (2007) in a line of work to train wide-coverage dependency parsers for English, which made use of the fact that CCGbank covers a broad swathe of English syntax.

While the journal paper linked to above covers a number of the constructions handled, see Section 3 of Julia Hockenmaier's thesis for a complete list of the constructions that receive CCG analyses in the corpus.

It does cover rarer syntax which is nevertheless required for wide-coverage English text, such as conditional inversion ("Had it existed then, ...") which it analyses by assigning a category like ((S/S)/(S[pt]\NP))/NP to the token "had", accounting for the inversion.

Is there a resource I can consult that has a big list/summary of different parts of speech of English with their corresponding CCG categories?

Note that syntax doesn't map cleanly onto parts of speech, but you can easily gather a list of all categories that receive a certain part of speech by running some scripts on CCGbank directly.

This site is temporarily in read only mode and not accepting new answers.

Not the answer you're looking for? Browse other questions tagged .