Morphological Analysis Can Improve a CCG Parser for English

Because English is a low morphology language, current statistical parsers tend to ignore morphology and accept some level of redundancy. This paper investigates how costly such redundancy is for a lexicalised grammar such as CCG. We use morphological analysis to split verb inflectional suffixes into separate tokens, so that they can receive their own lexical categories. We find that this improves accuracy when the splits are based on correct POS tags, but that errors in gold standard or automatically assigned POS tags are costly for the system. This shows that the parser can benefit from morphological analysis, so long as the analysis is correct.