Structured Knowledge for Low-Resource Languages : The Latin and Ancient Greek Dependency Treebanks

We describe here our work in creating treebanks – large collections of syntactically annotated data – for Latin and Ancient Greek. While the treebanks themselves present important datasets for traditional research in philology and linguistics, the layers of structured knowledge they contain (including disambiguated lemma, morphological, and syntactic information for every word) help offset the comparatively small size of extant Greek and Latin texts for text mining applications. We describe two such uses for these Classical treebanks – discovering lexical knowledge from a large corpus with the help of a small treebank, and identifying patterns of text reuse.