One way to improve dependency parsing scores for low-resource languages is to make use of existing resources from other closely related or otherwise similar languages. In this paper, we look at eleven Uralic target languages (Estonian, Finnish, Hungarian, Karelian, Livvi, Komi Zyrian, Komi Permyak, Moksha, Erzya, North Sámi, and Skolt Sámi) with treebanks of varying sizes and select transfer languages based on geographical, genealogical, and syntactic distances. We focus primarily on the performance of parser models trained on various combinations of geographically proximate and genealogically related transfer languages, in target-trained, zero-shot, and cross-lingual configurations. We find that models trained on combinations of geographically proximate and genealogically related transfer languages reach the highest LAS in most zero-shot models, while our highest-performing cross-lingual models were trained on genealogically related languages. We also find that cross-lingual models outperform zero-shot transfer models. We then select syntactically similar transfer languages for three target languages, and find a slight improvement in the case of Hungarian. We discuss the results and conclude with suggestions for possible future work.

This thesis produces a definition of consonant mutations, a hitherto relatively unexplored phenomenon in typology, using a sample of languages from the Celtic and Uralic languages. It is defined using Canonical Typology. The base of the phenomenon is established as ‘functionalized consonant alternations’, with seven dimensions of variation: conditioning elements>no conditioning elements; sole functional indicator>accompanied by other morphemes; more than two grades>two grades; some effect on surrounding vowels>only affecting the consonant; only leniting>also non-leniting sound changes; word-initial/-final>word-medial placement; and less regular>more regular. It is also argued that mutations’ phonology is less important than their function.