Information-Theoretic Characterization of Vowel Harmony: A Cross-Linguistic Study on Word Lists
This work addresses the challenge of analyzing vowel harmony for linguistic typology, particularly benefiting research on low-resource and under-studied languages, though it is incremental as it builds on prior quantitative methods with a new data approach.
The study tackled the problem of quantifying vowel harmony across languages by developing an information-theoretic measure based on phoneme-level language models trained on lemma word lists, achieving results that captured vowel harmony patterns in relevant languages despite using smaller datasets of up to 1000 entries per language.
We present a cross-linguistic study that aims to quantify vowel harmony using data-driven computational modeling. Concretely, we define an information-theoretic measure of harmonicity based on the predictability of vowels in a natural language lexicon, which we estimate using phoneme-level language models (PLMs). Prior quantitative studies have relied heavily on inflected word-forms in the analysis of vowel harmony. We instead train our models using cross-linguistically comparable lemma forms with little or no inflection, which enables us to cover more under-studied languages. Training data for our PLMs consists of word lists with a maximum of 1000 entries per language. Despite the fact that the data we employ are substantially smaller than previously used corpora, our experiments demonstrate the neural PLMs capture vowel harmony patterns in a set of languages that exhibit this phenomenon. Our work also demonstrates that word lists are a valuable resource for typological research, and offers new possibilities for future studies on low-resource, under-studied languages.