Learning to pronounce as measuring cross-lingual joint orthography-phonology complexity
This work addresses the challenge of cross-lingual pronunciation complexity for linguists and NLP researchers, but it is incremental as it builds on existing methods for language comparison.
The paper tackled the problem of measuring how hard it is to pronounce words from written text across languages by modeling grapheme-to-phoneme transliteration, finding that complexity depends on the mapping expressiveness or simplicity, with results based on training a transformer model on 22 languages.
Machine learning models allow us to compare languages by showing how hard a task in each language might be to learn and perform well on. Following this line of investigation, we explore what makes a language "hard to pronounce" by modelling the task of grapheme-to-phoneme (g2p) transliteration. By training a character-level transformer model on this task across 22 languages and measuring the model's proficiency against its grapheme and phoneme inventories, we show that certain characteristics emerge that separate easier and harder languages with respect to learning to pronounce. Namely the complexity of a language's pronunciation from its orthography is due to the expressive or simplicity of its grapheme-to-phoneme mapping. Further discussion illustrates how future studies should consider relative data sparsity per language to design fairer cross-lingual comparison tasks.