Probing Subphonemes in Morphology Models

TL;DR

This paper investigates how transformer models encode phonological and subphonemic features across multiple languages, revealing their strengths in local features and limitations in long-distance dependencies, informing future morphological modeling strategies.

Contribution

Introduces a language-agnostic probing method to analyze phonological feature encoding in transformers trained on phonemes across diverse languages.

Findings

Local phonological features are well captured in phoneme embeddings.

Long-distance dependencies are better represented in the transformer's encoder.

Insights inform strategies for training more effective morphological models.

Abstract

Transformers have achieved state-of-the-art performance in morphological inflection tasks, yet their ability to generalize across languages and morphological rules remains limited. One possible explanation for this behavior can be the degree to which these models are able to capture implicit phenomena at the phonological and subphonemic levels. We introduce a language-agnostic probing method to investigate phonological feature encoding in transformers trained directly on phonemes, and perform it across seven morphologically diverse languages. We show that phonological features which are local, such as final-obstruent devoicing in Turkish, are captured well in phoneme embeddings, whereas long-distance dependencies like vowel harmony are better represented in the transformer's encoder. Finally, we discuss how these findings inform empirical strategies for training morphological models, particularly regarding the role of subphonemic feature acquisition.