This study by Shojaei, R., (2026) investigates the complex lexical accent system of Tokyo Japanese, focusing on the diverse accentuation patterns induced by suffixes. The research challenges existing theoretical frameworks, such as Transderivational Antifaithfulness and head dominance, by proposing a novel representational account that integrates autosegmental representations with Gradient Symbolic Representations (GSR). This approach allows for a more comprehensive understanding of suffix-induced accent patterns, which include dominant, recessive, preaccenting, subtractive, and attractive suffixes.

Utilizing a detailed analysis of suffix behavior, the study reveals how different accentual classes interact with root accents to produce a range of outcomes. For instance, dominant suffixes can erase preexisting accents, while preaccenting suffixes can assign new accents to preceding syllables. The proposed model captures these dynamics through competition among tones and tone-bearing units (TBUs), allowing for a unified explanation of accentuation that previous models could not achieve.

The findings hold significant theoretical implications for phonological theory, particularly in understanding accentual variation and its representation. Practically, this research could inform computational models of Japanese phonology, enhancing applications in natural language processing and machine translation by providing a more nuanced framework for accent prediction and representation.

Source: glossa-journal.org