CNNs that robustly compute vowel harmony do not explicitly represent phonological tiers
University of Massachusetts (UMass) Amherst · 14.06.2025
DOI: 10.7275/scil.3189 ↗Kurz gefasst
Fachartikel in der Zeitschrift University of Massachusetts (UMass) Amherst, erschienen am 14.06.2025. Verfasst von Jane Li und 1 weiteren Autor:innen.
Abstract (Originalsprache)
Linguistic and model-theoretic analyses of long-distance phonology postulate the existence of phonological tiers (Goldsmith, 1976; Heinz et al., 2011). For example, vowel harmony may be analyzed as a process that projects vowels (but not consonants) onto a tier and ensures that all sounds on the tier have the same feature (e.g., [±front] in Turkish vowel harmony, Clements et al. (1982)). Li and Zhou (under review) recently demonstrated that convolutional neural networks (CNNs) learning a toy example of vowel harmony (§2) on short strings robustly generalize the pattern to much longer strings. One explanation is that these CNNs have independently recovered an “algorithm” that is consistent with the tier projection analysis. Alternatively, these models may have uncovered an approximation of this system, or an entirely different system that robustly generalizes to long lengths. This work investigates these hypotheses via various interpretability methods. In particular, we search for evidence for a “strong” implementation of tier projection, in which these CNNs exactly implement the tier-projection and feature-matching analyses described above.
Schlagwörter
Zitieren
APA
Li, J., & Zhou, A. T. (2025). CNNs that robustly compute vowel harmony do not explicitly represent phonological tiers. University of Massachusetts (UMass) Amherst. https://doi.org/10.7275/scil.3189
Vancouver
Li J, Zhou AT. CNNs that robustly compute vowel harmony do not explicitly represent phonological tiers. University of Massachusetts (UMass) Amherst. 2025. doi:10.7275/scil.3189
BibTeX
@article{Li2025,
author = {Li, Jane and Zhou, Alan Tiantian},
title = {CNNs that robustly compute vowel harmony do not explicitly represent phonological tiers},
journal = {University of Massachusetts (UMass) Amherst},
year = {2025},
doi = {10.7275/scil.3189}
}
EUVIMED