Enregistré dans:
Détails bibliographiques
Auteurs principaux: Jeong, Jinhong, Lee, Sunghyun, Lee, Jaeyoung, Han, Seonah, Yu, Youngjae
Format: Preprint
Publié: 2025
Sujets:
Accès en ligne:https://arxiv.org/abs/2511.10045
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866915873236713472
author Jeong, Jinhong
Lee, Sunghyun
Lee, Jaeyoung
Han, Seonah
Yu, Youngjae
author_facet Jeong, Jinhong
Lee, Sunghyun
Lee, Jaeyoung
Han, Seonah
Yu, Youngjae
contents Sound symbolism is a linguistic concept that refers to non-arbitrary associations between phonetic forms and their meanings. We suggest that this can be a compelling probe into how Multimodal Large Language Models (MLLMs) interpret auditory information in human languages. We investigate MLLMs' performance on phonetic iconicity across textual (orthographic and IPA) and auditory forms of inputs with up to 25 semantic dimensions (e.g., sharp vs. round), observing models' layer-wise information processing by measuring phoneme-level attention fraction scores. To this end, we present LEX-ICON, an extensive mimetic word dataset consisting of 8,052 words from four natural languages (English, French, Japanese, and Korean) and 2,930 systematically constructed pseudo-words, annotated with semantic features applied across both text and audio modalities. Our key findings demonstrate (1) MLLMs' phonetic intuitions that align with existing linguistic research across multiple semantic dimensions and (2) phonosemantic attention patterns that highlight models' focus on iconic phonemes. These results bridge domains of artificial intelligence and cognitive linguistics, providing the first large-scale, quantitative analyses of phonetic iconicity in terms of MLLMs' interpretability.
format Preprint
id arxiv_https___arxiv_org_abs_2511_10045
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Do Language Models Associate Sound with Meaning? A Multimodal Study of Sound Symbolism
Jeong, Jinhong
Lee, Sunghyun
Lee, Jaeyoung
Han, Seonah
Yu, Youngjae
Computation and Language
Sound symbolism is a linguistic concept that refers to non-arbitrary associations between phonetic forms and their meanings. We suggest that this can be a compelling probe into how Multimodal Large Language Models (MLLMs) interpret auditory information in human languages. We investigate MLLMs' performance on phonetic iconicity across textual (orthographic and IPA) and auditory forms of inputs with up to 25 semantic dimensions (e.g., sharp vs. round), observing models' layer-wise information processing by measuring phoneme-level attention fraction scores. To this end, we present LEX-ICON, an extensive mimetic word dataset consisting of 8,052 words from four natural languages (English, French, Japanese, and Korean) and 2,930 systematically constructed pseudo-words, annotated with semantic features applied across both text and audio modalities. Our key findings demonstrate (1) MLLMs' phonetic intuitions that align with existing linguistic research across multiple semantic dimensions and (2) phonosemantic attention patterns that highlight models' focus on iconic phonemes. These results bridge domains of artificial intelligence and cognitive linguistics, providing the first large-scale, quantitative analyses of phonetic iconicity in terms of MLLMs' interpretability.
title Do Language Models Associate Sound with Meaning? A Multimodal Study of Sound Symbolism
topic Computation and Language
url https://arxiv.org/abs/2511.10045