Do Language Models Associate Sound with Meaning? A Multimodal Study of Sound Symbolism

Sound symbolism is a linguistic concept that refers to nonarbitrary 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. Code - https://github.com/jjhsnail0822/sound-symbolism * These authors contributed equally. † Corresponding author. J.J. introduced sound symbolism and performed the semantic dimension prediction. S.L. conducted the internal attention analysis. J.L. introduced the idea of formulating the problem as multimodal interpretability. S.H. built the constructed word data and interpreted the linguistic implications.