Adversarially Probing Cross-Family Sound Symbolism in 27 Languages

TL;DR

This study demonstrates that phonological forms can predict size semantics across 27 languages, revealing universal sound symbolism biases through computational analysis and adversarial techniques.

Contribution

It provides the first large-scale cross-linguistic computational analysis of sound symbolism, showing phonological cues predict size semantics across unrelated languages.

Findings

Phonological form predicts size semantics above chance across languages.

Both vowels and consonants contribute to size prediction.

Adversarial training suppresses language identity while preserving size signals.

Abstract

The phenomenon of sound symbolism, the non-arbitrary mapping between word sounds and meanings, has long been demonstrated through anecdotal experiments like Bouba Kiki, but rarely tested at scale. We present the first computational cross-linguistic analysis of sound symbolism in the semantic domain of size. We compile a typologically broad dataset of 810 adjectives (27 languages, 30 words each), each phonemically transcribed and validated with native-speaker audio. Using interpretable classifiers over bag-of-segment features, we find that phonological form predicts size semantics above chance even across unrelated languages, with both vowels and consonants contributing. To probe universality beyond genealogy, we train an adversarial scrubber that suppresses language identity while preserving size signal (also at family granularity). Language prediction averaged across languages and settings falls below chance while size prediction remains significantly above chance, indicating cross-family sound-symbolic bias. We release data, code, and diagnostic tools for future large-scale studies of iconicity.