Iconicity in Large Language Models

TL;DR

This paper investigates how large language models encode lexical iconicity by generating and testing pseudowords in artificial languages, revealing that LLMs can recognize iconicity more effectively than humans and that such encoding differs from natural language processing.

Contribution

It demonstrates that LLMs can generate and recognize iconic pseudowords in artificial languages, highlighting differences from human processing of iconicity.

Findings

Humans better guess meanings of pseudowords in iconic language than in natural languages.

LLMs outperform humans in guessing meanings of iconic pseudowords.

Generated languages show universality and shared cues used by humans and LLMs.

Abstract

Lexical iconicity, a direct relation between a word's meaning and its form, is an important aspect of every natural language, most commonly manifesting through sound-meaning associations. Since Large language models' (LLMs') access to both meaning and sound of text is only mediated (meaning through textual context, sound through written representation, further complicated by tokenization), we might expect that the encoding of iconicity in LLMs would be either insufficient or significantly different from human processing. This study addresses this hypothesis by having GPT-4 generate highly iconic pseudowords in artificial languages. To verify that these words actually carry iconicity, we had their meanings guessed by Czech and German participants (n=672) and subsequently by LLM-based participants (generated by GPT-4 and Claude 3.5 Sonnet). The results revealed that humans can guess the meanings of pseudowords in the generated iconic language more accurately than words in distant natural languages and that LLM-based participants are even more successful than humans in this task. This core finding is accompanied by several additional analyses concerning the universality of the generated language and the cues that both human and LLM-based participants utilize.