The semanticity of catalan words: quantitative linguistics in the era of large language models

Abstract

The emergence of Large Language Models (LLMs) like ChatGPT has significantly transformed both theoretical and applied linguistics, raising a profound debate within linguistics. These models, such as GPT (Generative Pre-trained Transformer) series, have revolutionized the way linguists approach language analysis and comprehension. In contrast to traditional Quantitative Linguistics (QL) and conventional linguistic laws like Zipf's laws (Zipf, 1949), LLMs leverage massive datasets to generate linguistic patterns, syntactic structures, and semantic nuances in a comprehensive manner. In light of this, we recently introduced a novel quantitative concept called ”semanticity” which establishes a connection between a word’s potential meanings and its position within the linguistic network. To explore this notion, we conduct a comprehensive analysis of Catalan using extensive written corpora, leveraging the resources of the official dictionary (DIEC2). Our findings reveal that the semanticity of words provides a straightforward and quantitative classification for content and function words and for other word types in Catalan, allowing for the integration of both their semantic and syntactic attributes into this single quantitative parameter.