Reconceptualizing Multiword Expressions in Contemporary Turkic Languages: AI-Supported NLP Approaches

Abstract

Multiword expressions (MWEs) encompass idioms, proverbs, semi-fixed patterns, and fixed syntactic constructions that reflect both linguistic and cultural dimensions. In the context of contemporary Turkic languages, MWEs represent a shared historical heritage while simultaneously revealing the unique semantic worlds of each language. This study approaches MWEs within the framework of Construction Grammar (CxG), evaluating them as constructions defined by form–meaning pairings. In addition, the paper discusses the potential of Natural Language Processing (NLP) and artificial intelligence-based methods–particularly large language models (LLMs)–for the large-scale recognition and analysis of MWEs. By integrating CxG with NLP, the study proposes an interdisciplinary perspective that builds a bridge between linguistic theory and digital language technologies, opening innovative horizons for research on MWEs in contemporary Turkic languages.