Description
Phraseology (Cowie 1998; Granger & Meunier 2008; Mitkov 2017; Mel’čuk 2023; Polguère, 2002, 2014; Mejri 2018; Chen 2021), or multiword expressions (MWEs) (Savary 2008; Constant 2012) occupy a central position in linguistic description, language use, and language learning. Idioms, collocations, lexical bundles, and fixed or semi-fixed expressions constitute a significant part of natural language, yet they remain challenging to model, describe, and represent in lexicographic resources (Pecina, P. 2010; Mel’čuk 2011, Polguère 2014, Chen 2025). With the rapid development of new technologies—particularly corpus linguistics, Natural Language Processing (NLP), and Artificial Intelligence (AI)—phraseology and lexicography are currently undergoing a profound methodological and conceptual transformation.
From a modeling perspective, new technologies make it possible to move beyond traditional, intuition-based descriptions of phraseological units. Large-scale corpora (Mitkov 2017), both general and specialized, enable the systematic identification of MWEs through statistical, distributional, and syntactic approaches. Methods such as n-gram extraction, association measures (PMI, t-score), syntactic patterning, and embedding-based similarity allow researchers to capture degrees of fixedness, semantic compositionality, and contextual variability. These advances open new perspectives for representing phraseological knowledge in structured models, including ontologies, lexical networks, and standards such as OntoLex-Lemon (McCrae, Bosque-Gil et al. 2017 ; Bosque-Gil et al. 2019).
In terms of resource creation, technological tools have profoundly reshaped lexicographic practices (Atkins & Rundell 2008; Granger & Paquot 2012). Digital corpora, web-based data, and annotation platforms facilitate the semi-automatic extraction and validation of phraseological units (Mitkov 2017; Evert 2008; Gries 2008; Constant et al. 2017). New-generation lexical resources integrate rich metadata, usage examples, frequency information, and semantic relations, making them more dynamic and interoperable (Polguère 2014; Ci-miano et al. 2016; McCrae et al. 2017).. Moreover, computational approaches enable the development of multilingual and contrastive resources, which are essential for studying phraseological variation across languages and cultures (Paquot 2015; Mel’čuk 2011).
New technologies also play a crucial role in the design of pedagogical dictionaries and lear-ning-oriented resources (Bogaards & van der Kloot 2002; Lew 2012). Phraseology is often a major obstacle for language learners, as MWEs cannot always be interpreted compositionally (Wray 2002; Howarth 1998; Granger 1998). Corpus-based examples, learner-oriented definitions, and adaptive digital interfaces can significantly enhance the accessibility and usability of phraseological information (Sinclair 2004; Granger & Paquot 2015; Paquot 2015). AI-driven tools, such as intelligent tutoring systems or LLM-assisted lexicography, offer promising avenues for generating contextualized examples, usage notes, and difficulty pro-files tailored to learners’ needs (Heift & Schulze 2007; Godwin-Jones 2023; Bender & Koller 2020).
In translation lexicography, the impact of new technologies is equally significant (Hartmann 2007; Atkins & Rundell 2008; Tarp 2012). Phraseological units pose well-known challenges in translation due to their idiomaticity and phraseocultural specificity (Chen 2022a; 2022b). Parallel corpora, alignment tools, and machine translation systems provide valuable data for identifying translation equivalents, partial correspondences, and translation strategies (Chen et al. 2024). Digital translation dictionaries can now incorporate cross-linguistic mappings, semantic annotations, and real attested examples, bridging the gap between lexicographic description and actual translational practice.
Finally, the analysis of existing dictionaries through technological means (Béjoint 2010; Lew 2013) offers new insights into lexicographic traditions and practices. Computational methods allow for the large-scale comparison of dictionary entries, coverage, microstructure, and treatment of phraseological units. Such analyses contribute to a critical understanding of how dictionaries evolve in response to technological, linguistic, and societal changes (Hausmann 1989; Tarp 2012; Fuertes-Olivera & Bergenholtz 2011).
The intersection of phraseology, multi-word expressions, lexicography, and new technologies constitutes a fertile research domain, fostering innovative models, richer resources, and more effective tools for analysis, learning, and translation.