22nd Workshop on Multiword Expressions (MWE 2026)

[ https://multiword.org/mwe2026/ | https://multiword.org/mwe2026/ ]

Organized, sponsored, and endorsed by SIGLEX, the Special Interest Group on the Lexicon of the ACL, and by [ https://unidive.lisn.upsaclay.fr/ | UniDive ] COST Action CA21167

Half-day workshop collocated with the 19th Conference of the European Chapter of the Association for Computational Linguistics (EACL 2026, [ https://2026.eacl.org/ | https://2026.eacl.org/ ] ), Rabat, Morocco.

*

Workshop date: March 28 , 2026 (9:00 – 12:30) *

Early bird registration deadline: February 24 , 2026 (AOE) *

Information on registration: [ https://2026.eacl.org/registration/ | https://2026.eacl.org/registration/ ]

Multiword expressions (MWEs), i.e., word combinations that exhibit lexical, syntactic, semantic, pragmatic, and/or statistical idiosyncrasies (Baldwin and Kim, 2010), such as “by and large”, “hot dog”, “make a decision” and “break one's leg” are still a pain in the neck for Natural Language Processing (NLP). The notion of MWE encompasses closely related phenomena: idioms, compounds, light-verb constructions, phrasal verbs, rhetorical figures, collocations, institutionalized phrases, etc. Given their irregular nature, MWEs often pose complex problems in linguistic modeling (e.g., annotation), NLP tasks (e.g., parsing), and end-user applications (e.g., natural language understanding and Machine Translation), hence still representing an open issue for computational linguistics (Miletić and Schulte im Walde, 2024; Ramisch et al., 2023; Phelps et al., 2024; Mahajan et al., 2024).

For more than two decades, the topic of modeling and processing MWEs for NLP has been the focus of the MWE workshop, organized by the [ https://multiword.org/ | MWE section ] of [ http://www.siglex.org/ | ACL-SIGLEX ] in conjunction with major NLP conferences since 2003. Impressive progress has been made in the field, but our understanding of MWEs still requires much research, considering their need and usefulness in NLP applications. This is also relevant to domain-specific NLP pipelines that need to tackle terminologies most often realized as MWEs.

Topics

The dominating trend in MWE research presented at the workshop focuses on large language models, particularly examining their capabilities and limitations across diverse tasks. Papers present research on Swedish, Ukrainian, Romanian, Galician, Marathi, Turkish, Korean, Sinhala, and Chinese. Research discusses language-specific phenomena that challenge existing annotation schemes and identification methodologies, particularly in morphologically complex and typologically distinct languages.

The programme of the MWE 2026 Workshop is available here: [ https://multiword.org/mwe2026/ | https://multiword.org/mwe2026/ ]

Co-located Shared tasks

The workshop MWE 2026 will host [ https://unidive.lisn.upsaclay.fr/doku.php?id=other-events:parseme-admire-st-... | two shared tasks ] :

*

PARSEME 2.0, whose objective is to identify and paraphrase MWEs in written text, and *

AdMIRe 2 (Advancing Multimodal Idiomaticity Representation) , which explores the comprehension ability of multimodal models for MWEs in a variety of languages.

Organizing Committee

Verginica Barbu Mititelu, A. Seza Doğruöz, Alexandre Rademaker, Atul Kr. Ojha, Mathieu Constant, Ivelina Stoyanova

Anti-harassment policy

The workshop follows the ACL anti-harassment policy.

Contact For any inquiries regarding the workshop, please send an email to the Organizing Committee at [ mailto:mwe2026workshop@gmail.com | mwe2026workshop@gmail.com ] .