A Dual-Task Paradigm to Investigate Sentence Comprehension Strategies in Language Models

Researchers at the 64th Annual Meeting of the Association for Computational Linguistics introduced a dual-task paradigm combining arithmetic and sentence comprehension to study how language models allocate cognitive resources. Experiments with GPT-4o, o3-mini, and o4-mini showed that under dual-task conditions, models shifted toward plausibility-based comprehension, mirroring human rational inference. The findings suggest that constraints on memory and processing resources promote human-like sentence comprehension in language models.

Abstract Language models LMs behave more like humans when their cognitive resources are restricted, particularly in predicting sentence processing costs such as reading times. However, it remains unclear whether such constraints similarly affect sentence comprehension strategies, and existing methods do not directly target the balance between memory storage and sentence processing, which is central to human working memory. To address this issue, we propose a dual-task paradigm that combines an arithmetic computation task with a sentence comprehension task, such as "The 2 cocktail + blended 3 =...". Our experiments show that under dual-task conditions, GPT-4o, o3-mini, and o4-mini shift toward plausibility-based comprehension, mirroring humans’ rational inference. Specifically, these models show a greater accuracy gap between plausible sentences e.g., "The cocktail was blended by the bartender" and implausible sentences e.g., "The bartender was blended by the cocktail" in the dual-task condition compared to the single-task conditions. These findings suggest that constraints on the balance between memory and processing resources promote rational inference in LMs. More broadly, they support the view that human-like sentence comprehension fundamentally arises from the allocation of limited cognitive resources.- Anthology ID: - 2026.acl-long.552 - Volume: Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics Volume 1: Long Papers /volumes/2026.acl-long/ - Month: - July - Year: - 2026 - Address: - San Diego, California, United States - Editors: Maria Liakata /people/maria-liakata/ , Viviane P. Moreira /people/viviane-p-moreira/unverified/ , Jiajun Zhang /people/jiajun-zhang/unverified/ , David Jurgens /people/david-jurgens/ - Venue: ACL /venues/acl/ - SIG: - Publisher: - Association for Computational Linguistics - Note: - Pages: - 12065–12084 - Language: - URL: https://aclanthology.org/2026.acl-long.552/ https://aclanthology.org/2026.acl-long.552/ - DOI: - Cite ACL : - Rei Emura and Saku Sugawara. 2026. A Dual-Task Paradigm to Investigate Sentence Comprehension Strategies in Language Models https://aclanthology.org/2026.acl-long.552/ . In Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics Volume 1: Long Papers , pages 12065–12084, San Diego, California, United States. Association for Computational Linguistics. - Cite Informal : A Dual-Task Paradigm to Investigate Sentence Comprehension Strategies in Language Models https://aclanthology.org/2026.acl-long.552/ Emura & Sugawara, ACL 2026 - PDF: https://aclanthology.org/2026.acl-long.552.pdf https://aclanthology.org/2026.acl-long.552.pdf