A Transparent Model of Syntactic and Semantic Cue-based Retrieval

Shisen Yue, John T. Hale


Abstract
Human comprehenders have greater difficulty forming pairwise grammatical dependencies in cases where the earlier word competes with a "distractor" to which it is similar. Cue-based retrieval theories (see e.g., Lewis et al., 2006) address this "interference" phenomenon with explicit quantifications of memory retrieval difficulty. We propose a computational model, consistent with Cue-based retrieval, that separately quantifies two different kinds of similarity. A linear combination of the two reproduces the graded interference pattern reported in Van Dyke (2007). This simple account offers a more straightforward mechanistic interpretation than Attention-based predictors from opaque Transformer based models.
Anthology ID:
2026.scil-main.38
Volume:
Proceedings of the Society for Computation in Linguistics 2026
Month:
July
Year:
2026
Address:
San Diego, CA
Editors:
Rob Voigt, Alex Warstadt, Naomi Feldman, Tal Linzen
Venues:
SCiL | WS
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
411–422
Language:
URL:
https://preview.aclanthology.org/ingest-acl-workshops/2026.scil-main.38/
DOI:
Bibkey:
Cite (ACL):
Shisen Yue and John T. Hale. 2026. A Transparent Model of Syntactic and Semantic Cue-based Retrieval. In Proceedings of the Society for Computation in Linguistics 2026, pages 411–422, San Diego, CA. Association for Computational Linguistics.
Cite (Informal):
A Transparent Model of Syntactic and Semantic Cue-based Retrieval (Yue & Hale, SCiL 2026)
Copy Citation:
PDF:
https://preview.aclanthology.org/ingest-acl-workshops/2026.scil-main.38.pdf