aiqbench analysis
This study demonstrates that Large Language Models can independently uncover systematic semantic structures inherent in individual letters, revealing a multidimensional profile recoverable from text across diverse languages. The findings confirm that these letter-based associations are robust properties of the alphabet itself rather than artifacts of specific word contexts, offering new insights into how AI models encode linguistic primitives.
Evidence for systematic semantic structure in individual letters
arXiv:2603.17306v4 Announce Type: replace Abstract: Associations between speech sounds and meaning are well documented but have not been systematically mapped over a whole alphabet. Here we map them across the 26 English letters and find that each carries a structured, multidimensional semantic profile that is recoverable from text, perceived across languages, and predicted by articulatory features. Three large language models independently detected consistent semantic structure across nine perceptual dimensions in 220 pairwise letter contrasts, and the profiles they recovered were then tested in preregistered experiments with 1,388 human participants. Native English readers chose the predicted word above chance (85.3%, selected items; 65.7%, all contrasts), and the preference followed the letter rather than the words that carried it. Listeners of five typologically diverse languages showed the same preference (76.7%, selected pairs; 68.4%, randomly drawn contrasts), regardless of their English proficiency. Articulatory features assigned to each letter predicted both the model profiles and the human judgments. Letter-meaning association is thus a systematic, multidimensional property of the alphabet rather than a set of isolated effects.
arXiv:2603.17306 (cs)
Abstract:Associations between speech sounds and meaning are well documented but have not been systematically mapped over a whole alphabet. Here we map them across the 26 English letters and find that each carries a structured, multidimensional semantic profile that is recoverable from text, perceived across languages, and predicted by articulatory features. Three large language models independently detected consistent semantic structure across nine perceptual dimensions in 220 pairwise letter contrasts, and the profiles they recovered were then tested in preregistered experiments with 1,388 human participants. Native English readers chose the predicted word above chance (85.3%, selected items; 65.7%, all contrasts), and the preference followed the letter rather than the words that carried it. Listeners of five typologically diverse languages showed the same preference (76.7%, selected pairs; 68.4%, randomly drawn contrasts), regardless of their English proficiency. Articulatory features assigned to each letter predicted both the model profiles and the human judgments. Letter-meaning association is thus a systematic, multidimensional property of the alphabet rather than a set of isolated effects.
38 pages, 4 main figures. SI Appendix (6 figures, 5 tables) included
Computation and Language (cs.CL); Neurons and Cognition (q-bio.NC)
arXiv:2603.17306 [cs.CL]
(or arXiv:2603.17306v4 [cs.CL] for this version)
From: Gexin Zhao Dr. [view email] [v1] Wed, 18 Mar 2026 03:02:10 UTC (2,066 KB) [v2] Thu, 19 Mar 2026 23:27:05 UTC (9,859 KB) [v3] Fri, 29 May 2026 15:14:51 UTC (25,952 KB) [v4] Sun, 20 Sep 2026 16:35:02 UTC (14,510 KB)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 311 words
Related coverage
Links #5: 2026/07
LessWrong · Jul 31, 2026
The Continual Constitutional Convention: How language models revise the documents that govern them
AI Alignment Forum · Oct 13, 2026
Not Deciphering The Voynich Manuscript
LessWrong · Oct 11, 2026
Game Design Creativity evals
LessWrong · Oct 12, 2026