Scaling Self-Supervised Speech Models Uncovers Deep Linguistic Relationships: Evidence from the Pacific Cluster
Minu, Hoirin, David R · 2026
Read Scaling Self-Supervised Speech Models Uncovers Deep Linguistic Relationships: Evidence from the Pacific Cluster by Minu, Hoirin, David R (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-07238.txt.
Description
Kim, Kim and Mortensen test what happens to speech-model language maps as the system grows from a modest set of languages to thousands. Small-scale maps mostly mirror geography and recent contact, but the largest model shifts gear and recovers family lines plus long-term contact effects. A Pacific cluster stands out, grouping Papuan, Oceanic and Australian tongues that share sound patterns despite separate roots. The authors read this as a sign that very large speech models hold layers of language history.
Catalog record
| CREATOR | Minu, Hoirin, David R |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 18868 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.07238