aiqbench analysis
Scaling self-supervised speech models from 1,000 to 4,000 languages triggers a non-linear shift where the system begins resolving deep phylogenetic lineages and long-term linguistic contact previously obscured by surface similarities. This qualitative improvement enables the identification of a robust Pacific macro-cluster driven by shared acoustic signatures rather than genealogical proximity, suggesting that massive models internalize complex layers of language history. The findings offer a new computational approach for computational phylogenetics and the study of language contact dynamics.
Scaling Self-Supervised Speech Models Uncovers Deep Linguistic Relationships: Evidence from the Pacific Cluster
arXiv:2603.07238v2 Announce Type: replace Abstract: Similarities between language representations derived from Self-Supervised Speech Models (S3Ms) have been observed to primarily reflect geographic proximity or surface typological similarities driven by recent expansion or contact, potentially missing deeper genealogical signals. We investigate how scaling an S3M-based language identification system from 126 to 4,017 languages reshapes this topology, and find a non-linear effect: phylogenetic recovery stays flat up to the 1K scale, but the 4K model undergoes a qualitative shift, resolving both clear lineages and long-term linguistic contact. Most strikingly, a robust Pacific macro-cluster emerges, grouping genealogically unrelated Papuan, Oceanic, and Australian languages, and we trace its driver to a concentrated encoding that captures shared acoustic signatures such as global energy dynamics. These results suggest that massive S3Ms internalize multiple layers of language history, offering a promising perspective for computational phylogenetics and the study of language contact.
Abstract:Similarities between language representations derived from Self-Supervised Speech Models (S3Ms) have been observed to primarily reflect geographic proximity or surface typological similarities driven by recent expansion or contact, potentially missing deeper genealogical signals. We investigate how scaling an S3M-based language identification system from 126 to 4,017 languages reshapes this topology, and find a non-linear effect: phylogenetic recovery stays flat up to the 1K scale, but the 4K model undergoes a qualitative shift, resolving both clear lineages and long-term linguistic contact. Most strikingly, a robust Pacific macro-cluster emerges, grouping genealogically unrelated Papuan, Oceanic, and Australian languages, and we trace its driver to a concentrated encoding that captures shared acoustic signatures such as global energy dynamics. These results suggest that massive S3Ms internalize multiple layers of language history, offering a promising perspective for computational phylogenetics and the study of language contact.
Computation and Language (cs.CL); Audio and Speech Processing (eess.AS)
arXiv:2603.07238 [cs.CL]
(or arXiv:2603.07238v2 [cs.CL] for this version)
https://doi.org/10.48550/arXiv.2603.07238 Focus to learn more arXiv-issued DOI via DataCite
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 · 237 words