Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2601.07220 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911579560214528 |
|---|---|
| author | Shani, Chen Reif, Yuval Roll, Nathan Jurafsky, Dan Shutova, Ekaterina |
| author_facet | Shani, Chen Reif, Yuval Roll, Nathan Jurafsky, Dan Shutova, Ekaterina |
| contents | Multilingual language models (LMs) promise broader NLP access, yet current systems deliver uneven performance across the world's languages. This survey examines why these gaps persist and whether they reflect intrinsic linguistic difficulty or modeling artifacts. We organize the literature around two questions: do linguistic disparities arise from representation and allocation choices (e.g., tokenization, encoding, data exposure, parameter sharing) rather than inherent complexity; and which design choices mitigate inequities across typologically diverse languages. We review linguistic features, such as orthography, morphology, lexical diversity, syntax, information density, and typological distance, linking each to concrete modeling mechanisms. Gaps often shrink when segmentation, encoding, and data exposure are normalized, suggesting much apparent difficulty stems from current modeling choices. We synthesize these insights into design recommendations for tokenization, sampling, architectures, and evaluation to support more balanced multilingual LMs. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_07220 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | The Roots of Performance Disparity in Multilingual Language Models: Intrinsic Modeling Difficulty or Design Choices? Shani, Chen Reif, Yuval Roll, Nathan Jurafsky, Dan Shutova, Ekaterina Computation and Language Multilingual language models (LMs) promise broader NLP access, yet current systems deliver uneven performance across the world's languages. This survey examines why these gaps persist and whether they reflect intrinsic linguistic difficulty or modeling artifacts. We organize the literature around two questions: do linguistic disparities arise from representation and allocation choices (e.g., tokenization, encoding, data exposure, parameter sharing) rather than inherent complexity; and which design choices mitigate inequities across typologically diverse languages. We review linguistic features, such as orthography, morphology, lexical diversity, syntax, information density, and typological distance, linking each to concrete modeling mechanisms. Gaps often shrink when segmentation, encoding, and data exposure are normalized, suggesting much apparent difficulty stems from current modeling choices. We synthesize these insights into design recommendations for tokenization, sampling, architectures, and evaluation to support more balanced multilingual LMs. |
| title | The Roots of Performance Disparity in Multilingual Language Models: Intrinsic Modeling Difficulty or Design Choices? |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2601.07220 |