Why do language models perform worse for morphologically complex languages?

Result code in IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216208%3A11320%2F26%3AIBFTT2E4" target="_blank" >RIV/00216208:11320/26:IBFTT2E4 - isvavai.cz</a>

Result on the web

—

DOI - Digital Object Identifier

—

Result language

angličtina

Original language name

Why do language models perform worse for morphologically complex languages?

Original language description

Language models perform differently across languages. It has been previously suggested that morphological typology may explain some of this variability (Cotterell et al., 2018). We replicate previous analyses and find additional new evidence for a performance gap between agglutinative and fusional languages, where fusional languages, such as English, tend to have better language modeling performance than morphologically more complex languages like Turkish. We then propose and test three possible causes for this performance gap: morphological alignment of tokenizers, tokenization quality, and disparities in dataset sizes and measurement. To test the morphological alignment hypothesis, we present MorphScore, a tokenizer evaluation metric, and supporting datasets for 22 languages. We find some evidence that tokenization quality explains the performance gap, but none for the role of morphological alignment. Instead we find that the performance gap is most reduced when training datasets are of equivalent size across language types, but only when scaled according to the so-called “byte-premium”-the different encoding efficiencies of different languages and orthographies. These results suggest that languages of particular morphological types are not intrinsically advantaged or disadvantaged in language modeling. Differences in performance can be attributed to disparities in dataset size. These findings bear on ongoing efforts to improve performance for low-performing and under-resourced languages. © 2025 Association for Computational Linguistics.

Czech name

—

Czech description

—

Type

D - Article in proceedings

CEP classification

—

OECD FORD branch

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Project

—

Continuities

—

Publication year

2025

Confidentiality

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Article name in the collection

Proc. Main Conf. Int. Conf. Comput. Linguist., COLING

ISBN

979-8-89176-196-4

ISSN

—

e-ISSN

—

Number of pages

17

Pages from-to

6607-6623

Publisher name

Association for Computational Linguistics (ACL)

Place of publication

—

Event location

Abu Dhabi

Event date

Jan 1, 2026

Type of event by nationality

WRD - Celosvětová akce

UT code for WoS article

—