Are the LLMs Capable of Maintaining at Least the Language Genus?

TL;DR

This paper investigates whether Large Language Models (LLMs) are sensitive to linguistic genealogical structures and how training data influences their multilingual capabilities, revealing genus-level effects conditioned by resource availability.

Contribution

It extends analysis on multilingual behavior of LLMs by examining sensitivity to linguistic genera and the impact of training data imbalances.

Findings

Genus-level effects are present in LLMs.

Training resource availability strongly influences multilingual performance.

Distinct multilingual strategies vary across LLM families.

Abstract

Large Language Models (LLMs) display notable variation in multilingual behavior, yet the role of genealogical language structure in shaping this variation remains underexplored. In this paper, we investigate whether LLMs exhibit sensitivity to linguistic genera by extending prior analyses on the MultiQ dataset. We first check if models prefer to switch to genealogically related languages when prompt language fidelity is not maintained. Next, we investigate whether knowledge consistency is better preserved within than across genera. We show that genus-level effects are present but strongly conditioned by training resource availability. We further observe distinct multilingual strategies across LLMs families. Our findings suggest that LLMs encode aspects of genus-level structure, but training data imbalances remain the primary factor shaping their multilingual performance.