Towards Corpus-Grounded Agentic LLMs for Multilingual Grammatical Analysis

TL;DR

This paper presents a novel agentic LLM framework that leverages reasoning over annotated linguistic corpora to automate and interpret multilingual grammatical analysis, demonstrating promising results across diverse languages and features.

Contribution

It introduces an innovative agentic framework integrating reasoning, code generation, and data analysis for corpus-grounded grammatical tasks in multiple languages.

Findings

Achieved high accuracy in dominant word-order predictions

Provided comprehensive coverage of word-order features across 170+ languages

Demonstrated the system's ability to generalize and quantify linguistic variations

Abstract

Empirical grammar research has become increasingly data-driven, but the systematic analysis of annotated corpora still requires substantial methodological and technical effort. We explore how agentic large language models (LLMs) can streamline this process by reasoning over annotated corpora and producing interpretable, data-grounded answers to linguistic questions. We introduce an agentic framework for corpus-grounded grammatical analysis that integrates concepts such as natural-language task interpretation, code generation, and data-driven reasoning. As a proof of concept, we apply it to Universal Dependencies (UD) corpora, testing it on multilingual grammatical tasks inspired by the World Atlas of Language Structures (WALS). The evaluation spans 13 word-order features and over 170 languages, assessing system performance across three complementary dimensions - dominant-order accuracy, order-coverage completeness, and distributional fidelity - which reflect how well the system generalizes, identifies, and quantifies word-order variations. The results demonstrate the feasibility of combining LLM reasoning with structured linguistic data, offering a first step toward interpretable, scalable automation of corpus-based grammatical inquiry.