Typologically-Informed Candidate Reranking for LLM-based Translation into Low-Resource Languages

TL;DR

This paper introduces a typologically-informed reranking framework that improves translation quality for low-resource languages by leveraging linguistic typology without needing parallel data or retraining of large language models.

Contribution

It presents a novel framework combining linguistic typology profiles with reranking techniques to enhance LLM translation into typologically divergent low-resource languages.

Findings

Intervention rates correlate with typological distance from English.

Achieves 48.16% precision for conservative languages.

Operates without parallel training data.

Abstract

Large language models trained predominantly on high-resource languages exhibit systematic biases toward dominant typological patterns, leading to structural non-conformance when translating into typologically divergent low-resource languages. We present a framework that leverages linguistic typology to improve translation quality without parallel training data or model retraining. The framework consists of two components: the Universal Metalinguistic Framework (UMF), which represents languages as structured profiles across 16 typological dimensions with divergence-weighted scoring, and the Computational Engine, which operates through linguistic disambiguation during generation and typological compliance scoring during selection. Evaluation across nine language pairs demonstrates intervention rates strongly correlating with typological distance from English. In experiments on 341 English sentences each having different morphological and syntactic phenomena, the framework shows an intervention precision of 48.16% for conservatively treated languages, 28.15% for morphologically dense languages, and 86.26% for structurally profiled languages. The framework requires no parallel training data and operates with any LLM capable of producing multiple candidate outputs, enabling practical deployment for under-resourced languages.