Neural Unsupervised Reconstruction of Protolanguage Word Forms

TL;DR

This paper introduces a neural method for unsupervised reconstruction of ancient word forms, improving accuracy over previous approaches by capturing complex phonological and morphological changes.

Contribution

It extends classical expectation-maximization methods with neural models that incorporate monotonic alignment constraints and controlled underfitting.

Findings

Reduced edit distance in Latin reconstruction

Effective modeling of complex phonological changes

Improved accuracy over prior methods

Abstract

We present a state-of-the-art neural approach to the unsupervised reconstruction of ancient word forms. Previous work in this domain used expectation-maximization to predict simple phonological changes between ancient word forms and their cognates in modern languages. We extend this work with neural models that can capture more complicated phonological and morphological changes. At the same time, we preserve the inductive biases from classical methods by building monotonic alignment constraints into the model and deliberately underfitting during the maximization step. We evaluate our performance on the task of reconstructing Latin from a dataset of cognates across five Romance languages, achieving a notable reduction in edit distance from the target word forms compared to previous methods.