# Neural Unsupervised Reconstruction of Protolanguage Word Forms

arXiv: https://arxiv.org/abs/2211.08684 · HTML: https://mlsanity.com/paper/2211.08684/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- Builds on 2 of its 24 references that are in the map.
- First posted on arXiv 16 Nov 2022.

## Builds on (most cited first, 2 of 2)

- [Ab Antiquo: Neural Proto-language Reconstruction](https://mlsanity.com/paper/1908.02477/) (2019)
- [Morphological Inflection Generation with Hard Monotonic Attention](https://mlsanity.com/paper/1611.01487/) (2016)

## Built on by (most cited first, 5 of 5)

- [Jambu: A historical linguistic database for South Asian languages](https://mlsanity.com/paper/2306.02514/) (2023)
- [Semisupervised Neural Proto-Language Reconstruction](https://mlsanity.com/paper/2406.05930/) (2024)
- [Transformed Protoform Reconstruction](https://mlsanity.com/paper/2307.01896/) (2023)
- [Phonetic Reconstruction of the Consonant System of Middle Chinese via Mixed Integer Optimization](https://mlsanity.com/paper/2502.04625/) (2025)
- [Unsupervised Protoform Reconstruction through Parsimonious Rule-guided Heuristics and Evolutionary Search](https://mlsanity.com/paper/2506.10614/) (2025)

## Abstract (arXiv)

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.