Are Mutually Intelligible Languages Easier to Translate?

TL;DR

This paper investigates whether mutual intelligibility between languages reduces the data required for training neural machine translation models, finding a strong correlation between mutual intelligibility and learning efficiency.

Contribution

It introduces the hypothesis that mutual intelligibility inversely affects the data needed for translation models and provides empirical evidence supporting this in Romance languages.

Findings

Strong correlation between mutual intelligibility and translation model learning curves

Mutual intelligibility reduces data requirements for training translation models

Empirical validation on Romance languages

Abstract

Two languages are considered mutually intelligible if their native speakers can communicate with each other, while using their own mother tongue. How does the fact that humans perceive a language pair as mutually intelligible affect the ability to learn a translation model between them? We hypothesize that the amount of data needed to train a neural ma-chine translation model is anti-proportional to the languages' mutual intelligibility. Experiments on the Romance language group reveal that there is indeed strong correlation between the area under a model's learning curve and mutual intelligibility scores obtained by studying human speakers.