Measuring cross-language intelligibility between Romance languages with computational tools

2026-02-07

AI Summary

This study addresses the quantification of cross-lingual intelligibility among Romance languages—French, Italian, Portuguese, Spanish, and Romanian—with a particular focus on its asymmetry. The work proposes a novel computational metric that integrates orthographic, phonetic, and multi-source semantic embeddings to jointly capture lexical surface and semantic similarity. Leveraging parallel corpora and word embedding models, the resulting intelligibility scores exhibit strong correlation with human performance in cloze-task experiments, effectively capturing and validating the asymmetric nature of mutual intelligibility across these languages. This approach offers a scalable, data-driven paradigm for modeling cross-lingual comprehension, advancing both theoretical understanding and practical applications in multilingual NLP.

Technology Category

• Natural Language Processing: Machine Translation, Multilinguality, Cross-Lingual NLP
• Computer Vision: Language and Vision
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• Web Mining and Content Analysis: Mining multimedia, multimodal, multilingual, cross-lingual Web data
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

We present an analysis of mutual intelligibility in related languages applied for languages in the Romance family. We introduce a novel computational metric for estimating intelligibility based on lexical similarity using surface and semantic similarity of related words, and use it to measure mutual intelligibility for the five main Romance languages (French, Italian, Portuguese, Spanish, and Romanian), and compare results using both the orthographic and phonetic forms of words as well as different parallel corpora and vectorial models of word meaning representation. The obtained intelligibility scores confirm intuitions related to intelligibility asymmetry across languages and significantly correlate with results of cloze tests in human experiments.

Problem

• cross-language intelligibility
• Romance languages
• lexical similarity
• mutual intelligibility
• computational linguistics

Innovation

• computational metric
• mutual intelligibility
• lexical similarity
• semantic similarity
• Romance languages