On the Intelligibility of Romance Language Varieties: Spanish and Portuguese in Europe and America
Abstract
AbstractMutual intelligibility within language families presents a significant challenge for multilingual NLP, particularly due to the prevalence of dialectal variation and asymmetric comprehension. In this paper, we present a corpus-based computational analysis to quantify linguistic proximity across Romance language variants, with a focus on major Spanish (Argentine, Chilean and European) and Portuguese (Brazilian and European) varieties and the other main Romance languages (Italian, French, Romanian). We apply a computational metric of lexical intelligibility based on surface and semantic similarity of related words to measure mutual intelligibility for the five main Romance languages in relation to the Spanish and Portuguese varieties studied.