Aesop’s fable “The North Wind and the Sun” Used as a Rosetta Stone to Extract and Map Spoken Words in Under-resourced Languages

Proceedings of the 13th Conference on Language Resources and Evaluation (LREC 2022), pages 2072–2079 Marseille, 20-25 June 2022 © European Language Resources Association (ELRA), licensed under CC-BY-NC-4.0 2072 Aesop’s Fable “The North Wind and the Sun” Used as a Rosetta Stone to Extract and Map Spoken Words in Under-resourced Languages Elena Knyazeva, Philippe Boula de Mareüil, Frédéric Vernier Université Paris-Saclay, CNRS, LISN, Orsay, France {knyazeva, mareuil, frederic.vernier}@limsi.fr Abstract This paper describes a method of semi-automatic word spotting in minority languages, from one and the same Aesop fable “The North Wind and the Sun” translated in Romance languages/dialects from Hexagonal (i.e. Metropolitan) France and languages from French Polynesia. The first task consisted of finding out how a dozen words such as “wind” and “sun” were translated in over 200 versions collected in the field — taking advantage of orthographic similarity, word position and context. Occurrences of the translations were then extracted from the phone-aligned recordings. The results were judged accurate in 96–97% of cases, both on the development corpus and a test set of unseen data. Corrected alignments were then mapped and basemaps were drawn to make various linguistic phenomena immediately visible. The paper exemplifies how regular expressions may be used for this purpose. The final result, which takes the form of an online speaking atlas (enriching the https://atlas.limsi.fr website), enables us to illustrate lexical, morphological or phonetic variation. Keywords: semi-automatic spoken word alignment, linguistic atlas, under-resourced languages 1. Introduction Since the Rosetta Stone, which enabled Champollion to decipher Egyptian hieroglyphs, parallel text translations have made it possible to construct concordances, for the Bible especially. The Bible, with its many translations, has also been used for developing text-to-speech synthesis systems in 700 languages (Black, 2019; Boito et al., 2020). The second most translated book in the world, Le Petit Prince, has been translated in over 400 languages or dialects, including Gallo-Romance varieties (Quint, 2022). Another story, “The North Wind and the Sun”, one of Aesop’s fables, has been used by the International Phonetic Association, for more than a century, to illustrate a number of languages and dialects — the difference between languages and dialects being ill-defined. On the base of this Aesop fable, a speaking atlas of the regional languages of France has been designed (Boula de Mareüil et al., 2018), before being extended to the French Overseas (Boula de Mareüil et al., 2019a, 2019b) and various European countries. The online linguistic atlas, available at <https://atlas.limsi.fr>, allows visitors to hear and read this 1-munute story in over 800 versions, in minority languages or dialects. Through interactive maps, the aim of this atlas (which received over 800,000 visits) is to enhance the visibility of linguistic diversity among the general public and to provide an open access base for comparative studies. The automatic processing of such audio data, collected in the field in low-resourced varieties, is not straightforward, as exemplified by earlier projects involving recordings collected in Africa (Adda et al., 2016), Asia (Michaud et al., 2018) or Europe (Knyazeva et al., 2020). …

Content truncated to 500 words