CPLM, a Parallel Corpus for Mexican Languages: Development and Interface

Proceedings of the 12th Conference on Language Resources and Evaluation (LREC 2020), pages 2947–2952 Marseille, 11–16 May 2020 c⃝ European Language Resources Association (ELRA), licensed under CC-BY-NC 2947 CPLM, a Parallel Corpus for Mexican Languages: Development and Interface Gerardo Sierra, Cynthia Montaño, Gemma Bel-Enguix, Diego Córdova, Margarita Mota Instituto de Ingenieria Universidad Nacional Autónoma de México Ciudad de México {gsierram,cmontanor,gbele}@iingen.unam.mx diego.cordova.nieto@gmail.com, mmotam@iingen.unam.mx Abstract Mexico is a Spanish speaking country that has a great language diversity, with 68 linguistic groups and 364 varieties. As they face a lack of representation in education, government, public services and media, they present high levels of endangerment. Due to the lack of data available on social media and the internet, few technologies have been developed for these languages. To analyze different linguistic phenomena in the country, the Language Engineering Group developed the Corpus Paralelo de Lenguas Mexicanas (CPLM) [The Mexican Languages Parallel Corpus], a collaborative parallel corpus for the low-resourced languages of Mexico. The CPLM aligns Spanish with six indigenous languages: Maya, Ch’ol, Mazatec, Mixtec, Otomi, and Nahuatl. First, this paper describes the process of building the CPLM: text searching, digitalization and alignment process. Furthermore, we present some diffi- culties regarding dialectal and orthographic variations. Second, we present the interface and types of searching as well as the use of filters. Keywords: Parallel corpus, multilingual corpus, low-resourced languages of Mexico 1. Introduction Collecting corpora for endangered languages is crucial to preserve and expand language diversity, besides the appli- cations for NLP (Amel Fraisse and Fishkin, 2018). Mexico is a linguistically diverse country. There are 11 phyla (e.g. Uto-Aztec) that comprise 68 linguistic groups (e.g. Nahu- atl, belonging to the Uto-Aztec) and 364 varieties (Nahuatl from the Huasteca) present all around the country (INALI, 2008). About 6.5% of the population speaks at least one of these varieties, but according to the intercensal survey (IN- EGI, 2015), about 21% of the total population of Mexico identifies themselves as members of an indigenous group. However, these languages are not represented in education, government, public services and media and, therefore, they show high levels of endangerment. Languages facing this lack of large amount of data are called low-resourced and all linguistic varieties in Mexico are struggling with this situation. One of the current approaches for tackling the scarcity of data is the building of parallel corpora of Span- ish, that has a large amount of data, and the low-resourced languages of Mexico. There are three examples of online parallel corpora re- garding Mexican languages. The first one is Axolotl1 (Gutiérrez-Vasques et al., 2016), a Nahuatl-Spanish parallel corpus also developed by the Language Engineering Group. There is the parallel corpus Wixarika-Spanish2, that gathers translations of classic Hans Christian Andersen’s literature (Mager et al., 2018a). And Tsu¯nkua3, an Otomi-Spanish parallel corpus developed by Comunidad Elotl. The importance of the creation of parallel corpora is linked to the analysis of different phenomena that can be found in indigenous languages of Mexico that are not usually com- mon among the traditional ones used …

Content truncated to 500 words