Machine Translation of polysemic words: current technology in light of Cognitive Linguistics

In view of the constant improvement of machine translation technologies today and their use by students and additional language learners, this paper aims to investigate how free online Machine Translators (MTs) work with polysemic words undoing the ambiguity of meanings in the light of the theoretic...

Descripción completa

Detalles Bibliográficos
Autores: Borsatti, Débora Ache, Santorum, Karen Andresa Teixeira, Costa, Alan Ricardo
Tipo de recurso: artículo
Estado:Versión publicada
Fecha de publicación:2022
País:Brasil
Institución:Universidade Federal do Ceará (UFC)
Repositorio:Entrepalavras
Idioma:portugués
OAI Identifier:oai:ojs.localhost:article/2173
Acceso en línea:http://www.entrepalavras.ufc.br/revista/index.php/Revista/article/view/2173
Access Level:acceso abierto
Palabra clave:Machine translation. Polysemy. Cognitive Linguistics. Lexical semantics. Homonymy.
Tradução automática. Polissemia. Linguística Cognitiva. Semântica lexical. Homonímia.
Descripción
Sumario:In view of the constant improvement of machine translation technologies today and their use by students and additional language learners, this paper aims to investigate how free online Machine Translators (MTs) work with polysemic words undoing the ambiguity of meanings in the light of the theoretical approach of Cognitive Linguistics (CL). The study of the MTs functioning shows that these digital technologies have become increasingly sophisticated due to corpus-based systems and, more recently, the neural system. However, the MT systems still have limitations related to linguistic issues, among them, ambiguity, which can be described as one of the biggest challenges for MT nowadays. In this qualitative study, we analyzed comparatively the results of Portuguese-English automatic translations of polysemic words inserted in sentences with different contexts of use, comparing four free MTs available on the internet: (1) Google Translate, (2) Reverso, (3) Collis and (4) Wordlingo. The results indicate the importance of specifying the context for word disambiguation in machine translation.