Departamento de
Traducción e Interpretación


Tema:   Automática.
Autor:   Martins, Ronaldo Teixeira
Año:   2004
Título:   A nova liìngua do imperador [Tne new language of the emperor]
Lugar:   Campinas (São Paulo)
Editorial/Revista:   Universidade de Campinas (Unicamp)
Páginas:   309
Idioma:   Portugués
Tipo:   Tesis.
Disponibilidad:   Acceso abierto
Resumen:   This work addresses, in a linguistic-philosophical basis, the Universal Networking Language (UNL), a knowledge representation language that has been used, under many initiatives coordinated by the UNDL Foundation, for the development of information retrieval and machine translation systems. The text analyzes the status and the structure of UNL, whose expressions constitute hypergraphs, where nodes, referred to as Universal Words, represent concepts, and arcs between nodes, the so-called Relation Labels, describe oriented relations between pairs of concepts. Additionally, nodes and hypernodes can be annotated by attributes, the Attribute Labels, which introduce contextual information. It is claimed that, although formally equivalent to many other semantic networks, UNL comprises a numbe r of distinctive features that bring it to be more expressive and more suited to occupy the place of a pivot language in interlingua-based machine translation systems. It is stressed, among the distinctive features of UNL, its autonomy, comprehensiveness, transparency and functionality. From the point of view of autonomy, UNL is said to be an intensional, analytic, self-referenced language, with no empirical significance, whose object is the UNL Knowledge Base (UNL KB), which constitutes the repository where every relation holding between any pair of concepts would be registered. This self-sufficiency instrumentalizes UNL to the computational processing, and it does not prevent it from the ability of interpreting other external languages, because of the comprehensiveness of the UNL KB, which would comprise the sum of every possible (referable) world. This reference, in the case of UNL, would be made by means of a non-ambiguous syntax, which would hinder any opacity and assure precision to the translation process. Finally, UNL would be also functional, since its computational configuration would not sacrifice the readability, which would be preserved by mnemonic reference to the vocabulary of English. Despite all of this, many problems have been reported concerning the use of UNL in machine translation systems. Provided that it elects the sentence as the unity for representation, UNL under-represents, for instance, phenomena concerning the textual level, whose representation is essential for the translation process. The main restrictions, however, concern the dialectalization of UNL and the very scope of translation. It has been noticed that the signs of UNL has been differently used and interpreted, restricting the alleged autonomy of the translation process between UNL and natural languages. Finally, it has been observed that UNL operates with a conception of translation which has proved to be mainly naif, leading to a rather fantasy of translation, the same that has prevented machine translation systems from constituting an actual alternative to the human translator. [Source: Author]
Agradecimientos:   Record supplied by Katia Aily Franco de Camargo – (Universidade Federal do Rio Grande do Norte – UFRN).
2001-2021 Universidad de Alicante DOI: 10.14198/bitra
Comentarios o sugerencias
La versión española de esta página es obra de Javier Franco
Nueva búsqueda
European Society for Translation Studies Ministerio de Educación Ivitra : Institut Virtual Internacional de Traducció asociación ibérica de estudios de traducción e interpretación