Page 210 [210]
OCR
HUNGARIAN-ENGLISH LINGUISTIC CONTRASTS. A PRACTICAL APPROACH 11.7 PARALLELS BETWEEN LANGUAGE TEACHING AND TRANSLATION STUDIES In the 1950s, both FLT and TS wanted to solve its problems with help from Contrastive Linguistics. Both focused on structural contrasts at the sentence level, and both found that the problems of communication cannot be solved by linguistics alone. As a result, in the 1960s and 1970s both foreign language teaching methodology and translation studies switched to a multi-factor approach. 11.8 MACHINE TRANSLATION In the 1950s, researchers had high hopes of MT. Theidea of MT was based on the underlying assumption that translation was a matter of overcoming linguistic contrasts between language systems. Source language syntactic structures had to be exchanged for TL structures. However, it soon appeared that MT cannot deliver the goods, it cannot replace human translation. Indeed, it could not even approximate the quality of human translation. In the 1960s, in the wake of the ALPAC report", research on MT in the US was abandoned, and it re-started only much later. 11.8.1 Neural machine translation Recent machine translation engines, such as Google Translate and DeepL Translate, make use of artificial neural networks. These networks are trained on large bilingual text corpora, containing millions of translated texts. They are very good at translating non-literary, informative texts, e.g. news items or specialised texts (‘szakszovegek’). These text types contain a lot of formulaic expressions (collocations and lexical phrases) and the same technical terms and register-specific expressions occur in most texts, therefore it is relatively easy for the machine to find the correspondent for a given term or registerspecific collocation. Consider these examples: 101 ALPAC (Automatic Language Processing Advisory Committee) was a committee of seven scientists led by John R. Pierce, established in 1964 by the U. S. Government in order to evaluate the progress in computational linguistics in general and machine translation in particular. Its report, issued in 1966, gained notoriety for being very skeptical of research done in machine translation so far, and emphasizing the need for basic research in computational linguistics; this eventually caused the U. S. Government to reduce its funding of the topic dramatically. (From Wikipedia, the free encyclopedia) + 208 +
