Module 10: Translation Today
  Lecture 38: Machine Translation
 



Impact of TM

Translation tools help the translator, but how is it received? CAT is used predominantly for scientific and technical translation, and so the mechanical nature of finding ‘matches’ in the stored memory is successful. However, it is successful only to a certain extent. Even when TM is able to come up with exact matches for source text segments, the translator will have to consider the text as a comprehensive whole and judge if the segment will fit. TM is capable of creating what is called a ‘sentence salad’ effect, “when sentences are drawn (without adequate contextual information) from various translation memories created by different translators with different styles”(“Computer-aided Translation”, 50). Sometimes the cohesion and logical flow of arguments in the translation can be compromised for the sake of using TM. In their attempt to find suitable matches for source text segments, translators avoid using certain grammatical constructions. This will yield consistent matches, which is good as far as technical translations are concerned. But the end result might be a text that is monotonous, and lacks in readability and coherence.  

Michael Cronin draws our attention to an aspect of translation that can easily be forgotten in the eagerness to automate the process, which is that translation is not merely a synchronic process, but a diachronic one. He observes: “…students are increasingly adept at using the tools which allow for rapid information retrieval or the quick dispatch of text but deep, historical knowledge of languages and cultures which takes time and effort is not always valued in a culture of informational ubiquity” (21).Computers can provide exact matches for words, but do not have an awareness of the history behind works and the factors that have moulded them. This makes for a thin translation that is word perfect but nothing more.

Cronin quotes Reinhard Schäler to point out that one problem that has been noted in connection with the rise and spread of MT is the translators’ declining concern about the quality of translations. Schäler points out that “accuracy and consistency of the translation (features associated with MT output) take precedence over style, readability and naturalness, all associated with the traditional values and reference system of the human translators” (qtd in Cronin, 22). MT without human intervention can have undesirable consequences.