If you ask just about anyone—even someone with no linguistic training—what makes a translation good, most people will tell you that it has to be accurate. But what does accurate mean? Accuracy, on the other hand, has to do with the similarity of meaning. Surprisingly, while most people can identify that accuracy is important in translation, very few understand what it is. That’s because accuracy gets easily confused with literalness, even though they mean different things. Literalness has to do with the degree of similarity between linguistic forms (e.g. words and grammar).
The conceptual approach to the translation phenomenon is viewed as a deep integration of national cultures, and their interactions. Literary translation should be considered in the context of literary interaction as a part of multi-ethnic factor. Translation Studies in Kazakhstan has had many directions and common issues of prose, poetry and drama, the specifics of the translation process, and the place of translation studies in multicultural literary process has become the subject of translation studies. Automatic translators like Google Translate are great for quick, one-off translation in casual conversation. But Google Translate not only sometimes chooses the wrong translation of the several possible for a word, but it's not very good at putting the words together. In other words, it's prone to botching the grammar in a sentence. Literary translation schools reflect the evolution of transferability categories and contain modern concept of communicative equivalence of the original and the translated texts as a norm of translation accuracy. Modern communicative approach to translation is due to the facts of cross-language communication and translation dominants. Expansion of the original and the translated text communicative equivalence should be tolerant to the type of the receiving audience. The problem of interlinear translation was the object of translators’ attention for a long time.
Something always gets lost in translation. That’s what IKEA found out when a Reddit user slipped its “Gosa Raps” pillow into Google Translate and got back “Cuddle Rapes.” Now that Google Translate works in 50 languages offline for Android phones (which makes it sound like a great travel app), it seemed like a perfect time to test what works, and what doesn’t. Spoiler alert: Proper nouns, beware. And f you think common colloquialisms won’t pop up when you’re traveling or need a translation, think about how often you’re looking for a “cool” restaurant – how likely are you phrase this as a “popular with fashionable people” restaurant? We’re willing to bet not all that often. We decided to send the following snippet from the New York Times to a group of translators working in French, Spanish, and Mandarin. The bit is a challenge to Google Translate because of the various forms of verbs, proper nouns, and language that’s idiomatically American. Machine translation, sometimes referred to by the abbreviation MT (not to be confused with computer-aided translation, machine-aided human translation (MAHT) or interactive translation) is a sub-field of computational linguistics that investigates the use of software to translate text or speech from one language to another.
Here’s a humorous example that illustrates the difference quite well. Years ago I invited some Tanzanian friends over for dinner. I put the food out and said, “We’re going to eat ‘Canadian-style,’ so come to the table and just help yourselves.” However, my Tanzanian guests broke out into an awkward mix of laughter and horror. That’s because “help yourself” translated literally into Swahili has the same meaning as “relieve yourself” in English! So, yes, my translation was literal. But accurate? No way! A much better translation would have been for me to tell my guests, “serve yourselves.” All of this was stated in Swahili, but unfortunately, as a novice speaker, I translated it literally (i.e. word-for-word). When we talk about accuracy in translating God’s Word, we’re talking about meaning and the rule is: nothing should be added, deleted or changed. But it can be difficult to see how this gets applied if you’re only looking at the words. A good translation will, on the surface, look very different from its source text. That’s because meaning emerges out of a larger context than just single words or phrases. The translator must consider that readers bring a whole set of assumptions to the text. Now you may see no problem with what I said.
On a basic level, MT performs simple substitution of words in one language for words in another, but that alone usually cannot produce a good translation of a text because recognition of whole phrases and their closest counterparts in the target language is needed. Current machine translation software often allows for customization by domain or profession (such as weather reports), improving output by limiting the scope of allowable substitutions. This technique is particularly effective in domains where formal or formulaic language is used. It follows that machine translation of government and legal documents more readily produces usable output than conversation or less standardised text. Solving this problem with corpus statistical, and neural techniques is a rapidly growing field that is leading to better translations, handling differences in linguistic typology, translation of idioms, and the isolation of anomalies.
没有评论:
发表评论