Translation Fundamentals

Why translation and localization?

Translation is as basic a human activity as creating originals. It’s a natural trait of mind to express anything noticed new in the language one follows naturally. This has inspired the human beings to translate almost ever since the beginning of human civilization. Before the advent of modern technology, translated word was the only way to know that, in fact, there exists such a vast and diverse world beyond what we could see.

The world is as much indebted to the amateur translators as to the great scientists and inventors, revolutionary thinkers, prophets and social reformers. But for the translators’ efforts, the inventions, the thoughts, literature and mythology, and the revolutionary ideas that changed the world could not have traveled so far and wide.

The Challenge

In contrast to translation of literature and mythology, translation in technical domain has its own typical challenges and methodologies. It has developed a specialized process of migrating to other languages that we call localization. Localization is a broad sphere of activities of which translation is a major component. Technical adaptation with due respect to cultural sensitivities are the basic tenets of localization. This calls for a perfect blend of technical and linguistics skills.

Web and telecommunications are new and evolving domains. They are spreading at a speed that no other domains have achieved so far. Their user base is growing by leaps and bounds. Unlike other technologies, the web and telecommunications did not have time to mature first in the native land and language and then migrate to other lands and languages. Their technical maturity and widening of base have been simultaneous. This is the challenge for the translators: to be faithful to the domain and yet be welcome to the common user.

Almost every other day some new features are added to enhance mobile experience. Engineers and enthusiasts are working round the clock. Mobile manufacturers are in a cut throat competition to score over one another. The convergence of different technologies and domains has blurred the divide among them. So the translators now have to translate keeping this convergence in mind. This calls for a cross domain consistency.

With the evolution of Website localization as a specialized offshoot of translation, highly technical domains such web and telecommunications have developed industry specific standard terminology. This is a great help as well as a challenge for the translators. It’s a challenge because at times you have to accept a terminology which you do not find welcome from the language point of view. This underlines a fundamental rule of translation: consistency.

Translation always requires an active human involvement to come out in a welcome shape. Machines would always have a limited and supporting role in translation because they can only perform in a defined way. They cannot think. And here in comes fundamental 1.

Think

Think before you start. And think what? What is the domain? What is the text? What is the context? Who are the audience? To think clearly towards coming out with a good piece of translation, you need to read. This is fundamental 2.

Read

Read the text in total, not in isolation from the whole. This is a thumb rule. This would give you a feel of the content in its totality. At times, you get the actual message conveyed only at the end. Also read some existing domain translation to know what not to do. We should learn from others’ mistakes. It would be better if you just read the translation, not the source. This is very important. A reading would give you a feel whether this is an original text or a translation. General reading is immensely helpful. The more you read and know the better for you.

Post translation reading is equally important. Again read it in total, feel the flow and consistency of tone, style and narration. If you feel that it sounds like a translation and not an original, then the translation has failed the purpose and you have learned an important lesson: a translation has to sound like an original. This is fundamental 3.
Be close to original

The success of a translation is measured on how the audience accepts it in the target language. It must sound as if it is written in the target language itself. It must be as close to the original as possible in size, flow, and narration. An oversized translation with an entirely different narration is not a faithful translation. This underlines the need to be simple and precise, which is fundamental 4.

Be simple and precise

There is always a need and scope in translation to be simple and precise. The very purpose of translation is defeated if the message is not conveyed to the reader in the target language. You would never remember or fall back upon a complex document for guidance. We tend to use heavy terms while translating. We have to resist this and opt for the simpler terms wherever possible.

Golden Rules of Translation

Follow the source: A major challenge in translation is to reflect the tone, style, and structure of the source. This is simply addressed by respecting the source. It’s always safe and wise to follow the source.

Follow the rules: You are always correct if you follow the rules of language and grammar. A grammatically correct sentence is rarely a wrong translation. Being grammatically and linguistically correct is like speaking the truth. You don’t need to remember what you said the last time. This curtails half of the efforts at the review level. This ensure consistency of language and style.

Maintain Consistency: Consistency is the basic requirement of localization tasks. To a certain extent, it is more important than good quality. So much so that it can be acceptable if a translation is wrong but consistent but it is never acceptable if a translation is correct but inconsistent. And consistency is not restricted just to language, but it is also required in style, terminology, and narration. Beyond that, it extends to cross domain consistency. Ever growing convergence of technologies requires consistency in localization. Similar functionalities and concepts in telecom and software should be translated consistently.

Linguistically, consistency makes things easier for translators and reviewers. This saves efforts while making changes after the review. It helps in saving costs as we do not need to pay for anything that is already translated. It helps in developing logics and base doe machine translations.

Typical Challenges in Translation

To Translate or not

In technical translations, certain things are not to be translated. These are mostly company and product names, trademarks, sound marks and patents. They are not translated because they are known and identified in their original form, color, and shape. For example, Microsoft, as a rule does not translate its name and its product names. But this is not the case with all the companies. Certain companies like Nokia do not mind their brand name being written in target languages. Besides, untranslatable text may have technical significance. One has to be very judicious in making these decisions.

Translate or Transliterate

This is a major challenge for translators, particularly in the technical domain. At times one sees more transliterations than translations. This requires a very sound linguistic judgment, particularly when you do not have specific guidelines on what to translate or transliterate. At times, the client provides clear guidelines on what should be transliterated. This is simple to follow. As of now, there are no rules to guide the translators on this. However, there are generally agreed practices in the localization industry.

We just transliterate company and brand names (Nokia, Motorola), product names (Motorazr), domain specific terminology (modem, software, server, E-Mail, infrared, wireless, broadcast, and headset etc.), technology types (Multimedia) and proper names. Terms like file, folder, profile, call, settings, shortcut, operator, menu, media, gallery, card, video, clip etc. that are widely used in local languages in their adopted form are generally transliterated. However we have to be particular about their correct representation in the target language. We should use correct phonetic sounds to represent correct pronunciations in the target language.

Country names are a special category. Some country names have well adopted and accepted parallels in other languages like the United States of America (Sanyukta Raajy Amerikaa), South Korea (Dakshin Koriyaa), South Africa (Dakshin Afrikaa) in Hindi and other Indian languages. But certain country names with similar nomenclature like Saint Kitts and Nevis, Saint Pierre and Miquelon, and South Sandwich Islands are just transliterated because they are not well adopted or widely used in other languages. Certain country names are simply not translated as their adopted names conflict with other names. A typical example is United Kingdom which has to be transliterated as the exact translation (Sanyukt Raajya) would conflict with that of the US.

Another associated challenge is whether to transliterate the acronyms and how to transliterate them. The agreed practice in this regard is that we transliterate very popular acronyms like PIN and SIM. But when it comes to hard core domain specific acronyms like Encapsulated Post Script (EPS) Files, Push-to-Talk (PTT), File Transfer Protocol (FTP), we transliterate them and follow it up with the acronym in bracket in English. For ex. Push-to-Talk (PTT).

Cultural Challenges

We often come across typical terminology and names (Hangul Hanja, Katalan). It is pardonable if you misspell an unheard name unless it’s not culturally offensive. But a good translator would go and try to find native speaker to find the correct pronunciation of a typical term or name.

These issues are addressed by

1) following a common phonetic convention;
2) searching the web; and
3) speaking to a native.

Another important thing to remember is to respect the cultural sensitivities in translation. Never go ahead with what you find unparliamentary or offensive in your language.