Internet has become an essential part of our everyday life. As we follow faraway political developments that may still strongly influence our security or prosperity; or, alternatively, look for profitable investment opportunities in foreign lands – presented in incomprehensible languages, we increasing turn to online automated translation to somehow understand what it is all about. With all the current deficiencies of automatic translation this are the only translation tools that are easily available and allow us quickly to decide what is worth more thorough digging and possibly employing professional translators. The quality of even free online translators is permanently improving and one can imagine time when many human translators lose jobs as the bulk of translation will be done by machines. Indeed, machine translation as a practical sub-field of computational linguistics – which studies ways to use translation software for translating texts or speech from one language to another – is rapidly developing area of applied research.
Still, no matter how much the automatic translation technology will evolve, it is still many years away before translation process will become fully reliable and adequate, not requiring human intervention to check for misinterpretations and other errors (called post-editing). For years to come high-quality translation will remain at best a hybrid process in which human professionals revise and correct texts generated by translation software, as while machine translation’s advantages are low cost and high speed, it lacks context interpretation ability characteristic of human interpreters, necessary as the meaning of any text is strongly dependent on cultural and situational contexts.
Initially, machine translation was based on a simple substitution of words in one language for words in another, although even this is not an easy task due to such phenomena as polysemy, homonymy, metaphors and idioms. Naturally, such primitive approach was unable to produce good translation, as the textual meaning depends on how the whole phrases and their closest counterparts are arranged. Currently this problem is being partially resolved by applying data and statistical techniques from corpus linguistics, producing on average much better translations, even enabling translation of idiomatic forms and metaphors. Machine translation uses gathered corpus data to produce a statistical equivalent of what the best translation may be, still leaving a lot of potential for error.
Naturally, this is not yet enough, as anyone who has used online translation tools such as Google Translate or any other virtual translator – should have encountered absurd cases of machine translation – at best incorrect or funny, at worst, misleading or offensive. These tools are as yet only good for casual translation with an aim to grasp a general idea of a text, and not suitable for full-fledged documents. They are more satisfactory in case of standardized (official or legal) documents, but totally inadequate in case of translating a conversation or a literary text. Hardly anyone can afford having major mistakes to happen when it comes to translation for diplomatic, legal or business purposes. Hence, for a long while highly professional translators and interpreters will certainly remain in high demand.
Machine translation, automatic translation, automated translation – is the computer based technology using digital computer systems and translation software for automatically translating texts from one natural language into another. The first practical ideas of using digital computers for translation of natural languages was first proposed in Britain and US after the end of the World War II (respectively by A. D. Booth, Warren Weaver and others), with lead taken by the Birkbeck College in London, Georgetown University in Washington, DC, and the Massachusetts Institute of Technology, followed by work in the USSR and a few other countries. Currently machine translation is a rapidly developing area of applied studies, although the potential of automated translation and the possibility of achieving fully automatic machine translation of high quality have been questioned since its conception.
Translation tools – are any of systems that actually performs automatic translation, or assist and support human translators with language translation by offering possible solutions – the latter called computer-assisted translation or computer-aided translation.
Translation software – these are computer programs designed to provide high quality machine translation. At an early stage translation software conducted simple substitution of words in one language for words in another, leading to low quality product due to such phenomena as polysemy, homonymy, metaphors and idioms. Currently these difficulties are at least partially resolved by applying data and statistical techniques from corpus linguistics, producing on average much better translations.
Virtual translator – is a translator system that does not physically exist as such but is based on applying machine translation approaches and translation software to appear to do so:
Online translators – automatic translation tools available on Internet, increasing offered by web brousers, search engines and other globalized Internet resources. In most of cases free online translators are easily available, among the most popular such tools being Google Translate and Bing Translator, along with many others. Often online translation is conducted by specialized plug-ins, software components that add specific translation capacity to existing online resources such as web browsers.
Polysemy, homonymy, metaphor, idiom – Related linguistic phenomena characterized in case of polysemy by multiple meanings (or semes) of a word within certain contiguity of meanings (a semantic field), while in case of homonymy the multiple meanings of a word are as a rule unrelated; in the case of a metaphor, the word(s) may refer to one thing by mentioning some another thing, exploiting some form of analogy or similarity; finally, idiom refers to a fixed expression that in a given context conveys a figurative, rather than literal, meaning.