1、 The Possibility of Machine Translation姓名：曾秋丽 学号：20104478 班级：英语 1001Abstract:With the development of economy, politics, culture, science and language, the possibility of machine translation becomes more and more prosperous. Now there are different brands of translation machine in the market, such as
2、 Yixing, Yaxin,Tongyi and so on, and these translation machines plays an important part in translation between different languages and makes communication more convenient and more available. Based on the big success that translation machine has achieved in the past years, people then want to dig the
3、 potential possibility of machine translation which has much more advantages than people in doing the translation job. As for me, I am sure with the development of computer science, math and language, the possibility of machine translation could be realized in utmost degree.Key words: Machine transl
4、ation, computer linguistics, origin, development, problem, future 摘要：随着经济，政治，文化，科学和语言的发展，机器翻译的可能性越来越来被人们所看好。如今市场上有各种各样的翻译机器，例如译星，雅信，通译等等，这些翻译机器在不同语言的翻译过程中扮演了重要的角色，使得交流变得更加便利可行。基于过去翻译及其所取得的巨大成功，如今人们想要发掘机器翻译的潜在的可能性，而机器翻译与人工翻译相比具有更多的优点。就我本身而言，我相信随着计算机科学，数学和语言学的发展，机器翻译的可能性能在最大程度上得到发掘。关键词：机器翻译，计算机语言学，起源，
5、发展，问题，前景Introduction: When speaks of machine translation, we must talk about the Computational Linguistics and Natural Language Processing. Linguistics is not the only field concerned with language, there are many other areas such as law, sociology, psychology and many other fields that are also pre
6、occupied with language. And among all these fields is the Computational Linguistics, which means the use of computers to process or produce human language. Computer science provides theories and techniques for designing computer systems. One of the current application areas is machine translation (t
7、ranslating from one language to another). Machine translation is called Automatic Translation too, it refers to the process that translating the natural language to the target language, it is a branch of Natural Language Processing, and it connects very much to the Computational Linguistics and Natu
8、ral Language Understanding.Machine translation is based on the linguistics, math and computer science. Linguists provides words formation and grammar rules that computer can understand, mathematicians edit them into code, computer scientists provide software and hardware equipment and operate the pr
9、ocedure. Only these there areas work together, can machine translation be achieved. And the quality of machine translation is totally depending on these three factors.The origin and the development of machine translation: The history of machine translation can date back to the early thirties in the
10、last century. A France scientist presents the idea of machine translation, and then in 1933 a Soviet Union scientist designed a translation machine, but owing to the poor science techniques, his machine wasnt a success. In 1946 as the first modern computer ENIAC was invented, scientists widened the
11、areas that computer could work with, so the theory of machine translation was presented completely for the first time. In 1954, with the support of IBM Company, Georgetown University finished the machine translation experiment between English and Russian firstly, and this experiment also shows the g
12、reat potential of machine translation. From fifties to the early sixties in the 20th century, the research of machine translation presents a prosperous scene, and out of economy, politics and military reasons both of The USA and the Soviet Union supported a lot of foundation in machine translation.
13、But later in 1966 the Automatic Language Processing Advisory Committee announce the impossibility of machine translation, and deny the possibility of machine translation, and stop the research of machine translation. Then in the 1970s, with the development of science and technology and the prosper o
14、f communication between different countries make the language translation an urgency, but the traditional ways of translation can not meet peoples needs, so machine translation was brought into consideration again, at the same time, computer science has achieved great success especially the improvem
15、ent of hardware and the use of artificial brainpower in natural language processing motivates the research of machine translation, and a lot of practical translation machine such as Weinder and EURPOTRA and TAUM-METEO were designed. And now with the general use of internet, unify of the world econom
16、y and the rapid increase in international communication, translation machine is used in almost every area and every time. The process and the divide system of machine translation: The process of machine translation can be divided into three parts. Fist is analyzing the original language; this proces
17、s doesnt need to consider the different factors of the target language when analyzing the natural language. Second is the transition of the translated language, this process needs to consider the target languages factors without consider the natural language. Third is the production of the translate
18、d language. During this process, needs to consider both of the natural language and the target language.Machine translation system is developed as the development of corpus linguistics. The system of machine translation can be divided into two main parts. First is the rule-based system. This system
19、needs words formation and grammar rules forms its storage. It is mainly concentrate on grammar rules and the rules that form a sentence and solve the problem of form and logic integrity. It is like looking up to dictionary to find corresponding words and then establish their inner connection and fin
20、ally produce the corresponding target language. Second is corpus-based system, this system doesnt need words formation and grammar rules, its based on statistics regular patterns. The theory of this system is that the translation between natural language and target language is a problem of probabili
21、ty. Any of the target language can be the translation of the natural language, the difference is the probability, and task of the translation machine is to find the target language that has the maximal probability. And most of the machine translation systems that are using are rule-based system in t
22、he world. The most concerned problem of machine translation:Most people have a misunderstanding towards machine translation; they believe that machine translation has a totally different transition between natural language and target language, and then will lead to a false understanding of language
23、and cause some problem when people communicate with each other, so some people think that machine translation can do nothing and can not solve any problems. when they translate language, they can use linguistics knowledge, and then find the corresponding target language, that is how they works, but
24、sometime as some changes take place in grammar, rules, words formation and so on, it is unavoidable to make some mistakes during the translation process. In fact, the error is unavoidable, for as I just mentioned in how the machine translation system works, we can know machine is not human, when the
25、y translate language, they can only understand the superficial meaning rather than the real meaning that the speakers mean. Such as when people say some sentence, even translation machine can translate the meaning of the language, but it cannot express the feelings or emotions that the speakers real
26、ly want to express. Different cultures have different traditions and customs, they have specific sentence to express the same meaning, the words formation is huge, and people cannot put all the grammar rules and words into computer coding system. Moreover we all know that the language is developed a
27、ll the time, some new words and expressions can appear every time, and the old is disappeared all the time, although these changes take place, however we cannot put into the put them into use in machine translation. For example, Chinese language is multiple. Though we encourage using mandarin, but p
28、eople use dialect more often, different place have different dialects, such as we have Cantonese in Guangdong, Sichuan dialects in Sichuan, even in the same province, people may not understand each other owing to the accents and different tradition. It is difficult to understand Chinese for Chinese
29、people, even though for machine. Now the number 2 is popular through the internet, it is not only means number, but it can also refer to a person who acts silly or looks stupid, some young people could understand this very easily , however for some old people, they only consider 2 as a number.The fu
30、ture of machine translation:In fact, the most concerned problem is the quality of the translated language; it is far more satisfying for the eventual goal. It is not only connected with the machine translation system, but also the language itself. Since it is impossible to solve the languages proble
31、ms, the only way is to improve the translation system. We cannot stop researching how to improve the machine translation system, its advantages are more important than its disadvantages in the future. I also believe we may not make machine translation perfect, but with the development of computer sc
32、ience, computer linguistics, math and logic, we can make the most profits trough machine translation.附注：1 http:/ 胡壮麟。语言学教程。北京：北京大学出版社（第四版）Z，2011Title：The Possibility of Machine TranslationName: 曾秋丽Class: 英语 1001Student Number：20104478School of Foreign Languages and CulturesSouthwest University of Science and Technology
