Computer Aided Translation (CAT) helps human translators complete translation work efficiently and with high quality, organically combining machine translation and human translation. According to the 2019 China Language Services Industry Development Report, \"The ability to reduce translation costs is the main aspect of translation technology recognized by language services freelancers. The surveyed language service freelancers recognized 'the use of translation technology can reduce translation costs' by 80%, 'the use of translation technology can improve translation quality' by 67%, and 'the use of translation technology's use can improve translation skills' has an approval rating of 58%.\" At present, the vast majority of translation tools mainly used by most people are some type of machine translation, such as: electronic dictionaries, cell phone translation software, webpage translation and professional language translation software, and the percentage of people using paper dictionaries is getting smaller and smaller. This proves that translation technology has an important role in promoting human translation. Computer technology improves the efficiency of manual translation, frees translators from repetitive labor and provides translators with convenience. Human translation makes up for the shortcomings of machine translation, and human translation plays the role of human subject, and serves as a supplement in areas where machine translation is not mature enough. For example: obscure and difficult-to-understand literary texts, local languages, literary texts that need artistic processing, creative translation, etc. Machine translation tends to be accurate based on the data generated by a large number of human translators. The principle lies in the fact that \"corpus technology is a core concept of computer-aided translation, and the quality and size of the corpus are closely related to the quality of the translation.\"<\/p>\n\n\n\n
Needless to say, machine translation is an automated cross-linguistic conversion activity, and its translation speed and the scale of text translated at one time are far beyond the reach of human translation. However, machine translation is also a kind of mechanical two-dimensional imitation activity, the individuality and creativity presented by its translated text is far less than that of human translation, and it cannot translate the interpersonal meaning and pragmatic connotation of the source language text. Analyzing the principle of machine translation and its similarities and differences with human translation, it is not difficult to see that the relationship between machine translation and human translation is not contradictory and zero-sum, but complementary and mutually reinforcing.
First of all, machine translation is based on human translation. Leaving human \ntranslation, machine translation can not begin. Throughout the development of machine \ntranslation, it is not difficult to see that the reason for the successful implementation of \nmachine translation is that it has an important reason for having a bilingual aligned corpus. \nFor one thing, machine translation needs to use bilingual alignment corpus to analyze the \nbilingual corpus including the human translation corpus and depict the bilingual conversion \nrules based on it. Without the human translation corpus, there is no way to talk about the \nimportant material prerequisite for the implementation of machine translation, i.e., the \nbilingual alignment corpus. Secondly, the quality of machine translation depends largely on \nthe quality of the human-translated text included in the bilingual alignment corpus. After all, \nthe source language vocabulary or the target language counterparts or corresponding \nstructures of the utterances used in machine translation come directly from the human \ntranslations. Third, the translation strategies and methods adopted by machine translation \nare summarized and generalized on the basis of analyzing the manually translated corpus.
Secondly, machine translation can assist in solving the difficulties encountered by \nhuman translators. As we all know, the translation of abstract nouns, specialized terms and \nphrases is a major difficulty for human translation. These words usually belong to specific \nspecialized fields and are neither easy to understand nor easy to remember. Translators often \nface many difficulties in translating these terms. However, by utilizing the automation \nadvantage of machine translation, the difficulty of translating these terms can be easily \nsolved in a very short time.
In addition, the post-translation editing of machine translation and the touch-up of the \ntranslation need to rely on human translation. In recent years, machine translation \ntechnology has made rapid development and has been favored by users for its advantages in \ntranslation speed. However, there are still many problems with the quality of machine \ntranslation, although it has been significantly improved, such as mistranslation and omission \nof vocabulary or phrases, errors in word order and syntactic structure, etc. In addition, the \nmachine translation system can be used to translate the words and phrases by human \ntranslators, but it can be used to translate the words and phrases by human translators. In \naddition, when the machine translation system translates long or compound sentences, it \noften produces translations that are syntactically incoherent and difficult to interpret. For \nthis reason, to improve the quality of machine translation, in addition to improving and \noptimizing the machine translation software, it is also necessary to carry out human \ntranslation, post-translation editing and embellishment of translations, and correct the \ntranslation errors of machine translation. Obviously, machine translation supplemented by \nhuman translation can not only greatly improve the translation speed, reduce the intensity of \nhuman translation, but also significantly improve the quality of machine translation. It must \nbe pointed out that post-translation editing of machine translation can help us to understand \nthe correspondence between source language and target language more comprehensively \nand profoundly on the basis of analyzing the regular features of machine translation errors, \nso as to further improve the formal descriptions of bilingual conversion rules or translation \nmodels. These descriptions can be directly applied to improve the performance of the \nmachine translation system, and thus improve the quality of machine translation.
Finally, machine translation and human translation can carry out a reasonable division \nof labor and undertake translation tasks at different levels or of different natures. As \nmentioned above, machine translation has the characteristics of mechanical and \nsecond-degree imitation, etc. Due to the constraints of these characteristics, machine \ntranslation is often only applicable to the translation of programmed or informative texts in \nspecific fields, such as scientific and technological texts, legal texts, contracts, memorandums, \noperation manuals and so on. Informational texts refer to the texts that describe facts, \nconvey information, knowledge and opinions for the main purpose of communication. These \ntexts have more fixed words, clearer lexical semantics, and higher repetition rate in the \napplication of utterance structure. As Hutchins and Somers pointed out, non-literary texts \nsuch as scientific and technical literature are monotonous and repetitive, but require precise \nand consistent language expression. The translation of such texts is exactly the space where \nmachine translation can play a role. In addition, if the purpose of translation is to give the \nreader or user a general understanding of the basic content of the source text, without the \nneed for a precise and complete translation, machine translation can also be applied. In \ncontrast, texts with a high degree of creativity or expressive texts such as novels, poems and \nessays emphasize the emotional expression of the author or the characters, and the \nlexical-semantic expression is often unstable and vague, and a large number of metaphorical \nexpressions are used. Expressive text refers to the text that focuses on presenting the \nauthor's subjective emotions and is rich in imagination, which is mainly characterized by \nsubjectivity, emotion and imagination. The translation task of these texts can obviously only \nbe undertaken by human translators, but not by machine translators. According to Li \nYingjun's point of view, machine translation is mainly applied to scientific and technological \ntexts, and the application of literary texts is difficult and has a long way to go. As a matter of \nfact, machine translation cannot recognize and reproduce the emotions and attitudes \nembedded in these texts, nor can it effectively deal with the translation of semantic \nmeanings and metaphorical expressions in the source language texts.
It is necessary to point out that some informative texts presenting the results of the \nlatest cutting-edge research, such as academic papers or textbooks, often use a large number \nof completely new terminologies to disseminate the latest scientific ideas and principles. \nThese terms have been introduced for a short period of time and often lack counterparts in \nanother language. Obviously, machine translation is not capable of translating these terms, \nbecause machine translation's correspondence about source and target language words is \nbased on the analysis of the existing human translated corpus. The translation of these \ninformational texts can only be undertaken by human translators. If machines are relied upon \nto accomplish the translation of these texts, it will be difficult to obtain satisfactory \ntranslations.
In view of this, we recognize that machine translation will take up more and more \ntranslation tasks, and will partially replace human translation, but will not completely replace \nhuman translation in the end. The future of the translation profession will be an era of \nhuman-machine coexistence and human-machine complementarity.<\/p>\n\n\n\n
References: [1]Li Yamin and Feng Li. Pre-translation editing and post-translation editing in human-computer cooperative translation[J].Frontier Economy and Culture,2020,(01):101-104. [2]Hu Kaibao and Li Yi.A Study on Machine Translation Features and Their Relationship with Human Translation[J].China Translation,2016,37(05):10-14.<\/p>\n\n\n\n
<\/p>","protected":false},"excerpt":{"rendered":"
\u8ba1\u7b97\u673a\u8f85\u52a9\u7ffb\u8bd1(Computer Aided Tr …<\/p>","protected":false},"author":1,"featured_media":0,"parent":0,"menu_order":0,"comment_status":"closed","ping_status":"closed","template":"","meta":{"footnotes":""},"class_list":["post-408","page","type-page","status-publish","hentry"],"_links":{"self":[{"href":"https:\/\/2024group3.tianyun.li\/en\/wp-json\/wp\/v2\/pages\/408","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/2024group3.tianyun.li\/en\/wp-json\/wp\/v2\/pages"}],"about":[{"href":"https:\/\/2024group3.tianyun.li\/en\/wp-json\/wp\/v2\/types\/page"}],"author":[{"embeddable":true,"href":"https:\/\/2024group3.tianyun.li\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/2024group3.tianyun.li\/en\/wp-json\/wp\/v2\/comments?post=408"}],"version-history":[{"count":6,"href":"https:\/\/2024group3.tianyun.li\/en\/wp-json\/wp\/v2\/pages\/408\/revisions"}],"predecessor-version":[{"id":475,"href":"https:\/\/2024group3.tianyun.li\/en\/wp-json\/wp\/v2\/pages\/408\/revisions\/475"}],"wp:attachment":[{"href":"https:\/\/2024group3.tianyun.li\/en\/wp-json\/wp\/v2\/media?parent=408"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}