下载掌阅APP，畅读海量书库
立即打开
According to traditional translation theories，translators should hide behind the target language text，make themselves invisible and attempt to eliminate the exertion of their subjectivity.A target language text with no traits of the translator is regarded as a successful one according to the traditional principle of translation.However，with the development of translation studies，we find that the translator is not really invisible，for there is still some information about him in the target language text.Therefore，the translator's subjectivity is an objective reality in the process of translation.Wilss says，“Translating is a mental，multi-factorial activity which cannot exhaustively be investigated within a linguistic framework ignoring the person of the translator”（1982：217）.The translator himself，the process of translation，the target language text and other objective phenomena all prove the existence of the translator's subjectivity.
Subjectivity is a concept that accompanies the concept of subject.Like subject，subjectivity is also a hot topic in the field of philosophy.Then what is subjectivity？
Generally，there are five definitions of subjectivity：subjectivity is human nature；it is equivalent to subjectiveness；its essential character is its practice quality；subjectivity refers to a kind of relationship，i.e.，“for me”in the practical activities of subjects；it means possession of self-determination，a dynamic role and creative power，i.e.，it is a synthesis of independent self-determination and subjective dynamic role （丰子义，孙成叔，王东，1994：42-45）.These five definitions are arranged chronologically，and they actually show the development of the concept towards philosophical truth.Accordingly，the fifth definition is the most comprehensive one.
In short，subjectivity refers to the essential traits of a subject and the characteristics that a subject shows in practical activities.Here we would like to use another scholar's words to sum up the main characteristics of subjectivity：Subjectivity is the unification of practicality and conscious dynamic role，that of subjectiveness and objectivity，and that of the dynamic and the static.Subjectivity lays stress on practicality （王玉樑，1995：38）.
Now let's move to the translator's subjectivity.What is meant by the translator's subjectivity？The translator's subjectivity is reflected in his constant decision-making process.The translator's subjectivity includes two aspects：one is that the translator actuates his knowledge to correctly comprehend the source language text and successfully produce the target language text，which means that during the whole process the translator constantly makes proper choices according to his background knowledge；the other is that the translator carefully considers and conscientiously allows himself to be restricted by some elements related to translation，such as the source language text，the author of the source language text，the target language text readers，the idiomatic usage of the target language，etc.
When they are engaged in a practical activity，human beings are sure to exert their subjective dynamic role in order to carry on the activity successfully.At the same time they have to abide by some objective laws in the world，that is to say，they are subject to some objective elements.These are the ways in which the subjectivity of human beings manifests itself.Translators do translation practice just as human beings engage in other mental activities.It goes without saying that there exists the translator's subjectivity in the process of translation as a humanistic activity.
As a human being，the translator is bound to exhibit his subjectivity in a translation activity.It is impossible for the translator to be a slave of the author.What he translates must be inside the frame of the source language text，but in the meantime it is the result of his choices or even sometimes a kind of distillation.One grows up in a particular environment which is different from that of another in a certain way.For example，some grow up in a relatively open community，while others grow up in a relatively conservative environment.People have different education experiences.For example，some major in liberal arts while others specialize in sciences.What we want to stress here is that each person is a unique entity that is different from others.Translators are by no means exceptions.Since they are different，they have their own norms by which they make choices.If you ask different translators to translate the same work，they are bound to produce different versions.So the subjective elements of translators，such as personality，temperament，psychological gift，field of knowledge，the ability of language application，and the stand and morality of the translator，all have a direct and important effect on the translation practice （许钧，2010）.
The translation activity starts from the translator's selection of the source language text to be translated，and then according to his own aesthetic standards and his own judgment，the translator reaches a thorough comprehension of the whole text.The translator is a reader of the source language text.Each individual reader understands something on the basis of his “pre-understanding”.In translation practice，the “pre-understanding”of the translator is affected by his historical，cultural and social background.Therefore，there will appear differences among the understanding modes of translators and their comprehension of the content of the text.It is no wonder that there is a current saying，“If there are 1，000 readers，there will be 1，000 Hamlets.”With an understanding of the original，the translator will go on to the next stage—reproduction of the original.Of course，if translators understand the original differently，they will express it differently.Even though they share the same comprehension of the source language text，they may not express it entirely in the same way.Take the Chinese character “水”as an instance.Some may render it simply into “water”.But some may render it as “Adam's wine”.There are still some，maybe rarely，who render it as “H 2 O”.Each translator is affected by what he has picked up and acquired in the environment where he grows up.So each translator has his own values and aesthetic standards，hence different target versions of the source language text.The translator's style can be another point to support the existence of the translator's subjectivity.We often hear the famous statement in literary creation “The style is the man”.It is equally true of translation.Zhu Shenghao （朱生豪），a well-known translator of Shakespeare's works，was an excellent example.Influenced by Chinese culture，he adopted the domestication approach during translation.Thus，his versions are fluent or easily readable，and the words employed in his works are very gorgeous.Liang Shiqiu （梁实秋），another scholar who translated Shakespeare's works，served as another example.But his purpose to translate was to arouse the readers' interest in studying the original.In order to serve his purpose，he employed the foreignization strategy.We may find that his translated works are not as fluent as those of Zhu's.Evidently，different translators have different styles.
Different translators with different pre-understanding structures will process the source language text by employing different principles and approaches under various guidelines，and they will inevitably manifest their historical and national backgrounds，aesthetic values，literary competence in their translation，thus making their subjectivity visible.It can be asserted without any exaggeration that the subjectivity of the translator accompanies the whole process of translation.
The translator is in a central position in the process of translation.He guides and manipulates the whole process.Now let's look at his central position from the following aspects.
From the perspective of “geographical”position，there are two ternary relationships.On the one hand，we have “The source language text—The translator—The target language text”，which is acknowledged by translators as the three elements of translation.Clearly，in this ternary relationship the translator stands in the middle and makes great efforts to comprehend the source language text and produce the target language text.On the other hand，we have the ternary relationship of “The author—The translator—The reader”.In the line that is composed of the three human elements，the translator is in a position of fulcrum，who attempts to keep a balance between the other two elements and makes appropriate adjustments where necessary.No wonder many scholars impressively point out that the translator is the center of hermeneutics and the only element of subject.This idea is echoed by Xu Jun，who makes a similar statement，“When we try to define the subjectivity of translation，obviously we should consider the functions of the author and the reader.But who is in the central position？It is the translator”（2003：11）.
In fact，from the source language text to the target language text，the whole translation process is manipulated by the translator.As the translator is the recipient of the source language text as well as the creator of the target language text，he or she is in the central and key position in translation.Therefore，whether a translation is “good”or “bad”is determined by the translator's “feeling”and “thought”（Robinson，1991）.The central position of the translator is unquestionable.The translator plays a crucial role in the process of translation.It is the translator who is able to keep a balance between the source language text elements and the target language text elements.Thus studying translation from the angle of the translator can keep us away from going to the extreme：the theory of the source language text as the center and the theory of the target language text as the center.
From the aspect of human elements in translation，the translator plays an important role of regulating the relationship between different human elements and undertakes the translation practice directly.Although other human elements，such as the author，the reader，and the publisher，have their impact on the translation practice，they are not the manipulators of translation and they don't participate in translation practice directly.Instead，they indirectly influence translation practice through the agent—the translator.Take the author for an instance.Among Western literary circles，some theorists advocate the idea of “death of the author”in order to highlight the function of the reader in deciphering the meaning of the text.And in the process of translation，the translator has to be a reader first.He cudgels his brains to comprehend what the author means by the text before translating it.Thus the author influences the translation practice through a good understanding of the text by the translator.The target reader is another example.When the translator chooses a text to translate，he has to consider the need of the target reader.Anyway，he translates for readers.Thus，the target readers influence the translation practice through the translator's understanding of their needs and interests.Of course，we cannot ignore the effect of those other human elements，but we should make it clear that the translator is the most important element for the success of translation.
The central position of the translator determines that he must exert his subjectivity in order to make his translation activity proceed smoothly.
In reality，we often find different versions of the same original works，especially those classical masterpieces.It is really a common phenomenon that one literary work has different versions in a target language.Here are several concrete examples. Moby Dick ，which is the masterpiece of the American writer Herman Melville，shows the writer's great writing ability and excellent artistic craft.It （ Moby Dick ）has been translated into many languages.In China we have more than five versions （文军，曲艺，2003：63）. Tao Te Ching （《道德经》）is one of the most important philosophical and literary works China has offered to the world.That it ranks only next to Holy Bible among the world's most important translated works attests to its importance and popularity.“As far as I know，as a philosophical masterpiece Tao Te Ching has at least more than 250 versions in different languages，among which there are 45 English versions”（苗玲玲，2002：137）.In fact，many famous works have been translated by different scholars so that different versions of the same works have come out.For instance， Jane Eyre ， Rouge et Noir and most of Shakespeare's works have all been rendered by different translators and published by different publishers.Thus，retranslation has become a common practice.
No matter what attitudes people hold towards retranslation，here it stays.It exists in our social life and will exist unceasingly in the future.Language is historical，and people's understanding is historical，so it is necessary to retranslate those masterpieces in order to catch up with the times and meet the needs and tastes of people of different times.Each time a masterpiece is retranslated，it takes on a new look before readers，because each translator has his own understanding of the original，his own purpose for translation and his own style.As there are many scholars in different parts of China who are interested in translating Western masterpieces，and as translated masterpieces sell well on the domestic market，there have appeared a number of translated versions of the same source language text.As each translator understands the source language text differently and incorporates his own understanding into his version，we can say that the translator exerts his subjectivity to comprehend and translate the source language text.
The phenomenon that there exist many versions of the same source language text really proves from an indirect angle that the subjectivity of the translator does exist in translation.
Human beings never stop their natural desire to extend their physical and intellectual organs artificially.“A man who borrows a horse and carriage does not improve his feet，but he can travel by carriage as far as 1，000 li .A man who borrows a boat and paddles does not gain any new ability in water，but he can cut across rivers and seas.”（假舆马者，非利足也，而致千里，假舟楫者，非能水也，而绝江河。——荀子《劝学》）Machine translation （MT）is the translation of natural languages （like Chinese and English）through the use of computers.With rapid advances of science and technology，MT is developing very fast and becoming increasingly popular.
The conception of MT appeared almost at the same time as the appearance of computers.But the concept of machine translation may be traced back to the 17th century.In 1629，René Descartes proposed the development of a universal language with equivalent ideas in different tongues sharing one symbol.In the 1930s，several investigators envisioned the use of automatic machines to assist people in translation.In 1933，George Artsruni，an American French engineer，was issued a patent certificate for a translation machine called “Mechanical Brain”by its inventor.P.P.Trojanskij，a Soviet researcher，proposed a detailed process of translation from one language into another.During World War Ⅱ，a group of scientists and engineers led by Turing，working at the secret Bletchley Park to decode German's Encryption Machine—Enigma，solved an automatic translation problem，though the task was more popularly known as cryptography.Procedures from cryptography were useful.The cryptographic code displayed a one-to-one relation.
In 1947，the first electronic multifunctional computer ENIAC was born in the University of Pennsylvania.This heavy machine，which weighed 30 tons，and which contained 17，468 vacuum tubes，took up 1，800 square feet and consumed 150 kw of power per hour.It was primarily used to calculate artillery firing tables for the US Army；indeed，it was designed to solve a large number of numerical problems.
It seems that computers naturally found their way into translation，like many other human activities，for example，accounting and statistics.One of the most vivid impressions regarding the start of machine translation is that the idea of machine translation emerged during a conversation between Warren Weaver and A.D.Booth in 1947.However，it is not the whole story.For early computer models，it was not easy to implement non-numerical tasks.Messages and programs had to be keypunched on paper cards，and data had to be stored on high-cost magnetized steel tapes.
War and commerce were the two main drivers of translation technology.The end of the war symbolized the fall of the Iron Curtain.The Cold War began.In 1952，the growing general interest in MT led to the sponsorship by the Rockefeller Foundation of a conference devoted entirely to MT.In 1954，an experiment was conducted in Georgetown University.The successful launch of the first man-made satellite，Sputnik，was definitely a shock to the West，especially to the US.Large amounts of intelligence and information written in Russian needed to be processed.Soon，the spring of MT came，with enhanced support by the US government and army.There were 12 machine-translation groups in the US in 1962，mainly supported by several US government agencies such as the US Office of Naval Research，CIA，NSF，and the US Air Force （Josselson，1971）.It was not surprising that in both East and West MT research focused on Russian and English.The Soviets，of course，were not inclined to lag behind.In May 1958，the Russians held their first MT conference in Moscow with 340 participants from 50 institutions and universities，illustrating a huge number of languages being studied for possible MT systems.（Tezisy，1958）
Acceleration of MT research also attracted attention from other international agencies and institutions.The UNESCO held a conference on Information Processing in Paris in June，1960，including a special session on MT.Other countries，like PR China，France，Germany，Czech，also set up their own research groups.At the beginning，the aim of MT was to realize fully automatic translation.“Given a large special computer with direct access，one would be able to store both dictionaries and grammar rules，and thus carry out the whole translation process automatically”（IBM，1960）.
Progress was made，but not as expected；funds were invested，but the achievement was not as satisfactory as it might.Skeptics loomed，disillusions increased，and setbacks doomed.The years from 1956 to 1966 represented a very active period，due to the Cold War and strong financial support from governments and armies used for intelligence purposes.After the ALPAC report saw the light in 1966，it was found that the ten-year research had failed to fulfill expectations so that funds were reduced.
Failure is the mother of success.The flames of hope could not be easily extinguished and the fire of innovation started raging.According to a 1972 report by director of Defense Research and Engineering （DR & E），the feasibility of large-scale MT was reestablished by the success of the Logos MT system applied in translation of military manuals into Vietnamese during that conflict.
In the late 1980s，various MT companies were started，including Trados in 1984，which was the first to develop and market translation memory technology in the year of 1989.The first commercial MT system was developed at Kharkov State University in 1991，with two main language pairs：Russian-English and German-Ukrainian.
The birth of the Internet opened a new epoch.The World Wide Web connects millions of computers globally beyond any geographical and boundary obstacles.MT on the web started with SYSTRAN，which offered free translation of small texts in 1996，followed by AltaVista Babelfish，which racked up 500，000 requests a day，and more innovations emerged during this period like Moses，the open-source statistical machine engine in 2007 spurred by the giant search engine Google in 2012.
The main reasons for the recent significant and growing consideration given to machine translation fall into two categories：the practical need for large volumes of translation and the theoretical challenge posed by linguistic problems involved in machine translation and machine-aided translation.These factors have always，of course，been the driving force for development of product-oriented，experimental MT.
Though different research groups tend to adopt various methods and approaches，three basic types are the mainstream of this field：rule-based machine translation （RBMT），statistical machine translation （SMT），and example-based machine translation （EBMT）.
The earliest method for machine translation was rule-based.Three stages，including analysis，transfer，and generation，are essential steps of this approach.
All the translation problems cannot be solved by applying rules derived from grammar only.Researchers must know the rules under which the language operates and how the mechanism of this operation can be simulated by automatic means.
Systems （of the transformer and transfer kind）were designed in the 1970s and 1980s to run computers.They were then ported to become slow desktop applications for personal computers in the 1990s，and they have been running on high-performance web-based systems without changes in their basic design.
The rule-based system ascends up to the tip of the Vanquois Triangle and descends from there.
Statistical machine translation endeavors to generate translations by using statistical methods based on bilingual text corpora.
Statistical MT systems are trained in translated texts.Where such corpora are available，good results can be achieved translating similar texts，but such corpora are still rare for many language pairs.The quality of statistical MT is totally decided by the size of corpora.In 2005，Google improved its internal translation capabilities by using approximately 200 billion words from UN materials to train their system；translation accuracy improved.（Philipp Lenssen，2005）
The developer of any corpus-based system will encounter the following stages of development to run the system：corpus collection and clean-up，and system training including word and phrase alignment，and parameter tuning.
Phrase-based statistical MT is clearly the dominant paradigm in MT today.It is a two-way traffic，say，word and phrase alignment will determine the translational correspondences at syntactic level，while output of translated texts can be used as corpora as well.The problems caused by differences in word order of languages inevitably depend on the language pair at hand.
It might be better to learn how to translate from past translated examples since translating involves expressing meaning in different languages，so the example-based MT becomes a proper conception in some countries.This kind of system has been built especially in Japan since the 1980s.To produce new versions，example-based MT tends to use a range of data stored in the system's memory.
Besides，boasting a more comprehensive technical advantage，other hybrid systems play a more important role in this field，some of which have been already commercially available.
Consistent development of MT technology and the increasing need for translation at great speed with little cost has fueled the proliferation of online MT systems such as Systran，Google Translate，Babefish，and Windows Live Translator.These systems predominantly offer their services free of charge as part of a web-based platform.They provide real-time translation to the general public through web-based platforms that allow users to type sentences，paragraphs of text，or URLs for almost instantaneous translation into their chosen language.Although online MT systems may not be the best choice for highly accurate，large-scale，domain-specific translation，they adequately serve the small-scale，open-domain translation needs of the general public—as can be seen by the millions of hits per day that such sites receive.
Compared with human translators，two major advantages inevitably become visible，when users tend to use MT translation.
Thanks to the development of science and technology，especially along with the invention and proliferation of Large Scale Integration （LSI），computers are not limited in storage and speed，no longer expensive to use and widely available.They can work 24 hours a day，7 days a week，and 365 days a year free of charge and without any break or complaint.
Large savings are reported by many companies that have installed MT systems：in some cases there have been reductions in the costs of producing finished translations of up to 40% or 50%，and nearly all companies report much faster throughputs.（Hutchins，2003）
Like two sides of a coin，no system is perfect.Limitations and disadvantages are born along with the development of MT technology.For the last 50 years，the output of MT can never meet the expectations and requirements of inventors and users.Pre-and-post editing seems to be the necessary step in the whole process and the output by MT system is sometimes comic and strange though less catastrophic.Translation is no easy task，even for humans.The linguistic units include morphemes，units，phrases，and sentences.Complexity and diversity of human natural languages make it hard for a machine working according to fixed rules set by humans.
Before we start this section，one question should be asked：Are there any links and differences between these two types of translation？The answer is yes.
Phrases like human-aided machine translation and machine-aided human translation which refer to CAT later are definitely not the same thing.Some books argue that the main difference between these two depends on who will be responsible for the real translation task.If man plays an active role，that will be CAT，otherwise it will be machine translation.Maybe that is not true.First，no one can deny that a computer is a machine itself designed by man，no matter how big it is or what shape it is.Second，can a machine think like an experienced human translator？The answer is “no”.Does a machine really play a major role in machine translation？What will be the situation if a machine works under the circumstances without any program and procedure encoded by the human operator in advance？From this perspective，it is hard to say in MT，the real work has been done by the machine itself.Third，the initial and primary theory for MT is still being used effectively in CAT，that is to say，design ideas have been well preserved，like a son who inherited his father's genes.It seems ironic to define what is MT and what is CAT as if they came from different ages.The real difference only exists in the concept.New tools help human translators，reducing their workload and general costs.Undoubtedly，humans have always been in the center of translation，and have been playing a core role ever since.
A car cannot move without wheels，neither can software.Corpora，terminology-management system，translation memory part，and alignment programs are indispensable elements for smooth operation of CAT.
A corpus is a collection of machine-readable texts that have been produced in a natural language communicative process （Leech，1991）.
A terminology management system helps achieve consistent，accurate translations by using tools to store，retrieve and organize terms with a clear set of rules.
A translation memory （TM）is a database that stores “segments”，i.e.sentences，paragraphs or sentence-like units （headings，titles or elements in a list）that have been translated，in order to aid human translators.The TM stores the source text and its corresponding translation in language pairs called “translation units”.Individual words are handled by terminology bases and are not within the domain of TM.
An alignment program takes completed translations，divides both source and target texts into segments，and attempts to determine which segments belong together in order to build a translation memory or other reference resources with content.
Trados is a computer-assisted translation software suite，originally developed by the German company Trados GmbH and currently available from SDL International，a provider of translation management software，content management and language services.It provides translation memory and terminology management.You can visit its official website http ：// www . trados . com for further information.
The name Wordfast is used for any of a number of translation memory products developed by Wordfast LLC.The original Wordfast product，now called Wordfast Classic，was developed by Yves Champollion in 1999 as a cheaper alternative to Trados，a well-known translation memory program.The current Wordfast products run on a variety of platforms，but use largely compatible translation memory formats，and often also have similar workflows.The software is most popular with freelance translators，although some of the products are also suited for corporate environments.For more resources you can get at http ：// wordfast . com .
A variety of views of the development of MT are presented and discussed，which allows the readers to gain an overall view and to conclude personal judgments.We don't attempt to provide all of the answers for the future of this field surely depends on the scope and span of human mind and imagination，that is to say，possible solutions for better achievements of MT will be divergent and various.
When predecessors of machine translation played with their newly invented tools 50 years ago，few of them could imagine how these tools would be in the 21st century.Machine translation，generally speaking，belongs to the field of artificial intelligence which has been changed dramatically by the birth and proliferation of the Internet and new digital technology.The former enables people to get access to information anywhere，anytime in the world，and the latter can bring a new exciting hope for real-time communication.
Three new trends of MT will become true.The first is more flexible.Wireless Internet connections like Wi-Fi and 4G mobile phone grow rapidly，and new mobile computing and communications platform emerges based on smart phones to rival the PC platform.More compatible CAT software will be developed for mobile users.
The second is more accurate.A less perfect but useful translation output is the current state of art.The situation will be changed，thanks to the Big Data，which will share and analyze huge amounts of data across the world from several thousand samples to produce a more satisfactory output and get less human involvement.
The third is that the software used for translation and other natural language processing purposes will become cheaper and more user-friendly on account of fierce competition and technical innovation.The software will not be used alone，instead it will be integrated into other systems.Certainly，the future of MT will be both brilliant and tortuous.
It is not surprising that Snell Hornby should have stated，“Now there is no longer any doubt that the product of technology，however sophisticated，cannot compete with the creative power of the human mind”（1995：66）.No matter how complicated MT is，it cannot offer us a satisfactory translation，because it is unemotional and has no subjectivity.The fact that MT cannot take the place of human translation also proves that there exists the translator's subjectivity in the process of translation and this kind of subjectivity is crucial to successful translation.
1.What is subjectivity and what is the translator's subjectivity？
2.Why is the translator's subjectivity an objective reality？
3.How do you argue for translator-centeredness？
4.Why are there retranslations of the same work？
5.What do you know about the history of MT？
6.How many types of MT are available？
7.How much do you know about the outlook of MT？
8.Can MT replace human translation？
（Written by Zhang Lei and Zhang Yuhuan）