Editor’s Note: A previous version of this article was originally published at the ATA 42nd Annual Conference in Los Angeles in 2001.

Due to several unique linguistic and technical challenges, translating into Chinese – even when compared to other Asian languages – not only requires more energy and more time, but also calls for a much higher level of linguistic talent, expertise and professionalism. Current localization practice – without recognizing these important differences – has been generating negative results for all parties involved. This article, which includes the linguistic and technical basics that underlie the “Chinese challenge,” also proposes a win-win solution for the industry.

In part 1 below, Zhang provides wonderfully rich illustrative examples of what it is really like to create new terminology in Chinese. In part 2, he will discuss why Chinese poses challenges that Japanese and Korean do not, along with specific recommendations for project managers on how to select Chinese resources.

An Introduction to the Chinese Language

Of all the major European and Asian languages, Chinese is the only one that is entirely based on ideographic characters. The total number of Chinese characters is over 50,000. Xin Hua Zi Dian (新华字典, New China Dictionary), a small but authoritative Chinese dictionary designed for users with an intermediate level of education, contains about 11,000 Chinese characters, of which about 5,000 are generally considered as being frequently used.
In the early 1950’s, the government of Mainland China introduced Simplified Chinese in an effort to promote nationwide literacy. Meanwhile, Hong Kong and Taiwan continued using Traditional Chinese.

After a half-century, the differences between Simplified and Traditional Chinese are much more than simplified versus traditional characters. There now exist general style differences, along with vocabulary differences in many technical fields. In addition, there are all sorts of computing-related technical differences, such as for character encoding. For all intents and purposes, the localization industry now regards Simplified and Traditional Chinese as two separate languages. (Since we are addressing only the basic language features common to both Simplified and Traditional Chinese, these two versions of Chinese are not distinguished in this paper, although the examples are all given in Simplified Chinese.)

For English-to-Chinese translation, the text expansion rate is usually a negative value, meaning that the target Chinese text shrinks to a lesser number of bytes which then translates to less layout space. If the same text were to be further translated into the ancient Chinese language used during the Qin dynasty (221–207 B.C.), the text would shrink to about half its size. This is because most words in modern Chinese are two-character words, while they were all one-character words in ancient Chinese.

This transition from using only one-character words to a preference for two-character words in writing – one of the most important turning points in the the evolution of the Chinese language – happened when fu (赋), a gorgeously stylish Chinese prose genre, emerged and flourished during the Han dynasty (206 B.C.- A.D. 220). Since each character originally had and still has at least one meaning, each character is a simple word, and each two-character word (or word with more characters) is a compound word.

Chinese sentence structure is dramatically different from that of European languages in two ways. First, Chinese follows a “topic-description” pattern instead of the “subject-verb” formula. Second, Chinese is a synthetic, rather than an analytic, language. This means that details come before the general, and modifying elements come before the modified in sentences.

Technical Challenges

Terminology Development and Management

There are three types of compound words in Chinese, which are formed in three different ways. A class 1 compound word contains two characters of similar meanings. A class 2 compound word may contain two or more characters, and the character(s) before the last modifies the last character. A class 3 compound word contains a main character followed by a character that has no meaning within the word. This probably occurs due to a perceived need for a two-character word.

When English words are translated into Chinese, the same three ways of compound word composition are used. Table 1 lists some of these compound words translated from single English words.

Let’s use the example of translating into class 3 compound words to illustrate the unique challenge faced by Chinese translators. When translating an English word into another European language, the translator selects a specific word from all available words that could possibly be used in translating a particular source word. When translating an English word into Chinese, the translator selects a specific combination of two available one-character words from the many possibilities for a specific compound word. The task for a Chinese translator is obviously much tougher. First, he has to go through a much more complicated two-step mental process to translate one word. Instead of comparing directly between candidate words, he must first compose all possible combinations for all original candidate words, then compare all of these combinations. Second, because of this combination process, a Chinese translator must compare and select from many more possibilities. Assuming that n represents the number of available words (one-character words in the case of Chinese) that can possibly be used to translate a certain source word, and that this number is the same for Chinese and for a European target language. Theoretically, this means that the number of possible target choices for the European language translator is n, while it is n(n-1) for the Chinese translator. If n=10, the European translator would have 10 choices for the target term, whereas the Chinese translator would have 90 possible combinations. In practice, although many possible combinations are never used and can thus be discarded, there are still many more choices for a given target term when translating into Chinese.

The immediate result of this higher number of choices is that translating into Chinese requires more mental energy and more time. In today’s localization environment, this further causes a much higher level of difficulty in maintaining terminology consistency between different parts of a document, between different phases of a project, between different translators on the same project, and even between different pages for the same translator. The rationale is simple: if one source word can be translated into only one target word, the consistency problem simply does not exist. If there is more than one way to translate a word, the consistency problem becomes legitimate, and as the number of choices increases, the level of difficulty in maintaining consistency also increases – and not in a linear fashion, but in a geometric progression.

The considerably higher number of word choices for translating into Chinese therefore causes two results. It requires even more work, and a much higher level of professionalism, for Chinese translators to produce a translation at an equal level of consistency and linguistic quality. And, it requires a higher level of effective teamwork between the translators and the project manager than for other languages.

Technical Terminology Translation: Classifiers for Action Carrier Suffixes

The English language uses the action carrier suffixes –er and –or, which are very handy for technical writers. A person who prints is a printer, and a printing machine or a printing company is also a printer. We translators all know that we are human, but we cannot prevent someone from inventing a machine and calling it a Translator. When translating terms with an –er or –or suffix into Latin languages, it’s not necessarily a harder job than the original creation of these terms by the technical writer, for, technically speaking, the translator does not need to distinguish a machine from a person or a company. He can use the same action-suffix formula by simply applying the word stem and suffix counterparts in his target language.

Translating the same terms into Chinese, however, is a totally different story. The simple fact is that Chinese has neither an –er nor an –or construction, nor anything similar. The method employed is to use a character that classifies the object by the term to be translated, preceded by a one- or two-character word that modifies the classifying word by describing the function of the object. For the convenience of this discussion, we will call the character that classifies a classifier (as I said, it’s hard not to use a term with an –er or –or suffix when writing in English!), and the character(s) that modifies a modifier (here is another!). (In some grammar books, the Chinese term liang-ci (量词) is translated as classifier, which should not be confused with our usage of the term here. We are also going to discuss liang-ci in this article using the term counting word, which, I believe, is a more accurate translation of the Chinese term.)

The challenge for Chinese translators is that the information needed for classifying the object is not provided in the original term. If this information is not already in the translator’s knowledge base, then he must do research before he can translate the term.

Let’s consider an easy example, that of translating two English words, commander and bugler, into Chinese. A commander commands and a bugler bugles, so there is no problem in choosing specific modifiers for them. But what’s the difference between the –er in commander and the –er in bugler? They look, sound and are spelled exactly the same in English. But the Chinese translator must uncover the differences that will enable him to choose the correct modifiers. Luckily, the translator can use common sense to figure out that in a military unit a commander is usually the highest-ranking officer, whereas the bugler is probably one of the lowest. Therefore, he classifies a commander as an officer, or 官, a bugler as general personnel, or 员, and accordingly translates these two words as 司令官, or command-charging-officer, and 司号员, and bugle-charging-personnel, respectively.

If the differences between social roles of a human being need to be identified and reflected, then different attributes between machines or devices must be identified and reflected. It goes without saying that the difference between a human being and a machine (remember the example above of translator as a human being and as a machine) must be made clear in the first place. When it is clear that the object referred to by the term is a man-made device, then further attributes of the device must be identified in order to select the correct device classifier for translating the term.

Table 2 below lists several common classifiers used in translating device names. Each classifier implies certain attributes of the device, and in each class, a device name is translated using the principle discussed above. Typical devices in the first class 机 [ji], for example, are usually large and heavy machines. The term aircraft in the first class, for example, is translated as 飞机 [fei-ji], or flying machine, with the classifier 机 indicating that this is a rather large machine. The simplified Chinese translation of the term computer, 计算机 [ji-suan-ji] or counting-computing-machine, also uses the same classifier because an early computer could require a whole building to house it. The second classifier 器 [qi] also refers to the attributes of device size and weight, but a 器 is usually smaller and lighter. Therefore a calculator in Simplified Chinese is translated as 计算器 [ji-suan-qi] or counting-computing-device. Note that the modifiers used for translating computer and calculator are exactly the same, with the only difference being size, as indicated by the classifiers in their translated Chinese forms.

In other words, to be able to translate a device name into Chinese, just having the idea of what this device does, which is usually indicated by the stem of the original term, is far less than enough. In order to choose the right classifier in translating a device name, one has to know a number of extra attributes of the device, such as size, weight, application, design, level of operation preciseness, etc., which are usually not suggested by the original term. Therefore, a Chinese translator has to develop detailed technical knowledge about what he is translating. For example, if he is translating for a computer company that develops networking products, and he encounters a term with the –er or –or suffix, he has to think, ask, study, or do whatever he can to find the answers to a series of questions such as: Is this a human function name or a device name? If it is a device, is it a stand-alone device, an integrated part of a device, or a plug-in? Is it hardware, software or firmware? If it is hardware, how big and how heavy it is? What’s its shape? Et cetera. Et cetera. Why do we have to know all these attributes? Because they need to be imbedded in the Chinese translation, even though they are not reflected in the source term.

Technical Terminology Translation: Counting Words for Nothing

Added to the classifier issue is the counting word, a linguistic rule that does not exist in European languages. The usage of counting words in Chinese sounds like certain English expressions, such as “a piece of paper” or “a glass of water.” The difference is that in Chinese a counting word must exist between a number or pronoun and a noun. Therefore, one book or a book is translated as 一本书, or one ben of book; a car is translated as 一辆车, or one liang of car, etc. Table 3 below lists some commonly used counting words and their usage. Again, we see that in translating technical terms, the correct selection of counting words requires knowledge of some special attribute of the objects.

Therefore, it is not at all abnormal for a European language translator to translate a passage of technical text in two hours, while a conscientious Chinese translator would require four hours to translate the same passage. When I was in charge of Chinese localization quality at Intel, I devoted a considerable portion of my time to studying the technical details of the products being localized, in order to be able to verify that technical terms were being translated correctly.

Source Text Comprehension

A widely adapted practice in the translation industry is to have translators translate into their native languages. Since it is extremely hard to find a translator who is native in both the source and the target languages for a given project, this is probably the better choice for many projects. However, a translator’s source language comprehension ability should never be taken as less important than his target language expression ability. After all, if a source language expression is misunderstood in the first place, no matter how swiftly and beautifully it is translated, it will be translated incorrectly.

While this is true for all language pairs, source text comprehension constitutes a special challenge for Chinese translators. Primarily, this is due to the tremendous difference between Chinese and European languages. But added to this is the historical fact that China was completely closed to the Western world for over a quarter of century. During those years, China’s foreign language students could rely only on their textbooks and grammar books, with no exposure to any living Western language or the culture associated with it. The whole generation of foreign language students who graduated during those years is now teaching in Chinese colleges.

Here’s just one example from a news show that I saw on TV during a recent visit back to China. A city that is home to several dozen universities and research institutes renovated a shopping street and made it a pedestrian-only street. A sign in English was placed at the entrance to the street with the following words: Walking Street. An English language professor saw the sign and sent his comment to the local TV station. The station took it very seriously and broadcast it with the host’s supporting comment. The professor’s comment was as follows, “With more and more foreign tourists and investors visiting China, our whole society needs to pay attention to the correct English language expression, especially in public locations. The sign that says Walking Street is an example of poor English, because it is wrong in grammar and means a street that is walking!”

Therefore, if I were to name only one very serious problem in Chinese translation, it would be source text comprehension. In my editing experience, I found countless comprehension mistakes in all sorts of text and on all sorts of text elements. My conservative estimate is that this kind of mistake exists in 90-95% of original Chinese translations produced by the translation/localization industry everyday, and at least half of this is passing QA and being released. Table 4 below lists some examples taken from localized products of several leading hardware and software companies whose products we use every day, and who take localization very seriously. These examples are all from technical translations. In marketing literature translation, comprehension-based mistakes are more abundant – and to irrelevant parties, they are often more amusing as well.

This challenge calls for higher linguistic talent in Chinese translators. Among all U.S.-based translators, Chinese translators probably have the highest level of education. However, they are also the most challenged by the task of source text comprehension when translating from European languages, typically English, into Chinese. Although I have never tested this concept, I believe that if an average French native who has never studied English and an average Chinese native who has studied English for 3-5 years were shown the same newspaper article in English, the French person would understand more than the Chinese. In short, a Chinese translator needs more linguistic expertise than translators of European languages.

Acknowledgment

As the author’s long-term teammate in Chinese translation, desktop publishing, graphic processing and localization consulting, the author wishes to thank Kailian Li for her constructive input and ideas throughout the writing of this article.

---

Yu Zhang is a U.S.-based, freelance technical Chinese translator, localization consultant, and Co-Founder and Director of Locobyte. He has translated, edited and evaluated linguists for more than 50 of the Fortune 500. His former careers include being a Rockefeller Fellow in molecular biology, a localization trainer and Trademark and Brand Name Localization Committee member at Intel, and a lecturer in Chinese at the University of California. Zhang can be reached at yzhang@locobyte.com.