Peaceful Immigrants or Invading Hordes?
A Review of Three New Books
This review was commissioned and published by the on-line publication Translation Journal and appeared
in their October, 2004, edition.
José R. Belda Medina: El Lenguaje de la Informática e Internet y su Traducción, 328 pp., Publicaciones de la Universidad de Alicante, 2003. 18.00, available from:
Chan Sin-wai, editor: Translation and Information Technology, 215 pp., The Chinese University Press, the Chinese University of Hong Kong, 2002.
$18.00, available from Columbia University Press:
Jeffrey G. Brown: Thinking in Chinese, An American's Journey into the Chinese Mind, 133 pp., JB Linguistic Works, Philadelphia, 2002. $14.95, available from Amazon at:
All three of these books deal with a crucial problem underlying some translation tasks: precisely how do we handle a vast technical vocabulary unexpectedly imported into a language not necessarily prepared to deal with it? This question applies most urgently to the integration of computer terminology into many of the world's languages, but as we shall see this is not the only possible example. The answer may turn out to be that there is no one method guaranteed to be fully successful in all cases. And that unless there is a pressing need to for a culture to absorb such a vocabulary, no one standardized approach may immediately be acceptable.
Belda Medina's work is a brave, scholarly undertaking aimed at corraling the unruly herd of foreign interlopers that computer and Internet terminology has let loose into Spanish by classifying these thousands of new lexical items into meaningful categories. He sets out to isolate them in five major stockades, which comprise the five chapters of his book: Compounds, Derivations, Abbreviations, Specialized Terms, and Loan Words. Each of these stockades is subdivided into smaller holding pens according to principles that are on the whole gramatically and terminologically valid but can be a bit hard to follow.
For instance, "Compounds" are classed as orthographic, neoclassical, and what the author calls "syntagmatic," itself a technical term derived from Saussurian linguistics. (Note 1) Examples of orthographic compounds are "keyboard," "backslash," "clipart," "printout," "upgrade," "user-friendly," and many others. Examples of neoclassical compounds are all those words built up from classical prefixes, such as auto-, mono-, mini-, bi-, cyber-, etc., covering a wide range of well-known computer terms. Syntagmatic compounds differ in Belda's view from their orthographic cousins by juxtaposing two or more complex and/or longer words to create a new meaning, as in "assembler language," "analog computer," "stand-alone terminal," or even the humbler "junk mail." Each of these classes presents slightly different problems for translators into Spanish.
The work occasionally adopts a slightly prescriptive tone by suggesting preferred translations, for example it disapproves of rendering the word "run," as in to "run a program," as correr. And such a usage does make for somewhat inelegant Spanish, even though correr can be both transitive and intransitive, but Belda's complaint has not been fully heeded by translators.
One of the book's most interesting sections deals with the problem of translating -ware compounds into Spanish (hard-, soft-, share-, free-, vapor-) versus -tica compounds into English (informática, telemática, cibernética, robótica). Another is a detailed section on chat programs, which includes the information that "Hi, dude" translates into Spanish as Hola tío and also provides Spanish translations for most of the common chatroom and email abbreviations, such as ICQ, LOL, and RTFM, and even for smileys aka emoticons. The author seems a bit disappointed that hispanoparlantes have neglected the words charla and charlar in favor of Spanglish chatear and suggests tertulia or conversación as other possible improvements.
At the same time this professor of English Philology at the University of Alicante is scrupulously honest in pointing out that one of the main driving engines of English computer terminology, namely abbreviation, has a tendency to work less well in Spanish, sometimes favoring the use of the original English expressions instead of translations. He sees two reasons for this, first that so many words in Englishand so relatively few in Spanishare monosyllabic, which in itself promotes a form of brevity. And second that the accent in many English words falls on the first syllable, thus encouraging such further short forms as lab for laboratory and pic or pix for picture or pictures.
It turns out that there are three possible words for "web" in Spanish: red, telaraña, and malla, but this wealth does not end up being entirely helpful. In his concluding section Belda criticizes the Instituto Cervantes for suggesting that the correct Spanish translation of World Wide Web should be Malla Máxima Mundial, when the most frequent Spanish translations for the terms website and web page are sitio web and página web. He also points out quite correctly that computer terminology has gone through two generations since its birth, first a more truly technical vocabulary using such terms as del, dir, cls, list, DOS, etc., but in recent years a more accessible semi-technical vocabulary drawn from everyday words and tied to the Windows environment: window, folder, menu, dialog box, etc.
The author sometimes seems to be engaged in a private conversation with other published authorities on this subject, and the famous DRAE (Diccionario de la Real Academia Española) does not go unmentioned. Belda's volume is extremely learned, a genuine tour de force of its genre, but I would not necessarily recommend it for translators working on computer texts from English into Spanish. For this purpose, something a bit more rough and ready might be preferable, for instance a straightforward alphabetical English to Spanish computer glossary, such as those available from:
A closely related problem is every bit as alive in Chinese as in Spanish, as two papers from the middle section of Chan Sin-wai's volume clearly reveal. (Note 2) The main difference is that the basic structure of the Chinese language may be better endowedperhaps too well endowedto handle the thousands of invading new words and acronyms. The result can be a veritable profusion of possible Chinese translations for a specific Internet term. For instance, the word Internet itself can be conveyed in at least four different ways in Chinese, as transliterated in the following table in both Cantonese and Mandarin:
This paper presents an adjoining table with yet another nine possible translations. This is because meaning in Chinese is built up from monosyllabic characters, each usually with its own quite separable meaning, rather than from what we in Western languages call "words." As I have pointed out elsewhere (Note 3), if we credit Chinese with no more than 3,000 characters (though some estimates run far higher), it would be theoretically possible to pair every one of those characters with each other in a table reading 3,000 characters across and 3,000 characters down, for a total of nine million possible two-character compounds, each theoretically capable of possessing its own potential meaning. Three- and four-character compounds, though less common, are also possible. In other words, Chinese possesses semantic building blocks to spare as compared with most Western languages.
Not all of the thirteen possible compounds for Internet are widely used, and just as there are many national versions of Spanish, shortly after presenting the above table the authors of this paper point out that some Internet terms are "expressed differently in different Chinese-speaking regions, such as Hong Kong, Taiwan, China, Singapore, Malaysia, and so on." But at least one major terminological quarrel is rooted in politics alone.
Of the four terms in the above table, the most widespread is hu lian wang, or "mutually united nets." But during 1997 in Beijing the China National Committee for Terms in Sciences and Technologies (CNCTST) attempted to impose the term yin te wang, wang once again meaning "net" or "nets," but with the two characters yin te intended as nothing more than a phonetic filler supposedly replicating the English sound of "Inter-" in the word "Internet." This has not proved successful, as one critic points out:
I am in school every day, and in here you almost cannot hear the word yin te wang, since hu lian wang has already become part of our school life here. Every time when I return to the town or the village, I can only hear hu lian wang and not yin te wang. Let's imagine how difficult it is for people to understand you when you utter the suave word yin te wang to a worker or a farmer. Are you making fun of their knowledge? (Note 4)
Other attempts from above to impose Chinese translations for computer terms have met with resistance even from those professionals most familiar with the technology and may in the long run threaten the use of Chinese in this field:
Interestingly, using pure Chinese terminology in any discourse does not seem to be welcomed by the subjects. From their viewpoint, Chinese terminology causes comprehension difficulties. They have to recall the English equivalents in order to understand. (Note 5)
This book contains two other parts, each one longer than the section on terminology, and in a sense they tend to cancel each other out. Part I of the book is a survey of various Chinese projects to improve Machine Translation, including a system for translating Chinese cookbooks reminiscent of the limited vocabulary created by the Canadian TAUM-Météo weather prediction program during the Eighties. Other relatively familiar approaches include a "Chinese-English MT System Based on Micro-engine Architecture," a piece entitled "Example-based MT: A New Paradigm" and the use of "Bilingual Corpus Construction" to manage MT from Chinese into English.
Even though three authors concluded, after a careful study published in a recent issue of TJ, that MT will most probably never be able to translate the vast majority of texts and may serve best, where it serves at all, as one additional tool for human translators (Note 6), the papers in this section express mainly optimism for the future of this approach. Heard again are some of the same arguments favoring MT commonly voiced as early as the Eighties: that it will minimize "the shortage of human translators," that "the quality and marketability of MT have been improved considerably," that MT is "43,200 times faster than a human translator," that "MT systems will be widely used and web translation will be a leading trend in the future." Yet the editor also concedes, on the very same page where these statements appear, that "MT systems in the market are far from satisfactory." (Note 7)
Part III of this book goes considerably further in this direction. It is largely a compendium of articles either attacking MT or expressing doubts and/or fears about its ultimate efficacy (or simply ridiculing its use for translating any sensitive document). In many ways these two sections largely recapitulate the terrain covered by discussions of MT in such publications as Language Monthly, Language Technology, the ATA Chronicle, or TJ's paper precursor the Sci-Tech Translation Journal during the 'Eighties and early 'Nineties. This is not altogether surprising, since for all its current progress China was relatively slow during earlier decades in integrating the computer into its social and commercial structure.
The titles of some of these articles are revealing in themselves: "Computer Technology and TranslationFriends or Foes?" or "Globalization on Language: Death of the Translator in the Technological Age" or "Shall We Dance, When the Smart Machines Take Over, Virtually?" This last piece, written by Prof. Evangeline S.P. Almberg, contrasts two line-by-line Chinese poems as translated by one human and three MT programs, an exercise familiar both to the reviewer and the editor of this journal. One of the Chinese-English examples she chooses to make her point, however, is rather incisive:
Professor Almberg points out that this clever epigram could be translated in two different ways by human translators, either as:
Literature is that which expresses what language cannot express;
Science is that which clarifies what language cannot clarify;
And machine translation is that which makes unclear what language has made clear. (Note 8)
While literature expresses what language cannot otherwise express,
and while science makes clear what language cannot make clear,
machine translation befuddles anew what language has already made clear.
She then contrasts both or either of these human translations with the results obtained from three machine translation systems:
Language place cannot express to come out, express, is literature.
It is not clear to say language, it is clear to say, is science.
Say language understand, it is not clear to say again, is machine translation.
Language can't explain to come out of,/ Language can't express and come out of.
Express and come out of, is a literature.
Can not remember clearly the chu the language of,
Say clarity of, is a science;
Say the language clarity of,
Again to caned not remember clearly the chu of,
Is a machine translation.
And Software Three:
Can't express the language out,
Have expressed out, is literature.
Say and know at the pairs of language,
Say clearly, is the science;
Language make clear,
Is machine translation.
Both of the books discussed so far deal with computer terminology, both take for granted the basic linguistic and commercial realities surrounding us, and both assume that we know enough about the languages we speak to force-feed them a vast, new, and more or less foreign vocabulary. But our third book, which also deals with Chinese, questions all these assumptions and raises the possibility that there may be yet another technical vocabulary, equally daunting in its size and complexity, which we have almost totally failed to accept or even recognize, much less integrate into our language and culture.
The author is Jeffrey Brown, a young medical doctor who spent a year in China and unlike many American students of this idiom did his best to decipher the remarkable Chinese linguistic code for representing reality. He finds this code so challenging that he has come to question whether in some cases there can be any truly accurate translation between Chinese and English.
This is certainly a brave position to take, since the history of language studies is littered with the remains of Western theories about Chinese that had little connection with reality. Antoine Court de Gébelin, for example, a friend of Benjamin Franklin who during the 18th Century enjoyed a vogue for comparing Chinese characters with Egyptian hieroglyphics and claiming to see in both the influence of mystical Tarot fortune telling. Or a century earlier even the mathematical genius and co-inventor of the Calculus Gottfried Wilhelm von Leibnitz, who imagined Chinese to be a universal system of picture writing which all other languages could readily embrace (though even Japanese and Korean have never fully integrated it). Or more recently the late Sorbonne scholar René Étiemble, who in his book Parlez-Vous Franglais? savagely attacked the use of English words in French yet warmly welcomed the possible introduction of Chinese characters for writing his language [Note 9].
From the very beginning of his book, Dr. Brown insists on a close relationship between "Language and Thought," indeed this is the title of his very first chapter. He also believes that both Chinese language and thought patterns differ from Western ones, simply because they are based on Chinese characters. This too is a brave position (though one the reviewer agrees with), as it runs contrary to almost all current formal theory, especially the speculations of the MIT School of Linguistics. Dr. Brown is clearly aware of his temerity and proceeds with some restraint in his early pages, becoming almost apologetic as he challenges the following position of MIT guru Steven Pinker:
Knowing about the ubiquity of complex language across individuals and cultures and the single mental design underlying them all, no speech seems foreign to me, even when I cannot understand a word. The banter among Guinean highlanders in the film of their first contact with the rest of the world, the motions of a sign language interpreter, the prattle of little girls in a Tokyo playgroundI imagine seeing through the rhythms to the structures underneath, and sense that we all have the same minds. (Note 10)
This passage occurs as the very final paragraph of Pinker's The Language Instinct and can thus be viewed as an encapsulated overview of MIT linguistics. Since this approach to language has garnered so much recognition both from specialists and the general public, I find it refreshing that yet another generation of scholars is beginning to voice doubts about it. If this notion were true, then translation would present no real problems, MT would work perfectly for all texts, and there would be no need for translator-training courses, the ATA, or even Translation Journal. Even without MT, would-be translators into and out of whatever languages would merely need to connect themselves to "the single mental design underlying them all" and easily come up with a perfect translation, since "we all have the same minds."
But as we all know, much of the time translation doesn't truly work this way. As Martin Luther complained in 1530:
And it's often happened to us that we've searched and asked for fourteen days, even for three or four weeks, after a single word, and in all that time we haven't found it. (Note 11)
Dr. Brown takes the connection between language and thought a great deal further and questions whether Westerners can hope to truly understand the Chinese language unless they are willing to accept the cultural assumptions taken for granted by many Chinese. He links Chinese characters to Daoist, Confucianist, and Buddhist thought, but also to stir-fry cuisine and the history of China. And since he comes from a medical background, his main area of study and comparison soon becomes traditional Chinese Medicine.
It is here that he discovers an enormous new vocabulary of names, diagnostic procedures, and concepts of treatment, though one that has never fully made it into English usage and perhaps never will. He sums up his discovery at the beginning of Chapter 6:
The connection between language and thought, I imagined, lay somewhere in the mind of ancient Chinese medicine. The Chinese physicians, in pursuit of the health of their patients, had developed a sophisticated system of physiology, pathology, diagnosis, and treatment, completely outside the realm of Western medical science. It was a system of thought that had its own form of logic...
But the sense of urgent pressure so many nations feel to master computer terminology simply does not extend to Chinese medical terms, even though they might be more beneficial to our health, and even though Westerners trained in Chinese methods have been active for decades and during the last few years acupuncture, taiji, and qigong salons seem to have sprung up on almost every corner. Where only two decades ago New York City was busy shutting down hole-in-the-wall herb shops in Chinatown, not merely for selling herbs but for performing acupuncture in their back rooms, today over fifty state-certified colleges of acupuncture and Chinese medicine have burgeoned across the nation, and the terminology they and their graduates use, as Dr. Brown points out, has remarkably little in common with Western medical language. It might be interesting to hear Steven Pinker explain how these words and ideas connect to "the single mental design underlying them all."
Thousands of such terms are defined and explained in numerous highly technical reference books on Chinese medicine, though almost none of these terms is in currency among Western doctors and nurses. While they do not sound "technical" in the way we expect technical terms to sound, they nonetheless demarcate fairly precise territories within the realm of Chinese medicine. Here are just three examples, the names of three well-known syndromes in Chinese medicine:
It is important to add that from Laotse to Confucius all the way down to modern times, one of the goals of the Chinese language has been to reduce meaning down to four-character compounds similar to the first two above. Perhaps the technical character of these terms can be better appreciated from a more extended description of the first syndrome, as follows:
The pathologic changes due to accumulation and retention of pathogenic heat in the large intestine. The main symptoms are constipation, abdominal distension with pain and marked tenderness, yellowish [sic] and coating of the tongue, strong deep pulse, etc. (Note 12)
A typical Chinese medical text would of course go considerably further than this, providing details related to causal factors, differential diagnosis, and various methods of treatment. It is easy enough to see why Western doctors and patients are not eager to master this terminology. First of all, it overlaps but often fails to coincide precisely with Western medical definitions. Second, like computer terminology, for many it is not easy to master. And third, many patients in our society are not fully able to comprehend even Western medical terminology, and it is understandable that they would not welcome learning a second system.
This complexity extends to other fields as well. As soon as one translates the everyday Chinese compound tianqi as "weatherwhich is the main definition provided in most Chinese-English dictionariesone has done violence to the meaning of Chinese, since a more accurate translation would be "sky qi" or "qi of heaven." This is not at all the same as what we call etymology in Western languages, since the elements of meaning are not little-known Greek and Latin roots or obscure affixes like "ad-," "syn-," or "-hood" but everyday full-blooded Chinese characters, each with its own well-known meaning.
There are many phrases in Chineseand even in Japanesewhere such a shortfall in translation plays a role. It is therefore not surprising, just as English computer terms have made their way into both Spanish and Chinese, that some Chinese medical terms, like qi, yin, yang, and xue ("blood"), have begun to appear in English.
Lest it be assumed that Chinese medicine is perfect as it is now taught, it should be added that some of the same problems encountered in translating English computer terminology into Spanish can also be found in translating Chinese medical terms into English. Just as the use of three possible Spanish words for "web" have caused confusion, so many Chinese medical texts in English have relied on a variety of English words to convey the meaning of the same Chinese character. As Marnae Ergil, one of the main advocates for accurate translation in this field, points out:
...over the years a variety of terms have been used to translate any one Chinese character....Some authors make decisions based on extensive linguistic research, others choose a term because it sounds good, because it is what has been used before and so has become the accepted norm or because it is the western medicine translation of the concept. (Note 13)
Dr. Brown concludes by advancing further arguments against the assumptions of MIT linguistics and even suggests that our current definition of "science" may be self-serving, deficient, and not truly applicable to much of Chinese thinking, which through its Daoist roots has always been based on its own form of independent nature study. His final chapter is a spirited defense of Benjamin Lee Whorf's observations that language influences culture, a view MIT linguists have done their best to bury, though it keeps coming back to haunt them.
In the eyes of most terminologists and translators this campaign has surely been a foolish one. Vicious and even personal attacks on Whorf are frequently found in the literature of the MIT school of linguistics, and Steven Pinker, who has also boasted that the output of "translation engines" is definitely improving, is no exception, as can be found in laborious detail throughout more than nine pages of his book The Language Instinct. The reason is that Whorf's position (more properly the Sapir-Whorf Hypothesis), namely that language can influence culture, still stands in irrevocable opposition to fuzzy MIT claims that "a single mental design" underlies all languages and "we all have the same mind."
For instance, MIT spokesmen have repeatedly assailed Whorf for claiming that there are many words for snow in Inuit languages, yet their own literature defends what is essentially the same argument, that there are many different words for camel in Arabic. [Note 14] Here again the reason is simple: the claims for "snow words" are backed by the arch-enemy of the MIT school, Benjamin Lee Whorf. The claims for "camel words" are backed by another scholar of no particular importance to MIT linguists. But it is after all only logical that any language will possess an enhanced vocabulary and/or terminology for those aspects of reality of greatest importance to its speakers, as for example English has a large vocabulary for automotive words and/or terms.
To sum up, Belda Medina and the Hong Kong terminologists are perfectly correct in doing their utmost to seek out the best possible translations for computer terms. But the turnover in computer terminology has already been formidable, and there is no foretelling how many new generations of terminology in all fields may lie in our future. Or how many new generations of hardware, or perhaps even of currently unimaginable new devices, each with its own attendant wordware, may lurk just beyond the horizon. Or what unforeseeable new breakthroughs in our thinking might spawn, giving birth to entirely new terminologies quite beyond our powers to predict. Based on past history, all these developments could become so daunting as to turn our current terminologies in all fields into little more than dead languages, so that even parts of this review might become hard to understand as little as a hundred years from now.
In our "management" of terminology we may in fact be looking at an endless task, like the painting of the George Washington Bridge, where no sooner do the painters finish at the New Jersey side but they have to go back to work on the New York side and start all over again. Of the making of terminology there is no end, nor of the inventions that spawn it. Against this certainty we have but one consolation: what we face today is not truly a new problem at all. As Roger Bacon wrote back in 1268:
We must consider the fact that translators did not have the words in Latin for translating scientific works, because they were not first composed in the Latin tongue. For this reason they employed very many words from other languages. (Note 15)
The "other languages" he meant were primarily Greek, Arabic, Syriac, and Persian at a time when such words as alcohol, horizon, massage, zero, and many others were regarded by some people as difficult technical terms, not totally unlike the way many view new and unfamiliar vocabularies today.
1. Here is part of David Crystal's definition of "syntagmatic" in A Dictionary of Linguistics and Phonetics, p. 300: "The relationships between constituents ('syntagms' or 'syntagmas') in a construction are generally called syntagmatic relations. Sets of syntagmatically related constituents are often referred to as STRUCTURES... Syntagmatic relationships can be established at all levels of analysis."
2. Chiu and Björn Jernudd and Technological Problems and Language Management for Internet Language Professionals in Hong Kong by Charlotte To and Björn Jernudd.
4. Chiu and Jernudd, in Chan Sin-wai, p. 104.
5. To and Jernudd, in Chan Sin-wai, p. 124.
6. Machine Translation and Computer-Assisted Translation: a New Way of Translating? by Olivia Craciunescu, Constanza Gerding-Salas, and Susan Stringer-O'Keeffe, found at:
It seems to this reviewer that this detailed and balanced study deserves to become far better known, since only a remarkably small proportion (little more than one percent) of translation work has proved amenable to MT methods over the decades, nor does this market share appear likely to increase in the near future. Yet as Part I of Chan Sin-wai's book demonstrates, MT advocates have never ceased boasting of the vast power and prowess of their approach over the decades, even though this partakes of a continuing and unjustified slur on human translators. It also becomes apparent from this report that the reasons for these boasts (and perhaps one source for continued MT funding) springs from the minds of market analysts easily swayed by the prospect of MT one day gaining a far larger market share.
7. Chan Sin-wai, p. viii.
8. This and the next four cited passages come from pp.185-86 of Prof. Almberg's paper, in Chan Sin-wai.
9. Prof. Étiemble presents this argument on pp. 114-17 of his work L'Écriture.
10. Dr. Brown cites this passage on pages 13-14, near the beginning of his book. And as noted Steven Pinker places it as his final summing-up at the very end of his work The Language Instinct, on page 430.
11. Luther, p. 15.
12. Sung, pp. 238, 239, 216.
13. Marnae Ergil: Attentions in the English Translation of Traditional Chinese Medicine, available on-line at:
14. The repeated snortings and stompings of camels, accompanied by a voiceover explaining the linguistic nuances, can be seen in the videotapes The Human Language Series, the official audiovisual arm of the MIT School of Linguistics, directed by Gene Searchinger, Equinox Films/Ways of Knowing.
15. Bacon, I: 76, as translated by Robert Belle Burke.
Bacon, Roger. The Opus Majus of Roger Bacon (ca. 1268). Philadelphia: University of Pennsylvania Press, 1928. Translated by Robert Belle Burke.
Crystal, David: A Dictionary of Linguistics and Phonetics, Blackwell/Deutsch, Oxford, 1985.
Étiemble, René: L'Écriture, Gallimard, Paris, 1973.
Étiemble, René: Parlez-Vous Franglais?, Gallimard, Paris, 1964.
Luther, Martin: Sendbrief vom Dolmetschen. (originally published in 1530). Herausgegeben von Karl Bischoff. Halle: Max Niemeyer Verlag, 1951.
Pinker, Steven: The Language Instinct, W. Morrow & Co, New York, 1994.
Sung J. Liao: Chinese-English Terminology of Traditional Chinese Medicine (hanying shuangjie changyong zhongyi mingci shuyu), Hunan Scientific Technical Publishers, 1981.