PROMOTING MULTILINGUALISM ON THE INTERNET:
Mr Taik-Sup Auh
Graduate School of Journalism and Mass Communication
Republic of Korea
Since its birth a generation ago, the Internet has been dominated by the English language and North American culture. In a 1998 survey conducted by the Internet research group, eMarketer, two-thirds (68 percent) of a little over 60 million Internet users worldwide reside in just two countries, the United States, with 37 million users, and Canada, which has just over 4 million. About 60 percent of the Internet host computers are located in the United States. Nine out of 10 Internet users today are English-speaking. No fewer than 82 percent of home pages (web sites) are in English, according to the Internet Society’s survey of 60,000 computers with Internet addresses.
Yet some foresee an end to this electronic hegemony. The number of non-U.S. Internet users is about to outnumber those inside the country soon and increase by nine-fold over the next five years, from 16.4 million in 1997 to 143 million by the year 2002, representing an annual growth rate of 70 percent. In that case, the present practice of conducting business, presenting news and information, and performing discussion on the Internet will have to be drastically changed. The widespread use of English will eventually be contested and the Internet itself will become multicultural. This is already happening.
A consortium of American computer companies has developed a universal digital code known as Unicode to allow computers to represent the letters and characters of virtually all the world’s languages. Major search engines like Yahoo and Excite offer their services in multiple languages. Netscape Communications in partnership with the leading Latin American Internet service, Star Media Network, provides a free Internet guide in Spanish and Portuguese. Internet services in languages other than English, like Star Media, are starting to provide world and regional news, weather, stock listings, e-mail, chat rooms, Internet access and more, all in the users’ native language.
Given such developments, optimists argue that far from ending diversity, the Internet will promote it by allowing even small groups of people to disseminate their messages worldwide. By overtaking the “middle range” languages, it may actually protect minority languages threatened with extinction. A wider range of languages on the Internet means at least in theory that a wider range of ideas will be exchanged in a cyberspace, the long-promised global village.
Despite a tremendous influx of non-English languages in recent years, however, the Internet has a long way to go before it becomes a truly multilingual medium. As long as English can be understood by the largest number of Internet users, the cyberspace will continue to be dominated by English as the primary language for international discourse and commerce, European languages as a tool for regional and specialized communication, and many other minor languages for local communication.
At a glance there is an advantage to having one language dominant in the cyberspace. In a world of five to eight thousand different ethnic groups who reside in approximately 160 nation states speaking 5,000 distinct languages, some language must be the common language of the Internet. Many people believe that English must be by default the standard language on the Internet, much as it is the international language of aviation. At present there are clear indications of English becoming a lingua franca in the cyberspace.
The Internet, the network of networks, is most likely to be governed by the logic of the “Metcalfe’s Law,” the so-called magic of interconnections. Connect any number of machines, ‘n’, whether computers, phones or even cars, and you get ‘n’ squared in exponential value. Simply put, the values of the interconnected entities increase exponentially. To expand this logic further, as English-speaking countries have happened to take the central role in world politics and economy in modern period (namely, Pax Britannica and Pax Americana), the English language will deepen its hold on the world as more people go on line. As a recent British Council report shows, the evolution of the English language accelerates as it spreads beyond Anglo-Saxonism. The critical mass of English amongst providers and users is likely to drive the usage inexorably towards monoglot English. Even the French, famously fastidious about guarding their language against dilution by foreign words and phrases, has been forced to surrender to American vernacular on Internet matters.
As English continues to dominate the cyberspace with greater intensity and speed, global resistance to it accelerates with equal strength. Some countries already disgruntled by the encroachment of the American culture—from pop music, blue jeans, to videos—are worried that their cultures will be further eroded by an American dominance in the cyberspace. In their minds, English, by association, immediately evokes a negative image of colonial imperialism. They feel that it may threaten their cultures, their languages, even identities. For this ideological reason, many people in the Third World countries are opposed to having English as the lingua franca for commerce and trade, crisis management, and scholarly and intellectual discourse on the Internet. While the Internet is awash in information, almost all of it flows in one direction, i.e., from the United Sates to the rest of the increasingly wired world. It brings home the age-old controversy regarding unbalanced and unidirectional flow of information from the world’s richest countries to the Third World, an on-going controversy that prompted UNESCO to declare a New World Information Order (NWIO) on several occasions.
For the ideological as well as practical purposes, many countries have taken steps to protect the Internet from excessive American influence. The French government once filed a lawsuit against a Georgia Tech campus in Metz, France, whose web page did not translate its contents into French. Arab countries proposed creation of an Arab Intranet, a closed network in which indiscriminate access to pornography and political discussion could be blocked. Canadian content requirements are to be extended to the Internet to keep out the American “cultural vulture.”
Korean experience American culture has continued to exert its influence in Korea, first through Christian missions, trade and Hollywood productions, and more recently via the Internet. Thanks to the Korean government’s aggressive drive, an estimated 2.5 million Koreans are linked to the global medium. An equally large share of the credit for the phenomenal growth of the Internet goes to the media organizations in Korea, which have in recent years vigorously campaigned for introducing the “Net” to the school community through such programs as the KidNet (Chosun Ilbo) and Internet in Education (Joongang Ilbo).
In the vicissitudes of hopes and fears about the Internet, two isolated but interrelated events recently took place in Korea, all with a nationalistic overtone. One is the Microsoft’s abortive attempt to buy out a local Korean software company in exchange for a multi-million dollar investment. The other has to do with a somewhat radical proposal by a Korean writer to adopt English as a common language on the Internet.
Microsoft triggers off Korean nationalism In the midst of currency crisis in Korea these days came the shocking announcement that the Hangul and Computer (H&C) Company, a Korean software company, had agreed with Microsoft to stop development of its popular Arae A Hangul word processor program in exchange for a $20 million capital investment. Apparently, what Microsoft hoped for in the deal was to secure a monopolistic right of its Korean language version of Windows and Word in the Korean market. When it first appeared in 1990, Arae A Hangul-- named after the Korean alphabet invented by King Sejong more than five centuries ago--was the only word processor based on hangul instead of Roman letters with the hangul mapping. Arae A Hangul, “insanely great” in the words of Apple Computer’s founder Steven Jobs, was multilingual at a time when DOS versions of Word and WordPerfect barely went beyond ASCII. This made it easy to develop a number of attractive fonts frequently used in displaying old Korean characters.
Media reports of the deal aroused anger and frustration among Korean computer users and non-users alike. Defenders of Hangul, who were opposed to the Microsoft’s planned equity participation in H&C, quickly invoked the issue of Korean national pride to rally the public support for the program. A massive nationwide campaign saved Hangul at the eleventh hour from the threat of extinction as its developer, H&C, reversed its earlier decision to sell off the software, bowing to a nationwide campaign spearheaded by Korean scholarly community and venture capitalists.
As hindsight, invoking Korean pride as a defense against predatory foreign companies appears to be too far-fetched and even xenophobic. On the other hand, Microsoft’s aggressive marketing strategy to effectively terminate H&C’s further development of Hangul can be viewed as too selfish and insensitive to the Korean nationalistic sentiment.
English as a common language? Now that 2.5 millions of Koreans are linked up to form a global community, how will they communicate? Majority of Korean Internet users was found to use the global medium mainly for communication within the national boundary, according to an Internet user survey. To most of them the Internet remains largely an unexplored reservoir of knowledge and information due to language barrier. Their grasp of English is quite rudimentary, sufficient only for processing basic information such as the weather, sports, and erotic visuals. Only a fraction of the Internet users can, to whatever degree, comprehend and produce written or spoken utterances in English. As a result, they are denied of the tremendous opportunities that the Internet has to offer, namely, in-depth information and more serious discussions with the netizens all around the world. To be an effective Internet user one has to be equipped with both receptive and productive abilities in English. Such an expectation, however, appears to be neither practical nor realistic as far as Korea is concerned.
Bok Koh-Ill, a professional Korean writer, thinks otherwise. He came up with a radical idea to make English as an official language of Korea along with Korean. Touching on the sensitive theme of Korean nationalism, Bok argued that “putting ‘emotional’ nationalism under control is not enough.” “As the world is rapidly moving toward a ‘Terrestrial Empire,’” he further asserted, “not only the political supranational organizations such as the WTO and IMF but also the communication tools like the Internet are increasingly integrated into one unity.” Now comes his argument concerning English, the dominant language on the Internet. Bok observed that “the emergence of the universal language is interlocked with the enlargement of network.” He predicted that “the Korean language will soon be little more than a museum piece as more nations will adopt a bilingual system in which their own languages coexist with English.” His underlying perspective about language is that “language is a tool, and worshipping it as an idol is irrational.”
Sensationalism notwithstanding, Bok’s audacious proposal immediately touched off a flurry of verbal battle between the conservatives and liberals in Korea. In a wild exchange of high-pitched diatribes, Bok’s opponents argued that English was not panacea and accepting his idea would wither Korea’s unique traditions and hamstring the very cultural foundation. Bok’s supporters applauded accuracy of his observation that Koreans possess an excessive national pride, which Bok sees as the obstacle to the advancement of the Korean society. Bok even dared all his critics by asking, “If you had to choose between English and Korean for your child’s native language, which one would you select? If you choose English, your child is sure to have faster access to cutting-edge technology and information. If the Korean language is your choice, your child is doomed to lag behind in competition with others.” Whether he will succeed in his effort is far from certain. Bok says, “In our society, nationalism and national language are far too sensitive issues to be discussed with composure.” Yet his other key point—that Koreans are too sensitive to these kinds of issues—seems not that far from reality. A non-random online survey taken immediately after the verbal battle showed 58 percent of those polled were against Bok. Even more noteworthy is the fact that as high as 42 percent sided with Bok and his idea. The deepening dispute seems here to stay, at least for a while.
Technological aspects of multilingual translation For a truly multilingual Internet, the long-promised global village, to come of age, there are a host of difficult hurdles to overcome, including, among the many, technical difficulties of communicating in the majority of the world’s languages and development of hardware and software for machine-aided translations. Some progress has been made, and more in sight, in the development of hardware and software for processing texts, from seven-bit ASCII to ISO-Latin, and more recently, to Unicode (ISO 10646), a coding scheme for characters of most of the world’s scripts. Widely hailed as a significant breakthrough in electronic communication worldwide, Unicode character encoding, however, may not be the best one to use in our environment. For example, the Software Laboratory of Nippon Telegraph and Telephone Corporation (NTT) found that Chinese, Japanese, and Korean (CJK) ideographs share the same code space, when Unicode is used in a global search-engine context. Thus, if a Japanese searcher inputs a string for searching, it can equally match against Chinese and Korean counterparts. Another major problem with Unicode for CJK users is that it doesn’t contain enough code space to capture all ideographs. For the development of a perfect multilingual architecture, incorporating languages that build compound characters or right-to-left ordering needs further tests.
Transmitting a message in a language of reader’s preference is one thing; one’s ability to comprehend it is quite another. In order to send and receive a message, most readers will have to rely on human translators/interpreters. The Internet is home to many language translation sites that offer everything from simple online dictionaries to e-mailed translation services. From the desktop, one can request a translation by selecting to pay for human translation. The global nature of the Internet has proved a boon to translation services, such as TAR Communication in New York, which translated Web-based press releases into 28 languages during the Atlanta Olympic Games in 1996. Translation business via the Internet within the next five years is expected to account for 30 percent of their work.
Given the volume and variety of messages on the Internet, however, exclusive reliance on human translation appears to be an unrealistic proposition. It is too slow and costly to make it a sensible choice for maintaining multilingual websites. A viable alternative is machine-aided translation, which has been vigorously pursued mostly in Japan, Canada and Europe with somewhat mixed results. Systran, available on the French Minitel network since 1983 and Canada’s Meteo system, which translates meteorological bulletins between French and English, are considered to be success stories.
Korean translation softwares In Korea there are several translation softwares capable of machine-translating foreign language texts into Korean. In an attempt to assess the quality of these softwares, the author had a wide variety of texts in Japanese and English translated into Korean by means of King Sejong, one of the most popular translation softwares in the market.
To begin with a conclusion, the Japanese-Korean translation effort proved to be a resounding success, whereas the translated version of the English texts was dismally unsatisfactory. In terms of the syntax and grammar both Korean and Japanese share a lot of things in common, whereas there is a mile of difference between Korean and English. It seemed as if the intercultural difference involved in the English-Korean translation were an insurmountable wall indeed. The English texts included Abraham Lincoln’s Gettysburg address, a USA Today article headlined “White House Loses Round in Lindsey Case,” a passage in Jean Baudrillard’s La Pensee Radicale, excerpts from the novel Little Prince, and a paragraph in Nico Randeraad’s article “Authority in Search of Liberty.” This mini exercise yielded disconcertingly unsatisfactory results. The translated versions were totally unintelligible, and not good enough to be revisable even with the intervention of human editors. It appeared that the texts replete with sophisticated literary expressions—the Lincoln’s address and Baudrillard’s work in particular—simply defied machine-aided translation. Of the five different kinds of the English texts, only the USA Today article—concerning a U. S. court’s refusal of the White House appeal to block prosecutors from questioning a presidential aide about Monica Lewinsky--showed some sign of hope. Compared with the other texts, the news article revealed a relatively high level of fidelity in translation.
Why did the newspaper article alone fare reasonably well, while other texts simply failed to be translated into Korean? The translator’s familiarity with the subject matter, i.e. President Clinton’s sex scandal, could be an important factor. A more plausible explanation lies in the journalistic style in which the article is written. Journalists are required to adhere to the journalistic principle of brevity and precision when they write about an issue or event. All the important elements are to be clearly spelled without excessive literary ornamentation.
Such a finding could have a practical implication for the information producers on the Internet. For the benefit of a large number of population worldwide they are strongly encouraged to provide a brief and concise abstract of the full text in a machine-friendly manner, so that the abstract can be translated into many languages without losing fidelity. Also useful in multilingual translation will be a two-step process now being implemented by several groups including the United Nations University in Tokyo. In the two-step process, a text is first thoroughly analyzed into component parts (title, paragraph, sentence), clarified when necessary and possible by a dialogue with the author, then translated into an intermediate, abstract representation—which is used to generate translations in different languages. In this way, readers who have no receptive ability in a foreign language could at least get the gist of the material in hand with the help of translation softwares.
Human-assisted machine translation Given the low quality of machine translation and the expensive nature of human translation, the only logical option available to Korea--and for that matter to many countries in the world--appears to be human-assisted/validated machine translation. Experts at numerous regional and international conferences have already addressed to the crucial importance of human-assisted translation. To briefly summarize their recommendations and suggestions: (1) it is vital that a human validator is used to correct automatic translation; (2) it is important to take cultural-specific notions into account when undertaking translations between culturally different linguistic areas; (3) human translations can benefit from processing of files against terminology databases to ensure that technical phrases are translated in a domain-specific way; (4) as keyword matching does not work across domains or across languages, searchable concept-based terminology resources and thesauri are needed; and (5) it is important to integrate machine translation and domain-specific terminology sets with authoring tools to speed up translation services.
Technological solution to machine translation is one thing, financial difficulty associated with making machine translation services available through the Internet is another. Because these services take up too many CPU cycles, Information Service Providers would rather offer these services via specialist servers, not as a part of their mainstream operations. The question then arises as to who pays for such servers, and how. Also, the preponderance of English on the Web, with some estimates ranging as high as 95%, appears at odds with the huge investments in translation to be made by the software industry. Investments are likely to be made in the cost-effective European languages, but not in minor languages.
Preparing for the multilingual Internet calls for a concerted effort by both public institutions and industry players that produce and utilize language services, tools, and systems. Regional cooperation among nations that share cultural and linguistic similarities, such as Korea, Japan, and China, must be strongly encouraged not only by the governments concerned but also by regional and international organizations such as UNESCO. Through collaborative R&D arrangements, they can jointly develop multilingual translation tools and services in a much more cost-effective way.
From ephemeral to robust cyber community As personal computers become less expensive and user-friendlier, people outside North America and European countries are becoming increasingly linked. In this atmosphere diversity of languages will further enrich the on-line environment by making it possible for people of different cultures and languages to engage themselves in more serious in-depth discussions with other people. There are many advantages of a multilingual Internet: it would allow much more effective and wide diffusion of information and knowledge than would otherwise be possible; common mistakes and misunderstandings resulting from language barrier will be curtailed; and it could be a crack in Americana hegemony over Internet culture.
A multicultural and multilingual Internet has the potential to be simultaneously more universalistic and more particularistic, more global and more local, more cosmopolitan and more parochial, and more mass-oriented and more elite-oriented. The cybersociety created by the Internet as we know it today is drawn heavily from the rich and educated in the rich countries. By appearing to be global and universalistic, the Internet masks its very particular and elite characteristics. If it were allowed to continue to develop the way it had been, the global medium would fall a victim to its own success by further benefiting the rich and well-educated at the cost of the underprivileged.
The Web is an ephemeral community, to borrow words from Jim Falk, which is “unstable and transitory.” “Relationships within an ephemeral community, whether emotional or intellectual, are likely to be partial, satisfying only one or a few of the members’ needs.” In contrast, Falk maintains, a robust community is one in which “the members have not only a sense of interrelatedness and shared experience, but also share common ideals and believe that through by virtue of belonging to their community they can make great progress towards achieving their objectives than through belonging to other communities. Members will invest personal resources, energy and commitment into it because they consider it stable, growing, supportive and effective.” Multilingualism on the Internet is a necessary, if not the sufficient, condition for transforming an ephemeral cybersociety into a robust one.
REFERENCES Caldwell, B. (1997, June 30). In any language, it's still Cobol. InformationWeek, issue 637, p12,
Casselman, B. (1995). Our home and native tongue: A linguist's notes on the origins of
distinctively Canadian terms. Canadian Geographic, 115(6), p22.
CEN/TC304-Character Set Technology Workshop (1996, November) Providing multilingual
support in middleware: Implementing the universal character set ISO 10647 in the European
information society. (http://www.e5.ijs.si/il8n/ws-bled.html).
Coleman, F. (1997, April 24). A great lost cause: France vs. the Internet. US news & World
report, 122(15), p57, 3p, 1c.
El-Khodari, N. (1997). Is multilingualism important to Canada? (http://www.uottawa.ca/
Falk, J. (1996, November 5). The meaning of the web.
Fouser, R. (1998, June 24). Why Hangul is 'insanely great'. The Korea Herald.
Grewal, S. (1996). Networks and the worldwide web: Multilingualism on the worldwide web.
KIUSE: Korea Internet User Survey for Everyone (1998, June 16).
Korea Time staff (1998, July 23). 'English as official language' sparks intense debate. Korea
Time, 12. (http://www.kpi.or.kr/kinds-cgi).
Kwon, R. (1997, October 7). It's a small world after all. PC Magazine.
Nader, R. (1995). Virtual's search for reality on dictionary landscape. Public citizen, 5(3), p12.
Oudet, B. (1997 March) Multilingualism on the Internet. Scientific American, special report.
Pollack, A. (1995, August 7). A cyberspace front in a multicultural war. The New York Times.
Scientific American editors (1997, March). The Internet: Bringing order from chaos. Scientific
American, special report.
Shimizu, S., Kambayashi, T., Sato, S., & Francis P. (1997). A framework for multilingual
searching and meta-information extraction. (rodem.ingrid.org: 8080/inet97).
Stanko, G. (1997, March 10). Cultural imperialism online? Journal of Commerce, 411(28926),
The multilingual information society (MLIS) (http://www.isi.gov.uk/isi/europe/mils.html).
The multilingual information society (MLIS) work program for three years 1996 - 1998.
Troutener, J. (1995). Easing into the Internet. Technology & Learning, 16(3), p12.
Web Internationalization & Multilingualism Symposium (1996 November). Social, political and
cultural aspects. (http://www2.echo.lu/oii/en/w3c-int.html).
Young, N. (1997, December 29). Cultural imperialism aside, English spans linguistic gulfs.