The Language Teacher
November 2001
Machine Translation and the Future of English as a Global Language
Michael Cribb
Kansai Gaidai University
Imagine a Japanese speaker who has no working knowledge of English. He is browsing the Internet one day and comes across the New York Times. He navigates to an article in English, clicks the 'Translate' button on his browser and the page refreshes in Japanese. After reading the article, he decides to write a letter to the editor. He composes this in Japanese using his word processor, clicks the Translate button again and the letter is translated into English, which he then emails to the New York Times. A few days later, a reply in Japanese comes back from the editor confirming that his letter will be printed online. The following week, he receives a phone call from a French farmer congratulating him on the letter. The farmer knows no Japanese but the voice message is automatically translated and the two hold a natural conversation for several minutes.
Before you jump to the next article and dismiss me as another one of those crazy 'Nostradamus' soothsayers, please stop to think of the implications of machine translation (MT) and the effect it will have on the future of English as a global language and on our teaching (TESOL) profession. When, not if, such technology arrives -- and the machine translation of written text is already here -- the rise of English as the global lingua franca could be put into reverse and the teaching of foreign and second languages as we know it today could virtually disappear. Non-English speaking countries will be able to roll-up and shed their massive foreign language (mostly English) education programs, and the number of students visiting English-speaking countries solely for language education will drop to a trickle. Language teachers will go home or move into other professions and the world will get on with talking to each other on even terms, leading to true 'linguistic emancipation' (Eastman 1983: 101).
For many, the rise of the English language as the global lingua franca has been seen as the only solution to the increasingly problematic language barrier as the world moves toward globalization. David Crystal has illustrated this rise well in his book English as a Global Language in which he notes how English 'has repeatedly found itself in the right place at the right time' (Crystal, 1997: 110). The suggestion is that English is once again well placed to take advantage of the electronic revolution and establish itself in an 'impregnable' (p. 22) position as the international language. However, MT technology is rapidly developing and recently there has been an explosion of web sites offering translation services (e.g., AltaVista's 'Babel Fish', <www.world.altavista.com>). MT, then, looks like it will be a formidable 'opponent' to the English language in the near future and could well make second language learning redundant.
Interestingly, Crystal (1997) hints briefly in his book at the potential of MT as an alternative to English as a global language and the coming battle between the two:
It will be very interesting to see what happens then -- whether the presence of a global language will eliminate the demand for world translation services, or whether the economics of [machine] translation will so undercut the cost of global language learning that the latter will become otiose. It will be an interesting battle 100 years from now . . . . we may well be approaching a critical moment in human linguistic history. (Crystal, 1997: 22)
Crystal (1997) puts the battle some 100 years in the future but I believe that is a conservative estimate and sometime within the next two decades this battle will materialize in earnest. For example, by 2005 it is estimated that 57.3% of all Internet users will be non-English speaking, a 150% increase on 1999 (Transparent Language, 2000). It is unlikely that the world can afford to sit around for 100 years waiting for a solution.
In this coming battle, there are a number of reasons why MT has an advantage over English. First, there is the well-known progress of computer technology which follows Moore's law -- doubling in speed and capacity every 18 months. In comparison, new breakthroughs in second language acquisition and teaching methodology are few and far between. Nobody really believes that second language acquisition will be significantly quicker or less painful 50 years from now. Second, more and more material is appearing in electronic form. This includes newspapers, books, magazines as well as discussion, and chat groups. As the volume increases, MT technology will be well positioned to take advantage of this.
A third factor comes in to play when we consider the new type of user in the electronic world. In the past, anyone who needed to communicate at an international level (e.g., for business or academic study) usually had several solid years of foreign language study behind them. But the Internet allows rapid and instant access to the global village to just about anyone with a computer and modem. With such a rapid change in the way the world interacts, a new algorithm for the way it communicates is needed. It is unwise to believe that new users with little foreign language competence will embark on long-term courses in English, or that competence will develop simply as a result of surfing the Web. Rather, the companies and organization that form the backbone of the Web will need to take their products to the users in their native language, not only English. That some companies have already recognized this need is evident in the number of major internet companies offering multilingual services to their customers (e.g., Yahoo! <www.yahoo.com>).
No one is denying, of course, that MT is limited at the moment in its ability to produce quality translations between languages and that there are a number of hurdles to overcome. Web companies are all too keen to remind their customers that output from MT is nowhere near perfect, and words such as 'gisting' are common to explain how MT can only provide a general outline of a text to see if it is worth following up. We are likely to see great improvements in quality over the coming years with the help of example-based MT (Brace, 1993) and corpus linguistics. In example-based MT, a bank of bilingual phrases and sentences pairs are stored in memory and accessed when required, thus alleviating the need for word-for-word translation for the more idiosyncratic parts of the language. At the same time, large corpuses of language samples enable powerful, domain-specific collocations to be built regarding the lexical content of a language. These collocations can then be used to determine with high probability the sense of a particular word in any one context.
At the end of the day then, we need to ask ourselves whether it is easier to encode the rules of the English language in silicon-based chips (i.e., machine translation systems), or carbon-based matter (i.e., the human mind). I would suggest that the former is the more parsimonious, logical, and inevitable solution to the "Tower of Babel" problem, even if on an emotional level many teachers of English feel uncomfortable with this.
If MT does put the rise of English as the global language into reverse then it follows that the TESOL profession will see a similar decline. Many of the massive EFL education programs at secondary and tertiary level in Japan could be scaled back and language teachers will move on to other employment. This may horrify many teachers who, especially in the case of English teachers, have been able to travel the length and breadth of the world because of their profession. But change usually does not happen overnight and it often opens up more windows of opportunity than it closes, one particular window being the opportunity to communicate with the world on an equal footing; that is, 'linguistic emancipation' (Eastman, 1983: 101).
Obviously, nobody can accurately predict the future, but, as with Latin and French that were once thought destined to become global languages, we have to be prepared for the time when English too may become redundant in this role. That time may be approaching. In the words of David Crystal (1997: 113): "linguistic history shows us repeatedly that it is wise to be cautious, when making predictions about the future of a language." Or in the words of the AltaVista 'Babel Fish' translator (translated into French and then back into English): "the linguistic history proves to us on several occasions that it is wise to be careful, by making forecasts about the future of a language"!
References
Brace, C. (1993, September/October). Focus on Japan. Language Industry Monitor. Retrieved June 21, 2000, <www.eamt.org/archive /japan.html>.
Crystal, D. (1997). English as a global language. Cambridge: Cambridge University Press.
Eastman, C. M. (1983). Language planning for special purposes. In Language planning: An introduction (pp. 96-102). San Francisco: Chandler & Sharp.
Transparent Language. (2000, January). Machine translation and the power and purpose of "gisting" in the Internet era: A white paper from Transparent Language. Merrimack, NH: Author. Retrieved July 1, 2000, <www.transparentlanguage.com/ets/about/mtwhitepaper.htm>.
All materials on this site are copyright © by JALT and their respective authors.
For more information on JALT, visit the JALT National Website