Faculty of Information Quarterly Vol 2, No 3 (May/June 2010)

Babel 2.0: Bridging ostensibly global communications medium. Information and the Linguistic communication technologies (ICTs) in general tend to favour English, Digital Divide creating a "linguistic digital divide" wherein even those with access to technology find difficulties expressing themselves in their native languages. By: Brad Koegler The ASCII character standard creates difficulties for those using non-Latin Brad is a Master of Information (MI) characters, and solutions from candidate at the University of Toronto. standards organizations such as He hails from Waterloo, Ontario, ICANN have been both slow and Canada and holds a Bachelor of Arts piecemeal. Proprietary software (Honours) in Sociology from the makers have similarly lagged in University of Guelph. He has worked as localization efforts and are only now a bookseller, a residence assistant, beginning to catch up to open-source and a music journalist. His areas of projects. Machine translation has interest include language and advanced rapidly, but is still not a translation, cyberculture, social effects viable solution. In the meantime, of new media, and the information English's incidental supremacy makes economy. Plans for the near future it an ideal Internet lingua franca, but include writing a Master's thesis on the increased accessibility in other native linguistic digital divide, then pursuing a languages will benefit all. doctorate in Information Studies. [email protected] “La parole humaine est comme un chaudron fêlé où nous

battons des mélodies à faire danser les ours, quand on Abstract voudrait attendrir les étoiles.” (Language is a cracked kettle on which we beat out tunes for bears to dance to, The Internet's bias toward the while all the time we long to move the stars to pity) is not surprising -Gustave Flaubert, Madame Bovary considering its origins as an American military network, but it is a highly problematic access barrier for an

Page 1 of 10 Faculty of Information Quarterly Vol 2, No 3 (May/June 2010)

world to the Internet has revealed another Introduction digital divide, one which “mirrors traditional The rise of the information society and patterns” of division and oppression (Hindman, its backbone, the Internet, has been startling in 2009, p. 9). Even once citizens begin to use its speed and scope. The military programmers computers and the Internet, the same old who set up the first ARPANET packet- divisions and hierarchies remain. Nowhere is switching network in 1968 could hardly have this more true than in terms of language. The imagined its present size and ubiquity; in very infrastructure upon which the Internet is developed nations, it is now almost difficult to built is profoundly biased in favour of English participate fully in society without Internet content, at the expense of users wishing to access, and even developing countries are seeing communicate in other languages. This paper will notable increases in the use of information and examine the two most common manifestations communication technologies (ICTs). In 2009, of the linguistic digital divide: native-language and Sweden went so far as to declare support in software, and “ASCII imperialism” broadband Internet access to be an inalienable on the Internet (Pargman & Palme, 2009, p. right for their citizens (Stross, 2009, ¶ 4). The 197). abundance of such technology has meant democratization of the means of production in Linguistic Imperialism or what Benkler (2009) calls the networked Lingua Franca? information economy, “an economy of When unpacking and analyzing the information, knowledge, and culture that flow linguistic digital divide, it is important to stress through society over a ubiquitous, decentralized the difference between linguistic imperialism network” (p. 1246). and the rise of a lingua franca. Discourse about A dominant concept in the discourse the ascendancy of English often carries a surrounding this paradigm shift has been the cynical, suspicious tone, but Anglophone idea of the “digital divide,” a nebulously-defined supremacy in ICTs came about more as a matter problematization of the gap between those who of convenience than one of wilful oppression. use ICTs and those who cannot or will not. There are several possible outcomes Through government and corporate initiatives – when groups who do not share a mother tongue whether in terms of universal service like in attempt to communicate. If one group is Spain and Sweden or universal access like the dominant, the other might – whether willingly Bill Gates-fuelled public library project in the or coerced – learn the more powerful group’s United States – great strides have been made language. Two groups of more or less equal toward bridging the digital divide, at least as it is standing without any languages in common defined in the discourse of the neoliberal state might craft a “pidgin” language that blends the (Stevenson, 2009, p. 18). However, bringing the two mother tongues. However, if the groups are

Page 2 of 10 Faculty of Information Quarterly Vol 2, No 3 (May/June 2010)

lucky enough to share a common language, they communicate with other groups cannot use their will most often use that lingua franca (from the own languages to communicate amongst Latin for “Frankish language,” in reference to an themselves. Italian-based pidgin used by traders and Whither Localization? diplomats in medieval times) to communicate Musinguzi (2008) discusses Kiganira (Chirikba, 2008, p. 31). Kijambu, a Ugandan entrepreneur whose English has been a global lingua franca biography is a techno-utopian NGO’s dream: for upwards of a century. The supremacy of the using an ICT training centre near his home in British Empire, then the dominance of the Mayuge District, he was able to receive training United States, spread the use of English all as an accountant and use e-commerce around the world. It is one of six official technology to triple his farm’s profits (¶ 2). Mr languages of the United Nations (along with Kijambu’s accomplishments are impressive, but , Chinese, Russian, Spanish, and French), he had a decisive advantage: unlike all but ten and one of three working languages (along with percent of his fellow Africans, he speaks English French and German) of the European Union (Musinguzi, ¶ 6). (UN DGACM, 2002, ¶ 1). It is perhaps no Mr Kijambu’s mother tongue, LuSoga, surprise, then, that English has also become the is not supported by any of the major computer lingua franca of the Internet. Native speakers of operating systems, and is all but absent from the English only make up twenty-nine percent of Internet. This creates a vicious circle against the the wired world, and this number is dropping implementation of ICTs: lack of utility leads to rapidly – Chinese speakers will have overtaken lack of demand, which leads to a lack of English by the end of the next decade – but the incentive to develop software and content for content of the Internet remains four-fifths LuGosa, which only reinforces the lack of English (Currie, 2007, ¶ 3). demand by LuGosa speakers for computers Despite reflexive hand-wringing about with which they cannot use to effectively yet another expression of Western dominance, interact. While online business has allowed Mr having a common language on the Internet Kijambu to increase his business threefold actually serves to benefit its users. The (Musinguzi, 2008, ¶ 3), broader language Ethnologue language database counts 6 909 support would not only further improve his languages currently in use (Lewis, 2009, Table business, but would also allow his non- 1); to resist the use of a lingua franca would Anglophone peers the same kind of reduce the Internet to a new Babel. English’s opportunity. pre-existing ubiquity, however ill-gotten, makes Keniston (1997) defines “localization” it the logical choice. Allegations of linguistic as “the translation of programs originally written imperialism become more serious, however, in and for one language into intelligible and when the users who use English to

Page 3 of 10 Faculty of Information Quarterly Vol 2, No 3 (May/June 2010)

user-friendly versions in and for another programs that operate well in one language, with language” (¶ 14). In the early days of one set of characters, crash in another; non- programming, this was done as an “add-on,” if English speaking users may be presented with at all; software would be coded entirely in unintelligible, awkward, or English-language- English and, if unavoidable, a localizer might be only help files, tutorials, or documentation; and hired to translate the menus and manual into so on (Keniston, 1997, ¶ 16). Spanish or Japanese. The culture of early Meanwhile, open-source localization, computing was such that Linus Torvalds, a even in an established system, is as simple as Swedish-speaking Finn and the creator of Linux, editing and recompiling the code. The greatest did and still does all of his code comments and strength of FLOSS is the “L” for “libre," what annotations in English for simplicity’s sake Stallman (1993) defines as “free as in speech, (Torvalds, 1997, ¶ 1). In effect, the Anglophone not as in beer” (¶ 2). According to the Free “technology capitals” of Redmond, WA and Software Foundation’s definition of free Cupertino, CA were dictating the information software: policy of the world at large. You should also have the freedom to make modifications and use them Now, however, to only release a privately in your own work or play, program in one language is to greatly diminish without even mentioning that they exist...In this freedom, it is the user’s its appeal. Even Microsoft, historically a laggard purpose that matters, not the in terms of localization, has greatly broadened developer’s purpose; you as a user are free to run a program for your its language support; Windows XP supports purposes, and if you distribute it to over 140 languages, and the nascent Windows 7 someone else, she is then free to run it for her purposes, but you are not currently supports thirty-eight, with plans to entitled to impose your purposes on support “more than ever before” within the her...Freedom [one] includes the freedom to use your changed version in next year (Bennett & MacConnell, 2009, ¶ 34). place of the original” (Free Software FLOSS (Free/Libre Open Source Free Software Foundation, 2009, ¶ 6- 10). Software) tends to be more nimble in this The autonomous region of , regard. The source code for proprietary software Spain has been highly successful in using ICTs is rigid and uniform; adding support for other to improve their economic standing, and in languages later requires jury-rigged service packs using FLOSS to work around the linguistic and add-ons that often cause as many problems digital divide they might have otherwise faced. as they solve. Anyone who has tried to work Extremadura was the poorest region in Spain, with foreign-language translations of American predominantly agricultural and facing an exodus programs – be they operating systems or of its youth to more prosperous and/or urban applications – can attest that the problems regions. In 1997, the region’s president launched defined here as "technical localization" are far a “Regional Strategy on Information Society” from universally solved. Absurdities abound;

Page 4 of 10 Faculty of Information Quarterly Vol 2, No 3 (May/June 2010)

that examined how ICTs might be deployed to Perhaps most attractive, though, was the price. achieve “accessibility for all; Internet as a public The entire software project only cost €193 service; and stimulation of technological 562.84 for both development and deployment, literacy” (Van Disseldorp, 2008, ¶ 6). With the while the Spanish Ministry of Education, help of funding from the European Science, and Technology estimates that a Commission, the Extremaduran government comparable proprietary solution would have launched eExtremadura, a regional intranet with cost €30 million, and still might not have had three major components: “knowledge centres” the same degree of customization as gnuLinEx for promoting information literacy; a technical (Disseldorp, ¶ 24). As a result of eExtremadura, network linking schools, hospitals, and other the economic gap between Extremadura and the public buildings; and Vivernet, a “business rest of Spain has narrowed, the “brain drain” incubator” designed to encourage has slowed, the local quality of education and entrepreneurship (Van Disseldorp, ¶ 8-11). healthcare has improved, and many new This project would have been impossible, businesses and industries are starting up (Ghosh or at the very least prohibitively expensive, were & Schmidt, 2006, p. 4). it not for the use of open source software. Ghosh and Schmidt (2006) are Rather than approaching a proprietary software especially adamant about FLOSS as a tool for firm, the Spanish government hired bridging divides: programmers to download a copy of the Debian [The] link between open source and the rise of small ICT businesses is distribution of Linux and customize it for the especially important given the tendency needs of the Extremaduran people (Disseldorp, of proprietary vendors to ignore local needs, especially in developing 2008, ¶ 16). The resulting operating system, regions...Developing countries need to gnuLinEx, was customized far beyond what avoid being locked out of acquiring skills and competencies. The adoption would have been possible with closed software: of open source policies provides it is localized in Spanish and various other environments that promote skills development and the ability to create Extremaduran (many (p. 3). speak uncommon dialects of Portuguese, or the An initiative like eExtremadura might not endangered ), and the necessarily be exactly the right one for other software is all specially configured in developing regions like Mayuge District, conjunction with local laws and standards. Even Uganda, but it is difficult to argue that – for the programs have names particular to both cultural and economic reasons – any sort Extremaduran culture; the word processor, for of ICT solution should not be open source. example, is named for the famed local poet Espronceda (Disseldorp, ¶ 21).

Page 5 of 10 Faculty of Information Quarterly Vol 2, No 3 (May/June 2010)

ARPANET, ICANN, and the umlaut with an “e” after the bereft letter (e.g. the author of this paper is “Koegler” in ASCII Canada but “Kögler” in Germany). However, The near-anarchic democracy of today’s such solutions are not always possible, feasible, Internet belies its origins as a tightly secured or even desirable. German only needs to replace tool of the establishment. Before there was an the umlaut to conform to ASCII. Chinese Internet, there was ARPANET, a packet- speakers would have to discard their writing switching network connecting military computer system altogether. databases in the United States. In those early Of course, anyone who has used the days, little thought was given to the idea of Internet recently is well aware that websites can needing to use other languages, so the architects display non-ASCII characters ranging from of ARPANET did not bother programming Chinese and Cyrillic characters to smiley faces foreign-language support. The international and hearts. The left sidebar of Wikipedia is full standard for printing characters, ISO 646, had of links to other languages’ Wikipedias, rendered various international versions that allowed for in their own scripts. However, all of this is some non-English characters, but it was the artifice layered on top of the same ninety-five American variant (American Standard Code for ASCII characters that the Internet began with. Information Interchange, or ASCII) that Fonts like Unicode can support millions of became the backbone of the Internet (Pargman characters, more than in all the world’s scripts, & Palme, 2009, p. 183). but the information travelling between Nisselbaum and Friedman (1997) computers must first be translated in and out of enumerate three different categories of bias in ASCII. The German spelling of my name, typed computer systems: inherent, technical, and into an e-mail in HTML, would be transmitted emergent (p. 25). The use of the ASCII standard as “Kögler.” This seems harmless enough is a case of the third; the architects of the at first. Ghosh and Schmidt (2006) discuss the Internet were designing a national idea of a “natural monopoly,” a technical communication system, not an international standard that, while still monopolistic, is seen as one, and did not anticipate the later need to a lesser evil because it increases network accommodate all of the world’s languages. The externality (p. 3). We accept restrictions and problem, however, has become an incredibly regulations for things like rail gauge and wireless difficult one to fix. Problems with character bandwidth because we value the ability to use standards are as old as the printed word. our devices anywhere, prioritizing “how well it German characters bearing umlauts, for example interworks with other stuff [rather] than its own (ä, ö, ü), were uncommon on typewriters and intrinsic merits” (Levien, 1998, ¶ 4) It is difficult telegraph keyboards outside of Germany, so to imagine a functioning global Internet without Germans developed a convention of replacing a similarly strict transfer standard, but ASCII

Page 6 of 10 Faculty of Information Quarterly Vol 2, No 3 (May/June 2010)

simply does not handle foreign languages well. Internationalized Domain Name Fast-Track Pargman and Palme (2009) cite a study wherein Process, wherein they will now accept domain the translation company WorldLingo sent names that contain non-ASCII characters. customer service e-mails written in Japanese Despite all the fanfare, this was too little, too script to forty Fortune 500 companies (p. 179). late. The international domain names (IDNs) Only one replied in Japanese; most didn’t reply are encoded at the user level using a system at all. The remainder of the few who did reply called Punycode, but most of the said that the messages they received were aforementioned problems still apply: the “corrupted.” These ostensibly international Punycode will still be sent out to the Internet as conglomerates did not have support for ASCII, with the hope that the computer at the Japanese characters on their customer-service other end will be able to convert it back department computers and were baffled by the (Klensin, 2004, p. 2). It is perhaps telling of raw code they saw instead. ICANN’s level of commitment to The technology to fix this exists. The internationalization that, at the time of this Internet Corporation for Assigning Names and writing, the press release for the IDN Fast- Numbers (ICANN) and World Wide Web Track Process (2009) is still only offered in Consortium (W3C) could change the standards English. and begin using Unicode, allowing speakers of Conclusion all languages to interact with the Internet As with many other facets of the directly instead of through the artifice of ASCII. Internet, Google has been a loud and powerful It would be a costly process, and problematic presence in the field of computer translation. for the oldest computers, but the long-term Their Google Translate service uses a twenty- benefits on the world stage would far outweigh billion-word corpus of public-domain United the short-term troubles. Unfortunately, these Nations documents, statistical analysis, and “arcane standards boards,” once bastions of crowd-sourced corrections in an attempt to techno-utopians and open-source activists, have provide instantaneous translation of any page on been co-opted by “an influx of people from the Internet (Tanner, 2007, ¶ 2). It is arguably large commercial companies” (Pargman & the best such service available right now, but it Palme, 2009, p. 192). The inherent conservatism still has a long way to go before it can be used of committees, combined with reluctance on the for much more than approximations. part of the Microsofts and Googles of the world In the meantime, there will be a place to have to adapt their software, means that an for English as the lingua franca of the Internet. Its epochal shift to correct the emergent bias is current ubiquity makes it likely that it will highly unlikely. maintain this position for quite some time. Small concessions have been made. On However, ICTs in general and the Internet in October 30th, 2009, ICANN announced the

Page 7 of 10 Faculty of Information Quarterly Vol 2, No 3 (May/June 2010)

particular must expand support for other languages. All of the rhetoric about ICTs as new cornerstones of democracy is useless as long as there are people who cannot use those technologies in their own languages.

Page 8 of 10 Faculty of Information Quarterly Vol 2, No 3 (May/June 2010)

References

Benkler, Y. (2009). Freedom in the commons: Levien, R. (1998). The decommoditization of Toward a political economy of information. protocols. Retrieved November 20, 2009 from Duke Law Journal, 52, 1254–1276. http://www.levien.com/free/decommoditizing. html. Bennett, J., & MacConnell, J. (2009, July 17). Engineering Windows 7 for a global market [blog Lewis, M. P. (2009). Ethnologue: Languages of the post]. Retrieved from world, sixteenth edition. Dallas, TX: SIL http://blogs.msdn.com/e7/archive/2009/07/0 International. 7/engineering-windows-7-for-a- global- market.aspx#comments. Musinguzi, B. (2008, February 19). Mother tongue interference on the internet. AllAfrica Global Media. Chirikba, V. (2008). The problem of the Retrieved from Caucasian Sprachbund. In P. Muysken (Ed.), http://allafrica.com/stories/200802191077.htm From linguistic areas to areal linguistics. Amsterdam: l. John Benjamins. Nisselbaum, H., & Friedman, B. (1997). Bias in Currie, N. (2007, April 10). Culture flows computer systems. In B. Friedman (Ed.), Human through English channels, but not for long. values and the design of computer technology (pp. 21– Wired Magazine, 15. 40). Cambridge, MA: Cambridge University Press. Disseldorp, J. van. (2008). FLOSS deployment in Extremadura, Spain. Retrieved from Pargman, D., & Palme, J. (2009). ASCII http://www.osor.eu/case_studies/floss- imperialism. In M. Lampland & S. L. Star (Eds.) deployment-in-extremadura-spain. Standards and their stories: How quantifying, classifying, and formalizing practices shape everyday life Free Software Foundation. (2009). Free (pp. 177–199). Ithaca, NY: Cornell University software definition v1.80. Retrieved from Press. http://www.gnu.org/philosophy/free-sw.html. Stallman, R. (1993). The GNU manifesto. Ghosh, R., & Schmidt, P. (2006). Policy Brief: Retrieved from Open Source and Open Standards: A New Frontier for http://www.gnu.org/gnu/manifesto.html. Economic Development. New York. Stevenson, S. (2009). Digital Divide: A Hindman, M. (2009). The Myth of Digital Discursive Move Away from the Real Inequities. Democracy. Princeton: Princeton University Press. The Information Society, 25, 1–22.

ICANN. (2009). ICANN bringing the languages of Stross, R. (2009). Broadband now! so why don’t some the world to the global internet. Retrieved November use it? New York Times. Retrieved from 15, 2009 from http://www.nytimes.com/2009/10/18/busines http://www.icann.org/en/announcements/ann s/18digi.html. ouncement-30oct09-en.htm. Tanner, A. (2007, March 28). Google seeks world of Keniston, K. (1997). Software localization: Notes on instant translations. Reuters. Retrieved November technology and culture. Working paper for the MIT 21, 2009 from Program in Science, Technology, and Society. http://www.reuters.com/article/technologyNe ws/idUSN1921881520070328. Klensin, J. (2004). Application techniques for checking and transformation of names. Retrieved Torvalds, L. (1997). Linus Torvalds bio. November 17, 2009 from Retrieved from http://tools.ietf.org/rfc/rfc3696.txt. http://www.linux.org/info/linus.html.

Page 9 of 10 Faculty of Information Quarterly Vol 2, No 3 (May/June 2010)

United Nations Department for General Assembly and Conference Management. (2002). What are the official languages of the united nations? Department for General Assembly and Conference Management FAQ. Retrieved from http://www.un.org/Depts/DGACM/faq_lang uages.htm.

Page 10 of 10