Wednesday, July 07, 1999

Language Networks

Language Networks
Clay Shirky

The 21st Century is going to look a lot like the 19th century, thanks to the internet.
A recent study in the aftermath of the Asian financial storm ("Beyond The Crisis -
Asia's Challenge for Recovery," Dentsu Institute for Human Studies) found that citizens
of Asian countries who speak English are far more likely to be online than those who
don't. The study, conducted in Tokyo, Beijing, Seoul, Bangkok, Singapore, and Jakarta,
found that English speakers were between two and four times as likely to use the
internet as their non-English speaking fellow citizens. Within each country, this is
a familiar story of haves and have-nots, but in the connections between countries
something altogether different is happening -- the internet is creating an American
version of the British Empire, with the English language playing the role of the Royal

This isn't about TCP/IP -- in an information economy the vital protocol is language,
written and spoken language. In this century, trade agreements have tended to revolve
around moving physical goods across geographical borders: ASEAN, EU, OAS, NAFTA. In
the next century, as countries increasingly trade more in information than hard goods,
the definition of proximity changes from geographic to linguistic: two countries border
one another if and only if they have a language they can use in common. The map of the
world is being redrawn along these axes: traders in London are closer to their
counterparts in New York than in Frankfurt, programmers in Sydney are closer to their
colleagues in Vancouver than in Taipei. This isn't an entirely English phenomenon: on
the internet, Lisbon is closer to Rio than to Vienna, Dakar is closer to Paris than to

This linguistic map is vitally important for the wealth of nations -- as the Dentsu
study suggests, the degree to which a country can plug into a "language network,"
especially the English network, will have much to do with its place in the 21st century
economy. These language networks won't just build new connections, they'll tear at
existing ones as well. Germany becomes a linguistic island despite its powerhouse
economy. Belgium will be rent in two as its French- and Flemish-speaking halves link
with French and Dutch networks. The Muslim world will see increasing connection among
its Arabic-speaking nations -- Iraq, Syria, Egypt -- and decreasing connections with
its non-Arabic-speaking members. (Even the translation software being developed reflects
this bias: given the expense of developing translation software, only languages with
millions of users -- standard versions of English, French, Portuguese, Spanish, Italian,
German -- will make the cut.) And as we would expect of networks with different
standards, gateways will arise; places where multi-lingual populations will smooth the
transition between language networks. These gateways -- Hong Kong, Brussels, New York,
Delhi -- will become economic centers in the 21st century because they were places
where languages overlapped in the 19th.

There are all sorts of reasons why none of this should happen -- why the Age of Empire
shouldn't be resurrected, why countries that didn't export their language by force
should suffer, why English shouldn't become the Official Second Language of the 21st
century -- but none of those reasons will matter. We know from the 30-year history
of the internet that when a new protocol is needed to continue internet growth, it'll
be acquired at any expense. What the internet economy demands more than anything right
now is common linguistic standards. In the next 10 years, we will see the world's
languages sorted into two categories -- those that form part of language networks will
grow, and those that don't will shrink, as the export of languages in the last century
reshapes the map of the next one.