VoxEU Column Education Politics and economics

Lawrence Summers and the uselessness of learning foreign languages

English is the dominant language of the Internet, business, and world trade. Do we need another? This column applies an economist’s rationale to the question.

“I don't speak English. Kurdish I speak, and Turkish, and gypsy language. But I don't speak barbarian languages.”

“Barbarian languages?”

“English! German! Ya! French! All the barbarian”.

—Yasar Kemal, a Turkish writer whose words are quoted by Paul Theroux in The Great Railway Bazaar

In 2005, Larry Summers, then President of Harvard University, outraged 50% of the world by claiming that women are not as talented as men in science and mathematics. This time, he has outraged some 94% of the world’s population by suggesting that native speakers of English should forego learning other languages since, anyway, the rest of the world will soon become fluent in English (Summers 2012). English is indeed the native language of some 400 million people (that is 6% of the seven billion living on our planet) and is “spoken” by another billion, whatever “spoken” means (Crystal 2001).

Language is an essential expression of culture (and culture is, according to the Sapir-Whorf hypothesis, shaped by one’s native language). Read Shakespeare’s Richard II to see what happened to Thomas Mowbray whom King Richard exiled to Venice (“Have I deserved at your Highness’ hands/The language I have learn’d these forty years/For my native English, now I must forego;/And now my tongue’s use is to me no more/… What is thy sentence, then but speechless death,/Which robs my tongue from breathing native breath?”). Check Fernando Pessoa and his “my homeland is my language”, or Ngugi wa Thiongo: “The choice of language and the use to which language is put is central to a people's definition of themselves in relation to the entire universe” (Thiongo 1986).

Remember Sri Lanka, and the many lives it cost because one language group (out of the two main ones) decided that its language would become “more important” than the other. The reverse is also true, and the European Union is ludicrous in its defence that 24 languages (including Croatian, spoken in the recently admitted 28th member) are official, and that all official documents should be translated into all 23 other languages. This is by the way not the case in practice, but the EU still spends over $1.4 billion every year to interpret and translate from one language into all others. Just walk in the corridors of the many buildings of the European institutions in Brussels, and you will realise that the non-native English that is spoken is hardly understandable by a native English speaker, and that English native speakers lose others when they go into somewhat deeper discussions (Wright 2007). Is this what Professor Summers would like?

Worldwide, English is indeed the language that is most often used in international contacts and trade. But it is not the only one, as shown by Jacques Melitz (2008) who uses two measures of linguistic distances between trading partners and tries to estimate their effect. ‘Open-circuit communication’ (OCC) demands that the language be either official or widely spoken (at least 20% of the population knows the language). Spanish, for instance, will be an OCC between Bolivia (where 44% of the population knows Spanish) and Mexico (88%). A ‘direct communication’ (DC) language is any language common (that is, spoken by at least 4% in each country) in a pair of countries. In short, Melitz suggests distinguishing between two channels through which the trade-enhancing effect may take place: OCCs that depend on translation (which can be produced as long as there are enough people who can provide it in both countries) and DCs (which enable traders to communicate directly). He finds that ‘direct communication’ has the largest positive effect on trades: A 10% increase in the probability that two citizens, one in country A, the other in B, speak the same language increases their trades by 10%. Other European OCCs also contribute, but somewhat less. However, and interestingly enough, Melitz also shows that English as an OCC is no more effective than other European languages in promoting trade.

It is thus difficult to point to the number of languages that are needed, as well as to which ones are needed. There is no unique solution. It is clear today that globalisation has taken its toll, at least on some, if not on many of us. Do we need more of it by going to a unique language, whichever it is?

Editor’s Note: Victor Ginsburgh and Shlomo Weber are the authors of How Many Languages, Do We Need? The Economics of Linguistic Diversity, Princeton University Press, 2011.


Crystal, David (2001), A Dictionary of Language, Chicago: Chicago University Press.

Melitz, Jacques (2008), “Language and foreign trade”, European Economic Review, 52:667-699.

Summers, Lawrence (2012), “What you (really) need to know”, The New York Times, 20 January.

Thiongo, Ngugi wa (1986), Decolonizing the Mind, Oxford: James Currey.

Wright, Sue (2007), “English in the European Parliament: MEPs and their language repertoires”, Sociolinguistica, 21:151-165.

1,260 Reads