In the early 1940s at a Rockefeller-funded symposium at the University of Wisconsin on the fate of minority languages in America, the linguists present concluded that on the face of their evidence most of these languages, including French, would not survive to the end of the twentieth century. Today one can report that the status of French in Quebec is higher than it was at the time the prediction was made. Yet the same prediction had been made a century earlier in the 1840s by a British colonial administrator who had witnessed the dismemberment of the Napoleonic empire. Moreover, four centuries earlier, some English academics had come to the conclusion that once people were better educated (in Latin and French) English would not be needed as a language of learning.

If one looks at what happened to these past predictions one wonders if our present ones will fare any better. Could they go wrong and, if so, how and why? They could, indeed, go wrong in many ways and for different reasons. I have tried to group these into four broad categories concerning:

• their demographic assumptions;

• their choice of decisive factors;

• their methods of extrapolating present trends; and

• their appropriation of models from different disciplines.

Demographic projection

The future of a language depends on the number of people using it: if people can no longer speak or read a language its future is bleak.