Nicholas Ostler, Foundation for Endangered Language

The origin and early history of research and development in machine translation might suggest that it is only interesting, or applicable, to the greatest of major languages – Russian, Japanese, French, German, Chinese and above all English. But founders' effects do not persist in eras of furious technical change, unless they concern essentially arbitrary aspects, such as technical standards. (Not to be derided: in fact, an arbitrary carry-over from one language's technology may solve another's persistent problem.) This happens because deep learning, as well as de-skilling, can result from the introduction and acceptance of new technologies. Machine translation, and language technologies more generally, may yet be very useful to minority languages, promoting and extending both their use and their status.

It is now clear that the concept of MT's role which underlay the early research was not only insufficient but inadequate. But when technological ground is continually being ploughed up, there is cope for interesting new crops to germinate and flourish. Radical multilingualism may be one such crop, in a field-system (or a jungle) of pervasive digital technology. I shall draw on an autobiographical myth, personal experience of 'building Europe', and the global history of missionary linguistics, to illuminate the likely future of machine translation at large. The future is bright: the future might even be oral.

*Homage to F. Spencer Chapman.