Lessons from a past perdu

From Kam Patel's interview with John Hopfield (THES, October 21) we learn that his 1982 paper on associative memory "led to an 'explosion' of research in neural computing". Hopfield himself illustrates the concept as follows: "a memory of a party and everything in it the conversation, the food, the embarrassments can be triggered by something many weeks after the event."

As any student of French will -- or ought to -- be able to tell you, this view of the mechanics of memory is not a thousand miles away from an idea running through Marcel Proust's A la recherche du temps perdu, published between 1913 and 19. So much for the compartmentalisation of knowledge. One can but speculate on how many millions of pounds are largely wasted worldwide on research grants to look into questions already answered, at least in part, by research done in a discipline, and possibly published in a language, alien to the new researcher.

GEOFFREY PERRIN

Federal Language Centre

Cologne