"From about 1990, a group in the U.S. came up with the idea of just treating it as a statistical problem -- more of a machine learning problem where you essentially ignore linguistics and just look for patterns in data, " he said.

CNN: 'Star Trek' translators reach for the final frontier

To be clear, the problem is not with numerical and statistical methods, it is with their blind application in the absence of reason.

FORBES: Manage Risk Like Robert Rubin

Since the mid-1950s, there has been a massive effort by international agencies in co-operation with national statistical offices to deal with the kind of problem he raises.

ECONOMIST: Letters | The