Metaphor and the Intelligence Advanced Research Projects Activity

Oliver Burkeman alerts me to a fascinating article by The Atlantic‘s Alexis Madrigal, Why Are Spy Researchers Building a ‘Metaphor Program’?. According to Madrigal’s piece, the Intelligence Advanced Research Projects Activity (IARPA) is soliciting proposals for research that will discover what metaphors from non-U.S. cultures reveal about the worldviews of those cultures. The idea is to use brute computing force (analyze that metaphor, IARPA!) to deduce the metaphorical meanings associated with various concepts. The goal: To use the resulting insights as part of America’s overall intelligence effort.

Research into teaching computers to understand metaphors has been ongoing for some time. The most dramatic success to date has been I.B.M.’s Jeopardy!-winning supercomputer Watson, which can process the loose associations, punning relationships, sidelong and sidereal correlations characteristic of metaphors. Madrigal cites the IARPA solicitation brief as follows: “The Metaphor Program will exploit the fact that metaphors are pervasive in everyday talk and reveal the underlying beliefs and worldviews of members of a culture.” If computers become skilled in metaphor crunching, it would have an effect far beyond America’s counter-terrorism efforts. Computers sensitive to metaphorical meanings could be used, to take one example, in medical settings to assist in diagnosing conditions that patients can only describe using the vaguest of metaphors.

But the difficulties computers have had to date with metaphor reminds us of just how marvelous our own innate metaphorical abilities are. There is a subset of metaphors that are unique to individual cultures, metaphors that are impenetrable to people outside that culture unless they are provided with the necessary linguistic or cultural context. But, and here’s the amazing part, given even the slightest clues about an alien metaphor’s context, we can instantly figure it out and provide an analogous metaphor from our own culture. And, for now, we can do that much faster than any supercomputer. Which is not to say that IARPA funding might not lead to a metaphor-making and -understanding machine even more awesome than the human brain. But for the time being, there seems to me to be a much faster, more efficient, and less expensive way to understand the metaphorical worldviews of people from other cultures: Ask them.