The Art of Intelligence


New York Times

April 2, 2005

The years between 1950 and 1965 were the golden age of American nonfiction. Writers like Jane Jacobs, Louis Hartz, Daniel Bell and David Riesman produced sweeping books on American society and global affairs. They relied on their knowledge of history, literature, philosophy and theology to recognize social patterns and grasp emerging trends.

But even as their books hit the stores, their method was being undermined. A different group rejected this generalist/humanist approach and sought to turn social analysis into a science. For example, the father of the U.S. intelligence community, Sherman Kent, argued that social science and intelligence analysis needed a systematic method, "much like the method of the physical sciences."

Social research - in urban planning, sociology and intelligence analysis - began to mimic the hard sciences.

A new paper by a Yale undergraduate, Sulmaan Wasif Khan, contrasts these two ways of looking at the world. Khan compares the C.I.A.'s 1960's-era National Intelligence Estimates on China, which have been recently declassified, with the work of generalist scholars like Donald Zagoria.

The C.I.A.'s intelligence estimates are what you'd expect: bloodless compilations of data by anonymous technicians. They do not draw patterns based on an understanding of Chinese history or make generalizations about the ethos of the Chinese elite.

Zagoria's approach was quite different. Relying on a deep understanding of Chinese history and society, he made novelistic judgments about the Chinese leadership's hopes and fears. He imagined how we must appear to the Chinese, and how different American moves would be interpreted.

The C.I.A. analysts concluded on Nov. 12, 1970, that there was little prospect of improvement in Sino-American relations. Zagoria said China would be open to a rapprochement.

Zagoria was right. Henry Kissinger was in China within months of the C.I.A. report.

But the scientific method used by the C.I.A., and its technical jargon, can seem to have more authority (used to justify bigger budgets). Academic analyses of society and world affairs are now often quantitative, jargon-laden and hyperspecialized. Historical works have gigantic titles and minuscule subjects - think "Power and Passion: Walloon Shovel Making, 1723-1724."

So we get decades of calamitous intelligence failures. This week the presidential panel on intelligence pointed to the same failings found by other reports. It said intelligence analysts "displayed a lack of imagination." They created artificial specialties - separating regional, technical and terrorism analyses. They built layers of hard analysis on fuzzy and impressionistic information.

This commission does what so many others have done. It tries to reorganize the bureaucratic flow charts to produce better results.

But the problem is not bureaucratic. It's epistemological. Individuals are good at using intuition and imagination to understand other humans. We know from recent advances in neuroscience, popularized in Malcolm Gladwell's "Blink," that the human mind can perform fantastically complicated feats of subconscious pattern recognition. There is a powerful backstage process we use to interpret the world and the people around us.

When you try to analyze human affairs using a process that is systematic, codified and bureaucratic, as the C.I.A. does, you anesthetize all of these tools. You don't produce reason - you produce what Irving Kristol called the elephantiasis of reason.

The capping irony is that Sherman Kent and the other pseudoscientists thought they were replacing the fuzzy old generalists with something modern and rigorous. But, in reality, intuitive generalists like Jane Jacobs and Donald Zagoria were more modern and rigorous than the pseudoscientific technicians who replaced them.

I'll believe the intelligence community has really changed when I see analysts being sent to training academies where they study Thucydides, Tolstoy and Churchill to get a broad understanding of the full range of human behavior. I'll believe the system has been reformed when policy makers are presented with competing reports, signed by individual thinkers, and are no longer presented with anonymous, bureaucratically homogenized, bulleted points that pretend to be the product of scientific consensus.

I'll believe it's been reformed when there's a big sign in front of C.I.A. headquarters that reads: Individuals think better than groups.