It is such a mixed up, muddled up world we live in these days that verities seem harder than ever to come by. Even death and taxes may not stay inevitable for long, what with clever bioscientists trying to defy the former and the formidable Grover Norquist determined against the latter. So it's comforting to know that something of interest remains unarguably inevitable, and that something is error--particularly the errors of organizations. The old witticist who said that to err is human, but to really foul up takes a computer, might have added that to inadvertently get lots of innocent people killed generally takes a government or two.

Ultimately, the errors of organizations do come down to the errors of individuals, because abstract nouns don't literally make decisions--but they come down in no simple way. Context is crucial. Those guilty of poor judgment in one organizational setting might have been wiser in another. Thus we recall Irving Janis' warning against groupthink and, more relevant than ever, Solomon Asch's famous 1950s experiments on conformity. Interested in the psychological power of social context, Asch asked subjects to choose one of three vertical lines that matched in length a control line. Unbeknownst to the subjects, Asch planted shills to give close but wrong answers before the subjects were asked to give theirs. Asch wanted to know how many subjects would be influenced to give incorrect answers. Many were: at least one of four half the time, as many as three of four some of the time.

Asch's data suggested two interpretive possibilities: either the subjects knew they were giving wrong answers but did so anyway to avoid social tension, or they experienced genuine doubt and made genuine errors on account of social context. Asch suspected but could never prove the latter interpretation. Now we know: Gregory Berns of Emory University recently replicated the Asch experiments using MRI brain-monitoring devices to learn which parts of the brain processed respondents' answers. His research establishes that error in individual judgment can be and, we must surmise, often is induced by social context.

Of what practical importance is this knowledge? The hypothetical implications for understanding leadership dynamics and even the theory of democracy are significant. For the moment, however, let's limit ourselves to a not entirely hypothetical example from contemporary U.S. foreign policy.

Some people in the Department of Defense, the intelligence community and the White House in 2002 were about as sure as could be that the former Iraqi regime was hiding weapons of mass destruction and probably manufacturing still more. Those people increasingly engaged with like-minded others in government as Iraq became the focus of U.S. policy, and a powerful inferential logic that led...