Science rests on the assumption that the world can be known: that causes have effects, and logic applied on Thursday in Katmandu is valid still on Friday in Geneva. Psychology, and its claims to be a science, rest on the assumption that people can be known, in the same way we know about chemical bonds, the Jurassic and quarks. Individuals vary and minds are complex, but under all that complexity, psychologists promise to find underlying laws that explain why and how we think, feel, perceive and act as we do.
Their tools are supposed to be the same as those of other scientists: As particle physicists deduce the presence of a Higgs boson from measurements right after a high-speed collision, so psychologists measure walking speed to deduce that being reminded of stereotypes about the elderly will make students slow down afterwards. Such explanations, like many in science, often depend on relationships among measurements (for example, in hockey players wearing black correlates with more penalties for violence, suggesting that black has an effect on behavior (or at least how behavior is perceived). This can be done in retrospect, but the ultimate test of an explanation is prediction: After I've established that black is associated with aggression, I should be able to confidently tell you which team will get more penalties for violence next season (the ones dressed like ninjas).
Now, there are people who say psychology isn't a proper science (as this guy argues) because concepts like aggression, will-power and empathy can't be precisely enough defined to be measured by scientific methods. If they're right then academic psychology is simply a more systematic form of "folk psychology," which we all do every day: Trying to understand ourselves and others by means of correlations—"My kids do better on exams when they cram at the last minute" or "our people perform better after I berate the poor performers." Whether it's formal psychology or the folk variety, though, we all depend on this faith that it is possible to know other people—to explain their behavior in coherent ways, and to predict what they are more or less or never likely to do.
And then along comes James Holmes. He is accused of murdering 12 people and wounding 58 for no reason at all, as most of us would define a reason. And so our modern priesthood of experts is on the air to "explain" why this particular person would do such a thing. And they look to their correlations of variables, their indicators and theories, and find precisely nothing. Where are the indicators we want to see, the ones we can associate with senseless slaughter? Where are the traits whose presence would reassure us that it is possible to know who is vulnerable to the lure of mass murder?
There are none. No indications of mental illness earlier in life. No signs of a violent or troubled childhood. No signs of drug abuse. Instead, traces of a mild blank person, whose signature trait seemed to be that he left little impression at all on other people. A background guy, the sort who doesn't make us alert for trouble. His story threatens formal psychology and folk psychology, because it tells us that another person can't be known, not for certain. (That he was himself a grad student in neuroscience, aiming to elucidate how the brain causes behavior, adds a quality of mockery to the tale.) So far, at least, his life suggests that however well psychologists correlate variables and generalize their theories, we still have to assume that anyone is capable of anything—that whatever we may want to believe, Charles L. Mee had it right: Another person is a foreign country.
ADDENDUM, 7/24/12: Over at The Guardian, Harry J. Enten ran the numbers to see what Holmes has in common with other rampage killers. Maleness and being described as a loner were the only ones that leapt out. Enten also noticed a paradoxical fact about the numbers: Even as rampage killings have increased in number over the past twenty years, plain old murders have been declining. Moreover, as Patrick Egan points out here, when asked if they own a gun, fewer Americans now say yes than did in the 1970s. It's a huge drop, from around half of all respondents to only one-third. If we are a gun-crazed nation compared to, say, Canada, well, we are less gun-crazed than we used to be.
This week we all find it easy to imagine being killed by a berserk lunatic (because the media keep reminding us about such killings, and when something comes easily to mind we think it is more likely). But rampage killings, though increasing, are still so rare that your chances of being shot are lower by far than they were in the 1980s.