Last week's New Yorker has a great article by Malcolm Gladwell about the trouble with generalizations. Basically, Gladwell uses the examples of aggressive pit-bulls and policing and profiling tactics to show how problematic generalizations can be. It's really worth a read.

Gladwell's column in the New Yorker is called 'Fact', which may explain why his article is mostly a straight report of some interesting contradictions and oversights. The interesting thing that he didn't even get into, though, is why people choose to abide by generalizations, often without any supporting evidence at all. I suppose that's a combination of sociocultural and psychological, and it's at least as complex as what the article does get into. But my theory is that people choose to abide by generalizations, even when they are seemingly illogical (as they often are), for two main reasons.

1. Most people believe things that support things they believe already. In other words, people don't like to contradict themselves, at least not internally. It sets up too many upsetting complexities and ironies. Generalizations allow people to be consistent about their attitudes at a much lower resolution, which is nice for them.
2. No one likes to blame themselves when they can get away with blaming other people. Or, in the case of pit-bulls, no one likes to blame people when they can blame animals. Generalizations are based on extrapolating the details of a small number of situations out to a very large set of situations. Sometimes this is fair, granted. We wouldn't get far in the world without some degree of generalization. But in the process, all the pesky little details (like the negligent, animal abusing owner who made that pit bull go crazy) tend to just go away.