“He who is burnt by hot milk blows on ice cream.” An old Sufi saying. In other words, don’t over-generalize. Or to generalize further, don’t over-, under-, or otherwise mis-generalize.
But how do you know whether you are?
The ice cream quote would be called a false induction–generalizing from too little evidence. This milk is too hot, therefore all dairy products are too hot. Working from too little evidence is over-generalizing. But there’s the opposite too–under-generalizing. For example, “Though I’ve gotten four DUI’s and crashed three times in the last few years, that’s insufficient evidence that I should stop driving drunk.”
The ice cream quote also contains a false abduction–identifying the wrong dimension of analysis. It’s not the dairy-ness of milk that made it too hot.
False abductions are often funny:
Monday eve: Gin and Tonic; Tuesday morn: hangover.
Tues eve: Vodka and tonic; Weds morn: hangover.
Wed’s eve: Whiskey and tonic; Thurs morn: hangover.
I’ve got to stop drinking tonic.
Jokes like that tease about people’s motivations for mis-generalization. The drinker is biased against concluding he should stop drinking alcohol. He mis-generalizes to another conclusion to avoid an inconvenient one.
The potential for over-, under- and mis-generalizing is not into account when we claim that “those who don’t learn the lessons of history are forced to repeat it.” The question is which lessons or generalizations should we take from history? Both Hitler and Stalin were short guys with extravagant mustaches who killed millions. So is the lesson of history that as leaders we should never again have short guys with extravagant mustaches?
He who is burnt by mis-generalization… One can mis-generalize about mis-generalization concluding that since you can make mistakes in generalizing, it’s best to not generalize at all. People say “don’t generalize” as though that were possible. We all generalize. We have to, or else every moment would be alien to us in every way. The question, of course is how to generalize right. This is no trivial question. It’s basically how do you know what’s generally true? How do you figure out which generalizations are worth heeding?
A very common approach is to ignore the generalizations made by anyone with a bias, on the assumption that they can’t be trusted to generalize.
You get this a lot. For example, the people who deny global warming often say, “Oh, its just those pro-big-government types. They’re biased. They just use global warming as an excuse for imposing social control.” Tyrants do this too. They discount their opposition by saying, “Ignore them. They’re just biased.”
The impression they want to give is that bias is a rare and peculiar infection that should be quarantined instantly lest it spread. I think bias is more like common digestive bacteria. We’ve all got it. So simply pointing a finger and saying “Ooooo, bias! Close your ears!!” isn’t a good enough way of to prevent errors of generalization.
Obviously if you have no bias it can’t cause errors of generalization. But not all biases translate into errors of generalization. And besides biases are everywhere. It’s especially important to monitor those who claim or imply that they have no bias. Their anti-bias bias makes them doubly slippery.
If ignoring anyone with bias doesn’t protect you against errors of generalization, what does? It’s not a trivial question and the answer isn’t trivial either. One key is keeping an eye on those inductions and abductions. Here’s an audio powerpoint presentation on it.