Things You Don't Talk About in Polite Company

A lot of interesting material there. I zeroed on a debate: Sewell Wright claimed that small isolated communities are the best condition for rapid evolution, while Ronald Fisher championed large populations with random mating.

So, why does it have to be either/or? Perhaps the two extremes, for different reasons, maximize the probability of rapid evolution, while in-between situations don't. Maybe someone out there is making that argument. It's not my field.

I've noticed that people can get downright hostile whenever you suggest a multi-modal distribution in a hypothesis. I remember freshman year that my organic chemistry prof seemed a bit shocked that grades followed a bimodal distribution.

The Gaussian distribution is also known as the normal curve. It fits a lot of data or (through manipulation) data can be made to fit it. It must be called "normal" for a reason. If I were more rigorous I'd look that up. It seems so ingrained in people's minds that they assume that data fits the curve unless confronted with a lot of evidence to the contrary. Since the distribution is commonplace, it's a useful heuristic, but no need to be dogmatic about it.

I suppose I see democracy in the same light. It's a very useful heuristic for government rather than a value in itself. It doesn't guarantee a valid solution. It may even fail miserably. For now, our peculiar form of American democracy is still a relevant tool of justice. I don't see a reasonable replacement any time soon, but I'm always willing to think about it.