(For specialists.) The claim that physical processes (as opposed to mathematical models) are self-organized critical rests on finding power-law distributions in their behavior; in practice, straight lines on log-log frequency plots. But there are, if not precisely nine-and-sixty ways, then certainly a great many, of producing power laws, and every single one of them is right. As Deborah Mayo would put it, finding a power-law distribution for a process doesn't constitute a severe test of the hypothesis that it's self-organizing critical. I suspect that, if Bak had been in on it, we'd call the central limit theorem ``self-organized normality,'' and look for random walks (say) behind every Gaussian distribution.