The engineering geek in me gets impatient, from time to time, with talk of "risk-benefit analysis." As in, "Do the benefits of thus-and-so outweigh the risks?" Benefits are benefits, and risks are risks; they are unlike quantities, and to talk about whther one "outweighs" the other is about as reasonable as asking whether my apple is crisper than my pickup truck is dirty. Unlike quantities can't be compared in this way, and it's one of the shortcomings of the English language that we can so easily talk about such comparisons.
But a CNN story about the latest coal-mining disaster has an interesting sidebar: it bulletizes a handful of mining catastrophes, going back to 1984, in which a total of 63 people were killed. And yet, we go on generating electricity by burning coal. We don't build any nuclear power plants, because we were scared spitless by Three Mile Island (at which how many were killed or injured? that's right, zero) and Chernobyl, which ocurred in the essentially third-world context of the old Soviet Union.
In terms of excess deaths, those 63 miners in the past two decades alone are a drop in the bucket, compared with the excess deaths and illnesses caused by breathing the same air in which millions of tons of coal are burned.
The risks associated with various methods of generating a reliable electrical supply cannot meaningfully be compared to the benefits of the availability of that supply. But the risks associated with one kind of generation can be compared to the risks associated with other kinds of generation -- and should be. This may be a situation in which a class analysis is useful, however out-of-fashion Marxism may be. Could it be that the nuclear risk, while clearly smaller, is less tolerable to us because it applies to the population more generally -- while coal-mining accidents are bad mostly for people who are sort of Appalachian, and don't exactly live in our neighborhoods? Ugly ... but not implausible.