When experts and laymen disagree, the rational default presumption is that the expert is probably right—a presumption we rely on for mechanics, doctors, and historians—though it is only a starting point that can be overridden by a specific, testable account of expert bias.