Friday, November 9, 2007

"Extraordinary claims" defined

"Extraordinary claims require extraordinary evidence." This is a skeptical maxim, usually credited to Sagan (though others have said similar things before him). Basically what this means is that if you make increasingly fantastic claims, the skeptic is not increasingly impressed, but rather wants more evidence. But sometimes, there is confusion about this saying. What constitutes an "extraordinary claim"? And is the saying always true, or a rule of thumb?

There are several ways to define extraordinary claims. The most obvious definition would be those claims that require extraordinary evidence. That makes our maxim into a nice tautology; it is guaranteed to be true. But the problem with tautologies is that they do not tell us much. How do we tell an extraordinary claim from an ordinary one? We look to see whether it requires extraordinary evidence. How do we know whether it requires extraordinary evidence? We look to see whether it is an extraordinary claim. All in all, this is a useless definition.

There is another obvious definition that is not tautologous. An extraordinary claim is one that is already contradicted by extraordinary evidence. Under this definition, our maxim basically says that the only way to counter evidence is with more evidence. I think that this is self-evidently true, and quite useful. For an example application, take astrology. There are no known forces that might cause planets to affect our destinies, none that act even remotely similarly. Add to that the numerous studies that have shown astrology ineffective, and we've got quite an extraordinary claim here.

However, I think that skeptics usually mean much more by "extraordinary claims". The qualities that can make a claim extraordinary aren't completely arbitrary, but I would still have difficulty describing them. I would say that beyond this point, the maxim changes from an absolute truth to a rule of thumb.

The first, perhaps most important quality that can make a claim extraordinary is if it is unnecessarily complicated. This principle is called Occam's Razor, and it really deserves a separate discussion. Suffice it to say that inserting invisible unicorns into your theories usually does not improve them.

Another quality that makes a claim extraordinary is if it has important implications. If accepting a claim is a matter of life and death, you better be sure to supply lots of evidence. The caveat is that this often works both ways. For example, if I claim that children's vaccines are causing autism, this really is a matter of life and death, since vaccines save lives and autism harms them. However, the counter-claim that vaccines are not causing autism is also a matter of life and death. Both of these are extraordinary claims, but we obviously have to choose one, regardless of how little evidence we have. (In this case, the evidence says that vaccines do not cause autism.)

There are more qualities, but at this point it becomes difficult to pick them out. In reality, most skeptics just have an intuitive notion of what sort of claims are sensible and what sort are ridiculous. You know it when you see it. But I think skeptics would do better to try to explain a rationale for what makes claims extraordinary--which is what I just did.