One person says the medical establishment pushes autism-causing vaccines on the public, and “they” perpetrated a massive cover-up. A second person claims the tobacco industry colluded to distort the evidence that smoking causes cancer.

Each of these persons is pushing a “conspiracy theory,” a theory that says a group of people did something bad and is conspiring to cover it up. But the first person (the anti-vaxer) is badly confused. The second, however, is right: the tobacco industry did collude to distort and suppress science about the link from smoking to cancer.

这种对比凸显了阴谋论的思考难度。一方面，阴谋论往往是古怪的。(As John put it on theDecember 4 show, many conspiracy theorists are “paranoid nut-jobs.”) On the other hand, human beings do conspire from time to time; that’s often the instrumentally rational thing to do for a group that did something criminal. So the difficulty is in finding criteria that would help us say why some beliefs in conspiracies are cuckoo while others are reasonable.

Our guest, Brian Keeley, expressed some despair over finding such criteria. He seemed to think that there wasn’t anything you could do up front to tell the nonsense conspiracy theories from the reasonable ones—only time (and further evidence) would tell. John, however, was more optimistic. He seemed to think that there was something about the theories themselves that would allow you to distinguish which were wacko.

I agree with John. I don’t, however, think that there is acleandividing line between the crazy theories and the sane ones. Rather, there is a cluster of features that come in degrees, and the more of these features a given conspiracy theory has (and the greater degree to which it has them), the more it is the ridiculous kind.

What features do I have in mind?

1. Reliance on the amorphous “they” / Lumping Non-Cohesive Groups.Who’s covering up the information that the earth is flat?Theyare! Who’s responsible for chemtrails?Theyare! Who’s covering up the truth about J. F. K.?Theyare! And so on. Whenever a conspiracy theory doesn’t identify who exactly perpetrates the purported conspiracy, we should be suspicious of it. The general problem here is that if you think about who would have to be included in the “they” in order for the conspiracy to work, it often ends up being groups and institutions that are so different they would never work together. Just think about everyone who would have to be involved in the conspiracy to cover up the supposed flatness of the earth: it’s basically every government in the world, every scientist, every news editor, and every transport corporation (and many more). But why would all ofthemwork together? (Also, watch out for words or phrases thatsound likethey identify a group but aren’t informative at all. When you hear, for example, that “the establishment” is covering something up, you should ask to whom that phrase refers. If it’s a cohesive group, then fine. But if the only way to get the conspiracy theory to work is if the conspiring “establishment” includes such diverse people as Texas Tea Party members and liberal editors of newspapers—they’re all part of the establishment!—then that theory is probably incoherent.)

2. Lack of plausible motive.What would have to be true in order for Barack Obama to have been born in Kenya? In addition to his having been born there, two Hawaii newspapers and the Hawaii Department of Health in 1961 would have to have beenmotivatedto report that a baby born all the way over in Kenya was born in Hawaii. But—even setting aside the question of how the newspapers would know about this far away baby—why would any Hawaii newspaper want to convince people ofthat? The motive makes no sense at all.

3. Invention of new individuals.Let’s go back to chemtrails. Who’s putting the chemicals in the engines of the airplanes, for example? Who’s manufacturing those chemicals? Who designed he chemicals? The more previously-unheard-of individuals a conspiracy theory has to posit, the more suspect the theory is. (This is a somewhat different problem from point 1. Point 1 is that conspiracy theories are often incoherent for implausibly lumping together distinct groups; Point 3 says that conspiracy theories are often unparsimonious for having to posit individuals we have no independent reason to think actually exist.)

4. On the team / off the team.The really bad conspiracy theories rely on evidence from people who at other timesthose very theoriessay are part of the conspiracy. Conspiracy theories about global warming, for example, rely evidentially on the fact that during certain periods the size of polar ice caps has increased (the overall pattern, of course, is that they are shrinking). So for the tiny bit of evidence in their favor, they rely on sources like NASA. NASA, however, is clear about the fact thatglobal warming exists. So NASA has to be on the team then off the team: at one moment, NASA offers crucial evidence, but the next moment they’re part of the conspiracy!

None of these features is definitive of a wacko conspiracy theory. After all, sometimes theories are right to posit unheard of individuals, etc. But it’s fair to say that theories that have these features in spades are the crazy kind.

Let’s reconsider our opening examples.

In the case of the cancer cover-up, the four features are lacking. There’re no amorphousthey(我们知道是谁在做这件事，他们肯定会合作);其动机不仅貌似可信(而且邪恶);不需要创造新的个体;你不需要改变你的证据来源的状态(科学家就是科学家)。

But the vaccine conspiracy would have to include doctors, the CDC, drug companies, many scientific researchers…“they” is such a huge, amorphous group that “they” would never hang out together. Furthermore, the motives posited are bizarre. Why would anyone want to lie about what gives kids autism? And though it doesn’t seem to me that the anti-vaxers need to invent new individuals, they certainly need to have an oscillating attitude toward scientific sources. The journal that published the original anti-vaxer study hasretracted it; so it was authoritative at one point, but now it’s all of a sudden a part of the conspiracy.

So I think the concept of the wacko conspiracy theory is acluster concept. If the theory you believe in has this cluster of features, you’ve probably lost it. Time to take off the tinfoil hat.