Followup to: Talking Snakes: A Cautionary Tale

Related to: Explain, Worship, Ignore

Skepticism is like sex and pizza: when it's good, it's very very good, and when it's bad, it's still pretty good.

It really is hard to dislike skeptics. Whether or not their rational justifications are perfect, they are doing society a service by raising the social cost of holding false beliefs. But there is a failure mode for skepticism. It's the same as the failure mode for so many other things: it becomes a blue vs. green style tribe, demands support of all 'friendly' arguments, enters an affective death spiral, and collapses into a cult.

What does it look like when skepticism becomes a cult? Skeptics become more interested in supporting their "team" and insulting the "enemy" than in finding the truth or convincing others. They begin to think "If a assigning .001% probability to Atlantis and not accepting its existence without extraordinarily compelling evidence is good, then assigning 0% probability to Atlantis and refusing to even consider any evidence for its existence must be great!" They begin to deny any evidence that seems pro-Atlantis, and cast aspersions on the character of anyone who produces it. They become anti-Atlantis fanatics.

Wait a second. There is no lost continent of Atlantis. How do I know what a skeptic would do when confronted with evidence for it? For that matter, why do I care?

Way back in 2007, Eliezer described the rationalist equivalent of Abort, Retry, Fail: the trilemma of Explain, Worship, Ignore. Don't understand where rain comes from? You can try to explain it as part of the water cycle, although it might take a while. You can worship it as the sacred mystery of the rain god. Or you can ignore it and go on with your everyday life.

So someone tells you that Plato, normally a pretty smart guy, wrote a long account of a lost continent called Atlantis complete with a bunch of really specific geographic details that seem a bit excessive for a meaningless allegory. Plato claims to have gotten most of the details from a guy called Solon, legendary for his honesty, who got them from the Egyptians, who are known for their obsessive record-keeping. This seems interesting. But there's no evidence for a lost continent anywhere near the Atlantic Ocean, and geology tells us continents can't just go missing.

One option is to hit Worship. Between the Theosophists, Edgar Cayce, the Nazis, and a bunch of well-intentioned but crazy amateurs including a U.S. Congressman, we get a supercontinent with technology far beyond our wildest dreams, littered with glowing crystal pyramids and powered by the peaceful and eco-friendly mystical wisdom of the ancients, source of all modern civilization and destined to rise again to herald the dawning of the Age of Aquarius.

Or you could hit Ignore. I accuse the less pleasnt variety of skeptic of taking this option. Atlantis is stupid. Anyone who believes it is stupid. Plato was a dirty rotten liar. Any scientist who finds anomalous historical evidence suggesting a missing piece to the early history of the Mediterranean region is also a dirty rotten liar, motivated by crazy New Age beliefs, and should be fired. Anyone who talks about Atlantis is the Enemy, and anyone who denies Atlantis gains immediate access to our in-group and official Good Rational Scientific Person status.

Spyridon Marinatos, a Greek archaeologist who really deserves more fame than he received, was a man who hit Explain. The geography of Plato's Atlantis, a series of concentric circles of land and sea, had been derided as fanciful; Marinatos noted1 that it matched the geography of the Mediterranean island of Santorini quite closely. He also noted that Santorini had a big volcano right in the middle and seemed somehow linked to the Minoan civilization, a glorious race of seafarers who had mysteriously collapsed a thousand years before Plato. So he decided to go digging in Santorini. And he found...

...the lost city of Atlantis. Well, I'm making an assumption here. But the city he found was over four thousand years old, had a population of over ten thousand people at its peak, boasted three-story buildings and astounding works of art, and had hot and cold running water - an unheard-of convenience that it shared with the city in Plato's story. For the Early Bronze Age, that's darned impressive. And like Plato's Atlantis, it was destroyed in a single day. The volcano that loomed only a few miles from its center went off around 1600 BC, utterly burying it and destroying its associated civilization. No one knows what happened to the survivors, but the most popular theory is that some fled to Egypt2, with which the city had flourishing trade routes at its peak.

The Atlantis = Santorini equivalence is still controversial, and the point of this post isn't to advocate for it. But just look at the difference between Joe Q. Skeptic and Dr. Marinatos. Both were rightly skeptical of the crystal pyramid story erected by the Atlantis-worshippers. But Joe Q. Skeptic considered the whole issue a nuisance, or at best a way of proving his intellectual superiority over the believers. Dr. Marinatos saw an honest mystery, developed a theory that made testable predictions, then went out and started digging.

The fanatical skeptic, when confronted with some evidence for a seemingly paranormal claim, says "Wow, that's stupid." It's a soldier on the opposing side, and the only thing to be done with it is kill it as quickly as possible. The wise skeptic, when confronted with the same evidence, says "Hmmm, that's interesting."

Did people at Roswell discovered the debris of a strange craft made of seemingly otherworldly material lying in a field, only to be silenced by the government later? You can worship the mighty aliens who are cosmic bringers of peace. You can ignore it, because UFOs don't exist so the people are clearly lying. Or you can search for an explanation until you find that the government was conducting tests of Project Mogul in that very spot.

Do thousands of people claim that therapies with no scientific basis are working? You can worship alternative medicine as a natural and holistic alternative to stupid evil materialism. You can ignore all the evidence for their effectiveness. Or you can shut up and discover the placebo effect, explaining the lot of them in one fell swoop.

Does someone claim to see tiny people, perhaps elves, running around and doing elvish things? You can call them lares and worship them as household deities. You can ignore the person because he's an obvious crank. Or you can go to a neurologist, and he'll explain that the person's probably suffering from Charles Bonnet Syndrome.

All unexplained phenomena are real. That is, they're real unexplained phenomena. The explanation may be prosaic, like that people are gullible. Or it may be an entire four thousand year old lost city of astounding sophistication. But even "people are gullible" can be an interesting explanation if you're smart enough to make it one. There's a big difference between "people are gullible, so they believe in stupid things like religion, let's move on" and a complete list of the cognitive biases that make explanations involving agency and intention more attractive than naturalistic explanations to a naive human mind. A sufficiently intelligent thinker could probably reason from the mere existence of religion all the way back to the fundamentals of evolutionary psychology.

This I consider a specific application of a more general rationalist technique: not prematurely dismissing things that go against your worldview. There's a big difference between dismissing that whole Lost Continent of Atlantis story, and prematurely dismissing it. It's the difference between discovering an ancient city and resting smugly satisfied that you don't have to.



1: I may be unintentionally sexing up the story here. I read a book on Dr. Marinatos a few years ago, and I know he did make the Santorini-Atlantis connection, but I don't remember whether he made it before starting his excavation, or whether it only clicked during the dig (and the Internet is silent on the matter). If it was the latter, all of my moralizing about how wonderful it was that he made a testable prediction falls a bit flat. I should have used another example where I knew for sure, but this story was too perfect. Mea culpa.

2: I don't include it in the main article because it is highly controversial and you have to fudge some dates for it to really work out, but here is a Special Bonus Scientific Explanation of a Paranormal Claim: the eruption of this same supervolcano in 1600 BC caused the series of geologic and climatological catastrophes recorded in the Bible as the Ten Plagues of Egypt. However, I specify that I'm including this because it's fun to think about rather than because there's an especially large amount of evidence for it.

New Comment
15 comments, sorted by Click to highlight new comments since:

I agree with most of this, but I think you're skipping a really important issue:

"There's a big difference between dismissing that whole Lost Continent of Atlantis story, and prematurely dismissing it."

Well, sure, but we need some way of deciding when our dismissal is premature. I mentioned this in the comments on Talking Snakes as well... there is certainly some room for the absurdity heuristic - my time is valuable, I can't evaluate the evidence for every crank claim in the world (I know several academics who could easily spend their entire lives checking "proofs" that P=NP if they did this). I have to reject some of them out of hand - the issue is: which ones?

If someone tells me they've built a perpetual motion machine in the back garden from tin cans and elastic bands, I'm not going to waste even 10 minutes of my life trying to replicate it. If Steven Hawking tells me he's built one, I'll at least give the matter some consideration. The real issue is where we draw the line between claims which are too absurd to bother looking at the evidence for and those we should take the time out to evaluate, you don't seem to have said anything yet that will help us to make that decision (perhaps you're getting there).

Good points.

I don't have such a formula for drawing the line between "too absurd" and "go and evaluate" (I will be mentioning one technique in a while that might help a little, but it's a complicated technique and there's no reason to do it unless you've already decided something's important).

My point is that we should increase the amount of potential value we expect from investigating seemingly stupid claims for two reasons. One, the seemingly stupid claims may be true (ie evolution). Two, the seemingly stupid claims may be false, but there may be some hidden reason for their existence that is still a worthwhile discovery (ie religion leads to evo psych, Atlantis leads to Santorini).

But that's all I'm saying - increase the potential value you expect from investigating. In the case of the perpetual motion inventor, the case against perpetual motion isn't the absurdity heuristic, it's a physical law that's been proven about as well as anything in science - so no need to increase the value of investigation on those grounds.

You should consider the possibility that you'll discover something interesting about the human mind when trying to figure out why people keep inventing perpetual motion machines, but if you're a physicist who's very busy and doesn't have much interest in wild goose chases after some psychological phenomenon, then it's completely fair to leave that to the psychologists and tell the perpetual motion inventor to take a hike.

Likewise, you should assign nonzero probability to the possibility that the perpetual motion machine inventor has discovered something interesting that's not a perpetual motion machine - I think some people used to call the solar sail a perpetual motion machine, and even though they were wrong the solar sail is still quite an interesting discovery. This is worth consideration if the inventor is someone smart enough that you don't expect them to make elementary mistakes (ie Stephen Hawking).

If it was just some random guy, I would join you in not paying him any attention, unless I was feeling especially interested in cognitive biases relating to perpetual motion machine construction that day.

Summary: Correct for certain possibilities when deciding whether or not to reject something out of hand, but if you still want to reject something out of hand after making the corrections, do it. Highly informal, but I don't know any technique for formalizing it better.

I have attempted to design perpetual motion machines, so perhaps I can tell you something about the motivations involved. First of all, it is fun and educational. Each time I designed a perpetual motion machine, I learned something new about physics that I wasn't taught in school. This is probably not the motivation of someone who has actually tried to build one though.

Secondly, I have some doubts about the Second Law of Thermodynamics. It is almost assuredly not itself a law of physics, but rather of statistics; it is in a sense at best redundant. To my knowledge, there is not a single perpetual motion machine that cannot be shown not to work without invoking the laws of thermodynamics. I consider my existence to be proof or at least strong evidence that at some point the Second Law doesn't hold -- it seems incompatible with universe creation or recycling for eternity under increasing entropy.

Thirdly, I hate and despise the consequences of the laws of thermodynamics. The end of all life and civilization. The most certain threat to humanity's existence. Sure, in the unimaginably distant future, but it still bugs me.

Fourth, the idea of doing something really big that will make you famous. Note that worthy problems are really hard to find. I do like to poke at famous or impossible problems, though personally only attempt perpetual motion for educational purposes.

PS: my current best design for a perpetual motion machine, of the first type, is hollowed out permanent magnets plus a magnetic monopole. Note that some physicists seem to think magnetic monopoles might exist. It would have to be a fundamental particle though, as it can't be constructed from magnetic dipoles or electric monopoles. A similar design could be constructed from electrical monopoles and fundamental electric dipoles that can be held in a given orientation. Note that although "electric dipoles" are very common, they are not fundamental, they are made of electric monopoles, and the field inside them is incredibly strong and in the opposite direction. Active research is being done in an attempt to find a dipole moment for fundamental particles, so far with no luck. Unless my design has errors, it means that any attempt to find those is doomed to failure as a violation of conservation of energy.

Thinking about it some more, for me perpetual motion machine design is a special case of noticing I am confused. Specifically, that I hold the contradictory beliefs, "perpetual motion machines are to an incredibly high level of certainty impossible" and "my current understanding of [branch of physics] suggests I could build a perpetual motion machine". Then the perpetual motion machine design functions as a formal, step-by-step analysis so I can identify exactly where I think I'm breaking one of the laws of physics. Similarly, if I can't even formulate a design, not only am I confused but my understanding is too vague to apply to specific circumstances.


So, applying to some popular claim, say the resurrection of Jesus. The argument is that, instead of concluding, "It's highly unlikely that Jesus resurrected; his disciples were either mistaken, lying, or hallucinating; let's talk about the things we can test instead," attempts to assess the probability of surviving crucifixion according to available evidence are superior?

I agree on some level, but many "ignore" decisions are based on perfectly defensible Bayesian priors. Not ones that apply 0% probability to events, but I believe I could successfully argue that each of the 4 above explanations for the resurrection are ALL more likely than supernatural resurrection itself. And because of my knowledge of cognitive bias, I think I could well argue that the dismissive claims are more likely than surviving crucifixion as well.

Am I missing the point?

Edit: Actually, I think I see the error I made already. Explaining "hallucinating, mistaken, or lying" in terms of cognitive bias is much, much better than any of those explanations on their own.

I agree, many "ignore" decisions are based on perfectly defensible priors. I think this technique is only useful in certain circumstances.

If you care enough about the Resurrection that you want to be sure you have it exactly right (and many atheists don't care that much, and I don't blame them) then you need to spend some time looking for alternate explanations like Richard Carrier's.

It would be nice if there were some way to confirm Carrier's hypothesis in the same way Marinatos confirmed his Atlantis hypothesis, but I can't think of one. So you may end with a probability of epsilon for miracle, 70% for hallucination/mistake/lie, and 30% for Carrier. You now know more about the Resurrection than you did before, even though "hallucination/mistake/lie" is still your best guess (and you will know even more if you can break hallucination/mistake/lie down into the actual motivations or mechanisms behind lying or hallucinating in that situation).

I think this especially matters in cases where you have some reason to doubt the hallucination/mistake/lie hypothesis. If you personally knew St. Peter, and you couldn't possibly imagine him ever lying about anything, then restricting yourself to the dichotomy "Peter is a liar" or "Jesus rose from the dead" will tie you into epistemic knots. Under those circumstances, you'd be really happy you'd taken the extra few minutes to find Carrier's theory.

One thing I didn't make clear in the article and which you've caught is that the clever explanation isn't always or even usually better than the simple explanation. Sometimes people are just liars or mistaken, and to always use the clever explanation is as dangerous or worse than always using the simple explanation.

Ultimately this technique is about not prematurely stopping your search. If someone says "I saw Jesus three days after he was crucified", there's an overwhelming urge to ignore him, since your brain may leap to equate that with the absurd claim that he rose from the dead. But if you can tease apart those claims, you have another alternative. That alternative may be wrong, and you may end up rejecting it, but if you're unhappy with the previous dichotomy you ought to at least make sure you know it's there.

When skeptics say Atlantis is impossible, are they talking about Plato's story or the island with glowing crystal technologies?

Because I'd be willing to say that the probability of the latter is close enough to zero for government work.


From Wikipedia: "The prevalence of Charles Bonnet Syndrome has been reported to be between 10% and 40%; a recent Australian study has found the prevalence to be 17.5% [2]. Two Asian studies, however, report a much lower prevalence [5].[6]. The high incidence of non-reporting of this disorder is the greatest hindrance to determining the exact prevalence; non-reporting is thought to be as a result of sufferers being afraid to discuss the symptoms out of fear that they will be labelled insane[4]."

  • Yvain, your posts are an incredible wake-up call as to the astounding weirdness of the everyday world, to an extent that I ought to have expected but didn't. Entire countries, scientists, governments and all, are irrationally frightened of "fan death"? I wouldn't have believed it.

~10% of people see elves?! But don't report it because they think they'll get institutionalized? I would not have believed that.

I admit that I remembered hearing about Charles Bonnet syndrome before, and linked to the Wikipedia page without reading it. 10 to 40 PERCENT?!

Looking at the literature, I find some evidence that the real prevalence rate is closer to 1%, and that the higher numbers are only true of people with impaired vision. That's still impressive, though.

If it is the case that incidence is notably higher in people with impaired vision, and it is the case that in pre-modern societies impaired vision was far more common, then that would seem to provide a plausible explanation for why so many more otherwise rational people saw supernatural events in pre-modern societies than societies with access to better medicine.

Contrast the alternative hypothesis: people in pre-modern societies were all gullible and stupid.

I think the Ten Plagues of Egypt is more likely to be a fiction than to be based on a real event.

I think the entire story of Exodus from Egypt is more likely to be (mostly) fiction than based on a real event.

As Eliezer himself said in that post, the Egyptians were "known for their obsessive record-keeping". If anything remotely comparable to the Ten Plagues or the Exodus happened in Egypt, they ought to have recorded it. That they didn't is very strong evidence that nothing happened.

I agree with most of this, the problem is I'm not sure how useful it actually is:

"There's a big difference between dismissing that whole Lost Continent of Atlantis story, and prematurely dismissing it."

Well, sure, but we need some way of deciding when our dismissal is premature. I mentioned this in the comments on Talking Snakes as well... there is certainly some room for the absurdity heuristic - my time is valuable, I can't evaluate the evidence for every crank claim in the world (I know several academics who could easily spend their entire lives checking "proofs" that P=NP if they did this). I have to reject some of them out of hand - the issue is: which ones?

If someone tells me they've built a perpetual motion machine in the back garden from tin cans and elastic bands, I'm not going to waste even 10 minutes of my life trying to replicate it. The real issue is where we draw the line between claims which are too absurd to bother looking at the evidence for and those we should take the time out to evaluate, you don't seem to have said anything yet that will help us to make that decision (perhaps you're getting there).

I think it is fair to ignore things that have no consequences. And if there are consequences, you can get to testing whether the consequences happen or not.

If you ignore it, how do you know it has no consequences?