This is a good response, and very clear. Thanks. It's my own opinion that the Sagan standard has been controverted powerfully by the philosopher William Lane Craig. You can see his argument here ( reasonablefaith.org/writings/question-answer/… ) and also here ( reasonablefaith.org/podcasts/defenders-podcast-series-3/… ).

My central claim here is that the consilience of evidence has been called into question by the photos of cave paintings that had been presented. So, in my thinking, if we don't allow the consilience of evidence to be called into question then we are not acting as skeptics.

I tried reading Craig's argument. It seemed close to gibberish in the details. If his conclusion was true, then you should start believing in unicorns if anyone says they saw one because P(witness claim | unicorns) >> P (witness claim | no unicorns). It sounds like motivated reasoning to justify his other beliefs. Maybe it would make a Skeptics.SE question?

I disagree that the consilience of evidence has been seriously challenged. It has been very weakly challenged. We should allow the question, and we should approach it with an open mind. But when we realise how weak the counter-evidence is, we should reach a provisional conclusion, not simply shrug our shoulders and say "Well, it is in doubt. Who could know?"

Politely, but firmly: "No. [end of paragraph]" is much more than a provisional conclusion! It's a slap in the face for somebody with a photograph in his hand. Especially since the answer doesn't address the interpretation of the photographs at all. What justification is there for dismissing the claim to a precambrian rabbit? Doesn't skepticism mean that we take such claims seriously?

Thanks for reading the Craig article[s]! Regarding unicorns, yes: I suppose that we expose ourselves to rational belief in unicorns under Craig's Bayesianism. But applying the calculation again and again, we do have P(witness is crazy | unique claim of unicorns) >> P(witness is not crazy | unique claim of unicorns) and thus plenty of rationality at our disposal without tossing out the belief-math.

If it seems like gibberish, I do suggest that you work on it some more. Craig has debated various smart folks and not embarrassed himself.

Maybe part of that is that the longer article is in fact a transcription of a classroom presentation, discussion section included. It might make more sense in the audio.

It was "No. [end of paragraph]" but not "No. [end of answer]". The reasoning was justified with evidence.

I reject (a) that debating is a useful technique to find the scientific truth, rather than a useful technique to find who has better rhetorical skills, (b) whether his personal embarassment is a measure of whether he said something foolish, and (c) that he never said foolish things that weren't ripped apart later. I can rip his argument apart myself if you like - do you have a maths background?

And, I wasn't saying it would make a Skeptics Meta question about whether we through away skepticism and follow Craig's nonsense. I was saying it might make a good Skeptics.SE question, to show it was nonsense.

I do have a math background!

Great. Can start by agreeing that when he writes "Pr(h½e&b)" he means to write "Pr(h|e&b)", because the former is gibberish?

Can we agree that he adopts a non-standard expression of Bayes Theorem by introducing the "b = our background information about the world"? It isn't wrong, in the same way "E + 43 Joules = mc^2 + 43 Joules" isn't wrong, but it is a bit jarring, and it means that you need to make sure you consistently carry the additional terms all the way through.

Wait. P(h|e) doesn't have any meaning here, because he's dropped the b term. What does it mean to ask "What is the probability of a miracle, given an eye-witness account, where every other agreed truth in the world - including our understanding of Bayes Theorem - is not a given and might be false?" Nothing. It is gibberish.

At this point he quotes from the original author who is NOT using his weird expression of Bayes Theorem, and is NOT required to carry weird "b" terms around. By using different definitions to the original author, he can quote the maths of the original author out of context and make it look false.

Later he describes a perfectly reasonable statement as "plainly wrong", when it is plainly right! But neither side makes much traction with such arguments.

>as my friend Lydia McGrew prefers to put it, “the claim is either false or it is trivially true.”

The lottery winner argument is all confused, and I don't know how to make sense of it, except to say if someone told me "I will be winning $320 million in the lottery tomorrow" and then the next day said "I won $320 million in the lottery today." I would not believe them, without very strong evidence to support it - i.e. until the point that the chance that the evidence was false was more remote than the chance that the event happened.

But bringing it back to Bayes Theorem. Using the equation he provides, if we want to show that the miracle is likely true (i.e. the LHS has a large number), and the claim is extraordinary (i.e. the first term on the RHS is very, very low), the only way to achieve it is if the second term on the RHS is very very high.

That is, the probability of the evidence (given the claim is true) is very, very high and the probability of the evidence (given the claim is false) is very, very low.

^^^ This is just another way of saying the evidence is extraordinary.