Getting a Handle on Extraordinary Claims

The phrase “extraordinary events require extraordinary evidence” is largely inspired by Hume’s claim “that no testimony is sufficient to establish a miracle, unless the testimony be of such a kind, that its falsehood would be more miraculous, than the fact, which it endeavours to establish.”  If Hume is right, then it seems that he’s effectively vetoed any alleged evidence for purported miracles.  

But is he right?  Consider a short exchange: 

“Do Christians really believe in miracles? . . . they seem a bit unlikely.”

--”If miracles weren't unlikely, they wouldn't exactly be miracles, would they?”

The improbability of miracles is what enables them to serve as validation for religious claims.  If miracles were happening continually, they wouldn’t be able to serve this role.  

The prior probability for miracles is thus lower than normal events, so they do require more evidence to establish that they occurred than normal events require.  But admitting that an event has a lower prior probability does not mean that it has a zero probability.  If a purported event has some prior probability, then it can in principle be sufficiently supported to warrant belief.  Priors tend to wash out as evidence accumulates.

For some miraculous event P, the skeptic seems to have in mind that we need some equally extraordinary *single* piece of evidence to establish that P occurred.  But it is possible for *ordinary* evidence to *cumulatively* add-up to overcome the low prior probability of some miracle claim.  It may be the case that no individual piece of evidence is spectacular or sufficient, but added together they can offer enough evidence.  A cumulative collection of known reliable witnesses, can, in principle, give sufficient evidence to overrule a low prior probability.  

There’s a vagueness in the phrase’s use of “extraordinary.”  If it’s taken to mean “sufficient evidence,” then I agree.  Then the only relevance a purported miracle claim has to the phrase is that its miraculous nature reduces the prior.  Using the word “extraordinary” has the effect of raising the bar of evidence so high that it cannot possibly be met, and it tends to reinforce in its proponents minds the error that each piece of evidence should be taken singularly and not cumulatively.  

It seems possible to specify any event whatsoever to such a degree that’s prior probability is extremely low.  Joe married Sally, for instance.  The probability that Joe and Sally would evolve from single-celled organisms, exist at the same time, run into one another at some restaurant, and so on, is extremely low.  We can specify the event to such a degree that it’s wholly unprecedented.  Do we thereby need extraordinary evidence to believe that they are in fact married? Presumably not.  

But the claim that Jesus resurrected is assuredly different from the claim that Joe married Sally.  This difference seems to stem from the fact that Joe instead marrying Veronica also could have been so uniquely specified--and this chance probability seems to cancel out the improbability of him marrying Sally.  But if, in the case of the lottery, a person wins the lottery and they had a relative working on the lottery’s oversight committee, then it seems more probable that the hypothesis that they cheated is true.  There’s not a “cancelling out” hypothesis on offer.

In the case of the resurrection, it seems that it may also be explicable on a suspect-hypothesis.  Perhaps the apostles were hallucinating, or lying, or are mythical, etc.  This is a proper advancement of the discussion, as it allows us to get down to brass tacks and discuss the probability of each of these rival hypotheses (which, I think, can be adequately met by the apologist.)  

Further; it isn't just the prior that's relevant in this question.  We also need to gauge the probability of the evidence given that the relevant hypothesis is false.  What's the probability that the nightly news would announce that lottery number if that number were in fact false?  If it's quite low, this will wash out the prior improbability of that number having been called.  Similarly, what's the likelihood of the empty tomb, resurrection appearances, and belief in the resurrection in early Christianity given the falsehood of the resurrection?  Quite low, I would think.  And this can wash out the prior unlikeliness of the resurrection.  

The stock of prior beliefs that both believers and atheists bring to bear on the miracle of the resurrection are going to differ as well.  Christians, presumably, come to the question with the assumption that other miracles have in fact occurred.  This raises the prior.  Atheists will come to the question that no other miracles have occurred. This will lower the prior.  

Hume claimed that miracles violate the uniform experience of mankind that miracles do not occur.  But that’s simply false.  There are many miracle claims throughout history.  Craig Keener has documented many, too.  

Relevant factors in estimating the prior: 

(a) Arguments for the belief that God exists.  This raises the prior of any purported miracle, as the existence of a God capable of performing miracles surely increases the prior probability that a miracle has occurred.  (b) Religio-Historical Context; Historical reasons God may have for performing a miracle, such as verifying a certain message, or surprising interlocking of Jesus’s life with OT symbols (c) Argue from natural theology that God’s character would make it more likely that he would engage in certain miracles and courses of action that imply miracles. 

The above post is based on notes that I wrote when reading an excellent discussion of the topic by Jonathan McLatchie, which can be found here.