05 November 2010

Misuse of Bayes' Theorem to back up silly things

Over at Common Sense Atheism (one of my favourite blogs, recent posts notwithstanding), the redoubtable and really very clever Mr Luke Muehlausen has recently interviewed the Christian philosopher Lydia McGrew. A while back, Lydia and Tim McGrew (her husband) wrote an article in which they purported to use the mathematical theorem first derived by Bayes to weigh the evidence in favour of the resurrection of Jesus the Nazarene, in an attempt to show that it was extraordinarily *likely* to have occurred.

Regular readers of my blogarooney will have to peel their eyebrows from the ceiling, because given the stories we have in the bible, it does not at first glance appear possible that the McGrews could say such a thing. Yet that is what they argue, and they use a lot of fancy-looking mathematical footwork to show that to be the case.

Gadzooks! (I hear you cry). How in heck can they do that with some very contradictory passages that were only written down several decades after the death of Jesus the Nazarene by people who never even met him, and were not present at the supposed resurrection? And you would be right. At least part of the answer to that question lies in this statement from their article:
Our argument will proceed on the assumption that we have a substantially accurate text of
the four gospels, Acts, and several of the undisputed Pauline epistles (most significantly
Galatians and I Corinthians); that the gospels were written, if not by the authors whose names
they now bear, at least by disciples of Jesus or people who knew those disciples – people who
knew at first hand the details of his life and teaching or people who spoke with those
eyewitnesses – and that the narratives, at least where not explicitly asserting the occurrence of a
miracle, deserve as much credence as similarly attested documents would be accorded if they
reported strictly secular matters.

Now just hang on there a minute, my good man and lady - you are proceeding on some mighty questionable assumptions there! We actually have very very little evidence on which to base such a strong conclusion, and when you take these stories, such as they are, together, it is not at all clear what sort of resurrection we are talking about. At the very best, it seems that the body of Jesus the Nazarene might have gone missing, and in the cognitive dissonance, confusion and grief that followed, certain vulnerable folks claimed to have seen visions of him, and concluded that in some sense he had been "raised" from the dead. That we can make a connection between this sorry state of affairs and an *actual* miraculous resurrection seems rather hopeless.

But why do the McGrews do this? Why do they need to front-load their argument with such a contentious and unsupportable set of assumptions? The answer, it would seem, is that it helps buttress their argument and avoid the sorts of Bayes factors that undermine their conclusions, as long as they declare them and neutralise them up front, rather than factoring them into the argument.

And what constitutes a "similarly attested document"? The external attestation of pretty much anything reported in the gospels is nil (even leaving aside the fact that Matthew and Luke grabbed most of their narrative from the pre-existing gospel of Mark, plus other sources). It's all rather suspicious.

But there is a wider malaise at work here - ever since the dolorous Richard Swinburne let rip his own foray into Bayesian territory, some Theistic Christian apologists (not to be confused with the growing ranks of Atheistic Christians) have leapt on Bayes like squirrels on a Snickers to try to use its seemingly arcane powers of mathematical robustness to splint the legs of their teetering sacred cows, and try to bamboozle poor philosophers (who really have trouble dealing with mathsy things) into thinking that the arguments for ancient imaginary space pixies are a lot stronger than they actually are. But Bayes' Theorem is a wonderful little tool - we use it in Genetics all the time, and it is a common feature of many risk estimation algorithms and programmes. Do not try to fool Bayes - Bayes will find you out!

So things are not going well for the paper so far. In a future post I will proceed a little further into this curious article to see if any nuggets gleam from within the vein, and maybe explain a little more about the remarkable Bayes' Theorem as we go.

Spoiler alert: I will destroy Swinburne and the McGrews in the process, and show beyond reasonable doubt that the resurrection did not, in fact, occur. Sorry to give the game away so early! You're drooling already - I can tell!


  1. 'Spoiler alert: I will destroy Swinburne and the McGrews in the process, and show beyond reasonable doubt that the resurrection did not, in fact, occur. Sorry to give the game away so early! You're drooling already - I can tell!'

    I'll be interested to see how this turns out.

    Christ once said 'destroy this temple and I will raise it up again after three days' and we all know how that turned out...no wait...

  2. Can't wait for this. I take serious pleasure in the ruthless smackdown of Christian apologetic sophistry.

  3. Hi Andrew, well, that's one thing about a house of cards - it can get rebuilt :-) I think I may need a couple of intro posts about Bayes' Theorem and how not to do it wrong...

    Also, that "after three days" thing is fun - it was really only about 36 hours. But you knew that. Third day, not "three days and three nights"...

  4. It doesn't say much for the early church that they couldn't count to three.

    I'll leave it there, I have much to do and I have to go out in 3hrs 28 minutes.

  5. Andrew,

    And how many seconds? That would be critical to my trusting you!


    Don't worry about it, the early church obviously couldn't count to 2, never mind 3 ;-)

  6. Actually, this is one of the lines of evidence that I think could be used against the "mythicists". If the Jesus story was entirely fictional, it is unlikely that they would have had this "third day / three days" difficulty. However, clearly they were working off some dogmatic interpretation of a prophecy that "three days" were required (like Jonah in the whale, except that was "three days and three nights", which is a completely different concept, even in second temple Judaism).

    Anyway, it looks like I need to do a brief intro to Bayes. Swinburne should be a lot easier to demolish than the McGrews; I'll need to visit the library to see if I can get a copy of his 2003 tome...

  7. You know, Shane, if only I had time to answer, but I'm working day and night at the moment.

  8. It'll still be here. Look, I really do appreciate you and Andrew coming over to view my little blog, accustomed as you are to the hearty fare on W&T ;-)

    I don't want to get sidetracked, but you are getting the point that the gospel writers (of whom we know only one, remember, and even that is a bit dubious) quite frequently took the prophecy (perceived or otherwise) first, and twisted the story to make it fit. The most likely historical bits are, ironically, where the fit is least comfortable. This could be the topic of another post, but the "3 days" malarkey is one such possible example. The "This is Jesus the Nazarene, King of the Jews" is another.

    Come to think of it, the misfits should be incorporated into any proper Bayesian analysis - I'll have to see if Swinburne does this, but I'm not optimistic...

  9. I have a correction to make in my previous post.

    I said I was going out in 3hrs 28 minutes (for the record Peter it was 28 minutes exactly - how do you feel about milliseconds?).

    Well, I need to correct this; I was at least 3 minutes late and in the panic I lost count of the seconds. I did go out though.

    At least now I've committed perjury I can get out of jury duty. Tis as Cowper says, 'The Lord works in mysterious ways'.

  10. what didn't you like over at CSA recently?

  11. Hi Jeremy, no, no disagreements with Luke (major ones anyway); I was referring to the meteoric rise in 27b/6. Sorry for any confusion :-)

  12. Hi Jeremy, no, no disagreements with Luke (major ones anyway); I was referring to the meteoric rise in 27b/6. Sorry for any confusion :-)

  13. For the record, the "three days" fits a little better if we go with John's Gospel which has Jesus strung up Thursday afternoon. Not that any "Bible Believing" Christian would accept what's written in the bible...

  14. When a Christian starts using complex mathematical formulas and philosophical theories to defend his belief in first century corpse reanimation-transformation (aka:
    resurrections)…I yawn.

    I yawn because it is soooo silly.

    I know for a fact that if a Muslim attempted to use these same ploys to defend the veracity of Islam’s claim that Mohammad flew to heaven on a winged horse, the very same Christians would snicker and hand-wave away these arguments without giving them a second thought, believing that these tactics are nothing more than an obvious, desperate attempt to dress up a superstition as believable reality.

  15. Machine Learning Projects for Final Year machine learning projects for final year

    Deep Learning Projects assist final year students with improving your applied Deep Learning skills rapidly while allowing you to investigate an intriguing point. Furthermore, you can include Deep Learning projects for final year into your portfolio, making it simpler to get a vocation, discover cool profession openings, and Deep Learning Projects for Final Year even arrange a more significant compensation.

    Python Training in Chennai
    Python Training in Chennai

    Angular Training Project Centers in Chennai