• How (not) to argue about the resurrection


    About two years ago, in a post entitled How to argue about gods, I stated my intention to write a series of posts about the main arguments of natural theology – the Cosmological, Fine-Tuning, Moral, Ontological Arguments, as well as more Christian-specific arguments such as those from personal experience and from the historicity of the resurrection.  Well, life has been busier than I’d have liked (though I’ve been kept busy with good things) and I haven’t yet been able to devote the serious amounts of time I wish to give this project, though I’ve made a start on a few of the arguments – for example:

    So, although I haven’t discovered a mountain of free time, I’ve decided to bite the bullet and write a series on the resurrection.  Although I had originally planned to tackle arguments such as the fine tuning and moral arguments first (since the argument from the resurrection builds on the others, in the sense of moving from an almost deist god to the specific God of Christianity), various considerations have led me to treat the resurrection next.  Actually, it seems reasonably fitting to go straight for the resurrection as it is the arguably the heart of Christianity.  As the apostle Paul himself said:

    [I]f Christ has not been raised, our preaching is useless and so is your faith.  (1 Corinthians 15:14, NIV)

    Although it’s hard to get all Christians to agree on anything, the resurrection of Jesus is certainly a central pillar in the faith of most believers.  What’s more, quite a few of my Christian friends have told me that they are Christians because they believe there is good historical evidence for the resurrection.

    So the next few posts will be on the resurrection, and I intend to structure the series around the modern style of argument made by apologists such as William Lane Craig – see for example this post of Craig’s or this one from Gary Habermas.  You can actually get a good introduction to some of the arguments for the resurrection (and skeptical responses to them) by watching any of the many debates on the topic, and I highly recommend any of the following (featuring some of the more entertaining and well informed scholars going around):

    The modern argument for the resurrection is generally comprised of two parts:  it is argued separately that

    (1)  certain claims made in the gospels (that Jesus was crucified, that his tomb was found empty, that the disciples and others believed Jesus appeared to them after his death, and so on) can be established as historical facts, and

    (2)  these “facts” are best explained by the resurrection.

    This approach is known as the Minimal Facts approach and is utilised by a number of apologists (Craig, Habermas, Licona, etc).  I have considered the general validity of such an approach in another post, but in this series I will tackle both parts of the Minimal Facts argument head on.  Specifically, I will argue that:

    (1)’  there are good reasons to be suspicious of some of the proposed “minimal facts”, and

    (2)’  even if we grant the proposed “minimal facts”, the resurrection is not the best explanation.

    But this introductory post does not argue directly against the Minimal Facts argument.  Rather, it considers what a skeptic needs to do to deflect the Minimal Facts argument, or – more to the point – what a skeptic does not need to do.  Specifically, I want to take this opportunity to clear up a crucial (and far too common) misconception.  Consider the following words from the mouth of William Lane Craig:

    In order to explain that the resurrection is improbable, [a skeptic] needs not only to tear down all the evidence for the resurrection, but he needs to erect a positive case of his own in favour of some naturalistic alternatives.  (William Lane Craig, debate with Bart Ehrman, 2006 – transcript)

    I actually hear this kind of claim quite a lot, but it’s simply not true, and it’s very disappointing to hear someone of Craig’s reputation make it.  But since Craig and others do make it – all the time – it’s important to be clear about exactly why it’s wrong.  The problem is this: in order to explain why you think a certain event is unlikely to have happened, you don’t have to demonstrate that some other event happened instead.  You don’t even have to demonstrate that another event is more likely to have happened.  Let me illustrate with a couple of examples.

    Suppose I decide to play a guessing game with my students one day.  I ask Adam to go into another room where he finds a huge barrel filled with thousands of marbles each labelled with one of the numbers 1,2,3,4,5,6,7,8,9,0.  Adam is to pick 20 marbles at random and write the sequence of  labels on a blackboard in the order he draws them from the barrel.  When he returns, I announce to the class “I claim that Adam wrote 00000000000000000000”.

    Needless to say, my students are now questioning my mathematical ability.  There are 10^20=100,000,000,000,000,000,000, or one hundred quadrillion possible sequences Adam could have written on the board, each of which is just as likely as any other – how could I expect to guess correctly?  My chances of being correct are 0.000000000000000001%.  I am almost certainly wrong.  But could my students “erect a positive case” in favour of any of the other 99,999,999,999,999,999,999 possible sequences?  Certainly not!  What reason could you possibly give to suggest that 76989331655938885021, for example, is more likely than 00000000000000000000?  All of the possible sequences are equally likely.  So, even though such a “positive case” could not possibly be “erected”, my claim is still highly unlikely to be true.  The point is not that there is some single alternative sequence that is more likely than the claimed one – it’s that there are so many other possibilities.  To assess the likelihood of my claim, we don’t compare the “all zeros” hypothesis to some specific “sequence X is on the board” hypothesis.  We compare the “all zeros” hypothesis to the “anything other than all zeros” hypothesis.

    Let me stress that I’m not claiming the resurrection is just one of 100,000,000,000,000,000,000 equally possible explanations for the proposed “minimal facts”.  I’m simply explaining that one event can be exceedingly improbable even though it’s impossible to argue in favour of any rival event.  But even more can be said.  In fact, an event can be exceedingly improbable even if it is overwhelmingly more probable than any other given event.  Let me explain.

    In the next lecture, I decide to play another guessing game with my students, but this time I change the rules a little.  This time, Brenda is to go into the other room where she finds the same barrel filled with the same labelled marbles.  But there is also a coin on the table.  Brenda’s instructions are as follows.  Before touching any marbles, she is to toss the coin 20 times.  If she gets 20 consecutive tails, she is to simply write 00000000000000000000 on the board.  If she gets any other sequences of heads/tails, she is to use the marbles to produce a 20 digit sequence (just like Adam did) and write this on the board.  When Brenda returns, I again announce to the class “I claim that Brenda wrote 00000000000000000000”.

    What do we make of this situation?  What are the odds that I am right?  As it turns out, there is about a 0.000095% chance that the blackboard contains an all zero sequence – worse than one in a million [the boring details of the calculation are at the end of the post].  So I’m almost certainly wrong.  Again.  That small number may sound very unimpressive.  But appearances change a little when you compare it to the odds of some other given sequence – for any other conceivable sequence (76989331655938885021, say), the chance that this is the sequence written on the board is around 0.00000000000000000099%.  Again, this doesn’t seem all that meaningful – another tiny number.  But if you divide 0.000095 by 0.00000000000000000099, you discover that an all zero sequence is around 95 trillion times more likely than any other specific sequence.  (In particular, this blows out of the water any hope that a student might “erect a positive case” in favour of some rival sequence.)  So does this mean I’m likely to be right this time?  Of course not!  Even though the “all zeros” hypothesis compares extremely well to any other specific “sequence X is on the board” hypothesis, this is not the appropriate comparison to make.  As before, we need to compare the “all zeros” hypothesis to the “anything other than all zeros” hypothesis.  And the “all zeros” hypothesis has only a 0.000095% chance of being true – worse than one in a million – while the “anything other than all zeros” hypothesis has a better than 99.99% chance of being right.  Even though my prediction is a whopping 95 trillion times more likely to be true than any other specific prediction, I am almost certainly wrong.  My claim is crushed under the combined weight of the alternatives.

    Let me stress that the above examples are not intended to be directly analogous (numerically equivalent) to the question of the resurrection.  Rather, the stories about coins and marbles, and the ludicrous numbers chosen, are simply given to illustrate the fact that an event can be exceedingly improbable even if it is more probable than any other given alternative event.  Even if we could imagine that someone somehow demonstrated that the resurrection was more likely to be true than any other specific explanation of the proposed “minimal facts” (and I should hasten to point out that this is a gigantic “if”), this would not be enough to establish the resurrection as probably true.  It still might or might not be probably true, and this would depend on several other factors, including the number of alternatives.

    Let me also stress that I am not claiming the resurrection can be disproved merely by arguing that there are a huge number of alternative explanations for the proposed “minimal facts”.  As I said above, this post is not part of an argument against the resurrection.  Its purpose is simply to point out a crucial error in a claim of several apologists.  Contrary to the assertions of people like Craig, the skeptic does not have to “erect a positive case … in favour of some naturalistic alternatives”.  As it happens, I think there are plenty of perfectly plausible naturalistic alternatives that do at least as good a job of explaining the proposed “minimal facts” as the resurrection does (more on this in subsequent posts).  But I also think it would be extremely difficult, if not impossible, to pick a single such alternative hypothesis and argue strongly that this is really what happened way back then in the first century.

    But the burden of proof is on the apologist.  The apologist should not demand a skeptic prove some specific “non-resurrection” hypothesis.  Rather, the apologist needs to show that the “resurrection” hypothesis is more probable than the “anything but the resurrection” hypothesis.  In my opinion, no apologist has ever succeeded in doing this, and this series will outline my reasons for coming to this conclusion.


    Due to a couple of people objecting that the above examples don’t apply to the resurrection because the “alternative hypotheses” were all equally probable, I’ve posted another example that doesn’t suffer from this unfortunate distraction from the main point (which still stands) – check out Coins, marbles and the resurrection.


    Boring appendix – a little calculation

    There are two scenarios that could lead to an all zero sequence in the second “guessing game” story described above: either (i) Brenda got 20 tails in a row so the all zero sequence was chosen by the coin (so to speak), or else (ii) she got some other sequence of heads and tails but then, entirely by chance, drew 20 consecutive marbles labelled 0.  The odds of case (i) occurring are 1 in 2^20, or 1 in 1,048,576, ie, about 0.000095%.  The odds of case (ii) occurring are (2^20-1)/2^20 (the odds of getting anything other than 20 consecutive heads) multiplied by 1/100,000,000,000,000,000,000 (the odds of randomly drawing 20 consecutive marbles labelled 0), which works out to about 0.00000000000000000099%.  Adding the probabilities for outcomes (i) and (ii) yields approximately a 0.000095% chance that the board contains an all zero sequence.  For any other conceivable sequence (76989331655938885021, say), the odds that this is the sequence written on the board is also around 0.00000000000000000099%.  To get this value, we use the same argument used in case (ii) above – Brenda must first get anything other than 20 heads when she tosses the coins, and must then draw the marbles in just the right order.

    Category: AtheismChristianityFeaturedHistoryJesusResurrectionTheismWilliam Lane Craig


    Article by: Reasonably Faithless

    Mathematician and former Christian