Hitchens's Razor vs Bayes's Theorem
John W. Loftus has just endorsed my forthcoming book on the Resurrection in the following manner:
Hitchens’s Razor, not Bayes’s Theorem, is the proper tool to use against the “absolute baselessness” of the resurrection belief (per David F. Strauss, as quoted in this book). There’s no objective evidence for it. The testimonial evidence is abysmally poor. We should therefore dismiss this superstitious belief for what it is (per Hitchens). However, if you want to take such a belief seriously, read this thoroughly documented terminal case against the resurrection based on the latest research! This is the only book you'll need. Pearce is your expert guide on all the essential issues.
--John W. Loftus, author of Unapologetic: Why Philosophy of Religion Must End, and editor of The Case against Miracles.
I am hugely grateful to him and have returned the favour for his forthcoming book on the Problem of Evil and suffering, which is another fantastic (and his last) anthology. He really has written and edited a number absolute corkers, each and every one.I use Bayes's Theorem as a method to evaluate historical claims concerning the Gospel accounts. I think it formalises what we do every day in making probability evaluations (including, shall I cross the road now, in this busy traffic or, as one of my children asserted: "My Fortnite account has been hacked as it's not loading properly."**). However, and he does this in one of the chapters in his forthcoming book, Loftus believes that we should not use Bayes's. But not in the way that some Bayes's detractors claim. The way I understand him, he is saying it's simply not worthy of investigating, almost out of hand. Indeed, this quicker evaluation may just be a case of assessing the prior probabilities and dismissing the evidence as amounting to nothing! The idea, as he told me privately, is that it also "legitimizes that belief for people by taking it seriously." I agree.
You could see this as similar to dealing with Marjorie Taylor Greene: giving her airspace and column space is just giving oxygen to her ridiculous views.
I am torn. On the one hand, it is sensible to ignore such ridiculous claims (Greene or the Resurrection) but, on the other hand, when such claims reach a critical mass amongst a population, they need addressing. Although Loftus expands on his idea of Hitchens's Razor in his latest book, he originally expressed his views on his Debunking Christianity blog (where I used to write many years ago), the maxim being, "What can be asserted without evidence can also be dismissed without evidence". One of the key points is as follows. Bayes's Theorem (BT) is about probabilities. To do it formally, you actually need to plug in some numbers into a formula. If you don't have any numbers to plug in, you can't get any numbers out. It's not a case of garbage in, garbage out, but zero in, zero out. Loftus has five issues with using BT:
1) With miracles there is no previous data to work from.Bayes can only be useful when there is prior data to work from. We’re told every logically possible claim has a nonzero probability to it. But miracles don’t have any prior probability. A flying pig would be a miracle. So we need prior data to work from. How many pigs have ever flown of their own power? Without any previous data Bayes isn’t the proper tool to use here. All we know is that there is no objective evidence for such a thing. That’s more than enough.
In a sense, Loftus is spot on. However, it depends on your reference set. Perhaps, instead of "How many times have miracles taken place?", and getting a nil answer, it could be, "How many times have people made false claims about a miracle happening?" or similar. Loftus continues:
2) Bayes won’t help clarify our differences. We don’t need Bayes to know where our differences are to be found. We already know. The main difference between us is that believers value faith, blind faith, the only kind of faith there is, faith without objective evidence, while nonbelievers value sufficient objective evidence. That’s why we're nonbelievers, and that's why believers continue to believe based on ancient 2nd hand hearsay testimony about a miracle. 3) Bayes gives undue credibility to some miracle claims over others when none of them have any objective evidence for them. I’ve written a book called Unapologetic on why responding to fundamentalist arguments in kind gives their beliefs a certain undeserved respectability. To treat the resurrection story as if we have some objective evidence for it when we don’t, is to give it undeserved credibility over other miracles, especially the ones located in very gospel texts where we read of the resurrection. Why is no one doing a Bayesian analysis of the virgin birthed son of god? That's the point!
I think Loftus has some points here. There are certainly cultural biases that mean that certain miracle claims are given much more visibility and credence than others, in an epistemologically unwarranted sense. There is something to be said, though, for what if that resurrection miracle really did happen? Is Loftus merely expressing a metaphysical naturalism and unfairly dismissing the biblical claims?
4) Bayes won’t help convince anyone. Bayes is probably worse off in terms of convincing others, for the only people who would sludge through it are far less likely to be convinced by it. Just ask to see the objective evidence, and if it’s lacking, like it is, then dismiss it.
Loftus goes on to detail how different thinkers have produced wildly different percentage chances given varying numerical assignments. This variability and the nature of the argument means that, to him, it simply isn't that persausive. Loftus lays out the final argument against using BT:
5) Imagining what might convince us is largely an exercise in futility. Bayes asks us to imagine what might convince us of a given hypothesis. This is a reasonable request in criminal trials, and other kinds of scenarios where actual evidence is being considered. But in order to imagine what would convince us of miracles, it would require changing the past, and that can’t be done. If I could go back in time to watch Jesus coming out of a tomb, that might work. But I can’t travel back in time. If someone recently found some convincing objective evidence dating to the days of Jesus, that might work. But I can’t imagine what kind of evidence that could be. As I’ve argued, testimonial evidence wouldn’t work, so a purported handwritten letter from the mother of Jesus is insufficient. If a cell phone is discovered and dated to the time of Jesus, which contains videos of him doing miracles, that might work. But come on, this is as unlikely as his resurrection. If Jesus, God, or Mary themselves were to appear to me, that might work. But that has never happened, even in my believing days, and there’s nothing I can do to make it happen either. In any case, imagining what could convince us of a miracle only arises when using the hammer of Bayes on the nail of miracles. Imagining evidence that could convince us Mary gave birth to a divine son sired by a male god is a futile exercise, since we already know there’s no objective evidence for it. One might as well imagine what would have convinced us in 1997 that Marshall Applewhite of the Heaven’s Gate suicide cult, was telling the truth that an extraterrestrial spacecraft following the comet Hale–Bopp was going to beam their souls up to it, if they would only commit suicide with him. One might even try to imagine today what would convince us that he and his followers are now flying around the universe. Such an exercise is utter tomfoolery.
There is not too much about which I disagree with Loftus here. However, though I grant him much of what he says, in the same way there is intuitive value in Hitchen's maxim, there is also intuitive and rational value in Sagan's maxim that "extraordinary claims require extraordinary evidence".
And this, to me, is the value of BT. That if something is a really unlikely event, then to see it as actually having happened as claimed, you will need exceptionally good quality evidence.In this way, you could be forgiven for thinking that analysing how the four Gospels contradict each other on this claim here, and that claim there; that this theology of this pericope doesn't square with that theology there; or that this historical claim here doesn't cohere with known history over there. This is labour intensive and the sort fo work I have done. Was it a waste of time? Depends who you are and what you like reading. Will it convince anyone? Not nearly as many people as I would like. Will it give the oxygen of credibility to ridiculous ideas? I think that ship has sailed. There are many reasons I wrote the book, and I definitely enjoyed doing so, and it certainly does a good job of defeating the apologist claims of the historical reliability of the Resurrection claims (in my humble opinion). In a sense, what work like mine does is to establish the very premise upon which Loftus' argument is based: that the evidence quality level is, indeed, exceptionally low. What Hitchens is really saying is "What can be asserted with very low-quality evidence can be dismissed without analysing that evidence in great depth." There are two things to say to this.
- Loftus is right to take this approach if he has already got good reason to believe the premise that Christian claims amount to "very low-quality evidence". And he has written a huge collection of truly brilliant books, and has researched the bejesus out of all of these topics. He is justified from his point of view.
- For someone else, however, that belief that Christian claims amount to "very low-quality evidence" is something that might need to be established. My book is part of establishing this.
There is a chicken and egg scenario here.
For many people, they need to rationally evaluate that Christian claims do, effectively, amount to the epistemological equivalence of "without evidence" in Hitchens's Razor. That is what my book seeks to do.
Therefore, we're both right! Or something. Either way, thanks hugely to Loftus for endorsing the project, and remember:
...if you want to take such a belief seriously, read this thoroughly documented terminal case against the resurrection based on the latest research! This is the only book you'll need. Pearce is your expert guide on all the essential issues.
NOTES **The internet ethernet cable wasn't plugged into the PS4 properly. I used probability evaluation based on background knowledge about hacking and PS4s, and the prior probabilities of such claims he has often made, to evaluate that his claim was nonsense. I tried explaining this to him, but he's ten. [A note for Grammar Nazis: In my book, I use Bayes' Theorem, as is generally accepted due to what the grammatical style was at the time it was created. However, with Hitchens and Loftus having the standard 's added afterwards, I decided to keep it consistent and add 's to all of them. Personally, I prefer adding an apostrophe alone to words ending in s, but that's just me.]
Stay in touch! Like A Tippling Philosopher on Facebook:
A Tippling Philosopher
You can also buy me a cuppa. Or buy some of my awesome ATP merchandise! Please... It justifies me continuing to do this!
