Stephen Law’s paper Evidence, miracles and the existence of Jesus argues that the New Testament (NT) is not good evidence for the existence of Jesus. He takes an interesting approach: he argues that the evidence for the NT miracles isn’t good enough, and that the presence of the miracle stories contaminates the non-miraculous parts of the story such that we should be sceptical of those too.
Law introduces and defends two principles:
P1 Where a claim’s justification derives solely from evidence, extraordinary claims (e.g. concerning supernatural miracles) require extraordinary evidence. In the absence of extraordinary evidence there is good reason to be sceptical about those claims.
P2 Where testimony/documents weave together a narrative that combines mundane claims with a significant proportion of extraordinary claims, and there is good reason to be sceptical about those extraordinary claims, then there is good reason to be sceptical about the mundane claims, at least until we possess good independent evidence of their truth.
He then uses these in a deductive argument, concluding that “there’s good reason to be sceptical about whether Jesus existed”.
Most of the debate in the comments on Law’s blog is about P2. Law says that “Because once we know that a powerful, false-testimony-producing mechanism (or combination of mechanisms) may well have produced a significant chunk of a narrative (e.g. the miraculous parts), we can no longer be confident that the same mechanism is not responsible for what remains.”
Bradley C. came up with some counter-examples to P2. Bradley rightly says that the false-testimony-producing mechanism is key. What feels different about the ancient miracle reports (and perhaps Law’s “sixth islander” thought experiment) compared to Bradley’s examples is that in the ancient reports, we don’t really know what the mechanism was, we just know something has gone wrong. (In Bradley’s examples, we know that magicians and faith healers do tricks). If we don’t know quite what has gone wrong, we have to consider various possible mechanisms, which includes ones where the mundane testimony is also false. If we give such mechanisms any weight, that makes the mundane testimony less convincing (though it may still be positive evidence for the mundane events). But I think we’d have to consider how much weight to give them based on the circumstances, which makes it hard to come up with something general like P2.
So, I think Bradley’s come up with the equivalent of Gettier cases for P2 as it stands: even if they’re contrived, they show P2 needs changing.
Law responds to Bradley saying “You need to identify a mechanism as being the likely mechanism accounting for the false miracle claims, and then explain why that mechanism wouldn’t quite likely result in false mundane claims too.”
I don’t agree with Law here. If all we know is that something’s gone wrong with the testimony but the mechanism is obscure, perhaps it’s reasonable to say that it’s as likely that we’d have the testimony if it’s mundane parts were true as it is that we’d have it if the mundane parts were false. Then the testimony is no evidence for or against the mundane events: you should consider the events as likely as you did before you heard the testimony.
I’m not sure I’d want to go further than that and say that the burden of proof is on the people who believe the mundane portion of the testimony to show why it isn’t contaminated: mightn’t they equally well argue that the burden is on you to show that it is? But that’s what P2 says, I think: in P2, the testimony becomes evidence against the mundane events.
If you give a mechanism, though, maybe that’s just what you can argue: if you think Jesus’ disciples made it up, for example, who’s to say where the made up stuff ends? (Though why not make stuff up based on a real person, for verisimilitude?)
It looks like someone who wants to justify their belief in the mundane stuff has a motive to push the unbeliever to identify the mechanism so they can criticise it. The problem with my “average over possible mechanisms” idea, above, is that it’s pretty hard to identify them all. I don’t think we have a duty to do that with every weird testimony, though. Earlier, in defence of P1, Law correctly says that “the fact that it remains blankly mysterious why such reports would be made if they were not true does not provide us with very much additional reason to suppose that they are true.”
So, I’m not that convinced by Law’s general contamination principle, but I think he makes some good points along the way. For example, Law says:
It would also be foolish to try to construct a two part case for Jesus’ miraculous resurrection by (i) bracketing the miraculous parts of the Gospel narrative and using what remains to build a case for the truth of certain non-miraculous claims (about Jesus’ crucifixion, the empty tomb, and so on), and then (ii) using these supposedly now “firmly established facts” to argue that Jesus’ miraculous resurrection is what best explains them (yet several apologetic works – e.g. Frank Morrison’s Who Moved The Stone? – appear implicitly to rely on this strategy).
William Lane Craig’s rebuttal
The apologetical strategy Law talks about is used by William Lane Craig in his “4 facts” defence of the resurrection (see Craig vs Ehrman, for example). Craig read Law’s paper and attempted a rebuttal on his own blog, which I think was only partially successful.
Craig’s stuff about Ehrman is weird. I guess Craig’s point here is to show how reasonable he’s being by pointing out that even this bloke he beat in a debate (Ehrman) agrees with him. But Ehrman is not a radical sceptic, Law is not die-hard mythicist. The conclusion of Law’s argument is that we should be sceptical about J’s existence, not “Therefore J never existed”, so it’s not even clear that Ehrman’s ire applies to Law, or that we should care if it does, unless Ehrman’s arguments are made more explicit.
On Sagan’s dictum that “extraordinary claims require extraordinary evidence”, Craig writes: “This sounds so commonsensical, doesn’t it? But in fact it is demonstrably false. … Rather what’s crucial is the probability that we should have the evidence we do if the extraordinary event had not occurred. This can easily offset any improbability of the event itself.”
Craig makes a reasonable statement of Bayes Theorem. However, Sagan’s dictum can be read in a Bayesian way (by incorporating all the probabilities Craig mentions, so that the evidence is Bayesian evidence). Craig gives no good argument that the dictum must mean what Craig takes it to mean, or that Law’s argument relies on taking it to mean what Craig thinks it means.
Craig continues: “In the case of the resurrection of Jesus, for example, this means that we must also ask, “What is the probability of the facts of the empty tomb, the post-mortem appearances, and the origin of the disciples’ belief in Jesus’ resurrection, if the resurrection had not occurred?” It is highly, highly, highly, improbable that we should have that evidence if the resurrection had not occurred.”
This might be Craig’s attempt at that argument, namely, Craig saying that Law hasn’t considered that it’s unlikely we’d have the evidence we do if Jesus didn’t do miracles. But Craig plays fast and loose: the facts are that we have the gospel narratives (and whatever other historical documents we have to hand). The empty tomb and post-mortem appearances are not facts, and Law’s argument against the “bracketing” strategy is that they cannot be treated as facts. Craig cannot have the empty tomb or the post-mortem appearances as “facts” without addressing Law’s arguments.
Oddly, Craig doesn’t address really P2 or Law’s arguments for it at all: he just says “oh no it isn’t”. Craig’s strongest when he says that there is extra-Biblical evidence for Jesus’ existence. I’m not an expert, but my understanding is that Josephus’ mentions of Jesus is thought by historians to have a core around which Christian interpolations accreted, for example. Since even if we grant P2, Law’s argument fails without premise 6 (“There is no good independent evidence for even the mundane claims about Jesus (such as that he existed)”), perhaps this is a good tactic on Craig’s part. Law appears to agree that premise 6 is his weakest empirical premise: “6 is at the very least debatable”. In a way, it’s odd that everyone is concentrating on P2.
So, I think Craig casts doubt on Law’s conclusion about Jesus’ existence, but he doesn’t do much to convince us that Jesus rose from the dead or did any other miracles.
Jerry Coyne’s blog has some good comments on Craig’s rebuttal.
David B Marshall’s rebuttal
David Marshall also had a go at rebutting Law. He didn’t do as well as Craig, as his arguments relied on attempts to differentiate Law’s thought experiments (“Ted and Sarah”, and “The Sixth Islander”) from the claims about Jesus, but the distinctions he made between these, distinguishing magic from miracles, weren’t relevant to Law’s arguments, as far as I can tell. You can see my response to him here , his reply here and my response to that here, another attempt by Marshall here and my final response here.
I think this rebuttal is interesting for what it shows about what ordinary believers (rather than super-apologists like Craig) think are good arguments. Marshall appears to think that because the Jesus story is more fleshed out and more meaningful, it’s more likely to be true. I’m not sure whether this is a straightforward example of conjuction bias (obligatory Less Wrong link), or of the notion that the point of religion is to be in a meaningful story. Charitably, it might be an attempt at inference to the best explanation, but I don’t think the stuff that Marshall mentions means that the best explanation of the NT stories is that they are true.
So what do you think?
There was bloke called Jesus who was the basis of the NT stories. Pre-moderns had porous selves, so it’s pretty difficult to understand their writings in modern terms, but there is no good evidence that this bloke did miracles or rose from the dead. I don’t know how much of the NT is true, but I don’t accept Craig’s bracketing or 4 facts arguments: taking out the core miracle but leaving the context which points to a miracle does look like cheating without independent evidence of the context, because mechanisms where both the context and miracle are made up seem pretty likely to me.