Loke: Investigating the Resurrection

Also along these lines, I can rephrase the situation in light of @Faizal_Ali’s objection:

  • @Andrew_Loke’s main contribution in this new book is that he classifies all the logically possible naturalistic hypotheses according to a finite list of yes/no questions.
  • He also advances arguments for the small probability of each of these hypotheses. (A naturalistic resurrection hypothesis would also have vanishingly small probability.)
  • Assuming Loke’s arguments work, we are left with the large probability that a non-naturalistic (or supernatural) hypothesis must be true.
  • @Faizal_Ali argues that the (supernatural) Resurrection hypothesis is not the only possible supernatural hypothesis, but there are potentially a large number of these, each with their own probabilities.

My observations:
First even allowing his objection, @Faizal_Ali has to concede that there is a greater probability than normal that something supernatural happened in this case, even if we don’t know which supernatural being caused it or what it is. (Unless if he has some other objections to Loke’s specific arguments against the naturalistic hypotheses.) Thus, Loke’s argument is still effective against naturalism, even if it might not be effective for showing the truth of Christianity per se. This is I think what he is saying here (Chapter 8, pp. 197, emphases mine):

On the other hand, the miraculous resurrection of Jesus is not contrived given this religious context. Even if the resurrection of Jesus has a natural explanation which is yet unknown to twenty-first-century scientists, we still need to ask how it could have been known and utilized to resurrect Jesus in the first century and vindicate his claim to be truly divine. Such a knowledge and ability to manipulate natural laws would still require a supernatural agent in any case.

Thus, even if one might not be convinced that Jesus is the Son of God, one has to concede that in light of the historical evidence we have, it is likely that Jesus (or some other power behind him) had supernatural-like powers. (Again, assuming Loke’s arguments work.)

Second, what the objection shows is that these arguments do not work in a vacuum; as Loke himself continues after the above passage, other arguments for the truth of Christianity can improve the odds of the supernatural Resurrection hypothesis compared to other supernatural hypotheses (pp. 197-198, emphases mine):

Alternative naturalistic causes such as aliens or alternative supernatural causes such as demons are ad hoc, because there is no good independent reason for believing that an alien or a demon who had such powers to resurrect the dead exists. However, there are good independent reasons (viz. the cosmological and fine tuning arguments) for thinking that there is a God who created the universe with its laws of nature (Loke 2017b, forthcoming; Craig and Moreland 2009); a God with such powers would have no difficulty raising the dead. There are also reasons for thinking that such a God would interfere in history by becoming incarnate and that it is highly improbable that we would find the evidence we do for the life and teaching of Jesus, as well as the evidence from witnesses to his empty tomb and later appearances, if Jesus was not God incarnate and did not rise from the dead (Swinburne 2003, 2013a, 2013b; cf. Cavin and Colombetti 2013).

Third, apart from the above conclusion, even if he doesn’t spend time on the independent arguments for the existence of God (he has done that in other books), Loke does actually spend some time in Chapter 8 (specifically section 8.3) discussing some related problems in picking out a supernatural hypothesis, including Alvin Plantinga’s argument from dwindling probabilities and Michael Martin’s argument for the low probability of Jesus’ resurrection even if God exists. Both of these arguments are, in my view, more sophisticated versions of Faizal’s objection. However, I haven’t seen any engagement in this thread on this material, and I don’t have time to summarize them from scratch for everyone.

Fourth, one idea would be to consider all of the logically possible supernatural hypotheses similar to what we did with the naturalistic ones. For example, we would have the supernatural mass hallucination, supernatural Resurrection, supernatural body stealing, etc. It would be more difficult than with the naturalistic case, but perhaps one could argue against each of these supernatural alternatives by showing that they are ad hoc as opposed to the Resurrection hypothesis, by (for example) using other arguments for the truth of Christianity.

This is not a new endeavor at all: this is exactly what one would have to do to argue for the Resurrection to someone who is a Muslim, Jew, animist, or other religions. In fact in the Gospels we see the Pharisees who do not deny the reality of Jesus’ miraculous acts; they just argue that it came from Beelzebul (Matthew 12:24).


The list that I lay out is supposed to be a list of NATURALISTIC hypotheses alternative to Jesus’ resurrection. There is no gap where that is concerned. SUPERNATURAL mass hallucination is NOT a naturalistic hypothesis. Supernatural hypotheses can be evaluated using other considerations as Daniel Ang noted. One does not need to embrace Faizal’s unrestricted methodological naturalism to rule out all supernatural hypotheses, which as I have explained before is unreasonable because a supernatural hypothesis ex hypothesi is not supposed to be a hypothesis about how the natural world when left on its own operates. Rather, one should use methodological naturalism to evaluate naturalistic hypotheses only, and when one discovers that all of them failed to explain away Jesus’ resurrection, we have identified a miracle and refuted atheism, and we then go on and evaluate different supernatural hypotheses using other considerations as Daniel noted.


Daniel Ang has summarized the situation well in his latest post on this thread. His observations and his quotations from my book would be my reply to you.

1 Like

I found Dr. Loke’s syllogistic approach helpfully clarifying for someone interested in reviewing the available historical considerations (I realize that’s not everyone). I also appreciate this book generously being made available for free (and his willingness to engage with comments here)!

For anyone who has seen or read debates on this topic, it can seem like there are endless hypotheticals to consider. To help wrap your mind around this, I recommend starting by looking at section 1.4 “The approach of this book” and I think you will see more immediately how this can helpfully organize one’s thinking about it (whichever side of the debate one is on).

For Dr. Loke, I was wondering whether he felt eschatology was a fundamental component to the definition of resurrection (or understanding the resurrection as a concept). I have not finished the book yet, so I may have not come to that section.


I’ve been trying to avoid unlurking on this board, given previous experiences, but could not stand silent for the abuse of the field of Statistics occuring in this thread.

[Loke] also advances arguments for the small probability of each of these hypotheses.
What is the margin of error of these probability estimates? Given the length of times, the thinness of source data, the tangential nature of much of the evidence, and the inevitable caveats-and-disagreements-between-experts on the details of what they mean, I would expect them to be very high for the high-level 'analysis' contained in this book.

To put it another way, how certain can we be that the “small probabilities” are definitely as small as Loke thinks they are, given how murky and small the lens is that we’re looking through (figuratively speaking)?

Having, on occasion, had to calculate, and document, calculations based upon tenuous and ropy data, I can tell you that the margins of error overwhelm the value (or even the order of magnitude) of the calculated measure very quickly.

Assuming Loke’s arguments work, we are left with the large probability that a non-naturalistic (or supernatural) hypothesis must be true.
I would argue that this is false. Either we assume methodological naturalism, and the probability of the supernatural is zero by definition (so that the natural probabilities must be either underestimated, or non-exhaustive), or we reject methodological naturalism. In which case the probabilities of all outcomes (natural or supernatural) are incalculable, as they rely on the whim of whichever supernatural entity (or entities) control the outcome (whether that entity is Yahweh, Ra, Odin, or whoever, and no you don't get to cherry pick which).

We can estimate, given methodological naturalism, and a bunch of date, what the probability will be of rain tomorrow. Given the existence of the Goddess Demeter (for example), we can’t (it all comes down to whether she’s pleased with us or not).

I would further point out that, even if the probabilities of all the natural causes summed to 1, there is no requirement that the estimates of these probabilities must also needs sum to 1 (due to estimation errors). Therefore the claim that the difference between those estimates and 1 must be the the probability of a supernatural event, is demonstrably fallacious.

@Faizal_Ali argues that the (supernatural) Resurrection hypothesis is not the only possible supernatural hypothesis, but there are potentially a large number of these, each with their own probabilities.
As above, I would argue that the probability of all "supernatural hypothes[es]" are equally zero (assuming methodological naturalism), or equally incalculable (relaxing this assumption).

It seems to me that you haven’t read the book.

First, as far as I can see, Loke doesn’t give rigorous numerical estimates for the probabilities of each naturalistic hypotheses, only that they are “negligible”. I don’t know if Loke’s arguments against each of the naturalistic hypotheses are more convincing than other works that have already been published. You have to read it for yourself.

Personally, I am unconvinced that there is a rigorous way of assigning numerical probabilities to historical events (contra to what people like Vincent Torley, Richard Carrier, and the McGrews have done in their works). I have in fact argued in previous threads that the margins of error are often too large to say anything meaningful. At most you can confidently assign negligible probabilities to hypotheses that are clearly ridiculous. Applying this to Loke’s book, my takeaway is mainly that his logical framework helps to classify the naturalistic hypotheses systematically, even if in practice the probabilities attached to each can be very subjective.

Second, your point about supernatural hypotheses being incalculable doesn’t seem to contribute anything new to the discussion in this thread, because it does not engage with Loke’s arguments in section 8.3 nor even my comments about using other arguments for the truth of Christianity versus other religions.


Yes, and this is a big problem, particularly when his case for some of these probabilistic evaluations seem to be based entirely on his mere insistence.

For example, Loke argues in his book, as I point out in my post here:

First of all we are just fed the assertion that it would “not be unlikely”. To support a probability estimate, we are supplied evidence for a possibility. But the actual historical fact is proper burials of crucifixion victims was unlikely. It was rare. Crucified criminals were either just left to decay, or thrown in anonymous mass-graves.

To argue against legendary development, we get this nugget:

“On the other hand, it is implausible that the Gospels’ authors would invent a figure who was supposed to be a member of a well-identified group of their opponents (the Sanhedrin) and who could therefore be easily falsified by their opponents and thus discredit their own writings.”

Here it is merely asserted that something is “implausible”. That’s not even speaking probabilistically, Loke here seems to opted for talking about plausibility, as in whether something is believable or can be considered credible, as opposed to it’s likelihood. But the fact is legendary development, including wholesale inventions do in fact happen routinely. And they can inspire mass social movements and cults. Many Christians would of course agree that fatuous religious like Mormonism and Scientology are based on such obviously incredible claims. UFO cults have grown up around the idea that extraterrestrial aliens crash-landed their flying saucers in Roswell, New Mexico, in the 1940’s. Not to mention some of the ludicrous apologetics that have grown up around the idea that the american civil was was just about “State’s rights”, not about slavery. And these are legendary developments happening in an age of mass literacy, with mass-media, including news papers, television, and radio.

All it takes for an easily-discredited idea to spread is that it originates in your “in-group”, and for people in the “out-group” to criticize it. Rather than go and investigate for themselves, the basic psychological bias is to start defending it as somehow being intrinsic to your tribe, and that people who doubt it are somehow bad.


Of course we haven’t actually discovered that at all, since there are many naturalistic hypotheses that explain all of the relevant facts. For example, the hypothesis suggested in the Paulogia video I linked earlier. It really does in fact explain the relevant facts with observed phenomena.

1 Like

I did not state that Loke gives “rigorous numerical estimates”. However, his “negligible” inherently contains a margin of error, which expands each time he makes an assumption, has to take one expert’s opinion over another, has to infer from data not quite on point, and likewise whenever he relies on an author making similar judgements (and recursively back through their cited sources). The question is, is the Confidence Interval around negligible that of minuscule-to-tiny, or minuscule-to-small (with the potential for several ‘smalls’ to add to 1)? He can assert that it is negligible, but he cannot demonstrate it with any certainty.

Second, your point about supernatural hypotheses being incalculable doesn’t seem to contribute anything new to the discussion in this thread, because it does not engage with Loke’s arguments in section 8.3...
I encountered Baysian techniques early on in my university career. They are occasionally useful, but very subject to 'Garbage In Garbage Out' in their assumptions. I would particularly not agree that "the probabilities of the naturalistic alternatives" demonstrably render miracles not-"improbable" (see above).

Additionally, just because somebody has created a formula, is no guarantee that the the numbers they express are calculable (even to the point of approximate estimation). Neither Ehrman nor Craig are Philosophers of Statistics nor of Science, so I cannot rely upon their expertise for it to be meaningful.
The other arguments in the section seem less to prove that the probability of a miracle is high, so much as that the calculation problem is intractable (which was my point).


This is similar to Plantinga’s Argument from Dwindling Improbabilities, which is discussed in section 8.3.

And, as I keep trying to explain, my problem with your entire approach is the arbitrary and self-serving manner in which you define something as “supernatural”. What you are in effect saying is “The explanation that I want to prevail is ‘supernatural’, so I don’t need to provide any evidence that it could even possibly happen. But everyone else is obliged to demonstrate the likelihood of their explanation thru the rigorous application of history, psychology, science and other such disciplines. Oh, look, they failed and I win. Who’d have predicted it?”


True enough, it’s sort of similar, in that we are dealing with compounding numbers, but the particular problem Tim brings up isn’t actually discussed there. Compounding probabilities, meaning the probability invariably reduces as the number of factors increases, is not equivalent to compounding uncertainties.

His discussion of the problem in chapter 8.3 seems to hinge entirely on the point that the probability of the resurrection rests among other things on the probability that God exists, which he says the resurrection is evidence for. But I’ve dealt with that here.

1 Like

No, I do not have to concede anything of the sort. I am not even aware of any ongoing discipline or methodology by which the likelihood of “supernatural” factors being involved in some event is determined. Could you refer me to some of the scholarly publications or textbooks in which this methodology is described and validated?

1 Like

Well, Loke’s book was published by a serious academic publisher (Routledge)…
Seriously though, this is not a substantial reply. If you object to the methodology, give a substantial argument.

I agree it’s not exactly the same. But the similarities are enough such that likely a cogent reply to Tim’s argument would involve appropriating and modifying the gist of Loke’s (and McGrew’s) reply to Plantinga’s argument. And I just don’t have the time to reproduce that on the forum when it’s in a free book.

Secondly, I think I actually agree with a lot of what Tim says. Namely, there is a lot of uncertainty and disagreement in ancient history (and non-empirical science disciplines in general) apart from basic claims like that Jesus existed and died on the cross. So I don’t think a historical case for the Resurrection (or against it) is ever completely certain. But this is more of a meta-assessment of the resolving power of the tools, methodology, and evidence in the discipline of history. A historian arguing as a historian has a right to argue strongly that certain historical theses are very likely to be true, even if some other historians disagree with him.

You accuse Loke of arguing circularly there (i.e. God is more likely to exist because of the Resurrection, and the Resurrection is likely because God exists). First, you’re misreading Loke. He clearly has other arguments for the likelihood of God’s existence, which he has published in other books. In fact he mentions such arguments explicitly (see the quote I cited above). Secondly, circularity is different from updating of Bayesian priors based on new evidence.

Loke does not explain how he has determined that the resurrection was a supernatural event. His reasoning seems to be that 1) Resurrections do not happen as natural events. 2) Therefore, if the resurrection happened, it was a supernatural event.

But we could also argue 1) Mass hallucinations do not happen as natural events 2) Therefore, if a mass hallucination happened, it was a supernatural event.

I wish the fact this book was published by a legit publisher was sufficient reason to trust in its scholarly rigour. Unfortunately, that is not the case.

1 Like

@Faizal_Ali, I honestly don’t think you’re saying anything new here in response to the summary I made above (that Loke endorsed as accurate). You’re just reasserting your earlier point.

No, we can only make statements of relatively likelihoods. Something appears to be more or less probable. Ruling stuff out with non-zero probabilities is irrational. You can only have higher or lower degrees of confidence given the evidence.

I believe I explained that in the case of this thought experiment, we had documents where one or more people claim to have seen this happen. That’s kind of the point here. The conflict between our own experiences, as in our background knowledge about how the world is, and some claims made in some written accounts. Because of this background knowledge(which if God exists, includes evidence of what God usually refrains from doing, such as resurrections and changing the sky) we find the claims in the documents less credible.

We ALSO do NOT have the resurrected Jesus. We just have some claims in some documents, aka “for the Bible tells me so”. And no, they do not at all indicate that they truly saw something. It is highly doubtful there were ever more than 1 or two people who had such experiences. As the Paulogia video I linked earlier argues, two people could have had experiences of seeing Jesus alive for their own reasons, and they do not have to be identical at all. All they need to claim is to have seen him after his death, and legendary developments can do the rest.

I ignored nothing. I believe I am explaining that the evidence isn’t good enough to turn the low prior into a high posterior. The whole point is you have a prior given some background knowledge. Then you include the evidence to be considered, and if the evidence is good enough, it can render a hypothesis with a very low prior into one with a high posterior. But you have not done this work. Rather, you’ve attempted to skirt around doing this work by just handwaving in the direction of “when free agents are involved all bets are off and anything can happen”, and by a rather poor treatment of the disjunction of all naturalistic variants and combination hypotheses together, attacking a particular versions of it (one with some legendary development and numerous identical halluscinations, which nobody here actually advocates), and by completely ignoring all other supernatural hypothesis, apparently in the belief that the “God did the resurrection which actually happened” is the only alternative to all possible naturalistic ones.

But in point of fact, they haven’t. As I’ve shown examples of in this thread.

This is where I repeat my still-standing point about the nature of the factors that affect why God chooses to do something is irrelevant.

This is where I repeat my point that if the prior is indeterminate, it undermines your own argument. Question marks in your equations give you question marks in your results. Indeterminate probabilities means unbelievable conclusions. You can’t just make something up and decide to believe it as credible. Or well you can, but you’re believing it on faith. Because you want to.

This is where I repeat my point that all we are offered here is your opinion of what might be religiously special, unique, or religiously significant to God.

See above.

Yes I believe I’ve responded to all the points you’ve made at length. You’re just going around in a circle trying to make your different points support each other, but I’ve explained why all of them are bad.

Didn’t I already explain that when I said it?

I’ve already responded to this at length. As I explained your treatments of these different naturalistic hypotheses is poor, and you’re dismissing them on a combination of vague and subjective criteria and by only attacking particular examples and insinuating all combination hypotheses would fail on the same criteria as those you already discuss. But they would not, as various combination hypotheses have crucial differences from the particular examples you criticize in your book(such as the one shown in the Paulogia video linked earlier). Yes, I read your chapter on combination hypotheses, this one isn’t encapsulated by that discussion.

An example of a vague criterion is where you dismiss the idea that Jesus was left to decay without a burial, or being buried in an anonymous mass-grave, because you say it’s not unlikely Jesus would have been allowed to get a proper burial because this is known to have happened. But possibility doesn’t get you probability. In other words, a crucial aspect of a particular variant of naturalistic hypotheses have here been summarily dismissed by appeal to a mere possibility.

In you attempt to dismiss combination hypotheses that involve halluscinations, you bring out a particular version of it involving mass-halluscinations, which simply doesn’t apply to all hypotheses that involve halluscinations.

With respect to legendary development, you just subjectively declare aspects of it implausible. Yet there is overwhelming historical evidence, up to and including the present day, of legendary development of easily falsifiable facts.

Third point here is that you’re treating the resurrection as the only supernaturalistic hypothesis, as in you think the remainder left over when subtracting the probability of the disjunction of all naturalistic hypotheses, must be the probability of the Christian God resurrecting Jesus. But of course this is obviously false, as many, many other “supernatural” possibilities can be conceived of, and which you have done no work to show are less likely than the resurrection as you envision it. But if the probability of the resurrection is indeterminate, which it must be since you have no idea what the probability is that God would choose to resurrect some particular person, then you have no way of showing that the resurrection is more likely than many other supernaturalistic hypotheses. Hence you can’t subtract it’s probability from 1 by your (poor) attempt to calculate the probability of all possible naturalistic hypotheses.


I respectfully disagree. Your comments, it seems to me, address Loke’s argument for the acceptance of the supernatural in general as a valid ontological category, and that the means by which we evaluate claims for natural events should not apply to supernatural events.

My argument pertains to the arbitrary nature in which events and claims are categorized as “natural” and “supernatural.” I am conceding, for the sake of argument, that the supernatural is a valid ontological category and that it cannot be evaluated by the same standards as natural phenomena.

But with those issues put to the side, we are now left with what appears to be an insuperable problem: Any event, no matter how improbable or even impossible it may seem if we presume it to be “natural”, is no longer improbable or impossible if we presume it to be supernatural. Or, actually, it still is improbable, because that’s how believers in the supernatural say it works. But it is improbable in a special sort of way, in which its probability in any particular case cannot be evaluated based on its probability as a natural event.

That’s a problem that, as far as I can see, Loke does not even aware of, never mind address. Correct me if I am wrong.

So his approach is to arbitrarily decide six of the seven options are “natural” and therefore can be evaluated as improbable, which leaves the seventh as almost certainly true by default. I believe I have already explained why this is self-serving if not outright dishonest. But it is not even true on its own terms. If supernatural events remain improbable, then even the seventh option (resurrection) remains improbable. If we knew for a fact that God does resurrect people, but only once in every 100 trillion deaths, then option 7 remains no more likely than the other six and cannot be asserted as the winner by default. IOW, Loke and his fellow apologists are again just playing a self-serving game when they excuse themselves from having to quantify the probability of their “supernatural” explanations.

1 Like

Oh, I’m sure he does. Just as I am certain that people who believe the pyramids were built by aliens likely have other lines evidence for alien visitations to earth. So do we really have to deal with all those other arguments before we can reject the claim about the pyramids? Or would it be more scholarly to simply point out that the position being defended depends on a number of other propositions which are, themselves, far from a matter of consensus and that this is, therefore a further weakness of the argument being made?

1 Like