Was Christianity Too Improbable to be False? (2006)

Richard Carrier


Buy Not the Impossible Faith! Not the Impossible Faith: Why Christianity Didn't Need a Miracle to Succeed

Now available as a book, fully updated and reorganized. This is the definitive edition of “Was Christianity Too Improbable to Be False?” Even better than online, improved and revised throughout. Available at Amazon



A common argument for the truth of the Christian religion is that its origins were too improbable for it to be false. This argument has appeared in many forms over the years, but most of the usual ideas are combined into a single popular effort by James Holding.[1] The following article critiques that effort, by comparing Holding's arguments and claims there with the actual facts of ancient history, and identifying fallacies in his reasoning. Holding offers seventeen factors "where Christianity 'did the wrong thing' in order to be a successful religion" and concludes from this that "the only way Christianity" could "succeed" under those seventeen hostile conditions is "because it was a truly revealed faith," in particular "because it had the irrefutable witness of the resurrection." Besides those seventeen factors, Holding offers one additional critical assumption about "luck," making eighteen points altogether. Each of those points will be addressed in a separate chapter, in order, with his eighteenth underlying assumption counted last, followed by an evolving chapter responding to critics of the present work. In addition, I have added some preliminary remarks about method below (after the table of contents).


Table of Contents

01. Who Would Buy One Crucified?
02. Who Would Follow a Man from Galilee?
03. Was Resurrection Deemed Impossible?
04. Was the New Always Bad?
05. Who Would Join a Moral Order?
06. Who Would Join an Intolerant Cult?
07. Was Christianity Highly Vulnerable to Inspection and Disproof?
08. Who Would Want to be Persecuted?
09. Was the Idea of an Incarnate God Really Repugnant?
10. Would Groupthinkers Never Switch Groups?
11. Did No One Trust Women?
12. Did No One Respect the Opinions of Uneducated Laymen?
13. Would the Facts Be Checked?
14. Was the Apparent Ignorance of Jesus a Problem?
15. Who Would Follow an Executed Criminal?
16. Were Christian Teachings Too Radical for Anyone to Buy?
17. Did the Earliest Christians Encourage Critical Inquiry?
18. How Successful Was Christianity?
19. Responses to Critics


The Problem of Uncertainty

One thing that is missing from Holding's paper is any sort of formal logical or statistical argument. Despite his rather hyperbolic language, even Holding must admit that the odds of Christianity becoming successful without being true could not be zero even on all of his own assumptions. Human behavior is not that predictable, nor are there any demonstrated historical "laws" that make any conclusion about historical cause-and-effect beyond all probability of error. Rather, Holding can only mean that the probability of Christianity becoming successful, on all of his own assumptions and premises, is so low that we have no rational ground to believe it did--except by some divine aid. In Holding's version of the argument, this fact can only become reasonably probable if we accept as true the premise that the "witness of the resurrection" was (and therefore is) "irrefutable."

I will not quibble about what exactly "irrefutable" means, since I will assume he means the "witness of the resurrection" was (and therefore is) as irrefutable as the historical fact that Christianity was successful. All observers agree with the latter statement, and we certainly should believe any statement that meets the same standard, which is Holding's aim. However, how improbable would the success of Christianity have to be before we have to believe in the resurrection of Jesus to explain that success? Holding never says. Nor does he say how improbable Christianity's success really was. Yet without comparing those two estimates, it is not really possible to confirm the success of Holding's argument objectively. Many fantastically improbable things happen all the time, simply because so many things happen. For example, "that's about as likely as getting struck by lightning" is often used as a cliché of an event so improbable it never happens, yet over four hundred people are struck by lightning every year in the United States alone. Some people have been struck multiple times.[2] Hence our intuition often fails us when estimating the improbable.

Normally, this is not a barrier to historical inquiry, since we need only ascertain the most probable cause of an event, given all we know. And usually we can say that, given what we know, the most probable cause is the one that is most probably true, and therefore worthy of belief (though maybe only a tentative belief, depending on how much more probable it is than alternatives). However, in Holding's case this requires trying to sort out three crucial questions: (i) whether the "prior probability" of a miracle from God is greater than the prior probability of any alternative natural cause that is proposed to explain the same evidence (e.g. the prior probability of my being struck by lightning is a lot lower than my prior probability of catching a cold); (ii) the probability that a genuinely risen (and hence living) Christ would actually produce all the evidence we have (including a Church preaching immoral doctrines such as the subjugation of women and the persecution of doubters); and (iii) the probability that a qualified set of natural causes would still make Christianity as successful as it was.[3] We must also rule out the influence of a deceiving supernatural power, i.e. some force, such as Satan, who could bring about the same results through supernatural influence, as some Jews might allege for the success of Christianity.

Holding does not make any effort to answer these questions even vaguely. Thus, his conclusion can only be vaguely certain at best. We will nevertheless set this aside and assume Holding's argument succeeds unless we can show that some set of natural causes that we know for a fact happen more often than miracles do (i.e. natural causes that were not unusual or rare) were reasonably likely to have produced the same result (the actual success of the Christian Church). We will also assume for the sake of argument that all non-Christian supernatural causes that could logically be to blame are less probable than the most probable natural causes, whatever they may be. In other words, we will assume that if Jesus was not raised by God, then probably Christianity's success was due to natural causes, and not (for example) Satan.[4]


Proceed to Chapter 1



[1] James Patrick Holding, "The Impossible Faith: Or, How Not to Start an Ancient Religion" (n.d.). My critique is based on how that article appeared in October of 2004 (which I have saved and now keep on file). Holding responded to my original rebuttal as posted in 2004 ("Broken Vector Sinks Again: Or, TIF Vindicated") and I revised my work in response to that.

[2] See, for example: Robert Shmerling, "The Real Dangers Of Lightning" (Aetna InteliHealth.com, 2001). I have discussed these problems of relative probability before, in "Probability of Survival vs. Miracle: Assessing the Odds," which is Part 2 of Richard Carrier, "Why I Don't Buy the Resurrection Story" (5th ed., 2004).

[3] This follows necessarily from Bayes Theorem. See: Giulio Agostini, Bayesian Reasoning in Data Analysis: A Critical Introduction (2003) and Richard Swinburne, An Introduction to Confirmation Theory (1973). It is not always the case that a hypothesis with a lower prior probability (i.e. (i)) is less likely true than another, since a sufficiently high probability for (ii) along with a sufficiently low probability for (iii) can overcome any prior probability, no matter how low with respect to that of any other hypothesis. In laymen's terms, even though miracles must be extremely rare (since even at best we see few of them, and have yet to establish even one with anywhere near the same certainty as we have for countless other causes of even very bizarre events in history), and therefore miracles must be extremely improbable, it is still possible to have enough evidence to establish a high value for (ii) and an extremely low value for (iii), enough to make "miracle" the most probable explanation among all alternatives. Nevertheless, this does require a substantial scale of evidence, a fact that also follows necessarily from Bayes Theorem, which follows a deductively valid argument.

[4] My thanks go to Johnny Skeptic who fully funded the research and writing of all nineteen chapters of this present critique. Two previous critiques of Holding's "The Impossible Faith" have been published by other authors: Brian Holtz, "The Not-So-Impossible Faith" (Secular Web, 2002) and Robert Price, "James Patrick Holding's The Impossible Faith" (RobertMPrice Online, n.d.).

Copyright ©2006 by Richard Carrier. The electronic version is copyright ©2006 by Internet Infidels, Inc. with the written permission of Richard Carrier. All rights reserved.