The Probability of the Resurrection

Richard Swinburne

[Please also see Kai-man Kwan’s response from pp. 5-6 below.]
1)      Kinds of evidence for a historical hypothesis:.
a)      posterior historical evidence: the testimony of witnesses about and the physical data caused by what happened at the time and place in question.  In so far as the hypothesis is a simple one, and the posterior historical evidence is such as you would expect to find if the hypothesis at stake is true but not otherwise, that is evidence for the hypothesis.
b)      the crucial importance of simplicity: always an infinite number of possible theories to explain the evidence.  The theory that is most likely to be true is the simplest one.
c)      general background evidence of how likely the hypothesis is to be true, independently of the detailed historical evidence.  Ex.: supernova explosion. 
d)     The general background evidence may indicate that it is likely to be true only under certain conditions.  Evidence showing that those conducive conditions were or were not present: the prior historical evidence.

2)      A historical hypothesis h is probable in so far as it is intrinsically a simple hypothesis and (1) the posterior historical evidence e3 is such as probably would occur if h is true, but not otherwise, (2) the general background evidence k makes it probable that h is to be expected under certain conditions and not otherwise, (3) the prior historical evidence (e1 and e2) are such as probably would occur if these conditions are satisfied, but not otherwise. The stronger are (2) and (3), the less we need by way of (1).
3)      When we are dealing with a hypothesis H which would be not too improbable on one world-view T but would be immensely improbable on a rival world-view, to the extent to which the general background evidence supports most strongly the world-view T which makes H not too improbable, we need less by way of detailed historical evidence in order for the claim that H is true, to be probable overall.
4)      The Resurrection: if there is no God, it is a clear violation of laws of nature and so impossible. But if there is a God, he has the power to set them aside.  Hence, if Jesus rose from the dead, God raised him up.  But he will only do so in so far as he has reason to do so.
5)      So to determine whether Jesus rose from the dead, we need to investigate:
a)      the posterior historical evidence (what St. Paul wrote, that the original text of Mark’s Gospel ended at 16.8, or what was written in the original text of Josephus’s Antiquities): the kind of evidence to be expected if Jesus rose, but not otherwise.
b)      whether general background evidence supports the world-view that there is a God of a kind able and likely to intervene in human history in this kind of way in this kind of situation.
c)      the prior historical evidence - whether the nature and circumstances of the life of Jesus were such that if there is a God, he would be likely to raise this person from the dead.  In so far as our general background and prior historical evidence supports the view that there is a God who would be likely to raise Jesus from the dead, we shall need a lot less by way of detailed historical evidence for the Resurrection.
6)      But suppose the natural theologian is right that there is substantial positive evidence for the existence of such a God.  In that case clearly he could, if he so chose, raise Jesus from the dead.  Hence, to the extent to which, in virtue of his goodness, he has reason to do so, it is probable that he will. 
7)      In virtue of God’s goodness, he had reason to become incarnate and live a certain sort of life, and that if he does, God has reason to raise him from the dead. 
a)      to make atonement for human sins.  Various theories like vicarious punishment, compensation/satisfaction theory, and the sacrifice view.  God has no right to send anyone else to do such a formidable task (whatever its exact nature) on his behalf.  A good God might choose to become incarnate in such a way that he would be killed for leading a holy life.  Then the Resurrection would constitute God’s demonstration of his acceptance of that life as proper compensation, sacrifice or whatever, his signature on that life.
b)      God made humans subject to pain and suffering to serve greater goods. We humans sometimes rightly subject our own children to suffering for the sake of some greater good.  Under these circumstances we judge it a good thing to manifest solidarity with our children by putting ourselves in somewhat the same situation.  A perfectly good God would judge it a good thing to share our suffering by becoming incarnate.  In that case his Resurrection would show us that God has identified with our suffering.
c)      we need better information about how to lead good lives. Moreover, moral information needs to be filled out by moral example - we need to be shown what a perfect life consists in.  It would be good for us to have encouraging information, and extra help in leading the moral life - a community of encouragement, for example.  Again, God raising someone killed for certain teaching and living a certain life constitutes his signature on that teaching. 
8)      So, if God did become incarnate in a prophet, he would need to live a certain sort of life: a good life in difficult circumstances, perhaps ending in a judicial execution; showing us that he believes himself to be God; giving us deep moral teaching; founding a Church to make all this available to other generations and cultures.  So we have prior reason for expecting a resurrection of a human if he had led a life of the above kind.
9)      New Testament scholarship: the evidence is such as we would expect if Jesus led a good and holy life, gave us good and deep moral teaching, and founded a Church, which did teach that he was God Incarnate who atoned for our sins. New Testament scholarship is divided about whether Jesus proclaimed that his life and death was an atonement for sin, and whether Jesus taught his divinity.  My own view for which I have argued in my book is that the prior historical evidence is such as is to be expected with modest probability if Jesus taught both his atonement and his divinity.
10)  The stronger is the prior historical evidence that the life and teaching of Jesus was of a certain kind, the more reason we have for expecting a God to raise him from the dead. Only in the light of the general background evidence and the prior historical evidence can we approach the posterior historical evidence. 
11)  One further piece of evidence: Jesus led the life he did, when there is no other known serious claimant for satisfying either the prior or posterior requirements for being an incarnate God, in any way as well as Jesus.   By the prior requirements I mean living a good and holy life, giving us good deep moral teaching, showing us that he believed himself to be God Incarnate and that he was making atonement for our sins and founding a Church which taught the latter things.  By the posterior requirements I mean his life being culminated by a super-miracle, such as Resurrection.
12)  The non-existence of any other plausible candidate for satisfying either the prior or the posterior requirements shows that the coincidence of the prior and posterior evidence in one candidate is extremely unlikely - unless God brought it about. But if God did not become incarnate for the stated reasons in Jesus but became incarnate in some other prophet or plans to do so in future, it would be deceptive of him to bring about the existence of the kind of evidence of his incarnation in Jesus together with the kind of posterior historical evidence of his Resurrection. If God planned the coincidence in Jesus of the two kinds of evidence, then Jesus was God Incarnate; and it is very improbable that there would be this coincidence unless God planned it.
13)  So if there is a modest amount of evidence of natural theology that there is a God who might with modest probability be expected to become incarnate for the stated reasons and to have his life culminated by a super miracle such as the Resurrection, and there is only one plausible candidate (Jesus) for such an incarnation, you don’t need too much posterior historical evidence to make it probable that Jesus rose.  An analogy - if the background evidence gives a significant probability, say _, that John would do the crime; and so _ that he wouldn’t; and the clues are on the whole not such as you would expect if he did the crime (although there’s significant probability that they might occur), but are such that it is very improbable indeed that you would find them if he did not do the crime, then they make it probable that he committed the crime.

The Argument in Bayesian Form

Let h be the hypothesis that Jesus rose from the dead, k be the background evidence of natural theology, and e3 be the posterior historical evidence (the testimony of witnesses to what happened after the crucifixion). Let e1 be the detailed historical evidence about the life of Jesus, and e2 be the evidence that the one human about whom there is either a modest amount of e1 type evidence or a modest amount of e3 type evidence is the one human in whom these kinds of evidence are combined.
Bayes’s Theorem:  P(h½e&k) = P(e½h & k) P (h½k)
e = (e1 & e2 & e3)

Let t = theism, c = God becomes incarnate for reasons of the kind which I discuss. Suppose P (t½k) = 1/2, and P(c½t & k) = 1/2, then P(c½k) = 1/4

Let f1 = the evidence (as strong as it is with reference to Jesus) that the prior requirements for being God incarnate (living a holy life etc. etc.) are satisfied in one prophet.
f3 = the evidence (as strong as it is with respect to Jesus) that the posterior requirements (a super-miracle culminating his life) are satisfied in the same prophet.
f2 = the evidence that neither of these requirements is satisfied to any similar degree in any other prophet.
f = (f1 & f2 & f3).

Suppose P(f½ c&k) = 1/10. Then P(f½c&k) P(c½k) = 1/40.
P(~c½k) = 1 - P (c½k) = 3/4. Suppose P(f½~c & k) = 1/1000. P(f½ ~ c&k) P(~c½k)=3/4000.
By the calculus,  P(f½k) = P(f½c&k) P(c½k) + P(f½ ~ c&k) P(~c½k)
                                        = 1/40 + 3/4000 = 103/4000
Then by Bayes’s theorem, substituting f for e and c for h

P( c½f&k) = P(f½c&k) P(c½k)  =     1/40    = 100 = 0.97.
                                             P(f/k)           103/4000    103

If we add to f, the evidence that the prophet is Jesus, and that the evidence that the prior and posterior requirements are satisfied concerns Jesus, that will not make any difference to the support given by (f&k) to c.  So: P(c ½ e & k) = 0.97
Since an incarnation needs to be culminated by a super-miracle like the Resurrection and since it would be deceptive of God to bring it about, given the prior historical evidence, unless he was incarnate in Jesus:
Given (e & k), h is true if and only if c is true. So: P(h½e & k) = 0.97

Articles and Books by Richard Swinburne
1968.  "The Argument from Design."  Philosophy 43(1968): 199-212.
1971.  The Concept of Miracle.  London: Macmillan.
1977.  The Coherence of TheismOxford: Clarendon.  [revised edn, 1993]
1979.     The Existence of God. Oxford: Clarendon Press.  [revised edn, 1991]
1981.  Faith and Reason. Oxford: Clarendon Press.
1983.  "Mackie, Induction and God."  Religious Studies 19(1983): 385-391.
1989.  Responsibility and Atonement.  Oxford: Clarendon Press.
1991.  RevelationOxford: Clarendon.
1994.  The Christian God.  Oxford: Clarendon Press.
1996.  Is There a God?  Oxford University Press.
1998.     Providence and the Problem of Evil.  Oxford: Clarendon.
2003.  The Resurrection of God Incarnate.  Oxford: Clarendon.

Response to Prof. Richard Swinburne on The Probability of the Resurrection
Kwan Kai Man, Hong Kong Baptist University

1)      I think Swinburne has produced an original and ingenious argument for the Resurrection of God Incarnate which has made enormous improvement over previous versions, especially those offered by the evangelical apologists, e.g., Josh McDowell. Since I basically agree with his arguments, this response mainly constitutes of my appreciation of the merits of Swinburne’s case for the Resurrection.

2)      First, let me delineate the approach of traditional apologetics on the matter of Resurrection. Usually, the apologist will draw the attention of the readers to historical data like empty tomb and resurrection appearances on the basis of the New Testament records. Then he would argue that naturalistic accounts like the swoon theory and hallucination theory cannot explain all the data. Then he concludes that the only acceptable explanation is the truth of the Resurrection. So Christianity is vindicated.

3)      Now let me explain why I think this argument is not entirely satisfactory (while not denigrating its value).
a)      It usually adopts a rather pre-critical approach to the New Testament data. While I do not agree with the approach of the radical biblical critics, I think we need to recognize the fact that they are influential in some circles and their queries need to be addressed, e.g., the inconsistency of the accounts of the resurrection appearances. Some evangelical apologists sometimes give others the impression that they are not aware of these problems.
b)      When I was studying in the University, a fellow Christian who was doing a Master degree in Mathematics once told me his dissatisfaction about the conventional argument for the Resurrection. He felt that it was unfair to treat the hypothesis of the Resurrection on the same par with the alternative naturalistic accounts. The failure of the latter does not automatically vindicate the former, especially because it invokes a miraculous act of God. As I thought over the problem, I also became uneasy about the traditional approach. I think it is not philosophically sophisticated enough because it neglects the influence of worldview presuppositions on the assessment of evidence. In fact the naturalist may just insist that it was some inexplicable and unpredictable natural event (like a sudden outburst of energy or the action of ETs) which caused Jesus’ body to evaporate into thin air. 
c)      The traditional apologists seem to equate the Resurrection with the resuscitation of a corpse. Is it really identical to the glorious resurrection depicted in the New Testament which is pregnant with meaning? Jesus’ corpse has been resuscitated- so what? At least one Jewish scholar accepts this fact but still refuses to accept the Christian faith.

4)      On the other hand, I am also dissatisfied with the faith-alone approach to the Resurrection which eschews discussion of the historical facts. It introduces a dichotomy between the Jesus of history and the Christ of faith. I still feel that the traditional argument has some force, and it is important to Christians who want to uphold a particularist position amidst religious plurality.

5)      It seems to me Swinburne has successfully avoided the above problems:
a)      He adopts an in general conservative approach to the New Testament which is post-critical rather than pre-critical. It follows from his basic principles of credulity and testimony, and his understanding of the critical scholarship. While he rejects radical scepticism towards the New Testament, he is ready to engage with the queries raised by the more reasonable critics. For example, he has faced the alleged inconsistency of the resurrection appearance accounts, and offered a plausible reconstruction of the events which largely reconciles the apparent discrepancies (Swinburne 2003, 157). (Wenham has done a similar thing.)
b)      He clearly distinguishes different kinds of evidence for a historical hypothesis, and recognizes the importance of background evidence. However, he argues that the posterior and prior historical evidence together with a modestly successful natural theology can still render the Resurrection quite probable. He provides arguments for each of his claims, and presents the structure of the reasoning in a Bayesian form. While the critics will certainly dispute many of his claims, I think it is a great achievement to clarify the proper model of reasoning which takes into account various kinds of evidence in this complicated subject matter.
c)      By delineating the reasons for incarnation and the appropriateness of the Resurrection as a kind of divine signature, Swinburne has connected the fact of Resurrection with incarnation, atonement, God’s identification with our suffering, God’s moral revelation, and so on. It shows the coherence of the Christian worldview, and the significance of the Resurrection. The Jesus of history and the Christ of faith cannot be divorced after all.

6)      Plantinga has argued that Swinburne’s probabilistic natural theology faces the problem of the dwindling of probability. If it is only probable that God exists, and also only probable that given that God exists, He will become incarnate, and so on, then the resulting probability that the basic Christian doctrines are all true will become smaller and smaller. Swinburne has made a powerful reply to this criticism by the construction of the above argument (Swinburne 2003, 215 fn. 4). He shows that further empirical evidence can also raise the probability. In fact, he only gives a value of 1/2 to the probability of God’s existence here. However, after introducing the totality of evidence about Jesus, the probability of the resurrection of God Incarnate becomes 0.97. We may dispute the estimation of various probabilities but I think Swinburne has demonstrated the mutual support between natural theology and historical evidence in Christian apologetics. His cumulative argument for the existence of God provides background evidence for the Resurrection, and in turn the historical evidence reinforces the cumulative case.


Anderson, Norman. 1985. Jesus Christ: The Witness of History. Leicester: InterVarsity Press.
McDowell, Josh. 1979. Evidence that Demands a Verdict: Historical Evidences for the Christian Faith.  Revised edition. San Bernardino, CA: Here’s Life Publishers. 中譯:《鐵證待判》
Morrison, Frank. 1958. Who Moved the Stone? London: Faber and Faber. 中譯:《歷史性的大審判》
Swinburne, Richard. 2003. The Resurrection of God Incarnate.  Oxford: Clarendon.
Wenham, John. 1984. Easter Enigma. Exeter: Paternoster.