The Question That Probability Must Actually Answer

David Hume’s challenge to miracles concerns the relationship between ordinary experience and testimony about an extraordinary event. Human beings repeatedly observe that the dead remain dead, while human witnesses sometimes misunderstand, exaggerate, or deceive. Hume therefore argues that a report of resurrection faces a powerful presumption against its truth. The Christian response must acknowledge the force of ordinary experience without allowing that experience to decide a different question from the one under examination. Christianity does not claim that dead bodies naturally restore themselves to life as an occasional biological development. It claims that God raised Jesus from the dead, an assertion explicitly expressed in Acts 2:24 and Galatians 1:1. The relevant comparison therefore concerns divine action and the actual evidence, rather than spontaneous recovery and ordinary mortality. Any calculation that substitutes the probability of unaided biological recovery for the probability of God raising Jesus has changed the proposition before evaluating it.

Probability reasoning becomes misleading when the distinction between these propositions disappears beneath mathematical language. A statement about what nature ordinarily produces cannot, by itself, establish what the Creator has never done. Conversely, merely affirming that Jehovah possesses the power to raise the dead does not establish that He raised a particular person. The first mistake excludes divine action without adequate argument, while the second moves from ability to occurrence without adequate evidence. Christian apologetics must avoid both errors by considering background knowledge, the identity of Jesus, and the testimony bearing on His resurrection. Acts 26:8 brings the underlying issue into focus by asking why God’s raising the dead should be regarded as unbelievable. Paul’s question identifies the proposed agent and prevents the discussion from treating resurrection as an unexplained accident. It also leaves the particular historical claim open to examination rather than replacing examination with an assertion of mere possibility.

Hume’s Principle and Its Limits

Hume’s position should be represented accurately before it is criticized. His argument is stronger than the bare assertion that an unusual event cannot happen because it is unusual. He compares the established regularity supporting a law of nature with the reliability of testimony apparently opposed to that regularity. His familiar criterion requires testimony strong enough that its falsehood would be more extraordinary than the event it reports. In that formulation, the credibility of the witnesses and the improbability of the event both matter. Hume also advances objections concerning the actual character of miracle reports and the religious interests associated with them. A responsible rebuttal therefore distinguishes his general evidential principle from his judgment that the available testimony fails to satisfy it. Demonstrating a weakness in that judgment does not require denying that extraordinary claims deserve careful investigation.

The central difficulty arises when a strong presumption against an event becomes a rule that no evidence can overcome. Those are substantially different positions, and neither should be attributed to an opponent without attention to the argument actually being made. A defeasible presumption permits inquiry because sufficiently strong evidence can alter the initial judgment. An absolute exclusion settles the matter before the documents, witnesses, and circumstances receive consideration. If resurrection testimony is rejected simply because resurrection cannot occur, and resurrection cannot occur because all contrary testimony has already been rejected, the reasoning is circular. If the testimony is instead assessed and found wanting for specific reasons, those reasons must be examined on their merits. Hume’s challenge therefore cannot be answered solely by attaching the label of circularity to every skeptical judgment. The decisive question is whether the claimed uniformity of experience has been established independently of dismissing the very evidence under dispute.

What Ordinary Death Establishes

The regularity of death is part of the Christian account rather than an embarrassment to it. Genesis 3:19 describes the return of the human being to dust, and Ecclesiastes 9:5 identifies the dead as knowing nothing. Jesus’ resurrection does not depend upon an immortal human soul continuing its conscious life while the body remains permanently dead. The claim concerns God’s restoration of the person to life and His victory over death. Romans 6:9 makes the further distinction that the risen Christ does not die again and that death no longer rules over Him. Ordinary human experience thus supplies the background against which the resurrection has its exceptional significance. If the dead commonly returned to lasting life through ordinary processes, the apostolic proclamation would describe a familiar occurrence rather than an extraordinary act of God.

Nevertheless, the regularity of death must be interpreted within the limits of what observation establishes. Observations of dead persons remaining dead provide strong evidence against an ordinary biological process that routinely reverses death. They do not directly measure the frequency with which Jehovah would raise the Messiah in the circumstances of His redemptive purpose. Those circumstances are not interchangeable with every death throughout human history. An analogy clarifies the logical distinction without proving the miracle: observing thousands of locked doors remaining shut does not establish that an authorized person will never open one with a key. The observations accurately describe what happens without the additional agent’s relevant action. They fail only when they are expanded into a claim excluding that action altogether. Similarly, mortality supports a strong expectation under ordinary conditions, while the resurrection claim introduces a proposed cause whose reality and relevance require separate consideration.

A Presumption Is Not a Mathematical Law

The expression “Hume’s probability laws” must not suggest that Hume established a mathematical theorem proving miracles incredible. His argument offers a philosophical assessment of evidence, testimony, and the regularity of nature. Modern probability theory can clarify the relationships involved, but it does not automatically endorse his conclusion. A theorem states how specified quantities relate once the relevant assumptions and values have been supplied. It does not determine whether the assumptions accurately describe Jesus, His witnesses, or the world in which the events occurred. Calling a conclusion mathematically necessary therefore requires more than displaying a formula. The reasoning must establish the appropriateness of its inputs and the completeness of its alternatives. Otherwise, the formula expresses the consequences of a disputed judgment rather than independently settling the dispute.

A useful distinction separates a low initial probability from a low probability after evidence has been considered. Before learning anything about a particular incident, an investigator ordinarily expects familiar events rather than extraordinary ones. After receiving sufficiently discriminating evidence, the investigator may reasonably reverse that expectation. For example, a precisely specified sequence of twenty fair coin tosses has a probability of one in 1,048,576 before the tosses occur. A reliable recording can nevertheless provide strong grounds for believing that this particular sequence occurred. The example does not make resurrection equivalent to coin tossing, because the underlying processes and evidential questions differ. It demonstrates only that low probability in advance does not entail low credibility after observation. The evidential importance of a report depends on how strongly it distinguishes what happened from the alternatives.

Bayes’ Theorem and the Resurrection Claim

Bayes’ theorem provides a formal way to express the relationship between prior judgment and subsequent evidence. Let R represent the proposition that God raised Jesus, E represent the evidence under examination, and B represent the relevant background knowledge. In words, the posterior odds of R rather than its denial equal the prior odds multiplied by the likelihood ratio of the evidence. The prior odds express the comparison before E is incorporated, while the likelihood ratio compares how expected E would be if R were true with how expected E would be if R were false. The posterior odds express the comparison after the evidence has been incorporated. These terms distinguish the probability of a claim from the probability of encountering evidence if that claim is true. Confusing those two probabilities produces invalid reasoning even when the surrounding discussion appears sophisticated. Bayes’ theorem itself is neither Christian nor skeptical; its application depends upon the propositions, evidence, and background actually supplied.

A numerical illustration shows what the theorem does without pretending to calculate the resurrection. Suppose an investigator assigns a hypothetical claim prior odds of one to one thousand against its truth. Suppose further that the observed evidence is ten thousand times more likely if the claim is true than if it is false. Multiplying the prior odds by that likelihood ratio yields posterior odds of ten to one in favor of the claim. Those odds correspond to a probability of approximately 90.9 percent within the stipulated model. The numbers are illustrative and are not measurements of the resurrection, the apostles, or any ancient document. A different justified prior or likelihood ratio would produce a different result. The relevant point is that a substantial initial improbability can be overcome by evidence sufficiently more expected under the claim than under its denial.

The Difference Between a Tiny Prior and a Zero Prior

Assigning a probability of zero to resurrection has a consequence that must be recognized openly. Within ordinary Bayesian updating, a zero prior remains zero regardless of the strength of subsequently introduced evidence, provided the conditioning operation is defined. Such a model cannot discover that resurrection occurred because its initial structure excludes that outcome. That exclusion may express a philosophical commitment, but it cannot then be advertised as a conclusion produced by examining the resurrection evidence. A very small positive prior differs because sufficiently strong evidence can, in principle, overcome it. The distinction matters whenever a critic claims openness to evidence while defining every supernatural explanation out of consideration. An inquiry that cannot possibly reach one answer should disclose the restriction responsible for that result.

Christian reasoning must apply the same intellectual discipline to its own assumptions. Assigning certainty to the resurrection at the outset does not constitute a historical demonstration for someone who disputes it. A believer’s confidence in inspired Scripture and an argument intended to establish the credibility of Scripture occupy distinguishable roles. In the latter discussion, the apologist must explain why the testimony warrants belief rather than merely assuming that every disputed claim is already accepted. This does not require surrendering the conviction that Scripture is true. It requires presenting reasons appropriate to the question being discussed and identifying where an argument depends on previously established premises. Paul’s reasoning in 1 Corinthians 15:14–17 treats the resurrection as an event with decisive consequences, rather than as a belief insulated from the question of whether it happened.

Why the Reference Class Matters

A probability estimate depends partly upon the category of cases used to inform it. If the selected category is all human deaths, resurrection will appear maximally unlike the ordinary outcome. If the proposed event is God’s vindication of His promised Messiah, the category involves additional conditions that cannot be dismissed without argument. Selecting that narrower description does not automatically make the event probable, since the description itself requires evidential support. It does show why counting ordinary deaths supplies no ready-made numerical prior for the complete Christian claim. The investigator must ask which features of the case are relevant and whether they have independent justification. Without that work, an apparently objective frequency may measure a broader and materially different proposition.

The same caution prevents an apologetic abuse of the reference class. It would be arbitrary to define a category as “persons whom God certainly intended to raise” and then place Jesus in it without argument. Such a category would build the desired conclusion into the description. The proper approach considers evidence bearing on Jesus’ identity and mission without assuming that this evidence already proves the resurrection. John 5:36 presents His works as bearing witness concerning His mission, while John 20:30–31 explains the evidential purpose of the recorded signs. Those passages show that the New Testament locates the resurrection within a larger account of divine revelation. They do not supply a statistical table assigning numerical frequencies to divine decisions. The relevant background is therefore substantive and historical, while its conversion into precise numerical odds remains a separate and demanding task.

The Probability of Evidence Is Not the Probability of an Event

A recurring mistake treats the ordinary improbability of resurrection as though it directly measured the probability of the resurrection reports. These are different quantities. The question whether God raised Jesus differs from the question how likely the surviving testimony would be if He had done so. It also differs from the question how likely that testimony would be if He had not done so. A strong evidential argument must compare the latter two expectations while taking prior considerations into account. Merely repeating that people normally remain dead leaves that comparison unfinished. Conversely, merely observing that resurrection would explain resurrection testimony leaves the comparison unfinished from the other direction. The alternative explanations must also be assessed according to their capacity to account for the evidence.

For example, a report of an unusual event may be expected both if the event occurred and if a familiar misunderstanding occurred. In that case, the report alone may have little power to distinguish the two explanations. Additional details can change the comparison if they are substantially more expected under one explanation than the other. Luke 24:36–43 describes Jesus addressing the disciples’ fear, directing attention to His hands and feet, and eating in their presence. The probability question concerns why such specific claims appear and how credible the account is, rather than whether religious people sometimes report extraordinary experiences. The historical assessment of those details belongs to the investigation of the testimony itself. Their logical relevance is that evidence must be evaluated in its actual form. Reducing every account to the sentence “someone reported a miracle” discards potentially important information before the comparison begins.

Evidence Must Be Combined Without Being Counted Twice

Several pieces of evidence can support a claim together, but their combined force cannot be calculated responsibly by assuming independence without justification. Two witnesses repeating the same original report do not necessarily supply two independent observations. Likewise, two documents drawing upon shared information may agree because of that relationship. A correct cumulative assessment asks what additional support the second item supplies after the first item is already known. Dependence can reduce that additional support without making the second item worthless. The error lies in multiplying evidential weights as though each item arose independently when the relevant relationships remain unexamined. Christian apologetics gains strength from acknowledging those relationships because exaggerated arithmetic creates vulnerabilities that the historical evidence itself does not require.

The report of more than five hundred brothers in 1 Corinthians 15:6 provides a concrete example. Paul states that Christ appeared to this group and that most of its members remained alive when he wrote. The surviving letter is one documentary report of that appearance, rather than five hundred individually preserved written depositions. Its significance includes the scale of the claimed event, Paul’s assertion concerning living participants, and the report’s place within his argument. It does not authorize multiplying an imagined probability of individual error by itself five hundred times. Neither does its preservation in one letter justify treating the reported group as though Paul had mentioned only his own private experience. The evidence must be described accurately before its weight can be assessed. Careful probability reasoning preserves both the content of the claim and the actual form in which that claim survives.

The Denial of Resurrection Is a Composite Alternative

The proposition that Jesus was not raised includes more than one possible account of the evidence. It encompasses different combinations of error, deception, misunderstanding, and subsequent transmission. Those alternatives cannot all be assigned the probability of whichever one appears most convenient at a particular moment. In a formal model, the likelihood of the evidence under the denial of resurrection depends upon the relevant alternatives and their respective weights. A weak naturalistic hypothesis does not become strong merely because some other naturalistic hypothesis is imaginable. Equally, defeating one proposed alternative does not automatically establish the failure of every alternative. The comparison must remain consistent across the evidence under consideration.

This requirement also limits the appeal to an unspecified combination of ordinary causes. A critic may propose that one process explains one report while another process explains a different feature. Such combinations are logically possible, but their possibility does not establish their probability. The combined account must explain why those processes operated together and produced the particular evidence that survives. An apologist must not assign arbitrarily tiny probabilities to each proposed component merely to make the combination appear impossible. A critic must not treat a collection of unverified possibilities as a demonstrated historical explanation merely because every component is described as natural. Both sides must distinguish an available story from an account supported by reasons. The mathematical comparison cannot replace the historical work needed to justify the alternatives being compared.

Religious Testimony Requires Discrimination

Hume’s concern about religious enthusiasm identifies a real danger that Scripture itself recognizes. Proverbs 14:15 distinguishes gullibility from careful consideration, and 1 John 4:1 directs believers to examine claims concerning spiritual authority. The existence of religious deception therefore supplies no premise foreign to the Christian worldview. The question is whether a particular claim bears the characteristics of error or has adequate support. A general observation that religious people sometimes make false claims cannot determine that every religious claim is false. The same reasoning applies to political reports, financial statements, and ordinary personal testimony. A category containing falsehoods requires discrimination within that category rather than automatic rejection of all its members.

Luke’s account also prevents a simplistic portrayal of the disciples as persons who accepted resurrection merely because it satisfied a religious desire. Luke 24:11 states that the women’s report appeared to the apostles as nonsense and that they refused to believe it. That response shows the narrative acknowledging resistance to the claim rather than assuming that every sympathetic listener accepted it immediately. It does not establish that the disciples were incapable of error, nor does it provide a numerical measure of their reliability. It identifies an evidential feature that a psychological explanation must address. The relevant question concerns these persons, these reported encounters, and these circumstances. General statements about enthusiasm must be connected to those particulars before they can carry decisive explanatory weight.

Competing Miracle Claims Do Not Cancel by Arithmetic

Reports associated with competing religions do not invalidate one another merely by belonging to conflicting traditions. Two incompatible doctrines cannot both be true in the same respect, but that logical fact does not establish that every event reported by their adherents failed to occur. The occurrence of an unusual event and the doctrinal interpretation attached to it are distinguishable claims. Even the biblical account warns against accepting religious teaching solely because an impressive sign is alleged. Deuteronomy 13:1–4 requires faithfulness to Jehovah despite a sign associated with an invitation to serve other gods. This passage prevents the Christian argument from treating unusual power, standing alone, as sufficient proof of every accompanying assertion. Claims must be evaluated according to their evidence, content, and relationship to what God has revealed.

The resurrection argument therefore need not defend every extraordinary story ever told. It concerns a particular claim about Jesus and the evidence relevant to that claim. Rejecting an unrelated report does not establish that Jesus remained dead, just as rejecting a forged letter does not establish that all letters are forged. At the same time, a Christian cannot reasonably accept weak testimony merely because it favors a desired conclusion while rejecting equally weak testimony when it favors an opponent. Consistent standards require attention to access, reliability, transmission, and explanatory context in every case. The resulting judgments need not be identical because the evidence need not be identical. Probability reasoning becomes responsible when it preserves those differences rather than grouping all miracle reports into an undifferentiated class.

Numerical Precision Cannot Supply Missing Knowledge

Modern skeptical calculations sometimes appear decisive because they express a conclusion with several decimal places. Precision of expression, however, does not establish accuracy of judgment. An estimate that God would raise Jesus once in a trillion relevant circumstances requires some justification for the selected number. Ordinary mortality statistics do not directly provide that figure because they do not record repeated comparable decisions by God concerning the Messiah. An apologetic estimate assigning an equally impressive probability in the opposite direction faces the same difficulty. Divine purpose is not a mechanism whose frequencies can be read from a large sample of interchangeable events. Numbers can represent reasoned judgments, but the reasoning must remain visible and open to examination.

A more informative use of numerical models examines how conclusions change when disputed assumptions change. If a skeptical result depends entirely upon an extraordinarily small prior, that dependence should be identified. If an apologetic result depends upon an exceptionally large evidential likelihood ratio, the grounds for that ratio should receive equal attention. A calculation that reaches different conclusions across reasonable assumptions does not establish a result independently of those assumptions. Such an examination can clarify the location of disagreement even when the evidence does not justify a single precise percentage. The absence of an agreed numerical percentage does not imply that no rational historical judgment is available. Human beings routinely distinguish strong evidence from weak evidence without possessing a statistically measured probability for every relevant premise.

Extraordinary Evidence and the Actual Claim of Scripture

The demand for extraordinary evidence requires a clear explanation of what would count as adequate evidence. If it means evidence sufficiently strong to overcome a serious initial presumption, the demand expresses a defensible evidential concern. If it means that testimony is disqualified merely because its subject is extraordinary, the demand excludes a major route by which past events become known. The strength of testimony depends upon its particular characteristics, its relationship to other evidence, and the alternatives that explain it. No formula establishes that testimony must always lose whenever it concerns an event outside ordinary experience. Neither does the Christian position require every report of an exceptional occurrence to be accepted. The standard concerns adequacy to the actual claim rather than the automatic acceptance or rejection of an entire category.

Paul makes the stakes concrete when he writes, “And if Christ has not been raised, then our preaching is in vain, and your faith is in vain” (1 Corinthians 15:14). He proceeds to identify the apostles as false witnesses concerning God if their resurrection testimony is false. His argument refuses to preserve the Christian proclamation by reducing resurrection to a beneficial attitude or an inspiring religious interpretation. In 1 Corinthians 15:17–18, the consequences extend to forgiveness and the condition of believers who have died. The actual occurrence of resurrection therefore remains indispensable to the message being defended. A probability argument must address that occurrence and the evidence bearing upon it, rather than dismissing the claim through a substituted proposition about nature acting alone. The proper demand upon every proposed calculation is that it justify its assumptions, preserve the full question, and permit the evidence to affect the answer.

You May Also Enjoy

The Evidence for the Historical Jesus Christ: Investigating the Full Case

About the Author

CLICK LINKED IMAGE TO VISIT ONLINE STORE

CLICK TO SCROLL THROUGH OUR BOOKS