Miracles: The Credibility of the New Testament Accounts

by Craig S. Keener

I shall also suggest that today the testimonial evidence for miracles is overwhelming compared with what was available to Hume in his day; had he lived in our day, an argument based on the nonexperience of miracles would have proved much more difficult and much less persuasive to his contemporaries (and perhaps even to Hume himself).

Refusing to arbitrate the possibility of divine causation to avoid controversy merely takes a methodologically agnostic stance; to reject the possibility of divine causation, however, takes an atheistic or deistic one.[26] The latter (atheistic or deistic) position is not philosophically neutral.[27] Why is it so common in academia? Because academicians, like others, are often unconsciously shaped by the worldviews that we take over from our intellectual heritage, frequently uncritically.

More important, the multiple universe proposal appears less rationally parsimonious than appeal to a single source of information content.

It was reaction against this increased Christian apologetic from miracles, rather than a natural development from scientists’ empiricism, that led to a counteroffensive against miracle claims.

seventeenth centuries was not deterministic and did not exclude all divine causes; the tendencies that led to an emphasis on pure mechanism “were as much political and theological as philosophical,”[74] reflecting especially the polemic between Protestant and Catholic theologians with their respective political constituencies.[75] Empiricism itself is neutral regarding God’s existence and activity, and had been used by those who supported traditional belief in miracles as well as by those who opposed it. It is thus not the basic empirical approach per se but Hume’s a priori exclusion of God as a plausible cause that drives his argument.

For most thinkers, natural “laws” are merely systematized descriptions of normal activity, but under other conditions these patterns may prove subject to more complex principles.[159] Scientific method today normally understands laws as describing evidence rather than as prescriptive or comprehensive, since the latter approach would exclude the regularly discovered abundance of evidence that does not easily fit old paradigms.

Hume’s supporters and his detractors usually recognize that he harbored prejudice against contemporary religion[197] and nearly always recognize that he constructed his approach in antithesis to Christian apologetic.

Indeed, to do so would be to ignore alternative views held even by the majority of the founders of modern science, who most often believed in the Christian God, though in some cases the deistic one.[202] As I have noted, it was not new scientific data but shifting philosophic paradigms that altered that perception.

Despite claiming to argue inductively, Hume’s argument against miracles is deductive. He argues from nature’s uniformity against miracles, which is the point in question. He generalizes from the alleged lack of good testimony for miracles to exclude what may in fact be good testimony for miracles. If Hume intends such arguments to make miracles impossible, his case is circular. If Hume intends such arguments merely to predict the weight of finite probabilities rather than to render miracles impossible, his case falters against considerable evidence for miracles. If Hume argues that no evidence in principle can be sufficient to compel belief in miracles, his claim might succeed to the extent that the evaluator of the evidence held tenaciously to antimiraculous presuppositions, but is not logically necessary if the evaluator is genuinely open-minded on the question.[209] None of the various interpretations of Hume offer a successful argument against the possibility of miracles or even of finding his argument more persuasive than not.

According to a common reading of Hume (which I think most probable), he rejects in practice the possibility of any witnesses reliable enough to challenge the unlikelihood of miracles. He circularly bases this denial on the assumed uniformity of human experience against such miracles,[213] a uniformity that would deconstruct if there were any adequately clear instances of such miracles. Claiming uniform experience against miracles is not really an argument, scholars often note, because it “begs the question at issue, which is whether anyone has experienced a miracle.”[214] Or as one critic puts it, “Hume used the unproved conclusion (that miracles are not possible) and made it a datum of his argument (miracles do not happen).”[215]

Using this standard, and a priori suspicion of any antecedently improbable information, would undermine ordinary communication.[233]

This observation seems damaging to Hume’s argument; he advances the argument in terms of “general principles about evidence, reasonable credibility, and the like,” yet we clearly do not employ his approach outside of religion.[245]

rejected the sufficiency of any evidence. The evidence of testimony must be given ways to surmount prior improbabilities; otherwise “there is no way to underwrite the sorts of inferences made in everyday life and science,” such as a newspaper report of a winning lottery ticket (an illustration developed at greater length below).[267]

While claiming logic, Hume’s approach to testimony contravenes the form of logic normally applied to the value of witnesses.

Science is meant to address nature’s regularities, not anomalies, which scientific theories usually seek to accommodate within some higher regularities; history must deal with particulars, including the idiosyncrasies of human personalities whose behavior is not always predictable. If miracles are divine actions and not pure anomalies, they must be investigated on a case-by-case basis or in tandem with analogous cases (other miracles),[270] rather than weighed against the ordinary patterns of nature. Historiographic concern for testimony is thus far more relevant for investigating miracles than statements about norms, to which miracles are by most definitions not bound. Perhaps the sphere that weighs testimony most deliberately is that of the courts. Because of the stakes involved, legal approaches to testimony tend to reflect rigorous discipline, and are applicable to the matters at hand.

work of mathematician Charles Babbage (1792–1871), originator of the modern computer, regarding witnesses: where a significant number of witnesses agree without prior collusion, an event’s probability increases sufficiently to overturn any biases against it.

Already in Hume’s day, it was clear to some that he failed to take into adequate account that the confluence of multiple, independent, and reliable witnesses increases the probability of testimony’s accuracy.[275] In the case of multiple, independent witnesses, one no longer requires extraordinary proof to secure a probability against prior improbabilities.[276] If Hume’s challenge against individual testimony was in some respects questionable, he did not even address the force of multiple testimony, which seriously changes the probability calculus.

Indeed, Bayes’s essay was first published by Richard Price, who employed it against Hume’s own doubts concerning induction.[280] Price, largely forgotten today, was in their day nearly as famous as Hume in Britain, and more so in the American colonies.[281] Hume acknowledged that Price’s challenge was “plausible and ingenious” and might require him to rethink his position, but he never revised his essay in light of it.[282]

They also illustrate, with this approach’s supporters, that the probability of reliability is very high in multiple, independent testimony from persons normally deemed reliable.

hardly inductive. Scholars identify various problems with his approach; I will address the first of these at some length before turning to others. First, as I shall show later in the book, there is no such uniform experience against miracles; whether or not one trusts the claims of tens of millions of people that they have experienced miracles, so many claims, at least some of them substantive, offer a rather vigorous challenge to simply assuming a “uniform experience” against miracles.[296] The uniformity argument essentially rejects all miracle reports, which are inconvenient to the uniformity premise; yet such reports are very widely claimed, an observation that I shall document at greater length in subsequent chapters. Far from the spirit we normally associate with intellectual inquiry, a priori rejection of all these claims declares victory by fiat rather than by fair exploration. An argument that a miracle is never probable enough to allow any evidence to be cited in support of it must circularly presuppose that miracles have never been known to happen. Should testimony that would otherwise be accepted as reliable be rejected for an event because it is extremely unusual by the

standards of one’s or one’s control group’s experience?

Again, one cannot simply rule out miracles based on all of human experience, contrary to Hume’s suggestion, because human experience claims many examples of miracles, the point Hume wishes to challenge.

Granted, reports of unusual events must be more carefully examined than reports of ordinary ones,[299] but it is methodologically fallacious to correlate the unusualness of an event with “the unreliability of a report” concerning it.[300] Should one appeal to the general regularity of current events to judge special events in the past, while ignoring special events (at least what are regarded as anomalies) that are exceptions to that observed regularity today?[301] To argue against miracles on the supposition that they never happen, no matter

who claims them, is circular. Yet to argue against miracles based on their mere rareness would, as one critic observes, exclude rare events that are attested as having happened. This critic offers the example of someone “being dealt a perfect bridge hand,” against which he calculates the odds as “1,635,013,559,600 to 1,” noting that someone in his circle was dealt this hand.[302] Likewise, Life reported an extremely improbable event, this one conceivably in a religious context (for those who regard such contexts as particularly problematic): on March 1, 1950, all fifteen members of a church choir arrived late for choir rehearsal scheduled at 7:15 p.m., all for distinct reasons (e.g., one’s car failed to start). That church building exploded at 7:25 p.m., and their lateness prevented any liv...

argument fails,[308] retells this argument in modern terms. If the New York Times announced that a particular Mr. Smith won the lottery, the odds that this individual won the lottery (say, one in a million) are less than the odds that the Times made a mistake (say, one in ten thousand). On Hume’s apparent reasoning, the probability that the Times is correct would thus be only one in a thousand.

Thus the question of kinds of events should be framed thus: “How frequently are reports of similar miracles, given in similar situations, true?”[318] Clearly many miracle claims are not true, but many other claims would be deemed true by normal means of inquiry if the possibility is not a priori excluded (I defer this discussion for chs. 13–15). I will survey some such reports later in the book (chs. 8–12); reports in settings most similar to the Gospels and Acts appear fairly frequently and often stem from those that we would otherwise consider reliable eyewitnesses.[319] Moreover, whereas mere anomalies are unexpected, it is precisely theistic premises that allow us to anticipate the possibility of some miracles in a theistic context.

what the evidence is for their occurrence and only afterward a question of divine causation. While the collision of a comet and a planet are rare enough to appear improbable for a given time frame, one cannot rule out that they occurred in the past or might yet occur. Based on knowledge of planets and comets, we might even predict a very likely occurrence for a given finite time frame rather than denying its likelihood based on the usual failure of planets and comets to collide.[321] Similarly, the probability that a reliable witness tells the truth about an event that appears anomalous in view of current knowledge can be supported if other anomalies are observed; the additional question of divine causation then strengthens the plausibility of these observations by providing a coherent intellectual framework in which to explain them.

miracles, and multiple well-attested miracle claims in that context make them probable. As philosopher Richard Purtill emphasizes, unusual events might even be expected in the right contexts.[322] We do not expect athletic world records to be broken regularly, but we might expect them to occur at times at the Olympics.[323] Purtill then argues that while Jesus’s resurrection makes no sense as a random event, it fits the context of his life and teaching, which it climaxes.

miracles makes them probable, the general a priori improbability previously assigned to miracles decreases accordingly. Again, were Hume to object that rare events were merely improbable but miracles are impossible, he would be back to restating his (ultimately theological) presuppositions without offering evidence.[326] If it is the possibility of a miracle-working God rather than the rareness of miracles to which he really objects, he is offering the wrong argument. That Hume would accept extranormal phenomena not involving religion suggests a serious element of bias;[327] as one scholar notes, “Hume’s attempt to distinguish the miraculous from the merely extraordinary is scarcely justified on his own principles.”[328] His argument formally depends on “what always happens” (the point under dispute), not “what necessarily happens,” that is, not a priori limitations of what is possible.[329] Therefore his conclusion does not follow.

improbability of events in a particular class of event prejudges “the probability of the truth of an actual report of the event.”[330] As I have been noting, Hume implies that he is arguing inductively. He actually, however, argues deductively based on a conclusion that rests on an inadequate range of data, partly because it has a priori excluded disagreeable evidence. Rather than allowing genuine induction based on evidence, Hume produced a deductive approach that a priori virtually excluded the evidence for miracles. He cites experience against experience—typical experience against rare experience, though both are attested by witnesses.[331] As is frequently noted today (including

It is impossible to prove a negative by induction when one has observed a limited range of data, and it is precarious to infer an inflexibly negative rule by induction when abundant eyewitness claims exist that one merely refuses to admit as evidence.

Generalization should accommodate all the evidence; thus, this approach to natural law unfairly ruled out exceptional cases, which induction ought to have taken into account.

By contrast, appealing to evidence that Hume would have excluded, many scholars have countered that an inductive examination of evidence from most cultures in history includes abundant testimony to supernatural phenomena.

Scientific induction is useful; it is Hume’s epistemological reductionism that is problematic.

the view that there is in Hume’s essay, or in what can be reconstructed from it, any argument or reply or objection that is even superficially good, much less, powerful or devastating, is simply a philosophical myth. The mostly willing hearers who have been swayed by Hume on this matter have been held captive by nothing other than Hume’s great eloquence.[377]

The title of John Earman’s very different Oxford University Press monograph seems to concur: Hume’s Abject Failure: The Argument Against Miracles.[379] Yet despite its logical weaknesses, Hume’s essay provided the primary argument behind most prejudice against miracles today.

Given abundant and sometimes well-attested claims of miracles today, Hume’s argument on this point should appear even less persuasive in a twenty-first-century multicultural context than it appeared in his own day (see chs. 7–12). Hume’s argument proceeds mostly by definition rather than induction, and is logically unworkable even on his own philosophic premises. It succeeded historically largely on the weight of his intellectual prestige and its appeal to some popular trends in his intellectual milieu.

Among many academics. . . ., the belief that miracles are impossible in principle seems natural, normal, obvious, undeniable—rather like religious beliefs in close-knit, traditional societies. The conviction has an aura of neutrality and objectivity, as if dogmatic metaphysical naturalism were somehow not as much a personal conviction as is dogmatic religion, as if rejection of the very possibility of transcendent reality were the default

position, one obvious to any intelligent person. —Brad Gregory, in the 2006 theme issue of History and Theory[1]

Both John Wesley and George Whitefield secured the release of individuals from such asylums whom they believed to be merely religiously devoted; ninety patients admitted in Bedlam between 1772 and 1795 were said to be “insane because of religion and Methodism.”

Modern thought’s simultaneous a priori “exclusion of the supernatural from the domain of science” taints supernatural approaches with the suspicion of “falsity and irrationality.” Yet such identifications unfairly load the deck epistemologically.[7]

Yet it was metaphysical presuppositions, not empirical evidence, that drove scientism in a naturalistic direction,

That is, if a scientist or anyone else dismisses the possibility of supernatural reality, she is offering a metaphysical view and pontificating on metaphysics no less than someone who simply assumes the contrary. Moreover, as I observed in chapter 5, these assumptions depended on a scientific worldview quite different from ours today, and were offered not by scientists themselves but by philosophers (especially deists and Hume) with a religious agenda.

We do not speak of human actions as violating nature simply because the laws of physics do not predict them.

Nevertheless, this principle of analogy makes first-century miracle reports more rather than less plausible today, as many scholars now argue.

If neutrality is less in vogue for writers in the humanities today, tolerance is not; if advocacy against theism is acceptable, then advocacy for theism should be acceptable as well.

Such explanations were originally ruled out of court in some historiography, as in science, for philosophic reasons, yet the original reasoning for that exclusion no longer dominates the philosophy of religion on which such assumptions in other disciplines were based.

method of science (demanding replicability) does not work in history; the best that one could do is to exclude a class of events that lack analogy, but one may exclude miracles on these grounds only on the circular assumption that they have never happened. Historical evidence for particular miracle claims should thus be investigated rather than being dismissed as a priori impossible.

Against those who contend that one might not posit a cause with which one does not have direct experience, such a recourse is often necessary in science. For example, physicists often use particles and other factors to explain phenomena before such factors are more directly verified.

In sociological terms, antisupernaturalism supports itself circularly: our world construction shapes how we understand our experience, then we (like Hume) “make our experience normative for all experience,” and we allow no phenomena to challenge our construct.[113] Miracles may be viewed as unique events, but history, unlike science, is full of events that are unique in some respects.[114] Were historians to begin excluding unique or extremely unusual events, they would have to rule out much of history. Seeking to illustrate this point, in 1819 one scholar, parodying Hume, demonstrated that one could explain away a large measure of Napoleon’s life by this method.[115] Indeed, the first-century rise of a movement exalting Jesus as Lord within Judaism lacks prior analogy; yet no one denies that this movement began.[116]