An excerpt from Miracles: The Credibility of the New Testament Accounts, by Craig S. Keener. From Pneuma Review Fall 2013.
From Part 2, “Are Miracles Possible?”
Chapter 5, “Hume and Philosophic Questions”
The Circularity of Hume’s Approach
Houston challenges at length Hume’s belief that the general 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 above), Hume’s argument against miracles is thus circular, assuming what it claims to prove,332 an observation also offered by some of Hume’s contemporaries.333 His rejection of some experience on the grounds that it differs from usual experience actually contradicts his own empiricist tradition.334 The more genuinely inductive approach of English scientists of his era was to adjust research models and methods to accommodate new evidence from diverse phenomena, an approach that rendered their position immune to the arguments raised by the deists and Hume.335 Hume prevents his own argument from being falsified by rejecting evidence that contradicts his thesis. As Robert Larmer complains, Hume’s denial that any amount of evidence favoring miracles could ever be acceptable “commits him to holding that there are logically possible empirical events which no conceivable amount of positive evidence could ever confirm, but which a finite body of negative evidence disconfirms.” Hume at one point allows the relevance of empirical evidence for deciding the question but then contradicts his normal empirical approach by ruling it out of court.336
One may illustrate this predisposition in Hume’s own argument. As I have noted, he cites some strong testimony for some miracle reports but then uses the very strength of this testimony to argue that even strong testimonies are useless in favor of miracles, since (he asserts, without argument) these particular miracles may be dismissed!337 An early twentieth-century writer complained, “It is no use investigating these events, Hume says in effect, for no matter how conclusive your arguments for their occurrence, they cannot be accepted.”338
Various writers have noted the circularity of Hume’s denial of these reports of miracles, for example, among Jansenists, which he denied on the mere basis that miracles cannot happen,339 a denial that some today regard as “obscurantist.”340 Hume viewed the evidence for Jansenist miracles (often immediate, credible, and multiple testimony) as stronger than that for Jesus’s alleged miracles,341 so that denying the former made denial of the latter much simpler. But that at least some cures did occur is difficult to deny. After the pope condemned Jansenism in 1713, the ascetic Jansenist deacon Francois de Paris lived even more austerely and died in 1727. At his interment, a widow’s paralyzed arm was said to be healed, and many subsequent visitors experienced ecstasy and healing. Cardinal Noailles’s report in 1728 acknowledged genuine healings, and reported cures on the site included “cancerous tumors, … paralysis, deafness, and blindness.” Finally the king had the cemetery closed in 1732, whereupon one graffitist opined, “By order of the king, God is forbidden to perform miracles in this place.”342 Officials secured considerable evidence that the most dramatic of these particular miracle claims were false;343 their evidence may be correct, or it may reflect the use of political power in a propaganda war. Certainly the claims of the now-marginalized Jansenists were amenable neither to mainstream Catholics nor mainstream Protestants. But Hume fails to note either these detailed challenges to their credibility (differentiating their claims from some other historical miracle claims) or the potential political motivation for the challenges (allowing that the healing claims possibly could be more reliable than their critics conceded).
Hume, like most Catholic and Protestant critics, could dismiss Jansenist reports, but some intellectuals closer to the events felt differently—that is, those who were closer to “direct experience” than Hume was. Consider the influential mathematician Blaise Pascal, who devised a calculating machine that was the forerunner of modern computers, invented the syringe and the barometer, devised the mathematical theory of probability (hence his famous wager about faith), demonstrated the possibility of vacuums, and so forth.344 Pascal’s commitments to Jansenism (and his reasons for penning the Pensees) were reinforced precisely by the healing of his niece, Marguerite Perrier, in a Jansenist setting, long before the more controversial cures noted above. A severe, long-term fistula in her eye disappeared during the touch of a consecrated relic on March 24, 1656, at the Jansenist Port-Royal monastery. From all the evidence available, the cure must have been organic and not merely psychosomatic. The repulsive odor from her wound, which had forced her separation from the other girls, and her apparent bone deterioration vanished immediately. Her case provided significant medical evidence and was verified by the diocese. The royal physicians examined Mlle Perrier, and the Queen Mother herself was persuaded by their positive verdict of a miracle.345 In the next few months, some eighty further miracle claims followed.346
The vast majority of us today would question the relic’s authenticity (a thorn from Christ’s crown),347 but the dramatic recovery is difficult to deny. Even the Jansenists’ critics acceded to the official recognition of the miracle’s genuineness, but whereas Jansenists cited it as a sign of divine approval, their detractors treated it as a warning.348 For Hume, however, writing in a period of established Protestant and Catholic polemic over whose miracles were authentic, all miracle claims were religiously partisan and thus unreliable.349 Though these miracles were recent, public, and attested by many witnesses—that is, they fulfilled Hume’s evidential criteria—he dismissed them as irrelevant because they would have entailed what he considered a violation of nature.350 His dismissal, then, rested on his argument challenged above; Jansenist claims were rejected by Hume, Conyers Middleton, and others not because of lack of evidence but simply because they were miracle claims.351
Hume follows the line of argument established by a deist predecessor challenging recent miracle claims surrounding the Huguenots (1705). Deist Thomas Chubb emphasized the vast number of otherwise reliable witnesses, arguing that they were better attested than nt miracle claims, yet dismissed their credibility by appealing to the authority of consensus: “not one in ten believe it now.”352 Other deists advanced the same form of argument, which Hume merely dressed in a special outfit, with Jansenists as the target.353 Hume could scoff at Jansenists with impunity, since they were too Catholic for Protestant tastes (Hume’s primary English readership) yet not acceptable to Catholics either.354
Hume recognized abundant reliable witnesses for such a case and that by all normal means of inquiry, one would conclude that a miracle occurred. Nevertheless, Hume felt justified in dismissing such evidence by appealing to his premise that miracles are impossible.355 It is difficult to comprehend what would qualify as circular reasoning if this approach does not. Hume could logically deny that any evidence for a miracle can be compelling only if he could a priori show that miracles are “logically impossible (that is, conceptually impossible, like a ‘square circle’ or a ‘married bachelor’)”; yet Hume does not do so.356 Some critics further counter that Hume’s own approach is epistemologically flawed in that it proves referentially self-defeating.357
Hume’s argument is not inductive; rather, it is designed to support his conclusion. When he cites the need for public events attested by many credible eyewitnesses and then dismisses even their testimony, his language is too general to function as a full argument in itself. Rather, he is simply listing ad hoc characteristics that Christian apologists cited in favor of the apostolic witnesses, which he then finds deficient.358 His failure to provide a complete argument at these points invited severe critique from his contemporaries.359
Other Noninductive Elements in Hume’s Approach
Other factors also indicate that Hume is not arguing inductively. I have noted first that Hume does not argue inductively, but constructs a deductive argument against miracles based on a probability rigged by his nontheistic starting assumptions. A second observation is that Hume’s explicit exclusion of beliefs of “ignorant and barbarous nations”360 reflects ethnocentric bias that the vast majority of scholars would reject as unacceptable today. This is a serious problem, but I reserve a more extensive response to it for my discussion in later chapters (most explicitly in ch. 7). Suffice it to note now that he was again adopting a typical deist argument; John Toland, for example, condemned superstitions that flourished among “ignorant and barbarous” peoples.361 Third, Hume explicitly mentions even some European miracle claims from his own era (i.e., the Jansenists) but then rejects them, because, he contends, miracles cannot happen. I have already commented on the circular character of this reasoning.
Fourth, Hume uses many bogus claims of miracles (already rejected by many Christian critics) to deny the reality of any miracles.362 This guilt-by-association approach, however, reflects the logical fallacy of false analogy, of generalizing based on specific cases without examining other cases that may differ in relevant details.363 Hume thus effectively argues here against a straw man. To proceed genuinely inductively, Hume would have to examine each miracle claim and show it to be false; and he still would not have foreclosed the possibility of some miracle claim. So long as he proceeds inductively, a single confirmed miracle would disprove his case.364 Indeed, reliable witnesses for sufficiently numerous different miracles, if genuinely independent, support the class of events.365
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. Inferring from superstitious supernatural claims that all supernatural claims must be rejected is logically analogous to rejecting any form of theism because we have found earlier forms of polytheism wanting. The latter argument would have been more scandalous in his day, however; his milieu was better prepared to reject direct divine action in nature than to reject theism in general.
This excerpt is from Craig S. Keener, Miracles: The Credibility of the New Testament Accounts, 2 volumes, Baker Academic, a division of Baker Publishing Group, 2011. Used by permission. All rights to this material are reserved. Material is not to be reproduced, scanned, copied, or distributed in any printed or electronic form without written permission from Baker Publishing Group.
Footnotes appear in the full digital issue of Pneuma Review Fall 2013 and in the book from which this excerpt is derived.