Introduction: Locke and Hume on competing miracles

Locke, like other Christian apologists, cites the testimony of the miracles of Jesus as evidence for Christianity. Since Hume, most of the philosophical literature on miracles has focused on whether it would be rational to believe the testimony of a miracle. But that is not my topic. Suppose instead, just for the sake of argument, that Locke is right that it is rational to believe the testimony of the miracles of Jesus and that these miracles provide reasonably good evidence for Christianity. The issue considered in this article is whether the testimony of miracles in competing religious traditions undermines the argument from miracles for Christianity. Hume, among others, argues that other religious traditions can make the same kind of argument from miracles, in which case, he thinks, we would have no reason to believe Christianity over the equally plausible alternatives. Call this the ‘competing miracles objection’ to the argument from miracles. In ‘A Discourse of Miracles’, Locke tries to undermine the force of this objection. The aim of this article is to explain and defend Locke's reply.

My main contribution here will be to clarify the role of defeaters in miracle arguments, both in Hume's version of the competing miracles objection and in Locke's (pre-emptive) reply. Christian apologists present the testimony of the resurrection of Jesus as evidence for Christianity, which we are assuming (for the sake of argument) initially justifies a belief in Christianity. Hume then tries to undermine this argument by appealing to miracles in other conflicting religious traditions. He suggests that the evidence from miracles for one religion is, correspondingly, evidence against another incompatible religion. Thus, on Hume's view, competing miracles act as a defeater that diminishes the evidence for Christianity.

While Hume's competing miracles objection is well known, it has thus far gone unnoticed that Locke takes himself to identify a defeater for the competing miracles defeater (i.e. Locke identifies a defeater-defeater). Locke argues that there is more evidence from miracles for Christianity, and so given our evidence we ought to believe in Christianity. Yet, even if there is more evidence from miracles for Christianity, we might worry, as Hume does, that miracles in other religious traditions at least weaken the evidence from miracles for Christianity. However, in the most interesting move in this exchange, Locke denies that miracles supporting other religions weaken the evidence for Christianity. He argues that God will ensure that the greatest evidence from miracles supports genuine revelation, and since the greatest evidence supports Christianity, we can dismiss the inferior evidence for other religions as ultimately misleading evidence. The result, he thinks, is that the superior evidence for Christianity is a defeater for the competing miracles defeater.

My goal in this article, then, is to explain why Locke thinks that competing miracles do not weaken the evidence for Christianity, and to defend the plausibility of his general strategy. To do this, I will first use Hume to clarify what exactly the objection is supposed to be. (Although Hume is after Locke, Locke is responding to the same kind of objection.) I will then explain how, on Locke's view, the greater evidence from miracles for Christianity acts as a defeater to the competing miracles objection (i.e. as a defeater for the competing miracles defeater). Finally, I suggest that Locke's argument, if successful, has broader application for religious epistemology: it suggests that evidence for competing religious traditions does not always diminish the justification for belief in a specific religion.

The competing miracles objection

Hume takes the competing miracles objection to undermine the argument from miracles that Locke and other Christian apologists use as evidence for Christianity.Footnote ¹ The argument for Christianity has two stages. First, Christian apologists appeal to the testimony of the disciples, recorded in the New Testament, as good evidence that Jesus was resurrected. They argue that the testimony of the resurrection is more likely to be true than false because it is very likely that the witnesses would say that they saw the resurrected Jesus if it was true but very unlikely if it was false. It follows that the testimony of the resurrection is very good evidence for the resurrection.Footnote ² Finally, if Jesus was resurrected, then Christianity is almost certainly true. Thus, the apologists argue, the testimony of the resurrection is good evidence for Christianity.Footnote ³ While both steps of the argument are controversial, for purposes of this article we are just granting that the argument is successful. The point of this article is to evaluate the objection that even if the testimony of the miracles of Jesus provides good evidence for Christianity, the argument still does not succeed because other religious traditions can make the same kind of argument. Let us assume that miracles can provide evidence for a religious tradition, then, and see whether competing miracles undermine the argument for Christianity.

Even if the miracles of Jesus taken on their own justify a belief in Christianity, sometimes the evidence that initially justifies a belief that p later turns out to be mistaken; a defeater is the further evidence that, once we become aware of it, makes it so that the belief that p is no longer justified.Footnote ⁴ A rebutting defeater is new evidence we become aware of that p is actually false. For example, suppose an argument for p appears to be sound, and so the argument initially justifies a belief that p, but we later are presented with a decisive argument that p is false. The argument that p is false would be a rebutting defeater for the argument for p. By contrast, an undercutting defeater is evidence that the initial evidence for p is unreliable and hence not good evidence for p after all (i.e. the evidence for p is undermined). For example, suppose an argument for p appears to be sound, and so the argument initially justifies a belief that p, but we later discover a mistake in the logic of the argument. The conclusion may or may not be true, but our reasons for believing the conclusion turned out not to be very good evidence. An undercutting defeater for the argument from Christianity, then, would not be evidence that Christianity is false, but rather that belief in Christianity is unjustified. Hume thinks that competing miracles are an undercutting defeater for the argument from miracles for Christianity.

Hume argues that a miracle that is evidence for one religion is also evidence against another religion, and so the evidence from miracles ‘destroys itself’ (Enquiry, 121). He claims that ‘all the [miracles] of different religions are to be regarded as contrary facts, and the evidences of these prodigies, . . . as opposite to each other’ (ibid., 122). Christianity and other competing religions, such as Islam and Buddhism, etc., are incompatible. So, insofar as the resurrection is good evidence for Christianity it is also evidence against Buddhism, and miracles supporting Buddhism undermine Christianity, and so on for other religions. In this way, Hume suggests, ‘In destroying a rival system, it likewise destroys the credit of those miracles, on which that system was established’ (ibid., 121–122).

David Johnson (Reference Johnson1999, 80–88) objects to Hume's claim that the miracles of one religion are evidence against another. He points out that the evidence for one hypothesis is not always evidence against an incompatible hypothesis (ibid., 81). For example, suppose there is a lottery with tickets numbered 1–1000. Compare two incompatible hypotheses: one hypothesis is that 209 is the winning ticket, and another is that 207 is the winning ticket. Now suppose we learn that the winning ticket is between 201 and 210. This new information is evidence for the hypothesis that the winning ticket is 209, but it is not also evidence against the incompatible hypothesis that the winning ticket is 207; the new information is evidence for both hypotheses. Hence, Johnson concludes, the evidence for one theory is not always evidence against an incompatible theory. Further, he suggests that this is true for miracles (Reference Johnson1999, 81).

However, sometimes the evidence for one theory is evidence against an incompatible theory. In Johnson's lottery example, we do not have any more reason to believe that the winning ticket is 209 over the incompatible hypothesis that the winning ticket is 207. In that case, the evidence that the winning ticket is between 201 and 210 raises the probability of both incompatible hypotheses. But whenever the evidence makes one hypothesis most likely to be true, the evidence for that hypothesis is also evidence against any incompatible hypothesis. For example, if the coin lands on heads, then it does not land on tails. Suppose we learn that a coin is weighted in such a way to land on heads more frequently. The higher the probability is that the coin will land on heads entails a correspondingly low probability that the coin will land on tails. So likewise, if Christianity is probably true given the evidence from the resurrection, then that entails incompatible religions are probably false given the evidence from the resurrection.Footnote ⁵ (One and the same miracle cannot be evidence that Christianity is most likely true and evidence that Buddhism most likely true and hence that Christianity is most likely false.) We might take ‘good evidence’ to be, at a minimum, evidence that makes something more likely to be true than not. Christian apologists claim to have such evidence from testimony for the resurrection and that the resurrection is good evidence for Christianity. If so, then Hume is right to claim that the resurrection is evidence against other religions.

While Hume is right that if one miracle is evidence for a religion then that miracle is also evidence against an incompatible religion, it is less clear that the combined evidence from competing miracles undermines the argument for Christianity. Like Johnson (Reference Johnson1999), Jordan Howard Sobel (Reference Sobel2003, 596–597 n. 24) constructs an example that calls into question Hume's claim that the evidence for one religion is evidence against an incompatible religion. Suppose one urn, call it urn A, has three marbles: one rough and white, one smooth and white, one smooth and green. A second urn, urn B, also has three marbles: two rough and green, one smooth and white. Suppose all that we know is that a white marble is drawn from one or the other urn. Since two of the three marbles in urn A are white, the probability that the marble came from urn A given that it is white is 2/3 (whereas the probability that it came from urn B is only 1/3). But, separately, suppose all we know is that a rough marble is drawn from one or the other urn. Since two of the three marbles in urn B are rough, the probability that the marble came from urn B given that it is rough is 2/3 (whereas the probability that it came from urn A is only 1/3). If the fact that the marble is white and the fact that it is rough are taken separately, we have equally good evidence for two competing theories. But when the evidence is combined, it is obvious that the marble must have come from urn A because there is only one marble that is both white and smooth and it was in urn A. Thus, Reference SobelSobel claims: ‘when evidence that would alone establish a theory is combined with evidence that would alone to the same probability establish another theory, the evidence does not “destroy itself”’ (Sobel (Reference Sobel2003), 596 n. 24). It is therefore not guaranteed to be true that when we combine the evidence from competing miracles they would undermine the argument from the resurrection for Christianity.

Again, though, sometimes the evidence for one theory is evidence against an incompatible theory even when the evidence is combined. Suppose we are not in a position ourselves to see the results of a coin flip and we rely instead on two equally reliable witnesses, one of which says the coin came up heads while other says the coin came up tails. The probability that it landed on heads given a witness said so, taken in isolation, is quite high (say, 9/10). But the other witness is equally reliable, and so the probability that it is heads given the other witness said it was tails, again taken in isolation, is quite low (say, 1/10). When evidence from these two witnesses is combined, we have no more reason to believe it was heads than tails.Footnote ⁶ Hume (perhaps wrongly) thinks the argument from miracles is more like this coin example than Sobel's urn example.

Hume claims that equally good arguments from miracles can be made in competing religious traditions, and this undermines any one version of the argument from miracles. He says that the testimony of miracles in different religions each ‘has it the same force’ and should be received ‘in the same light’ and ‘with the same certainty’ (Enquiry, 121–122). For example, some Buddhists have claimed that the Buddha levitated while meditating and that this miracle is evidence for Buddhism. If Christianity and Buddhism have equally good evidence from miracles, then the evidence for Christianity (the resurrection) and against Christianity (the Buddha levitating) are equally good evidence, and we thus have no more reason to believe Christianity than Buddhism, and no more reason to believe Buddhism than Christianity. ‘Every miracle, therefore, pretended to have been wrought in any of these religions’, Hume claims, will ‘overthrow every other system’ (Enquiry, 121).

The controversy about competing miracles can be summarized in the following way. The argument from miracles for Christianity can briefly be stated as:

The Argument from Miracles for Christianity Footnote ⁷

1. 1. The testimony of the resurrection of Jesus is good evidence for the resurrection.

2. 2. The resurrection is good evidence for Christianity.

3. 3. So, the testimony of the resurrection is good evidence for Christianity.

Hume objects that if the testimony of the resurrection is good evidence for Christianity, then the testimony of competing miracles is equally good evidence for competing religions:

The Argument from Miracles for a Competing Religion

1. 1. The testimony of a competing miracle is (equally) good evidence for the competing miracle.

2. 2. The competing miracle is (equally) good evidence for a competing religion.

3. 3. So, the testimony of the competing miracle is (equally) good evidence for the competing religion.

But the Christian apologists cannot grant the latter argument is successful. For, as Hume argues, if there are equally good arguments from miracles for incompatible religions X and Y, then we should not believe religion X (nor religion Y) on the basis of miracles. The argument from competing miracles, then, is intended to show the Christian apologists that the inference from the testimony of the resurrection to the truth of Christianity is not a good one. That is, the competing miracles objection is an undercutting defeater for the argument from miracles for Christianity.

Locke accepts that if there are (what Sextus Empiricus would call) equal and opposite arguments from miracles, then we cannot believe one argument over the other. He recounts the story of Moses going to the court of Pharaoh. As a sign that he had received revelation from God, Moses miraculously turns his staff into a snake, but then the priests of Pharaoh do the same. Locke says that ‘if matters had rested there’, then who had ‘the truth on their side, could not have been determined’ (‘Miracles’, 260). Since there is equal and opposite evidence for each of their claims for divine authority, there is no more reason to believe one over the other. So, as Shelley Weinberg (Reference Weinberg2020, 258) explains, we ‘begin to reflectively question whether . . . [our] initial assent was warranted given that [we] now have a reason to doubt’. That is, at this point in the story, the competing miracle of the priests is an undercutting defeater for the initially justified belief that Moses had divine authority. But the story does not stop there. Moses later performs miracles that the priests could not replicate, and at that point it is ‘plain [the priests] acted by an inferior power’ and so ‘the decision was easy’ to believe Moses (‘Miracles’, 260).

The competing miracles objection depends on equal and opposite arguments from miracles, and thus there are two ways for the objection to fail. First, some miracles are not really opposed because they do not provide contrary evidence. For example, a miracle recorded in the Hebrew Bible may be evidence for Judaism, but this is not really evidence against Christianity because Christianity accepts the Hebrew Bible as scripture (ibid., 258). Second, when there are opposing arguments, the arguments must be equally plausible. If the argument from miracles for Christianity is a better version of the argument, then we can accept that version of the argument over the alternatives, just as Christian apologists claim. In short, the objection requires there to be equal and opposite arguments, and the objection would fail if the arguments from miracles are not really opposed or not equally plausible.

As we will see in the succeeding sections, Locke rejects the claim that there are equal and opposite arguments from miracles. In his first reply, he limits the scope of the competing miracles by arguing that most miracles are not opposed to Christianity. In his second reply he denies that, even when there are competing miracles, the arguments from competing miracles are equally plausible; instead, he argues for the superiority of the evidence for Christianity. If he is right about either of those claims, then the competing miracles objection fails.

Locke's first reply: limiting the scope of the objection

Miracles are said to have occurred in every world religion, but most of these miracles do not conflict with Christianity. While Hume claims that the argument from miracles for Christianity is ‘opposed by an infinite number of witnesses’ of competing miracles (Enquiry, 121), Locke claims that the ‘miracles for the confirmation of revelation, are fewer than perhaps imagined’ (‘Miracles’, 257). Locke argues that most religions do not appeal to miracles as evidence for revelation and so the extent to which the competing miracles objection poses a problem is rather limited.

Some religious traditions do not have miracles that conflict with Christianity because they do not appeal to miracles as evidence for the religion. For example, Locke claims that there are no competing miracles in Islam because Muhammad ‘pretends to no miracles for the vouching of his mission’ (ibid., 258). This is controversial, both because the writing of the Quran is regarded as itself a miracle and because oral tradition, recorded later, includes miracles by Muhammad. However, some Muslims do regard the writing of the Quran as the one miracle by Muhammad. If that view is adopted, then Locke would be right that there is no other miracle that serves as evidence for the revelation recorded in the Quran. In some cases, then, there is no competing argument from miracles because there is no testimony of a miracle that serves as evidence for revelation.

Further, even if miracles occur in different religious traditions, it still does not follow that those miracles provide conflicting evidence. Some miracles are neither performed nor interpreted as evidence for a specific religious tradition, and in such cases, Locke says, ‘how many or great soever [these miracles are], revelation is not concerned in’ (ibid., 257). Travis Dumsday suggests that ‘Locke is correct in restricting the scope of competing miracles’ in this way (Dumsday (Reference Dumsday2008), 419). Dumsday imagines ‘a Hindu mother who fervently prays to Ganesha to cure her child of a fatal disease’ and suggests that ‘one can easily imagine [the Christian] God granting this prayer by a miracle’ (ibid., 420). Miracles like this do not conflict with Christianity, and so do not provide evidence against Christianity.

Other miracles are used as evidence for revelation but still do not conflict with Christianity. Locke thinks that the miracles of Moses provide evidence for his claim to have received revelation, but even though these miracles are evidence for Judaism they are not also evidence against Christianity because Christians accept the revelation of Moses (ibid., 258). Put differently, if the miracles of Moses really did happen, this would not lower the probability of Christianity being true.

Finally, there are miracles that are can be taken as evidence for a competing religion and yet still these miracles are not necessarily evidence against Christianity. Locke argues: ‘The heathen world, amidst an infinite and uncertain jumble of deities, . . . had no room for a divine attestation of any one against the rest’ (ibid., 257). If those who worship Athena receive a revelation confirmed by miracles, that does not necessarily conflict with those who worship Brahma confirmed by other miracles. Given the non-exclusive nature of polytheism, ‘no one of [the gods] could be supposed in the pagan scheme to make use of miracles to establish his worship alone, or to abolish that of the other’ (ibid.). So, Locke seems to imply, miracles in polytheistic religions are not evidence against Christianity.

Most miracles do not provide evidence against Christianity because they are not relevant to the truth of a specific religious tradition. When the Hindu mother's prayer is miraculously answered, for example, this might raise the probability that a general religious worldview is correct, but it is not the case that Hinduism is most likely to be true given this miraculous healing. Similarly, the Buddha is sometimes said to have levitated while meditating. I suppose that, if true, this miracle is evidence for Buddhism. Yet, it is not the case that Buddhism is most likely to be true given this miracle. This is true in general: for most miracles, it is not the case that the religion is most likely true given the miracle. And if so, then, as with Johnson's (Reference Johnson1999) lottery example, it is not true that the evidence from miracles for one religion is evidence against another.Footnote ⁸ Although Locke probably meant to imply something stronger, his point about the non-exclusive nature of polytheism could be taken to imply, at the very least, that miracles in polytheism are not good evidence for a specific religion (i.e. the religion is most likely true given that miracle). If that is the point, then Locke is right that the miracles in polytheism do not constitute evidence against Christianity.

As we have seen, Locke thinks that the competing miracles objection has a rather limited scope, and so the objection is overblown. In fact, the argument from miracles is uniquely important to Christian apologetics, and the emphasis that Christianity alone places on the argument from miracles supports Locke's contention that the competing miracles objection is overblown. Christians have always relied on the miraculous resurrection of Jesus to support belief in Christianity.Footnote ⁹ By contrast, there are not many miracle arguments in non-Christian religious traditions. There are some. But there are comparatively few. Locke is largely correct, then, to emphasize the limited scope of the competing miracles objection.

Notwithstanding the limited scope of the objection, though, there remains some conflicting evidence from miracles. Locke's next move is to argue that, even when there is some genuine conflict, the arguments from miracles are not equally plausible.

Locke's second reply: denying miracle arguments are equally plausible

The competing miracles objection is that there are equal and opposite arguments from miracles and that this acts as a defeater for the argument from miracles for Christianity. As we have seen, Locke's first response to the objection is to limit the number of opposing arguments from miracles. His second response is to develop a standard for judging between any two competing arguments from miracles. By doing so, Locke denies that competing miracle arguments are equally plausible.

The argument from miracles has two steps. First, argue that a miracle probably took place given the evidence from testimony. Second, show that if this miracle occurred then a specific religion is probably true.Footnote ¹⁰ So, when Locke denies that there are equally plausible arguments from miracles (i.e. he insists that there is more evidence from miracles for Christianity), he has two options: (1) he can argue that the probability of the resurrection is greater than the probability of competing miracles, or (2) he can argue that the probability of Christianity given the resurrection is greater than the probability of a competing religion given its competing miracles. Locke focuses mostly on the latter.

When there are competing claims to revelation, both supported by miracles, Locke claims that we should believe the revelation backed up by ‘the marks of a greater power than appears in opposition to it’ (‘Miracles’, 259). He argues:

Locke here appeals to God's power and goodness to argue that a genuine revelation from God will always be backed up by the greater of two competing miracles. As an omnipotent being, God can perform greater miracles than any contrary power. Further, because God is good, he will not allow ‘the appearance of a greater power [i.e. greater miracle] on the side of an impostor, and in favour of a lie’. Since God wants us to believe, on the basis of evidence, genuine revelation from him, Locke thinks that God will ensure that the greater evidence (i.e. the greater miracle) will always support genuine revelation.Footnote ¹¹

Some worry that Locke's greater-miracle criterion will open the door to being easily deceived.Footnote ¹² In reply, Locke argues that God will guarantee that opposing arguments form miracles are not equally plausible:

Locke suggests that because God is good he would not allow the greater evidence to back up a false claim to revelation and so God, being more powerful than any other being, would provide greater evidence, by way of a greater miracle, for genuine revelation. God simply would not allow false priests or tricksters to deceive others with greater evidence from miracles than God performs in support of genuine revelation.Footnote ¹³ For this reason, ‘the marks of a superior power accompanying [the event], always have been, and always will be, a visible and sure guide to divine revelation’ (ibid., 262).

Another potential worry about the greater-miracle criterion is that ‘Locke does not spell out exactly what makes one miracle greater than another’ (Dumsday (Reference Dumsday2008), 419). Locke mostly relies on an intuitive grasp of what would require ‘greater power’. Locke identifies miracles by what seems to be a violation of the law of nature caused by God (‘Miracles’, 256),Footnote ¹⁴ and so it should be unsurprising that he identifies the greater miracle by what seems greater: ‘the marks of greater power should have a greater impression on the minds and beliefs of the spectators’ (ibid., 263–264). Locke does, though, have a little more to offer than bare intuition. First, Moses can perform miracles that the priests cannot replicate, implying that Moses has a greater power (ibid., 260).Footnote ¹⁵ Second, Locke says that more miracles are evidence of greater power, ‘two supernatural operations showing more power than one, and three more than two’ (ibid., 263).Footnote ¹⁶ So miracles that one person can do that another cannot, and a greater number of miracles, both contribute to an assessment of a ‘greater’ miracle.

Having established the greater-miracle criterion, Locke then uses it to argue for the truth of Christianity. He argues, ‘the number, variety, and greatness of the miracles wrought . . . by Jesus Christ’ make the argument from miracles for Christianity stronger than any other competing religion (ibid., 261). Jesus performed the most and the greatest miracles, and so there is more evidence for Christianity than any rival religious tradition. Locke concludes, then, that Christianity has a stronger argument from miracles than other religious traditions.

Although Locke emphasizes the greater-miracle criterion, he also seems to think that there is better evidence for the miracles of Jesus than for competing miracles. He claims that the evidence for the miracles of Jesus and the apostles is so strong that ‘the Enemies of Christianity have never dared to deny them’ (Reasonableness, 146). By contrast, the evidence for competing miracles is questionable. Locke says: ‘what the . . . Indians say of their Brama (not to mention all the wild stories of the religions farther East), is so obscure, so manifestly fabulous, that no account can be made of it’ (‘Miracles’, 258). Mooney and Imbrosciano (Reference Mooney and Imbrosciano2005, 164) describe Locke's dismissive attitude here as ‘colonialist arrogance’, arguing that any competing miracles are ‘conveniently dismissed a priori as inevitably inferior to Christianity’.Footnote ¹⁷ But perhaps Locke's point is that Hindu scripture is not very good historical evidence, and so the historical evidence for the miracles of Christianity is better evidence than the (non-historical) descriptions of miracles in Hinduism. At any rate, other Christian apologists make this argument and there is no reason to think Locke would deny it.

The competing miracles objection depends on the premise that opposing miracle arguments are equally plausible, but Locke denies this. Further, this is not accidental. Locke thinks that God is committed to ensuring that there is more evidence from miracles for Christianity (if true) than for other religious traditions. In his view, there is both better evidence for the resurrection than competing miracles and the resurrection is better evidence for Christianity than competing miracles are for other religions. So, Locke can deny that there are equal and opposite arguments from miracles, undermining the competing miracles objection.

Locke's defeater-defeater: greater evidence

Even if Christianity has the strongest argument from miracles, there is one lingering worry about competing miracles. Any miracles supporting a competing religion, Hume and others argue, diminish the confidence we ought to have in the argument for Christianity. So, even if there is comparatively better evidence for Christianity, competing miracles can still significantly weaken the evidence for Christianity. Locke, though, thinks that he can grant that miracles in other religious traditions occur without undermining the argument for Christianity.

As I indicated above, I think Hume is best interpreted as taking competing miracles as an undercutting defeater that undermines the argument from miracles for Christianity. Hume's intention, I think, is to completely undermine the argument for Christianity: he argues that, given competing miracles, ‘the testimony [of miracles] destroys itself’ and ‘destroys the credit of those miracles, on which that system was established’ (Enquiry, 121–122). Charity Anderson (Reference Anderson, Benton, Hawthorne and Rabinowitz2018, 19–20, 24) agrees that Hume tries to completely undermine the argument from miracles, but she thinks what he really does is reduce the evidential force of the argument. Justification comes in degrees and, she suggests, so might the strength of undercutting defeaters.

Compare two cases. In one case, the testimony that a person committed a crime might give us some evidence that there was a crime, but if we later acquire decisive evidence that the ‘eyewitness’ was nowhere near the location of the supposed crime, then this completely undermines the evidential weight of her testimony; we should give her testimony no weight when making a judgement about what happened. In a second case, the testimony that a person committed a crime again might give us evidence of the guilt of the suspect, but we later acquire decisive evidence that the eyewitness has a vendetta against the suspect. Perhaps to some degree the vendetta undermines the evidential value of the witness's testimony, but it does not completely undermine the evidence from his testimony. A defeater, then, might partially undermine the initial evidence.

Even if competing miracles are not equally plausible arguments, Hume implies that any competing miracle would at least partially undermine the evidence from miracles for Christianity. He claims that when there are two sources of conflicting evidence ‘the superior [evidence] only gives us an assurance suitable to that degree of force, which remains, after deducting the inferior’ (Enquiry, 116). Following Locke, suppose that Christianity is most likely to be true given the argument from the resurrection. Still, Hume insists that our confidence in the truth of Christianity ought to be diminished proportionally to the strength of the evidence from miracles for Islam, and so on. In short, even if Christianity is believed on the basis of the argument from miracles, the competing miracles objection could be taken as an undercutting defeater that weakens (but does not completely undermine) the evidence from miracles for Christianity.

Hume's claim that when we have evidence for two competing theories we somehow ‘subtract’ one probability from the other has been widely panned. Earman (Reference Earman2000, 43) calls it ‘nonsense’, Johnson (Reference Johnson1999, 81) thinks it is ‘patently false’, and Sobel (Reference Sobel2003, 319) ‘despair[s] of finding a ‘saving’ symbolization’ of the claim. Still, other interpreters are tempted to agree with Hume.Footnote ¹⁸ Locke's response, then, can be seen as an explanation for why Hume is wrong about the need to ‘subtract’ probabilities in the case of competing miracles.

Locke denies that competing miracles weaken the evidence for Christianity. Even if ‘the most knowing spectator’ acknowledges that a miracle for a competing religious tradition has occurred, this ‘cannot at all shake the authority’ of the argument from the greater miracle:

For example, the miracles done by the priests of Pharaoh, being inferior to the miracles of Moses, ‘could not in the least bring in question Moses's mission’ and, in fact, his mission ‘stood the firmer for this opposition, and remained the more unquestionable after this, than if no such signs had been brought against it’ (ibid.). According to Locke, competing miracles do not at all undermine the evidence for Christianity. Locke, in effect, argues that the superiority of the evidence for Christianity is a defeater for the competing miracles defeater.

Locke's response works, if at all, because God is committed to providing the most evidence from miracles for Christianity (if true, or Islam if true, etc.). By way of analogy, suppose theories A and B are mutually exclusive and given our evidence theory A is most likely true. Normally, we are not assured that if A were true then we would have more evidence for A. In that case, as Hume suggests, the evidence for B may weaken the evidence for A. But imagine that, for whatever reason, we knew in advance that if theory A were true then there would be more evidence for A than B, but if theory B were true then there would be more evidence for B. In that case, the greater evidence for A would be an undercutting defeater for the evidence for B. Even if there is prima facie evidence for theory B, we would know that this evidence is ultimately misleading; for if B were correct, then (we are assuming) there would be more evidence for theory B, but there is not. So, evidence for B can be dismissed. In this case, the greater evidence for A is a defeater-defeater for the competing evidence for B.Footnote ¹⁹ Locke thinks the same holds for religion.

Locke thinks God would not allow there to be greater evidence from miracles for a false claim to revelation; instead, God will guarantee that genuine revelation will be backed up with sufficient evidence for us to rationally believe it. Further, he thinks that the evidence comes from miracles. And since there is more evidence from miracles for Christianity than for other religious traditions, this assures us that any evidence from miracles for other religious traditions is misleading evidence (‘he cannot but know they are not seals set by God’, ibid.), for only the greater miracle is evidence for revelation. Once it is shown that there is more evidence for the miracles of Jesus, this greater evidence is then an undercutting defeater for the competing miracles defeater.

While Locke emphasizes the greatness of the miracle, the greatness of the miracle is relevant only because, in his view, the greater miracle provides greater evidence, and there are two ways for miracles to provide greater evidence from miracles. The first step of the argument from miracles is to provide evidence that the miracle occurred, and the second is to show that this miracle is good evidence for a specific religious tradition. Consequently, the argument from miracles can be strengthened by either providing more evidence for the miracle or the miracle providing more evidence for the religion. Locke's greater-miracle criterion concerns the second step of the argument, but both steps are relevant to evaluating an argument from miracles.

A notable implication of Locke's view is that it is open to revision. If a new ‘greater’ miracle occurred in support of a competing religion, and the competing religion is also consistent with natural theology,Footnote ²⁰ then Locke implies that we ought to believe that competing religion instead of Christianity. I take it that Locke here admits this implication: we should believe Christianity ‘till any one rising up in opposition to him shall do greater miracles than [Jesus] and his apostles did’ (ibid., 261). As already noted, though, Locke thinks the evidence for Christianity is very strong, in terms of both the types of miracles performed (ibid.) and the quality of that evidence (Reasonableness, 146).Footnote ²¹ So, he thinks it is unlikely, though possible, that we would discover a competing religious tradition that has more combined evidence for the miracle and then for the religion than the miracles of Jesus provide for Christianity.

An interesting possibility is to imagine an argument from miracles that is exceptionally strong in one step but weak in another.Footnote ²² For example, imagine there are thousands of witnesses of a trivial miracle that, even if true and miraculous, would not provide substantial evidence for a competing religious tradition.Footnote ²³ Alternatively, perhaps there is almost no evidence whatever for a really big miracle, but if actual then that would provide decisive evidence for a competing religious tradition.Footnote ²⁴ Locke could grant that one or the other step of a competing miracle argument could be greater than the corresponding step of the argument for Christianity (if true), but not both; otherwise, he could not claim that the greater-miracles ‘always have been, and always will be, a visible and sure guide to divine revelation’ (‘Miracles’, 262).

Another interesting implication of Locke's argument here is that, insofar as the quality of the evidence matters to the argument from miracles, God will ensure that the evidence from the miracles of Jesus is preserved. If there were eyewitnesses to the miracles of Jesus, but this was never written down, then later generations would not have evidence from miracles for Christianity. Likewise, if the evidence for competing miracles were better preserved, then that could tip the scales against belief in Christianity. So, for Locke, God has a reason to ensure that the historical evidence for the miracles of Jesus is preserved. I do not find it wholly implausible to think that, if Christianity is true, God would take a special interest in preserving the evidence of miracles of Jesus.

One nice feature of Locke's view is that he can grant that competing miracles occur and provide prima facie evidence for their respective religious traditions without undermining the evidence from miracles for Christianity. For example, he does not deny that the priests of Pharaoh performed miracles, nor that these miracles provide prima facie evidence for their claim to divine authority. Later, though, Moses performs miracles that the priests cannot replicate. At that point, the greater evidence in support of Moses' claim to revelation is a defeater for the prima facie evidence in support of the priests's claim to divine authority. Similarly, Locke could grant that the Buddha miraculously levitated without diminishing the evidence for Christianity because only the greater miracle is evidence of genuine revelation.Footnote ²⁵

Locke's response to the competing miracles objection is interestingly different from those given by other Christian apologists. Christian apologists often try to undermine the evidence from competing miracles on a case-by-case basis.Footnote ²⁶ For example, the earliest reports of the miracles of Muhammad are first recorded long after he lived, and miracles do not fit within the Buddha's own ideology, which makes any report of his miracles suspect; these are taken as undercutting defeaters for the testimony of these competing miracles.Footnote ²⁷ But notice that, on this strategy, the Christian apologists needs to look for undercutting defeaters for competing miracles one at a time: each miracle is undercut by independent considerations. By contrast, Locke provides one general consideration that acts as a defeater for all (inferior) competing miracles.

So, Locke's strategy is different in two important ways. First, rather than finding defeaters one by one for each competing miracle, Locke argues that the greatest evidence from miracles is a defeater-defeater for all other arguments from miracles. Second, Locke's response does not depend on rejecting the testimony of competing miracles. Just as he grants that the priests of Pharaoh performed miracles, he can grant that Muhammad or the Buddha performed miracles. Yet, he insists, these competing miracles do not weaken the evidence from miracles for Christianity. That's interestingly different from the typical strategy used by Christian apologists.

Locke's argument can be described narrowly or more broadly. Narrowly, the greater miracles of Jesus defeat the competing miracles defeater. Yet, the greater miracles of Jesus act as a defeater-defeater only because greater miracles provide greater evidence. More broadly, then, Locke's point can be described as the view that the greater evidence is a defeater for any (lesser) competing evidence.

On the broader view, the evidence from miracles is not the only thing that matters. Indeed, Locke rejects any revelation that is ‘inconsistent with natural religion’ (‘Miracles’, 261). For example, Locke argues that ‘the Existence of more than one GOD’ is ‘contrary to Reason’ (Essay, 4.17.23), and that we should never accept a purported revelation that is contrary to reason.Footnote ²⁸ I agree with Locke that revealed religion should be consistent with natural theology. Thus, even if the miracles of Jesus provide important evidence for Christianity, belief in Christianity does not depend solely on the evidence from miracles. Locke's more general position, then, is that God will ensure that the greater total evidence will support Christianity (if true) and that this greater evidence is a defeater for any (lesser) evidence for a competing religion.

It might be objected that Locke's reliance on the greater-miracle criterion is circular or arbitrary. If he is picking the greater-miracle criterion because Christianity has the greatest miracles, then that would be circular. Alternatively, we might wonder why a proponent of another religion cannot appeal to some other criterion that would be more favourable to that religion. For example, a Buddhist could take enlightenment, a type of subjective religious experience, as better evidence for religion than an outward miracle, and then conclude on the basis of such an experience that Buddhism is to be preferred over Christianity.Footnote ²⁹ Unless Locke has a good reason for picking the greater-miracle criterion specifically, then his choice of criterion seems arbitrary.

Again, though, Locke's view is that we should believe whatever is most likely to be true given the total evidence and that God will provide the greater evidence for genuine revelation. The commitment to evidentialism does not require any religious commitment, and Locke thinks there is compelling evidence from natural theology that God exists and is perfectly good (Essay, 4.10.1–6). His theism does not commit him specifically to Christianity, though, since this commitment is shared with other monotheistic traditions. What is perhaps unique to Locke is the view that God is committed to providing the greater evidence for Christianity (if true, or Islam if true, etc.). However, Locke thinks this follows from theism generally, not Christianity specifically. He argues that God wants us to be rationalFootnote ³⁰ and so if God reveals something then he will provide the evidence we need to rationally believe it.Footnote ³¹ Further, since God is good, God will not allow us to be rationally deceived by evidence for a false claim to revelation, and so will ensure that genuine revelation is supported by greater evidence.Footnote ³² Finally, Locke is highly critical of private religious experiences, and so is driven instead to miracles as publicly available evidence.Footnote ³³ For Locke, then, the greater-miracle criterion follows from general considerations in natural theology about the nature of God and the available evidence. Also, in his view Christianity is supported by the total evidence (including these reflections in natural theology), and not solely by the evidence from miracles. So, his selection of the greater-miracle criterion is neither circular nor arbitrary.

We have seen how Locke's greater-miracle criterion can be a defeater for the competing miracles defeater, and now we have seen how Locke's view can be generalized to a greater-evidence criterion. Accordingly, Locke's response to the competing miracles objection may also provide a model for responding to religious disagreement more generally.

It is sometimes argued that religious disagreement undermines the justification for religious belief. It is assumed that there are equal and opposite reasons for believing different (mutually exclusive) religious traditions. Yet this is exactly what Locke denies: he insists that there is more evidence for Christianity than for other religious traditions. Let us assume, just for the sake of argument, that Locke is right about that. Still, objectors might persist, the evidence for competing religious traditions is an undercutting defeater that at least weakens the evidence for Christianity. Again, though, this is exactly what Locke denies. Since God would not allow there to be better evidence for a false claim to revelation than there is for genuine revelation, the fact that there is more evidence for Christianity than any competing religious tradition is a defeater for the religious disagreement defeater.

The upshot of Locke's position is that, if there is more evidence for Christianity (from miracles or otherwise) than for other religious traditions, then we can be confident that Christianity is correct. Religious disagreement (about miracles or more generally) is often thought to be an undercutting defeater for the evidence for Christianity. However, Locke argues that God will ensure that genuine revelation will be supported by the greater evidence, and this greater evidence then acts as an undercutting defeater for the defeater from religious disagreement.

In conclusion, Locke pursues an interesting and plausible strategy for replying to evidence for competing religious traditions. While he takes the argument from miracles to support belief in Christianity, the equal and opposite arguments from miracles for other religious traditions present a potential defeater for the argument from miracles for Christianity. As Locke recognizes, though, the competing miracles objection can fail in two ways. First, most miracles do not actually conflict with Christianity, and hence do not undermine the argument for Christianity. Second, even when there are competing miracles, Locke denies that those arguments are equally plausible; he insists that there is better evidence from miracles for Christianity. Finally, if he is right that there is greater evidence for Christianity, this can act as a defeater for the competing miracles defeater. The result, Locke thinks, is that competing miracles do not undermine or weaken the evidence from miracles for Christianity. So, notwithstanding religious diversity and conflicting evidence, we can confidently believe in Christianity.