If you read a book by an alleged eyewitness to a miraculous event, should you believe that the event occurred? Hume’s argument against the plausibility of believing in miracles on the basis of testimony is probably the most famous contribution to the philosophical debate on this question. It is also the most hotly contested and debated. Some people think that Hume’s argument is eminently reasonable, a shining example of his insatiable common-sense approach to philosophical argument. Others are less persuaded, believing Hume’s argument to be either question-begging or an ‘abject’ failure.
I’ve been dimly aware of these debates for a number of years. And I used to think I had a reasonable grasp of what Hume had to say. Indeed, in my recent post on Arif Ahmed’s case against the resurrection, I noted that Ahmed’s first argument is, to all intents and purposes, a reformulation of Hume’s. Nevertheless, there is one book that has been resting on my shelves for a number of years that argues we need to pay much closer attention to what Hume actually said in his original discussion.
That book is Robert Fogelin’s A Defense of Hume on Miracles. It provides a close reading of Hume’s original text and, as the title suggests, a defence of Hume from two common misreadings. These are:
MR1: Hume thinks that no testimony could ever be sufficient to establish the occurrence of a miracle.
MR2: Hume thinks that an a priori argument suffices to make his case against miracle claims.
These misreadings stem from the failure to appreciate the intimate connection between the two parts of Hume’s original text. Part one deals with standards for evaluating testimony relating to miraculous events. Part two looks at specific historical examples of such testimony. Many critics (and supporters) of Hume treat the two parts as hermetically sealed wholes. And when read in this light it can indeed appear as if the first part presents a purely a priori argument against miracles and that this argument is based entirely on the notion that testimony could never establish the truth of a miracle.
But these really are misreadings. Hume openly admits that testimony could be sufficient to establish a miracle (he just thinks the standard of proof is pretty high); and he doesn’t limit his discussion to the a priori realm. Over the next two posts, I want to examine Hume’s argument, as filtered through the lens of Fogelin’s book, and consider whether it has continuing merit. I start today by outlining the key principles from part one of Hume’s discussion.
1. Hume’s Method for Evaluating Testimony
Let me start with a general interpretive caveat. The modern trend when evaluating arguments about miracles is to cast them into the formal language of probability theory, specifically the language of bayesianism. Thus, we get lots of talk about conditional probabilities, prior probabilities, background evidence and posterior probabilities. This is all well and good, and for anyone who is well-versed in this language, it is relatively easy to see how Hume’s argument can be recast in those terms. But it is important to bear in mind that Hume himself did not adopt this language. In fact, Bayes theorem was only formulated in response to Hume’s work. Instead, Hume writes in an intuitive and colloquial fashion about the nature of probability. In what follows, I stick with Hume’s original language and abjure the formal garb of probability theory.
With that caveat out of the way, I’ll proceed to discuss the main features of part one of Hume’s Of Miracles. The first, and most important, of these is Hume’s proposed method for evaluating testimonial claims. Hume starts with the general guiding principle that ‘a wise man proportions his belief to the evidence’. The question then becomes the relative weight that should be assigned to testimonial claims in favour of miracles.
As Fogelin notes, there are two methods mentioned in Hume’s text when it comes to assigning weight to testimonial evidence. The first is the ‘direct method’ of assessment, so-called because it has to do with the reliability of the witnesses who present the evidence:
We entertain suspicion concerning any matter of fact, when the witnesses contradict each other, when they are but few, or of a doubtful character; when they have an interest in what they affirm; when they deliver their testimony with hesitation, or on the contrary, with too violent asseverations. There are many other particulars of the same kind…
To rehash this, we can say that the direct test works like this (NB - this is taken, with some modifications, from Fogelin):
Direct Test: Testimonial evidence for some event X is generally more reliable (i.e. more likely to be true) if it exemplifies the following (non-exhaustive) list of properties:
(i) There are many witnesses, not few.(ii) The witnesses concur with one another rather than contradict one another.(iii) The witnesses are of unimpeachable, rather than of doubtful character.(iv) The witnesses are disinterested, not interested, parties.(v) The witnesses present their testimony in measured tones of confidence, rather than with hesitation or too violent asseveration.
This direct test is to be contrasted with another method for assessing testimony. This other method focuses less on the character of the witnesses and the formal properties of their testimony, and more on the actual content of their testimony. This is what Fogelin calls the ‘reverse test’. Hume himself describes it in the following manner:
The reason, why we place any credit in witnesses and historians, is not derived from any connexion, which we perceive a priori, between testimony and reality, but because we are accustomed to find a conformity between them. But when the fact attested is such a one as had seldom fallen under our observation, here is a contest of two opposite experiences, of which the one destroys the other, as far as its force goes, and the superior can only operate on the mind by the force which remains.
We have here Hume’s take on the role of prior probability when it comes to assessing the value of testimonial evidence. He is saying that the degree to which reliable testimonial evidence will raise the probability of an event’s occurrence is always to be determined relative to how improbable that event was, prior to the testimony in its favour. Again to rehash this in a slightly more formal style, we can say that the test works like this:
Reverse Test: The probability-raising potential of reliable testimonial evidence for X must be assessed relative to the prior probability of X. If X was highly improbable, then the effect of reliable testimony is proportionally diminished.
Hume’s point is that whenever we evaluate the testimony for a claim we must do so by factoring in both the direct and the reverse tests. We need to consider whether the evidence ‘passes’ both tests, where ‘passes’ is understood to mean something like ‘renders X more likely than not’. We can represent this testing process using a two-by-two matrix. Whenever testimony fails or passes both tests, we have a relatively straightforward situation. In the former case, we’ll assume that the event did not occur; in the latter case, we’ll assume that it did. The tricky cases arise whenever there is a clash between both tests. What do we do then?
2. Applying the Method to Miracle Claims
Answering that question is the key to Hume’s argument in the first part of Of Miracles. It is also key to the misreadings of his position. As we shall see, when we move on to discuss part two, Hume thinks that testimony for miracle claims typically fails both the direct and reverse tests. In this sense, dealing with testimonial evidence for miracles is relatively trivial. Nevertheless, in part one, he tries to imagine the scenario in which the testimonial evidence for a miracle passes the direct test but then rubs up against the the reverse test.
In dealing with such a scenario, Hume asks us to consider just how improbable the miraculous event really is (e.g. the raising of someone from the dead). This leads to one of the most quoted passages in the entire text:
A miracle is a violation of the laws of nature; and as a firm and unalterable experience has established these laws, the proof against a miracle, from the very nature of the fact, is as entire as any argument from experience can possibly be imagined…There must, therefore, be a uniform experience against every miraculous event, otherwise the event would not merit that appellation. And as a uniform experience amounts to a proof, there is here a direct and full proof, from the nature of the fact, against the existence of any miracle; nor can such a proof be destroyed, or the miracle rendered credible, but by an opposite proof, which is superior.
There is a lot in this short passage, and a lot that can be misunderstood. It is important that we unpack it carefully. First, note the general definition of a ‘miracle’ that Hume employs. He says that a miracle is something that is contrary to a law of nature and that there will tend to be uniform experience in favour of that law. For example, we have almost uniform testimony establishing that people, once dead, do not rise from the dead three days later. This means that it is extremely improbable that the alternative should happen. If somebody testified that a person did rise from the dead (or if even several people testified to that effect) it would not be sufficient to establish the truth of that miracle. The reverse test would overwhelm the direct test.
But this is not to say that it is impossible to establish the truth of a miracle claim. Some people think that Hume is begging the question in this passage by assuming that there is uniform experience against the existence of miracles. Admittedly, Hume doesn’t help himself much in this respect with his talk of “direct and full proofs”, but as Fogelin makes clear, Hume uses this language in an informal, colloquial sense. He is not appealing to ‘mathematical proof’ or logical certainty. He is merely appealing to the notion that it is so well established that no one assumes the contrary to be true in the course of their everyday lives. Furthermore, as the last sentence of the quoted passage makes clear, Hume thinks it would be possible to establish the truth of a miracle, but only by ‘an opposite proof, which is superior’. This leads him to his famous maxim:
[N]o testimony is sufficient to establish a miracle, unless the testimony be of such a kind, that its falsehood would be more miraculous, than the fact, which it endeavours to establish: and even in that case, there is a mutual destruction of arguments, and the superior only gives us an assurance suitable to that degree of force, which remains, after deducting the inferior.
This is the nub of Hume’s argument in this first part. It might benefit from being reformulated in the language of probability theory, but its gist is tolerably clear. He is not mounting a full argument against the plausibility of testimonial evidence for miracles. He is simply pointing out that the standard is exceptionally high. The direct test would have to completely overwhelm the reverse test for us to be in any way confident about the occurrence of a miracle, based solely on testimonial evidence. The probability that the witnesses are telling a lie (or otherwise misleading us) would have to be less than the probability that the miracle took place.
Okay, this brings us to the end this post. To briefly recap, Hume’s argument in part one of Of Miracles is relatively modest in its scope and effect. It is focused purely on methods for evaluating testimonial evidence and how those methods might apply to testimonial evidence in favour of miracles. He is not mounting an a priori argument against the occurrence of miracles; nor is he assuming that testimony could never be sufficient to establish the truth (probabilistically speaking) of a miracle. He is merely arguing that the standard of proof is exceptionally high.
I’m pretty sympathetic to what Hume has to say. I certainly concur that the standard of proof is high when it comes to accepting testimony in favour of miracles. But I think there are problems with the application of Hume’s evaluative method. I think many religious believers are inclined to disagree that the prior probability of miraculous events is exceptionally low. Indeed, I have read religious believers who think that miracles are a common everyday occurrence. It might be possible to test their convictions on this score. The best way to do this would be by pointing out how they reject most miracle claims. For instance, one could highlight the fact that the Catholic church rejects most miracle claims associated with holy sites and putative saints. Evaluating the rate of approval of such miracle claims would establish a fairly low prior probability for any particular miracle claim (albeit maybe not as low as the one Hume claims). But there is a good deal of argumentative work to do on that front before Hume’s method will become acceptable to the typical religious believer.
I’ll continue the analysis of Hume’s argument in the next post.