David Hume On Miracles

The following was an essay I turned in for class. There were footnotes, but those didn't copy over. The copying also made the Bayesian formula a little weird, but I did my best to make it look like a legit equation. 

David Hume has a two pronged argument against belief in miracles. The first is granting for the sake of argument that a full proof could be given for a miracle. Since the full proof for a miracle stands against a full proof against it, the miracle cannot be believed. The second is that in actuality, there has never been a full proof mounted for a miracle, so even more so miracle cannot be believed. 

Hume sets out as a principle of knowledge that one proportions his belief to evidence. When one comes to two competing propositions that needs to be judged, one proposition may tip the scales, so to speak, and the weightier proposition will be the more probable one. Now miracles are defined as a violation of the laws of nature. What establishes a law of nature is our constant experience, which constitute a full proof. Supposing for the sake of argument that a full proof could be given for a miracle, that is a violation of a law of nature, which then is to believed? As Hume seems to imagine a scale of sorts in which to weigh these two propositions, one side which has a full proof against a miracle and another side which has a full proof for a miracle, neither side tips the scale and thus neither side demands our beliefs. Therefore, since miracles are given to us by testimony, no testimony can properly establish the truth of a miracle since it cannot in principle overpower its contrary proof. 

Of course, Hume believes, no such full proof has ever been given for a miracle. He gives four problems that miracles claims have faced that prevented them from constituting full proofs. First, that the testimony of miracles comes from uneducated men, this counts against the veracity of their miracle claims. Learned men have much to lose when making suspicious claims, while unlearned men, who don’t have a high reputation anyways, do not. So it would be better that a miracle claim come from an educated man rather than an uneducated one, and no miracle claim has ever come from an educated man. 

Second, what is usual is more probable, and what is more usual is established by past observations. Since miracles are unusual, they are therefore not as probable and you are supposed to proportion your beliefs towards the more probable propositions, which would therefore always be against belief in miracle claims. To add to this, since there have been forgeries of miracles, we ought to be suspicious of miracle claims. 

Third, since miracle claims tend to originate among “ignorant and barbarous” people, this constitutes a strong presumption against their claims. These claims don't happen in Hume’s present day, and since he presumably believes his day and civilization is more advanced than where miracle claims originate, the difference of education and civility are what explains the differences of miracle claims frequencies. 

Fourth, contrary miracle claims discredit one another. Many religions make many miracle claims, so that when we put them on the scale of judgement, they destroy one another and so neither are believed. Since we treat miracle claims the same, that is having the same weight, they would not tip the scales in any of their favors, and so none should be believed. For these reasons, Hume declares that no testimony can establish a miracle and certainly not be justified in making it the foundation for a religion. 

Grounds for objections are fertile. First, it would be a mark of bad epistemology that one could not believe something true. His argument, to be clear, isn’t that miracles could not have occured, but that the testimony upon which they depend on for our belief is always insufficient. 

Hume does allow the possibility that miracles have happened. So his argument is epistemic. However, the purpose of a good epistemology is to properly know that which is knowable. If your epistemology doesn’t allow you know that which is knowable, you have a bad epistemology, and bad epistemology should be discarded. Since Hume’s epistemic argument doesn’t allow for belief of something true, it is a bad epistemology and should not be believed. 

Hume may respond by saying that such is a misconstrual of his project. Hume does allow for belief in miracles, but doesn’t allow it on the basis of testimony. He writes, “Our most Holy religion is founded on Faith, not reason;...And whoever is moved by Faith to assent to it, is conscious of a continued miracle in his own person…” So it may be that Hume allows belief in miracles on fideistic grounds, but not on rational ones, and particularly not on ones dependent on testimony. 

Second, it seems as though Hume begs the question against miracles. A law of nature is established by constant regularity. However, it can only be constantly regular if we already know that a break in that regularity hasn’t happened. And since miracles are a break in that regularity, a law of nature is already defined as something that doesn’t allow for the possibility of miracles. But to know that a break in that regularity hasn’t happened, that is, to know a miracle hasn’t happened, is to assume something we are trying to figure out, which is whether a miracle has indeed occurred. So Hume begs the question against the knowability of miracles. 

Hume may respond by saying that because the laws are established by regularity and there is no connection of necessity between the causes and effects of these events, as argued in earlier parts of the Enquiry, a law of nature isn’t supposed to be understood as some ironclad, involitle rule that precludes some future anomaly. These proofs may be fallible, and so do allow for such exceptions, which is why proofs for a contrary are possible. Contradictions aren’t real or actualized, but we only conceive them as such when weighing contradicting proofs and determining the truth between them. Because laws of nature is a term broader than objected as, it needn’t be understood as begging the question. 

Third, by no means is Hume’s definition of a miracle uncontroversial. Hume defines a miracle as a violation of nature. Hume does qualify his definition by adding that it is an interposition of the laws of nature by some invisible agent, which includes a deity. But there are few problems with this understanding. First, a miracle has been commonly understood to be a sign rather than a violation of the laws of nature, and second, agent interposition doesn’t always entail a miracle. In the first case, suppose you were walking in the park and you said, “God, if you’re real, give me a sign.” Suddenly, a flock of birds with twigs in their beaks come flying down and drop their twigs in front of you in such a way that it reads, “I am real.” Firstly, our use of the word “sign” signifies that we commonly understand that a miracle is a kind of sign. “Signs” and “miracle” would be synonyms in that sentence. Secondly, though you could explain the movement of the birds by their brain chemistry and the landing of the twigs by a careful examination of the physics of the drop and wind, you would still see that as an act of God even though you could explain every act in a physical way, and thus no violations would have occurred. So it seems as though a miracle needn’t be defined as a violation of the laws of nature. In the second case, in some anthropologies, every human act would count as a miracle. In some views of human nature, man is an invisible agent. He is an invisible soul that is in an extended body. So everything that body of the human soul does is a result of an invisible agent, and thus every human act counts as a miracle. But this is not how we conceive miracles, so the understanding of miracles is incorrect. 

Hume may respond by saying that granting a miracle is a sign, it is only a sign because it plays against the backdrop of the regularity of nature, so all signs from the divine must also be a violation of the laws of nature. If an event could be explained by naturalistic causes, then naturalism is the explanation for a cause, not supernatural causes, and definitely not a supernatural agent. He may also add that in the case of birds moving, those birds would not have moved by their own volition if God did not command them in some sense to do such, and as it is an agent interposition would count as miraculous. If they did do it by their own volition, then it may count as extraordinary, but not miraculous. And finally, in Hume’s view, there is no human substance, only a bundle of properties. But ones anthropology is also controversial and is outside the scope of this paper.

Fourth, because miracles presuppose the regular workings of nature, the regularities of nature cannot be counted against miracles. If miracles are supposed to be a violation of the laws of nature, and miracles are only identifiable by their violations of the laws of nature, then the laws of nature are presupposed. If a miracle presupposed or predicts regularity, then it cannot be counted against miracles anymore than any other theory which makes accurate predictions cannot be counted against that theory. Analogously, most of the space on this paper is white. You can only see these letters because they are different from the white background. If these letters were typed in white ink then there would be nothing to identify. The differences between the letters and the backdrop are part of what makes this intelligible and so cannot be counted against the letters. In the same way, miracles can only be identified against a regularity and so that regularity cannot be counted against the miraculous. 

Hume may respond by saying that once again, these aren’t supposed to be explicitly contradicting proofs, but probabilities. That black ink happens to be on a dominantly white background doesn’t count against black ink because we also regularly see black ink upon it, and we clearly know there is regularly an agent cause behind it. In other cases however, we do have to determine whether a miracle is more probable than not. While a miracle does require a regularity of the contrary behind it to be considered miraculous, it doesn’t say much about whether that claim is true or not. 

Fifth, that Hume would seem to want to limit this epistemic approach to the miraculous is suspect. Hume, as mentioned earlier, does distinguish between the extraordinary and the miraculous. But what exactly is the difference? Because Hume seems to contextualize the discussion with religion, it seems that what distinguishes an extraordinary claim from a miraculous one is a religious context. So for example, a man coming back from the dead by way of resuscitation is extraordinary, but a man coming back from the dead by way of God’s command is miraculous. But is this qualification justified? Perhaps not. For example, belief that stones could fall from the sky could be interpreted religiously and non religiously. So, we put a Humean scale to it, did a rock fall from the sky or not? Suppose a tribe experiences a rock falling from the sky and then began to worship the rock. Because there is a religious context surrounding it, Hume would have us believe that no such thing happened, precisely because of the religious context. The problem is that we now know that such events have happened, and so Hume would not have us believe an event because what seems to be an anti-religious prejudice, and would stifle scientific progress at the same time.

Hume may respond by saying that what differentiates the two is agent causation. That a rock falls from the sky may be anomalous and extraordinary, but when you attach a religious context to it, you are further positing an agent behind the effect. And that is an additional claim that needs justification, whereas rocks falling from the sky for unknown reasons may be open to natural and predictable causes that are regularly seen elsewhere. If you posit a divine agent, you stifle scientific investigation, whereas being open to other natural options do not. 

Sixth, it is not a strike against the proofs of miracles that testimony came from relatively uneducated men. Hume argues that it makes the claims of the miraculous suspect that it comes from ignorant men, barbarous and uncivilized people. But why would it? While it may be granted that they may be relatively ignorant according to our standards, is expertise of our level necessary to recognize when a man is dead or alive? How much of an expert does one need to be to know that at some time T1 Queen Elizabeth was alive, at T2 she was dead, and at T3 she was alive again? All that is required here is for someone to know the difference between a dead person and an alive one. So that miracle claims come from the relatively uneducated shouldn’t count as a strike against the claims since expertise isn’t always needed, just common sense which everyone has. 

Hume may respond by saying that what is in question is their ability to know the difference between a dead person and a live one, but whether they can detect a hoaxed death as opposed to real one. It is more likely that they be deceived than to truly see things as they are. Add to this fact that the barbarous and uncivilized tend to like miracle claims, it throws further suspicion on their claim.

Seventh, that miracle claims tend to contradict each other doesn’t mean they cancel one another out. It is true that those who believe in miracles don’t believe in all miracle claims. But do they do so justly? If they believe that evidence for some miracle A is strong while evidence for some miracle B is weak, they would be justified in believing in one over the other. Not all miracle claims have the same amount and the same kind of evidence behind them. 

Hume may respond by saying that while that may be true, it’s a more detailed analysis of what he is lying down. His first argument is more in principle which can fit whatever facts and evidence you can muster, while the second argument is more of a sketch of all miracle claims in general, and so getting into the details of this miracle or that one is really not his point. He is only putting down a methodology of why no miracle claims can be established by testimony, not whether one miracle claim is worse than another. 

Eighth, and perhaps most technically, Hume may only be considering prior probabilities which is only one part of determining the entire probability of some hypothesis. In applying Bayesian probability theory to epistemic claims, we are able to quantify how likely some belief is. As put forth in John Earman’s book Hume’s Abject Failure, the probability Pr of some hypothesis H is formulated by some evidence E and acquired knowledge K. It looks like Pr(H/E&K). So how does one determine the probability? The first step is to consider the probability of the hypothesis against the acquired knowledge. So Pr(H/K) is called the prior probability. Hume pounds this particular point hard. He insists that the hypothesis of some miracle, which would be H, has to be determined by the acquired knowledge of our uniform experience of the laws of nature acting only a certain way, which would be K. And Hume correctly notes that this prior probability is going to be very low. This is fine as far as it goes, but unfortunately for Hume, this is only as far as it goes, and we need to go further. The next step is to multiply that against the probability of the evidence with the acquired knowledge and the hypothesis, which is Pr(E/K&H). This is called the posterior probability, or it’s explanatory power. These two factors, the prior and posterior probability constitute the numerator of the Bayesian formula. The denominator is constituted by the numerator plus the probability of the negation of the miracle hypothesis against the acquired knowledge Pr(￢H/K), multiplied by the probability of the evidence against the negation of the hypothesis and the acquired knowledge Pr(E/￢H&K), which would be the explanatory power of the naturalistic explanations of the hypothesis on the evidence given. So finally, it looks like this: 

Pr(H/E&K)=

Pr(H/K) x Pr(E/K&H)

_____________________________________________

[Pr(H/K) x Pr(E/K&H)] + [Pr(￢H/K) x Pr(E/￢H&K)]

So we see then that the only factor that Hume wishes us to consider is the first part of the numerator Pr(H/K), which is the prior probability of a miracle occuring. But the ultimate probability of a miracle hypothesis on the evidence and acquired knowledge Pr(H/E&K) could still end up being fairly high even though the initial probability of the hypothesis on acquired knowledge Pr(H/K) starts off fairly low. So granting everything that Hume says, it may still be justified to believe in miracles on the basis of testimony if you consider the evidence seriously enough.

Hume may reply by saying that attributing concrete numbers to knowledge is a category error. Because we are trying to quantify historical events, and historical events are what testimony tries to establish, there would be a category error here as well. In any case, the probability of a past event happening has to be either 1 or 0 since it either happened or it didn’t. It’s not like it kind of happened so we give it a .5.