HOME ABOUT ME EXAMPLES OF WORK  FEES AND SERVICES LINKS CONTACT ME FEEDBACK

What would it mean for an event to be a ‘miracle’ in the sense that Hume describes it? Is Hume’s definition of ‘miracle’ accurate?

 

In this essay, I hope to show that some of the criticisms levelled against Hume, especially those by Robert Hambourger (1980), are not effective in their attempts to erode Hume’s argument. These issues were discussed in an article by Dorothy Coleman (1988), which I will use to outline the problems with Hambourger’s arguments.

Firstly, I will give an interpretation of Hume’s definition of a miracle, as discussed in his Enquiries Concerning Human the Understanding (Selby-Bigge ed, 1975).

Hume’s argument is not in principle against the possibility miracles, but rather against the possibility that it could be reasonable to believe that a miracle had ever occurred. Hume’s argument is trying to weigh up various probabilities against each other in order to determine the credibility of testimony for a miracle. In order to measure this credibility, the probability of the event must be weighed against the probability that the report is mistaken. An event’s probability is determined by the degree to which it conforms to known laws of nature, which are formulated on the basis of uniform experience. Assuming, as Hume does, that a miracle is an event which is a violation of the laws of nature, any judgment that an event is indeed a miracle presupposes a judgment that there exists a uniform pattern of causation to which it is opposed. This pattern would count as empirical evidence against the occurrence of the miracle. However, testimony is not so reliable, because it has been shown by experience to not be uniformly reliable in the same way that a law of nature is. Therefore, no report of a miracle will be credible or rational to believe, because there is always a uniform experience or law as empirical proof against it, which is obviously going to outweigh the merits of the testimony. As Hume remarks:

[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 endeavors to establish.... When anyone tells me, that he saw a dead man restored to life, I immediately consider with myself, whether it be more probable, that this person should either deceive or be deceived, or that the fact, which he relates, should really have happened .... If the falsehood of his testimony would be more miraculous, than the event which he relates: then, and not till then, can he pretend to command my belief or opinion (Hume, Enquiries, pp115-116).

 

This idea is criticised by Robert Hambourger (1980), who believes that there are counterexamples to this idea. Hambourger’s counterexamples involve lotteries, which he believes show that Hume’s argument is unsound. Hambourger believes that Hume’s idea relies upon an idea known as “the principle of relative likelihood”. He restates Hume’s ideas, basically saying what Hume did about weighing the probabilities of the event and the testimony against each other. If it were more likely that the people testifying are mistaken than it is that the event has occurred, then it would not be reasonable to believe that the event has occurred.

To counter this principle, Hambourger gives an example of someone winning a lottery. He says that there is a lottery in which there are one million entrants, each of whom has a one in a million chance of winning. A reliable newspaper such as the New York Times reports that someone called Smith has won the lottery. The Times only makes mistakes one in ten thousand times. Therefore, if we apply Hume’s principle as laid out by Hambourger, the conclusion would be that the Times report is mistaken, as this is the more likely of the two probabilities. But, as Hambourger says, this does not seem right. It seems more likely that the report is correct. He says that the credibility of the report should not be assessed in terms of its statistical probability, but rather its track record in giving correct reports. Hambourger therefore concludes that Hume’s principle is invalid, and so his argument against the possibility of having a rational belief in miracles is false.

To apply these considerations to miracles and their credibility, Hambourger says what we want to know to evaluate the testimony is the non-relative likelihood that the testimony is true. Just as in the case of the lotteries, we should not look at the probability that the testimony is true, but rather, how often are examples of that type of testimony true. So, for the case of miracles we should ask “How often are reports of miracles true?”. Hambourger does agree that this question is not easily answered in any way, but it is definitely not possible to answer it from Hume’s relative likelihood principle. This is because when you look at testimony you cannot “factor out the plausibility of what one reports, consider the remaining factors without it, and then, by weighing the two together, arrive at the probability that a report is true.” (Hambourger, 1980, p599).

The point of this argument “is to show that Hume falsely assumes that the predictability of an event must have a bearing on the credibility of a report of the event once it has occurred” (Coleman, 1988, p5). However, Coleman believes, and I am inclined to agree, that Hambourger’s criticism is not as damaging as he makes it out to be.

One reason for this is that Hambourger overlooks the two senses of the word probability. The first being the sense of probability of events pertaining to unique occurrences, and the second is in the sense of probability of events pertaining to the instances or tokens of event types. This, Coleman says, echoes the distinction made by Hume is his Treatise of Human Nature (Mossner ed., 1986) between “probability of chances” and “probability of causes” (Bk. 1, Pt. 111, Sect, XII and XIII).

Hume’s argument against the rationality of believing in miracles invokes the second sense, whereas Hambourger’s argument invokes the first. Following the first sense of probability, the likelihood of an event occurring is measured by its degree of predictability as a unique occurrence. But in the second sense it is measured by its degree of conformity to causal laws applicable to events of its type. It seems that an event that has low predictability may still be credible if it conforms to the relevant causal laws. Coleman (1988, p8) gives an example. She says that the credibility of Caesar’s crossing the Rubicon has little to do with its predictability. Clearly, the historical testimony relies upon the causal probability of the event it reports, i.e. the conformity of the events to causal laws pertaining to events of its type. The story of Caesar’s crossing is credible in part because it is not said to have violated any known natural laws, for example his crossing by walking on the water.

If it is granted that Hume’s idea invokes the second sense of probability, which I would agree with, then it can be seen why Hambourger’s lottery example does not prove to be a counter- example to Hume’s relative likelihood principle in the way it was assumed. The principle will only be invoked by reports of events that do not conform to laws pertaining to their event type. But this is not the case in the lottery example, because Smith winning the lottery is not an exception of the rules governing lotteries. Smith has a small chance of winning, but it is an equal chance to everyone else, and somebody will win the lottery, so no laws are broken. Therefore, when evaluating the Times’ report about Smith, we would not use the principle at all, but do as Hambourger himself says, and look at the reliability of the Times in terms of report of this nature, or look for further evidence to corroborate the story.

I will now look at Coleman’s (1988, pp9-10) lottery example, which she believes does invoke the relative likelihood principle. This example, as Coleman says, must be analogous to evaluating the credibility of a report of an improbable event in the sense that the event does not conform to causal laws pertaining to its event type, rather than in the sense that the statistical odd are against it. Furthermore, she says it must be analogous to our situation of not fully knowing the laws of nature, as in the only way we can know anything about nature is to form generalisations from recurring patterns within our experience. So, the lottery example must be one in which the participants of the lottery do not know the precise rules of it. An laws about the lottery, therefore, will be generated from observed uniformities throughout the length the game is being played.

The example is as follows: let us assume that there is a small country called Lottovania that has sponsored a lottery every month for the past one hundred years. It is known that the government of the country print up an indefinite number of tickets each month, but only distribute some of them. Each month a winning ticket is announced, but for all the hundred years it has been going, the winning ticket has always been one that is not distributed. Based on this, it can be generalised as a law that the winning ticket will always be one that is undistributed, since this is the only uniform observation we have on which to base a law. Now suppose that the newspaper known as the Lottovania Times prints a report that Smith of Lottovania has won the lottery this month. The paper is usually accurate, and only prints something wrong once every ten thousand times. In reporting this, however, the paper is reporting an event that is not just unlikely, but an event that is not conformable with past experience, and therefore not consistent with the laws pertaining to its event type. To determine the credibility of this report requires more than looking at the track record of the newspaper, it also requires looking whether the event is an instance of a regular pattern or type known to occur. While it may be possible that the lottery laws have changed, or that the rules were not what they were thought to be, it still seems that on the current evidence, the report does not conform with past experience, and so should be looked at skeptically. More evidence in favour of the event would need to be put forward to corroborate the story.

I will now look at how this principle can be applied to the evaluation of miracle reports. Hambourger believes that it may be rational and reasonable to infer that a miracle has occurred provided it meets two factors:

  1. The event is sufficiently well testified to warrant the belief that the event has occurred

  2. Without any plausible explanation, and,

  3. It is the type of event that would be appropriate for God to cause.

Hambourger does not give an argument to support these factors, but uses examples that he believes establish his case. I will look at the Hambourger example that Coleman (1988, p11) analyses, which is as follows:

“If all records and accounts from the beginning of the last century agreed that on January 1, 1800, in all parts of the earth, the clouds of each region began to spell out the Old Testament with perfect precision and in the language of the region, and that they continued to do so for several weeks until the new Testament was completed, then I think it would be hard to escape the conclusion that a miracle had occurred. Thus I believe that, at least in principle, testimony could give one adequate reason to believe in a miracle.” (Hambourger, 1980, p603).

 

It will be granted straight off for argument’s sake that what Hambourger says could have occurred and that the testimony would provide sufficient evidence to support a belief that it did occur. Now it comes down to deciding whether the aforementioned event, which is obviously anomalous, is indeed a miracle. This, as Coleman says, is more a question about the nature of the event rather than its occurrence. It comes down to whether it is more reasonable to believe that an event has a supernatural cause rather than a natural cause. It seems that of the factors Hambourger states, (1) and (3) on their own do not sufficiently give justification for the event being a miracle, as many ordinary events can meet these conditions. It is therefore up to (2) to separate the natural from the miraculous. However, I would agree with Coleman here that it does not seem that just because an event has no readily available natural explanation that it is indeed miraculous. For an event to be a miracle, it must be inexplicable in terms of what the laws of nature actually are, not just what they appear to be or we believe them to be. If the aforementioned event has no readily available natural explanation, then it could be that the laws of nature were not what they appeared to be. For example, someone could have learnt, without anyone else’s knowledge, how clouds really work, and used this to write the words of the bible into the cloud. Or it could be possible the event was of a different nature altogether. For example, the clouds were not really clouds but laser images produced by aliens. If we apply Hume’s law of relative likelihood, we must ask whether it is more likely (i.e. conformable) that the reporters of the events are mistaken than it is that the events are genuine miracles. It seems that the former is more likely to me in most cases, although there could be exceptions. However, counter-instances of what are taken to be natural laws are not by themselves enough to establish that no natural law could explain them. At most, Coleman (1988, p13) says, it provides some evidence for revising the laws or seeking a better understanding of the particular phenomena in question. And at the very least they simply provide grounds for suspending judgements about the nature of their cause until we have more evidence. Past evidence shows us that events, which were once considered violations of natural law, are now seen as parts of nature. Therefore, it seems that it is always more reasonable to believe that any claim about a miracle is mistaken than it is to believe that the miracle occurred.

It might be seen that this approach is a form of dogmatic objection of the supernatural in that it always puts the weight of evidence in favour of rejecting the claim that an event is a violation of natural law. It is claimed that if one assumes that God exists, then an event that has no natural explanation must be an appropriate event for God to cause. However, it seems that this is a mistake, because unless every event that has no natural explanation is seen as an appropriate act of God, there is no way to differentiate between the appropriate and non-appropriate ones. Also, it seems that the evidence we have for God’s existence makes it no more plausible than to say aliens or some other unknown force caused the events.

In conclusion, I believe I have shown that Coleman’s ideas about Hume are correct, and that Hambourger’s critique of Hume is not damaging in the slightest once it is properly analysed. Therefore I believe that if a miracle is defined as a violation of natural law, which seems like a sound definition, then it is never rational to believe that a miracle has occurred. This does not rule out miracles actually occurring however, just that belief in them can never be justified.

 

Bibliography

 

Coleman, Dorothy. (1988). Hume Studies Volume XIV, Number 2 (November 1988)328-346.

Hume, David (1986), A Treatise of Human Nature, ed. by Ernest C. Mossner. Viking.

 

Hume, David. (1975). Enquiries Concerning the Human Understanding and Concerning the Principles of Morals. Edited by L.A. Selby-Bigge (3rd. ed.) revised by P.H. Nidditch. Oxford: Clarendon Press.

 

Hambourger, Robert. (1980) "Belief in Miracles and Hume's Essay," Nous 14 (1980): 587-604.

 

©David Friesen 2004