Saturday, January 17, 2015

Book Review: The Concept of Miracle by Richard Swinburne

In The Concept of Miracle, Richard Swinburne (RS) examines Humean arguments against miracles[i] and finds them wanting. His modest goal in this book is not to argue for the truth of miracles, but rather to defend the possibility of justified belief in the miraculous. He approaches this goal by rebutting anti-miraculous Humean arguments and discussing criteria which, if met, would ground justified belief in the occurrence of a miracle.

As with any work of Swinburne’s, even a short book like this one (a mere 71 pages) has plenty of ‘meat’ to be slowly digested, and plenty of links between earlier and later material. This presents a reviewer with a decision—to follow RS closely, or to restructure to make what seem to be the central points clearer. I’ve chosen the latter approach.

The review proceeds as follows. In the first section I present “The Humean Picture’. This consists of an operational definition of ‘miracle’, and the Humean arguments directed against specific portions of that definition. In the second section I present ‘The Swinburnean Picture’, which adopts the Humean definition of miracle and attempts to rebut, point-by-point, the Humean attacks on belief in the miraculous. RS also expands the discussion by arguing that, even if the Humean arguments were correct as originally presented, this is not the end of the story for the simple reason that there are additional sources of evidence for belief which Hume did not discuss. That will provide a segue to the third stage of the review, which I call ‘Background Knowledge’. In the fourth section, I summarize the book as concisely as possible. In the fifth and final section, I compare/contrast the contents of The Concept of Miracle with current philosophy of religion, focusing on Antony Flew and Richard Swinburne.

The Humean Picture

For Hume, a miracle is ‘a violation of a law of nature by a god’. On this definition, therefore, justifiable belief in miracles requires (a) justifiable belief in a violation of a law of nature and (b) justifiable belief that this violation can be attributed to a god. Both (a) and (b) have been challenged by Humeans.

Problems with justifiable belief in a violation of the laws of nature. In “Of Miracles”, Hume outlines four arguments against justifiably believing in testimonies[ii] to the miraculous. Three of these are dismissed by Swinburne on various grounds[iii] as beyond the scope of the book, so he focuses on what is, in essence, the argument from religious pluralism. That is, even if some miracle were well attested to, because miracles are usually situated in a specific religious context, they provide counter-evidence to other religious systems. (Put differently, Islam and Christianity assert contradictory propositions. A ‘Muslim’ miracle would, according to Hume, provide counter-evidence to the truth of Christian revelation—and vice versa.) To paraphrase a famous statement, ‘each is refuted by all’.

Another Humean argument, which I’ll term the presumption of natural law[iv], was put by Antony Flew. Flew argued that natural law and miracles possess different characteristics which should decisively tilt the scales against belief in the miraculous. Put baldly, the idea is that a miracle is a singular historical event which cannot be repeatedly tested by later persons—not so for a natural law.

Problems with attributing (purported) violations of laws of nature to a god.[v] Finally, Humeans pressed what I will call the presumption of causal closure[vi]. The idea here is this: suppose you see someone walk on water. How might you interpret this event? One inclined to belief in the miraculous might see this as a violation of the laws of nature. One disinclined to belief in the miraculous might argue for a refined understanding of the laws of nature. There is not a violation of the laws of nature, but rather there is something about this particular situation which should press us to more accurately describe the situation (without, obviously, involving supernatural entities). Once we’ve done so, we will see that this is a ‘natural’ occurrence after all.[vii]

The Swinburnean Picture

Although Swinburne defends the coherence of alternate definitions of miracle[viii], for the purposes of argumentation he accepts Hume’s definition of the miraculous. As always, Swinburne begins by providing important definitions and distinctions. In turn, he (i) analyzes what would constitute a violation of laws of nature (ii) analyzes what would constitute evidence for believing that some event E (a purported violation of laws of nature) has occurred and (iii) what would constitute evidence that the violation was due to the actions of a god.

I’ll briefly recap (i) through (iii) and then show how Swinburne applies those results to ‘The Humean Picture’.

What would constitute a violation of laws of nature?[ix] The key to understanding a violation of a law of nature, Swinburne thinks, is that a violation must be a non-repeatable event. One way of unpacking this is as follows: say we are an ‘ideal’ observer—that is, we are omniscient. We know all the laws of nature which apply at a given point in time. We can describe all the powers and properties which are possessed by all material objects in the universe. On January 1, 1970, we see Richard Swinburne walk on water. If all of the same laws of nature applied (and those are the only causally relevant factors) on January 2, 1970, would Richard Swinburne be capable of walking on water again? If so, then this is not a violation of the laws of nature. There must be some other natural causal factor at work, not previously measured or understood by the physical scientists of our day, which is the reason why Swinburne was capable of walking on water. Hence, any tendency to attribute this to ‘miracle’ merely reflects an epistemological gap—not an ontological one.[x] There is a law of nature which is at play and which is the reason that, in this particular circumstance, walking on water is possible.

The scientist, Swinburne opines, “seeks the simplest formula from which past results can be deduced”.[xi] So two key criteria here are simplicity and explanatory power (hence ‘can be deduced’).[xii] Understanding these criteria provide the key to describing situations which contain a violation of the laws of nature. First, we must be justified in believing that the purported law of nature is, in fact, a law of nature. Second, we must be justified in believing that some event E, a violation, occurred. For now, focus on the first. If some purported law of nature L is simple[xiii] and has vast explanatory power[xiv], then we are justified in believing it. If we attempt to ‘refine’ the laws of nature (so that L is now law of nature L*) so as to encompass E, but L* is vastly more complex (or suffers a vast reduction in explanatory power) compared to L, then we are justified in believing that L is a real law of nature.

Say, for example, we are examining claims that someone rose from the dead. Surely natural law L (people do not rise from the dead) is more justifiably believed as a law of nature as compared to natural law L* (people do not rise from the dead, unless they are first-century Palestinian Jews who were crucified). The question in such a case would not be “Is this a violation of a law of nature?”[xv] but, rather, “What grounds do we have for presuming that this event (the purported violation) occurred?” Put differently….

What would constitute evidence for believing that a violation occurred?[xvi] Humeans typically consider two sources of evidence for miracles (a) testimony of others and (b) our understanding of what is physically (im)possible (in other words, our current understanding of the laws of nature). Humeans argue further for the presumption of (b) over (a). I’ll turn later to Swinburne’s rebuttal of this presumption. For now, though, note that Swinburne’s view is that this is too truncated a vision of evidence—there are at least two other sources[xvii] of possible evidence, namely (c) physical traces and (d) our own memories. While Swinburne provides some examples to illustrate how these sources of evidence can be weighed against each other, the point for now is that the Humean picture is at best incomplete. Even if the presumption of (b) our understanding of the laws of nature over (a) testimony is granted (which Swinburne does not grant), this could be counterbalanced, in principle, by evidence from (c) physical traces and (d) our own memories.[xviii]

What would constitute evidence that the violation was due to the actions of a god?[xix] Here, Swinburne notes that we rightly invoke two types of explanation—inanimate or scientific (in terms of laws of nature) and the animate or personal (in terms of purposes or intentions). He easily shows that these two types of explanations are very different from one another, primarily by arguing that mental properties (purposes) exhibit privileged access[xx] whereas laws of nature and physical objects do not. The purposes we have are tied to the character we have. And purposes (while not reducible to our actions[xxi]) can be inferred from the actions we take. We can have evidence that some person S1 (an embodied agent, ie.., a human) is favorably inclined toward S2 (another human). S2 requests money from S1, and S2 receives the money. We thus can reasonably posit that S1 lent S2 the money—S1 (we can suppose) had the ability to lend the money, and the tendency to do so, and the lending occurred. Similar grounds could lead us to posit that some effect E was due to the action of some non-embodied agent S3 if we have other evidence for the existence of an entity like S3 (i.e., an entity with the capacity [‘could’] and the tendency [‘would’] to do a certain action.[xxii]

Problems with justifiable belief in a violation of the laws of nature.[xxiii] Recall that there were two Humean arguments presented earlier under this subheading: the argument from religious pluralism and the presumption of natural law. In responding to the former, Swinburne simply notes that it is overblown—many miracles may simply indicate the existence of a god exhibiting a fairly narrow concern for humanity, and not provide the kind of counterevidence Hume claims.[xxiv] In addressing the latter, Swinburne argues that—although the nature of animate explanation (purposes) involved in miracles and inanimate (laws of nature) are very different, they can be assessed by similar criteria (simplicity, explanatory power). Further, he argues that miracles can potentially be addressed in a similar way. Although physical traces can fade with time, we can get accruing evidence about the reliability of testimony under such and such circumstances. And he feels that Flew exaggerates the extent to which laws of nature can be tested—all such testing is always partial in that, like miracles (see footnote xxiv) experiments and observations provide evidence for specific propositions and not whole bodies of thought (whether religious systems or scientific theories).

Problems with attributing (purported) violations of laws of nature to a god. In providing criteria for attributing violations to the action of a god, Swinburne has described ways in which such attributions could justifiably be made. Indirectly, he also is making a point I’ll return to shortly—that the Humean is once again adopting a too-narrow-view of reality. By ignoring the point of personal explanation, they have failed to grasp how such attributions could be made.[xxv]

Background Knowledge

Up to this point in the book, Swinburne has been focused solely on the evidential force that miracles might have for theism. He notes, however, that this is too truncated a view. If we have prior reason to believe that God does (or does not) exist and would (or would not) ‘interfere’ with the Laws of Nature, then we have reason to be more (or less) inclined to accept whatever evidential force miracles have. (This point has already been indirectly indicated by a discussion of how evidence that there is a God who could and would do some action X is grounds for being ‘predisposed’ to accept evidence that X occurred.)

Philosophy of Religion-Then and Now

Many commentators more capable than I have commented at length upon the shift in philosophy of religion which has taken place in the last few decades.[xxvi] This observation is relevant to the current review, for the simple reason that it is easy to trace the evolution of thinking which has occurred since the 1970 publication of The Concept of Miracle. I make this point by first discussing one of the factors which helped ‘convert’ Antony Flew from atheism to deism, and show how this illustrates a sharp departure from his presumption of natural law cited earlier.

Natural Law vs. Miracles. Antony Flew lists three factors which led to his adoption of deism “Why do I believe this, given that I expounded and defended atheism for more than a half century? The short answer is this: this is the world picture, as I see it, that has emerged from modern science. Science spotlights three dimensions of nature that point to God. The first is the fact that nature obeys laws.”(emphasis added) [xxvii]

Now this is interesting. In the presumption of natural law, the regular behavior of physical objects was seen as a pattern which was evidence against the occurrence of miracles (and thus, indirectly, evidence against the existence of a supernatural agent, e.g., God). Here, natural laws are seen as evidence for theism. What is happening here?

Natural Law and Theism. Flew explicitly cites Richard Swinburne’s argument from temporal regularity[xxviii] on page 110 of There Is a God. His quotes Swinburne with “What is a law of nature? (This is not an issue faced by any of my critics.) To say that it is a law of nature that all bodies behave in a certain way (e.g., attract each other in accord with a certain formula) is, I suggest, just to say that each body of physical necessity behaves in that way (e.g., attracts each body in that way). And it is simpler to suppose that this uniformity arises from the action of one substance which causes them all to behave in the same way, rather than to suppose that all bodies behaving in the same uniform way is an ultimate brute fact.” Flew then goes on to state that “Swinburne’s central argument is that a personal God with the traditional properties best explains the operation of the laws of nature” and indicates his acceptance of Swinburne’s conclusion.

The Problem of Induction vs. The Laws of Nature. Flew also came to appreciate that there was an inconsistency in the thought of David Hume with regard to the laws of nature. Hume famously proposed that the regularities of behavior shown by physical bodies (stones falling to the ground, e.g.) were not representative of ‘real’ cause/effect relationships, but rather acts of the mind which ‘imposed’ laws on reality. However, “there is…no trace of the thesis that causal connections and necessities are nothing but false projections onto nature in the notorious section “Of Miracles” in the first Inquiry. Put another way, if the Problem of Induction is correct then causal connections and necessities (Laws of Nature) are ‘false projections’ which cannot possibly furnish grounds for rejecting miracles.


In providing criteria for understanding what constitutes a violation of natural law, grounds for believing that some purported violation occurred, and attributing that violation to the actions of a god, Swinburne has plausibly shown that (1) it is possible in principle for the evidence for miracles (even based on eyewitness testimony alone) to outweigh the ‘presumption of natural law’ (2) even if this was not so, there is no knockout punch for atheism here as other sources of evidence (memory, physical traces) could swamp the evidential force of natural law (3) arguably, natural laws are not evidence against miracles but evidence for theism, thus showing that Hume is ‘stealing’ from the theistic worldview and (4) the role that natural laws played in the early vs. later thought of Flew is an illuminating example of how much more theism-friendly philosophy of religion has become. The Concept of Miracle is highly recommended, then, on both its own merits as a stand-alone work and for its function as a before/after comparison of trends in philosophy of religion.

Apologetics 315 Book Reviewer Latter Day Inkling is a U.S.-based research psychologist for the military. He is especially interested in epistemology and natural theology.

[i] By ‘Humean’ I mean both arguments taken directly from David Hume and his successors (notably, Antony Flew).
[ii] The alert reader will realize that there are, potentially, other sources of evidence for belief in the miraculous. Swinburne raises that very point later on in the book.
[iii] The three are, briefly put, as follows—(1) such testimony has never been provided by prima facie trustworthy eyewitnesses (2) humans in general, and the religious in particular, have a tendency to exaggerate, gossip, and prevaricate about certain matters and (3) miraculous testimony occurs in ignorant times and places. While Swinburne acknowledges that these may have some force, he feels that Hume is engaged in hand-waving when a case-by-case judgment must be made. Further, Swinburne argues forcefully that for (3) to have any power, Hume must provide a non-circular definition of ‘ignorant’—he can’t simply define ‘ignorant’ to mean ‘believe in the miraculous’. Finally, Swinburne judges that these arguments are to be judged on grounds that are more historical than philosophical in nature. Hence his focus on the ‘argument from religious pluralism’.
[iv] Meant to play off of Flew’s (in)famous ‘presumption of atheism’ principle.
[v] Swinburne thinks that Hume does a much more thorough (albeit still incomplete) job of discussing the criteria for justifiable belief in violations of laws of nature than he (Hume) does of criteria for justifiable belief that such (purported) violations were due to the actions of a god. E.g., “On Hume’s definition, a miracle is not merely a violation of a law of nature but one affected by a god, but Hume does not discuss what further evidence would be needed to show the violation to have been effected by a god”. Discussion of how such attributions can be justifiably grounded in evidence provides one of the major ‘bridges’ between Swinburne’s arguments in this book and his later works.
[vi] I apologize for this potentially confusing terminology. I’m aware that this may be conflated with the argument from causal closure, but I really do see this as more of a presumption of causal closure than an argument. See Swinburne’s rebuttal of this presumption for why I think so.
[vii] Multiple problems with this proposal, I think. The first is that, as stated, it does not make clear what would distinguish this from a ‘naturalism of the gaps’. Swinburne does provide such a criterion, however-see the section discussing ‘repeatable’ results.
[viii] Chapter 1.
[ix] Chapter 3.
[x] Again, as noted in footnote vii, this smacks of ‘naturalism of the gaps’.
[xi] Page 23.
[xii] For more extensive discussion of such criteria, see Swinburne’s Epistemic Justification and Simplicity as Evidence of Truth.
[xiii] It posits few entities, few properties, few relations, all mathematically describable in simple ways.
[xiv] It unifies, explains, or leads us to expect many things which we do see—and does not lead us to expect things which we don’t see. Put differently, L ‘gels’ with reality.
[xv]“We know quite enough about how things behave to be reasonably certain that....these events are physically impossible." (page 32)
[xvi]Chapter 4.
[xvii]“What, one wonders, would Hume say, if he himself apparently saw a man walk on water? And Hume says nothing at all about traces, the kinds of evidence on which detectives...rely a great deal." (p. 35)
[xviii] “So I conclude that although standards for weighing evidence are not always clear, apparent memory, testimony, and traces could sometimes outweigh the evidence of physical impossibility.” (page 51) Swinburne also argues, in a premonition of his later ‘principle of credulity’, that we should always trust our own memory over the testimony of others for the simple reason that we have privileged access to the (apparent) truth of our own memories—we know *we* really believe such-and-such, but some other person could be lying about that fact. As Swinburne notes in Is There a God (page XX) “If we could not trust our senses when they seemed to show us what otherwise on a slight balance of probability we had reason to believe not to be so, we would always remain the prisoners of what we initially believed. If a slight balance of my evidence suggests that you are in London today (e.g. you told me yesterday that you would probably go to London today) and then I seem clearly to see you in Oxford, I ought to believe my senses, to believe that you are in Oxford today (and so believe that you changed your mind about going to London).”
[xix] See Chapter 5.
[xx] This is a central plank in his ‘argument from consciousness”. See his The Existence of God and Mind, Brain, and Free Will.
[xxi] Our actions or public behavior provide rational but defeasible evidence for our purposes or intentions. In this case the physical (our public behavior) is evidence for, but not constitutive of, the mental (our purposes).
[xxii] This is the way in which Swinburne constructs his cumulative case for believe in Christian revelation. See my review of Was Jesus God? at
[xxiii] As ‘The Swinburnean Picture’ section consists [largely] of point-by-point rebuttals to ‘The Humean Picture’ section, I re-use the ‘Problems with…’ subheadings.
[xxiv] Swinburne discusses a possible case in which miracles would provide counterevidence to specific religious doctrines. He does so as follows—say we have S1 (a Catholic priest) and S2 (a Protestant clergyman). S1 prays for a miracle to demonstrate transubstantiation and the miracle occurs. S2 prays for God, via miracle, to demonstrate the idolatry of transubstantiation by lightning from a cloudless sky striking the tabernacle holding the Sacrament. These would provide the kind of conflicting evidences Hume claims miracles provide. But “most alleged miracles, if they occurred as reported, would show at most the power of a god or gods and their concern for the needs of men, and litle more specific in the way of doctrine.” (p. 61)
[xxv] It should be noted that in The Concept of Miracle Swinburne has a somewhat unrefined view of what a Law of Nature is. He seems to be accepting the idea that a Law of Nature is an extra ‘substance’, above and beyond the physical constituents, in the universe which has causal power. He has subsequently come to reject that view. (See,%20or%20Divine%20Laws.pdf). In his most recent work Mind, Brain and Free Will he argues for a conception of natural law which makes substance dualism more plausible. (For more on this point, see Angus Menuge’s review of Mind, Brain, and Free Will at
[xxvii]Page 88 of There is a God: How the World’s Most Notorious Atheist Changed His Mind. [xxvii] See ‘The Datum of Temporal Order’ of Chapter 8 in The Existence of God.


J. Steve Miller said...

I'm amazed at how influential Hume's "Of Miracles" still is, especially when I look at the syllabi of those teaching introductory philosophy. They'll hit Plato, Aristotle and perhaps a few other Western philosophers, then discuss (very positively) "Of Miracles" and jump from there to Buddhist philosophy. Many seem to assume that any possibility of intellectually defending theism was destroyed by Hume. Moreover, we hear the echoes of Hume in so many of today's "New Atheists."

So this review of Swinburne is very relevant and needed.

I also often hear people say that if miracles happened in mutually exclusive religions, that this shuts down our ability to say that a verified miracle would point to God. Swinburne's response seems adequate to me. Why should people take it that any legitimate miracle is a confirmation that their entire theology is correct?

And I love the phrase, used in the endnotes: "naturalism of the gaps."

Good sum. Many thanks to the author!

Post a Comment

Thanks for taking the time to comment. By posting your comment you are agreeing to the comment policy.

Blog Archive