In The City of God, Book XXI, Chapter 8, St. Augustine quotes Marcus Varro, Of the Race of the Roman People:
There occurred a remarkable celestial portent; for Castor records that, in the brilliant star Venus, called Vesperugo by Plautus, and the lovely Hesperus by Homer, there occurred so strange a prodigy, that it changed its colour, size, form, course, which never appeared before nor since. Andrastus of Cyzicus, and Dion of Naples, famous mathematicians, said that this occurred in the reign of Ogyges.
The Bishop of Hippo comments:
So great an author as Varro would certainly not have called this a portent had it not seemed to be contrary to nature. For we say that all portents are contrary to nature; but they are not so. For how is that contrary to nature which happens by the will of God, since the will of so mighty a Creator is certainly the nature of each created thing? A portent, therefore, happens not contrary to nature, but contrary to what we know as nature. (Modern Library, p. 776, tr. Dods, emphasis added.)
Augustine's approach is thus epistemic. It is because of our ignorance of nature's real workings that we take as contrary to nature what in reality is not contrary to nature. The contrast is this:
ONTIC: Whether or not event M is a miracle does not depend on what any finite mind thinks, believes, opines, expects, takes to be the case, etc. There is a fact of the matter as to whether or not an event is a miracle; whether or not M is a miracle is not relative to us.
EPISTEMIC: There are no events contrary to nature; there are no "transgressions" (to use Hume's word) of laws of nature. M is a miracle only in the sense that it does not comport with our understanding of nature, does not fit our picture of nature, thwarts our expectations as to how nature will behave, etc. To a perfect understanding there would be no miracles.
So far, so good. But the epistemic approach to miracles has an untoward consequence noted by Antony Flew in his entry "Miracles" in The Encylopedia of Philosophy, vol. 5, 346-353. Miracles have an apologetic function: they can be cited as attesting to the reality of God or as supporting the credibility of a putative divine revelation. For example, the miraculous multiplication of loaves and fishes, if accepted as fact, serves the apologetic purpose of attesting to the divinity of Jesus. Only a divine being could do that, or change water into wine, or walk on water, or raise the dead, etc. But if these events are merely inexplicable to us at present, then we have no reason to take these events as having any special divine origin. If theism is true, everything other than God has a divine origin. But for miracles to have probative force in respect of specific theses such as the divinity of Jesus, they would have to be brought about by a special divine intervention. They would otherwise be no different than any other event.
Suppose you see a man walking on water, and suppose your seeing is veridical: the man really is walking on water. (And to preempt the unserious, we are talking about water in the liquid state.) That is not possible given the laws of nature as we understand them. The surface tension of water is not great enough to support a man's weight. But it may be that our understanding of the laws of nature is very incomplete. There may be special psychophysical laws, unknown to us, that allow certain human beings possessing great powers of concentration to affect by force of will alone the surface tension of water. Suppose that is so in the case of Jesus. Then there would be nothing ontically miraculous about his walking on water. If so, Jesus' walking on water would not give one reason to infer that he was divine. He might simply be a man with special powers.
For the miracles associated with Jesus to attest to his divinity they must be construed as genuine ontic miracles, as being special divine interventions that are contrary to nature. But if nature is whatever God wills, then there cannot be any ontic miracles. For nothing can act contrary to the will of God. The problem is similar to the problem we confronted before in connection with Hume. If laws are or entail exceptionless regularities, then there cannot be any miracles given that miracles are violations of, e.g., exceptions to, laws.
It looks as if Augustine's position faces a dilemma. Either we construe miracles epistemically or we construe them ontically. If we construe them epistemically, then in Flew's words, they "provide no good ground at all for believing that doctrines associated with these occurrences embody an authentic revelation of the transcendent." (348) But if we construe miracles ontically, then we face a version of the difficulty pointed out by Hume. Just as there cannot be exceptions to exceptionless regularities, there cannot be any occurrences contrary to nature given that what occurs in nature is willed by God.