(Note that this post is extremely tentative and exploratory; I have not framed the issues as well as I would like, and my sketch of the two approaches to divine action is not the best formulation they could receive. I'm simply jotting down some thoughts that I will expand more carefully in future posts)
As a person who takes the current scientific consensus very seriously in the way I understand the world, one of the most challenging issues I face in theological reflection is how to understand God's action in the world, not primarily his creating and conserving the world in existence but those 'special' acts we ordinarily call miracles. The problem is that the narrative of modern science-certain controversies over the implications of quantum mechanics notwithstanding-is one of finding ever more precise regularities in the goings-on of the natural world, which many scientists are tempted to summarize as laws which govern the behavior of all objects in the natural world. On one account of physical laws, called necessitarian, physical laws tell us what must happen in any given situation. Many scientists are probably intuitive necessitarians. If we accept this account, and if the necessary laws we discover do not leave room for events we would call miracles to occur, God would either have to suspend the order of nature to perform a miracle, or limit himself to working only through these laws once he has created and set the world in motion. Both conclusions are theologically unpalatable, the former because it would seem imprudent of God to create a world which he has to override in order to accomplish his purposes, the latter because the current inventory of natural laws does not allow for most events usually understood as miracles.
Some theologians look for 'loopholes' in the laws of nature, for example in the indeterminacy of quantum mechanics or the unpredictability of chaotic systems, that would allow God to 'break in' and influence events in one direction or another. All these proposals suffer from severe difficulties, however. To take just one example, the scope of quantum indeterminacy is so limited that for God to simply nudge an asteroid in a certain direction (perhaps away from Earth), he would have to tamper with quantum outcomes over the course of millions of years!
There are, however, at least two good options on the table for Christian thinkers who take both science and special divine action seriously. They are very different from each other, yet both seem to allow for a very robust theology of divine action while at the same time respecting the integrity of science.
The first takes its cue from the history of science. Time and time again we have seen laws which were originally assumed to be universally valid subsumed as special cases of more general laws, which apply under special conditions (usually called 'limit' conditions), or as approximations to more general laws which are 'valid' enough in those conditions. For example, Newton's laws of motion, once thought to be universally valid, are now seen merely as a 'good enough' approximation of the more general relativistic laws of motion, valid only when the objects being studied are moving slowly enough and are not too massive. Once the limit conditions are transcended, however, general relativity predicts (and experiments confirm) strange behavior never anticipated by Newton's laws, and in fact quite unintelligible within that framework.
By analogy, we can think of divine action, not in terms of God violating the laws of nature, but of his taking advantage of a limit condition, in which events occur that are not covered by our current understanding of the laws of nature, but which are still lawful according to the most general laws of nature, which by definition we have not discovered yet. Even the physicists' coveted 'theory of everything' would probably not reveal these most general laws, because if the world is God's creation, it has a depth which empirical science cannot penetrate, at least the empirical science which is the product of finite, fallen human beings. In this perspective, miracles would not be an anomaly but, as C.S. Lewis argued, manifestations of the deeper, truer order of nature, of which our ordinary, 'natural' experience is just a special case. In our ordinary experience water does not turn into wine. But that may be simply because in our ordinary experience the right conditions do not hold for this to happen.
The second approach does away with the idea of general laws of nature which explain the particular events we see occur. Rather, as Wolfhart Pannenberg has argued, God's action would consist primarily of particular, individual, contingent events (that is, events which did not necessarily have to happen), which only secondarily manifest the regularities that natural science studies. Chesterton probably has a view like this in mind when he suggests that certain events, like the rising and the setting of the sun, predictably occur, not because they obey a general law, but because God does the same thing over and over again. In his own delightful words:
A child kicks its legs rhythmically through excess, not absence, of life. Because children have abounding vitality, because they are in spirit fierce and free, therefore they want things repeated and unchanged. They always say, Do it again; and the grown-up person does it again until he is nearly dead. For grown-up people are not strong enough to exult in monotony. But perhaps God is strong enough... It is possible that God says every morning, Do it again, to the sun; and every evening, Do it again, to the moon. It may not be automatic necessity that makes all daisies alike: it may be that God makes every daisy separately, but has never got tired of making them. It may be that He has the eternal appetite of infancy; for we have sinned and grown old, and our Father is younger than we.In this understanding, natural laws are not so much laws as regularities: they do not specify what must happen, only what has been observed to happen with some reliability. But each event is actually unique and completely contingent, connected to what came before and to what will come after only through the faithful action of God, who reserves the right to introduce an event completely unlike all that have come before, such as the Resurrection. In this understanding, history is not subsumed under nature, but the other way around: history is primary, the history of God's particular, contingent acts, some of which exhibit the regularity that allows science to 'work'.
Pentecostal scholar Amos Yong has attempted a very interesting variant on this approach in dialogue with Charles Sanders Pearce, who understood the laws of nature, not as laws but more as 'habits' that nature acquires as it evolves unpredictably from one state to the next. These habits, once established, are not set in stone (this notion receives some support from recent work in theoretical physics; several scientists are now suggesting that the laws we formulate today may have evolved over time). Just like human beings can change their habits over time in response to a 'jolt' which sets them off in a new direction, so nature can acquire new habits in response to a supernatural jolt. We might think of Jesus' resurrection, for example, as the beginning of nature's forming habits more in line with God's ultimate purpose for Creation, when it will finally exhibit God's glory and life the way he originally intended.
Again, this is just a very rough sketch and much more should be said about these two approaches. In future posts I may interact with good statements of these views in more detail. For now I will just say that these two approaches seem to be the most promising for an understanding of special divine action that respects the integrity of science but also allows for genuine miracles, not just misunderstood 'natural' events (natural under the current understanding of the working of nature) or metaphors for spiritual truths.