Christine Overall famously argued that miracles, conceived as violations of the laws of nature, would be evidence against the existence of the traditional God. A lengthy debate with Robert Larmer ensued, in which Larmer argued that only slight modifications to the law-breaking account of miracles are necessary in order for miracles to serve as evidence for, rather than against, the existence of God. Larmer tries to argue that miracles do not violate the laws of nature, but nevertheless holds that they are different from ordinary events in that they don't follow from the laws of nature. (I don't have Larmer's book handy to remember the exact details of his account.)
The Overall-Larmer debate in some respects replays one dialectical thread from the Leibniz-Clarke correspondence: Clarke defends the view that any sufficiently widespread natural regularity should be regarded as a law, and any event that violates such a regularity should be regarded as a miracle. Furthermore, Clarke argues, miracles of this sort occur from time to time. Leibniz argues that God, as traditionally conceived, would not create a world of the sort Clarke envisions and, furthermore, that Clarke's weak conception of laws does not allow a theologically adequate distinction between miracles and ordinary events.
I think Overall pretty decisively won the debate with Larmer, and Leibniz pretty decisively won the debate with Clarke on this and most other points. (One point where Leibniz clearly loses: his insistence that if there were not a unique best possible world God would be unable to create a world is clearly false.) However, there are a lot of people who seem to disagree, who continue to hold that miracles are best understood as somehow in tension with laws, and that such events can serve as evidence for the existence of the traditional God. I in fact think that miracles should not be conceived as in any sort of tension with laws, so, instead of speaking of miracles, I'll speak of 'lawless events'. Lawless events are those which don't follow, either probabilistically or deterministically, from the laws of nature. (interpret 'follow from' in whatever sense your favorite theory of laws requires.) In this post I am concerned with arguments from the traditional divine attributes against the occurrence of lawless events. These arguments will of course work backward to show that lawless events would be evidence against the existence of a being with those attributes.
The intuition behind this general line of argumentation is best stated by Leibniz:
Sir Isaac Newton and his followers have also a very odd opinion concerning the works of God. According to their doctrine, God Almighty wants to wind up his watch from time to time; otherwise it would cease to move. He had not, it seems, sufficient foresight to make it a perpetual motion. Nay, the machine of God's making is so imperfect according to these gentlemen that he is obliged to clean it now and then by an extraordinary concourse, and even to mend it as a clockmaker mends his work, who must consequently be so much the more unskilful a workman, as he is oftener obliged to mend his work and to set it right. (First Paper to Clarke, sect. 4)
I do not say the material world is a machine or watch that goes without God's interposition ... But I maintain it to be a watch that goes without wanting to be mended by him; otherwise we must say that God bethinks himself again. No, God has foreseen everything. He has provided a remedy for everything beforehand. There is in his works a harmony, a beauty, already pre-established.
This opinion does not exclude God's providence or his government of the world; on the contrary, it makes it perfect. A true providence of God requires a perfect foresight. But then it requires, moreover, not only that he should have foreseen everything but also that he should have provided for everything beforehand with proper remedies; otherwise he must want either wisdom to foresee things or power to provide against them ... According to this doctrine, God must want either power or good will. (Second Letter to Clarke, sects. 8-9)
Here, as in much of philosophy, the difficult question is to turn this vague intuition into a fully articulated argument. Here are two versions of the argument, one deductive and one inductive.
- A perfectly rational being who could create a world would choose a world which was optimally simple relative to the class of worlds which (a) are among his options, and (b) achieve all of his ends. A world is optimally simple relative to a class of worlds iff it achieves the best balance of (i) having few laws, (ii) having simple laws, and (iii) having few lawless events. (There may be more than one optimally simple world relative to a class, if there is a tie.)
- An omnipotent being's class of options for worlds to create would include every possible world.
- For any set E of rationally permissible ends, and world w in which those ends are achieved, there is a possible world w' in which all members of E are achieved at least as well as in w, and whose laws are as few and simple as those of w, and in which there are no lawless events.
- A being who was perfectly rational and omnipotent would create a world in which all of his ends were achieved without lawless events.
All 3 of the premises here are controversial, but I find them all plausible. Something like (1) is going to have to be true if predictability or consistency is partly constitutive of rationality, and it is. Furthermore, we could probably get the argument off the ground with some sort of doctrine of divine constancy. Nevertheless, (1) as stated might be strictly false, if the being's ends compete with each other in some way. (2) has been much debated in the literature, so I won't discuss it here. I'm not sure I endorse it. (I do believe that if God should will to create any world, he would create that world, but (2) doesn't follow from that claim.) It should be possible to weaken (2) and strengthen (3) to compensate, so that the argument still goes through. (3) seems to me to be on the shakiest ground. I think a defense could be mounted by a combination of modal intuitions about plenitude of worlds and the view that any end that, necessarily, could only be achieved through disorder would not be rationally permissible. However, I think modal intuitions tend to be shaky, so we perhaps shouldn't be too confident in (3). Anyway, we can mount an inductive version of the argument with much less controversial premises:
- A perfectly rational being who could create a world would, ceteris paribus, create a world in which there was as little disorder as possible.
- Lawless events would be instances of disorder.
- It is (subjectively) highly probable that, among all the worlds an omnipotent being could create, there are some which are just as good as the actual world in other respects and have no lawless events.
- On the hypothesis that the world was created by an omnipotent and perfectly rational being, it is highly probable that there are no lawless events.
Here (1) and (2) seem unassailable. (3) is perhaps somewhat controversial, but it at least matches my intuitive evaluation of the situation, and I'm sure I'm not the only one.
A skeptical theist response to this argument might be tempting. However, using this strategy to prevent miracles from being evidence against the existence of God will likewise prevent it from being evidence for the existence of God, since the result will be that the probability of lawless events on the hypothesis of theism is inscrutable. So it seems that lawless events are at least not evidence for the existence of the traditional God, and may even be evidence against.
Now, I don't think lawlessness is part of the ordinary religious believer's conception of a miracle in the first place. But even if it is, a rational reconstruction is in order. If the concept of a miracle is taken to include lawlessness, then miracles just can't play the roles religious thought takes them to.
[cross-posted at blog.kennypearce.net]