This is the latest post in a series about Bertrand Russell's essay Why I am not a Christian, in which I run through his various arguments and show that for the most part (there are a few exceptions) they either don't apply to classical theism, or are just completely invalid in the first place. But another striking thing about Russell's work are the arguments he omitted. In the last post, I gave a brief overview of some of these; in this post I discuss in detail what I consider the most important, the argument from miracles.

The purely philosophical arguments for the existence of God are useful, in that they provide us with most of the necessary intellectual framework in which to understand revelation. But they cannot take us all the way. Real-life religions are founded on revealed truth; that is particular messages from God that provide information about either God's nature, ourselves, or the relationship linking God with material living creatures such as ourselves. These revelations are what really founds each religion. Later on theologians combine what those revelations revealed with the philosophy that best fits them to create the various systems of doctrine and dogma which define each religion.

Revelation is seen by the devout as more secure than philosophy. The reason is straight-forward: philosophy is based on reason, the movement from premise to conclusion. The problem is: how do we know which premises are correct? Reason alone cannot guide us. Of course, we have the condition that the philosophy needs to be consistent with empirical science, which narrows the field of possible premises down, but (it is claimed) not enough to decide the big metaphysical questions. I would not wholly agree with this statement, since modern quantum physics is far more limiting than many philosophers and theologians realise, but it is still partially correct. Revelation, on the other hand, comes direct from the source. There is always some ambiguity in interpretation (but then there is also ambiguity and imprecision in experimental data), but this can be quantified and stated within sometimes quite narrow bounds.

But the problem is that how do we know that a revelation is truly from God, and not, for example, merely an epileptic hallucination , or some other such thing? The answer is that it needs to be associated with some dramatic sign that could not be explained by natural means. Composing a book, any book, no matter how great is not enough (because if we can speak some words then we can also think them for the first time). A dramatic healing, the dead coming back to life, suspension of gravity: we need something of that magnitude. Then, if the witness statements are credible and the context of the life associated with the event consistent with them and what we expect from a man or woman of God, we can believe the testimony that accompanies them. Events such as these, which authenticate and expand upon a revelatory testimony, are what we mean by miracles.

The analysis of whether any particular miracle occurs is not a matter of philosophy. That means looking at the sources, and coming to the best decision possible. That miracles have occurred , and continue to occur, is I think, undeniable given the overwhelming weight of evidence. But some philosophies claim that the miraculous is impossible, or at least so unlikely that we need not consider the possibility in practice. This, I think, is enough to show that those philosophies are untenable. Since atheistic philosophies all deny the possibility of miracles, the occurrence of miracles is thus a strong argument against atheism.

But I am getting ahead of myself. What do we mean by a miracle, and how do we know that something is miraculous? A common definition is that a miracle is a breaking of the laws of physics by some supernatural power. This was, for example, the definition used by Hume in his attack on the possibility of the miraculous, and it has been taken up by most atheists since then. I do not, however, believe that it is a particularly useful definition.

The crux comes in what do we mean by the "laws of physics". There are two definitions which I want to discuss here, which tend to be used by the majority of the atheist apologists who adopt this definition. The first is that they are empirically derived laws; the second some possibly mechanistic rule that is independent of God and obeyed by all matter.

The empirical rule was perhaps the first of these that came onto the historical scene. The simplest form of rule can be thought of as follows. Some experimental physicist measures something. He varies one parameter of a closed physical system, keeping everything else constant as well as he can, measures the response of some other observable, and plots on a graph the original variation against the response. For example, he might enclose a gas in a container, and measure the pressure against the temperature. Then, when he has enough measurements and data points, he will try to fit a mathematical function to the data. If he finds a simple relationship, then he will declare this a law of physics, and move onto the next experiment. Examples of this sort of law are the original statements of Gilbert's studies of magnetism, Boyle's studies of gases, Kepler's laws of planetary motion, Snell's law of refraction, and so on. It was this understanding of the laws of physics as a statement of empirical correlation that inspired Hume's philosophy, and it is what he meant when he formulated his definition of the miraculous.

The second understanding of the laws of physics started from mathematical premises, and used them to predict how matter behaved. This had its roots in the medieval mathematicians, but didn't really reach maturity until Galileo. It is what most people today understand by the laws of physics. Examples include Newton's laws of motion and gravity, Fermat's least time principle, or the Euler-Lagrange principle of least action. Generally these laws are used to calculate the empirically derived "laws", and if they do so well then they are vindicated and accepted. For example, Kepler's laws of motion are deduced from Newton's laws; Boyle's laws of Gases from the laws of Maxwell-Boltzmann statistical mechanics; Snell's law of refraction from Maxwell's equations of electromagnetism. And generally that is how classical physics proceeded, with experimentally derived "laws" deduced from and superseded by laws derived from more fundamental, metaphysical (albeit expressed mathematically) premises (for example, we could use the premise that Newton's laws are obeyed by all matter). And, I would say, it is this understanding of law which influences most of those who define a miracle as a breaking of the laws of physics today.

Now of these two approaches, the mathematical is to be preferred, since it explains the empirical law and can also be used to predict under what circumstances it will be valid and what circumstances it will break down. But neither will do as part of the definition of the miraculous. The reason is straight-forward: both of these understandings of the laws of physics had their origin in the late medieval period, and came to prominence during the Renaissance or later. Before then, people had a very different understanding of physics, one which did not depend on such concepts of laws. And yet these people also spoke of the miraculous. The classical, Biblical, early Christian and early Islamic writers must have had a different definition of what a miracle entailed. If we want to be consistent with these writers (always a good idea when conversing with them), then we should adopt a definition which might be expressed in a more modern language but nonetheless consistent with their own.

To these writers, a miracle was something that induced wonder, or a sign from God that displayed something about his character or power. One can't just separate the universe into that subject to physics which is independent of God and the miraculous which depends on God because to this mindset everything is caused by God, both the non-miraculous and the miraculous. It is why I define physics as the description of how God would sustain the universe were He wholly indifferent to it. There is no law independent of God, there is just God and His deeds.

So it is agreed that a miracle is something out of the ordinary; it is agreed that the universe is usually regular and a miracle is a break in that irregularity. But it is not an intervention by God because everything is God's work; that definition would not distinguish between the miraculous and non-miraculous. Instead we should regard it as a sign. But a sign of what? If nothing else, that God (or some other supernatural being with similar power) takes an interest in human affairs. Each miracle might well, and usually would, mean more than this; it cannot mean less than this. A miracle, in short, is an event that provides evidence that God (or whatever we want to call the foundation behind physics) is not wholly disinterested in human affairs.

We can construct physical theories on the basis of a small number of premises and assumptions. We can take some assumptions, and calculate what physics would resemble; with different assumptions we get a different picture. By comparing these assumptions against reality, we can narrow down to the correct picture. If we add to these premises the assumption that God is indifferent to mankind, then we obtain a theory of physics. If some observations deviate from that theory, then one of the premises behind the theory must be wrong, and we can in principle eliminate the alternatives and conclude that those deviations provide evidence that God is not indifferent to our species. These events thus provide evidence, sometimes strong evidence, for the existence and nature of God.

Quantum physics is indeterminate (which ultimately is a consequence of God's free will), but that does not mean that anything can happen. There are various conservation laws, the conservation of momentum, electric charge, and so on that must be obeyed by matter (if God is indifferent to mankind); equally there is the second law of thermodynamics which complex systems are also bound to. Each of these constraints ultimately arise from the principle of locality -- that interactions between particles (i.e. the creation and annihilation of material particles from other particles) occur at the same point in space-time. What happens in one place doesn't affect the probabilities of each possible outcome elsewhere. If God were indifferent to nature, that would be what we expect: being outside time and space, God relates to each point in the universe in the same way, and since he doesn't care about events at point A, that would not affect what he does at point B, or point C. Thus, if they are otherwise equal, the probabilities of a particular decay occurring at B or C would be the same, regardless of what is happening elsewhere in the universe.

But suppose that God were not indifferent to mankind; we would expect to see some over-reaching goal or purpose to the universe. Because God cares about the fate of a man at point A, he will act at B at an earlier time to make that goal realised. This breaks the principle of locality, and therefore momentum need not be conserved or the second law need not be satisfied if God so wills it. Thus examples where these principles do seem to be violated, such as when water turns into wine, or a man comes back from the dead, are evidence for God's caring about mankind and arranging events at a microscopic level to bring about those goals. It seems idiotic to me for someone to admit the possibility for God's existence, and yet deny that He can perform the miraculous.

But does the regularity of nature show that miracles are immensely improbable? Isn't a law of nature induced from numerous experiments, so to say that the law is violated to deny that conclusion? No. The law of nature found by induction from experiment is just one model to explain that data. There could be many other models which also explain that data. So let us say we have some data D, which everyone agrees on. There are two models consistent with this data, X which forbids miracles in any circumstances and model Y which permits them in certain circumstances. We then have the evidence E for an event which is claimed to be miraculous. Both models predict the experimental data D which was used in the initial analysis. How then can one use the experiments to judge between the models? That some experiments seem to apply a certain regularity in a particular circumstance is all well and good; but to link them to the particular circumstance of E depends on the model we use to perform the extrapolation. If we attempt to calculate the probability that the miracle actually occurred, this will be conditional on the both the evidence E, and our choice of model. Those who demand atheism will choose model X and conclude that the probability of the miracle occurring conditional upon the evidence and the model is vanishingly small no matter how good the evidence E. Agnostics would give models X and Y equal weight, and if the evidence is good enough conclude that it is quite probable that the miracle occurred (and consequently that model Y is to be preferred as it better explains the totality of the evidence). Both of these conclusions are correct: probability is, after all, always conditional and objectively computed from its premises.

We were asking the wrong question. The issue is not whether or not the miracle occurred, but which model is most consistent with all the data, both D and E. X which is consistent with atheism, is strongly contradicted by the evidence E. Y, which implies some form of theism, is consistent with all the data available. The obvious conclusion is that model Y and thus theism should be preferred.

The problem with arguments of this sort is that they get probability completely wrong, and fail to realise that the probability one computes for an event to have occurred depends on one's model and metaphysical assumptions, and there are different choices of model consistent with the original data. The only way in which one can say that the probability of the miracle is tiny is to demand that we use a model that denies the miraculous, or assert that there is no model that allows the miraculous consistent with our initial data, before examining the evidence. This approach is not only circular, but a wholly incorrect methodology. We should let the evidence judge our atheism (or theism, or whatever), not let our atheism colour how we view the evidence.

But if an event claimed to be a miracle occurs, doesn't that just prove that we got our understanding of physics wrong? There is no need to invoke God to explain the event. We just need a better physics. Isn't this so?

No. It might have been a reasonable argument in the pre-Newtonian days when physics was just about connecting empirical data, or even the period between Newton and Einstein. But not know, because the most fundamental physical theory is now directly computed from various philosophical premises (which enter the theory as various symmetry rules, the requirement of locality and cluster decomposition, and so on). We are thus able to test these premises directly. One such premise is that the power that underlies physics (whether it is God as traditionally conceived or something else) is indifferent to the fate of mankind. But whether or not that is true is the question we want to answer in the debate between theism and atheism. Miracles thus provide a direct channel to knowledge of God. Whether we call these events as part of physics or not is irrelevant to their role in providing evidence for God's nature. The question assumes that physics acts independently of God; theism is built on the idea that physics is a description of God's acts. This is the real question, and whether we define physics so that it includes miraculous acts or excludes them (I define it to exclude them) is an irrelevance.

But isn't belief in miracles just a relic of a superstitious age? Hasn't science disproved them?

No. Firstly, people in the past might not have known about the second law of thermodynamics, but they didn't need to to know that dead people don't usually come back to life. The amount of science needed to recognise a miracle can be gathered by just basic observation of the world. Secondly, miracle claims didn't only occur in the past or less educated communities, but are witnessed today and by experts. While not every claim is a genuine miracle, that there are some mistakes and frauds does not mean that they are all mistakes and frauds. Thirdly, science proves nothing of the sort. Physical theory is a description of how the universe would be sustained if God were indifferent to its fate. But, if so, then we cannot use it to claim that the source underlying Physics is indifferent to us. It is an assumption put into the framework used to develop that formulation of physical law; to subsequently argue against miracles because they violate that physical theory would be begging the question. The law simply was only designed to be applied in certain circumstances. To use it outside those circumstances is to draw more from it than it allows. The assumption of indifference is what is tested by claims of miracles.

For example, one can claim that evolution by natural selection is undirected (except in the sense that populations have a tendency to adapt forms which are most likely to pass on their genetics). But that is not a conclusion read from the theory, but an assumption put into it. Evolution is a combination of unpredictable mutations and natural selection. The mutations are only unpredictable to us because they are, for the theist, the direct result of God's free will. There is therefore nothing inconsistent in saying that mutation is both unpredictable (to us) and guided by God. The theory describes how life forms evolve under if certain circumstances are correct and if certain assumptions hold. To suggest that those circumstances and assumptions are universal is an assumption added to the theory by atheists. It does not follow from the experiment (since no experiment that fits the assumption of universality can be used to refute the denial of universality: it is consistent with both the assumption that the law is universal and that it isn't); and it is logically impossible to deduce the universal applicability of some particular physical theory from the theory itself.

But isn't it demeaning for God to have to keep intervening? Wouldn't it be greater if He just designed the universe to not need miraculous intervention?

Why would it be? This objection again assumes a set of physical laws independent of God; it states that wouldn't God have been greater if He could have designed those laws so that he didn't need to keep intervening. But, for the theist, this is based on an incorrect understanding of God's sustaining of the universe; it puts an intermediary between God and the movement of matter. To God, there is no difference between the usual evolution of matter and the special case. Instead, miracles showcase God's power and love for humanity; they show God to be greater than He would be if there were just the usual movement of matter.

But doesn't that the universe mostly seem as though God were indifferent to us show that God is wholly indifferent to us, and thus disprove the possibility of miracles?

Not for all forms of theism. Much of the universe evolves consistent with the assumption that God is indifferent to it because God genuinely indifferent to it. But what about this small pinprick in the galaxy we live on? Here the claim is that God largely leaves things to get on as though He did not care, not because He does not care, but because we demanded and continue to demand independence. God's intervention comes at a price: that we become good people. For almost all of us, that is too much; it means giving up all the pleasurable wrongdoing we are so addicted to. God seems to be indifferent because we demand that He keeps out of our lives, and buy and large God respects our freedom. Why then are there any miracles at all? Because, firstly, even though we reject God's help, we still need it, and He has to give us enough information to still understand that and His plan to rescue us. Secondly, there are a few not in rebellion, who desire a genuine goodness and are willing to rely on and trust God to provide it.

The question is what is the optimal balance between the regularity of nature and the miraculous. If men frequently came back from the dead, then the resurrection of Lazarus would lose its significance. If the principle that dead men stayed dead extended even to Lazarus, then we would also not be able to learn the lessons from that event. It is difficult for us to say that the frequency of miracles is as we would expect if Christian doctrine is correct; it is even harder for us to say that it isn't.

But doesn't quantum mechanics allow anything to happen? So aren't miraculous occurrences possible in an atheistic world-view, just incredibly unlikely. Yet the unlikely will happen on occasion if there are enough examples. So aren't miracles just the working of chance?

Firstly, there comes a point when the odds are so stacked against something (not just the occurrence of the miracle, but the coincidence of it coinciding with the prayer) that the probability (conditional on the assumption of atheism) of it happening is so close to zero that we may as well forget about the difference and seek alternative explanations. But secondly, and more importantly, quantum mechanics is not the most fundamental physical theory; quantum field theory is the fundamental theory, and transitions between states in field theory are bound by various conservation laws. It is not the case in quantum physics that anything can happen. Some things, even most things, are forbidden. If one of those principles are violated, then we know that quantum physics is only a partial, but not complete, description of the universe. Precisely as the theist has always claimed, and the atheist and deist always denied.

If one believes in an omnipotent God who actively sustains all aspects of the universe, then to deny the possibility of the miraculous is as insane as to deny the usual regularity of nature (and a belief in miracles is just as dependent on that usual regularity as a disbelief in them). The usual arguments against the miraculous, from Hume's to Spinoza's to those of the present day are built on premises drawn from an atheistic or deistic world-view. To refuse to consider the argument from miracles in an apologetic against theism believing those arguments to have been refuted therefore requires that we assume theism is false before looking at the evidence: a classic case of begging the question.