There is no presumption of atheism

In a recent post I excerpted some quotes from Robin Le Poidevin's recent book Agnosticism on the moral argument. In this post I want to share some more quotes from the book on the presumption of atheism. I wanted to title it, 'An atheist denies the presumption of atheism', but the more I read, the more convinced I am that he cannot possibly be an atheist. The book seems to be, not merely a description of agnosticism, but a strong positive argument in its favor. Therefore, unless Le Poidevin is merely writing 'speech in character', it seems that he is an agnostic. Nevertheless, he provides a very convincing rebuttal to the presumption of atheism, which is relevant to the discussion between Christians and atheists.

Le Poidevin begins by describing the presumption of atheism as follows:

Atheism doesn't require defense. Rather, it is up to theists to convince us that there is a God. Unless they can do so, we can remain comfortable in our disbelief. Only if they produce a really compelling argument in favor are we obliged to stir ourselves and show just where the argument fails. If there's room for doubt (and there's always room for doubt over arguments for God) then the rational thing is to be an atheist. (p.46)

He goes on to suggest, however, that atheism "is the default position only if some general principle that is clearly correct says that it is the default position." (p. 47) In other words, that atheism is the default position cannot necessarily be taken as the default position! It must be justified on the basis of a more general, sound epistemological principle. The rest of Le Poidevin's discussion consists of his analysis and critique of several such principles.

1) Perhaps the default position should be whatever common ground exists between two disputants. One way to argue for the presumption of atheism, then, might go as follows: both the theist and the atheist grant the existence of the physical world. Since atheism does not go beyond this common ground, whereas theism does (by postulating a non-physical, transcendent cause of the physical world), it should be taken as the default position, and the theist must provide arguments for going beyond it. Le Poidevin replies:

This line of thought might initially be tempting, but it is very definitely wrong-headed. It isn't just theism that goes beyond the common ground. Atheism does so too: it says that there exists nothing more in the world than what both theist and atheist could agree exists. So atheists are not excused, on these grounds, from providing reasons for their position. (p. 47)

An illustration from Abbot and Costello Meet Frankenstein is apt here: in an early scene the titular comedians are unloading Dracula's coffin. In a sequence repeated for comedic effect, whenever Chick (Abbott) goes into the other room, Wilbur (Costello) sees something strange near the coffin which convinces him that Dracula is really in there. And every time Chick is called back he insists that the coffin (still closed) is empty, because of course no such person as Dracula exists. Now the coffin itself is common ground to both of them, but it is not clear which presumption should be the correct one: that it is empty or that it is occupied!

2) Perhaps the default position should always be the negative position, the one that says such-and-such is not the case, and the positive position always has to be argued for. As Le Poidevin notes, however,

This would be a very dangerous principle to put into practice! Everyday life requires us to have countless positive beliefs about the world, some so obvious that we barely think about them: that we have bodies that allow us to move around, that whatever we see in front of us exists, that there are other people similarly situated with whom we can communicate, and so on. But the philosophical skeptic shows that even these beliefs can be challenged. The skeptic invites you to contemplate the following: despite appearances, your brain does not reside in a body at all, but is being kept artificially alive in the laboratory of some unhinged neuroscientist, who cunningly stimulates your brain in such a way that you have a continuous series of entirely illusory experiences...The idea is ludicrous, of course. But can you think of a reason that enables you to rule it out completely?...Now consider again the proposal that all positive beliefs have to be justified, and in the absence of a totally convincing justification, the negative belief is the default position. Since we cannot conclusively defeat the skeptic, we would have to concede that we should give up our belief that we have bodies that can move around, and so on. We don't want to do this, so we shouldn't accept the principle. (pp.47-48)

3) Most promisingly, perhaps the default position in any debate is the most intrinsically likely one. One might argue that theism is intrinsically less likely than atheism, and so should bear the burden of proof.

The issue here, of course, is how we understand 'intrinsic likelihood', or 'prior probability'. Le Poidevin defines the prior probability of a hypothesis as "how inclined we should be to believe it" (p.49) before we start evaluating the specific evidence for it, and suggests that we can measure its prior probability according to how much the hypothesis rules out: the more it rules out, the lower the prior probability. In other words, the more specific a hypothesis, the lower its prior probability. For example, if a fair deck of cards is randomly shuffled, the hypothesis that the cards will all be arranged in ascending value according to their suites is highly specific and hence highly unlikely: in fact there is only one arrangement out of about 10^68 (1 with 68 zeros after it) possible arrangements that satisfies this hypothesis! On the other hand, the hypothesis that the cards are merely arranged according to their suites, but not necessarily in a particular order, while still quite unlikely, is a bit more likely than the previous one, since it does not rule out as many possible arrangements.

So by this analysis, which has the greater prior probability, theism or atheism? This would depend on how specific each hypothesis is taken to be, in other words how many possible states of affairs each leaves out. Starting with theism, Le Poidevin notes that there may be no one answer to this question, since some versions of theism are more specific than others. He suggests, however, that "this is an area where one simply cannot be too specific. It would be absurd, for instance, to suppose that God wears a red bow tie with white spots. More seriously, although we may be relatively clear about the role God needs to play in the cosmos in order to be worthy of the title 'God', we would be wise to be fairly unspecific about the way in which those functions become realized." (p.51) He then suggests that we restrict our understanding of theism to the hypothesis that there exists a being who fills the following roles:

i) the ultimate and intentional cause of the universe's existence

ii) the ultimate source of love

iii) the ultimate source of moral knowledge

If theism is understood in this way, the debate between theists and atheists boils down to the following situation: theists affirm the existence of a being who fulfills all three roles, while atheists deny that roles (ii) and (iii) are fulfilled by the same being, and deny that role (i) is filled at all. For example, atheists might posit the reproductive drive as the ultimate source of love, and a biologically useful delusion as the ultimate source of moral 'knowledge' (it probably should not be called knowledge on this view, hence the quotation marks).

Le Poidevin then concludes,

Thought of in these terms, does theism have a much lower initial probability than atheism? It isn't at all clear that it does. Of course, once we start filling in the details, going beyond the basic role that God is supposed to play and describing how exactly he plays those roles, then the specificity of the hypothesis goes up, and the initial probability consequently goes down. And the same is true of atheism. Once we go beyond a denial that there is a being that plays the roles in question, and start to fill in the alternative explanations, then the initial probability goes down. It looks, then, as if theism and atheism start on pretty much the same footing. There should be no presumption of atheism, and indeed no presumption of theism either. The initial position should be an agnostic one, which means that theists and atheists share the burden of proof. (p.53)

A couple comments here. There seems to be a tension between Le Poidevin's conclusion here and his reply to the suggestion that all positive positions should bear the burden of proof. Recall that he explicitly rejected that position: some positive positions or hypotheses, including our having a body and being in contact with persons having minds similar to our own, do not have to be argued for in order to be accepted. The question is, which positive positions enjoy this prima facie acceptability (what Alvin Plantinga calls 'proper basicality')? Unless Le Poidevin provides an argument that neither theism or atheism qualify as properly basic, he should not conclude that both theists and atheists share the burden of proof. But he provides no such argument.

Second, it might be suggested that Le Poidevin's conclusion, while perhaps congenial to the theist because it defuses the presumption of atheism, nevertheless puts the Christian theist in the uncomfortable position of holding to a much more specific hypothesis than 'mere' theism as Le Poidevin describes it, which therefore has a very low prior probability. I have three comments by way of reply. First, this suggestion ignores the likelihood that, if a good case can be made for mere theism, the prior probability of Christianity given mere theism would be greatly increased. Take the deck of cards again: while the prior probability of an arrangement where all the cards are arranged in their suites and in ascending order of value is fantastically low (1 in about 10^68, to be exact), the prior probability of that arrangement is considerably higher, given that the cards are at least arranged according to their suites. Second, even if the prior probability of Christianity given mere theism remains low, a low prior probability can be nullified if the actual evidence is strong. For example, even though the unique arrangement of cards we have been discussing is fantastically unlikely, if I see the arrangement emerge with my own eyes when I am of sound mind, perhaps with the concurrence of several others in the vicinity, I should certainly believe the evidence of my senses. It would be absurd to deny that evidence on the basis of the negligible prior probability of the hypothesis. Finally, we should keep in mind the fact that judgments of specificity are not always as clear-cut as in the case of a pack of cards. The attempt to enumerate all possible versions of theism, in order to assign specificity to a certain subset of those versions, is probably a hopeless undertaking.