The purpose of a cumulative case argument is to collect and categorize evidence (experiences, data, etc.) into respective piles and then see which pile is bigger, implying that one worldview is more probable than another.
The cumulative case for atheism says we would better expect things like evolution and evil and non-belief in God if God did not exist (and so the fact that these do exist favors atheism). The atheist may want to add religious diversity, the failure of prior supernatural explanations, and the advance of scientific progress in support of his position. After all, atheism appears to be a leaner theory and explains just as much as theism does, in which case—Occam’s Razor. Bam.
On the contrary, atheists are putting evidence—virtually all the evidence, in fact—into the wrong pile. They are too quick in their sorting. A more thorough investigation reverses most if not every point from favoring atheism to favoring theism. I have, for example, already argued how the most important superficial evidential datum for atheism, evil, is actually evidence for the existence of God upon more substantial analysis. If this is correct, then the strongest piece of evidence in the atheist’s cumulative case has been flipped in the other direction.
I am with Alvin Plantinga: whereas there may be apparent discord between theism and certain life experiences or scientific findings, there is, in fact, deep concord when each is properly considered. and the reverse is true of atheism: superficial harmony, serious conflict.
Unfortunately, because cumulative cases often include wide varieties of evidence, I cannot provide a comprehensive assessment in a single article, which means I must be content to suggest how reversals in favor of theism can be made on several points, and then link further resources for those wanting to engage each point more substantially. In other words, get ready for links.
This article can set the stage. Here I want to make two observations—one about the nature of cumulative case apologetics compared to metaphysical demonstration, the other about explanatory simplicity. These observations will be useful for rebutting what I take to be the argument for atheism, which is:
- If two theories explain just as much, then believe the simpler.
- Theism and atheism explain just as much; atheism is simpler.
- Believe atheism.
First of all, cumulative cases against God do not invalidate metaphysical demonstrations for God—if there are, in fact, metaphysical demonstrations for God. If we have a truly robust philosophical argument for God—which is to say, an argument that starts from premises that cannot be coherently denied (that there is a caused reality, that change occurs, etc.) and proceeds deductively to the conclusion of a purely actual being—then cumulative (evidential) arguments are inconsequential in comparison. (Examples of such arguments here, here, here, and here.) For even if one admits that evil counts as evidence against the existence of God, other things equal, if that person has in place a metaphysical demonstration for God’s existence, his conclusion should be that God must have a reason for allowing evil, even if it is not clear what that reason is.
Why? Because we take what is clearer to reason, and metaphysical demonstrations (which provide certainty), over cumulative case apologetics (which provide probability). To be fair, this could work the other way: if it were shown that God and evil are truly contradictory, then it wouldn’t matter what other evidence we claimed for God’s existence, like miracles. Contradictions are impossible, so until the logical case were dealt, with the cumulative argument would be moot.
Fortunately, the logical problem of evil has been dealt with, which is a point acknowledged among modern skeptics. Hence the move to more evidential arguments.
The atheist may deny that we have any such metaphysical demonstration for God, but that would be a different argument. The point is that cumulative cases are often ushered in on the assumption that God’s existence or non-existence is, or only could be, established by making a cumulative-evidential case in either direction.
However, traditional natural theology denies this and argues that God’s existence can be deductively—rather than inductively, or abductively—established. In this case, we should privilege the position that God is compatible with whatever experiences (or data) we encounter unless 1) the metaphysical arguments can be defeated and 2) the cumulative case favoring atheism is stronger than the cumulative case favoring theism. (Indeed, not only do we have successful metaphysical arguments for God, but also, the cumulative case for theism far overpowers the cumulative case for atheism.)
My second observation is that the claim that atheism is a simpler theory for not including God and other supernatural phenomena is not necessarily a plus for atheism. As philosopher Tim McGrew notes, the situation here isn’t as simple as comparing one theory of just A (say, material stuff) and another theory of A and B (God). Rather, what atheism says is A and NOT B (God does not exist), and there is no uncomplicated way to evaluate which theory is simpler when the atheist is not simply agnostic about B, but actively denies B. Thus, in response to Graham Oppy, who appeals to naturalistic simplicity, McGrew says, “There is no simple and straightforward ordering of these two positions with respect to simplicity. . . . It seems to me, therefore, that in claiming that metaphysical naturalism has greater simplicity—and therefore greater initial plausibility—than Christian theism, Oppy has given his own view an illicit head start” (p. 56 in Four Views on Christianity and Philosophy).
Part of this is because theories that appear simpler initially may not actually be simpler overall. Here is a theory about world history: Bill Clinton never existed. This theory is comparatively simple; it contains one fewer entity, one fewer Bill Clinton, than other theories. However, it is explanatorily impoverished; it would be forced to either leave a profuse amount of historical data unexplained or require ad hoc explanations (convoluted conspiracy theories, for example) that would themselves be more complicated (in the sense of having more ways they could go wrong, which is an explanatory cost, all things equal). In other words, having a leaner theory at the beginning (fewer existing entities) may not lead to a simpler theory (fewer ways something could go wrong) in the end, especially once that theory is strained to explain what it lacks the resources to explain.
Now, the atheist may retort that Bill Clinton is not a fundamental entity, and our considerations regarding simplicity must be focused there. But this puts the atheist in no better a position, because it is not clear that naturalism actually does posit fewer fundamental features than theism. This depends on the brand of naturalism at play; however, classical theism is hard to beat in this respect, because it posits only one (supremely simple) fundamental feature: God.
As Kenny Pearce explains:
According to the naturalist, there is only one kind of fundamental stuff, physical stuff. God, however, is not physical, so the theist has fundamental physical stuff plus God, which is extra complexity. This, however, is a mistake. According to the version of classical theism I’ve developed, the physical is grounded in God who is the only truly fundamental entity. So at the most fundamental level, this view not only has only one kind of thing, it has only one individual thing, namely, God. If the naturalist objects that I have introduced a new layer in the grounding hierarchy, this is quite correct, but when we introduce a more fundamental layer of reality that unifies and explains the layers to which we were already committed, this is not an increase in complexity. As a result, it is actually not clear that the naturalist has any advantage in simplicity at all.
Either way, it seems looking at theories in the abstract and counting how many entities they have is insufficient for determining which theory is more explanatorily virtuous (that is, able explain as much with less). To determine that, we must move into the theoretical details of the respective worldviews—that is, we have to see how well they actually explain things and with what resources. It is difficult if not impossible to see how the first claim—that atheism is a simpler theory—can be adjudicated without staging an investigation of the second claim—that atheism explains things just as well, or even better, than theism.
Once we get into the explanatory details, atheism will inevitably face one of the two difficulties mentioned above: leaving important aspects of experience unexplained (or having to explain important aspects away) or adopting far more convoluted explanations than theism to accommodate the relevant data, thereby losing any advantage as a simpler theory overall.
An example of explanatory inadequacy are atheistic theories positing brute facts—as Bertrand Russell once said of the universe, it’s . . . “just there, and that is all.” This leaves unexplained something which stands in need of explanation: namely, why any contingent thing exists instead of nothing. Or perhaps the theist will adopt “unparsimonious extravagances” (to lift a phrase from Richard Dawkins) such as the multiverse to explain cosmic fine-tuning. In either case, what might have started out comparatively leaner becomes either explanatorily inadequate or incredibly complex (and contrived), all of which is to say simplicity—although a consideration—is not the consideration.
So on the whole, it is far from clear that atheism enjoys any advantage in terms of being explanatorily more virtuous, since that is almost impossible to determine without getting into the theoretical weeds.
This article is the first in a three-part series. You can read part two here.