[Note: This post is an excerpt from the draft of a book I hope to have published sometime next year.]
Naturalism has been described as belief that the universe is a self-contained system, consisting of strictly natural, material or physical phenomena, constituting all of reality that is knowable in principle. As C.S. Lewis remarked, naturalism means nature is "the whole show": there are no agents external to the natural system (or if there are such agents, they are incapable of interacting with or influencing the system). This belief is commonly said to enjoy two major strengths relative to Christian theism: (1) It is more consistent with observable evidence, since we at least know that nature exists; and (2) in keeping with the principle of Occam's Razor, it is more parsimonious (it contains fewer explanatory elements), since we do not know with any certainty that any entity outside of nature exists.
Those certainly sound like reasonable assumptions at first blush. But they are far from self-evident. For example, naturalism entails much more than the modest and wholly uncontroversial claim that "nature exists," but claims further that nature is all that exists (or at least all that can be known). To say that nature exists is no more evidence for naturalism than it is for Christian theism, since both explicitly posit the existence of nature. It's not saying much, really: When you get right down to it, most any serious claim about anything must take into account that the observable universe exists. (There are exceptions, like ontological idealism and solipsism, but these are generally taken seriously by neither Christians nor atheists.) Similarly, to say that naturalism is necessarily more parsimonious than Christianity asserts much more than can be demonstrated. For many observers (like me) naturalistic explanations for the origin of the universe, of life on earth, or of morality, appear to be exceedingly and even hopelessly complicated. Since these complicated explanations have been devised specifically as alternatives to more basic theistic explanations, there is no need for theism to invoke them and therefore it does not follow that theism entails more entities than naturalism.
But my rejection of naturalism extends beyond recognition that its strong claims cannot be verified. In addition I find naturalism less than fully coherent. I say this not because I believe naturalists are irrational people, but because the ambitious claims made for naturalism simply do not appear to square with nature itself. Naturalism stipulates that the universe (nature) is somehow self-existent, and perhaps infinite. But two of the most thoroughly verified fundamental properties of the universe, energy conservation and entropy, indicate that the universe is both finite and in need of an external power source. Or as the philosophers would say, the natural order is not necessary but contingent. This curious philosophical situation is made worse in light of the frequent appeals of naturalists to science and the scientific method. For if nature is not caused, how can it be amenable to the scientific method even in principle? Science is about physical (primarily causal) explanations, not metaphysical assertions. To say that the universe "just is," or "always has been," is to repudiate, or at least temporarily abandon, the scientific method, and thereby undercut the very foundation of naturalism. Christian theology of course not only permits but demands the repudiation of the scientific method regarding decidedly supernatural claims like the creation of the universe. But as a science-driven belief system naturalism can make no such allowances. On the issue of cosmology, then, Christianity appears more internally consistent than naturalism.
Another objection to naturalism arises from consideration of rational thought itself. The "argument from reason," famously explicated by C.S. Lewis in Miracles and revised by contemporary philosophers like Victor Reppert, basically holds that non-rational physical or material causes cannot be expected to create rationality. In principle, a thoroughly natural system would not be capable of producing rational thought, capable in turn of reliably determining whether or not naturalism is true. Or as Plantinga suggests, "the probability of our cognitive faculties being reliable, given naturalism and evolution, is low." This may not make naturalism completely self-contradictory, but it certainly provides fodder for skepticism: If our thoughts have been produced by mindless mechanisms of evolution, then our thoughts cannot at the same time be a product of reason and reflection. Besides, given evolutionary naturalism our brains—hence our thoughts—are still evolving at this very moment. Therefore the very principles of logic we consider rational and true today could be considered crazy and false tomorrow, as evolution dictates. Alternatively, what we believe to be logical or veridical might be false right now. Natural selection couldn't care less either way: If believing what is false confers short-term reproductive advantages upon our species, so be it.
A similar problem holds for the origin of morality: That is, non-moral physical or material causes cannot be expected to create morality, or at least not the sort of universally applicable objective morality that acknowledges love good and sadism bad for all people at all times. As a young man I recognized the truth of this. While a freshman in high school I recall witnessing a couple of bullies overpowering a smallish self-identified Jewish kid and literally shoving him into a trash can, laughing as he screamed in protest. My first thought (and one of my first serious reflections on such matters) was that if there were a God he would not have allowed such evils; but a second thought immediately followed: What exactly is evil if there is no God? My answer was, and remains, that "evil" conveys little meaning unless there is some ultimate, transcendent moral authority. In more sophisticated forms the same question confronts naturalists to this day.
Now it is just possible that rational thought and morality are somehow emergent properties of a wholly naturalistic system, but that seems intuitively implausible. In any event there appears to be no evidence for such a proposition, nor any way to test it. For all its accomplishments and merits, the scientific method cannot presently verify the hypothetical past emergence of rationality and morality from non-rational and amoral physical matter (whether by evolution or by divine creation from the "dust of the earth"). Per the epistemology of naturalism itself—that only scientifically rigorous beliefs are justified—naturalism is therefore an unjustified belief. That situation would perhaps be more epistemically tolerable if naturalism were also a properly basic belief, since even for a naturalist properly basic beliefs cannot be justified by external evidence. But no one thinks the truth of naturalism is strictly self-evident or incorrigible, and consequently naturalism is not properly basic. It is therefore a belief simpliciter, an article of faith. It turns out that naturalists have no leg to stand on when deriding theists for believing without sufficient evidence.