When we begin to discuss the existence of a god, generally one of the first things that comes up, especially from theists, are arguments for the existence of a deity. Of the three classes, ontological, teleological, and cosmological, the latter two are the ones that are asserted as being "scientific" proofs. The arguments in and of themselves may be sound, but are they do not seem to be valid in terms of scientific hypotheses. But this may also play to the advantage of theists as well, as it seems to be the case that such things do not even fall into the realm of science, and therefore to attempt to falsify them scientifically is a categorical mistake.

A scientific hypothesis is an explanation in the form of an educated guess for a particular natural phenomenon. The hypothesis, then, inherently is testable given the nature of the phenomenon in question. Through inductive methods, the hypothesis is verified. Why it is certainly true that induction cannot prove anything beyond the shadow of a doubt, the conclusions are, nevertheless, quite certain. Another often cited criterion for valid scientific hypothesis is the hypothesis' falsifiability, popularized by Karl Popper. To be falsifiable, the hypothesis should have a means to show that the educated guess and its conclusions are false. Suppose one was to say, "All observed chicken eggs are white, therefore chicken eggs are white." This is a reasonable inductive conclusion based on the observations. To falsify the claim that "chicken eggs are white" one only needs to find an egg that is another color, such as a brown egg. This simple mode of deductive falsification helps buffer scientific explanations against unfalsifiable claims.

The recent fall of intelligent design was largely contingent upon the issue of falsifiability, and one can learn much from this. Intelligent design is but one of many sorts of teleological arguments. No one knows for sure how old such arguments are, as they date to antiquity. The modern intelligent design movement probably started with William Palley's teleological argument. Palley argued from analogy that design was an observable phenomenon based on the fact that if one were to find a watch in a field that one would not assume the watch to have showed up there by natural means. The opponents of design arguments, such as Hume, suggest that design may be a phenomenon, but one needs a basis to recognize it. Humans are the creator of watches, and therefore know a designed watch when one sees it, but humans, so it seems, are not the creators of the nature, but rather are bound by it. Hume suggests that detecting design from within the natural order is only possible from outside the order. The only way to suppose that some natural phenomenon has design, then, is to assign design to it arbitrarily.  The critics of intelligent design suggest that the proponents arbitrarily assign design to phenomenon, then use this arbitrary standard and prove it. This amounts to proving the assumptions, which is question begging.

Two recent books, The God Delusion by Richard Dawkins and Breaking the Spell by Daniel Dennett attempt to grapple with the idea of religion and scientific evaluation of it. Dawkins takes a bottom up approach. He attempt pits the quote "god-hypothesis" against natural selection and asks which one provides a better explanation for the existence of life. In his "Ultimate Boeing 747 Gambit", Dawkins attempts show that a something more akin to natural selection provides a better explanation for the existence of life. I personally think Dawkins treatment of the whole thing is sloppy, but that aside, even if an argument for a designer was successful in defeating natural selection as an explanation, it in and of itself is question begging, as we've already established. On the other hand, there is Dennett's top-down approach. Dennett attempts to reverse engineer religion, breaking it down into smaller pieces, suggesting that the complexity of religion evolved from simpler ideas. Dennett's approach is more grounded, in that he starts with an analogy from evolution. He suggests that religions are similar to how certain wild animals have been domesticated for human purposes. The same, Dennett suggests, holds true in religion in that seemingly paranormal phenomenon evolved into religion. The problem with this is that it precludes, rather than concludes, religions are superstition. Under this paradigm, revelation is not a possibility, as many theists would posit it is. Without taking this into account, the Dennett's enterprise amounts to question begging again.

The problem with inductively proving the existence of a god has to do with the preeminence of a perfect being. A perfect being would be an uncreated creator god, which by definition exists independently of created time and space and therefore, so it seems, exists tautologically. What seems to be the case, then, is if one allows for such a god to exist, then approaches from this point of view will always prove to be true. Similarly, one cannot falsify it. But at the same time one cannot assume religion to be false, because this approach denies the possibility its truth, unscientifically. For these reason, it seems that teleological and cosmological arguments fall somewhere outside the realm of scientific inquiry.