I don't know about that. In D&D, a lot of people see things like people disintegrating, or being consumed in holy (or unholy) fire, or wounds being healed with a touch and a prayer. So, the question is whether the mass of people could rationally trace the phenomena of clerical spells and divine visitations back to their divine source. It's possible in principle; the question is whether it is possible for most people.
To a great extent, the answer to this will depend upon the D&D campaign in question since the frequency, nature, and power of divine magic varies from campaign to campaign. In fact, even the source of divine magic varies from campaign to campaign, so, for instance, in a campaign where belief in the gods grants powers because it comes from commitment to the principles the gods represent (rather than inhering in the gods' persons), anyone who rationally traced the power back to its source should find that it is
not the gods. However, if one assumes that the power really does inhere in the gods (whether or not similar power inheres in abstract principles), it's possible to look at the question in a world like Greyhawk.
The first principle in extending the analysis is that reliable testimony of supernatural events has to be rationally possilbe. Of course, everyones' knowledge of the natural world relies upon testimony and, in a world like Greyhawk, the notion that it would require an impossible amount or kind of evidence to believe in a supernatural event but not in an unusual natural event is ludicrous. After all, which is more common in Greyhawk, a natural event like an unmarmed and unarmored man beating a Pyrohydra to death in twelve seconds or a supernatural event like someone miraculously healing after drinking a potion? Any standard of evidence that would allow a Greyhawker to believe in 20th level monks but not in cure light wounds is manifestly wrong.
Once it is acknowledged that characters in a D&D world can accept reliable testimony of supernatural events, the question is how much knowledge they have and what form it comes in.
Observations or reliable reports of spells would probably be the most common. Even spells like cure light wounds are fairly common and are likely sufficiently remarkable to be noticed and sufficiently unexplainable as to demand a non-natural explanation. Given DMG and Greyhawk module demographics, it's probably that reliable testimony of low level magic is well-nigh universally available and firsthand knowledge is fairly common. Reliable testimony of specific higher level magic is probably rarer but still relatively common--especially with the number of wars in Greyhawk that would group large numbers of people together and give them a chance to see some higher level magic. (Not that they are likely to see it as a "chance"--they're probably more interested in not seeing it unless it's on their side).
There are, however, two related obstacles to logically inferring the existance of spiritual beings from spells. The first is the weaker one and will vary in strength depending upon how DMs interpret divine verbal components. If they are interpreted as something like "May Heironeous smite you with fire from the heavens!" (Flame Strike) then they connect to a named entity other than the cleric. If, however, they are, like arcane verbal components, then a common observer has no way of determining that the power comes from beyond the cleric, druid, or paladin, rather than from within him as it would with a sorcerer, wizard, or bard. It's quite possible that educated individuals with Knowledge: Arcana or Spellcraft might recognize the divine spells as depending upon an outside influence (the rituals lack the necessary power to call fire or healing into being on their own), but again, it's not certain. In general, however, I think a that there's ample reason to think of divine verbal components as taking the form of prayers rather than arcane syllables and further reason to think that, even if they don't, educated individuals would be able to tell the difference between arcane and divine magic.
So, the weaker form of the objection might run like this: spells are connected to individual characters rather directly. In the example of Moses and the burning bush, there is no spellcaster, prophet, or intermediary and there is a voice and a burning bush to indicate the presence of an unobservable entity. For spell-evidence, however, the story of Moses and the Pharaoh's magicians may be a more appropriate analogue. Moses casts sticks to snakes. So do they. Moses casts transmute water to blood, so do they. Moses casts frog plague, and they polymorph a pebble into a frog. The logical conclusion for the Pharaoh when Moses casts his widened, extended insect plague and they can't follow suit is not "this must be the work of a god" but "Moses must be higher level than my magicians." (Curiously, his reaction in the biblical story is not very different from that). Spell evidence in a D&D world is strong evidence of the power of the individual but may well not be sufficient to prove the existence of the beings called upon.
In order to move from spell evidence to the existence of a specific spirit that would cause those spells, I think some further interaction is necessary. In order to conclude that a spirit with intellect and will was behind the spells, I think something more than a recognizable verbal component is needed. The phenomenon of ex-clerics and ex-paladins is probably not as widely known, but would provide evidence that something other than the cleric or paladin has some control over granting their powers. Spells like Commune and Divination which provide answers to questions would provde a some evidence for an entity behind the spell-evidence since they provide for (very limited) two-way communication. (Cross referencing with the similar contact other plane might also provide knowledgable people stronger evidence that those spells provide contact with a real entity). Large scale public events like an angelic messenger descending to gift the founders of a city with an olive tree (as Athena did in the myth of the founding of Athens) or a voice speaking in concert with an individual who had demonstrated spell-evidence (like the voice at Mt. Horeb) would also provide some evidence from which to infer. That, however, is rare, even in Greyhawk. (Though not unheard of; for instance, it's quite possible that credible testimony of St. Cuthbert's actions in the Temple of Elemental Evil would be available to loremasters of the Flanesse). Furthermore, the presence of non-deific extraplanar entities who nevertheless influence the world of Greyhawk (night hags being one example) would also make it more credible that there might be more powerful entities that are similarly unseen. Ironically, the rather irrefutable presence of Iuz on Greyhawk is probably some of the strongest evidence for other deities. (After all, he operates like them in nearly all manners except location and would thus provide evidence that, for at least one deity, there is definitely a person behind the spell evidence).
So, I think that most people in a Greyhawk like world would have to accept the existence of gods on authority--that of experts and rulers. The common folk would have spell evidence that was consistent with the beliefs they accept on authority but would not be sufficient to justify them, given the other information they probably would not have. The experts would have sufficient evidence for a reasonable belief in the gods. But that's how most everything works in the real world anyway. Most people IRL have to accept their information that way (and are happy to do so as well). We have ample evidence that fits with our worldview and renders it believable but if we want to justify everything from first principles, we'll find that a rather daunting task.
Umbran said:
The biggest flaw is that most folk (even in D&D worlds) are not presented with events quite as remarkable and unexplainable as unconsumed yet burning bushes that speak and transform sticks to snakes. His argument holds pretty well for Moses - his is not merely a case of confirmation bias, but holds poorly for anyone but Moses. His point simply doesn't scale to the world at large.