This would be true if everyone insisted on the same criteria, but bear in mind that the insistence on sense datum as the exclusive basis for knowledge is a relatively recent set of criteria, and the persistence of religious belief would seem to indicate that it isn't yet universal. All these attempts to get religious believers to justify their belief in accordance with methodological naturalism is a bit misguided until you can get everyone to agree that methodological naturalism ought to be the exclusive standard for all knowledge.
This assumption doesn't break down when discoursing about immaterial phenomenon--we don't have knowledge of such phenomena because we have no evidence of such phenomena.
I'm fine with that conclusion, but I think you ought to recognize that it paves the way for other conclusions which are not so easily brokered. Descartes pointed to one of the dilemmas, which is that we ultimately have no evidence that our senses are reliable indicators of what is happening around us. The gist is that evidence, even if we limit its efficacy to epistemology, cannot lead us back to any solid conclusion on its own.
You seem to think either alternative is just as plausible as the other--naturalism is just as plausible as supernaturalism, and we can justify this on pragmatic grounds regarding the benefits we accrue from assuming one or the other.
That's the philosophical pragmatist's view of the situation. I've been reading some philosophical pragmatism lately, so that probably bled into my discussion.
Frankly, though, I'm a theist, so yes, I do think either alternative is just as plausible -- in very general terms.
And the atheistic conclusion seems warranted.
It's warranted in so much as the atheist assumes that lack of evidence is tantemount to lack of existence. I think that's a valid point of view, but an atheist who is honest with himself should recognize that such an assumption is something of a personal preference, and cannot be proven logically.
If we want to insist on only drawing conclusions that are supported by logic and evidence, then the non-atheist agnostic's conclusion is the most warranted, because he recognizes the limitations of his access to the knowledge that could decide the question, and reserves judgement in accordance with that limitation.
Clearly, the first hypothesis is superior to the second because it does not make needless and unjustified assumptions about the world.
It does make an assumption about the world, though. What makes the other assumption "needless" is your interpretation of what is necessary to explain phenomenon within the world. What makes it "unjustified" is the same thing which makes the atheist assumption of an exclusively naturalistic world unjustified as well -- we are in no epistemic position to say for sure, one way or the other.
An agnostic wants absolute proof for his beliefs, while an atheist is content with adequate justification, even though it is possible to be otherwise.
I don't think either characterization is implicit or explicit in the terms. Atheism itself doesn't provide a methodology -- the atheist could just as well believe there is no God because his cat told him so, and she would know. And agnosticism doesn't depend on absolute proof -- just on evidence furnished by the senses.
