I have written a number of pieces on the thought of atheist and notable libertarian George H. Smith (b. 1949)  with ranging philosophical subjects. Ever since Valentines Day of this year (when I fist bought one of his books), I have been fascinated with the writings and arguments Smith uses against theism and Christianity more particularly. In his volume, Atheism: The Case Against God (1979), he touches on multiple subjects throughout the philosophy of religion, epistemology, and metaphysics. In his later book, Why Atheism? (2000), the back cover I think adequately accounts for this point:
Why Atheism? tackles a wide range of subjects, some of which have never been thoroughly analyzed from an atheistic point of view. [ … ] He also analyzes a number of classical philosophical issues, such as the nature of knowledge and belief, concluding that modern atheism is largely an unintended consequence of the religious diversity brought about by the Protestant Reformation. 
Aside from the criticisms one might aim at the theses or arguments Smith presents against theism or Christianity, it is hard not to make the case that Smith’s work contains one of the most fruitful discussions on atheism that has ever been written from the pen of an atheist. Now, although I believe his understanding of atheism is misguided and taken from a false “historical” definition (as he calls it), his work is surely something to address. To first be clear on what I mean by this, Smith’s view as to the epistemic responsibility of atheism is as follows:
The atheist is not obligated to answer arbitrary assertions, unproven assumptions and sloppy generalizations concerning the nature and consequences of the atheistic position. Atheism is the absence of belief in a god, nothing more. If the theist wishes to draw monumental implications from this lack of belief, he must argue for his claims… It is the atheist who demands proof from the theist, not vice-versa. 
This is what’s known as the burden of proof (onus probandi). Smith suggests that this principle “falls on the person who affirms the truth of a proposition” . Thus, “[i]f the theist claims to know that God exists, then we have the cognitive right – indeed, the responsibility – to ask this person how he acquired this knowledge and why we should take him seriously” . As the argument goes, if the theist cannot adequately meet his burden of proof, then “atheism is left standing by default as the only rational alternative” .
Thus, the atheism of George Smith can be a sort of epistemological or metaphysical one, once the religious person gets caught up in the language and knowledge of what he claims as true.
Mortimer J. Adler in his interesting book How To Think About God (1980) wrote in respect to the “uniqueness” of the word “God” once people from multiple perspectives consider the definition of such a word (agnostics, pagans, atheists, theists, etc). For instance, “[e]veryone concerned with the question whether X exists must attach the same meaning to the word that names or designates X; and that meaning must be made as clear as possible” . This is the focus Smith has on the language of religious persons. For isntance, Smith (1979) in chapter 2 of his book first comments that “[k]nowing what one is talking about is of inestimable value in any dialogue,” so
the theist, before he sets out to explain why we should believe in god, must first explain what he means by the word “god.” What is the theist attempting to establish the existence of? How are we to identify him (or it)? 
The thrust of Smith’s argument is to say that the “theist must present an intelligible description of god [ … ] Nothing can qualify as evidence for the existence of a god unless we have some idea of what we are searching for. Even if it is demanded that the existence of god be accepted on faith, we still must know what it is that we are required to have faith in” . He explains his position in full:
When the atheist is told that God is unknowable, he may interpret this claim in one of two ways. He may suppose , first, that the theist has acquired knowledge of a being that, by his own admission, cannot possibly be known; or, second, he may assume that the theist simply does not know what he is talking about. If the atheist regards the second assumption as far more likely, if he suspects the theist is uttering nonsense, this is partially because the first assumption is appallingly convenient.
-  You can see my post An Analysis of George H. Smith and the Onus Probandi and George Smith and the Objection to Analogy.
-  George H. Smith, Why Atheism? (Promotheus Books: 2000) back cover.
-  George Smith, Atheism: The Case Against God (Promotheus Books: 1979) p. 27
-  Smith (2000), p. 31
-  Ibid.
-  Ibid.
-  Mortimer Adler, How To Think About God (Touchstone: 1980) pp. 50-51
-  Smith (1979), p. 29
-  Ibid., p. 30