1. Eshleman, Andrew ‘Can an atheist believe in God?’, Religious Studies, 41 (2005), 183–199.CrossRefGoogle Scholar All in-text references are to this paper.

2. The idea that pretending to believe in God could be a significant lifelong motive seems problematic. For a discussion see Robin LePoidevin Arguing for Atheism: An Introduction to the Philosophy of Religion (New York: Routledge, 1996), 113ff.

3. For a thorough classification of different types of religious anti-realism and a sustained criticism of religious anti-realism as a species of different types of anti-realism in general, see Peter Byrne God and Realism (Burlington VT: Ashgate, 2003). See also Christopher Insole The Realist Hope (Burlington VT: Ashgate, 2006). For a further critique of Wittgensteinian anti-realist approaches see Graham Oppy and Nick Trakakis ‘Religious language games’, in Andrew Moore and Michael Scott (eds) Realism and Religion (Burlington VT: Ashgate, 2007), 103–130.

4. J. S. Huxley Religion without Revelation (New York: Harper and Brothers, 1957).

5. The following discussion brackets the question of whether the reasons given successfully support atheism.

6. The position canvassed in the text is similar to how Feuerbach depicts religion at a certain stage in its historical development. My own view is that the logic of the problem of depicting God generates a strong pressure to depict humanity in an extravagantly negative manner. I do not believe that it is inevitable that any depiction of God's greatness must be accompanied by an extravagantly negative depiction of humanity – we are simply too creative to be forced into such a box. However, it is such a reasonable way of depicting God that it can be expected to be a highly dominant method over the long-term.

7. Tyler Burge ‘Belief de re’, The Journal of Philosophy, 74 (1977), 338–362.

8. Representative of the position of the realist believer here is William Alston who argues that anti-realism subverts the Christian faith. See his ‘Realism and the Christian faith’, International Journal for Philosophy of Religion, 38 (1995), 37–60.

9. See section 12 of Hume's Dialogues Concerning Natural Religion and essay 4 of Kant's Religion within the Bounds of Reason Alone.

10. This is related to a problem with other forms of anti-realism. While Eshleman denies the need for a religious practitioner to speak the truth, other anti-realists hold that practitioners speak truth but that truth is to be understood in terms of something like ideal rational acceptability (in various works, Putnam and Dummett have presented views like this). The problem is that standards of rational acceptability cannot in principle single out a true account from among those that conflict with it. It is a formal requirement on any theory of truth (if Boolean logic holds) that of a set of conflicting accounts, at most one can be true (they could all be false). In principle, any number of conflicting accounts could satisfy the standards of rational acceptability. Perhaps if the situation is ideal a single account could be determined – but that's because ideally our standards of rational acceptability should single out the true account, not because truth is ideal rational acceptability. That they help us find truth is a measure of just how ideal our standards are. Some anti-realists have this backwards: instead of recognizing that the ideal epistemic standards are to be defined in terms of truth, they define truth in terms of these ideals.

11. Truth here is taken in a fairly robust realist sense. For a reason why, see n.10 above.

12. I am grateful to many helpful comments from an anonymous reviewer for Religious Studies and from the Editor.