^1. Peter van Inwagen ‘What is the problem of divine hiddenness?’, in Daniel Howard-Snyder and Paul K. Moser (eds) Divine Hiddenness: New Essays (New York NY: Cambridge University Press, 2002), 30.

^2. Ibid., 31.

^3. The period and degree of the doubt can vary among individuals.

^4. There are many uses to which the argument could be put, but it seems to us that Schellenberg's own aim is to argue that the agnostic possesses sufficient evidence in virtue of being a rational agnostic to warrant atheism. That is, once the agnostic – who takes the state of the evidence to be roughly a wash – realizes that God would not allow such an evidential state to obtain, she thereby acquires new evidence against God's existence which is sufficient to tip the scales in favour of atheism. We think this is an ingenious strategy even though we think the argument ultimately fails.

^5. J. L. Schellenberg Divine Hiddenness and Human Reason (Ithaca NY: Cornell University Press, 1993), 83. Schellenberg notes that this is just one formulation of the problem of hiddenness. In a recent article Schellenberg presents the analogical argument from hiddenness. See ‘Does divine hiddenness justify atheism?’, in Michael L. Peterson and Raymond J. VanArragon (eds) Contemporary Debates in Philosophy of Religion (Malden MA: Blackwell, 2004), 30–41.

^6. Since the primary issue is non-culpable unbelief, we will henceforth drop ‘and able’ but it shall be understood to apply throughout.

^7. It should be noted that ‘reasonable’ also admits of multiple admissible precisifications. The reasonableness at stake could anywhere on a scale from purely subjective to purely objective. We think the method of the paper works for any sensible account of reasonableness. The two main interpretations which seem to fit Schellenberg's use are: (1) sufficiently probable on correct evidential standards, or (2) non-culpable in a robustly deontological sense. On either reading our argument remains unchanged. Thanks to Mike Thune for discussion on this.

^8. Belief de se is not relevant to our argument; hence we ignore it.

^9. If the reader prefers to think of a gradeable property of beliefs – confidence – we have no objection. Also, we are not committed to there being precise degrees of belief.

^10. Not that we think betting behaviour defines belief: it is, rather, defeasible evidence for it.

^11. This could be reasonably interpreted either as an interval during which reasonable non-belief is not to occur or as some specific time – a sort of ‘deadline’ – by which reasonable non-belief must be overcome.

^12. Naturally, the corresponding version of (3) will be weaker: (3a) There is some time at which reasonable non-belief occurs. This is no help, however, since we think (7a) is wholly inadequate. This will reveal a general pattern: on each reading one premise comes out better, but another wholly implausible.

^13. J. L., Schellenberg ‘The hiddenness argument revisited (I)’, Religious Studies, 41 (2005), 201–215, 206Google Scholar

^14. William J. Wainwright ‘Jonathan Edwards and the hiddenness of God’, in Howard-Snyder & Moser Divine Hiddenness: New Essays, 118.

^15. This is an a fortiori argument, for notice that premise (3) in the second version has a maximality property which is lacking in the previous argument. Essentially this point is made by Schellenberg in his response in ‘The hiddenness argument revisited (I)’, 208. Furthermore, the first argument just isn't Schellenberg's. Wainwright provides a good summary at the beginning of his article ‘Jonathan Edwards and the hiddenness of God’, (98) which is equivalent to our own. In light of this it is strange that he would then offer the present pair of arguments.

^16. J. L., Schellenberg ‘The hiddenness argument revisited (II)’, Religious Studies, 41 (2005), 287–303, 299–300Google Scholar.

^17. Ibid., 299.

^18. Richard Swinburne Providence and the Problem of Evil (Oxford: Clarendon Press, 1998), 257–258.

^19. Note that the idea that this is even possible seems to assume a kind of behaviourism not consonant with a humanistic understanding of personhood.

^20. Schellenberg ‘The hiddenness argument revisited (II)’, 288ff.

^21. Ibid., 288.

^22. Swinburne Providence and the Problem of Evil, 210–211.

^23. Schellenberg ‘The hiddenness argument revisited (II)’, 298–299.

^24. We wish to make it perfectly clear we think that most of Schellenberg's rebuttals fail in part or in whole. We offer the considerations we do because they differ from anything explicitly considered by Schellenberg, though they are closest to Robert McKim Religious Ambiguity and Religious Diversity (New York NY: Oxford University Press, 2001), 34–48, and we think they are some of the best reasons for hiddenness.

^25. Note that Schellenberg explicitly affirms that it is acceptable for a relationship with God to vary through time. In another response to Wainwright's parallel argument Schellenberg writes, ‘But one only gets an analogue for the mentioned claim and the odd result in my argument if one supposes it to say that God should at all times provide us with a fully salvific life, or at all times make the deepest possible human–divine communion available, or something along those lines. And I have said nothing of the sort. All I have said is that we might expect at all times to be in possession of belief, and to have at all times the opportunity to be involved in some level of explicit relationship with God. Indeed, in DH I emphasize that “the relationship I am thinking of is to be understood in developmental terms”, that were it to obtain, “it would admit of change, growth, progression, regression”, that it might be “shallow or deep, depending on the response of the human term of the relation”.’ (Schellenberg ‘The hiddenness argument revisited (I)’, 208).

^26. In fact it may be the case that the parties are both more confident than not that the other does not exist but because of the utility of there being another person they act as if there was another person in the adjacent cell.

^27. For a concise treatment of the lottery paradox, see Jonathan Kvanvig ‘The epistemic paradoxes’, in Edward Craig (ed.) Routledge Encyclopedia of Philosophy (London: Routledge, 1998).

^28. For an alternative view, see for example, as Mark Kaplan explicates it in Decision Theory as Philosophy (Cambridge: Cambridge University Press, 1996), pp. 107–110.

^29. See Paul Weirich ‘Belief and acceptance’, in Ilkka Niiniluoto, Matti Sintonen, and Jan Woleński (eds) Handbook of Epistemology (Dordrecht: Kluwer, 2004), 499–520.

^30. For a defence of this account, see Richard Swinburne Epistemic Justification (Oxford: Oxford University Press, 2001), 34–38.

^31. A good place to start is Jeff Jordan Gambling on God (Lanham MD: Rowman and Littlefield, 1992). An extensive bibliography follows a critical discussion in Alan Hájek ‘Pascal's Wager’, in Edward N. Zalta (ed.) The Stanford Encyclopedia of Philosophy (Spring 2004 edition), URL=< http://plato.stanford.edu/archives/spr2004/entries/pascal-wager/>.

^32. In Snyder and Moser Divine Hiddenness: New Essays, 113.

^33. Ibid.

^34. A fictional example of this kind of relationship may be found in the US TV series Magnum, P.I. Private investigator Thomas Magnum is hired by the rich, mysterious Robin Masters. Magnum lives on Robin's Hawaiian estate, run by Jonathan Higgins. The TV show suggests that Robin Masters is Jonathan Higgins. Although Magnum does not realize it for some time he is in a meaningful relationship with Robin Masters in virtue of his attitudes to Robin (e.g. gratefulness) and his current relationship with Higgins. Thanks to Kevin Meeker for suggesting this example.

^35. This was suggested by an anonymous reviewer for this journal.

^36. Mark 9.24, AV.

^37. Jeremiah 29.13, NKJV.

^38. 1 Corinthians 13.12, AV.

^39. We are grateful to the audience at the 2005 Midwest Regional Meeting of the Society of Christian Philosophers for helpful comments. We also thank the participants at The Prosblogion, a weblog for the philosophy of religion (http://prosblogion.ektopos.com/). For in-depth comments on earlier drafts we thank Jon Kvanvig, Kevin Meeker, Mike Thune, and two anonymous reviewers for this journal.