The Problem of Divine Hiddenness

The tradition has long had a name for the phenomenon at issue here. The prophetic literature speaks of a God who hides, and the Latin phrase for it, deus absconditus, has a history running through the medieval mystics, the Reformers, and Pascal. For most of that history the hiddenness of God was treated as a devotional difficulty: something the believer must endure, and sometimes a sign of spiritual seriousness rather than a mark against the belief.

J. L. Schellenberg turned it into an argument. His claim, first developed in Divine Hiddenness and Human Reason (Cornell University Press, 1993) and restated compactly in The Hiddenness Argument (Oxford University Press, 2015), is that the existence of people who are open to relationship with God and yet unable to believe that God exists is straightforwardly incompatible with the existence of a perfectly loving God. The argument has been the most active area in this part of philosophy of religion since, and it is now standard to treat it as a second argument alongside the argument from evil rather than as a variant of it.

The datum

Begin with what the argument uses as its evidence, because a good deal depends on getting it right.

The shift to nonresistance is a deliberate strengthening. An earlier way of framing the problem asked whether there is reasonable nonbelief, which invites the reply that the nonbeliever has assessed the evidence badly. Nonresistance sidesteps this. Whether the seeker's reasoning is any good is irrelevant to whether they are open, and openness is what love is supposed to respond to.

The category is also not empty, which is the second thing the argument needs. It is difficult to maintain that everyone who does not believe is refusing, given the existence of people who have looked hard, wanted to find something, and reported finding nothing, and given the many centuries and regions in which the relevant conception of God was simply not on offer.

The premise about love

Two features of the thesis matter for how the argument runs.

It concerns openness rather than actual relationship. God is not required to force anyone into anything. What is required is that the door be open from God's side, so that a person who wants the relationship can have it.

And openness requires belief. This is the step the argument depends on, and it is stated as a conceptual point: one cannot enter into a personal relationship with someone one takes not to exist. Whatever else a relationship with God involves, it involves addressing God, trusting God, and being conscious of God, and none of these is available to someone for whom the question of God's existence has come out negative.

Put those together and the conclusion is quick.

The argument is deductive, which distinguishes it from the evidential arguments of the last several rungs. It does not claim that hiddenness makes theism improbable; it claims that the two are inconsistent. This makes the theist's task clearer than it is with evidential arguments, since exactly one premise has to be denied, but it also makes the argument brittle, since a single successful denial defeats it entirely.

What being able to believe amounts to

The phrase "unable to believe" carries more weight than it looks, and clarifying it removes a common misreading.

Belief is not under direct voluntary control. I cannot believe that there is a coin in my hand by deciding to, and the point generalises: beliefs form in response to how things seem, and the will can influence them only indirectly, by directing attention or by putting oneself in the way of new evidence. This was the difficulty that made the pragmatic arguments of the beginner tier awkward, and it matters here for a different reason.

If belief cannot be adopted at will, then a person's inability to believe is not a choice they are making, and the theist cannot treat it as one without denying a fairly secure point about how belief works. The seeker who cannot believe is in the position of someone standing in a room where the light has not come on, not the position of someone with their eyes shut.

What the argument therefore requires of God is not that God install a belief but that God make the situation one in which belief forms naturally in a person who is open. On the theistic picture there is no shortage of means: an experience, a sense of presence, a conviction of the kind rung 10 discussed. None of these needs to be coercive, and none needs to be publicly available. The argument's demand is minimal in exactly the way that makes it hard to answer, since a God with the resources theism ascribes could meet it in countless ways without producing certainty in anyone.

Denying the third premise

This has strong precedent in the tradition. The Reformed doctrine of the sensus divinitatis holds that awareness of God is naturally available to everyone and that its absence is explained by the corrupting effect of sin on the faculty rather than by any failure of God's part, and Paul's remark that what may be known of God is plain has been read the same way for a long time.

Three difficulties face this reply, and together they are why it is now a minority position.

The first is evidential. The claim is a very strong empirical generalisation about the inner lives of billions of people, made by someone with no access to those inner lives, and made in a way that conveniently protects a favoured conclusion. Any nonbeliever who reports having sought sincerely is, by the terms of the reply, either lying or self-deceived, and the reply has no independent test to distinguish those from the honest case.

The second is moral. Telling a person who spent years seeking that their failure to find was a fault in them is an accusation, and it is one the theist is not in a position to substantiate. The anti-theodicist's charge from the previous rung transfers directly.

The third is that the reply overreaches historically. Entire populations across long periods had no exposure to the conception of God at issue, and describing their nonbelief as resistance is not credible. The theist may say that God's dealings with such people run through some other channel, but that concedes the point: if relationship is available without belief in those cases, the second premise is what should have been denied, not the third.

Denying the second premise

The more promising line grants that nonresistant nonbelief exists and denies that a loving God would prevent it. The claim is that some good depends on the availability of hiddenness, and that a loving God might permit it for that good's sake. This is theodicy transposed into a new domain, and the candidate goods are recognisably related to the ones met earlier.

The best developed version concerns coercion. Michael Murray, in his contribution to Divine Hiddenness: New Essays, edited by Daniel Howard-Snyder and Paul Moser (Cambridge University Press, 2002), argues that evidence of God's existence sufficiently overwhelming to remove all doubt would function coercively on any finite agent, since the awareness of being under the constant observation of an omnipotent being who has attached infinite consequences to one's conduct would swamp all other motives. A creature in that condition could not develop a genuinely virtuous character, because it could never act from any motive other than the recognition of who is watching. Hiddenness, on this account, creates the epistemic distance within which free moral development is possible.

Kierkegaard's treatment belongs to the same family, though the emphasis differs. The idea there is that a relationship of trust has a value that a relationship of certainty does not, and that a God who made himself epistemically unavoidable would have made the specific virtue of faith impossible. Pascal made a related point when he observed that the evidence available is enough to guide those who are willing and not enough to compel those who are not, and treated that as an intelligible design rather than an accident.

There is also a cost to the coercion reply that is worth naming. If awareness of God would swamp all other motives and make virtue impossible, then the many people within the traditions who report a vivid and settled sense of God's presence are in exactly the compromised condition the reply describes. The theist is then committed either to saying that those reports are exaggerated, or to saying that the most devout are the least free, and neither is a comfortable thing for a theist to say.

Schellenberg's reply is that these accounts prove too little and the wrong thing. They may explain why God does not provide overwhelming proof, but the argument does not require overwhelming proof; it requires only that the person be able to believe, which is compatible with the evidence remaining modest and resistible. And the goods on offer all presuppose that the person is already in a position to respond to God, which is precisely the position the nonbeliever is not in. A hiddenness that protects the moral development of believers by excluding some people from the relationship altogether has not been explained.

Denying the first premise

The most interesting recent work attacks the analysis of love itself.

Michael Rea, in The Hiddenness of God (Oxford University Press, 2018), argues that the argument's picture of what love requires is more parochial than it appears. Our sense that a loving person must be constantly available, must communicate in ways the beloved finds legible, and must not permit long silences, is shaped by particular expectations about relationships, and there is no evident reason why a being whose nature is as unlike ours as God's is supposed to be must express love within them. Rea also presses the point that God, on the theistic conception, has a personality, and that a personality includes modes of self-presentation that are not chosen to suit the preferences of the other party.

This is a serious challenge, and its cost is equally serious. The more the theist insists that we cannot say what divine love would look like, the closer the position comes to the sceptical reply of rung 15, with all the consequences catalogued there. And it sits badly with the devotional literature of the traditions themselves, which is not reticent about what God's love involves and which frequently uses precisely the parental and marital analogies the argument exploits.

Where it stands

Three observations are fair.

The argument has succeeded in establishing that hiddenness is a distinct problem rather than a special case of evil. The two arguments have different data, different premises, and different available replies, and a theodicy that handles suffering may leave hiddenness untouched. Theistic philosophers have largely accepted this framing.

The structure of the debate now mirrors the one over evil closely enough to be predictable. There is a strong deductive version, replies that identify goods secured by hiddenness, sceptical replies about our access to what love requires, and disputes about whether the sceptical replies can be contained. The parallels are not accidental, since both arguments run from a feature of the world to a claim about what a perfect being would do.

The deductive form is not the only one, and its brittleness has led to the same migration seen at rung 14. A weaker version claims only that the pattern of belief and nonbelief we observe is more expected on the hypothesis that no loving God is arranging matters than on the hypothesis that one is. That version cannot be defeated by finding one flaw in one premise, and it draws on a further datum the deductive version does not use: religious belief correlates strongly with where and to whom a person was born. Whatever else that fact is evidence of, it is hard to fit with the claim that availability is being managed person by person by someone who loves each of them. The pressure this puts on the truth claims of any single tradition is taken up at rung 20.

There is one point at which hiddenness presses harder. The argument from evil is about what God permits to happen; the hiddenness argument is about God's own conduct toward a particular person. Appeals to the free choices of others, which do a great deal of work in responses to evil, have no purchase here, since nobody else's freedom is at stake in whether a given seeker is able to believe.

Next

The last several rungs have asked whether God's conduct toward creatures is consistent with perfect goodness. The next takes up a difficulty about creation as a whole, and it arises even if every question about suffering and hiddenness were settled. If for every world God might have made there is a better one available, then whatever God made, a better being could have done more, and the notion of the best that omnipotence could do may turn out to have no content.