Divine hiddenness, the demographics of theism, and mutual epistemic dependence: a response to Max Baker-Hytch

Abstract

In his article ‘Divine Hiddenness and the Demographics of Theism’ Stephen Maitzen (2006) develops a permutation of the argument from divine hiddenness which focuses on the uneven distribution of theistic belief around the globe. Max Baker-Hytch (2016) responds to this argument by providing a theodicy which appeals to the fact that humans are epistemically interdependent. In this article I argue that Baker-Hytch's response is at best incomplete and at worst relies on a faulty modal judgement. After exploring some ways Baker-Hytch might salvage his theodicy and maintaining their failure, I conclude with the success of Maitzen's argument.

Introduction

Stephen Maitzen (2006) has developed an innovative permutation of the argument from divine hiddenness which focuses on the uneven distribution of theistic belief across the globe.¹ He notes that contemporary demographic data show theism to be exceptionally common in some countries and exceptionally rare in others.² For example,

The populace of Saudi Arabia is at least 95 per cent Muslim and therefore at least 95 per cent theistic, while the populace of Thailand is 95 per cent Buddhist and therefore at most 5 per cent theistic. The approximate total populations are 26 million for Saudi Arabia and 65 million for Thailand. (Maitzen 2006, 179)

Maitzen argues that this lopsided distribution of theistic belief – which occurs in other countries as well – is quite surprising given theism. For according to this view, there exists a God who seeks to be personally related to everyone he creates and who has the power to make his existence universally known. But Maitzen adds that on an alternative view such as naturalism, the uneven distribution of theistic belief in the world seems less surprising. For on naturalism, the spread of theistic belief is highly influenced by historical, cultural, and geographical proximity of persons, and there is no force for spreading theism that has the power to transcend these influences. Maitzen's conclusion is that the lopsided distribution of theistic belief in the world is much more surprising and improbable given theism than naturalism and therefore constitutes powerful evidence for the latter over the former.³ So where ‘E’ stands for ‘evidence’ and is the proposition that theistic belief is unevenly distributed in the world, we can capture the core of Maitzen's argument as follows:

(D): Pr(E|naturalism) is much higher than Pr(E|theism).

Max-Baker Hytch (2016) has responded to this argument by emphasizing that humans are mutually epistemically dependent (MED). He argues that MED is fairly unsurprising on theism and, moreover, that MED leads us to expect the geographical clumping of religious beliefs we see in the world.

In this article I critique Max-Baker Hytch's appeal to MED. I argue that his response is at best incomplete, and at worst, relies on an implausible modal judgement. I begin by briefly explaining his theodicy. Next, I contend that it fails to undermine Maitzen's argument. Since Baker-Hytch also draws heavily upon the concept of de re relationship with God, I address this idea in a separate section. Finally, I consider ways Baker-Hytch might salvage his response. I conclude with their failure and the success of the demographic argument.

A response to the demographic problem: mutual epistemic dependence

As mentioned previously, Baker-Hytch (2016) has tried to cast doubt on (D) by appealing to MED. MED refers to the fact that humans are

cognitively constituted in such a way that it is a practical necessity that we rely upon testimony for much of what we know about the world and in such a way, moreover, that we are liable to be significantly influenced by what those around us believe, particularly when it comes to matters that aren't readily susceptible to empirical investigation, including religious matters. (Baker-Hytch 2016, 377)

Baker-Hytch adds that MED is a result of, among other factors, two types of cognitive biases displayed by humans. These two biases, which are sometimes known as context and content biases, greatly influence how likely we are to accept any given piece of testimony. Context biases, including conformity, prestige, and similarity bias, all jointly entail we will more likely accept the testimony of those trusted and admired within our immediate social contexts (Barrett 2011, 42–47). And content biases, which include ‘minimally counterintuitive ideas’ (Boyer 2001, 87–100), entail that concepts with memorable or interesting characteristics, such as religious concepts, will be socially contagious (Gervais et al. 2011, 398).

Hence given MED and these biases, Baker-Hytch thinks it is hardly surprising to see considerable divergence both in the content of religious beliefs in the world and in the contexts in which they arise. For if humans are dependent upon each other for information and are more likely to trust the testimony of those they know and respect, then given that communities are somewhat geographically and culturally isolated, something like E is to be relatively expected. Baker-Hytch therefore claims that:

(A1): Pr(E|MED) is high.

Now Baker-Hytch grants that MED is not very surprising given naturalism.⁴ So to cast doubt on (D), he attempts to show that MED is fairly unsurprising on theism. He argues that if God exists and creates a world with embodied persons – which, following Swinburne (2004), we might believe God has good reason to do – then God has good reason to bestow these embodied persons with the sort of cognitive constitution humans have. This is because the cognitive powers which humans possess strike a favourable balance between two undesirable extremes.

To elaborate, imagine that God decides to create embodied persons. Available to him are (perhaps) an infinite number of types of creatures with varying degrees of cognitive powers. On one end of the spectrum, there are creatures with very low cognitive powers. Creatures who, for instance, lack the ability to make any intelligent inferences about the world and what other people say. On the other end of the spectrum, there are creatures with very significant cognitive powers. Creatures who, for instance, can read minds or acquire new information at will. But humans in the actual world, Baker-Hytch alleges, have cognitive powers situated between these extremes. And this permits a favourable balance between the following competing goods: (i) exercising interpersonal trust versus being invulnerable to deception; (ii) sharing responsibility for one another's acquisition of epistemic goods versus practising epistemic self-reliance; and (iii) having opportunities to acquire, practice, and perfect intellectual virtues versus freedom from intellectual challenges (Baker-Hytch 2016, 380–384). Thus, Baker-Hytch says it would be fairly unsurprising for God to bring about MED, as it makes possible a broad range of goods God would probably want to realize. Baker-Hytch therefore claims that:

(A2): Pr(MED|theism) isn't much lower than Pr(MED|naturalism).

Moreover, from (A2) and (A1) Baker-Hytch infers the following conclusion:

(C): Pr(E|theism) isn't much lower than Pr(E|naturalism).

And this proposition is logically equivalent to (C1):

(C1): Pr(E|naturalism) is higher than Pr(E|theism).

Recall Maitzen's contention, (D). As Baker-Hytch has now supplied an auxiliary hypothesis which makes E more expectable on theism, (C1) reflects a more accurate comparative likelihood than (D). While Pr(E|naturalism) is still higher than Pr(E|theism) on Baker-Hytch's account, it is no longer much higher than Pr(E|theism). So contrary to Maitzen's claim, the uneven distribution of theistic belief in the world is not powerful evidence against theism. At the most, it only weakly supports naturalism over theism. Or so Baker-Hytch believes.

A critique of the mutual epistemic dependence response

So far I have introduced the demographic problem of divine hiddenness and a response to it from Baker-Hytch. It is important to note that while Baker-Hytch develops his theodicy specifically in relation to the demographic problem of divine hiddenness, it could also be applied mutatis mutandis to other hiddenness arguments as well. It is therefore especially deserving of attention and in this section I offer a critique of the response.

For the sake of argument, let us concede to Baker-Hytch that God has reason to create a world with persons who are significantly reliant upon each other in roughly the way that humans are, as this epistemic interdependence promotes a balance of the goods listed in (i)–(iii). Of course, some may wish to contest these points, but let us push that to the side for now. For even granting them, it remains unclear why God would also let these individuals be reliant upon each other for one piece of knowledge in particular, namely God's existence.⁵

By opting for a world where knowledge of God's existence falls within the purview of MED, God anticipates the demographic problem and frustrates his aim of seeking relationship with every person that exists. For assume the world is much like our own. Unless there is special intervention by God to make his existence and nature universally manifest, whether any particular individual comes to possess theistic belief will be significantly (if not wholly) determined by the individual's historical, cultural, and geographical happenstance. That is, by their immediate social context. Thus, it may well be that a vast number of people end up lacking belief in God. And since the belief that God exists is necessary for a reciprocal relationship with God, many people might therefore be precluded from participating in a relationship with God.⁶ So why would God create a world like this, when on the version of theism we are entertaining, God loves every person equally and desires relationship with any finite person there may be?

At times Baker-Hytch seems to think that if God actualizes any world with persons who are epistemically interdependent to roughly the extent humans are, then any proposition these persons come to believe about God will be highly contingent upon MED. Or in other words, Baker-Hytch appears to pick out ramifications of MED in the actual world and assume the same ramifications hold in every possible world in which MED-like affairs obtain. He therefore appears to assume something like the following modal judgement:

(M): Necessarily, for any world W in which there exist embodied finite persons S and in which S are significantly epistemically interdependent, whether any S believe any propositions about God in W will be highly contingent upon their immediate social contexts.

Something like (M) seems to be in the background of Baker-Hytch's thought when he writes the following about creatures who have roughly the same cognitive powers that humans have:

More specifically, it should not come as a surprise that certain propositions concerning the nature of ultimate reality would be believed by many such creatures in some communities and by hardly any of them in other communities. (Baker-Hytch 2016, 378; emphasis added)

What's more, propositions concerning such matters as metaphysics, religion, politics, and morality tend to be such that, for creatures who are subject to such cognitive limitations as we are, it is frequently far from obvious how to adjudicate the truth or falsity of these sorts of propositions with reference to perceptual experience. (Baker-Hytch 2016, 378; emphasis added)

Therefore, whereas one would not expect much divergence among such creatures with respect to their beliefs about the physical properties of publicly observable, medium-sized physical objects, it would not be at all surprising to find that they diverge considerably with respect to the contents of their religious beliefs. (Baker-Hytch 2016, 379; emphasis added)

If (M) is true, Baker-Hytch's theodicy would be complete.⁷ But there is, however, very little reason to accept (M). An unsurpassably powerful God has (perhaps) unlimited resources at his disposal. Surely there are any number of ways in which God could set up the world or cognitively constitute epistemically interdependent persons so that they would not be significantly influenced by and reliant on those around them for knowledge of God's existence. Take, for example, a world in which God provides clear manifestations of himself to every individual by way of miraculous wonders or theophanies (Hanson 1972, 313–314).⁸ By creating a world like this, God need not let knowledge of his existence be so highly contingent upon any person's immediate social context. The demographic problem would thus probably disappear. But the point missed by Baker-Hytch is that people in such a world could still possess a robust kind of MED. Excluding perhaps their inner mental states, immediate perceptual experience, a few incorrigible truths, and their belief that God exists, these humans could still be cognitively constituted so as to rely upon each other for much of their knowledge of the world. In this way, the same goods Baker-Hytch describes could all be realized within this type of world. For people in this world will not be invulnerable to deception and will still be able to exercise significant interpersonal trust. They will also have a choice between helping each other acquire epistemic goods or choosing epistemic self-reliance. And there will still be many opportunities for them to overcome intellectual obstacles and challenges.

Let us call the type of world I have attempted to describe a Limited MED world. This is a world similar to the actual world but where the following three conditions are met:

1. (a) Persons are not significantly dependent upon each other for acquiring belief in God.

2. (b) Belief in God is secured for each person irrespective of their immediate social (historical, geographical, and cultural) context.

3. (c) Persons are significantly epistemically interdependent.

And let an Extended MED world designate any world like ours where (c) is true but not (a) and (b). The main objection I see Baker-Hytch raising to a Limited MED world is that there would be an inappropriate epistemic distance between God and humans and that human freedom would therefore be inhibited. Consider again a world where God reveals himself to every individual through powerful miraculous encounters. Baker-Hytch might allege that people in such a world would either be irresistibly drawn to God's goodness (Hick 2010, 372–373), too greatly wish to be well-liked by God (Swinburne 2004, 268–269), or would be compelled under the context of threat (Murray 1993).

Nevertheless, one obvious reply is to point to other Limited MED worlds that are less obviously freedom-inhibiting. For God could reveal himself in many subtler ways so as to not crush human autonomy. Consider, for example, a scenario delineated by J. L. Schellenberg. He asks us to imagine a world where evidence for God's existence is made available to each person by way of personal religious experience.

Suppose, then, that the world is one in which all human beings who evince a capacity for personal relationship with God have an experience as of God presenting himself to them, which they take to be caused by God and which actually is caused by God presenting himself to their experience. This experience, let us say, is non-sensory – an intense apparent awareness of a reality at once ultimate and loving which (1) produces the belief that God is lovingly present (and ipso facto, that God exists), (2) continues indefinitely in stronger or weaker forms and minimally as a ‘background awareness’ in those who do not resist it, and (3) takes more particular forms in the lives of those who respond to the beliefs which it gives rise in religiously appropriate ways. (Schellenberg 1993, 48–49)

As Schellenberg extensively argues (1993, 95–130), this experience need not be so powerful as to overwhelm human freedom. Whereas an experience of a grand miracle may be coercive or forceful in nature, a subtle religious experience need not be. The latter experience could merely be persuasive instead of overwhelming and so humans could choose to ignore the experience should they desire to do so. But as Schellenberg rightly notes (1993, 52), this choice would clearly be one of resistance.⁹

But even should one reject the world advocated by Schellenberg on grounds that such a religious experience would be unacceptably freedom-inhibiting, other Limited MED worlds remain. God could, for example, directly implant the belief that he exists into our cognitive make-up (Drange 1998). Or the ability to form belief in God might have arisen early in infancy, much like how young children universally possess the ability to detect faces and learn language.¹⁰ Or as Nick Trakakis suggests (2007, 222), God could have created humans so that the proposition ‘God exists’ strikes nearly everyone as self-evidently true in much the way that items like ‘1 + 1 = 2’ or ‘if a person is a bachelor, then he is an unmarried man’ strike nearly everyone as obviously true.¹¹ Even John Hick, a staunch defender of the epistemic distance concept, is willing to admit that coming to belief in God in this kind of way – detached from any miracles or personal encounters with God – would probably not compel a person, or even necessarily incline them, to seek relationship with God.¹² Of course, in the aforementioned worlds, epistemic distance between God and humans would be diminished. But so long as this diminishment did not infringe upon human freedom, which even Hick seems to concede it would not, then it is difficult to see how this lessening of epistemic distance would be negative.

Consider still other ways in which God could make his existence known while safeguarding human autonomy. Take, for example, propositions that are neither necessary truths nor impossible to resist but which receive almost universal acceptance. Propositions I have in mind might be those such as: ‘other minds exist’, ‘the world wasn't created five minutes ago’, ‘it is morally permissible to have children’,¹³ ‘people sometimes act for reasons’, ‘some events have causes’, or ‘human beings know various things’. While it is possible to reject propositions like these, and some individuals do, these items are such that they are widely accepted across cultures – that is, there is no meaningful lopsidedness with respect to them being mostly believed in some countries and mostly disbelieved in others. The point here is that if God exists, humans could have been created in such a way that we accord the proposition ‘God exists’ a similar status accorded to propositions like those just mentioned. And if this were the case, then we would expect belief in God to be ubiquitously spread. But just as some individuals reject the aforementioned items in the actual world, so too would there be room to cognitively resist belief in God in such a world.

Given the foregoing discussion, it is relatively evident that not all Limited MED worlds would be freedom-inhibiting. Moreover, in many Limited MED worlds, including the ones just described, the inhabitants could still be reliant on each other for most of their information of the world. As such, MED and its concomitant goods can be preserved in these worlds. But the demographic problem, or E, would not be, because whether the inhabitants believe in God would not be so highly determined by their immediate social contexts.¹⁴

However, perhaps Baker-Hytch did not have (M) in mind when crafting his response. Perhaps Baker-Hytch instead thinks that if God actualizes any world like ours with MED-like affairs, then even though knowledge of propositions about God need not be so highly contingent upon MED for any person, God will still desire to make them so. For Baker-Hytch might have a specific good in mind which is made possible by God doing so, which would make an Extended MED world overall preferable to God.

Unfortunately, Baker-Hytch fails to mention such a point. As it stands then, his account is at best incomplete. At worst, it implicitly relies on the implausible modal judgement (M). In the rest of this article I will look at various goods Baker-Hytch might appeal to in order to complete his account. But before moving on to this, we need to clear up a point Baker-Hytch raises about the possibility of de re relationship with God.

De re relationship with God

Halfway through his article Baker-Hytch argues that individuals can participate in an implicit relationship with God without the explicit belief that God exists. Following common usage of the term, let us call this a de re relation to God. While the precise import of this idea on his own theodicy is left unspecified by Baker-Hytch, I take it that he views it as an auxiliary hypothesis which, in conjunction with his own account, helps lighten the conceptual burden his theodicy has to bear.¹⁵ For if de re relationship with God is available to individuals in geographical regions where theism is largely absent, then non-theists in these areas will at least be able to participate in a kind of rudimentary relationship with God. And if this is true, then perhaps God will be less concerned about making awareness of his existence more widespread, given that non-theists will not be completely cut off from relationship with God. So given theism, something like E would be less surprising. To fully address Baker-Hytch's theodicy then, a couple of comments about this idea are in order.

The concept of de re relationship with God should be familiar to most readers of divine hiddenness by now, as it is a prominent idea in the literature. Regrettably, theists who employ the response seldom seem to engage with Schellenberg's (1993, 41–43; 2007a) longstanding critiques of it. Regardless, authors such as Imran Aijaz and Markus Weidler (2007), Andrew Cullison (2010), and Jeff Jordan (2006) all point to other propositional attitudes besides belief, such as acceptance or hope, which can motivate and sustain relationships. Likewise, Ted Poston and Trent Dougherty (2007), C. Stephen Evans (2006), William Wainwright (2002), and Daniel Howard-Snyder (2015) all argue for the notion of ‘implicit’ or ‘nonconscious’ relationships. The common thread between all these authors is consensus that a person S need not believe that God exists for S to participate in a relationship with God.

While Baker-Hytch supports this thesis as well, he fleshes it out in terms of reference-theory – more specifically, in terms of how one might pick out an object with a definite description without getting the description right. His primary concern is to show that, so long as one thinks about the divine under certain characteristics possessed by God (e.g. the creator of the universe, the moral lawgiver, and so on), one might still refer to God despite holding substantially false beliefs about the divine. Baker-Hytch suggests two ways this reference may occur. The first is if a description of God is being used in a purely referential sense. This is where the description acts merely as an instrument to call attention to a person or thing. In this case,

it might be that the subject who is using the description has some kind of direct cognitive contact with the object and is merely using the definite description as a means of ostension. By ‘direct cognitive contact’, I mean that the object is in some way present to the consciousness of the subject, perhaps as an object of her perception or as a proximate cause of some salient aspect of her perceptual experience. (Baker-Hytch 2016, 387)

So where a person experiences direct cognitive contact with God like this, they might use a description that is false of God. But since this description is merely an instrument to draw attention to the referent of their experience, they may still refer to God.

The second way Baker-Hytch thinks reference may occur is as follows:

Suppose the subject employs a description involving several properties – for instance, the x which is F, G, and H – some of which are possessed by the object in question and some which are not. The mechanism by which the subject succeeds in referring to the object with that description might be the greater salience, in the context at issue, of the properties that are possessed by the object as compared with the properties that aren't. (Baker-Hytch 2016, 387)

Imagine then that a person attributes to God some properties that God possesses and some that he does not. Imagine moreover that the person predicates more of the latter than the former. So long as the properties God does possess are the most important properties associated with the divine for that person or are playing the primary role in generating the person's response of worship, reverence, or whatever, to the divine, then the person may still succeed in referring to God.

To respond, let us assume for the sake of argument that Baker-Hytch is correct: that some non-theists may refer to God despite having substantially false beliefs about the divine. In this way, these individuals can participate in a relationship with God. What conclusion follows for Baker-Hytch's MED-theodicy? I would suggest not much of significance. Granted, the possibility of individuals participating in de re relationship with God does go some way in establishing (C). For E would certainly be more surprising on theism without this possibility. But it must be remembered that Baker-Hytch employs the idea as an auxiliary consideration. His main aim is to show that (C) gets most of its support from an appeal to MED, not from the notion of de re relationship with God.¹⁶ So the success of Baker-Hytch's account rides more on the former than the latter. Of course, Baker-Hytch could always drop his appeal to MED or relegate its importance in favour of the de re response. But I think Baker-Hytch would not readily do so, because of two further problems.

First, from the text and my conversations with Baker-Hytch, it is relatively clear that a limited number of individuals will be able to relate to God via the mechanisms described above. More specifically, only individuals who conceive of ultimate reality in personal terms will be likely to do so. For recall that individuals who refer to God must, according to Baker-Hytch, think about the divine under certain descriptions that are possessed by God (e.g. the creator of the universe, the moral lawgiver, etc.). Presumably this is because if someone does not think ultimate reality satisfies even these fundamental descriptions, it's difficult to see how they could be referring to anything like the God of traditional theism.¹⁷ As such, it's hard to see how the mechanisms of reference Baker-Hytch suggests will be relevant in cases where individuals do not think about the divine under these descriptions – where, for instance, ultimate reality is conceived of by an individual in non-personal terms. In fact, non-theistic religious believers are not the only ones in this boat: atheists and agnostics will also not be able to refer to God under such mechanisms. So, on Baker-Hytch's account, the fact that a whole swathe of religious adherents and atheists and agnostics are left without the opportunity for either a de re or more explicit relationship with God is still surprising.

Now Baker-Hytch is aware of such a problem, which is why he suggests that any leftover individuals may be able to participate in a kind of proto-relationship with God. The idea is that, where God is causally responsible for some aspect of a person's experience,

God might very well count the expressions of gratitude, reverence, and obedience that a human being directs towards some non-existent deity as though they were directed towards God, so long as God knows that she would direct those same attitudes towards him were she apprised of the true identity of the moral lawgiver, the creator of the universe, and so on. (Baker-Hytch 2016, 388)

Baker-Hytch (2016, 389) gives an example of an atheist named Anders who is grateful for his existence. While Anders does not believe that he owes his existence to any personal source, such as a God, Anders abides by an exemplary moral code and lives with a deep sense of appreciation for what he has. Baker-Hytch's point is that, if Anders was apprised to the fact that God is the source of his existence, Anders might very well direct his gratefulness towards God. And God, knowing all counterfactual truths about Anders and whether Anders would (probably) do so, could decide to accept Anders's sense of gratefulness as directed towards God. If so, we can imagine that once God decides finally to reveal himself to Anders as the benefactor of his life, Anders might proclaim, ‘it was you all along!’ (Baker-Hytch 2016, 389). And the ensuing relationship between Anders and God could take on a deeper significance than it would have otherwise. Of course, Baker-Hytch recognizes that prior to this point, Anders would not be in a relationship with God in the ordinary sense of the word. But he does think this proto-relationship could contribute retrospective significance to the more explicit, future relationship between Anders and God.

However, while this is all well and good, it highlights the second main problem with the de re response: the question of why an unsurpassably loving God would choose to relate in such implicit ways to non-theists. Remember that individuals who relate de re to God or participate in a proto-relationship with God would not describe their situation as ‘relating to God’. In fact, the point is that such individuals will be engaging in a type of communion with God without explicit awareness that this is what they are doing. But notice how this is quite foreign to what we ordinarily consider a loving relationship to entail.

Consider an analogy. Take, for example, a loving relationship between a child and their parent.¹⁸ If a parent loves their child, we ordinarily expect the parent to seek to relate to their child in a way that involves conscious recognition of the parent. We would be extremely surprised, for example, if a parent were to relate to their child only anonymously through letters, letting the child believe that the parent either did not exist or was someone else entirely. In fact, for us to continue viewing the parent as loving, we would need to assume that the parent had an incredibly weighty reason for staying distanced from their child. For without such a reason, the parent could not rightly be defined as loving.

Analogously, some non-theists might be able participate in a de re relationship with God without believing that he exists. And some individuals like Anders might be able to partake in a proto-relationship with God that takes on retrospective significance. But if a loving parent seeks to be related to their child in a way that involves conscious recognition, how much more should we expect an unsurpassably loving God to seek to be related to any persons he creates in a way that involves conscious recognition.¹⁹ So unless God has an overriding reason to relate to non-theists in such an anonymous fashion, then all other things being equal, he would probably not relate in such a way.

It is for the foregoing reason that the ‘de re relationship’ response is parasitic on theodicy. In the case of our present topic, in the absence of a successful explanans (theodicy) for the explanandum (demographic problem), the response adds limited value to the discussion. So, unless Baker-Hytch has first established his theodicy on independent grounds, which I have argued that he has yet to do, then appealing to this auxiliary hypothesis lends little plausibility to his overall account. For without God having an overriding reason to relate ‘de re’ to non-theists in geographical areas such as Thailand, God would probably not, all other things being equal, relate to them in such a way. In the remainder of this article I will therefore put this objection to the side and look at several ways in which Baker-Hytch might seek to complete his theodicy.

Notice moreover that nothing of the foregoing discussion is undermined by the fact that Baker-Hytch makes use of an ‘inclusive’ theistic hypothesis for his account – that is, a version of theism where post-mortem salvation for humans does not necessarily require ante-mortem belief in God. Of course, if ante-mortem belief in God were necessary to avoid, say, damnation, then Baker-Hytch's account could hardly get off the ground, as Baker-Hytch himself admits (2016, 376). But it's important to note that the demographic argument and concerns about the impoverished nature of de re relationship can be run without any appeal to afterlife considerations. In fact, the argument from divine hiddenness as originally formulated by Schellenberg (1993) arose solely from reflection on the nature of love and its implications for ante-mortem human life.²⁰

Goods made available by making belief in God subject to MED

If we recall my earlier argument, I contended that even if God has reason to create a world in which MED obtains, we cannot merely assume, like Baker-Hytch does, that God would let MED greatly influence whether any of the persons in this world acquired belief in God. Rather, we have good reason to believe that these persons would acquire belief in God independently of their MED, as there is a possible world in which God could eliminate the demographic problem while also obtaining MED and its concomitant goods. Still, Baker-Hytch might object to my criticism by pointing to an additional good made possible by God allowing the proposition ‘God exists’ to fall within the purview of MED. This would be a good for which God would be willing to permit the demographic problem – a problem which entails that millions or even billions of people are precluded from participating in a conscious relationship with God. To my mind, there are two such possible goods most relevant to Baker-Hytch's theodicy. In this section I evaluate each in turn.

First, borrowing a point from Swinburne, Baker-Hytch might argue that by opting for an Extended MED world, God expands the range of morally significant actions available to theists.²¹ As Swinburne writes:

The agnosticism of the agnostic also makes possible a great good for the religious believer. It allows the believer to have the awesome choice of helping or not helping the agnostic to understand who is the source of his existence and of his ultimate well-being (helping the agnostic not merely by verbal preaching but by an example of what living a religious life is like). (Swinburne 2004, 271)

So Baker-Hytch might argue that by opting for an Extended MED world, God thereby grants theists the great responsibility of proselytization. Theists have the choice either to enlighten others of spiritual truth or to leave them in spiritual darkness. Formally, we might state this argument as follows:

1. (1) An Extended MED world makes available to theists the opportunity of proselytization.

2. (2) Theists would not have the opportunity of proselytization in a Limited MED world.

3. (3) The good of proselytization outweighs the good God sacrifices by allowing the demographic problem.

4. (4) Therefore, other things being equal, it is plausible to suppose that God would prefer an Extended MED world over a Limited MED world.

But there are significant problems with (2) and (3). Let us consider (3) first. Recall that persons in a Limited MED world can still be epistemically interdependent for much of their knowledge of the world. Hence, apart from knowledge of God's existence, persons in such a world can still possess significant responsibility for each other's acquisition of epistemic goods. For example, individuals might choose to deprive others of knowledge or share knowledge with them, distort the truth or faithfully relay the truth, and cooperate with others while searching for the truth or thwart anyone on this journey. If we add that such people can possess significant responsibility in other areas of their lives as well, much as they do in the actual world, then there will be no shortage of ways in which these people could contribute to their world for good or ill. There will therefore be ample opportunity for people in a Limited MED world to exercise their moral freedom.

Given this, why think the good of proselytization outweighs the many number of people who are precluded from participating in conscious relationship with God? A relationship with God is, by many theists' own admission, the greatest conceivable good in this life. Moreover, a relationship with God brings with it untold moral, experiential, and relational benefits (Schellenberg 1993, 18–26).²² Surely God would prioritize relationship with everyone over providing one more opportunity to theists to exercise their moral freedom, seeing how ample opportunity for exercising moral autonomy is already available to theists in the absence of this choice. This becomes increasingly obvious when we remember that the good of proselytization is available to only a subset of humans, namely theists. God has less reason to permit the demographic problem then, for it is not a universal good made available by MED's extension, but only a good experienced by some individuals. (3) is therefore doubtful. The value of proselytization seems far less great than the value of creating a world where the demographic problem is eliminated.²³

But even if (3) were true, (2) remains problematic. The good of proselytization could still exist in a world where Limited MED obtains. For if there are any resistant non-believers in the world, theists will have the opportunity either to help convert these individuals or to leave them in their non-belief. In this way, proselytization regarding God's existence would remain on the table. Granted, the number of opportunities for proselytization would presumably decrease. But given the value of a world without the demographic problem, it is at least difficult to see how a Limited MED world – which can still preserve the good of proselytization to some extent – would not be preferred by God. In fact, if proselytization was also tied to the dissemination of special revelation (Schellenberg 1993, 194), God could still secure a bountiful number of opportunities for proselytization. For if God wanted to reveal to humanity anything beyond bare facts of his existence, he could do so in a non-universal way. God could then grant some individuals the responsibility of disseminating this special revelation to others. Of course, whether God would prefer to tie proselytization to the dissemination of special revelation in this way would presumably depend on how important the contents of the special revelation in question was; how much the possession of that special revelation would enhance the quality of an individual's relationship with God.²⁴ Too much confusion about God's nature, desires, or commandments that might allow for this type of proselytization could compromise the divine–human relationship. But perhaps there is valuable yet significantly less consequential information God could disseminate via special revelation. One example might be God's relationship to time – whether God is timeless or temporal. So God could potentially retain even the good of widespread proselytization opportunities and eliminate the demographic problem in a Limited MED world.

A second way Baker-Hytch might complete his account is by appealing to the value of religious diversity. The gist of the argument would be as follows:

1. (5) An Extended MED world makes available the good of religious diversity.

2. (6) The good of religious diversity cannot be realized in a Limited MED world.

3. (7) The good of religious diversity outweighs the good God sacrifices by allowing the demographic problem.

4. (8) Therefore, all other things being equal, it is plausible that God would prefer an Extended MED world over a Limited MED world.

To make this argument convincing, we must fill in details of (5) and specify the good that religious diversity brings to the table. Otherwise, the truth of (7) will be difficult to ascertain. Of course, one suggestion is that religious diversity provides individuals with the opportunity to respect and tolerate one another. That is, religious diversity makes available certain moral goods, namely the good of exercising moral autonomy in how one treats others with differing religious viewpoints. But this suggestion falls prey to the first criticism levelled against the good of proselytization. People would still have ample opportunity to exercise their moral autonomy even if religious diversity were absent. So (7) is plausibly false, as the value of eliminating the demographic problem seems much greater than the value of granting humans an additional moral choice.

But perhaps this reply is not totally satisfactory. It could, for example, be charged with resulting in a slippery slope. To illustrate, let us assume that apart from religious diversity, there are n number of opportunities for a person to exercise their moral autonomy (which we can represent as A₁, A₂, A₃, and so on). Of course, God need not permit religious diversity to allow A₁–A₃ . . . . But likewise, God need not permit any particular opportunity (such as A₁ or A₃) either, as there will always be other opportunities (such as A₂ or A₄) available. The worry then is that by reasoning in this way, we tacitly commit ourselves to the view that God would grant a person a very minimum number of opportunities to exercise their moral autonomy. And this is problematic if we think God would grant a person many opportunities to exercise their moral freedom.

However, this concern is misplaced. Let us represent opportunities to exercise moral autonomy which arise from religious diversity as R₁, R₂, R₃, and so on. Let us also represent opportunities to exercise moral autonomy which arise in the absence of religious diversity as A₁, A₂, A₃, and so on. Assuming God desires to grant humans an abundance of opportunities to exercise their moral autonomy, God has prima facie reason not to eliminate opportunities in either set, as this would diminish the number of opportunities for humans to exercise moral freedom. However, each opportunity in R₁–R₃ . . . has as its logical entailment the demographic problem, whereas no such entailment relation holds for opportunities in A₁–A₃ . . . . As such, while there is significant motivation for God to eliminate opportunities in R₁–R₃ . . ., no parallel motivation holds for opportunities in A₁–A₃ . . . . So the response above does not entail a slippery slope or a commitment to the view that God would only grant humans a very minimum number of opportunities to exercise their autonomy.

Regardless, even if (7) is true, another formidable problem remains with (6). Turning our attention to affairs in the actual world, we see that major religious traditions often exhibit an array of internal diversity and disagreement. In fact, disagreement between the same type of religious adherents is characterized by such polarization at times as to result in severe violence. We need only be reminded of, for example, the long-standing bloodbath between Christians in Northern Ireland in the twentieth century. With this in mind, even if religious believers were universally united under the conception of a personal and loving God in a Limited MED world, so that the existence of other major religions like Hinduism, Buddhism, Taoism, etc., were excluded, it does not follow that the exercising of religious tolerance would be precluded. For I see no reason why significant disagreement could not or would not arise within one major religious tradition in such a world, much like it arises in traditions like Christianity in the actual world. Thus, inhabitants of a Limited MED world would have to tolerate and respect internal diversity and disagreement that would be likely to arise within their tradition. So (6) is false.

Another way the good could be specified in (5) is in line with biodiversity. Similar to biological biodiversity which aids in the adaption and survival of a species, religious diversity might be the key to religion's survival against ever-changing historical and cultural tides. In the words of Victoria Harrison:

Each of these religious families [Judaism, Hinduism, Christianity, etc.] contains, embedded within its sacred texts and liturgical traditions, a diversity of conceptions of God. Different times and places give rise to different needs on the part of religious adherents and these can influence which conceptions of God are brought to the fore (God as Father/Mother/Friend/Teacher/Lover, to skim the surface of possible examples). At this level it seems advantageous to have a reservoir of diverse religious conceptions at the disposal of the faithful. If just one conception of God prevailed, then faith in God might lack resilience if times changed to the extent that this conception was no longer viable. (Harrison 2013, 488)

Pace Harrison, even if diverse conceptions of God offer some benefit on this front, what is easily overlooked is how this diversity is simultaneously detrimental to some religions. For one, many authors view religious diversity as an obstacle to religious knowledge and a justification for religious scepticism.²⁵ For another, diversity helps give rise to the problem under discussion, namely divine hiddenness, and therefore itself constitutes a reason for some theists to abandon their worldview. So (7) is questionable, as the diversity of religion may not contribute to the ‘survivability of religion’ to the extent Harrison indicates.

But more importantly, there is no reason to think various aspects of the traditional God of theism – a being infinitely rich in depth and beautiful in every conceivable respect – could not be emphasized by different historical people groups in a Limited MED world. These aspects of God could play a similar role that Harrison suggests diverse conceptions of God play in the actual world. While these aspects would all fall under the umbrella of a loving and personal God, some groups might emphasize God more as a father figure, others a mother figure, some a friend figure, and still others a teacher or lover figure, just to name a few. This variety could therefore cater to the tastes of different historical people groups, thereby promoting the ‘survival of religion’. Granted, some conceptions of God may not exist in a Limited MED world, such as impersonal or polytheistic permutations. But why should this be undesirable, when these conceptions remain an obstacle to consciously meaningful and reciprocal relationship with God? (6) is therefore implausible.

However, what if the value of religious diversity lies not in providing opportunities to respect those different from us or in facilitating biodiversity, but in its aesthetic beauty. Few would deny, for example, that the countless expressions of art, ritual, and ways of life a diverse world of religions gives rise to is wonderous and beautiful. And some authors argue that God has good reason to make beautiful things, by virtue of their intrinsic and instrumental goodness.²⁶ Might the beauty of religious diversity be a sufficient reason to permit the demographic problem?

Of course, we might follow Maitzen (2006, 183) and grant that religious diversity is beautiful but contend that no theistic explanation is forthcoming regarding why this type of beauty fails to flourish within certain cultures, such as Saudi Arabia or Thailand. But I suspect that when we try to sharpen this complaint up, it is equivalent to the claim that God must maximize religious diversity and beauty. And such a claim is highly contentious if it has any coherence at all.

Instead, a better tack is highlighting the falsity of (6). Of course, in a Limited MED world, the diversity of religions we see in the actual world would probably be eliminated. But even if this diversity and its accompanying beauty were eliminated, other forms of religious diversity might exist instead. It seems that many diverse forms of religious art, rituals, and ways of life could easily sprout up in a Limited MED world given cultural diversity and reflection on the infinite depth and nature of the God of theism. Hence, a Limited MED world could exhibit great religious beauty without precluding anyone from participating in a meaningful and reciprocal relationship with God.

Finally, it might be argued that the value of religious diversity lies in it making human cooperation necessary when investigating metaphysical questions about ultimate reality, including the question of whether God exists. In an Extended MED world, it is good that humans should have need of each other to investigate such questions. But in a Limited MED world, there would be no need for this cooperation, as God's existence would already be apparent.

Nonetheless, even if it were unnecessary for people collaboratively to settle the question of God's existence, there would very likely remain much need for collaboration on questions about God's nature and any related revelatory claims. Take, for example, the massive literature that has arisen in philosophical theology over the precise nature of God's attributes, such as God's omnipotence. In a Limited MED world, discussion of such topics would also probably take place. And if so, there would be great need for cooperation between individuals to settle these matters. Moreover, even if everyone believed in God's existence, there could remain a paucity of evidence for any revelatory claims (i.e. God being revealed through certain persons), and hence, cooperation needed to sort these claims out too. Again, if the revelation is particular and not universal, God might need to balance this against preventing too much confusion about God's own nature, commandments, or desires. But surely an omnipotent and omniscient being would be up to the task if any person is. So (6) seems false. So much then for Baker-Hytch appealing to the value of religious diversity.

Conclusion

The demographic problem of divine hiddenness deserves more attention than it has received. My aim in this article has been to show why an appeal to MED fails as a response to this argument. While I have not covered every conceivable way an appeal to MED might be salvaged, I have addressed those avenues which I believe are most salient. Moreover, other objections might be levelled against the demographic argument besides the one I have pursued here.²⁷ Unfortunately, entertaining these would take me well beyond the scope of this article.