Introduction
This paper is concerned with J. L. Schellenberg's argument from divine hiddenness. The basic idea of this argument is that there exists non-belief which is incompatible with the existence of God, and by ‘God’, I mean a being with at least the four properties of being all-good, all-powerful, all-knowing, and all-loving. (Perhaps a less ambiguous name for this argument would be ‘the atheistic argument from non-belief (in theism)’.) In this paper, I will examine the relationship between that argument and the value of divine–creature relationships. Given the apparent great value of divine–creature relationships, it is natural to wonder whether an awareness of that value can motivate a certain premise in Schellenberg's argument. This paper shows that the relevant premise is not supported by the value of divine–creature relationships, even if it is supported by other considerations.
I make one important assumption in this paper, namely that God would create a world in which a significant number of free creatures have the sort of free will that we take ourselves to have. In particular, He would create creatures that have libertarian free will concerning a significant range of choices including, at the very least, the choice between doing good or evil and the choice of whom (God included) to love. Of course, we don't have complete control over whom we love; we can't instantaneously choose to love someone, for example. Nonetheless, we can make free choices which make us morally responsible for loving or not loving that person.
Schellenberg's defence of the argument from divine hiddenness proceeds as far as possible granting this assumption,Footnote 1 but he nonetheless rejects it. In fact, I think one of Schellenberg's most interesting arguments, namely the free-will offence, pushes the idea that God would not create creatures with the sort of free will we take ourselves to have.Footnote 2 There are three reasons I make this assumption, despite Schellenberg's opposition. First, it is desirable for Schellenberg to have independent arguments for atheism. But if the divine hiddeness argument can't succeed without appealing to the free-will offence, then these arguments are not independent. Second, this assumption is so widely held, it is interesting to see if Schellenberg's argument from hiddenness can grant it and still succeed. I think the first two reasons partly explain why Schellenberg grants this assumption as much as possible. Finally, a satisfactory reply to Schellenberg's free-will offence and the arguments that drive it would require an additional paper.
Schellenberg's argument from divine hiddenness
The first and most controversial premise of Schellenberg's argument from divine hiddenness is:
(DH1) Necessarily, if God exists, anyone who is (i) not resisting God, and (ii) capable of meaningful conscious relationship with God, is also (iii) in a position to participate in such relationship (able to do so just by trying).Footnote 3
Let's say that someone is in the relating position iff she is in position to participate in a meaningful and conscious relationship with God just by trying to do so. The rough idea of DH1, then, is that, necessarily, if God exists, then every creature with the cognitive and affective capacity for a meaningful and explicit relationship with God will be in the relating position unless she resists that position in some way or another.
(DH2) Necessarily, one is at a time in a position to participate in meaningful conscious relationship with God only if at that time one believes that God exists.
DH2 follows largely from how Schellenberg wants ‘conscious relationship’ to be understood. A conscious relationship with God in part just is a relationship in which one is aware that she is relating to God. Several objections to Schellenberg's hiddenness argument appeal to weaker kinds of meaningful relationship that needn't be conscious;Footnote 4 however, the considerations I adduce below avoid this move. At any rate, the conjunction of DH1 and DH2 entail:
(DH3) Necessarily, if God exists, anyone who is (i) not resisting God and (ii) capable of meaningful conscious relationship with God (iii) also believes that God exists.
DH3 maintains that if we can find one capable and non-resistant person who does not believe that God exists, then God does not exist. The next premise affirms that there have been such people:
(DH4) There are (and often have been) people who are (i) not resisting God and (ii) capable of meaningful conscious relationship with God without also (iii) believing that God exists.
The conjunction of DH3 and DH4 entails atheism:
(DH5) God does not exist.Footnote 5
Some of what Schellenberg says suggests that the value of divine–creature relationships (perhaps together with certain justified background assumptions) can provide an independent justification for DH1.Footnote 6 But, even if Schellenberg doesn't think that the value of such relationships can justify DH1 by itself, it is still an independently interesting question whether it can. Let us call any argument that reasons from the value of divine–creature relationships to DH1 a version of the value argument. Then, this paper is concerned with whether there is a successful version of the value argument, such that it can constitute an independent justification for DH1.
Before we consider the structure of any plausible version of the value argument, we need to have a basic understanding of why divine–creature relationships would be so valuable. First, a personal relationship with God would itself be intrinsically good, that is good in itself.Footnote 7 Second, a personal relationship with God would be extrinsically valuable, that is, it would help bring about other intrinsic goods. For example, relating personally with God would greatly enhance human wellbeing. It would do so at least ethically: God would seemingly have a positive moral influence on human beings and this relationship would provide resources for both dealing with moral weakness and having improved relationships with other human beings.Footnote 8
We can conclude, then, that divine–creature relationships would be very valuable. Yet how might their great value justify DH1? Exactly what line of reasoning might take us from their value to DH1? Consider:
The sufficient reason thesis (SRT) Necessarily, if God exists, the value of possible divine–creature relationships would always provide God with sufficient reason to put all capable subjects in the relating position at all times at which they are capable, except insofar as they are culpable for their contrary position.
SRT entails DH1, so if we can find a good reason for SRT, then we have a good reason for DH1. Apparently, the most plausible way for the value argument to succeed is to argue from the value of divine–creature relationships to SRT and then from SRT to DH1. Since SRT entails DH1, the latter part of the argument is unproblematic. On the other hand, the details of the former argument aren't clear. But, however the sub-argument for SRT goes, it must provide an independent justification for SRT, for otherwise the value argument could not provide an independent justification for DH1. Thus, the potential for the value of divine–creature relationships to justify DH1 depends crucially on this assumption:
The independent justification assumption (IJA) An appreciation of the great value of divine–creature relationships, perhaps in conjunction with certain justified background assumptions, can constitute an independent justification for SRT.
Two things are worth discussing concerning the background assumptions to which IJA refers.
First, it is worth stressing the importance of the background assumptions being justified. Unless we have a good reason to believe the necessary background information, the argument for SRT (and so the whole value argument) won't be capable of justifying its conclusion. Second, we need to clarify exactly which background assumptions IJA is talking about. That is hard to say precisely, but we can put some conditions on which assumptions are acceptable in this context. Since this paper is attempting to determine whether the value argument can justify the first premise independently of the free-will offence, any key premise or conclusion from that argument seems off limits. Also, if the value argument is to motivate Schellenberg's argument from hiddenness independently of arguments from evil, then any key premises or conclusions needed for those arguments are also off limits. Furthermore, the relevant background information can't assume that SRT is justified already. If the value argument for SRT is successful only if we already have an independent justification for SRT, then it might be able to enhance SRT's justification, but it won't be able to make it justified for the first time, that is, it won't constitute an independent justification. Still, there are a wide variety of claims which might provide acceptable background assumptions.
We haven't yet actually identified (a version of) the value argument, but we at least have a general outline. The argument first proceeds from the value of divine–creature relationships to SRT and then from SRT to DH1. The argument for SRT, whatever it is, can succeed only if the independent justification assumption is true. Unfortunately, (any version of) the value argument fails because IJA is false: the inference from the value of divine–creature relationships to SRT depends for its success on background assumptions which we have no reason to believe (at least not independently of arguments we are ignoring in this paper, such as the free-will offence). The next section, ‘Value and God's knowledge’, sets the stage for the main presentation of my argument which I provide in the third and fourth sections, namely ‘The presumptuous version’ and ‘The modest version’, respectively. In the fifth section, ‘God's love and His “calculations”’, I respond to a Schellenbergian objection to the material in the preceding three sections.
Value and God's knowledge
Intuitively, God's knowledge affects how much reason is provided by the value of a divine–creature relationship (for God to put some person in the relating position). Indeed, it seems that something like the following is true: other things being equal, the more certain God is that His efforts will not attain some good, the less reason that good provides God to do what He can to attain it. (Note that this principle is compatible with God's being absolutely certain of the truth-value of every proposition.) To see this point, let's consider an analogy between human and divine reasoning.
Suppose that there is a lottery which will pay $1,000,000 to the winner. The value of this money gives us reason to buy a ticket even if the cost of the ticket provides a far greater reason not to purchase it. Yet the amount of the prize is not the only factor in determining how much reason it provides us to buy the ticket; the odds of winning the lottery are also important. The value of this money will provide more reason to buy this ticket if the odds are 1 in 2 instead of one in 10 million. This explains why we would be willing to sacrifice more to acquire a ticket when the odds are the former. Similar considerations apply to God. Suppose that, were Mary to relate to God from t 1 to t 2, her relationship would be immensely valuable. If God knows that, were He to put Mary in the relating position, there is a 1 in 2 chance that Mary relates to Him, then the value of that relationship would provide God with strong reason to put her in the relevant position; however, if the odds were 1 in 10 million, the immense value of the relationship would provide God with very little reason to put her in that position. Hence, other things being equal, the more certain God is that His efforts will not attain some good, the less reason it provides God to do what He can to attain it.
Some philosophers have objected to my analogy by relying on the resources of decision theory. Decision theory tells us that, in situations of this kind, the rational agent will choose the action with the highest expected utility/value. To calculate the expected value of some action, such as putting someone in the relating position, we must (i) identify the various amounts of (dis)value that the action might result in; (ii) multiply each of the amounts from (i) with the probability that the action would yield that value; (iii) add the products from (ii); and (iv) subtract any costs in performing the action. Hence, when an agent is deliberating over whether to perform some action the cost of that action is a relevant factor. The objector tries to take advantage of this fact by pointing out an alleged disanalogy: playing the lottery costs us something, but it isn't clear that there is a cost for God to put some person in the relating position.
This objection is irrelevant. The cost of performing an action is relevant to determining whether to perform that action; however, it is not relevant for determining how much reason some valuable thing provides to perform it. To see this, maybe it will help to have a formal procedure for determining how much reason the value of one's potential divine–creature relationship gives God to put her in the relating position. This following is a plausible candidate for such a procedure: (i*) identify the possible (dis)values that one's divine–creature relationship could have were God to put her in the relating position; (ii*) multiply each of the values from (i*) by the probability that the action would have that value; and (iii*) add the products from (ii*). This procedure has two important differences from the procedure for determining the expected value of some action: this procedure ignores both the cost of the action and all value the action could result in, except the value of the divine–creature relationships. Hence, the objection fails and my analogy supports the claim that, other things being equal, the more certain God is that His efforts will not attain some good, such as a relationship with some person, the less reason it provides God to do what He can to attain it.
More specifically, it seems correct to say that, if God is knows that He cannot get some good thing no matter what He does, the value of that thing provides God with no reason to attain it.Footnote 9 More formally and applied to our present concerns, the idea is that:
(PV1) Necessarily, if, at some time prior to t, t -1, God knows that from t to t 1 S would not have a relationship with God (even if S were in the relating position), then the value such a relationship would accrue (during that period of time) gives God no reason to put S in the relating position from t to t 1.
(The rationale for the name ‘PV1’ will become clear in the next section.)
Some may believe that PV1 is true, but unhelpful for our present concerns. For these philosophers believe that God doesn't know the counterfactuals of freedom for some reason or another. They will therefore say something like: ‘Yes, if God knew that some particular person would not have a relationship with Him, then the potential value of that possible relationship would not provide God with any reason to put the person in the relating position. But God never knows things of the sort – or at the very least he only knows them in exceptional cases – because he doesn't know the counterfactuals of freedom.’ This point is well taken, and PV1 won't be very useful to those philosophers who believe (mistakenly in my opinion) that God doesn't know the counterfactuals of freedom.
Nonetheless, even if God doesn't know the counterfactuals of freedom, God's probabilistic judgements will still affect how much reason He has to put a person in the relating position. And it seems that when God judges that, say, there is a.3 probability that a person will relate personally to Him, then He knows that there is an objective.3 probability that the person will do so. We, therefore, have the resources for formulating a similar principle whose utility does not depend on the supposition that God knows the counterfactuals of freedom:
(MV1) Necessarily, if, at some time prior to t, t -1, God knows that from t to t 1 S in all likelihood would not have a relationship with God (even if S were in the relating position), then the value that relationship would accrue (during that time) gives God some, but not much reason to put S in the relating position from t to t 1.
(The rationale for the name ‘MV1’ will become clear in the next section.) I should shed some light on what I mean by ‘in all likelihood’, and ‘some, but not much reason’. For our purposes, we could tentatively stipulate that when God knows that x in all likelihood will not occur, then God knows that there is a probability equal to or less than.01 that x will occur.
But what does it mean for something to give God ‘some, but not much reason’ to do something? If x provides God with some but not much reason to do y, then that reason will be outweighed if either God is sufficiently confident that He can acquire a weak good that is incompatible with doing y or avoid a weak evil that He is sufficiently confident will result if He does y. (In the rest of this section, I focus on how weaker goods can outweigh stronger ones, but similar considerations apply to how weaker evils can outweigh stronger goods.) If God knows that some person in all likelihood will not have a relationship with Him, then God is in a position analogous to someone who knows that, in all likelihood, he will not win the $1,000 lottery. Assuming money is the only relevant good and that the expected value of not buying a lottery ticket is $0, then the person should not play the lottery if the ticket costs over $10.Footnote 10 Even though $1,000 is nearly 100 times more value than $11, the latter would outweigh the former. Likewise, even though some person's relationship with God would be a relatively great good, some relatively weak good which is incompatible with God's putting the person in the relating position, can provide offsetting or outweighing reason not to put that person in the relevant position. (Though, of course, for a relatively weak good to provide such reasons, God must be much more certain He will attain it than He is that He will attain the relationship.)
The presumptuous version
In this section, I begin my attack on the independent justification assumption (IJA). My attack comes in two versions, the presumptuous and modest versions. The former begins by considering PV1 and the latter begins by considering MV1. The presumptuous version gets its name from its presumption that God knows the counterfactuals of freedom. I present two different versions of the argument because each has a virtue the other doesn't. The presumptuous version seems to be simpler and more decisive, but the modest version can be accepted by those who reject Molinism (the view that God, if He exists, knows the counterfactuals of freedom). In this section, I present and defend the presumptuous version. I turn to the modest version in the next section.
Preliminaries
As I already mentioned, the presumptuous version of the argument begins with:
(PV1) Necessarily, if, at some time prior to t, t 1, God knows that from t to t 1 S would not have a relationship with God (even if S were in the relating position), then the value such a relationship would accrue (during that period of time) gives God no reason to put S in the relating position from t to t 1.
To see that the value argument's sub-argument for SRT fails, it will be helpful to bring PV1 into discussion with SRT. To do so, we need to make two changes to PV1. Notice that PV1 states that, in certain circumstances, the value of some relationship during some period of time provides God with no reason to put the person in the relevant position. SRT, on the other hand, is concerned with whether there is sufficient reason to put the person in the relevant position. The transition from ‘gives God no reason’ to ‘does not give God sufficient reason’ is easy, as it is obvious that, necessarily, if x gives God no reason to do y, x cannot give God sufficient reason to do y.
Once PV1 is modified in the above way, it is concerned with whether a relationship with S from t to t 1 would provide sufficient reason for putting S in the relating position during t to t 1. But SRT's concern is more general: it is whether the value of any subject or subjects' relationship during any period of time provides God with sufficient reason for God to put S in the relating position from t to t 1. As Schellenberg himself points out, the value of S's relationship at some later time might give God reason to put S in the relating position now.Footnote 11 It may be that S would ultimately have a better relationship than he otherwise would have had were he to be in the relating position now,Footnote 12 despite the fact that he wouldn't relate personally with God until much later. It also might be that putting S in the relating position now would benefit someone else's relationship with God, despite the fact that S resists belief in God. It is important to stress, though, that whether S's being put in the relating position for some period of time would benefit someone else or benefit S at some later time depends crucially on what the counterfactuals of freedom are. Just as being put in the relating position now might leave S more open to relationship later, his continued resistance may make his heart harder than it otherwise would have been.
There are then two ways that the value of divine–creature relationships might give God reason to put S in the relating position during t to t 1. First, putting S in the relating position would make possible relationship with God during that period of time. Second, it might in some other way benefit or promote some person's relationship with God. These seem to be the only two ways in which the value of divine–human relationships might give God reason to put someone in the relating position during some period of time. Hence, if God knows both that S would not relate to God during the relevant period of time and that putting him in the relating position will not otherwise benefit anyone's relationship with God, then the value of divine–creature relationships does not provide God with any, much less sufficient, reason to put S in the relating position from t to t 1. But SRT requires that the value of divine–creature relationships provide God with sufficient reason to put all capable subjects at all times at which they are capable in the relating position. Hence:
(PV2) SRT is true only if this conjunction is impossible (given the counterfactuals of freedomFootnote 13): there is a subject who, for some period of time, would fail to relate to God were she in the relating position during that period of time,Footnote 14 and that subject's being put in the relating position during that period of time would not otherwise benefit someone's relationship with God.
If we can show that it is possible, given the counterfactuals of freedom, for that conjunction to be true, then we will have shown that SRT is false, which would suffice to show that the value argument fails.
One might think that we can show, not only that it is possible for this conjunction to be true, but also that it is actually true. Many people in this world are such that, for some period of time, they are ‘too busy’ with their own agenda to care about God. There seem to be many people from the history of the world that would prioritize something else over their relationship with God (for some period of time or another). Apparently, many religious traditions are sensitive to this fact, because they bemoan the human tendency for idolatry, that is, the human tendency for in some way worshipping, prioritizing, or valuing created things more than God. And it seems unlikely that for every such person in the actual world, that putting her in the relating position would in some way benefit someone's relationship with God. One might think, then, that we have good reason to believe that the conjunction from PV2 obtains in the actual world.
But Schellenberg will insist that this way of securing the conjunction from PV2 constitutes cheating. The above argument for that conjunction simply assumes that God would create people like those of the actual world, but that assumption is illegitimate in this context. In order to secure that conjunction, we need to show that God would create a person who would resist belief in Him, such that putting that person in the relating person wouldn't benefit someone's relationship with God. Noting that the actual world contains such people doesn't show that God would create them. Yet I don't know how to secure the relevant conjunction without assuming that God exists and so begging the question against Schellenberg. On the other hand, the value argument does not give us any reason to deny it.
Why would the conjunction from PV2 be impossible? Assuming, as we are, that God would create a significant number of creatures with the sort of freedom we take ourselves to have, the only reason I can think of is this: the counterfactuals of freedom work out so that the best (i.e. creation-worthy) feasible worldsFootnote 15 contain a significant number of free creatures who never resist belief in God, such that their being in the relating position wouldn't benefit someone's relationship with God. Since God is good and desires relationships with his free creatures, He would create one of these better worlds rather than creating a world where some free creatures resists belief.Footnote 16
The main argument
We are now in position to show that independent justification assumption (the value of divine–creature relationships in conjunction with justified background assumptions can provide an independent justification for SRT) is false. PV2 tells us that SRT is true only if it is impossible (given the counterfactuals of freedom) that there be a subject who would resist belief in God, such that his being in the relating position would fail to benefit or promote anyone's relationship with God. But this impossibility would obtain only if the conjunction is never true in the best feasible worlds. Thus, the truth of SRT depends on the counterfactuals of freedom working out in a very particular way.
More importantly, that the counterfactuals of freedom work out this way is an assumption that needs to be justified in order for IJA to be true. In other words, the value argument can succeed only if we have some reason to believe that the best feasible worlds never contain a subject who would resist belief in God, such that his being put in the relating position wouldn't benefit anyone's relationship with Him. Remember, the value of divine–creature relationships doesn't automatically provide God with any reason to put some person in the relating position. In particular, it won't do so when the relevant conjunction is true of that person. Hence, if we have no reason to believe that the conjunction fails to be true of some person, it is hard to see how we can know that the value of divine–creature relationships provides sufficient reason to put that person in the relating position. An appreciation of the value of divine–creature relationships, therefore, can justify SRT (the claim that the value of divine–human relationships provides sufficient reason to put all non-resistant capable people in the relating position at all times at which they are capable) only if we have some reason to believe that the conjunction is never true in the best feasible worlds, the sort of worlds God might actualize.
Although we need some reason to believe that the best feasible worlds never make the relevant conjunction true, we can't get this reason without undermining the independence of the value argument. First, the value of divine–creature relationships doesn't seem to provide any insight into whether the best feasible worlds work out that way. Second, it is very hard to see how we can know they work out that way unless we already have some other reason to believe SRT. Hence, the value of divine–creature relationships can't provide an independent justification for SRT, i.e. the independent justification assumption is false.
We can summarize the argument from this subsection as follows:
(PV3) In order for IJA to be true, we need to have good reason to believe that the conjunction from PV2 is never true in the best feasible worlds, and we need to have this reason without relying on arguments that are off limits (e.g. the free-will offence). But we can't get such a reason. Therefore, IJA is false.
The modest version
Preliminaries
Now that I have presented the presumptuous version, it is time to consider the modest version, the version that does not assume that God knows the counterfactuals of freedom. Earlier, we secured:
(MV1) Necessarily, if, at some time prior to t, t -1, God knows that from t to t 1 S in all likelihood would not have a relationship with God (even if S were in the relating position), then the value that relationship would accrue (during that time) gives God some, but not much reason to put S in the relating position from t to t 1.
To see that Schellenberg's argument for SRT fails, it will be helpful to bring MV1 into discussion with SRT. To do so, we need to make two changes to MV1. Notice that MV1 states that, in certain circumstances, the value of some relationship during some period of time provides God with some but not much reason to put S in the relevant position. SRT, on the other hand, is concerned with whether there is sufficient reason to put the person in the relevant position. Recall the definition of ‘some but not much reason’: if x provides God with some but not much reason to do y, then that reason will be outweighed if either (i) God is sufficiently confident He can acquire a weak good that is incompatible with doing y, or (ii) avoid a weak evil that He is sufficiently confident will result if He does y. We can make our forthcoming discussion less wordy if we replace (i) with ‘there is a competing good’ and (ii) with ‘there is a competing evil’. Since goods and evils that compete with x (if they are present) outweigh the reason x provides for y, x will not provide a sufficient reason to do y when such goods and evils are present. Hence, the value of some relationship during some period of time will not provide sufficient reason for God to put S in the relating position during that time if in all likelihood S will not relate to God and there are competing goods or evils.
Once MV1 is modified in terms of ‘sufficient reason’ instead of ‘some but not much reason’, it is concerned with whether a relationship with S from t to t 1 would provide sufficient reason for putting S in the relating position during t to t 1. But SRT's concern is more general: it is whether the value of any subject or subjects' relationship during any period of time provides God with sufficient reason for God to put S in the relating position from t to t 1. As mentioned in the last section, the value of someone else's relationship or the value of S's relationship at some later period of time might give God reason to put S in the relating position now. We should remember, though, that whether S's being put in the relating position for some period of time would benefit someone else's relationship or S's relationship at some later time depends crucially on what the counterfactuals of freedom are.
There are two ways, then, that the value of divine–creature relationships might give God reason to put S in the relating position during t to t 1. First, putting S in the relating position would make possible relationship with God during that period of time. Second, it might in some other way benefit or promote some person's relationship with God. These seem to be the only two ways in which the value of divine–human relationships might give God reason to put someone in the relating position during some period of time. Thus, if God knows that S would likely not relate to God during the relevant period of time, that there is no competing good or evil, and that it will not otherwise benefit anyone's relationship with God,Footnote 17 then the value of divine–creature relationships does not provide God with sufficient reason to put S in the relating position from t to t 1. But SRT requires that the value of divine–creature relationships provide God with sufficient reason to put all capable subjects at all times at which they are capable in the relating position. Hence:
(MV2) SRT is true only if this conjunction is impossible (given the counterfactuals of freedom)Footnote 18: there is someone that in all likelihood would not have a relationship with God for some period of time, and there is a competing good or evil, and that S's being put in the relating position during that period of time would not otherwise benefit anyone's relationship with God.
If we can show that it is possible, given the counterfactuals of freedom, for that conjunction to be true, then we will have shown that SRT is false, which would suffice to show that the value argument fails. One might think that we can show not only that it this conjunction is possibly true, but also that it is actually true. Such a person will insist that we keep two things in mind. First, we don't need to identify a case in which God would fail to have sufficient reason for never putting some subject in the relating position; rather, all we need to do is identify a case in which God lacks sufficient reason for putting a person in the relating position for some period of time, no matter how short. Moreover, we need identify neither a good at least as valuable as a relationship with God (during the relevant period of time) nor an evil that is at least as bad as the relationship is good. As I argued earlier, if God is much more confident that he can obtain the weaker good, the value of a good relationship with God can be outweighed by a good almost 100 times (or more) less valuable (similar considerations apply to an evil that God might avoid).
One may think the actual world provides examples in which the value of the relationship fails to provide sufficient reason to put the person in the relating position. Some people claim that their experiences fit the following pattern. These people become very aware of their spiritual need, prompting them to discover its source. Yet, instead of immediately putting this person in the relating position (or instead of already having done so), God remains hidden. This person's spiritual journey then becomes an arduous struggle at the end of which God puts them in the relating position. After yearning for the source of their spiritual need (a relationship with God) for so long and having worked so hard to achieve it, the person appreciates and values the relationship more than she otherwise would have. As a result, she has a deeper relationship and a more positive influence on those around her. If we are to take people's description of their spiritual lives seriously, we might have reason to believe that there are goods that provide reason not to put some person in the relating position for some period of time.
Or consider a hypothetical fellow named Fernando. Suppose that God also knew that, if Fernando had a relationship with Him, he would most likely be severely persecuted by those around him, such that Fernando would either die very soon or would live with an oppressed and dismal quality of life. Given the great amount of religious persecution in the history of the world, it seems reasonable to believe that some people have been in situations analogous to Fernando, which means that there are evils that provide reason to refrain from putting someone in the relating position.
Examples like the above two suggest that there are a significant number of people from the history of the world, such that God has reason to refrain from putting them in the relating position. But then it seems likely that at least one person satisfies the conjunction from MV2. That is, it is likely that some person in all likelihood will resist relationship with God, there is a good or evil that competes with the value of that relationship, and her being put in the relating position does not benefit anyone's relationship with God. It may seem, then, that we have at least some reason to believe SRT is false.
Schellenberg will protest that such an argument against SRT (even in outline) falls prey to what he calls the accommodationist strategy. Roughly, this strategy argues that the alleged outweighing goods and evils can be accommodated within a divine–creature relationship, and, if so, then that good (evil) doesn't provide God with reason not to put a person in the relating position. Take the first case involving the spiritual seeker. We described this case so that the person had a deeper relationship because God failed to put her in the relating position for some period of time; however, those same benefits can be achieved within a divine–creature relationship by what is sometimes called a ‘dark night of the soul’. In such a case, the believer has sufficient evidence for belief in God, but the subject might still long and struggle for a deeper relationship with God because God ‘feels’ absent.Footnote 19 In the case of Fernando, God could have put him in the relating position and then prevented the persecution from taking place or in some other way preserved his wellbeing (e.g. by providing him with special comfort and strength, etc.). Hence, Fernando's wellbeing can be accommodated from within the divine–creature relationship.
What motivates this accommodationist strategy, in Schellenberg's mind, is God's greatness, i.e. His power, His knowledge, His resourcefulness, His ingenuity, etc. Says Schellenberg:
If God exists, then there must at any stage along the way be literally an infinite number of ways of developing in relationship with God that God could facilitate, despite obstacles to continuing relationship of the sort to which reasons for hiddeness often make reference [including the two I mentioned above] … . It follows, given that it is a necessary condition for a reason for God to remain hidden that it would not be thus capable of being accommodated within a relationship-conducive set of conditions, that theistic replies to the hiddenness argument of the sort we have considered do not provide us with such a reason.Footnote 20
For our purposes, the most important part of this passage is Schellenberg's claim that ‘it is a necessary condition for a reason for God to remain hidden that it would not thus be capable of being accommodated within a relationship-conducive set of conditions’. Call this claim the outweighing reason condition (ORC). We shall return to this claim shortly.
The main argument
We are now in position to show that the independent justification assumption (the value of divine–creature relationships in conjunction with justified background assumptions can provide an independent justification for SRT) is false. MV2 tells us that SRT is true only if it is impossible (given the counterfactuals of freedom) that there be a subject who would likely resist belief in God, such that there exists a competing good or evil, and his being in the relating position would fail to benefit or promote anyone's relationship with God. But this conjunction would be impossible (given the counterfactuals of freedom) only if the conjunction is never true in the best feasible worlds. Thus, the truth of SRT depends on the counterfactuals of freedom working out in a very particular way.
More importantly, the truth of the conjunction from MV2 is an assumption that needs to be justified in order for IJA to be true. For the sake of argument, let's assume that (a) the value of divine–creature relationships provides God with at least some but not much reason to put all capable subjects in the relating position (at all times in which they are capable). But the truth of this claim (if it is true) does not automatically license an inference to (b) the value of divine–creature relationships provides God with sufficient reason to do so. To make this inference appropriate we need some reason to believe that (c) the reason provided by the value of divine–creature relationships is never outweighed by something else. To have this reason, we need to know whether the conjunction from MV2 obtains in the best feasible worlds. For if that conjunction does obtain, then there will be competing goods or evils that outweigh the reason provided by divine–creature relationships.Footnote 21
For the value argument to be successful, then, we need to have some reason to believe that the best feasible words never make the conjunction from MV2 true. I suppose it might be such that we can find some completely independent justification for SRT and then infer that the relevant conjunction never obtains in creation-worthy worlds; however, that would be irrelevant for our present concerns, because we are attempting to determine whether IJA is true, i.e. whether an appreciation of the value of divine–creature relationships can constitute an independent justification for SRT. But, under these constraints, there is no way of knowing whether the relevant conjunction fails to be true in any of the best feasible worlds.
Notice that it seems dubious that our appreciating the great value of divine–creature relationships would give us genuine insight into whether the counterfactuals of freedom work out so that the conjunction from MV2 is never true in the best feasible worlds. To find the required insight, we need to look elsewhere.
Recall that one conjunct of the relevant conjunction is that there is some competing good or evil. Schellenberg may think that the outweighing reason condition (i.e. X is a reason to refrain from putting S in the relating position only if X cannot be accommodated within a divine–creature relationship) provides the resources to show that this conjunct (and so the whole conjunction) never obtains in the best feasible worlds. Yet it is hard to see how we can justifiably believe ORC under the constraints imposed by IJA. Just as we can't know that the conjunction from MV2 never obtains in the best feasible worlds without relying on something off limits (e.g. the free-will offence), we can't know that ORC is true without relying on something that is off limits. Let me explain.
Notice that the mere possibility of accommodation does not disqualify something as a reason for God to remain hidden. Suppose that the only way(s) God can put Fernando in the relating position and preserve his wellbeing (for the relevant period of time) would make the world overall worse, even if it makes Fernando's relationship more valuable. In such a case, the possibility of accommodation would not prevent Fernando's wellbeing from being a strong reason not to put him in the relating position. In general, in order for the possibility of accommodation to disqualify something as a reason for not putting someone in the relating position, it must also be that accommodating that thing would lead to an overall better world.Footnote 22 Yet whether accommodation of some good or evil would lead to an overall better world depends on what the counterfactuals of freedom are.Footnote 23 God's accommodating some good or evil, whether it directly affects some agent or not, might nonetheless lead to free choices that make world overall worse, and so God would prevent those world-worsening choices by refusing to accommodate that good or evil.
ORC says that something can be a reason to refrain from putting some person in the relating position only if that thing can't be accommodated from within a divine–creature relationship. What we have seen is that accommodation by itself doesn't prevent something from constituting such a reason. In order for ORC to be true, then, the possibility of accommodation must, in some sense, guarantee that the world would be better than if that thing were not accommodated. Yet it is hard to see how we can know that without cheating. Therefore, this hypothetical proposal fails to show that the conjunction from MV2 never obtains.Footnote 24 Apparently, then, the value of divine–creature relationships can't provide an independent justification for SRT, i.e. the independent justification assumption is false.
We can summarize the argument of this subsection as follows:
(MV3) In order for IJA to be true, we need to have good reason to believe that the conjunction from MV2 is never true in the best feasible worlds, and we need to have this reason without relying on arguments that are off limits (e.g. the free-will offence). But we can't get such a reason. Therefore, IJA is false.
God's love and His ‘calculations’
My arguments in this paper suggest a certain picture of God that Schellenberg finds problematic. I am picturing God in such a way that He constantly evaluates, for every person, whether it is good or best to put her in the relating position, that is, a position to relate personally with Him just by choosing. Schellenberg protests that
The best human love – the best love of parent or spouse or friend – involves seeking meaningful personal relationship with the beloved … . The lover clearly wants this for its own sake – she reaches out to the one she loves naturally and spontaneously and without depending on some calculation of advantages and disadvantages for either party.'Footnote 25
The analogy to divine love is supposed to show that it is somehow inherently inappropriate for God to engage in those calculations: God's perfect love is incompatible or in tension with this behaviour. Hence, it is misguided to picture God in this way.
I admit that it is inappropriate for human beings to engage in constant calculations of this sort. Take the mother–child relationship for example. It would seem quite odd for a good mother to constantly wonder whether, despite her love, she has outweighing reasons to not make relationship available for her child. A big reason why it would be inappropriate for mothers to do this is that it would require tremendous time and resources to do so. Second, even if, say, a mother's relating with her child would set in motion a string of events that would have terrible consequences, it is highly unlikely that she would know it. (How could Hitler's mother know that taking good care of her son would ultimately lead to so much evil?) For humans, constant calculations are time- and resource-demanding and aren't likely to identify outweighing reasons even if they are present. Hence, the best use of the mother's resources (assuming she has sufficient resources to care for her child) is to relate personally with the child and hope for the best.
Notice, though, that God's epistemic position is entirely different. For He doesn't expend any energy to make His constant ‘calculations’; He can't help having well justified beliefs about what it is good or best to do. Furthermore, God is likely (or at least far more likely) to identify outweighing reasons when they are present. The best strategy for God, then, is to take into account what He knows when He is choosing whether to put a person in the relating position. (In a way, this is the best strategy for mothers too; it's just that they almost never, if ever, have any reason to think they should remain hidden from their children.) After all, it would certainly be odd to demand that God put a person in the relevant position when He knows (and perhaps when He knows it is highly likely) that it would be better if He did not put the person in the relevant position.
In short, there is nothing inherently peculiar, odd, or unsatisfactory about claiming that God constantly ‘calculates’ whether it is best to make relationship possible. In fact, this constant ‘calculation’ contributes to God's being an ideal lover: His well-justified beliefs about what is good and best make Him ideally suited to care for His beloved. I therefore deny my arguments in this paper paint a poor picture of God. God's knowledge affects how much reason is provided by the value of divine–creature relationships and God's decisions about whether to remain hidden will be sensitive to this information.
Conclusion
In this paper, we have criticized the value argument for the first premise in Schellenberg's argument from divine hiddenness. The first stage of the value argument attempts to establish the sufficient reason thesis, i.e. necessarily, God has sufficient reason to put all capable, non-resistant subjects in the relating position at all times at which they are capable. In particular, it is supposed to establish SRT on the basis of the value of divine–creature relationships in such a way that the value of divine–creature relationships can provide an independent justification for SRT (and so DH1). Yet this sub-argument depends for its success on our having reason to believe that certain conjunctions are never true in creation-worthy, feasible worlds. The problem is that we can't know that those conjunctions don't obtain in the relevant worlds without compromising the independence of the value argument. Hence, the value argument cannot justify the first premise of Schellenberg's divine hiddenness argument. Perhaps there are good reasons to believe this premise, but the value of divine–creature relationships is not one of them.Footnote 26
