Impurity

Schellenberg's P1 can be challenged by considering a certain notion of impurity, one that is rooted in the Old Testament but refined in Paul's letters. Jonathan Klawans argues that biblical texts refer to two distinct types of impurity – ritual and moral (Klawans (Reference Klawans2000), 36–38). He also argues that ancient Jews and Israelites saw both types of impurity as defiling in a literal sense; references to the defiling force of impurity are not merely metaphorical (ibid., 32–36).

According to Klawans, ritual impurity and moral impurity are distinct kinds of impurity that function analogously (ibid., 38). They differ in that ritual impurity originates from contact with certain natural sources (e.g. scale disease), whereas moral impurity originates from a person's sinful action. They also differ in that moral impurity is directly associated with sin, whereas ritual impurity is not. However, the two kinds of impurity are analogous in that they both lead to the defilement of other entities beyond the person from whom they originate. Ritual impurity can defile other things that are physically contacted by the ritually impure person (e.g. holy foods). And moral impurity can defile ‘the land of Israel’ and ‘the sanctuary of God’ (ibid., 23–26). This article focuses strictly on moral impurity, leaving ritual impurity aside. The key point is that moral impurity can transfer to other entities beyond the sinner. Henceforth, all mention of impurity refers strictly to moral impurity, unless otherwise noted.

What other entities can impurity defile? Can other people be defiled by the sinner's impurity? According to Christine Hayes, ‘[m]oral impurity is not communicable to others’ anywhere in the Old Testament (Hayes (Reference Hayes2002), 23). But, as Hayes points out, the letters of Paul give us a first glimpse of the view that impurity can transfer from person to person: ‘Paul's new “carnal impurity” conflates the characteristic features of moral and ritual impurity: It is a moral impurity, and yet it resides in the flesh and may be communicated to others by (carnal) contact’ (ibid., 14).Footnote ³ The person-to-person transfer of impurity is crucial for this article, so let's pause on it.

As noted, Old Testament authors believe impurity defiles the sanctuary or temple. This is problematic because the temple is seen as the special dwelling place of God. But, according to Michael Newton, Paul transforms these ideas by identifying the temple with the community of believers. For example, writing to the Corinthians, Paul says that: ‘God's temple is holy, and you are that temple’ (1 Cor. 3:16–17). Here, Paul is identifying the temple with the people among whom God's spirit would dwell if they remain pure (Newton (Reference Newton1985), 54–58). Paul is greatly concerned to maintain the purity of this community-based temple (e.g. 1 Thess. 4:7). But how exactly does impurity affect the community?

Hayes interprets Paul as holding that impurity spreads from person to person via carnal (i.e. fleshly) relations. She supports this by focusing on Paul's conception of impurity within marriage. In 1 Corinthians 7:10, Paul explicitly prohibits divorce. Moreover, according to Hayes, Paul prohibits marriage between believers and impure unbelievers.Footnote ⁴ But these two prohibitions raise a question: Should a new convert who is already married to an unbeliever now seek divorce? Paul answers ‘no’:

Paul's rationale for advising against divorce in this situation is that the new convert makes the unbeliever holy. This is important, because holiness is conceptually incompatible with impurity (Hayes (Reference Hayes1999), 5). That is,

EP has the status of a conceptual truth, and it will be assumed throughout this article. Given EP, one can infer from 1 Corinthians 7:12–14 that the unbeliever, who was previously impure, is made pure through his or her relationship with the believing spouse. In other words, Paul thinks there is a transfer of purity from the new convert to the unbelieving spouse via their spousal relation (Hayes (Reference Hayes2002), 94–96).

In other passages Paul suggests that impurity can also transfer from person to person. For example, Paul frequently advises pure believers to separate from impure believers:

Immediately prior to this call to shun impure believers, Paul uses an analogy to convey the defiling effects of impurity within the community of believers:

The spread of yeast throughout dough is seen as a metaphor for the spread of impurity throughout a community (Hayes (Reference Hayes2002), 93; Liu (Reference Liu2013), 132). Paul is here exhorting the community to ‘clean out’ the impurity so that it does not infect the whole community.

The above passages reveal that Paul is cautious about various kinds of relationships with impure people. He recommends the avoidance of sexual and spousal relations with impure people (1 Cor. 6:15–20). He advises church members not to eat with, or associate with, impure believers (1 Cor. 5:11). And he claims that the children of two married unbelievers ‘would be unclean’ (1 Cor. 7:14), suggesting a concern for parental relations with children. Thus, Paul is cautious about at least five different relations – sexual relations, spousal relations, table fellowship, ‘associating’ or socializing, and parental relations. What unifies these?

The above biblical scholars do not answer this question. But here's a plausible suggestion: the above relations are personal relationships, in Schellenberg's sense – they are explicit, reciprocal, intimate, conscious, and meaningful relationships. Indeed, Schellenberg himself illustrates personal relationship by using similar examples – e.g. spousal, parental, sibling, and friendship relations (Schellenberg (Reference Schellenberg2007), 203).Footnote ⁶ And it is plausible to interpret Paul as warning against the same sort of relations with impure people because he thinks personal relations facilitate a transfer of impurity between persons.

At this point, readers might be baffled: what exactly is impurity? And how does it transfer between persons? As will be argued later, impurity partly consists in having passions that make one prone to sin; and one avenue that facilitates its transfer is the moral influence within personal relationships. These claims are substantiated below. For now, let's take stock on Paul's view.

Like his predecessors, Paul believes impurity can defile the temple. But, unlike his predecessors, he identifies the temple with the community of believers. And he thinks impurity can spread like yeast throughout the community by way of personal relations. These ideas can be summed up with a principle:

Paul's acceptance of PP explains why he advises believers to avoid specific forms of personal relationship with impure people. If true, PP would be a conceptual truth – a partial definition of Paul's concept of moral impurity. For Paul's predecessors, it was a conceptual truth that both ritual and moral impurity are defiling – they transfer via certain relations (e.g. contact) to certain relata (e.g. holy foods, the sanctuary). And Paul assimilates his predecessors’ concept of moral impurity by keeping its basic structure but specifying a new transfer relation (i.e. personal relationship) and new relata (i.e. persons).

As stated, PP universally quantifies over any y, including God. Should God be excluded from the domain over which PP quantifies? Not in Paul's thinking. If God were excluded from that domain then we would lose an explanation of an important fact about Paul's view. Yulin Liu (Reference Liu2013, 234), Hayes (Reference Hayes2002, 97), and Newton each interpret Paul as holding that purity is required for God's presence within a community. Quoting Newton,

But why does Paul believe an impure community cannot enjoy God's presence? If we allow that PP quantifies over a domain that includes God, then we have our explanation: Paul believes that an impure community cannot enjoy God's presence, because he accepts PP even when ‘God’ is substituted for ‘y’ and he believes that it is impossible for God to become impure, i.e. the consequent of PP is false whenever ‘God’ is substituted for ‘y’. These claims require him to believe that an impure community cannot be personally related to God. But this explanation is unavailable unless PP quantifies over a domain that includes God.

To be sure, we could restrict PP so that it does not quantify over God. But then we would lose the above explanation of Paul's belief. And the only reason to place this restriction on PP is if we have reason to think PP is false when left unrestricted. But we shall see that there is no such reason, once objections are addressed below.

Is impurity impossible?

Let's now consider two types of objections against the possibility of impurity – an evaluative objection and some metaphysical objections.

We begin with the evaluative objection. It is plausible that Paul's concept of impurity is an evaluative concept. After all, this concept is closely related to other evaluative concepts (e.g. sin), and those competent with the concept took impurity as a reason to separate from others. Under this assumption, one might argue against the existence of impurity by arguing that it entails a false evaluative claim. By analogy, consider the concept of sexual perversion, which is also evaluative. This concept may entail that the unnaturalness of a sexual behaviour is sufficient to make it bad. But this evaluative claim might be false, which would mean that sexual perversion does not exist (Priest (Reference Priest1997), 371). Similarly, it could be argued that impurity entails a false evaluative claim. Perhaps it entails that impure people have a feature F such that anyone who has F is therefore bad. And, once we find out what F is, we will see that this evaluative claim is false.

What could F be? An initial suggestion is that F is a historical property regarding how the impurity came about. For example, F might be the disjunctive property of having sinned or having been personally related to a sinner. Admittedly, it is implausible that this property would be sufficient for badness. However, this is not the property that makes an impure person bad. After all, this disjunctive property cannot be removed once one has it, and this means that the badness associated with impurity cannot be removed through purification. But the possibility of purification is essential to the concept of impurity (Klawans (Reference Klawans2000), 27).

Purification provides a clue as to what property F is. The way purification is depicted in the Bible suggests that it involves a change to the heart. For example, in Ezekiel 36 we find that the nation of Israel has become impure, and God voices his intention to purify them as follows:

Here, God removes the impurity by changing the heart.Footnote ¹⁰ Plausibly, then, F is some property of the heart. But what is the heart? And how can changing the heart remove impurity?

In Romans 6, Paul clarifies what the heart is by using a slavery metaphor to describe the control that sin exercises over a person:

Here Paul contrasts being ‘slaves to impurity’ with being ‘obedient from the heart’ in avoiding sin. And prior to this passage Paul explains the slavery metaphor as ‘sin [exercising] dominion in your mortal bodies, to make you obey their passions’ (Rom. 6:12). For Paul, then, the heart consists in the passions that control a person's actions. Obedience from the heart entails that one is controlled by passions that make one prone to righteousness. And slavery to impurity entails that one is controlled by passions that make one prone to sin. From this, we can infer that impurity itself entails that one has passions that make one prone to sin, whether or not one is controlled by them. Let's call this the passion account of impurity.

For clarity, we can construe passions as affective states, including desires and emotions. According to the passion account, a person is impure only if she is in an affective state that inclines her towards sinful action.Footnote ¹¹ Henceforth, the phrase ‘sinful affective states’ will be used to refer to affective states that incline the agent towards sinful action.

The passion account of impurity plausibly gives us the relevant feature F, and allows us to address the original objection. In particular, we can now see that there is no reason to think impurity entails a false evaluative claim. It is not implausible that being in a sinful affective state is sufficient to make a person bad. To use an analogy, desires to do something morally wrong are widely seen as bad-making features of an individual. There is no reason to think differently of sinful affective states. Moreover, contrary to Klawans's label ‘moral impurity’, we cannot assume that the sort of badness entailed by impurity is moral badness; it may just be that the impure person is bad in some particular way, without being morally bad.

Of course, impurity may require the possibility of sin, and sin might be seen as problematic if it entails the existence of God (e.g. if sin requires alienation from God). But, in the present context, this objection would only justify us in suspending judgement on whether impurity exists; it would not justify us in believing that impurity is impossible. And the present response to Schellenberg's argument can succeed even if we ought to suspend judgement on whether impurity exists.

Let's now turn to some metaphysical objections. Conclusion 3 might be rejected by claiming that Paul's concept of impurity (expressed by Conclusion 3) entails an impossible causal relation. Paul's concept is partly explicated by PP. And according to PP, one person's impurity can cause another person to become impure. If this causal transfer is impossible then Conclusion 3 is false.

This objection comes in different forms. First, one might hold that it is impossible for impurity to be causally efficacious at all. For an initial reply, recall the passion account: a person is impure only if she is in a sinful affective state. This account is neutral on whether impurity is identical to the property of being in a sinful affective state. But let's temporarily assume they are identical.Footnote ¹² In this case, impurity is causally efficacious if and only if the property of being in a sinful affective state is causally efficacious. And surely sinful affective states are causally efficacious – e.g. a desire to commit adultery can cause a person to commit adultery.

However, scepticism may persist. On the present assumption, impurity is not identical to any particular sinful affective state. It is rather identical to an existentially quantified property: the property of being in some sinful affective state or other. And if impurity is causally efficacious in addition to the sinful affective states it quantifies over, then there is causal over-determination, which is unacceptable. So, according to this objection, we should reject the causal efficacy of the higher-order property – impurity.

For our purposes, this problem can be met with a concessive response, one that allows higher-order properties to be causally relevant even if not efficacious. Frank Jackson and Philip Pettit (Reference Jackson and Pettit1990) have developed the notion of a ‘program explanation’ in response to the problem of higher-order causation. According to Jackson and Pettit, a program explanation (i) refers to a higher-order property that ensures the presence of a lower-order, causally efficacious property, and (ii) provides modal information about how the causal history might have gone, information that's not available on the lower level. For example, suppose a container is filled with water, and that the container cracks just as the water reaches boiling point. It is plausible that the momentum of water molecules is causally efficacious in cracking the container, and this may imply that the water's temperature is not efficacious (Jackson & Pettit (Reference Jackson and Pettit1990), 110). But surely the temperature property is not causally irrelevant. So, Jackson and Pettit provide a program explanation that refers to the temperature property, thereby giving it causal relevance. The temperature property satisfies (i) and (ii). First, this property ensures the presence of some causally efficacious momentum property – for the water to have its temperature there must be certain molecules moving with such and such momentum. And second, the explanation referring to the temperature property provides additional modal information not provided by the lower-order momentum explanation. The higher-order explanation tells us that the container would crack in relevantly similar circumstances (e.g. if the molecules were slightly different). Someone who knows the lower-order causal explanation could lack this additional information supplied by the program explanation. Thus, according to Jackson and Pettit, it is plausible that the temperature property is causally relevant even if not efficacious.

Jackson and Pettit point out that this approach ‘is clearly going to work with any explanations which are higher-order in virtue of using existential quantification’ (ibid., 115). So, if impurity is identical to the existentially quantified property of being in a sinful affective state, then impurity is guaranteed to be causally relevant to some effect. Moreover, they point out that if the higher-order property programs for a property that is relevant but not efficacious then ‘the story will go a level or more deeper until we find an efficacious property for which the original higher-order property programs indirectly, via the programing of intermediate features’ (ibid.). Thus, it turns out that we need not – and should not – assume that impurity is identical to the property of being in a sinful affective state. Impurity can gain its causal relevance by programing for an intermediate-level property that eventually programs for a lower-order efficacious one. And it is more plausible to treat the property of being in a sinful affective state as an intermediate-level property. After all, impurity ensures the presence of this intermediate-level property; and explanations that cite impurity provide modal information that's unavailable on the intermediate level – they inform us of the possibility of a person-to-person transfer, and this modal information is lacking from explanations that merely cite the property of being in a sinful affective state.

In short, the initial objection can be accommodated. The consequent of PP holds that an instance of impurity can occur ‘as a result of’ another instance. We have assumed that ‘as a result of’ picks out a causal relation, but this relation could be one of mere causal relevance, not efficacy.

However, even if impurity is causally relevant to certain effects, an objector might claim that it is impossible for impurity to be causally relevant to the impurity of another person. This objection is also implausible. According to PP, personal relationship is one nexus that facilitates the causal transfer of impurity from one person to another. And personal relationship often involves moral influence – a causal relation where one person's instantiation of a moral property is causally relevant to another person's instantiation of a moral property.Footnote ¹³ Moral influence occurs, for instance, when a teenager becomes prone to misbehaving as a result of his personal relationship with a troublemaker. The transfer of impurity described in PP can be understood similarly. Perhaps x's impurity is causally relevant to y's impurity, because of the moral influence involved in their personal relationship. The present attempt at arguing against PP claims that this is impossible, but it is clearly not.Footnote ¹⁴

At this point, one might grant that it is possible for impurity to cause another human to become impure, but object that it is impossible for impurity to cause God to become impure. However, Conclusion 3 does not entail that this is possible. Indeed, two premises in the Pauline argument – EHP and EP – entail that it is impossible for God to become impure. So what exactly is the objection?

The impossibility of God's impurity could be used to reject PP, as follows: (1) it is impossible for God to become impure, but (2) it is possible for impure people to be in personal relationships with God; therefore, PP is false. The first thing to note is that this argument obviously does not show that impurity is impossible, since premise (2) presupposes that impurity is possible. Moreover, it is unclear how the falsity of PP would show that Conclusion 3 is false, since the falsity of an argument's premise does not guarantee the falsity of its conclusion. However, as explained in the next subsection, premise (2) can be used in an argument that directly targets Conclusion 3. And we shall see that there is no reason to accept (2). So the claims in the next subsection also amount to a defence of PP from the present argument.

Can impurity put someone in a contrary position?

Conclusion 3 states that S's impurity explains why she's not in a position to relate personally to God. But one might argue against this explanation as follows: (1*) S's impurity can explain why she's not in a position to relate personally to God only if it is impossible for impure people to be in such relations to God (for, if such relations were possible, we would expect an all-loving God to bring them about). But (2) it is possible for impure people to be in personal relationships with God. Therefore, S's impurity cannot explain why S is in a contrary position.

As noted, (2) can be used to reject PP. But, more importantly, (2) and (1*) entail that Conclusion 3 is false. In response, it will be argued that there is no reason to accept (2). If so, then we cannot use it to reject PP or Conclusion 3.

The first thing to note is that we're not justified in accepting (2) merely on the basis of intuition or personal experience (e.g. the experience of being impure while in divine–human relationship). After all, for an intuition or experience to support (2), it would need to employ Paul's concept of impurity. But since we are not fully competent with Paul's concept, this intuitive or experiential support would be dubious. Nevertheless, we may have a competent grasp of the concept of a sinful affective state, and perhaps we might find the following intuitive: (2′) it is possible for people with sinful affective states to be in personal relationships with God. The trouble is that there is no reason to think (2′) entails (2). As noted earlier, we should not assume that the property of being in a sinful affective state is identical to impurity. We've also seen no reason to think it is sufficient for impurity (but only necessary). So, we have no reason to think (2′) supports (2).

Does omnipotence entail that it is possible for God to be in personal relationships with impure people? No. Omnipotence is, roughly, the power to bring about all possible states of affairs, not impossible ones (Flint & Freddoso (Reference Flint, Freddoso and Freddoso1983) ). And the question we are trying to answer is whether God's personally relating to impure people is a possible state of affairs. If it is impossible then God would still be omnipotent even though God cannot bring it about.

Do biblical texts suggest that impure people can be in personal relationships with God? Not clearly. Recall that personal relationships are to be understood as explicit, reciprocal, intimate, meaningful, and conscious. But there is no biblical text that clearly suggests that such relationships occur between impure people and God. Moses may be the only biblical figure recorded as having this kind of relationship with God: ‘the Lord used to speak to Moses face to face, as one speaks to a friend’ (Ex. 33:11). But Moses had the purity required for such relationship (Ex. 19:19–25). And the Bible describes this kind of relationship as extremely rare: ‘Never since has there arisen a prophet in Israel like Moses, whom the Lord knew face to face’ (Deut. 34:10). More importantly, there is no straightforward mention of an impure person enjoying this kind of relationship. Indeed, impure people were barred from highly impersonal relations to God – they could not even ‘touch the edge’ of the mountain on which God's presence was said to dwell (Ex. 19; Ps. 24:3–4). Thus, it is unclear that any biblical text supports (2). But even if we find a passage that can be interpreted as supporting (2), the above considerations would justify us in doubting that this interpretation is an adequate basis for accepting (2) – after all, (2) seems to conflict with the clearest biblical depictions of divine–human relations. Biblical support for (2) is doubtful if not altogether absent.

Of course, biblical texts do suggest that God can be personally related to those who have sinned (e.g. Moses). But sinful action is distinct from the state of impurity. A person's sinful action may be sufficient to make her impure for a time. But past sinful action does not entail that the agent is impure in the present, since she may have been purified in the meantime (e.g. Isaiah 6:1–8). So, the fact that God is personally related to those who have sinned, and to those who have been impure, does not entail (2).

Initially, one might think that purification entails (2), because God's purifying a person seems to require that God be personally related to her while she is impure. However, we've seen that purification can take place through a change to the heart. And clearly it is possible for God to change a person's heart without having an explicit, reciprocal, intimate, conscious, and meaningful relationship with her while she is impure. Thus, purification does not entail (2).

One might seek to support (2) by claiming that Christ was personally related to impure people (according to biblical accounts); and since God and Christ are one, it follows that God was also personally related to impure people. However, this argument assumes that God and Christ are one, but exploits a key problem for this assumption. In particular, there are many predicates that seem to apply to Christ but not to God (e.g. ‘was born’, ‘died’, ‘was buried’, etc.). Initially, this makes it difficult to hold that Christ and God are one. But the Christian theological tradition provides at least two strategies for circumventing this problem. And these strategies can be employed to undermine the argument for (2).

One strategy emphasizes that Christ and God-the-Father are distinct persons (albeit one in substance), and then provides enough ‘logical space’ between them so that the problematic predicates can apply to the former without applying to the latter. The other strategy claims that these problematic predicates are not literally true of Christ, but only of the human body that Christ used for his earthly mission.Footnote ¹⁵ Either strategy can be used to undermine the argument for (2). It can be claimed that ‘is personally related to impure people’ is among the problematic predicates, and is therefore not true of God-the-Father (on the first strategy) or not true of Christ (on the second strategy). The present argument for (2) must provide reason to exclude this predicate from the problematic ones. But it is unclear what reason this would be.

Of course, these strategies may fail in dealing with the above problem. But this would mean that the oneness of Christ and God is problematic, and should not be used to support (2). So, even if these strategies fail, we're still left with no support for (2).Footnote ¹⁶

Is inculpable impurity impossible?

IIP was initially supported by the idea that impurity can arise without its bearer performing any sinful actions; it can instead be acquired via personal relations with an impure person. Still, one might object that a person must be culpable for acquiring impurity in this way.

In addressing this, we should first take note of a claim that Schellenberg relies upon: ‘only for voluntary actions could we ever legitimately be blamed’ (Schellenberg (Reference Schellenberg1993), 63).Footnote ¹⁷ From this, it follows that a person cannot be directly culpable for her impurity, since impurity is a state, not a voluntary action.

Still, one might claim that, if a person S contracts impurity via personal relations, S must be culpable for some action that causes her to be impure. For what action must S be culpable? Here are two salient possibilities:

1. (i) S must be culpable for the act of entering into, or continuing, the relationship with the impure person;

2. (ii) S must be culpable for some act (e.g. adultery) that occurs within her relationship with the impure person.

But, as we will see, neither (i) nor (ii) seems to follow from the fact that S contracts impurity via personal relations.

Regarding (i): Must S be culpable for the act of entering into, or continuing, the relationship with the impure person? This act need not be considered wrong or sinful. But it is plausibly harmful to S, since it leads to S's impurity and alienation from God. Still, even if the act is harmful, S need not be culpable for it, since harmful action does not entail that the agent is culpable (i.e. blameworthy). In particular, an agent of harmful action can have an excuse that eliminates deserved blame (Zimmerman (Reference Zimmerman1997) ). For example, it is widely accepted that subjects can be excused for harmful actions that they've performed while non-culpably ignorant of the harmful effects (Ginet (Reference Ginet2000) ). And it is possible that S enters into the relationship while non-culpably ignorant of the fact that she could contract impurity from the other person. Or, S could be non-culpably ignorant of the fact that the other person is impure. Thus, even if it is harmful for S to enter into, or continue, the relationship that makes her impure, S need not be culpable for this act.

Regarding (ii): Must S be culpable for an act that occurs within her relationship with the impure person? The only reason for thinking so would involve the sinful affective states required by S's impurity. So, let's suppose that S's personal relationship with the impure person causes her to form a desire to commit adultery, and that this desire is the relevant sinful affective state. Even so, it is possible that S is not culpable for any act that leads to her forming this desire; again, she could be non-culpably ignorant of the fact that she would form the desire. Moreover, S could altogether resist the temptation to commit adultery and never act culpably as a result of the desire.Footnote ¹⁸ So, the sinful affective state, which partly constitutes S's impurity, need not be the cause or effect of any culpable acts. And there is no reason to think S's impurity must involve something else that would be the cause or effect of a culpable act. So, we have no reason to accept (ii).

Neither (i) nor (ii) seems to follow from the fact that S contracts impurity via personal relations. There appears to be no reason to think a person must be culpable for becoming impure.

Is Conclusion 3 incompatible with perfect love?

If impurity precludes divine–human relationship, we might expect an all-loving God, who seeks personal relationship, to create a world of universal purity (i.e. containing no impurity). But it will now be argued that there is a reason an all-loving God could have for allowing some impurity.

First, consider that personal relationships vary in quality – some instances of personal relationship are better than others. Schellenberg agrees. According to him, personal relationship with God ‘would admit of change, growth, progression, regression. It might be shallow or deep, depending on the response of the human term of the relation’ (Schellenberg (Reference Schellenberg1993), 28).

Second, consider that an all-loving God would have more reason to pursue better personal relations than to pursue lesser ones, other things being equal. This is true because better relations would have more of the value that Schellenberg attributes to personal relations. They would have more intrinsic value (being more worthy of pursuit in their own right), and more instrumental value (contributing more to human well-being). And third, consider that there are possible persons who are capable of better relations with God only if they are purified rather than prevented from ever becoming impure. God could not attain higher-quality relations with these people in a world of universal purity. This means that God would have reason to create a world containing these people along with their impurity. And, given the second point above, this reason could override any reasons that God has to create a world of universal purity.

Why should we think there are possible persons who can attain better relations with God only if they are purified? First, notice that there are factive mental states (e.g. knowing that P, being glad that P) that help to improve personal relationships, provided they have the right content. For instance, if a wife knows that her husband has been faithful then this can improve their relationship much more than if she does not know this.Footnote ¹⁹ Moreover, knowledge is factive: S knows that P only if P.

Now imagine a level of relationship that one party cannot attain without a certain factive mental state. For example, a wife might be incapable of emotional intimacy with her husband unless she knows that he has been faithful. Now consider a divine–human case:

1. (1) There is a possible person S, degree of value D, and factive mental state M such that S surpasses D in relationship with God only if S has M regarding the proposition that God has purified S.

There is no reason to reject (1), especially if we imagine that D is a very high degree of value. Moreover, given that M is factive and purification requires past impurity, (1) entails

1. (2) There is a possible person S and degree of value D such that S surpasses D in relationship with God only if S is impure at some time.

Now, if (2) is true then God would have reason to create a world containing S along with some impurity – the reason being that God desires S to surpass D. And this reason could be overriding since, all else being equal, an all-loving God would have more reason to pursue a relationship that surpasses D than to pursue one that falls short (e.g. in a world of universal purity). Therefore, if we have no reason to reject (1), then we have no reason to expect an all-loving God to create a world of universal purity.

One might resist (1) by claiming that, for every possible person S and every degree of value D, there is no factive mental state that is required for S to surpass D in relationship with God. But this claim is inconsistent with Schellenberg's argument. Recall that a central premise in his argument is that divine–human relationship is ‘logically impossible for one who does not believe that God exists’ (Schellenberg (Reference Schellenberg1993), 30). So, every degree of value of divine–human relationship requires one to believe that God exists. Moreover, every degree of value of divine–human relationship requires this belief to be true, since one cannot be personally related to God unless God exists. Therefore, every degree of value of divine–human relationship requires a factive mental state – the state of true belief that God exists. The objection, then, is incompatible with Schellenberg's argument.

In short, there is no reason to think an all-loving God is incompatible with the existence of impurity. Furthermore, there is no reason to think an all-loving God is incompatible with inculpable impurity. If God allows S to be impure so that S can attain a higher level of divine–human relationship, it is implausible that God should require that S be culpable for her impurity. Clearly, S's impurity could be contracted inculpably (as outlined earlier). Thus, there is no reason to reject the compatibility of Conclusion 3 with an all-loving God.

Is Schellenberg's argument defeated?

If Conclusion 3 is true and compatible with an all-loving God then Schellenberg's P1 is false. However, the modest approach in this article is not sufficient to establish that Conclusion 3 is true and compatible with an all-loving God. It has only been argued that there is no reason to disbelieve these claims, but there still may be no reason to believe them. The rational response might be to suspend judgement. Nevertheless, even if we should suspend judgement, Schellenberg's argument still fails, because this means we should (at most) suspend judgement on P1, which means P1 cannot justify Schellenberg's atheistic conclusion.

Of course, we may recall that Schellenberg provides positive reasons in favour of P1. Do these reasons justify P1 above the level of suspension of judgement? No. Schellenberg's reasons for P1 leave open the putative counterexample that's been levelled against it. Schellenberg bases P1 on the idea that an all-loving God would seek the intrinsic and instrumental values associated with divine–human relationship. But it was already argued that there is no reason to think inculpable impurity is incompatible with an all-loving God who seeks these values. In particular, it was assumed (in the preceding section) that an all-loving God would seek these intrinsic and instrumental values; though it was pointed out that God would have more reason to seek the greater levels of such value associated with higher-quality relations. Moreover, we saw that higher-quality personal relations might be attainable only if the subject is impure for a time; and clearly this impurity need not arise culpably. So, Schellenberg's reasons for P1 do not exclude inculpable impurity and therefore do not exclude the putative counterexample levelled against P1. Even if divine–human relations have the values that Schellenberg attributes to them, there still might be inculpably impure people who would not be in a position to relate personally to God.

In short, Schellenberg's reasons do not support P1 over and above the counterexample levelled against it. At most, we should suspend judgement on P1, which means that P1 provides no support for his atheistic conclusion.