Prologue

Matthew records that in the ninth hour of Christ's crucifixion, Jesus cried out with a loud voice, ‘Eli, Eli, lama sabaachthani?’ or ‘My God, my God, why has thou forsaken me?’ (Matt. 27:46). It is a dark moment. G. K. Chesterton captures its drama in a way that is worth reproducing at length:

The problem of the atonement

For something as terrible as Christ's suffering to be vindicated as morally good, it seems that there must be something truly horrifying that would happen in its absence. Perhaps motivated by an idea of this kind, Christians have long sought to understand Christ's suffering as making possible the forgiveness of human sins.Footnote ⁷ Without it, humans would be consigned to remain in a sinful state, and would thereby be subject to hell forever. That quickly starts to sound like a problem that could demand a solution as serious as the atonement.

But connecting the atonement to the possibility of forgiveness runs quickly into a dilemma, famously pointed out by St Anselm. The first horn of the dilemma is that if God could not forgive sin except by way of the death of a perfect person, then God is not omnipotent. The second horn is that if God could have forgiven sin in some other way but decided to require the death of a perfect person anyways, then God did something horribly wrong (Anselm (Reference Anselm, Davies and Evans1998); Porter (Reference Porter2004), 229). Satisfaction theories respond by holding that sin creates a debt that must be repaid if persons are to be saved, either (on the first horn) because there is no other way to do it, or (on the second horn) because other ways of making payment would be worse. Substitution theories hold that it is valuable for sins to be punished, and it is either impossible (on the first horn) or undesirable (on the second horn) to forgive sins without punishment. Thus, satisfaction theories are informed by an ideal of compensation, whereas substitution theories are informed by an ideal of retribution (Murphy (Reference Murphy2009), 254 n. 1). Although St Anselm himself took the first horn, it seems strange to think of God as hostage to some kind of causal force that forbade him from forgiving without exacting the terrible suffering of a perfect person first. Most contemporary defenders of satisfaction and substitution views have preferred to take the second horn, and have tried to explain why the atonement was not morally unacceptable even if alternatives were available. In doing so, they confront two central challenges. First, why should we think that punishing a third party can offer compensation for another's sins, or satisfy the debt of guilt created by someone else? Second, even if punishing someone could pay for another's sins or satisfy another's debt, why suppose that a loving God would demand this? We, despite our moral weaknesses, can forgive each other without demanding payment or suffering. Why couldn't God do the same? The first poses a problem about why the atonement of Christ should help at all, and the second asks why it should be necessary for achieving that aim. Call the first the problem of helpfulness and the second the problem of necessity.Footnote ⁸

The problem of helpfulness squarely confronts substitution theories of the atonement. If substitution theories presuppose a value to retributive punishment, then they must explain why the goods of retribution can still be had by punishing someone other than the wrongdoer. That seems to be a tall order, since retributivist theories are often motivated in part by the thought that retribution would disallow punishing the innocent, while other theories of punishment (for example, utilitarianism) would not. The reason retributivist theories disallow punishing the innocent is that the putative value of retribution is in bringing something bad upon someone who deserves it. But the innocent do not deserve it at all. Thus, penal substitution seems to be a perversion of the retributivist ideal.

Satisfaction theories skirt the problem of helpfulness by denying that the atonement was an act of penal substitution at all. The locus classicus of satisfaction theories is Richard Swinburne's (Reference Swinburne1989). By offering reparation (payment) and penance (sorrow for sin) when he was not obligated to do so, Christ made available these resources to us, who are not able to offer them on our own. Swinburne explains:

For Swinburne, Christ's death provides us with resources to receive God's forgiveness, rather than substitute punishment. But this proposal immediately raises the problem of necessity: Why would God require the incredible suffering of the atonement before being willing to forgive his children? Within our ordinary moral practice, it is possible to forgive another person who has wronged us without ever wanting that person to suffer, let alone without that person ever actually suffering. Forgiveness involves forswearing resentment and other negative reactive emotions, and in that way sets aside the prospect of holding the person accountable in the future (Allais Reference Allais2008). Even if I do resent your wronging me before forgiving you, my resentment need not be partly constituted in a desire that you suffer. Rather, it may just express my commitment to the moral principle which you have violated, as well as the seriousness of my moral standing as an agent. Reactive attitudes can be negative without being punitive.Footnote ⁹ Indeed, it is possible to even forgive offenders who display no penitence at all. In general, our moral practice gives wide discretion to wronged parties in deciding when to forgive (Allais (Reference Allais2008); Sussman (Reference Sussman2005) ). This practice is relevant because it underscores the present point: if ordinary moral agents can forgive without requiring any suffering, then wouldn't we expect a perfect moral agent to have a much greater capacity for forgiveness than we do? Why, then, think an atonement would be necessary?

There is an answer to this question common to several theories that regard the atonement as a means to forgiveness. The answer is that granting forgiveness without exacting any punishment, reparation, or suffering from anyone would – in some sense – betray the significance of the moral law. Swinburne challenges the claim of the preceding paragraph, holding that even in ordinary cases of wrongdoing the victim errs in forgiving the offender if penance and reparation are not forthcoming. This is a mistake, Swinburne believes, because it involves ‘not taking the other seriously in the attitudes expressed in his actions’ (Swinburne (Reference Swinburne1989), 148). The idea of suffering as a way of taking each other seriously comes up frequently in related explanations of the atonement. Steven L. Porter writes:

David T. Williams makes the point in more explicitly scriptural language:

The shared idea here is that just forgiving people – without demanding anything of anyone – would be a kind of moral mistake. By treating the offender as someone whose actions could just be dismissed rather than dealt with, it would be failing to hold the sinner responsible and, thus, not taking her or him seriously as a moral agent. Or, it would fail to treat the offended party as a serious moral agent by suggesting that wrongful treatment of the offended party was not important enough to require some sort of future attention. Or, it would forgo opportunities for the moral education of the offender. Requiring some suffering might be thought of as one way of expressing our common commitment to the moral law; by demanding punishment or reparation, we publicly exhibit our shared understanding of moral requirements and prohibitions as authoritative in our practical reasoning. At the extreme, disregarding morality by failing to demand something prior to dispensing forgiveness might show we do not actually think that sinning against God is morally impermissible.

Challenges for substitution and satisfaction theories

Taken together, these points sketch an argument that, roughly, it is a moral mistake to grant forgiveness too easily. Call it the argument from moral seriousness. I think this argument contains a shallow truth and a deeper mistake. The truth, to my mind, is that it might sometimes be possible to forgive too easily. If I forgive everyone who wrongs me without any second thought, I might be revealing either that I do not regard offences against morality as important, or that I do not regard myself as important. Imagine if I had a friend who was wronged by some third party. But when my friend expressed his sorrow over this offence, I nonchalantly replied that I had already forgiven the offender and had moved on. Perhaps I might even recommend that my friend do the same. If I were to respond in this way, I could be accused of not regarding my friend's moral standing as appropriately serious, or as not regarding my relationship with my friend as significant to me in an emotionally attentive way. When those close to us are hurt, we express our commitment to them by showing our indignation at those who have trespassed against them. For analogous reasons, it might be a mistake never to resent those who have wronged us personally. If we did not have any attitudes against them, it would show that we were not emotionally attentive to moral agency instantiated in our own person. That is sometimes associated with adopting a posture of servility.Footnote ¹⁰ So, the argument from moral seriousness is right to suppose that if we love someone – including ourselves – we should not just disregard offences against her (Frankfurt Reference Frankfurt2004).

The mistake in the argument from moral seriousness is to suppose that our only two choices involve requiring that something bad happen to someone, or disregarding an offence altogether. The value of suffering or punishment, according to the argument, is either expressive or educative. It could display our commitment to something valuable (a person, a principle, or a relationship), or it could teach someone something valuable. To say, with the quotation above, that a sin ‘cannot be forgotten but must be expiated’ slides towards the thought that if there is no punishment or suffering, then we must have just decided to forget. But these are not the only possibilities.

Consider an example used by Stephen Darwall. Darwall recounts the true story of Laura Blumenfeld, whose father was shot (and nearly killed) while on a tourist trip in Jerusalem. Blumenfeld visits her father's almost-assailant in prison. Under the pretence of being a journalist, she comes to know the attacker – Omar Khatib – and develops a personal relationship with him and with his family. Throughout this time, Omar remains unrepentant, seeing his actions as part of an otherwise justified political movement. Omar's family agrees, seeing the shooting as ‘nothing personal’, but instead as a tool for garnering media attention (Blumenfeld (Reference Blumenfeld2000); Darwall (Reference Darwall2006) ). A couple of years later, Laura attends a legal hearing for consideration of Omar's early release, on account of his poor health. Darwall describes what happens:

The general problem with accounts of the atonement that require suffering or punishment is not that they make forgiveness out to be more difficult than it really is. Holding another accountable in a way that respects the moral agency of everyone involved may be very challenging – perhaps sometimes even more challenging than old-fashioned suffering. The problem with these accounts of the atonement is that they are insufficiently sensitive to the diversity of ways or methods or tools at our disposal for holding each other accountable while taking each other seriously. In the story above, Laura Blumenfeld holds Omar accountable for his crime. She treats Omar and his family – as well as herself and her father – as morally serious agents. She does this by coming to know and understand them, and showing herself to be a person whom they could trust. She does not disregard the seriousness of Omar's crime, but uses it to guide her actions. At a crucial moment, she asserts her own authority to speak about the crime; she does not treat herself or her father as morally servile. What is of interest is that she accomplishes all of this without demanding punishment or suffering. In fact, her act of holding Omar accountable is in part accomplished by requesting that his punishment be ended. It is by extending forgiveness, rather than insisting on reparation or payment, that she calls Omar and his family's attention to the gravity of their moral mistake.

The problem with atonement theories connecting forgiveness with punishment or suffering is that they don't recognize this kind of freedom in finding novel and appropriate ways of holding others accountable for wrongdoing while taking themselves seriously at the same time. Laura Blumenfeld restored the broken moral relationship between herself and her father's attacker by transforming the incident from a piece of arbitrary destructive violence to an opportunity for forging understanding. At no time does she even express indignation or resentment, let alone call for suffering. If atonement can be won by ordinary moral agents without resorting to punishment or suffering, why suppose that God cannot also forgive without suffering or punishment? Theories requiring that God insist on punishment or suffering of someone seem to me unduly sceptical about God's moral creativity.

I believe this is a quite general problem for theories of the atonement that require some penance before forgiveness can be accessed. Some of the most recent philosophical theories of the atonement have departed from both the satisfaction and substitution theories, but retain a focus on punishment. I will illustrate this point with one recent, philosophically sophisticated example. Mark Murphy has argued that the atonement is about punishment, but not substitution. Instead, he holds that Jesus's death is a form of vicarious punishment. Vicarious punishment involves inflicting hard treatment on one person in order that another person be punished. According to this theory, Christ is innocent of any crime, and suffers hard treatment not because culpability can somehow be transferred from us to him, but because his suffering is a way of punishing us. And indeed, it does seem that we can certainly imagine vicarious punishment: we count the suffering of those close to us as bad for us as well – perhaps even worse than our own suffering. As Murphy writes, ‘it may be that my imprisonment would not be as bad for me as would be my wife's, or my daughter's, imprisonment’ (Murphy (Reference Murphy2009), 264).

This account has a number of virtues. For one, it does not attribute any injustice to God, provided – as Murphy explains – that no hard treatment is imposed on the innocent person without that person's consent. There is also an appealing moral unity towards which the account adumbrates. Jesus dies for people who all share a common vicarious punishment – the suffering of a person they commonly love. This shared experience of punishment might then offer a basis for a shared loyalty, to Christ as well as to fellow members of the religious community. Borrowing a metaphor, Murphy imagines a sports team whose captain is made to run laps for the mistakes of other team members. Realizing ‘our captain is suffering out there, for what we did’ is the refiner's fire of team unity. Murphy writes: ‘We humans, or we Christians, are unified through the fact that our punishment is through the suffering of our Lord, Jesus Christ’ (ibid., 267).

This picture provides a moving ideal of the atonement, but it also contains the seeds of puzzlement about how the details are meant to be worked out. It is clear that the hard treatment of someone we care about a lot is bad for us as well. And so it is correspondingly clear that, for Christians who care about Jesus Christ, his hard treatment can count as a punishment. But what about for other people, who may have never heard of Christ or may even have negative evaluative attitudes of him? Murphy's ambiguous rephrasing of ‘we humans, or we Christians’ suggests this point. Is Christ's death important to everyone in the human family, or only to Christians?

Regardless of which way one wishes to answer, a dilemma lurks behind this question. On the model of atonement as vicarious punishment, it seems that people are either not punished at all, or else they are punished unnecessarily. On the one hand, consider people who commit vicious crimes. Although such persons may be deserving of punishment (at least if, as Murphy supposes, retributivism in some form is true), it seems very unlikely that the hard treatment of another innocent person will succeed in punishing them. Indeed, their viciousness may be just the thing that will insulate them against being made to suffer by the suffering of another person. Of course, it is possible that a few people who commit vicious acts will also be committed to Christ in such a way as to make vicarious punishment succeed. Just as likely, though, they will be indifferent to Christ's suffering. They may not value Jesus at all, or their viciousness may extend even to the figure of Christ. These might be those who would crucify Christ afresh, if only they had the chance (Hebrews 6:6).

Perhaps, then, Murphy intends the account of vicarious punishment to apply only to Christians. If that is right, then it turns out – perhaps surprisingly – that Jesus did not really die for the sins of the world. He died for the sins of those who have, or form, commitments to him. This version of the view would at least explain how Christ's suffering counted as a punishment. There will be many people for whom the suffering of Christ is a grave moral calamity. These people would be punished by the hard treatment of Christ. The problem is that there is no reason to think that these people also deserve to be punished. Here the problem of necessity resurfaces. For a great many Christians, their love of Christ will make his suffering a very serious punishment, but they will have done nothing to earn so serious a punishment as this. Christians believe that all fall short of full compliance with the law, but this hardly means that all need to be punished. To repeat the now familiar refrain: we can forgive each other our trespasses without imposing suffering, so why think that God cannot? Notice that this need not involve denying the truth of retributivism (so we can keep intact all of Murphy's premises). One could accept that it is sometimes good for the guilty to suffer without thinking that it is always good for everyone who has done anything wrong to suffer. But even if you hold the latter, implausibly strong view, it would still not vindicate the atonement as vicarious punishment. This is because the suffering of those who love Christ would remain – for many Christians who had not done anything particularly terrible – deeply disproportionate to the offences perpetrated.

To put the point a different way: retributivists usually think that those whose crimes are greater should be punished more, but on the current proposal, it could well turn out that those whose crimes are least are punished more. The more innocent a Christian's heart, the more vulnerable she or he might be to the suffering of Jesus, and so the greater the punishment it would impose. The more hardened one is in one's viciousness, the more likely the punishment is to be completely insignificant, or not even a punishment at all. My purpose in making this point has not been to disparage Murphy's vicarious punishment theory, which is novel and interesting. Instead I have just been trying to underscore how difficult it is for models of the atonement based on punishment and suffering to explain why Jesus's death was necessary. It seems that however such a view is specified, there is a fundamental mismatch between the severity of the event and the conditions for appropriate forgiveness. There are just not many cases in which something as serious as the atonement is needed to justify something as simple as forgiveness.

The authority of God

What would be wrong with God forgiving without punishment? Recall the passage from Donald Williams quoted in the first section: if God just forgave, that would suggest that sinning against God was morally permissible. In the last section I argued that this inference was mistaken. Forgiving freely does not imply that the offender did not do something morally wrong. Never mind that for now. Suppose it were true that free forgiveness implied no wrongdoing had occurred. Why should that be a problem? The implicit idea seems to be that anything that is a sin must also be morally impermissible. But this seems to me to be obviously mistaken, at least on a variety of ordinary religious views. Christians believe in a commandment to love one another, but they seldom think they have been morally wronged by those who fail to love them. We do not resent those who don't think to walk with us two miles or give us their cloaks as well as their coats – but we do resent those who wrong us (Wallace Reference Wallace1994). Mormons, Jews, and Catholics observe certain religiously motivated dietary practices. Though they regard the violation of these principles as sins, they are seldom inclined to think that others outside their religious communities should feel guilty about their non-compliance with these rules.

In short, there seems to be a gap between thinking that some transgressive act is a sin, and thinking that it is plain morally wrong. This is not to say that sin is not a normatively important category; it is just to say that it is not extensionally the same category as moral requirement. I doubt this result is surprising; plenty of norms are important to our judgements about what we have reason to do, even when we do not regard them as enacting moral requirements or prohibitions. They may be norms of etiquette, politeness, or club rules.

The puzzle I aim to raise is about why philosophers and theologians sometimes suppose that sinning must also be violating a moral requirement, when this does not seem to match our moral practice. Here is a possible diagnosis. We think that God's commands must be in some way important to how we ought to act. That is, we believe that God must be a kind of authority in our lives.Footnote ¹¹ And we also believe that morality is authoritative in our lives. We ought to comply with moral requirements. So if God's commands doubled as moral requirements, then we would have some purchase both on why God was an authority, and on how moral requirements were grounded or sourced. What this picture suggests is a divine command metaethics, according to which moral requirements are explained by appealing to God's commands. Although there is a venerable history of trying to defend this type of view, I will not take it up here. I will accept the longstanding metaethical scepticism towards divine command theory.Footnote ¹² Either the commands of God are sensitive to moral reasons, in which case morality is based on something that precedes God's commands; or else God's commands are not sensitive to moral reasons, in which case they are arbitrary.

Let me take stock. Punishment-based theories of the atonement try to use the suffering of Christ as a way of shoring up the connection between moral impermissibility and breaking God's commandments. In this way, they try to preserve God as a moral authority. But that is a mistake twice over. Punishment is not needed to show something was morally impermissible, and God's authority is not the authority of moral requirement anyway. Nevertheless, I do think there is something right about this approach. We want God to have some kind of authority in our lives. And I also think that the atonement is centrally connected to conferring this authority on God. For the rest of this section, I will try to explain what I take the authority of God to be.

Consider two basic – if simplistic – ways in which one agent can have authority over another. First, A might have authority over B if A can issue commands backed by credible threats of force. A has authority over B in the sense that if B failed to comply, A could either force B to comply or else harm B's interests in some way. Call this Hobbesian authority, based on Hobbes's insight that the sovereign had authority in virtue of holding power to alter the incentives faced by members of a political community (Hobbes Reference Hobbes and Curley1651/1994). Alternatively, A might have authority over B if A knows more than B about some domain of practical interest to B. In this case, A has authority over B in that A's assertions about the domain provide B with reason for acquiring beliefs about what A says (Hinchman (Reference Hinchman2005); Keren (Reference Keren2007) ). Call this epistemic authority.

Now, it would be an impoverished view of God to regard God only as a Hobbesian authority or as an epistemic authority. We do not want to believe that God has authority in our lives only because he is very powerful; nor do we want to accept God's authority only because he is very smart. The first is inadequate because it does not distinguish God's authority from that of a very powerful demon who could visit something terrible upon us if we did not comply with his demands. Our reasons to heed God's commands are not just because we rationally fear God's power. The second is inadequate because it restricts God's authority to our beliefs. If God were just an epistemic authority, we would have strong reason to apportion our beliefs to his assertions, but we would not thereby have reason to do anything in particular. It would relegate God to the status of a divine weatherman.

One might counter that epistemic authority often marches in step with practical authority. In less metaphorical words, if you have epistemic authority about some subject matter that I regard as relevant to my decisions about what I will do, then your authority will predictably extend to my specifically practical reasoning. However, this need not be so. Suppose you have a lot more knowledge than I have about how to swing at a baseball pitch. If I recognize this, I rationally ought to accept your assertions about how one best swings a bat. However, it might be that if I tried to act on your advice, I would do relatively worse than if I did not attend to your advice at all. (For example, I might become nervous by thinking of your complicated instructions, and my nervousness might make me more likely to miss the ball.) So, there is no guarantee that epistemic authority, even when it bears directly on some practical domain, will translate into practical authority.Footnote ¹³

A third conception of authority is designed to solve this problem. Consider what Joseph Raz calls the ‘normal justification thesis’:

Raz's conception of authority is genuinely practical. You have authority over me, according to this view, when I would do better at following the reasons I have by accepting your directives as authoritative than by trying to follow the reasons directly. When the normal justification thesis applies between you and me, it follows that it would be rational for me to treat you as an authority. But one might still worry that this is not adequate as an expression of God's authority. Imagine that you have Razian authority over me (that is, the kind of authority just described), and that you and I share the knowledge that you possess this authority. Now suppose that you give me some authoritative directive, but I tell you that I have no intention of following it. Although I am making a rational error of a sort, it does not seem like you have any basis for complaint. I might say, ‘Look, it is true that I would do better to follow your advice than to go on my own, but that fact does not mean that I owe it to you to do as you say. How I act is none of your business!’ It may be (as we sometimes imagine) that a baseball coach or politician would do better to follow our advice than to act on some other set of considerations, but no one supposes that this gives us standing to demand that they do as we prescribe.

Raz's conception of authority also feels too epistemic (Darwall Reference Darwall, Sobel and Wall2009). We want God to have a kind of authority whereby God would also have standing to hold us accountable for our actions. Such authority is not conferred merely by a superior epistemic vantage point. What is required is that the two agents have some relationship that is morally important in a way that gives them the ability to hold each other accountable: to blame them when they do not act on authoritatively issued reasons and to praise them or feel grateful to them when they do. These kinds of practices are not made appropriate by any of the conceptions of authority on offer so far, but they are essential features of the ordinary Christian conception of a relationship with God.

What, then, is lacking? It may help to think of a typical personal relationship wherein the parties have a non-epistemic authority over each other's lives. Perhaps the paradigm example of this type of authority is in a marriage. When two persons share a marriage, they treat the other's intentions and plans as providing them with reasons for action as well. They have reason to want each other's ends to be satisfied, and to coordinate their actions in such a way as to satisfy each other's ends as well as their own. Even if one party to a marriage thinks the other's end is impractical or unnecessary, that does not compromise its reason-giving status. In this sense, being married to someone involves treating that person as a source of practical authority in your own life. Notice that it need not extend to the epistemic realm at all. You need not judge that the other person's values are in fact valuable, but you will still feel some normative pressure to promote your partner's ends, however silly or inconsequential you may find them.Footnote ¹⁴ Nor need you think that your partner's preferences set up moral requirements for you to comply with, as divine command theorists suppose about God's will. Indeed it would be a strange perversion of love, rather than an expression of it, to treat your spouse's will as if it were the moral law itself. Still, there is a perfectly ordinary sense in which the purposes and requests of one's spouse carry an implicit reason-giving authority. You cannot just dismiss the requests of your spouse on a whim, at least without ignoring morally serious reasons that you have.

It should then not be surprising that the Bible so often invokes the metaphor of a marriage in describing God's relationship with his people. Likewise the other most common metaphor – that of God as a loving parent – also picks out a relationship of practical authority.Footnote ¹⁵ We regard parents as having authority over their children, and we do not generally think that this authority is reducible to the epistemic superiority of parents (Keller Reference Keller2006). Instead, we treat parents' directives as giving us reasons for action in virtue of their relationship with us, much as we treat God's word as reason-giving. These metaphors have in common the idea of a non-instrumentally valuing relationship (Scheffler Reference Scheffler2010). When we value someone non-instrumentally, we care about them for their own sake. We are vulnerable to them. Our attitudes and emotions respond to theirs. We are hopeful of their successes and committed – to some degree at least – to contributing what we can to them. In virtue of our emotional and intentional entanglements, the other person occupies a position of authority in our lives, and we similarly occupy a position of authority in theirs. This shared authority requires – in the most intimate cases – deciding together how we will lead our lives with each other.

Of course, not every case of shared practical authority is like a marriage. The most typical example of such a relationship is probably friendship. We regard our friends as having a standing to ask things of us, and to hold us accountable for how we respond to these requests. This is to say that our friends occupy a position of authority in our lives. It is not just that we care about their well-being. You might care about the well-being of someone with whom you have no relationship at all – the way that I hope for the best for my favourite athlete or musician. It is the further fact that the reason-giving force of our friends' requests follows from their authority itself, rather than from facts about their well-being. I have moral reason to help a stranger carry a heavy bag of groceries in from his car, just because I see that this would be good for him (Ebels-Duggan Reference Ebels-Duggan2009). But I have reason to do what my friends ask of me, sometimes even if I think it will not be good for them in the end. To modify a case from Ebels-Duggan, suppose my close friend is applying to graduate school, but I know that graduate school will be a frustrating and unsatisfying experience for him. My protests notwithstanding, he tells me he is still going to do it, and he asks for my help. Most of us, I think, would hold that I still have reason to help. However, my reasons cannot arise from facts about my friend's welfare. They follow, instead, from his standing of practical authority in my life.

Atonement as a source of authority

Just as with our friends, we do not usually think that our reasons for acting on God's commands follow from the fact that we will thereby improve God's well-being. God will do fine either way. For reasons canvassed above, it is also unsatisfying to think that our reasons for following God's commands are just for our well-being. That mistakenly treats our relationship with God as instrumental. We do better, I have proposed, to follow biblical metaphor in regarding the authority of God in our life as irreducibly practical – like the authority of a spouse, or a parent, or even a friend. In this final section I will try to show that the atonement places God in a position where this type of practical authority is possible.

‘Ye are my friends,’ Jesus says to his disciples, ‘if ye do whatsoever I command you’ (John 15:14). To say that Jesus's qualification sounds a little stringent is to be modest. Being friends with Jesus is apparently pretty demanding. But this reaction obscures a deeper point that the verse is making: friendship is always about giving someone else authority in your life. And in asking for that from his disciples, Jesus is only inviting his disciples to share a profound commitment he himself has with his Father (John 15:10).

The question I now want to consider is: why should Jesus's disciples – or anybody, for that matter – have reason to want to accept God's offer of friendship? One might think this question is in danger of being too obvious. Why ever would we not want God's friendship? God is all-knowing and all-powerful, not to mention perfectly loving. Remember, though, that the question of friendship is not like the question of team membership or business dealings. It would definitely be useful to have a being of infinite power and knowledge on one's side, but that is the kind of instrumental thinking that has already been ruled out. Once such considerations are eliminated, it is much less obvious that we would immediately want to be friends with a perfect being. Think again to ordinary cases. It is a long-considered observation in moral philosophy that we are just as likely to be annoyed as awed by moral saints (Wolff Reference Wolff1982). We can easily find morally heroic personalities as other-worldly, devoid of the personal flaws and quirks that we appreciate in our friends (Velleman Reference Velleman1999). We can also find their motivations to be foreign or rigid. In short, the fact that an agent is morally better than we are can make us less, rather than more, likely to want to be friends with them.

The reason for this phenomenon is plain. One thing we want out of friendship is a shared feeling of empathy or understanding. Stronger still, we want to have friends that we regard as like us in some fundamental way. Accordingly to an old and venerable tradition in moral philosophy, we want to be able to see ourselves in our friends (cf. Brink Reference Brink1999). Only when we identify our friends as being like us can friendship provide a ‘mirror of the self’ (ibid., 264). And that is something worth wanting, because the self is a difficult thing to understand. A variety of cognitive pressures all converge on pushing us to achieve a coherent self-understanding (Velleman Reference Velleman2004). We feel guilty or ashamed when our beliefs and emotions stand in tension with our desires or plans. We say that people who are inconsistent lack integrity, or are being ‘untrue to themselves’ (Knobe Reference Knobe2005). We try to figure out what we are like by looking closely at other members of our families (Velleman Reference Velleman2004). We have, in short, strong and urgent moral reason to want to see ourselves more clearly.

Friends who are like us can give us a helpful perspective on who we are. This understanding makes it easier to grant them practical authority in our lives. There is a common reluctance to grant other persons control over our lives, but such fears can be allayed by seeing the other as relevantly similar to ourselves.

The trouble with being a moral saint is that it makes it difficult to relate to people. But even moral saints are still human. The trouble with being God is so much the more pressing. If we have a hard time connecting with people who are morally better than we are, how difficult must it be to forge a relationship of deep personal commitment with God? Jesus tells his disciples that it is actually not hard at all. His assurance comes as a surprise, but there is a reason for it. He has done the difficult part already. Even before his death, Jesus announced that he had ‘overcome the world’ (John 16:33). Jesus saw in himself the fulfilment of Messianic prophecy. He would come into Jerusalem and liberate Zion from oppression (Exodus 24:8; Zechariah 9:9–11; Hosea 10), albeit not in the way that anyone anticipated. The story of Jesus's life had been one of turning conventional teachings on their heads. The way to get out of conflicts with your enemies is not to destroy them but to love them (Matt. 5). The greatest were those whom everyone regarded as the least (Matt. 18:1–5). Those with official religious authority knew nothing about God (Luke 22:53). The Kingdom could only be found in places where no one suspected it (Matt. 12). Jesus would not overcome the captivity of foreign powers in the way that people expected the Messiah would. But then he had done nothing in the way that people expected the Messiah would. Jesus would deliver Jerusalem like the hen, envisioned by Ruth, who allows herself to be burned to protect her chicks (Matt. 23:37; Ruth 2:12). He might not prevent Israel from suffering, but at least he was going to be there too. As N. T. Wright has suggested, Jesus was a revolutionary revolutionary: he would bring about a revolution in a way that overturned all expectations about how revolutionaries should act.Footnote ¹⁶

By intentionally bringing about his own death, Jesus would complete a kind of radical identification with his people. He would choose to suffer the worst of what could happen to them, just as he had chosen to live the life to which he had invited them. Because he had shared in their experiences – including the worst experiences possible – he could forge the kind of relationship that enabled practical authority in another's life. However, the practical authority of Christ in the life of his followers would not just make it the case that they would be better off following his advice. Instead his authority would be based on shared understanding. That, after all, is what makes friends different from servants (John 15:15).

The earliest account of Christ's authority in the Gospels comes from the first chapter of Mark, who reports that Jesus spoke ‘as one that had authority, and not as the scribes’ (Mark 1:22). His authority is there on conspicuous display: meeting a man with an unclean spirit, he rebukes the spirit and frees the man (Mark 1:23–26). The authority of Jesus is in two parts: he understands what afflicts the man, and his words make a difference to changing it. So it is for all who follow Christ (John 6:35; 14:6). Christians can trust and confide in God in the same way that they can trust and confide in their closest friends, and for the same reason: God has shared their experiences. Even more, Christians believe that – in some sense – God has gone through exactly the same suffering that they undergo. It makes more sense to trust in God than in any other person (Ephesians 3:4).

And not only has Christ suffered what we suffer, but he has fitted it already into a meaningful life. We all understand that the significance of events in our lives is shaped by how they are connected to other events (Velleman (Reference Velleman2000); Portmore (Reference Portmore2007) ). Failure that precedes great success can be understood as part of a process of learning. Personal tragedy can be remembered either as ruinous or as character building, depending on what we take from it. But we might sometimes wonder whether there are limits to the extent that independently bad events in our lives can be transformed by what follows them. Are some things too terrible to be redeemed? The message of the atonement is that the answer is no. No matter how bad our suffering is, it is still amenable to being part of a choice-worthy life narrative.Footnote ¹⁷ In fact it already has been made part of a choice-worthy life narrative – that of Jesus himself (Luke 24:1; Colossians 4:5; Galatians 3:13).

Because God has freely chosen to share our experiences, we can see that it is possible to make our experiences part of a life that we would want to have. We can also have the benefit of treating God as a practical authority in our life – the same way as when we have loving relationships with our spouses, or parents, or friends. We can perhaps take comfort in times of suffering by thinking that our pain makes our life more like Jesus's life. Our capacity to understand God may be greater for the sake of our greater shared experience with his. So even seemingly pointless suffering can have at least that meaning. Suffering can be sacralized with devotional import. I can hope, with the hymnist, to be ‘nearer my God, to thee, nearer to thee, E'en though it be a cross, that raiseth me’.Footnote ¹⁸

Epilogue

For a moment God himself was an atheist. So says the Christian narrative of the atonement, anyway. If the story is taken seriously, then the darkest moment of Jesus's life was about abandonment rather than physical pain or even humiliation. We sometimes are wont to say that the atonement must remain incomprehensible. There is some truth to this thought, because in his experience of complete isolation Jesus endured something which, thanks to him, we do not have to confront as well. In fact – if the story is true – we cannot confront it. No matter how isolated we may feel, there is someone else who has faced a loneliness very much like ours. If we stare into the universe and experience it as totally devoid of a supreme being of any kind, God can appreciate even that deeply mortal experience. After all, he has been there, too.Footnote ²¹