Substitution, Justice, and Love: Reflections on Penal Substitutionary Atonement

Editor’s Note: This is part 6 in our Lyceum Disputation series considering the necessity of Penal Substitutionary Atonement as described in this 2017 SBC resolution. Stay tuned for further installments. As with all our work, the London Lyceum publishes a range of viewpoints to encourage thinking.

I am grateful for the opportunity to contribute some reflections on the SBC statement on penal substitution. Although I am not a Southern Baptist and so in one respect have no dog in this fight, I am a Roman Catholic and as a fellow Christian I have a desire that the gospel of Christ be proclaimed faithfully and wisely by whomever purports to proclaim it. The atonement is at the heart of our shared faith and so statements about it from Christian bodies should be faithfully and wisely written; what they say affects what both believers and nonbelievers think—and so how they react to—the Christian Church and God himself.

The SBC statement, in my view, rightly emphasizes important aspects of the doctrine of the atonement: Christ functioning as a substitute (in some sense) and the justice of God being honored through the atonement. These are Biblical ideas that should be reflected in a Christian church’s doctrines. However, the statement, in my view, errs in pushing a specific interpretation of these two aspects: God’s justice must be honored through punishment, and thus Christ honors God’s justice by being punished in place of humans in his death on the cross. In this essay I will argue not that penal substitution is a false or inadequate account of the atonement, but that it is unwise for a Christian body to assert through a declaration that penal substitution forms an essential part of atonement.

The statement does not define how it understands penal substitution and, as other authors in this series have noted, there are several different ways of developing this sort of view. Penal substitution is perhaps better seen as a family of views. However, members of this family share common features. One such feature is that they rest on a particularly strong view of retributive punishment. Penal substitutionary views hold not just that sinning against God merits punishment, or that it is sufficient to satisfy the demands of justice that God carry out proper punishment. They go beyond these weaker retributionist claims and assert a stronger retributionist claim that the demands of justice require that proper punishment be delivered. Richard Purtill (1990) calls this latter claim ‘strict retributionism’. Other theories of atonement, including satisfaction theories, are consistent with the former two retributive views. Anselm himself seems to endorse both views as he holds that punishment is an appropriate or fitting response to violating God’s will and that God’s justice would be upheld through punishing sinners. Anselm rejects strict retributionism, however, because he holds that God’s justice may also be upheld through the provision of satisfaction for (or reparation for—something done to make up for and repair the effects of) human sin. If adequate satisfaction is provided for human sin, God’s justice is honored just as much as if humans had been punished. (Indeed, on Anselm’s view, it is better that God do things this way; if God punishes humans appropriately, then they will be eternally separated from him, thus ruining his original plan for humanity.) However, defenders of penal substitution think that satisfaction isn’t enough to do proper justice; punishment is required. That further claim requires strict retributivism.

How plausible is strict retributivism? Think about the role of punishment in our lives. Why do parents punish? Not simply retributively, but to send a message about the seriousness of our actions and to educate. However, parents do not think that punishment absolutely must be given, or that their own justice requires that they punish children for disobedience or wrong actions. Punishment can be waived not just for practical reasons (such as its being counterproductive in certain cases), but because the child has repented and offered reparation. In the latter sorts of cases, parents seem to reason as follows (as a parent, I can testify that I reason this way): my child understands the seriousness of what he or she did and worked sufficiently to make up for it. That’s enough. They’ve learned what they need to learn and have rectified things; there is no need to punish them any further.

Let’s now turn to look at punishment in another important context. Why do governments punish? Arguably also not simply retributively. But here, contrary to parental justice, it appears that governmental justice requires that when a government issues a punishment for a given sort of action, the punishment be delivered no matter who performs the action. Why this difference? There’s a lot to say here, but at least part of the answer, plausibly, is that when finite humans govern a large and diverse bunch of finite humans, the more opportunity and incentive there is for injustice to be done when attempting to uphold justice. It is practically impossible for a government to track and fairly rate the repentance and reparative attempts of the many people it governs. Government officials don’t know their citizens nearly as well as parents know their children, which puts the former in a worse position than the latter to judge when someone under their authority has adequately repented and made reparation. In addition, bias and rigging the system for advantage is a much greater threat with governmental activity. And even if a government could manage these issues super well and justly and fairly punish some, while not punishing others who had adequately repented and made reparation, the epistemic distance between the judges and lawyers in a given case and the general citizen will inevitably lead the citizens to misinterpret some judgements as unfair. A system of justice that works needs legitimacy. These pitfalls can be largely avoided by requiring punishment for illegal activity. Of course, parents can fall into these pitfalls too. But given there are fewer people involved, who know each other and the relevant situations far better, and there is far less motivation to favor some children over others, it is generally easier for parents to make judge that a child has genuinely repented, made adequate reparation, and learned their lesson.

Should we think that God is a strict retributionist? God isn’t subject to the above limitations of governmental authorities—limitations that explain (at least in part, perhaps largely) why strict retributionism is plausible for the government. In addition, God is quite often presented in scripture as a parent who has unruly and disobedient children—more often, I would argue, then he is presented merely as an enforcer of law. There is at least a decent case, then, that God treats us more like parents when it comes to dealing with wrongdoing, and strict retributionism is false in that realm.

William Lane Craig, a prominent recent defender of penal substitution, has a response to my argument here, which he presents as an objection to Eleonore Stump’s recent work on atonement. He writes that her view (and by implication my argument above) “is based on construing God on the analogy of a private party involved in various personal relationships rather than as a Judge and Ruler” (2019: 539). He thinks this is a mistake and it fails to do justice to the legal analogies of atonement in scripture. God should be viewed as a ruler and a judge for—and here he traces his argument back to a quote from Grotius—“to inflict punishment, or to liberate any one from punishment . . . is only the prerogative of the ruler as such, primarily and per se; as, for example, of a father in a family, of a king in a state, of God in the universe” (2019: 540).

Craig’s argument, however, fails to touch my argument above and his quote from Grotius explains why. One can at one and the same time regard God as both a party of a personal interaction and a judge and ruler. Grotius counts parents as people who, when they punish their children, act as judges and rulers. Parents are also in personal relationships with their children. And yet here strict retributionism is false. The case against strict retributionism doesn’t at all turn on ignoring God’s role as a judge and ruler.

Here is a summary of my argument so far:

1. Scripture does not demand a penal substitutionary understanding of atonement.

2. There are understandings of justice that do not require penal consequences in order for justice to be served.

3. One of our core scriptural metaphors for understanding God—as a parent who knits humanity together into a family—supports one of these alternative understandings.

4. Given 1, 2, and 3, one of the alternative understandings of justice is a live option for theological reflection. Therefore,

5. One of the alternative understandings of justice is a live option for theological reflection.

(2) and (3) are explicitly supported by my discussion above. The evidence of scripture using parental language to describe God provides some support for (1), but I’m not here able to fully defend (1). However, I believe there is very strong support for (1). (See Belousek-Snyder 2011 and Eberhart 2011 for detailed support.) (4) is compelling; if scripture doesn’t rule out a view, and a key feature of scripture supports the view, then the view is a live option for theological reflection (perhaps provided it isn’t ruled out by clear empirical evidence or evidence from church tradition, but a rejection of strict retributionism isn’t ruled out by either). (5), I would suggest, suffices to show that it would be unwise for a Christian body to issue a statement that endorses penal substitution. It rules out live options for theological reflection, which is bad theologically, but also potentially can result in crises of confidence for Christians who might come to a view of atonement other than penal substitution—and who might find penal substitution deeply unsatisfying. We should avoid creating such crises when there is a case that alternative views are live theological options—especially when there are already enough stumbling blocks in our cultures to create crises of confidence. Let’s not ourselves put a potential stumbling block in the way of believers and potential future believers.

But the argument can be pushed even a little further using an idea deep in the heart of the Reformed tradition. The goal of atonement is reconciliation. Reconciliation involves both parties being in a relationship of peace and harmony; both feeling that justice has been served. Could we humans feel that justice has been served on the penal substitutionary model? Arguably no, for reasons that have been much discussed in the literature. We have no experience with, or intuitions supporting penal substitution where a guilty person deserving death has his penalty of death paid entirely be an innocent third party. Why would God use a method of reconciling us that, on our reflection, involves what we would regard as a great violation of justice—especially if there are other understandings of justice on which a penal substitute is not required? Something like Calvin’s notion of divine accommodation is relevant here. If God and we are to be reconciled, the conditions for reconciliation must be something that we can on reflection endorse as just while in a state of reconciliation. God, then, because he desires reconciliation with us, will want to use a means of creating reconciliation that we can on reflection endorse—he will accommodate his methods to our human perspective. Penal substitution of this sort we seem unable to reflectively endorse. One can always respond that perhaps in the eschaton we will come to reflectively endorse it. But one would only think that if one had excellent reason to believe in penal substitution in the first place, which is precisely what is at issue.

This extension of the argument can be summarized as follows:

6. We have strong intuitions that penally substituting the death of an innocent person for the wrongdoing of a third party would be deeply unjust.

7. If alternative understandings of justice (i.e. alternative to strict retributionism) are live options for theological reflection and we have strong intuitions indicating that penal substitution of Jesus’s death for human sin would be unjust, God would likely atone in a way that is consistent with an alternative understanding of justice. Therefore,

8. God would likely atone in a way that is consistent with an alternative understanding of justice.

If (8) is true, two important things follow. First, the notion of justice required to ground the penal substitution view is likely false. Thus, one of the key supports for the theory is undermined. Second, if one of the key supports for the penal substitution theory is undermined, and there are other theologically live understandings of justice around which a different theory of atonement could be built, it is once again, and even more so, unwise for a Christian body to issue a statement that endorses penal substitution so strongly as to say that “the denial of penal substitution constitutes false teaching that leads the flock astray” (SBC resolution).

It is good that the SBC has focused more attention on teaching about the doctrine of atonement. Christians can benefit from being led to a deeper understanding of this core aspect of our faith. And a different statement that emphasized clear Biblical notions—such as substitution of Christ and the justice of God, perhaps along with others such as God’s parental love—while excising endorsement of penal substitution as necessary for any valid Christian understanding of atonement, could be a great benefit to Christians.



Darrin Belousek-Snyder, Atonement, Justice, and Peace. Eerdmans, 2011.

William Lane Craig, “Eleonore Stump’s Critique of Penal Substitutionary Theories.” Faith and Philosophy 36.4 (2019): 522-544. DOI: 10.5840/faithphil2019364136. Available at:

Christian Eberhart, The Sacrifice of Jesus: Understanding Atonement Biblically. Wipf and Stock, 2011.

Richard Purtill, “Justice, Mercy, Supererogation, and Atonement,” in Christian Philosophy, ed. Thomas Flint. Notre Dame Press, 1990: 37-50.


  • Joshua Thurow

    Joshua Thurow is an associate professor of philosophy at the University of Texas at San Antonio and the Executive Director of the Society of Christian Philosophers. He has published several essays on atonement, the cognitive science of religion, debunking arguments, and a priori knowledge in edited collections and journals such as Philosophical Studies, American Philosophical Quarterly, Philosophical Quarterly, Journal of Analytic Theology, and Faith and Philosophy. He is also a co-editor of The A Priori in Philosophy, and author of the forthcoming book, God and the Problem of Epistemic Defeaters

Connect With TLL

Podcast Channels

You might also enjoy