Divine Forgiveness As Legal PardonWilliam Lane Craig
I argue that divine forgiveness of sinners, bzw. sins, is more akin to legal pardon than to forgiveness as it is usually understood in human relationships. For divine forgiveness involves not merely a change of attitude (e.g., relinquishing resentment or bitterness) on God’s part but the cancelation of our liability to punishment, which is the function of legal pardon. Pardons are wholly compatible with the prior satisfaction of justice, usually being granted, in fact, long after a criminal sentence has been served. Moreover, pardons may be conditional, requiring for their effect the acceptance of its conditions by the potential pardonee. These features serve to invalidate the neo-Socinian objection to the compatibility of divine forgiveness with atonement theories featuring the satisfaction of divine justice.
The distinction between personal forgiveness and legal pardon is well-known and widely recognized in the literature on forgiveness. In this paper I wish to explore the analogy between divine forgiveness and legal pardon, particularly as it exists in the American justice system. There are at least two reasons for thinking that divine forgiveness implies a legal pardon of sinners on God’s part.
Divine Forgiveness as Legal Pardon
First, God stands in a governmental relationship to human beings. In his classic A Defence of the Catholic Faith concerning the Satisfaction of Christ, against Faustus Socinus (1617) the famed international jurist Hugo Grotius identified Socinus’ “fundamental error” in his critique of traditional atonement theories as his assumption that God is to be construed on the model of an offended party in a personal dispute, such as between a creditor and a debtor (II). For such a private person has no right to punish another. Certainly, God is offended by sin, but He does not act as merely the offended party in punishing it. Rather God should be considered to act as a Ruler. “For 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” (II). God as Supreme Ruler is responsible for the administration of justice in the universe and so has the right of punishing and the right of forgiving wrongdoing. Although God has the right to forgive sins, Grotius thinks it would be unjust of God to let certain sins go unpunished, such as sins of the unrepentant. Therefore, it would be inconsistent with the justice of God that He should remit all punishment whatsoever.
On the contemporary scene legal philosopher Jeffrie Murphy has made a similar distinction between the private and public spheres in an effort to carve out conceptual space for exercises of mercy consistent with the demands of retributive justice. Distinguishing between a creditor in a civil lawsuit and a judge in a criminal case, Murphy maintains that as a litigant in a civil lawsuit, the creditor occupies a “private role” and so does not have “an antecedent obligation, required by the rules of justice, to impose harsh treatment” by demanding repayment of the debt owed (Murphy, 1988, pp. 175-6). He is therefore free to show mercy without prejudice to justice. By contrast a judge in a criminal case “has an obligation to do justice—which means, at a minimum, an obligation to uphold the rule of law. Thus if he is moved, even by love or compassion, to act contrary to the rule of law—to the rules of justice—he acts wrongly” (Ibid., p. 175). Murphy thinks that the judge qua judge cannot, like the creditor, act mercifully without prejudice to the demands of justice. Like Grotius Murphy thinks that the executive power can exercise mercy but only within the limits of individualized justice.
Given God’s status as Judge and Ruler of the world, it is more accurate to think of divine forgiveness on the analogy of a legal pardon by a Ruler rather than on the analogy of the forgiveness extended by a private person. The philosophical literature typically treats forgiveness as a subjective change of attitude or judgement on the part of the person wronged, a determination to put away feelings of resentment, bitterness, or anger, a relinquishing of the desire for revenge or a claim to requital. But God’s forgiveness accomplishes much more than a change of attitude toward sinners on God’s part. Kathleen Moore has made the point forcefully by observing that when people ask God to forgive their sins, they are clearly hoping that God will not inflict the full measure of punishment they know they deserve. “These people would discover the seriousness of their conceptual confusion if God forgave their sins and punished them nevertheless–which is always an option for God” (Moore, 1989, p. 184).
The work of contemporary Christian philosophers exhibits a discouraging Socinian tendency to think of God in terms of a private person involved in a personal dispute, so that they miss the legal character of divine forgiveness as pardon. For example, Eleonore Stump’s approach to the doctrine of the atonement is based entirely on construing God on the analogy of a private person engaged in various personal relationships rather than as a Judge and Ruler (Stump, forthcoming). She frequently compares God and human persons with two friends Paula and Jerome, who have to deal with wrongs committed by one against the other. Focused as she is on private, interpersonal relationships, Stump overlooks entirely the character of divine forgiveness as legal pardon. In fact, Stump’s characterization of forgiveness in subjective terms implies that God’s forgiving sinners is compatible with His exercising retributive justice by punishing those sinners. He both forgives their sins and punishes them for those sins, thereby realizing the nightmarish scenario envisioned by Moore above. By ignoring pardon, Stump can make no sense of the way in which divine forgiveness frees one from the liability to punishment.
My point is that we are apt to make more progress in understanding divine forgiveness by conceiving of God along Grotian lines as Ruler (as well as Lawmaker and Judge) than along Socinian lines as an offended party in a private dispute. God’s forgiving sins will thus have the character of a legal pardon by the executive power of the state.
Second, the consequences of divine forgiveness as described in biblical revelation imply God’s pardon of sinners. The Levitical system of sacrificial offerings in the Tabernacle and Temple, offerings which New Testament writers took to prefigure Christ’s own death as the ultimate sacrificial offering (Rom 3:21-26; 8.3; Eph 5:2; Heb 9.6-14; 10.1-18), aimed, not merely at the cleansing of consecrated objects from impurity, but more fundamentally at the expiation of the sins of the people and their forgiveness. Repeatedly the promise is given, “the priest shall make atonement on your behalf for the sin that you have committed, and you shall be forgiven” (Lev 4.35; cf. 4.20, 26, 31, etc.). At the heart of the new covenant prophesied by Jeremiah lay the forgiveness of sins: “I will forgive their iniquity, and remember their sin no more” (Jer 31.34). Christians considered Jesus, by his sacrificial death, to have inaugurated that new covenant (Mt 26.28; Mk 14:22-24). So in the Acts the consistent apostolic proclamation is that “everyone who believes in him receives forgiveness of sins through his name” (Acts 10.43; cf. 2.38; 5.31; 13.38; 26.18). In short, in Christ “we have redemption, the forgiveness of sins” (Col 1.14; cf. Eph 1.7).
It is noteworthy that the object of divine forgiveness is just as often said to be sins as sinners. Not only are people forgiven for their sins, but their sins are forgiven. God is said to “take away” (aphaireō) our sins (Rom 11.27). This fact makes it evident that divine forgiveness is not (merely) a change of attitude on God’s part toward sinners. Divine forgiveness has as its effect, not (merely) God’s laying aside feelings of resentment or bitterness or anger (or what have you, according to one’s favorite analysis of forgiveness), but rather the removal of the liability to punishment that attends sin. As a result of divine forgiveness, a person who formerly deserved punishment now no longer does so. Because of the forgiveness that is to be found in Christ, one is no longer held accountable for one’s sins. “There is therefore now no condemnation for those who are in Christ Jesus” (Rom 8.1). On the contrary, they are now reckoned by God to be righteous in His sight (Rom 4.5-8). The biblical concept of forgiveness thus entails God’s pardoning people for their sins, freeing them of liability to punishment and constituting them righteous before God.
On the basis of God’s role in the government of the world and the biblical consequences of God’s forgiveness of sins, we ought to think of divine forgiveness, at least in part, on the analogy of a legal pardon. Now, of course, there will be significant disanalogies between divine pardon and the pardoning power as it exists in the American system of justice—for example, the President may issue pardons for personal political advantage—but, still, given the similarities between divine forgiveness and legal pardon, we may expect to gain a good deal of insight into divine forgiveness by exploring the pardoning power vested in heads of government.
Pardon and Its Effects
From ancient times, including the New Testament era, heads of state have exercised the power to pardon crimes. So when the framers of the U.S. Constitution met in Philadelphia in 1787 they naturally included in the Constitution the pardoning power. Since this power is not defined in the Constitution, U.S. courts have interpreted the presidential power to pardon on the model of the pardoning power of English monarchs, which the framers doubtless presupposed. The power of English monarchs to pardon was, in turn, understood as a divine right, an act of grace reflecting God’s ability to pardon sins. In Moore’s pithy conclusion, “Presidents used pardons as they chose, having been given a pardoning power patterned after that of the English Kings, which was patterned after God’s” (Moore, 1989, p. 51). So it is not surprising that the power of the executive to pardon strongly resembles divine pardon.
Chief Justice John Marshall, in a landmark decision, describes a pardon as follows:
A pardon is an act of grace, proceeding from the power entrusted with the execution of the laws, which exempts the individual, on whom it is bestowed, from the punishment the law inflicts for a crime he has committed (United States v. Wilson, 32 U.S. 150 (1833)).
Marshall’s description was later cited by the Supreme Court as a correct characterization in Burdick v. United States, 236 U.S. 79, 89 (1915). According to this characterization a pardon is an act of mercy, coming from the person(s) possessing the power of the executive, which removes a criminal’s liability to punishment for a specific crime he has committed.
Marshall’s description seems an apt characterization of a divine pardon as well. God is the power Who executes His divine torah, and His pardon is an act of grace by which He exempts elect sinners, who have violated His law, from the punishment they deserve. Every element of Marshall’s definition finds a theological analogue. No wonder Daniel Kobil characterizes Marshall's vision of a pardon as “something akin to divine forgiveness” (Kobil, 1991, p. 594)!
What are the effects of a pardon? Marshall says that it exempts the individual from the punishment prescribed by the law for his crime. This much is uncontroversial. But pardons do much more than merely exempt a convicted criminal from punishment for his crime. A pardon removes all the legal consequences of the criminal’s conviction. A pardon thus restores to a person any civil rights which were restricted as a result of his conviction, such as the right to vote, to serve on a jury, or to obtain a business license (Knote v. United States 95 U.S. 153 (1877)). We shall return to the effect of a pardon in restoring a person’s civil rights, a feature of pardons which is also uncontroversial, even if in some cases difficult to adjudicate.
The truly controversial question is whether a pardon serves to remove the criminal’s guilt. Following the English model, the U.S. courts were at first emphatic as to the effect of a pardon in expiating guilt. In Ex parte Garland (1866) the Supreme Court famously declared:
. . . the inquiry arises as to the effect and operation of a pardon, and on this point all the authorities concur. A pardon reaches both the punishment prescribed for the offence and the guilt of the offender; and when the pardon is full, it releases the punishment and blots out of existence the guilt, so that in the eye of the law the offender is as innocent as if he had never committed the offence. If granted before conviction, it prevents any of the penalties and disabilities consequent upon conviction from attaching; if granted after conviction, it removes the penalties and disabilities, and restores him to all his civil rights; it makes him, as it were, a new man, and gives him a new credit and capacity (Ex parte Garland, 71 U.S. 333, 380-1 (1866)).
Like Marshall’s description of a pardon, this characterization of the effects of a full pardon is an apt description of a divine pardon. God in His mercy is similarly said to “blot out my transgressions. . . and blot out mine iniquities” (Ps 51. 1, 9 KJV). Paul exults, “If anyone is in Christ, he is a new creation; the old has passed away, behold, the new has come” (II Cor 5.17). The pardoned sinner’s guilt is expiated, so that he is legally innocent before God.
But as a description of the effects of human pardons, Garland’s sweeping assertions have been eroded by subsequent court decisions. In the Harvard Law Review of 1915 Samuel Williston published what has been called a “seminal” and “landmark” article, “Does a Pardon Blot Out Guilt?,” in which he criticized Garland and its judicial progeny and which has been frequently cited by the courts. Williston complained, “Everybody. . . knows that the vast majority of pardoned convicts were in fact guilty; and when it is said that in the eye of the law they are as innocent as if they have never committed an offense, the natural rejoinder is, then the eyesight of the law is very bad” (Williston, 1915, p. 648). The truth, says Williston, is rather as Lord Coke wrote: Poena mori potest, culpa perennis erit. A moment’s reflection suggests that Williston must understand by “guilt” simply the property or fact of having committed the crime. On this understanding, to be guilty of a crime is just to have committed the crime.
That this is how Williston understands guilt is evident from the remainder of his article. He blames the verdict of the English Court in Cuddington v. Wilkins (80 Eng. Rep. 231 (K.B. 1615)) as laying the main foundation for the view that after a pardon the law could not see the criminal’s guilt. Cuddington had brought an action against Wilkins for calling him a thief. Wilkins justified this appellation because Cuddington had once been convicted of theft. But Cuddington replied that he had been pardoned by the king for the alleged felony. The Court decided for Cuddington, “for the whole court were of opinion that though he was a thief once, yet when the pardon came it took away, not only poenam, but reatum.”
Williston disagrees. According to Williston,
The true line of distinction seems to be this: The pardon removes all legal punishment for the offense. Therefore if the mere conviction involves certain disqualifications which would not follow from the commission of the crime without conviction, the pardon removes such disqualifications. On the other hand, if character is a necessary qualification and the commission of the crime would disqualify even though there had been no criminal prosecution for the crime, the fact that the criminal has been convicted and pardoned does not make him anymore eligible (Williston, 1915, p. 653).
The point is this: a pardon removes the legal disqualifications (abridgement of civil rights) resulting from the fact of conviction; but a pardon does not affect any disqualifications resulting from the commission of the crime. The fact that a crime has been committed cannot be erased. It is this fact that Williston identifies as guilt. Though pardoned, the person still stole or lied or acted recklessly and so remains guilty of the crime he committed. As such he may, despite his pardon, be disqualified from certain activities, such as giving testimony or practicing law.
Henry Weihofen in a later review, citing Williston’s criticism, complains of “the mischief that results when a court applies literally the unfounded dictum of Ex parte Garland that a pardon ‘blots out’ guilt, and makes the offender a ‘new man’, etc.” (Wiehofen, 1939, p. 181; cf. pp. 189-90). The effect of a pardon (other than on grounds of innocence) is “to absolve from further punishment and restore civil rights, but not to undo what is past or blot out of existence a fact, namely, that the person has committed a crime and been sentenced and punished for it” (Ibid., my emphasis).
An examination of various district, state, and appellate court cases walking back the assertions of Garland reveals that the courts in such cases tend to presuppose this same understanding of guilt as the property of having committed a crime. These cases have typically to do with whether a pardon serves to expunge one’s criminal record or to remove a particular disqualification (such as disbarment, banishment from the trading floor, or denial of veteran’s benefits) suffered by the pardonee as a consequence of his being convicted of the crime for which he received a pardon. In holding that Garland overstepped in asserting that a pardon blots out guilt because a pardon does not blot out the past conduct leading to the conviction, these courts equate guilt with having carried out the conduct which led to the conviction.
While such an understanding of the word “guilt” may accord with much of ordinary language, a little reflection reveals that, given standard retributive theories of justice, such a conception of guilt has bizarre consequences. For on this view a person’s guilt could never be expunged, whether by pardon or punishment. Even if a person has served his full sentence and so satisfied the demands of justice, he remains guilty, since it will be ineradicably and forever the case that once upon a time he did commit the crime. But then on standard theories of retributive justice,  he still deserves punishment! For it is an axiom of retributive theories of justice that the guilty deserve punishment. Such an understanding of guilt would thus, in effect, sentence everyone to hell, even for the most minor of crimes, since guilt could never be eradicated and, hence, the demands of justice satisfied. Indeed, even a divine pardon would not serve to remove guilt and save us from punishment, since even God cannot change the past. But such a conclusion is incoherent, since it is the function of pardon to cancel one’s liability to punishment. Therefore, this understanding of guilt is incompatible with standard theories of retributive justice.
The Garland court and its progeny should not be thought to consider a pardon to be a sort of judicial time machine, capable of erasing the past. It is logically incoherent to bring it about that an event which has occurred has not occurred, and it would ungracious to attribute to our courts the absurd opinion that a pardon can erase from the past a person’s wrongdoing or conviction for a crime. Rather what the Garland court was doing, and what its detractors have failed to do, is what contemporary philosophers of time call “taking tense seriously.” When the Supreme Court declared that a pardon “blots out of existence the guilt, so that in the eye of the law the offender is as innocent as if he had never committed the offence,” it takes seriously the tenses of the verbs involved. It recognizes that the offender was guilty, but as a result of his pardon he is now innocent in the law’s eyes. Moreover, the counterfactual conditional “as if . . .” reveals that the law is not blind to his offense. The law can see his offense, but as a result of the pardon the offender is now as innocent as he would have been if he had never committed the offense.
From the beginning courts which held that a pardon expunges a person’s guilt recognized the importance of tense. In Cuddington v. Wilkins, for example, the court opined that while Cuddington was once rightly called a thief, as a result of the king’s pardon he should no longer be called a thief. In Hobart’s report on the case, we read, “It was said, that he could no more call him thief, in the present tense, than to say a man hath the pox, or is a villain after he be cured or manumised, but that he had been a thief or villain he might say.” The court’s decision turns upon taking tense seriously.
Moreover, contrary to the opinions of several lower courts, Garland is wholly consistent with the Supreme Court’s opinion in Burdick v. U.S. that the pardon of an accused person, if accepted, actually implies his guilt (otherwise there would be nothing to be pardoned), for Garland has no interest in denying that the offender was guilty, so that the pardon, in taking away his guilt, implies that he was guilty. A pardon does not have an “appellate” function, as the courts have recognized, in that it does not imply a miscarriage of justice; the correctness of the guilty verdict rendered is not undermined. But now the person is pardoned, and so the effect of that verdict is canceled: though once guilty, the pardonee no longer is.
The opinion in Garland was properly explicated in In re Spenser (1878) as follows:
This is probably as strong and unqualified a statement of the scope and efficacy of a pardon as can be found in the books. And yet I do not suppose the opinion is to be understood as going the length of holding that while the party is to be deemed innocent of the crime by reason of the pardon from and after the taking effect thereof, that it is also to be deemed that he never did commit the crime or was convicted of it. The effect of the pardon is prospective and not retrospective. It removes the guilt and restores the party to a state of innocence. But it does not change the past and cannot annihilate the established fact that he was guilty of the offence (In re Spenser, 22 F. Cas. 921, 922 (1878)).
The opinion in Garland is thus fully in accord with the prevailing view that a pardon has no effect upon the criminal conduct and conviction of the person pardoned. Garland is thus in accord with the prevailing opinion that a pardon serves to release a person from all the legal consequences of his conviction, including punishment, taken in abstraction from the wrongdoing itself.
It is obvious that the Garland court has a very different conception of guilt than lower courts which see themselves as departing from Garland. Rather than equate guilt with the facticity of a past event, Garland assumes that guilt is a property which can be temporarily exemplified and then lost though pardon or appropriate punishment. So what is this property? It seems to me that the most perspicuous understanding of guilt in this sense is that it is liability to punishment. Guilty verdicts in cases of strict liability (in which there may be nether wrongdoing nor culpability) show that guilt cannot be equated merely with culpable wrongdoing. Rather a verdict of “Guilty” is plausibly a declaration that the person is legally liable to punishment. To be guilty of a crime is to be legally liable to punishment for that crime. Such an understanding of guilt makes it perspicuous why punishment or pardon serves to expiate guilt. A person who has served his sentence has “paid his debt to society” and so is now no longer guilty, that is to say, no longer liable to punishment. Similarly, a person who has been pardoned is by all accounts no longer liable to punishment for the crime he committed.
To return, then, to the concerns of theology, it seems to me that Garland’s statement of the effects of a pardon is a marvelous description of the effects of a divine pardon of a person’s sins. By taking tense seriously, we understand how a person who was once guilty may, in virtue of a pardon, be no longer guilty, despite the ineradicable fact that he did commit the sin for which he was justly condemned. The decisions of certain lower U.S. courts do not compromise Garland, for they are assuming a different understanding of guilt which equates guilt with the facticity of the past offense, which Garland would not think to deny. Like punishment, pardon expiates a person’s legal guilt, so that he is no longer condemned and liable to punishment.
These debates over the effects of a pardon provide insight into the nature of divine justification. Our legal pardon by God no more transforms our character and makes us virtuous people than does a human pardon a convicted criminal. Again and again, the courts have insisted that a person may suffer various disabilities, despite his pardon, because of the flawed character that led to his conviction. The conviction alone, now pardoned, may not serve as grounds of disability, but it may serve as evidence of a corrupt character and conduct that are disabling. So, for example, in the case In re Abrams Elliott Abrams was deemed unfit to practice law despite his pardon because a pardon did nothing to restore the moral character necessary for him to continue to practice law. Such cases nicely illustrate Williston’s point that “while pardon dispenses with punishment, it cannot change character, and where character is a qualification for an office, a pardoned offence as much as an unpardoned offence is evidence of a lack of the necessary qualification” (Williston, 1915, p. 657).
Similarly, while a divine pardon makes us legally innocent before God, free of liability to punishment, it is powerless of itself to effect moral transformation of character. To that end we need regeneration through the Holy Spirit and His sanctifying influence to make us over time into the men and women that God wants us to be. Sanctification is not a forensic transaction but a moral transformation of character and is not therefore wrought by divine pardon alone.
The Justification of a Pardon
Another controversial question is whether pardons are acts of mercy, and if so, what justifies such an act of clemency. H. R. T. Roberts provides a rough working explication of acting mercifully: “In all justice I am entitled to A from x, but it is mine to exact and I choose not to” (Roberts, 1971, p. 353). Alwynne Smart would add that the choice is made “solely through benevolence,” and not, for example, out of constraint, self-interest, or ulterior motives (Smart, 1968, p. 359). Samuel Morison makes the application to executive pardons:
The institutional expression of mercy through executive clemency means . . . the partial or complete mitigation of justly imposed punishment (including the removal of the collateral consequences attendant upon a felony conviction) by the chief executive on non-retributive grounds, that is to say, for reasons which do not necessarily have anything to do with what a criminal justly deserves as punishment for the commission of a particular offense (Morison, 2005, pp. 18-19).
The central question to be answered here, in Moore’s words, is this: given a retributivist theory of justice and of the role of the state, under what conditions is a pardon justified and under what conditions is it not justified? (Moore, 1989, p. 9). I hope to show that this question has profound theological significance.
As we have seen, early Supreme Court opinions considered pardons to be acts of mercy on the part of the executive power. The landmark decision in this respect was United States v. Wilson (1833), in which Chief Justice Marshall wrote:
A pardon is an act of grace, proceeding from the power intrusted with the execution of the laws, which exempts the individual, on whom it is bestowed, from the punishment the law inflicts for a crime he has committed. It is the private, though official, act of the executive magistrate, delivered to the individual for whose benefit it is intended, and not communicated officially to the court (U.S. v. Wilson, 32 U.S. 150, 160-1(1833)).
In this opinion we see, in reliance on English precedent, the characterization of a pardon as an act of grace and as a private, though official, transaction between the executive and the criminal.
According to Humbert, “In virtue of the stress which Marshall placed upon the grace and upon the private character of the presidential act, mercy or grace became, in strict legal theory, the reason for a pardon” (Humbert, 1941, p. 22). Pure retributivists like Kathleen Moore have, however, sharply challenged the validity of pardons issued solely on grounds of mercy. These theorists argue that pardons given for any other reason than furthering justice is of necessity unjust and therefore immoral, even if legal. In particular, pardons given out of mercy violate the principles of (positive) retributive justice because in such cases the guilty do not receive their just desert. On retributive theories of justice it is axiomatic that the guilty deserve punishment. To pardon someone out of mercy is therefore to subvert justice and so to act unjustly. Pardons are appropriate on retributive grounds only in cases of innocence, excusable crime, justified crime, and prevention of undeserved suffering.
The claim of the pure retributivists has enormous theological implications for divine pardon. For God is portrayed in the Bible as acting mercifully toward us and His pardoning our sins as an act of grace (Eph 2.8-9; Rom 9.16). Although we deserve condemnation and death, having neither excuse nor justification for our breaking of His law (Rom 1.32; 2.1-3; 3.20), God out of His great mercy has pardoned our sins and graciously reckoned us righteous. Thus, a divine pardon is, indeed, in Marshall’s words, “an act of grace.”
At the same time, the Bible portrays God as a positive retributivist with respect to justice (Exod 34.7). God’s judgement is described in the Bible as ultimately eschatological. The ungodly are “storing up wrath” for themselves for God’s final day of judgement (Rom. 2.5). Punishment imposed at that point could seemingly serve no other purpose than retribution. In any case, the biblical view is that the wicked deserve punishment—“those who do such things deserve to die” (Rom 1.32)—, so that retributive justice belongs to God’s character.
Indeed, it is plausible, I think, that retributive justice belongs essentially to God. Brian Leftow observes that “the more central and prominent an attribute is in the Biblical picture of God, the stronger the case for taking it to be necessary to being God, ceteris paribus: this is the only reason philosophers usually treat being omniscient or omnipotent as thus necessary” (Leftow, 2012, p. 412). It is hard to think of an attribute more central and prominent in the biblical picture of God than His righteousness or justice (Owen, 1653). “Shall not the Judge of all the earth do right?” (Gen 18.25). “Is there injustice (adikia) on God’s part? By no means!” (Rom 9.14). It would have been inconceivable to the biblical authors that God might act unjustly.
But then God faces “the dilemma of the merciful judge:” when a judge tries to treat an offender mercifully, either the offender is given the penalty he deserves (in which case he is being shown justice, not mercy) or the offender is not given the penalty he deserves (in which case the judge acts unjustly). Thus, a judge in his official capacity cannot exercise real mercy; his choice is between being just or unjust. God in His capacity of Judge acts in conformity with the strict demands of justice, so that we sinners find ourselves condemned before His bar (Rom 3.19-20).
The official remission of punishment can be justified only through pardon by the executive power. Since God is both Ruler and Judge, He is as Ruler in the rather odd situation of undoing His own verdict as Judge. Not that He second guesses the Judge’s determination of guilt, for He as Judge is infallible in His determination of justice. Justification should not, despite careless statements by some New Testament scholars, be thought to be a verdict of acquittal. The guilty verdict stands. But as Ruler God pardons us, so that whereas we were once guilty, we are now innocent before Him.
But now God faces a similar dilemma to the one above: if the pardon is given by the executive to rectify some injustice, then the pardon is not an act of grace given out of mercy but is an expression of justice; but if it is given out of mercy, then the executive violates the principles of retributive justice and so is unjust. Clearly, God cannot give pardons to rectify some injustice, since His judicial condemnation of sinners is perfectly just. If He pardons, it must be out of mercy. But then He would seem to be acting unjustly. But given that retributive justice belongs to God’s character, it is impossible that He so act. He must give people what they deserve, on pain of acting contrary to His own nature.
Critics of the pure retributivists have argued that the demands of retributive justice can be overridden by other considerations, so that the executive who pardons out of sheer mercy is not immoral. In a recent, lengthy review of the question Samuel Morison argues that sometimes leniency is morally justified when satisfying the prima facie demands of retributive justice would be immoral or practically impossible. What is striking about Morison’s concerns is that none of them, such as protecting people against self-incrimination, preventing unreasonable searches and seizures, and so on, is remotely relevant to the case of God’s administration of justice. In fact, it is telling when Morison quotes approvingly Murphy’s declamation, “The liberal tradition would thus view it as silly (and perhaps impious) to make God’s ultimate justice the model for the state’s legal justice; and thus any attempt to identify criminal with sinner is to be avoided” (Murphy, 1985, cited by Morison, 2005, p. 84). In fact, Morison states plainly, “the pursuit of the legitimate interest in securing social peace via state-sponsored legal punishment (as distinguished from divine retribution) does not entail any prima facie obligation to exact the full measure of morally justified punitive suffering merely because the offender deserves it” (Morison, 2005, p. 86). Morison thus recognizes God’s obligation to exact the full measure of morally justified punitive suffering.
Morison’s defense of pardons on grounds of mercy is the fullest I have encountered in the literature, and yet it is stunningly irrelevant, as he recognizes, to the case of divine pardon. In the end Morison rejects “the implicit conflation of morality and justice, which assumes that the legitimate exercise of mercy always must be consistent with the demands of justice” (Morison, 2005, p. 100). He cites George Rainbolt: “The fact that mercy counsels unjust acts on occasion does not imply that it is a vice. It only reflects the unfortunate fact that mercy and justice can conflict” (cited by Morison, 2005, p. 101). But that is precisely the problem for the Christian theist: God’s justice and mercy are both essential to Him and so neither can be sacrificed. We can agree with Morison “that the moral basis for the merciful extension of clemency is thus whatever ‘is right and good as judged against all moral considerations, rather than only those of justice. Any pertinent moral consideration may be taken into account’” (Morison, 2005, p. 104, citing Andrew Brien, with emphasis added). One should not, indeed, simply identify morality with justice. But none of the considerations that Morison has adduced for tempering justice with mercy in the case of the state applies to God. So how can God legitimately exercise mercy if doing so is inconsistent with the demands of His justice? Morison admits that that “there is no tidy conceptual solution to the problem of reconciling justice and mercy in the abstract” (Morison, 2005, p. 102). He concludes that “the practice of punishment is informed by a plurality of values that may not be ultimately commensurable” (Ibid.).
If none of the reasons that go to justify pardons based on mercy rather than on justice applies in the case of divine pardon, then it is difficult to see how God can mercifully pardon sins; indeed, it is difficult to see how divine pardon is possible at all, since neither can it be justified on grounds of justice. What seems to be needed is a way of reconciling divine mercy and justice which justifies a pardon without sacrificing the demands of either virtue.
We thereby seem to have backed into a persuasive argument for the conviction of Anselm and the Reformers that the satisfaction of divine justice is a necessary condition of salvation. Theologians have long debated the question of whether God could have simply pardoned our sin without Christ’s atoning death or, more broadly, without the satisfaction of divine justice. Thomas Aquinas followed most of the early Church Fathers in thinking that this is possible, although less suitable for God’s purposes. Following the rise of Socinianism, most Protestant theologians, with the notable exception of Hugo Grotius, followed Anselm’s lead in holding that divine justice had to be satisfied if salvation from sin were to be possible. Our inquiry suggests the following argument in support of the necessitarian perspective:
1. Necessarily (Retributive justice is essential to God).
2. Necessarily (If retributive justice is essential to God, then God justly punishes every sin).
3. Necessarily (If God justly punishes every sin, then divine justice is satisfied).
4. \ Necessarily (Divine justice is satisfied.)
5. \ Necessarily (If some human beings are saved, divine justice satisfied).
Let me say a word about each of the premises.
In support of (1) we have seen that the centrality and prominence of divine retributive justice in the biblical scheme supports its being essential to God. Moreover, to mention an ad hominem consideration, neo-Socinian opponents of penal substitution need (1) if they are to argue successfully for the injustice of penal substitution, for otherwise God may determine that it is not unjust to punish a substitute in our place. Given that there is no higher law to which God must conform, He will be bound only by His own nature in determining what is just or unjust.
The support for (2) lies in the absence of any apparent justification for pardons of sheer mercy on God’s part. It is difficult to see what would justify waiving the demands of retributive justice essential God’s nature. We say “justly punishes” to ensure the truth of (3), since only proportionate punishment of sins committed will satisfy the demands of retributive justice.
From the three premises, (4) follows. Divine justice is satisfied so long as no sin goes unpunished. This will be the case whether there are no human beings and, hence, no sin, or whether there are in fact sinners. (5) in turn follows, since any proposition implies a necessary truth. It also follows that if divine justice is not satisfied, then no human beings are saved; indeed, that it is impossible that any human beings are saved.
If this is right, then God’s pardoning us for our sins demands the satisfaction of God’s justice. This is exactly what the atonement theories of Anselm and the Reformers offer. On the Reformers’ view Christ as our substitute and representative bears the punishment due for every sin, so that the demands of divine retributive justice are fully met. The demands of divine justice thus satisfied, God can in turn pardon us of our sins. God’s pardon is thus predicated on Christ’s satisfying for us the demands of divine retributive justice. Indeed, in a sense, such a divine pardon meets the requirements of even the pure retributivists, for given Christ’s satisfaction of divine retributive justice on our behalf, nothing more is due from us. God’s pardon of us is therefore required by justice. On the other hand, God’s provision of Christ as our penal substitute is an active expression of God’s mercy and grace, giving us what we did not deserve. The whole scheme is motivated by and justified by God’s grace: “For by grace you have been saved through faith, and this is not your own doing; it is the gift of God—not the result of works, so that no one may boast” (Eph. 2.8-9). This atoning arrangement is a gift of God to us, not based on human merit. In this sense God’s pardon of us, while consistent with divine justice, is a pardon grounded ultimately in mercy.
The Anselmian and Reformation theories of the atonement came under withering attack by Socinus in his On Jesus Christ our Savior (1578). In part III of that work, Socinus argues that divine forgiveness is incompatible with the satisfaction of divine justice. He not only disputes the contention that satisfaction of divine justice is a necessary condition of the remission of sins (III.1), but argues further that satisfaction is actually logically incompatible with the remission of sins (III.2). For remission, by definition, entails that the creditor forgoes satisfaction of the debt owed him and that the debtor is forgiven of his obligation. If, on the other hand, the debt is fully paid, then there remains nothing to remit and so nothing to forgive. “There is no need for remission—indeed, remission is an impossibility—where the debt no longer exists” (III.2). So it is an impossibility that our debt be simultaneously both paid by Christ and remitted by God.
Although Socinus’ claim that satisfaction of divine justice is incompatible with divine forgiveness of sin is not an objection raised by many contemporary critics, an exception is Eleonore Stump, who in her recent book At-Onement presses several Socinian objections, including this one, to atonement theories featuring Christ’s satisfaction of divine justice. Stump claims that such theories, despite their asseverations to the contrary, “do not in fact seem to present God as forgiving human sin” (Stump, forthcoming). She argues, “For someone to forgive a debt or to forego a penalty or a penance is for him to fail to exact all that in justice is owed him. But, on interpretations of the Anselmian kind, God does exact every bit of what is owed him by human beings; he allows none of it to go unpaid” (Ibid.). To illustrate: “Suppose that Daniel owes Susan $1000 and cannot pay it, but Susan’s daughter Maggie, who is Daniel’s good friend, does pay Susan the whole $1000 on Daniel’s behalf. Is there any sense in which Susan can be said to forgive the debt?” (Stump, 1988, p. 62). It “is hard to see what constitutes forgiveness on this claim” (Stump, forthcoming). Therefore, “the penal substitution theory of the atonement does not, in fact, present God as forgiving human sin” (Stump, 1988, p. 57).
We have seen, however, that Socinus and contemporary neo-Socinian thinkers err in thinking of God merely as a party in a private dispute, such as a creditor to whom a debt is due, and neglecting God’s role as Judge and Ruler. Given God’s role(s) in the government of the world as described in the New Testament, divine forgiveness is much more akin to an executive pardon than to the remission of a debt or the forgiveness of an offense. So can God pardon us of our sins if Christ has satisfied divine justice by being vicariously punished for those sins, as the Reformers maintained?
Pardons granted on grounds of innocence and wrongful conviction already show that a pardon is wholly compatible with the demands of justice being satisfied. Pardons to achieve remedial justice do not imply the failure of the person involved to satisfy the demands of justice. Indeed, quite the opposite is the case. Moreover, the vast majority of pardons are granted after the criminal’s sentence has been fully paid. The U.S. Office of Pardon Attorney will not even permit applications for a presidential pardon until at least five years have elapsed since the sentence of the criminal has been fully satisfied. A pardon in such a case does not imply that the pardonee has failed to satisfy justice’s demands. As in a case of double causation, both his punishment and his pardon are sufficient to annul his guilt and restore him to innocence. But a pardon also serves to restore to him all his civil rights voided by his conviction. Similarly, a divine pardon serves to bestow upon us the full rights and privileges of a child of God, such as adoption into God’s family (Eph 1.5), an inheritance in heaven (I Pet 1.4), citizenship in God’s Kingdom (Phil 3.20), access to the Father (Rom. 5.2), and so on (all, interestingly, legal notions). Indeed, precisely because Christ has met the demands of divine justice, God can be both “just and the justifier of him who has faith in Jesus” (Rom 3.26). Because God’s justice has been fully satisfied, God can pardon us on the basis of Christ’s sacrifice without prejudice to His justice. Paul says, “when you were dead in trespasses. . . , God made you alive together with him, when he forgave us all our trespasses, erasing the record that stood against us with its legal demands. He set this aside, nailing it to the cross” (Col 2.13-14). Forgiveness in this legal sense is based on the fact that the penalty has been fully paid and therefore we may be pardoned.
Because it is Christ and not we who has discharged the sentence for our sins, our guilt is not expiated unless and until we receive God’s pardon. In contrast to the criminal who has fulfilled his sentence, we remain in our state of judicial condemnation until we accept the pardon offered us by God.
U.S. courts remain unclear on whether a pardon, in order to be effective, must be freely accepted by the criminal who is the intended beneficiary of the pardon. Courts and legal theorists have tended to answer this question based on whether a pardon is considered a private communication of the executive to the criminal or a public proclamation of the executive which is known to the court. For better or worse, this question is unlikely to come before the Supreme Court today, since pardons are virtually always given in response to applications to the Office of Pardon Attorney and not bestowed upon unwitting criminals. People who receive pardons are those who want them.
The theological analogue to this question is whether a divine pardon must be freely accepted in order to be efficacious. Taking divine pardons to be acts of grace does not serve to resolve this question, since theologians have differed on whether grace is intrinsically efficacious and so irresistible by him upon whom it is bestowed or whether grace is extrinsically efficacious and so requires the free consent of the creaturely will in order to produce its effect. Given God’s love for those He pardons, God’s pardons are intensely personal and in this sense private. Though official acts, they are motivated out of concern for the individual and not just for the general welfare. Whether a divine pardon requires acceptance by the person to whom it is granted is going to depend more on theological considerations such as freedom of the will and the nature of divine grace than upon the nature of a pardon.
In any case, what has not been mentioned thus far is that pardons may be conditional, in which case they undisputedly depend for their effect upon the pardonee’s agreeing to the conditions of the proffered pardon. While the President may not demand just any condition for a pardon—for example, to vote forever after for the President’s political party—, nevertheless the conditions which the President may lay down for a pardon are endless. In fact, no federal court has ever held any condition invalid. A divine pardon, then, can be granted on the conditions of repentance and faith. If God desires people to come freely into His Kingdom, then He may offer His pardon to everyone who will freely accept it. Anyone who refuses a divine pardon therefore does not benefit from it and so remains liable to punishment. If anyone refuses the pardon offered by God, then Christ’s sacrifice avails him nothing, for he has rejected the satisfaction of God’s justice wrought by Christ.
A divine pardon is rooted in God’s love and grace, for it is by His mercy that God determines to supply a satisfaction of His justice that we might in turn be pardoned. Stump realizes that the penal substitution theorist will insist that “God’s justice precludes his overlooking the debt and that therefore he has shown mercy and forgiveness. . . by he himself paying the debt owed him” (Stump, forthcoming). Her response falters at this point; rather than show that such an act does not count as mercy and forgiveness, Stump instead turns to a different Socinian objection, namely, that it would be unjust for God to punish an innocent person like Christ. She thereby fails to sustain her objection that God’s legally pardoning us on the basis of Christ’s payment does not plausibly count as mercy and forgiveness of sins. As for the justice of God’s punishing a substitute in our place, that is a discussion for another day., 
 See, e.g., Hughes, 2014, §3.1.
 For a helpful survey of the literature on the nature of divine forgiveness, see Warmke (forthcoming a & b). Taking divine forgiveness to entail pardon does not preclude taking God to experience a change of attitude as well.
 On Stump’s view love involves both (i) a desire for the good of the beloved and (ii) a desire for union with him. Just as a person can love unrequitedly, so he can forgive unilaterally, despite the wrongdoer’s rejection of that forgiveness. So on her account of love and forgiveness, she concludes, “it is possible to hold that imposing retributive punishment on a wrongdoer is sometimes required by justice, and still to maintain that love and forgiveness are obligatory even for wholly unrepentant wrongdoers” (forthcoming, p. 67).
 We encounter here the debate over whether the Levitical sacrifices and Christ’s sacrificial death served to propitiate God, to change His attitude toward sinners from wrath to acceptance. It has become conventional wisdom among contemporary theologians that because the New Testament authors use katalassō (“reconcile”) and its cognates only with respect to human beings, not God, God does not need to be reconciled to humanity, but only humanity to a welcoming God. I leave aside whether such an argument from silence is cogent. But if God does not need to be reconciled to sinners, that fact shows all the more that divine forgiveness is not a change of attitude on God’s part, in the way that forgiveness is usually understood by contemporary philosophers analyzing human relationships.
 For a thorough review of the relevant judicial decisions see In re Sang Man Shin, 125 Nev. 100, 104-9 (2009); Robertson v. Shinseki, 26 Vet. App. 169, 176-9 (2013).
 “Punishment may expire, but guilt will last forever.”
 Hob. 67, 81, cited by Williston, 1915, p. 651.
 See, e.g., Groseclose v. Plummer 106 F.2d 311, 313 (9th Cir.1939); People ex rel. Prisament v. Brophy 287 N.Y. 132, 137-8 (1941); State Ex Rel. Wier v. Peterson, 369 A.2d.1076, 1080, 1081 (Del.1976); Dixon v. McMullen 527 F. Supp. 711, 717-18 (N.D.Tex.1981); In re Abrams, 689 A.2d 6, 7, 10-11 (D.C. 1997); R.J.L. v. State, 887 So.2d 1268, 1280-81(Fla.2004); Hirschberg v. Commodity Futures Trading Com'n, 414 F.3d 679, 682, 683 (2005); Fletcher v. Graham, 192 S.W.3d 350, 362-363 (Ky.2006); In re Sang Man Shin, 125 Nev. 100, 110 (2009); Robertson v. Shinseki, 26 Vet. App. 169, 179 (2013)). For discussion of some of these cases see Steiner (1997), who makes the same equation.
 Theories of justice may be classified as broadly retributive or consequentialist. Retributive theories of justice hold that punishment is justified because the guilty deserve to be punished. Consequentialist theories of justice hold that punishment is justified because of the extrinsic goods that may be realized thereby, such as deterrence of crime, sequestration of dangerous persons, and reformation of wrong-doers. Retributivism may be either positive (“the guilty deserve punishment”) or negative (“the innocent ought not to be punished”). There has been over the last half-century or so a renaissance of theories of retributive justice, accompanied by a fading of consequentialist theories.
 The phrase was apparently inspired by the great Oxford tense logician A. N. Prior, who, in reaction to W.V. O. Quine’s extolling the tenselessness of modern logic, praised medieval logic because it “took tenses far more seriously than our own common logic does” (Prior (1958), 117). I’m grateful to Prior scholar David Jakobsen for alerting me to Prior’s article, which was originally Prior’s presidential address to the New Zealand Congress of Philosophy in 1954.
 Hob. 81, 82 (1615), cited in Williston, 1915, p. 652. Williston notes that “The principal case was followed in Leyman v. Latimer, 3 Ex. D. 15 (1877), on very similar facts, and the court upheld the validity of the distinction taken in Cuddington v. Wilkins, between the legality of using the present and the past tense” and yet fails himself to appreciate the importance of this distinction.
 E.g., In re Sang Man Shin, 125 Nev. 100, 105 (2009).
 A number of scholars have noted that pardons differ from other forms of executive clemency in that the latter, unlike pardons, do not negate the criminal’s conviction but leave intact the judgement of guilt. For example, President Carter, in proclaiming an amnesty for Vietnam War draft-dodgers, said poignantly that their crimes have been forgotten, not forgiven. Similarly, recipients of commutations and reprieves remain guilty (Kobil, 1991, p. 577; Stacy Caplow, 2013, p. 299: Messing, 2016, p. 672; Schoenburg, 2016, p. 924). This distinction seems to make sense only if a pardon annuls the guilt of the offender.
 On strict liability see Leigh, 1982; Ormerod, 2011, chap.7.
 For literature see Morison, 2005, p. 4.
 For a theological version of the argument of the pure retributivists against divine pardon see Londey (1986); along with responses by Brien (1989); Geuras (1992).
 See Moore, 1989, p. 192; cf. Murphy, 1988, pp. ***; Roberts, 1971, pp. 352-3; Smart, 1968, pp. 349-53.
 See discussion by Weihofen, 1939, pp. 178-9.
 The situation remains the same as when Humbert concluded that the President “cannot under existing law make a full pardon effective without the consent of the prisoner. The latter must be willing to receive and accept a full pardon before it can be put into effect” (Humbert, 1941, p. 135).
 See Ex parte Wells, 59 U.S. 307, 314 (1855); for comment, see Humbert, 1941, p. 72.
 Humbert, 1941, pp. 74-5, explains that in a conditional pardon the conditions can be either precedent or subsequent. If the conditions are precedent, the pardon becomes operative when the recipient has fulfilled the conditions, but not until then. If the conditions are subsequent, the pardon takes effect upon delivery and acceptance but becomes void upon the violation of the specified conditions. Both precedent and subsequent conditions have theological analogues with respect to justification and perseverance.
 See my (forthcoming).
 I am grateful to Dr. E. Descheemaeker of the University of Edinburgh School of Law for helping to direct me to legal literature on pardon and to Shaun McNaughton at Brown & Streza, LLP for help in obtaining court opinions.