Suppose that a person has been convicted of a crime and comes before a judge to be sentenced. This person has had a hard life, and the judge feels compassion for him. At first she is tempted to show mercy and impose a light sentence. She resists this temptation, though, and imposes a heavier but just sentence instead. Did the judge act rightly? When, if ever, ought judges to show mercy during sentencing? After the judge imposes her sentence, the criminal asks the head of state (or the appropriate executive body) to pardon him. Moved by the criminal’s plight, the head of state shows mercy and grants his request. The criminal is not free yet, however. His victims go to the judge and ask her to rule that the pardon is invalid. On what basis ought judges to evaluate—or, as lawyers say, “review”—a decision by a head of state to grant (or withhold) mercy?
The first set of questions concerns the judge in her role as sentencer. The second set concerns the judge in her role as reviewer. When thinking about mercy in the criminal justice context, philosophers and criminal law scholars have tended to focus on judges as sentencers. Public law scholars, who on the whole have had less to say about mercy, have focused on judges as reviewers. There has been little dialogue between these groups of scholars, but I want to show we can better understand how judges ought to act in each of their roles by considering them alongside each other and alongside the roles of other institutional actors, particularly heads of state. Most of the discussion below is pitched at a general enough level to be relevant to any legal jurisdiction. When the discussion becomes more specific, I will confine my remarks to common law jurisdictions.
Philosophers disagree as to the proper analysis of many concepts, but the disagreement is especially fierce when it comes to mercy. Some scholars have tried to explain why there is so little consensus about what mercy is; unfortunately, even they disagree.1 This is not the place to propose another analysis of mercy, and I am not able to review, let alone assess, all of the existing ones. What I shall do instead is set out a common core of mercy then add to it by making several somewhat controversial assumptions. This will give me an account that few would accept without qualification but that captures what is often meant by mercy. This chapter is best read not as a defense of any understanding of mercy, but as a reflection on the implications for judges of one possible understanding of mercy.
Everyone accepts that, for you to now be merciful to me, several conditions must be met. They are that (i) there are two acts A and B you can now perform; (ii) you perform A; and (iii) your performing A harms2 me less than your performing B would have harmed me.
Conditions (i)-(iii) are often satisfied even when no act of mercy has been performed. Suppose that a maniac roams the street with a gun in his pocket. Spying me waiting at a traffic light, he decides I will be his first victim. He approaches with the gun in his hand. Before he shoots, he changes his mind and moves away. Conditions (i)-(iii) are met in this scenario, but the maniac has clearly not acted mercifully. What is missing? When you show mercy, you choose not to impose a harm even though doing so is, in a certain sense, permissible. The difficult question is what this “certain sense” is.3 Indeed, this may be the most difficult issue mercy poses. We do not need to resolve it to know that the maniac has not acted mercifully, because it is hard to imagine any notion of permissibility that would sanction killing me in cold blood. But we do need to know what the relevant notion of permissibility is to understand mercy in a general way.
Here I shall make my first somewhat controversial assumption. I shall assume that the relevant notion of permissibility is permissibility according to justice, such that (iv) your performing B would have been just. By “justice” I mean justice according to morality, not (necessarily) justice according to law. I assume that justice in this sense fixes the relevant notion of permissibility for two reasons. First, it is consistent with how most philosophers now think about mercy. Alwynne Smart set the tone for much of the contemporary discussion of mercy when she wrote that a person who shows mercy “decides that a particular punishment would be appropriate or just, and then decides to exact a punishment of lesser severity than the appropriate or just one.”4 Second, even if mercy does not always pose a challenge to justice, it does in the cases that create real dilemmas for judges, and these cases are my concern here.
It is uncontroversial that mercy and compassion are closely connected, but the precise nature of that connection is controversial. One view is that compassion is essential to any merciful act, such that you act mercifully toward a person only if you act out of compassion for that person. Another view is that compassion is essential to a merciful character, such that you are a merciful person only if you are disposed to act mercifully out of compassion for others. If you act out of self-interest, then according to the first view, you have not acted mercifully; according to the second view, you might have acted mercifully, but you will not have shown yourself to be a merciful person. I find the second view more attractive because it explains why a person can act mercifully without being a merciful person. I cannot argue for that view here, however, so I shall assume in what follows that compassion is relevant to whether a person is merciful, not to whether she acts mercifully.5
Some philosophers treat its normative status as part of the concept of mercy; others separate the two issues. I shall take the latter approach for the sake of simplicity. To act mercifully is to act in a way that is (at least prima facie) morally good. This is consistent with either of two possibilities: that mercy is supererogatory or that mercy is the subject of a moral duty. A person who never performs a supererogatory act is not, for that reason, morally blameworthy. By contrast, a person who never acts mercifully is, for that reason, morally blameworthy. So mercy must be the subject of a moral duty.6
There are different kinds of moral duties. Some moral duties give you a degree of discretion or latitude as to when, how often, in what manner, or with respect to whom you must perform some act (or perhaps latitude along other dimensions); others give you no latitude at all. This is the distinction between imperfect and perfect moral duties. Which is the duty to act mercifully: imperfect or perfect? Were the duty to act mercifully a perfect duty, a person would be (at least prima facie) morally blameworthy for not performing a particular merciful act. As George Rainbolt has pointed out, though, this is not how we think about mercy.7 We think people can be criticized for never acting mercifully, but not for failing to act mercifully on this or that occasion. So it would seem that mercy is the subject of an imperfect moral duty, one that requires us to act mercifully some of the time but not all of the time.
There is another reason to think of mercy as the subject of an imperfect moral duty. Mercy is commonly thought to be a “gift” or an act of “grace.”8 It is up to you whether to act mercifully on a particular occasion. No one has a moral right that you show mercy to him specifically; no one may demand mercy, only ask for it. This is exactly what we would expect if mercy is the subject of an imperfect moral duty, given the latitude it confers.
Almost all legal jurisdictions give the head of state (or another executive body9) the power to pardon criminals, i.e., to remit all or part of a criminal’s sentence. In the United Kingdom and some other jurisdictions, this pardoning power is called the “royal prerogative of mercy,” and lawyers and legal scholars in these jurisdictions understandably tend to describe any granting of a pardon as an act of “mercy.” Let us call the grant of a pardon an act of “royal mercy.”
An act of royal mercy always meets conditions (i)-(iii). The head of state will have two acts open to her: granting or refusing a pardon. When she grants a pardon, she harms a criminal either not at all or less than she would have by refusing the pardon. Condition (iv) is more difficult. If a judge has sentenced a criminal to a just punishment, then refusing a pardon to the criminal would be consistent with justice. Condition (iv) is then met. If the head of state pardons the criminal anyway (as in the example in the introduction), then the head of state shows royal mercy and genuine mercy. A judge may impose an unjust sentence, however. (This may not be the judge’s fault. The law may not permit the morally just sentence.) In such a case, for the head of state to refuse a pardon would be inconsistent with justice. So, by granting the pardon, the head of state shows royal mercy but not genuine mercy. In short, if judges have acted justly, then royal mercy will be genuine mercy but otherwise not. In what follows, my concern shall be acts of royal mercy only insofar as they are also acts of genuine mercy.
Suppose the law allows a judge to sentence a thief to either 1 or 2 years in prison. A sentence of 2 years would be just. A sentence of 1 year would be unjust, so imposing that sentence would be merciful. Which sentence ought the judge to impose?
It is often said that a judge has a duty to do justice within the law. This does not mean that a judge has a duty to do justice all the time. She has a duty to do justice while acting as a judge. Doing justice is part of a judge’s “job description,” to use Jeffrie Murphy’s phrase.10 In the role of sentencer, a judge has a duty to impose just sentences, i.e., to sentence criminals to just punishments as far as the law allows. Now, I take it to be intuitive that a judge has a duty to act justly and a more specific duty to sentence justly. (If you are sceptical, though, I ask you to wait until the next section when I give some reasons for these claims.)
As I said, there is also a duty to act mercifully. The duty to act mercifully applies to everyone, so it applies to judges. A judge is not under a duty to act mercifully because she is a judge (or, at least, not only for that reason11). She is under a duty to act mercifully because she is a human being. Thus, with this personal duty in mind, Antony Duff writes that “[m]ercy … is not a virtue of sentencers, qua sentencers; it is not a virtue internal to the role of sentencer within a system of criminal law. It is, rather, a virtue of the human beings who fill that role.”12 A judge who imposes a merciful sentence is guided not by the demands of her office but by her emotions toward the person before her, whom she treats “simply as a suffering fellow.”13
So, judges have two duties: a role-based duty to sentence justly and a non-role-based or personal duty to act mercifully. In cases like the one above, these duties conflict. Which sentence the judge ought to impose—the 2-year just sentence, or the 1-year merciful sentence—depends on how this conflict is resolved.
Some conflicts between duties are resolved based on the relative importance of the duties. Some other conflicts can be resolved by the availability of alternative ways of fulfilling one of the two duties—and this is usually the case when it comes to the conflict between a judge’s role-based duty in justice and a judge’s personal duty in mercy.
To fulfill her duty to sentence justly, a judge must impose a just sentence on every criminal she sentences. If she sentences even one criminal unjustly, she violates her duty. The judge’s duty to sentence justly gives her no latitude as to whom to treat justly, or on what occasions. Her duty to sentence justly is a perfect duty. Her duty to act mercifully, by contrast, is an imperfect duty. The different character of these duties is crucial, because it means a judge can fulfill both duties. All she needs to do is to always impose just sentences, thereby fulfilling her duty to sentence justly, and to perform a considerable number of merciful acts in her personal life, thereby fulfilling her duty in mercy. Although a judge must sometimes choose between justice and mercy in her role as sentencer, she need not choose between fulfilling her duty to sentence justly and fulfilling her duty to act mercifully.
As a general principle, you ought to act in a way that fulfills two duties rather than in a way that fulfills only one of them.14 This suggests that a judge ought to always impose just sentences, even when they are merciless. In short, a judge ought to do justice at work, reserving mercy for her personal life. In terms of the example, above, this would mean that the judge ought to sentence the thief to 2 years. She is acting mercilessly by doing so, but she can make up for it by acting mercifully on other occasions, in her personal life.
To be clear, this is not an argument to the effect that a judge’s role-based duty to sentence justly is more important than her personal duty to act mercifully. The argument is that her duty to sentence justly takes priority because there is an alternate, equally good way for the judge to fulfill her duty in mercy, but no alternate, equally good way for the judge to fulfill her duty in justice.
The “general principle” that a person ought to fulfill two duties rather than just one of them is merely that—a generalization, which may not always hold true. In particular, it may not hold true when there are more than two duties in play.15 A judge might restrict mercy to her personal life to avoid violating her duty in justice, only to find that she cannot act mercifully in her personal life without violating some other duty (e.g., a promissory duty). In such a case, what ought the judge to do?
Without trying to answer this question definitively, there are a few points that stand out. First, these cases will be unusual. Judges have roughly as many opportunities to show mercy in their personal lives as the rest of us, and we do not tend to find it impossible to reconcile our duty in mercy with our other duties. Second, asking “what ought the judge do?” amounts to asking “which duty ought the judge violate?” The answer to that question will depend on which duty is least important. Third, the least important duty may be neither the judge’s duty in mercy, nor her duty in justice, but rather her “third” duty, i.e., the duty she cannot fulfill while acting mercifully in her personal life. In that case it will be unnecessary to decide whether her duty in mercy or justice is more important.
If the judge’s choice comes down to fulfilling her duty in mercy or her duty in justice, there is at least one reason to think she ought to prioritize her duty in justice. Were the judge to do otherwise and give priority to her duty in mercy, she would be using the powers and discretion entrusted to her by the public not for the public good, but to fulfill her personal duty. Normally, we condemn that kind of act as an “abuse of power.” Suppose, by way of comparison, that a friend of the judge’s comes before her for sentencing. She can fulfill her duty in friendship, or she can impose the just sentence. To my mind, there is no genuine dilemma here: The judge’s public duty comes first. As it is with friendship, so it would seem to be with the judge’s other personal duties—including mercy.
I have been taking it as given that a judge’s only duty to act mercifully is the duty we all have to act mercifully. In other words, I have been taking it as given that a judge does not have any additional, role-based duty to act mercifully. Maybe that is wrong, though. Suppose a judge defends her merciful sentencing like this:
It is true that everyone has a duty to act mercifully and that I am no different in this respect. It is true, too, that I ought not to let my personal duty in mercy interfere with my duties as a judge. But it is wrong to set mercy in opposition to the official role of a judge and to imply in this way that by passing a merciful sentence I am failing to act as a judge qua judge ought to act. For it is not only individuals who should act mercifully. The state should, too. The state should temper justice with mercy—and as an agent of the state, so should I. Therefore, by showing mercy, I am not putting a personal duty ahead of a professional one; I am recognizing that justice does not define the whole of my duty as a judge.
The idea is that the state should sometimes act mercifully, and a judge should act to advance or fulfill this aim or duty by sometimes acting mercifully herself. In that case, justice would not exhaust a judge’s official duties, nor could mercy properly be confined to a judge’s personal life. A judge qua judge would properly be guided by considerations of both justice and mercy.
One objection might be that, contrary to what the judge has argued, in fact the state should always act justly, never mercifully. Some philosophers do say this, but it seems doubtful to me. It would mean that the state had a duty to punish (that is, to serve justice by giving wrongdoers the hard treatment they deserve), whereas what the state seems to have is a right to punish, which it can waive if it wishes. Moreover, there seem to be strong reasons for the state to waive this right and to act mercifully. Those reasons are similar to the reasons we as individuals have for acting mercifully: to reduce human suffering, to show an awareness of human frailty and weakness, and to show that we are strong enough not to need to react to every wrong.16 This is not to say the state should act mercifully in all cases. That would lead to too much injustice, for one thing. Instead the state should strike a balance between justice and mercy.
These remarks are sketchy, and the relationship between justice and the state is a core concern of political philosophy. So in what follows I shall assume that the state should sometimes act mercifully and consider what other objections to the judge’s argument might be made.
A second objection is that, even assuming the state should sometimes act mercifully, it is not the case that judges should sometimes act mercifully on behalf of the state. For this objection to be sound, there must be some other official who should act mercifully on behalf of the state instead. I think there is, and it is the head of state. I say that for four reasons.
First, even when the state shows mercy to a criminal, it should make known that punishing him would have been just. Otherwise, people might think the state treated the criminal justly, rather than unjustly but mercifully. The state would be seen as condoning the criminal’s behaviour instead of condemning and forgiving it. These are very different messages to send. Moreover, by making the just punishment known, the state indicates what treatment people can expect from it in similar cases in the future, when mercy may not be forthcoming.
There are different ways the state can communicate what justice requires even when it ultimately shows mercy, but some ways are more effective. A judge might decide that a just punishment for a thief is 2 years imprisonment, decide that justice should yield to mercy, and order the thief to be released. We can find out that the just punishment was 2 years—but only by reading the judge’s reasons. Instead, the judge could sentence the thief to 2 years, and the head of state could pardon him. Here we have two legal orders made by two legal institutions. Dividing responsibilities in this way allows the state to more clearly communicate the different demands made by justice and mercy.
Second, giving the job of deciding what is just to one institution and the job of deciding between justice and mercy to another institution has symbolic value. It shows clearly that there is a tradeoff between justice and mercy. In the time between the judge sentencing the criminal and the head of state pardoning him, the criminal’s fate is clearly in the hands of the state; he is at the state’s mercy. We do not want the state to revel in its power over us. But we do want it made clear that mercy cannot be expected. Justice should be reliable; mercy should be exceptional. When different officials perform justice and mercy at different times using different procedures, it draws attention to the difference in what the state is doing.
Third, an institution needs different legal powers when its job is to decide what is just than when its job is to decide between justice and mercy. Judges may have discretion to decide on a just punishment, but their discretion is usually limited. In Canada, for example, a judge must sentence a person convicted of second-degree murder to between 10 and 25 years in prison. The judge can show mercy (by imposing a sentence of 10 years), but only up to a point. By contrast, the Governor General, who represents Canada’s head of state, can grant even a convicted murderer a full pardon. In common law jurisdictions generally, sentencing judges are able to show mercy to a lesser degree than heads of state. On the other hand, heads of state are unable to increase the punishment a judge has ordered, even when it would be just to do so. Both facts suggest that sentencing judges have the better tools with which to do justice and heads of state the powers more suited to showing mercy.
There is a fourth and more subtle reason that heads of state are better suited than judges to decide when to give priority to mercy. One way the state can act mercifully is by not insisting that a criminal serve his just punishment. But the state can act mercifully in other ways, too. To give just two examples:
(i) Some former students are burdened by large student loans. It will take them decades to pay these loans back, if not longer, and their lives are harder as a result. By “forgiving” these debts, the state would show mercy.
(ii) During the recent housing crisis, many people ended up with mortgages they could not hope to pay back. When states were not legally obligated to forgive them, but did so anyway, they showed mercy.17
In the same way that an individual has a duty to act mercifully sometimes but not always, the state should act mercifully sometimes but not always. The state should show mercy often enough, but not too often. So, there are two points: The state has opportunities to show mercy outside the criminal justice context, and the state should not act on all of its opportunities to show mercy.
As a result, a state needs to somehow coordinate or ration its merciful activity. The state’s merciful activity in one context must be set against its merciful activity in other contexts to ensure that, overall, it acts mercifully often enough but not too often. Thus, the more often the state acts mercifully in other contexts, the less often it needs to act mercifully in the criminal justice context and vice versa. This may sound too calculating, but the idea is a familiar one: Just as an individual should decide whether she is acting mercifully enough by considering all aspects of her life, the state should decide whether it is acting mercifully often enough by considering all aspects of its activity.
The most obvious way for the state to coordinate its merciful activity is by centralizing much of its merciful activity in one institution. That institution will not be the judiciary. Judges do not have opportunities to show mercy as officials except in the courtroom. They are deliberately kept separate from the rest of government. The executive, by contrast, has many opportunities to show mercy beyond the criminal justice context. Indeed, in many states, the acts in (i) and (ii) would be overseen by the executive.
I have given several reasons judges should leave mercy to the executive. I should add that this is often exactly what judges do. Consider the Canadian case of R v. Latimer.18 Latimer had killed his severely disabled daughter for (he said) compassionate reasons. After his conviction for second-degree murder, the trial judge imposed a sentence of 10 years, which was the minimum sentence the law allowed. The trial judge also recommended that the executive consider showing Latimer mercy. The Supreme Court, which upheld the sentence, said:
[T]he prerogative [of mercy] is a matter for the executive, not the courts. The executive will undoubtedly, if it chooses to consider the matter, examine all of the underlying circumstances surrounding the tragedy of Tracy Latimer that took place … some seven years ago. Since that time Mr. Latimer has undergone two trials and two appeals to the Court of Appeal for Saskatchewan and this Court, with attendant publicity and consequential agony for him and his family.19
The Supreme Court is clearly of the opinion that Latimer’s suffering and the suffering of his family are relevant to whether he should receive mercy. Equally clearly, the Court thinks these are factors for the executive to consider, because it is the executive that ought to decide whether to show Latimer mercy.20
In the last section I said that a judge should not use her public office to fulfill her personal duty in mercy. One of my arguments was that she had an alternative way to fulfill her personal duty. My argument for why judges should not seek to show mercy on behalf of the state has a parallel structure. Judges should not try to show mercy on behalf of the state because there is an alternate and better way for the state to show mercy, namely, through the executive. In short, judges should not act mercifully as sentencers because mercy should be given over to other spheres of their lives and to other state officials.
Suppose that the executive announces a policy that says, in effect—no pardons, no exceptions. The executive keeps all of its powers and capabilities, but it refuses to issue any more pardons. Judges know that if they do not show mercy to criminal wrongdoers, no one will. Now, we would need to know more to be sure (including why the executive refused to issue pardons), but it seems that the argument I outlined for why judges should stick to doing justice no longer holds true. The executive should show mercy on behalf of the state, and if it does, judges should not. But if the executive is unwilling to do its job, then it seems to be better that judges do it for them than that no one does.
In the scenario in the last paragraph the executive says it will never show mercy. That is unrealistic.21 But is easy to imagine that the executive will show mercy in the criminal justice context less often than it should. Alex Tuckness and John M. Parrish have described a “decline of mercy in public life,” including a decline in the number of executive pardons in the United States. They write:
Perhaps the most stunning contemporary use of the power to pardon and commute sentences was the decision of Governor Ryan of Illinois to pardon or commute the sentences of all 171 inmates on the state’s death row shortly before he left office. To put this into perspective, that one act was more than four times the number of pardons and commutations in death penalty cases by all American governors combined over the previous twenty years.22
There are no doubt many causes of the decline in the number of pardons, and it may be that this decline is to be welcomed. That depends on how often the state should be merciful. But if the executive is not acting mercifully as often as it should, as Tuckness and Parrish seem to believe, perhaps judges should ensure the state continues to show mercy to criminal wrongdoers.
Even if judges should let the executive show mercy on the state’s behalf, this does not mean that judges have no role relevant to mercy. They do—the role of reviewer. At this point we move from the domain of the criminal law to that of public law. What follows is a brief sketch of the relevant basics of public law.
The executive is given legal authority to make certain decisions, but this authority comes with conditions attached. A decision that does not meet these conditions is “invalid,” i.e., without legal effect. Judges have the responsibility to determine the validity of the executive’s decision, an activity that is known (in many but not all common law jurisdictions) as “judicial review.” The “grounds of review” are the conditions a decision must meet to be valid.
With respect to executive decisions to grant or refuse pardons, judges traditionally exercised a very light touch review. This is still the case in the United States. Elsewhere in the common law world, though, things are changing. Judges in Canada,23 England and Wales,24 India,25 Northern Ireland,26 Singapore,27 and other jurisdictions are now willing to review pardoning decisions on a range of grounds. It would take more space than I have to discuss the different approaches these jurisdictions have taken. So I shall focus on the ground of review that raises the most interesting philosophical issues, namely, consistency.
Suppose there are two convicted thieves, Thief 1 and Thief 2, their cases alike in all relevant respects. Each has been sentenced to a just punishment of 5 years in prison. The head of state shows mercy to Thief 1 and grants him a full pardon but refuses Thief 2’s request for a pardon. Thief 2 asks a judge to review the head of state’s decision. His argument is simple. There was no more reason for the head of state to show mercy to the other thief than to him. The head of state has acted inconsistently and hence arbitrarily. And this, the thief says, she should not be permitted to do.
The thief’s argument raises an ancient problem regarding mercy. The thief’s objection is to inconsistent legal mercy, but St. Anselm had much the same concern with seemingly inconsistent divine mercy. Anselm wrote:
But if it can be comprehended in any way why you can will to save the wicked, yet by no consideration can we comprehend why, of those who are alike wicked, you save some rather than others, through supreme goodness; and why thou dost condemn the latter, rather than the former, through supreme justice.28
How, Anselm asked, could it be right to show mercy to some of those who deserve punishment, but not to others who are equally deserving?
Perhaps an answer can be found with respect to divine mercy, as Anselm thought. In the legal context, however, courts are often hostile to inconsistent acts of mercy by heads of state. In India, for example, the Supreme Court has long held that pardoning decisions are reviewable for arbitrariness.29 In Northern Ireland, the Court of Appeal has been willing to conduct detailed comparisons of pardoning decisions to see if they have been made in an “unequal” or “inconsistent” way.30 Courts in England have been willing to invalidate pardoning decisions if they are not based on “relevant” factors or are based on “irrelevant” factors.31
The question for us is, should courts review pardoning decisions for inconsistency? The answer will be “yes” only if it is objectionable to make pardoning decisions inconsistently. Intuitively it seems that it must be. Yet it turns out to be hard to say what exactly the objection is.
Let us start with justice. Some philosophers, including Joel Feinberg, say that there are two kinds of justice.32 Non-comparative justice requires that people be treated as they deserve. Comparative justice, by contrast, requires that people similarly situated be treated similarly. Punishing Thief 2 is comparatively unjust, given that Thief 1, similarly situated, was let free. That makes the head of state’s decision to refuse a pardon to Thief 2 objectionable, in a sense. But we should not condemn her decision yet. Letting Thief 2 go free might be comparatively just, but it would be non-comparatively unjust (because he deserves a punishment of 5 years in prison). There is a conflict between non-comparative and comparative justice. There is no dilemma here, though, because as Feinberg says, “injustice by noncomparative standards tends to be a much more serious thing than comparative injustice.”33 When they conflict, non-comparative justice prevails over comparative justice. So, even assuming there are two kinds of justice, justice does not favor pardoning Thief 2 in an overall sense.34
Another possible objection is based on rationality. Rationality is often said to require like cases to be treated alike. Were this indeed what rationality requires, the head of state would have acted irrationally by showing mercy to Thief 1 but not to Thief 2. This is not what rationality requires, however. Suppose there is a bowl of red and green jellybeans, and you have a reason to take any two. You take a red one, then you take a green one. You have acted “inconsistently,” but not irrationally. What rationality requires is for you to respond appropriately to your reasons. When a choice is undetermined by your reasons, as the choice of jellybeans is, you act rationally whichever choice you make. The state has a reason to sometimes show mercy. But it has no greater reason to show mercy to one wrongdoer than to another. As long as the state shows mercy to enough people, it responds appropriately to the reason mercy provides it with. So, the state can show mercy inconsistently without acting irrationally.35
We might instead object to the head of state’s inconsistency on practical grounds. If the head of state makes consistent decisions, then her decisions will be easier to predict. In general, predictability in legal decision-making is valuable, because it allows citizens to plan their lives better. However, as Andrei Marmor points out, this general point does not apply with equal force in all contexts.36 It is important that tax decisions be made consistently because people should be provided with a stable basis on which to plan their economic lives. But the value of predictability is less obvious with respect to criminal punishments and less obvious still when it comes to mercy. What valuable plans are likely to be disrupted or inhibited by unpredictable pardoning decisions?
There is a second and stronger practical objection to inconsistent pardoning decisions. Suppose you, as a private individual, only show mercy to people who pay you to do so. There is something wrong here, but the fault lies not with what you have done but with you, and specifically your character. To act mercifully is one thing; to be a merciful person is another. To be a merciful person means being disposed to act mercifully out of love or compassion. By only showing mercy for money, you show yourself not to be a merciful person, and that is grounds for criticizing you. Just as we care about a private individual’s motives for acting mercifully, we care about the state’s motives for acting mercifully. We want the state to not just act mercifully, but to be merciful. That is to say, we want the state to act mercifully out of compassion or sympathy for us, its citizens. If the head of state’s pardoning decisions are the result of improper motives like nepotism, racism, or corruption, instead of compassion, the head of state and the state she represents are open to criticism. If pardoning decisions are made inconsistently, even taking into account the features of cases that properly evoke compassion, then people may tend to believe that the inconsistency is the result of improper motives. If it is valuable to avoid such suspicions, then it may be valuable to ensure that pardoning decisions are made consistently.
This seems to me to be a sound, but limited, objection to inconsistent pardoning decisions. It is limited because, while a requirement of consistency is one way to avoid suspicions of impropriety, there might be other, even better ways. For example, the process of making pardoning decisions could be made highly public. Or, heads of state could be required to prepare guidelines setting out the kinds of factors they will consider. Or, judges could simply ask whether pardoning decisions have been made for improper purposes—as judges in some jurisdictions in fact do.37 It would be worthwhile to compare these alternatives in detail. For now, the somewhat surprising result is that there seems to be nothing wrong with arbitrary pardoning decisions in themselves. If they are objectionable, it is because they are a symptom of other problems.
Discussions of judges and mercy have tended to focus narrowly on sentencing. I have discussed sentencing, but I have also widened the focus to include judges’ personal lives and the opportunities they present for showing mercy; the institutional context within which judges operate; and judges’ power to review the merciful decisions of the executive. My major claims are, first, judges should not fulfill their personal duties to act mercifully by imposing merciful sentences. Second, judges should not try to act mercifully on behalf of the state. They should let the executive do that job. But a possible exception is if the executive is unwilling to do that job. Finally, there is reason to think that judges are entitled to review the executive’s pardoning decisions for consistency as a means of avoiding the appearance of improperly motivated pardons.
1 See, e.g., Alex Tuckness and John M. Parrish, The Decline of Mercy in Public Life (Cambridge: Cambridge University Press 2014) c 9 (arguing that mercy is a family resemblance concept); K. Bell, ‘Mercy and criminal justice’ (unpublished PhD manuscript) (arguing that there are actually two concepts of mercy).
2 I am using ‘harm’ in a broad sense to mean hard or undesirable treatment, such that even deserved punishments count as harms.
3 See N. Markosian, ‘Two Puzzles about Mercy’ (2013) 63 Philosophical Quarterly 269 for an excellent discussion of the different notions of permissibility that have featured in debates about mercy.
4 A. Smart, ‘Mercy’ (1968) 43 Philosophy 345, 350. See also J. Murphy, ‘Mercy and legal justice’, in J. Murphy and J. Hampton, Forgiveness and Mercy (Cambridge: Cambridge University Press 1990); A Duff, ‘Mercy’ in J. Deigh and D. Dolinko (eds.), Oxford Handbook of the Philosophy of Criminal Law (OUP 2011) 480.
5 Others have argued for this view, e.g., A. Brien, ‘Mercy, utilitarianism and retributivism’ (1995) 24 Philosophia 493, 498–99.
6 As far as I know, this argument was first made in G. Rainbolt, ‘Mercy: an independent, imperfect virtue’ (1990) 27 American Philosophical Quarterly 169, 171.
7 G. Rainbolt, ‘Mercy: in defense of caprice’ (1997) 31 Nous 226, 231–36. Others who have claimed that mercy is an imperfect duty include S. Garvey, ‘Is it wrong to commute Death Row? Retribution, atonement, and mercy’ (2004) 82 North Carolina Law Review 1319, 1331–1334; S. Kershnar, ‘Mercy, retributivism, and harsh punishment’ (2000) 14 International Journal of Applied Philosophy 209.
8 That mercy is an act of grace is a common theme in literature. It is often expressed in the law, too. See, e.g., United States v Wilson, 2 U.S. (7 Pet.) 150 (1883) at 160; Reckley v Minister of Public Safety and Immigration (No 2) [1996] AC 527 at 540.
9 In some states, including the United States, the head of state has the legal power to issue pardons, and exercises real discretion as to the exercise of that power. In other states, including the United Kingdom, the head of state has the legal power to issue pardons, but de facto and indeed constitutional control over the exercise of that power lies with the government. To avoid awkward constructions, I shall write as if the decision lies in law and in fact with the head of state.
10 J. Murphy, ‘Mercy and legal justice’ in J. Murphy and J. Hampton, Forgiveness and Mercy (Cambridge: Cambridge University Press 1990) 175.
11 I consider whether showing mercy is part of a judge’s responsibilities as a judge in the next section.
12 A. Duff, ‘Mercy’ in J. Deigh and D. Dolinko (eds.), Oxford Handbook of the Philosophy of Criminal Law (Oxford: Oxford University Press 2011) 480.
13 Ibid. 479.
14 S. Hale, ‘Against supererogation’ (1991) 28 American Philosophical Quarterly 273, 276 (‘when the demands of a perfect duty conflict with the demands of an imperfect duty, it is morally forbidden to neglect the demands of the perfect duty; whereas, the demands of the imperfect duty are not neglected in a morally permissible manner by acting on the perfect duty rather than on the imperfect duty’). For a similar ‘principle of conflict resolution’ with respect to reasons, see J. Raz, Practical Reason and Norms (Princeton, NJ: Princeton University Press 1990) 188.
15 A point made by J. Raz, Practical Reason and Norms (Princeton University Press 1990) 188.
16 G. Rainbolt, ‘Mercy, justice and the death penalty’ (unpublished manuscript).
17 This example is from Alan Tuckness and John M. Parrish, The Decline of Mercy in Public Life (Cambridge: Cambridge University Press 2014) 3.
18 R v. Latimer [2001] 1 SCR 3.
19 Ibid. [89].
20 Ultimately Latimer did not apply for mercy.
21 Apparently only one regime has officially abolished pardons, and that was for a short time, during the French Revolution of 1789. See K. Moore, Pardons: Justice, Mercy, and the Public Interest (New York: Oxford University Press 1989) 24–25.
22 Alan Tuckness and John M. Parrish, The Decline of Mercy in Public Life (Cambridge: Cambridge University Press 2014) 2. The authors add that, by contrast, in many European countries pardons and amnesties remain common.
23 Thatcher v Attorney General of Canada [1997] 1 FC 289 (reviewable for procedural and substantive unfairness).
24 McGeough v Secretary of State for Northern Ireland [2012] NICA 28 (inconsistency).
25 Maru Ram v Union of India [1980] INSC 213 (arbitrariness, bias); Epuru Sudhakar v Government of Andhra Pradesh [2006] INSC 638 (relevancy, failure to exercise discretion).
26 R v Secretary of State for the Home Department, ex p Bentley [1994] QB 349 (error of law); R (Page) v Secretary of State for Justice [2007] EWHC 2026 (Admin) (relevancy).
27 Yong Vui Kong v Attorney General [2011] SGCA 9 (procedural impropriety).
28 St. Anselm, Proslogium XI (translated by SN Deane, Chicago: Open Court 1903).
29 Maru Ram v Union of India [1980] INSC 213 at [62].
30 McGeough v Secretary of State for Northern Ireland [2012] NICA 28.
31 See, e.g., R (Page) v Secretary of State for Justice [2007] EWHC 2026 (Admin).
32 J. Feinberg, ‘Noncomparative justice’ (1974) 83 Philosophical Review 297.
33 Ibid. 317.
34 Another response would be that this kind of example shows that the distinction between comparative and noncomparative justice is flawed, and that Thief 2 has not been treated unjustly in any meaningful sense: see P. Montague, ‘Comparative and noncomparative justice’ (1980) 30 Philosophical Quarterly 131, 133.
35 For a related argument, see J. Tasioulas, ‘Mercy’ (2003) 103 Proceedings of the Aristotelian Society 101, 129–130.
36 A Marmor, ‘Should like cases be treated alike?’ (2005) 11 Legal Theory 27, 33–34.
37 See, e.g., Epuru Sudhakar v Government of Andhra Pradesh [2006] INSC 638 (invalidating a pardon given to a person for being ‘a good Congress worker’, at [55]).
Bell, K. ‘Mercy and criminal justice’, unpublished PhD manuscript.
Brien, A. (1995) ‘Mercy, utilitarianism and retributivism’. 24 Philosophia, 493: 498–99.
Duff, A. (2011) ‘Mercy’, in J. Deigh and D. Dolinko (eds.) Oxford Handbook of the Philosophy of Criminal Law, Oxford: Oxford University Press, p. 480.
Epuru Sudhakar v Government of Andhra Pradesh [2006] INSC 638.
Feinberg, J. (1974) ‘Noncomparative justice’, Philosophical Review, 83: 297.
Garvey, S. (2004) ‘Is it wrong to commute death row? Retribution, atonement, and mercy’, North Carolina Law Review, 82: 1319, 1331–1334.
Hale, S. (1991) ‘Against supererogation’, 28 American Philosophical Quarterly, 28: 273, 276.
Kershnar, S. (2000) ‘Mercy, retributivism, and harsh punishment’, International Journal of Applied Philosophy, 14: 209.
Markosian, N. (2013) ‘Two puzzles about mercy’, Philosophical Quarterly, 63: 269.
Marmor, A. (2005) ‘Should like cases be treated alike?’, Legal Theory,11: 27, 33–34.
Maru Ram v Union of India [1980] INSC 213.
McGeough v Secretary of State for Northern Ireland [2012] NICA 28.
Moore, K. (1989). Pardons: Justice, Mercy, and the Public Interest, Oxford: Oxford University Press.
Montague, P. (1980) ‘Comparative and noncomparative justice’, Philosophical Quarterly, 30: 131, 133.
Murphy, J. (2011) ‘Mercy and legal justice’ in J. Murphy and J. Hampton Forgiveness and Mercy, New York, NY: Cambridge University Press.
Rainbolt, G. (1990) ‘Mercy: an independent, imperfect virtue’, American Philosophical Quarterly, 27: 169, 171.
Rainbolt, G. (1997) ‘Mercy: in defense of caprice’, Nous, 31: 226, 231–236.
Rainbolt, G. ‘Mercy, justice and the death penalty’, unpublished manuscript.
Raz, J. (1990) Practical Reason and Norms: Princeton, NJ: Princeton University Press, p. 188.
R v Latimer [2001] 1 SCR 3.
R (Page) v Secretary of State for Justice [2007] EWHC 2026 (Admin).
R v Secretary of State for the Home Department, ex p Bentley [1994] QB 349.
Reckley v Minister of Public Safety and Immigration (No 2) [1996] AC 527 at 540.
Smart, A. (1968) ‘Mercy’ Philosophy, 43: 345, 350.
Tasioulas, J. (2003) ‘Mercy’, Proceedings of the Aristotelian Society, 103: 101, 129–130.
Thatcher v Attorney General of Canada [1997] 1 FC 289.
Tuckness, A and Parrish, J. M. (2014) The Decline of Mercy in Public Life, Cambridge: Cambridge University Press, c 9.
United States v Wilson, 2 U.S. (7 Pet.) 150 (1883) at 160.
Yong Vui Kong v Attorney General [2011] SGCA 9.
St. Anselm, Proslogium XI (translated by SN Deane, Chicago: Open Court 1903).