I begin with an observation: many people think we need the institution of punishment because of some failing of humans to be good. That is, if humans were perfectly good—if they never behaved badly—we could dispose of punishment. I think this belief is wrong. In what follows, I argue that a need for punishment flows instead from our intrinsic human goodness.
My argument has several parts. I begin by using a thought experiment to consider whether a society without punishment is even possible, and, as a strictly logical matter, I concede it is indeed possible. I then consider what kinds of agents would populate such a society. Such agents would quell—or not have—narrow self-interest and, in the philosophical literature, are often called angels. But if that is the nature of angels, then, I argue, angels lack something of human goodness. I realize that this is an odd claim: what could be wrong with doing the right thing without ever needing to be threatened with punishment? To answer that question, I introduce two, conflicting virtues which I call ‘love’ and ‘decency’.As I use these terms, they evoke their colloquial meanings, but they’re not quite the same—so I take some time below to explain exactly what I mean. In the next part of the argument, I suggest that human goodness requires the exercise of both love and decency—and I suggest the two virtues frequently conflict. I propose two ways of avoiding such conflicts, one of which is the institution of punishment. Punishment, I argue, allows us to make compatible two fundamentally conflicting virtues—two virtues that, I think, bound and define our moral, human lives.
So, what kind of society could exist without recourse to punishment? Let me begin by considering how we might eliminate legal punishments (as opposed to, say, social or familial ones). Consider a calm and well-off island community with three policemen. In the community’s first year of existence, the policemen patrol, but there are no complaints or arrests. A new community budget is approved, and the community cuts the police budget. There are now two policemen. In the following year, they again cut the police budget. This time, there is only one, part-time officer. And, the next year, they cut the police budget entirely.
What does this lack of law-enforcement entail? Note that it does not entail a lack of law—merely a lack of enforcement. Law could still be necessary to solve coordination problems: which side of the road to drive on, etc. The lack of enforcement also doesn’t entail that the community members never have legal disputes. Community members could still disagree about what the law is, and they could still take matters to court. Courts would act as arbitrators not backed up by threats—much as certain religious courts often act today. Drawing these implications makes such a non-coercive system seem tenable:certainly at least logically tenable—and perhaps, for a specific kind of agent (which, again, philosophers might call angels) also practically so.
Excising non-legal punishment is much the same. Rules can still exist: school rules, rules in sports,company conduct codes, explicit rules within families, and social rules of etiquette or morality. All that is required for this to work is that most agents, most of the time, follow the rules—and that none of the few breakages are particularly damaging to the system. Note that community members could still point out to a rule breaker when a break has occurred, so long as no actual negative consequences follow. No social ostracism (its own kind of punishment), no shaming, no removal of jobs or status, no lowered grades, and no extra chores. Nevertheless, if you saw me do something wrong, you could still tell me you thought it wrong. And, if I were the kind of agent who populates this society, I’d probably listen, and correct my behaviour. So: a lack of punishment (either legal or nonlegal punishment) does not necessarily entail a lawless society, or a society without morality, or a society without social protocols. So long as its inhabitants were—let’s say—angelic, it could run quite smoothly. Why, then, do I think this society is so bad?
To explain myself, I want to set all the preceding work aside and consider something which may, for a few sentences, seem irrelevant. Any liberty is fundamentally binary. My liberty to do X at T (what I shall call (X,T)) is a binary: I am either free to do it or I am not. A post-facto sanction operates by limiting the chains of acts which I am free to do. For example, if I do X at time T, then I might be imprisoned and unfree to do Y at time U. So, it is not possible to engage in actions (X,T) and (Y,U). Without the sanction, I would have been free to perform the act-chain (X,T), (Y,U). That is, punishments don’t make me less free to do the single, punished act but rather limit my set of possible act-chains.
In normal society, punishments operate on our choices by closing act-chains. These act-chains would remain open in the non-punishing societies: the community members would simply choose not to engage in certain act-chains. Each ‘angel’ would understand the needs of the community and put those needs above her own, and above the interests of those close to her, despite never being forced to.
All of us subordinate our interests and the interests of those close to us in our day-to-day lives. An example: I don’t assault my friend’s competitor in a running race despite knowing my friend’s life would be better if her competitor were slowed so she could win. It might feel wrong to abstain from tripping the other runner—say, if winning means a great deal to my friend (maybe the prize is money to be used for her sister’s cancer treatment). But, I might rationalize: if I assault the runner I would be imprisoned and unable to help in other ways, say, by sharing my wages.
The best option for my friend might be for me to engage in act-chain (assault, today) and (share wages, in future). But, the law stipulates that I can only do the first act or the second, not both. I am unable to perform—coerced out of performing—the act pair.
In the Society of Angels, there is no such coercion. Instead, it would be up to me whether to engage in the assault, the wage-giving, or both. Each angel would have to subordinate her friends’ interests to the community’s—a choice which a coercive legal system typically allows us to avoid.
That choice illustrates a broad category of moral conflict—conflicts between what I call ‘love’ and what I call ‘decency’. By ‘decency’, I mean that regard which we give (or ought to give) not just to those close to us or to those we like, but to all people (or all sentient beings, or all beings, or all potential beings) as well as to those we dislike. Decency is typified by principles such as protecting the commons, giving serious weight to the happiness of those one has never met, or being fair to someone even if one justifiably despises that person. In the general case, decency is a calm virtue: it is complimented by cool rationality, and it is—usually—a virtue that does not inculcate obsession. Our angels, of course, are decent.
By ‘love’, I mean something often at odds with this decency. Love is the kind of regard one gives to another because one thinks of that other as special—as more important than others. Parents love their children, of course, and we often love our friends. In the example above, with the running, it is love that tempts me to trip the other runner, and it is decency that convinces me to refrain. Love is not, at least not usually, a rational virtue. It is often unjustifiable, it often compels us to an act where reason would advise we do something else, and it often—unlike decency—inculcates obsession. Our angels lack this kind of love. Love, in a human society with no punishment, would compel us to put the interests of our friends over that of society. The angels are not so moved—and so we see they lack this quality.
I occasionally explain this love-decency distinction to people, and a few times I’ve got responses arguing that love—as I define it—isn’t a virtue at all. That, insofar as love and duty conflict, well, so much the worse for love. In defense of unconstrained love, I don’t have much to say. But I will make one small observation.
People crave love—or, at least, the ones I know do. They do not just crave good treatment, or material security, or good health—they crave personal love. And, the kind of love offered universally—that I call decency—does not fulfill that desire. We crave—at least most of us—to be loved specially. I take that as a simple empirical fact about humans, and that special love is what constitutes our human friendships and familial relations.
I know that point is little more than my observation. I suppose I take the force of love, or the idea that love is good (that friendships are valuable even if the friends are to everyone else indecent), as axiomatic. So too do I take as axiomatic the idea that decency is good. There is something about human life that one misses if one misses either of these qualities.
Note also that love and decency operate on us in similar ways: they both can oblige us to take particular actions. Decency obliges me not to pollute a nearby river; love obliges me to make enough money (however that money is made) to pay for my parents’ healthcare. I think it is central to a moral, human life to be compelled by both these kinds of reasons. And—as I have said now several times—they can conflict: it is good to want to put your friend over your society—that is what makes one a good friend; but it is also good to have real care for all those around one, however distant in space or time, and to want to do what is right for them too. Such conflicts, unavoidable as they sometimes are, tend to wrench at agents in ways other moral conflicts (at least in my experience) have not wrenched. In essence, one chooses between being a good person and being a good friend, and that is a rather miserable choice. As I see it, there are two ways of avoiding such conflicts.
The first is that one should avoid foisting such conflicts on one’s friends. For example, I might have a secret that—were I to tell you—you would feel obliged to reveal. Suppose you and I are friends, and Maddie is my enemy. I’ve stolen Maddie’s dress and plan to shred it tomorrow. If I—in confidence—tell you this, you might feel obliged to break my confidence to stop me from enacting my plan (say, if Maddie couldn’t afford a replacement). In that kind of situation, I might have a love-based obligation not to tell you my secret (along, of course, with my decency-based obligation not to steal and destroy things to begin with). That love-based duty to you flows from my obligation to you, as my friend, not to put you in positions where being a good friend to me means sacrificing some of your decency. Being a good friend should align with being a good person—and it is a friend’s role to try to make that, usually, the case.
This solution isn’t perfect. It is sometimes not up to one whether or not a friend is put in such a position. And sometimes friends fail in their obligations of love just as most people sometimes fail in their obligations of decency. One friend’s failure to avoid causing another friend such conflict doesn’t remove the second friend’s obligations of love for the first. Mistakes should not, in general, undo friendships—even when those mistakes are ones precisely in the realm of that friendship.
The other way of avoiding such conflict is, as I have discussed throughout, the institution of punishment. By limiting the set of possible act-chains, the institution of punishment ensures that the incentives to exercise the love virtue and the decency virtue are aligned. It allows us, mostly and most of the time, to both be loving and be decent.
So, where are we? I began by arguing that societies can, if made up of a certain kind of agent, dispose of punishments. But those agents lack a kind of desperate, flowing love that I think constitutes part of what it is to be human. Without it, we might be more decent—but we would miss something valuable about what it is to live as a human: in accord with human goodness or human nature. I then discussed conflicts between love and decency—and how some of these painful conflicts can be avoided. First, we can avoid causing them in others; and second, we can set up societal incentive structures such that they don’t arise quite as often.
Neither solution is perfect: friends fail and allow their friends to face miserable moral dilemmas. And not all conflicts between love and decency can be detangled by appropriate threats of punishment. But it is, I think, crucial that we try to allow people to be more than loveless angels, and part of that can be done with punishment. Punishment—far from being inhuman, or from reflecting some human flaw—allows us to exercise two most human of virtues.
The Society of Angels thought experiment appears in various places. Notable examples can be found in Joseph Raz’s Practical Reason and Norms, H. L. A. Hart’s “Theory and Definition in Jurisprudence”, John Gardner’s Law’s Aim in Law’s Empire, and John Finnis’s Natural Law and Natural Rights (though Finnis calls it a ‘society of saints’ rather than of angels).
Matthew Kramer’s idea of binary liberties is set out in several places. A good example is his “Freedoms Do Not Exist by Degrees”.
My ideas on love flow in part from W. H. Auden’s poem September 1, 1939 (his repudiation not withstanding). My thoughts on decency are mostly taken from my personal life. But see also Auden’s Law Like Love—on both love and decency and on their fundamentally similar structures.
For the idea, threaded loosely throughout, that we should live in accord with our essences, see Korsgaard’s The Constitution of Agency at 129-150.