1.1. Shared Structural Features: Significant Interactions, Dispositions, and Minimal Autonomy

Perhaps the most obvious relation of friendship is that of interaction: friends must interact with each other and that interaction must be directed at the people involved. We can’t be friends with the dead, except in a Pickwickian sense, because we can’t interact with the dead. And the DMV clerk I see every seven years for two minutes isn’t my friend, even if we exchange pleasantries each time. For the kindness displayed isn’t directed at the other due to who the other is; we are each trying to make a tedious exchange less miserable.

This relation is also a structural feature of enemyship. There are figures in history or in the news whose character or interests we condemn. And if we were positioned to do so, we would foil their plans. We might call such people ‘enemies in principle,’ but such people aren’t our enemies because we can’t interact with them, either directly or indirectly. One can despise Pol Pot, but it is too late to aspire to be his enemy. Similarly, the driver of the farm combine who is preventing me from making it to the most important meeting of my life is not my enemy, even though he is preventing me from fulfilling my goal, for his actions aren’t directed at me. He would have driven criminally slowly even if I hadn’t been there.

The nature and extent of that interaction will vary depending on whether the relationship is one of enemyship or friendship, and it will depend on what kind of friend or enemy relationship obtains. But in order for you to be either my friend or foe, you must be able to intentionally affect my well-being, either directly or indirectly. The notion of interaction at play here has to be expansive and include both direct and indirect action. For often our enemies act through intermediaries to foil our interests. Consider the junior high bully who barely even acknowledged you but made sure you were excluded from the best parties. Her minions knew you had been designated a social pariah and knew not to include you, even if some had once been your friend. Or think of the mob boss who has his underling kill his rival. This subordinate need not have any beef with the rival boss and might just be doing his own boss’s bidding. The enemyship here is between the rival bosses, not the subordinate, even though it is the subordinate who acts directly against the rival.

In fact, the interaction between enemies can be wholly indirect. Consider the early medieval queen regents and sisters-in-law, Brunhilda (c. 543-613) and Fredegund (529-597) of Austrasia. These two never met and never communicated. But they attempted (sometimes successfully) to have each other and each other’s family members killed, even arranging the assassinations of their respective husbands, Sigibert and Chilperic. This sequence of aggression and retaliation was not just geopolitical maneuvering: Brunhilda believed that Fredegund, a royal servant and mistress to Chilperic, had arranged for the death of Galswintha, Brunhilda’s sister, so that Fredegund could assume her place at Chilperic’s side. It was Galswintha’s death that precipitated the war between the two houses, a war that led to the sequential assassinations of these queens’ respective husbands (see Puhak Reference Puhak2022). Though these two were clearly enemies, their interactions were conducted wholly indirectly. In contrast, while a person can outsource some of the work of friendships to others, she can’t outsource all of that work. Say roommates, Beth and Susan, once very close, move across the country from each other. Beth never has time to speak with Susan, what with her important job and new family, but she sends Susan lovely cards and gifts. If Susan learns that every card and gift she received was the work of Beth’s administrative assistant, she would be right to suspect that the two were no longer really friends. To be friends, Beth has to do some of the work of her relationship herself. The structural requirement of suitable interactions also holds for enemyship; this is one of the grains of truth in the opposition view. But there is already an important difference: those interactions can be wholly indirect in enemyship, though not in friendship.

The second structural relation we also find in enemyship is the disposition to behave in a certain way. A person wants her friends’ interests, especially those interests around which the friendship is built, to be satisfied. And for the most part, she supports and helps her friends in the satisfaction of those interests. Where a friend tries to further the other’s interests, an enemy must be motivationally disposed to thwart at least one of the other’s interests. Someone might not succeed at thwarting her enemy’s interests; she might not even have the opportunity to do so; but if she does have the opportunity, all things being equal, she will act against the interests of her enemy.Footnote ⁴

Finally, both parties must take each other to have at least a minimal degree of autonomy. We aren’t committed to any particular account of this concept. Yet it does seem to us that, at least at the point at which humans enter into friendships and enemyships, each must take the other to be self-guided to some degree. For these relationships are ones in which we take the other party to be capable of acting either to support or frustrate our interests. It is laudable to visit patients in the dementia ward, but it would be bizarre to visit those very advanced patients with the goal of making new friends.Footnote ⁵ Similarly, conspiring to thwart the interests of one of these patients would be not only conceptually but morally inappropriate. The requirement of minimal autonomy equally means that it makes no (literal) sense to take the weather, or a white whale for that matter, as one’s enemy.

1.2. Divergences: one-sidedness and affective states

Thus far, the opposition view seems in good shape: the requirements of significant interactions, appropriate dispositions, and minimal autonomy apply to both friendship and enemyship alike. But there is a straightforward divergence between friendship and enemyship that scuppers the opposition view: where friendship has to be reciprocated (Aristotle, NE VIII.2, 1156a4-6), enemyship can be entirely one-sided.Footnote ⁶ Remember that the opposition view claims that enemyship has the same structural relations as friendship, while inverting the emotional and practical valences. Asymmetry violates any of the structural relations that require reciprocity. For example, Ted might want to thwart the interests of Carol, whereas Carol is simply indifferent to Ted. Ted is Carol’s enemy, even if the disposition to thwart is entirely on his side. Friendship, by contrast, requires that each party reciprocate; each friend desires to further the other’s ends. Given the possibility of one-sided enemyship, we should distinguish between the ‘antagonist,’ who wants to thwart the interests of the other, and that other herself, whom we will call the ‘target.’

Next, consider that friends declare to each other in some fashion that they are friends, and this declaration needs to be accepted by each party (Telfer Reference Telfer1971: 230). No such declaration needs to be made or accepted in the case of enemies. It is entirely possible to have an enemy one doesn’t recognize. Given that the target might not know her enemy, enemyship, unlike friendship, need not be acknowledged by both parties. A target can even count as her friend someone who is in fact her antagonist. Martha confides in Lynn the details of the affair she is having. Lynn acts as a trusted confidante, but she is merely feigning affection in order to collect information that she can use against Martha later on.

We take another important structural relation of friendship to be equality: friends must start off on a roughly equal footing and work to preserve that equality (Kant Reference Kant and Infield1963: 204). While there is much room for variation here, it seems to us that, at the limits, there is no room for friendship: a sixth-grader cannot be friends with the adult single woman who lives next door, though they can certainly be friendly.Footnote ⁷ But any requirement of equality, however loosely interpreted, does not obtain with enemyship. Nothing stops that same sixth-grader from standing in mutual enemyship with the curmudgeonly neighbor who lives down the block. The curmudgeon just wants peace, but is convinced he will have none until that juvenile delinquent neighbor is dealt with by the authorities, which he has called—many times. For his part, the juvenile, feeling harassed by his neighbor, believes his only option is to live up to the reputation the curmudgeon has bestowed on him: so all papers delivered go straight into the bushes and all bottle rockets are aimed in one direction.

Someone can seek to thwart the interests of those who are socially, politically, and morally above or below them. Here again, we see that enemyship is not simply the opposite of friendship: it is not that one requires equality and the other inequality. Rather, enemyship is largely indifferent to whether its relata are equals or not. That said, we should note that there are certain social roles that (in principle at least) prevent a person from assuming the antagonist role with respect to specific individuals. As exhausting and irritating as a particular student might be, a teacher shouldn’t take that student as a target for her antagonism, regardless of the age and status of the student. The same can be said of a doctor or therapist and her patients. All of these professions require practitioners to support the patient’s or student’s pursuit of their best interests, not thwart them.

We have left the best case for the opposition view for last: affective states. Here, if anywhere, one would expect to find the simple inversion of friendship: where friendship requires philia, enemyship ought to require enmity. And in most cases, enemyship does involve negative feelings. Nevertheless, enemyship does not require them. There can be cases where all the requirements one might place on enemyship are satisfied, and yet there are no negative affective states involved. Suppose at a time of great financial difficulty, the Dean seeks to eliminate the German program, but Professor Frank considers it central to the mission of any legitimate liberal arts college and works to preserve this program, involving current students, professors, and alumni in his efforts. For good reasons, each is frustrated with the other. The Dean is frustrated that Professor Frank has used his personal contacts to involve donors in an internal affair, while the professor is frustrated with what he sees as the Dean’s short-sightedness. But neither hates or despises the other; in fact, they respect each other’s dedication to the institution. Still, they have competing visions and each believes the other’s efforts put the college’s future in jeopardy. Each sees the other as playing an important role in this struggle over the college’s future, such that to thwart the efforts of the other will help them achieve their goals. But nevertheless they have no animosity towards each other; both believe it is unfortunate that they are on opposite sides, for if not they could have been friends. Although such cases are perhaps rare, we see no reason to deny that these two are enemies. (We discuss a similar case in section two below.)

All this has gotten sufficiently complicated that a score-card may help:

Friendship requires

1. i. significant interactions

2. ii. appropriate actions and dispositions

3. iii. minimal autonomy

4. iv. mutual acknowledgment

5. v. equality

6. vi. positive affective states directed at one another

Enemyship requires

1. i. significant interactions

2. ii*. appropriate actions and dispositions (possibly one-sided)

3. iii. minimal autonomy

4. iv*. acknowledgment (possibly one-sided)

Although enemyship retains i and iii, and even ii and iv when adjusted to allow for one-sidedness, it does not require v or a valence-reversed equivalent of vi.

In cases where only some of the requirements of enemyship obtain, we generate nearby relationships that, although not strictly enemyship, are interesting in their own right. A case that fails to meet i generates what we have called ‘in-principle’ enemies: the absence of significant interactions makes it impossible for the antagonist to thwart the target’s goals. A case that fails to meet ii* might be called ‘parties to a truce.’ Here consider a variation on the story above of the Dean and Professor Frank. Sharon and Roman don’t share the same vision for the large non-profit for which they both work. Each thinks the other is wrong-headed on all fronts. Unlike in the previous case, they respond to their overwhelming differences by avoiding each other at all costs. Each leaves the other to do what she will in her own particular sphere. Either could, if she chose, make the work life of the other unbearable, but neither has the stomach for doing so. Thus, they satisfy conditions i and iii, but not ii*, and hence are not enemies.

We have said little about condition iv or iv*. Although there is no magical incantation that makes two people friends, it seems clear that both parties need to acknowledge the other as their friend (Telfer Reference Telfer1971: 230). We need not exchange friendship bracelets, mercifully, but we must in some way ‘dub’ the other a friend. We think a parallel requirement holds for enemies: it is not enough that our interests clash and that one is disposed to thwart the other’s pursuit of their clashing interests. The antagonist has to recognize her target as her enemy, though of course the relation can be one-sided, and the target need not recognize her antagonist. Cases where i, ii*, and iii are met without iv* are not impossible: indeed, many an enemyship begins with those first three conditions and only curdles into enemyship when iv* is met.

Suppose the manager of a gym, Agatha, finds herself routinely opposing the proposals of an employee, Tim. Tim suggests adding weekend hours, creating a Halloween party for children, and so on, in order to attract and retain customers. Sometimes, Agatha cannot find the resources to carry out these projects; at others, she thinks countervailing considerations make them unwise (does the gym’s insurance cover injuries incurred in bobbing for apples?) Agatha does not engage in the anointing step iv*, and even worries that Tim will (wrongly) think himself the target of her antagonism. Only if Agatha takes step iv*, and takes Tim as her target, does he come to stand in the enemyship relation. That fact reflects a deeper one, which we explore in the next section: enemyship is a relation between persons, not merely an opposition among interests and dispositions.

2. An Aristotelian Taxonomy

Though we have argued against the opposition view, that doesn’t mean we reject any kind of comparison between friendship and enemyship. We want to leave the opposition view behind and take inspiration instead from Aristotle’s taxonomy of friendship. For as we will show, this taxonomy is a useful tool in discovering roughly corresponding distinctions among enemies.

An important caveat here: The concepts we consider have distinct lineages, and there is no reason to expect that our distinctions will perfectly map on to a common conceptual inheritance. So our project necessarily includes an element of stipulation: if our use of ‘nemesis,’ for instance, does not perfectly accord with other uses, we are happy to say that we are stipulating what our terms will mean. What matters is that the phenomena we aim to capture are real.

Aristotle characterizes friendship most broadly as reciprocated good will, where each party is aware of the good will each has to the other: “[t]o be friends, then, they must be mutually recognized as bearing goodwill and wishing well to each other” (NE VIII.2, 1156a3-5). He recognizes three kinds of friendship: of pleasure, utility, and character. Let us begin with friendships of pleasure. Such friends have good will towards each other because each finds pleasure in the other’s company. Consider the person you are happy to see show up at a party, a person who is equally happy to see you. You two like sharing stories and jokes, and you mean to get together outside of the party, but you never do. You have affection for this person and hope she is well. But if your life changes, what is fun or pleasurable can change as well. And because of such changes, you no longer have fun with this person. That old drinking buddy is out of place among the moms at the pool. The friendship survives as long as each finds pleasure in the other, but when that pleasure ends, the friendship ends.

Is there a corresponding category among enemies? It seems possible that some of our acquaintances should fall into such a category, targets we might call ‘enemies of irritation.’ Consider Bobby, the party-goer at whose visage you recoil. You know he will tell you the same not-very-funny joke he told you last time, and you will be forced to feign mirth and surprise at the punchline. You are well aware of his uncle’s sciatica and irritable bowel, but you will hear all about them again. Is Bobby your enemy?

First, to the degree that Bobby’s interests include conversing with you, you aim to thwart those interests. And that might not be the only interest you could be thwarting. Say you have anointed him your living, breathing objection to party-going. By avoiding the parties he might attend, you are acting in a way that can lead to one or both of you being isolated from your social group. So you can also be thwarting his (as well as your own) interest in being social. Second, the reasons you have for fleeing his company have a moral element: at a minimum, he is insufficiently attentive to the preventable (if minor and temporary) suffering he causes his audience. The conditions for enemyship are met, if just barely. Just as philosophers have wondered whether a friend of pleasure is a ‘true’ friend, for Aristotle, so we may wonder whether enemies of irritation are true enemies. Nothing for us turns on how that question is settled. But it should be clear that, if the irritant is an enemy, he is at the bottom of any hierarchical ranking.

Let us move to our next Aristotelian category, the friend of utility. Such friends have good will towards the other because each is useful to the other. Here, consider an illustrator and a writer who are collaborating on a children’s book. Each might enjoy working with the other and want the other to be well as they work through this project. But when the project ends, the relationship ends. It isn’t that they now wish each other ill; but in time, each will have no or very little thought of the other.

We think there is a corresponding kind of enemyship, what we call ‘enemyship of interests.’ Here, consider brothers Alexander and James Campbell, who fought on opposing sides during the American Civil War (see J.T. Power Reference Power1994). Each was present at the battle of Secessionville, with James manning the fortifications of a fort that Alexander and his regiment stormed. When James learned of this he wrote to his brother: “I was in the Brest work during the whole engagement doing my Best to Beat you[…] but I hope you and I will never again meet face to face Bitter enemies in the Battle field […] but if such should be the case You have but to discharge your deauty to Your caus for I can assure you I will strive to discharge my deauty to my country & My caus’” (Power Reference Power1994: 135). Alexander shared this letter with his family, writing “‘It’s rather too bad to think that we should be fighting him on the one side and me on the other for he says he was in the fort during the whole engagement [.] I hope to god that he and I will get safe through it all and he will have his story to tell about his side and I will have my story to tell about my side’” (Power Reference Power1994: 135).

These brothers recognized each other as enemies, and each was willing to fight the other if they met on the battlefield. But each hoped such a tragedy would not come to pass. They differed mightily with respect to the so-called Great Cause, and acted accordingly, but each wanted the other to survive the war. They meet all the conditions of enemyship, but their enemyship was not global in nature. These brothers did not reject each other; instead, evidence suggests that they loved each other. After the battle, each made efforts to see the other, to assure themselves that the other had lived and had not been wounded, though a face-to-face reunion never happened. Instead, their enemyship was seemingly localized to the conflict between the North and the South.

Both friends of utility and enemies of interests are characterized by their local nature. Each relationship is founded on the goals of the participants, and when these change, the relationship changes. Just as the friendship of our writer and illustrator depends on their converging interests and, if founded solely on that basis, will dissolve when those interests are satisfied, so the enemyship of Alexander and James was without basis outside of the context of the war. Prior to the war’s end, Alexander was medically discharged and James was in a military prison. They corresponded, not about the war, mostly about family and mutual friends, with James at times requesting news, money, and clothes from Alexander. Unlike anonymous soldiers on a battlefield, the enemyship between these two men was personal. Here consider Alexander’s claim that he was trying his “Best to Beat you.” But because the two were kin and had a personal relationship, they were able to sequester this conflict from other aspects of their lives. Their correspondence continued after the war, ending with Alexander contemplating joining James in Charleston in hopes of securing work as a stone cutter.

Opposition of interests can generate a kind of relationship that is not quite enemyship. To see this, it is important to remember that it is not enough that the interests of the two brothers are incompatible. For one element, iv*, requires that one brother anoint the other his enemy. James hopes he never meets Alexander again as “[b]itter enemies in the battle field.” He also declares that his cause is not his brother’s, and that he will fight his brother for the sake of his cause. Such a declaration functions as a condemnation of his brother’s cause and his anointment of his brother as an enemy, meeting condition iv*. There can be opposition of interests without this anointment, however; in such cases, we have not enemies but rivals. Both John McEnroe and Bjorn Borg want to win the men’s singles title at Wimbledon in 1980. Obviously, their interests are incompatible. But this doesn’t make them enemies. In fact, insofar as they share an interest, it would be incoherent for one to condemn the other’s: neither can believe there is anything odious about wanting to win Wimbledon, since that is precisely what he wants, too.Footnote ⁸ Now, Borg wants himself, rather than McEnroe, to win; but, unlike our Civil War brothers, Borg does not want McEnroe to relinquish his interest in winning. It would be a desultory match if McEnroe stopped caring about winning; the rivalry depends on the opposition of interests. We want our rivals to have the interests they have, but we want our enemies to abandon their interests.

Enemyship of interests involves the antagonist’s condemning the interests of the other. Perhaps the limiting case of such enemyship is the relation of ‘nemesis,’ from the Greek νέμεσις, to give what is due, and the goddess of the same name. Her role is to correct the wrongs of man and to punish mortals for their hubris. A nemesis in our sense of the term is a long-standing enemy of interests, opposed to one or more important interests of the target.

Imagine that Chad is the trustee of a small college who is convinced that a group of faculty members has led the institution to near ruin: by adding positions in archaic fields, by having too generous a sabbatical policy, and so on. In Chad’s view, these faculty members are out of touch with the demands of today’s marketplace and the interests of the students. So he continually works to right the wrongs and undo what the faculty members have achieved. It seems fair to say that Chad is the nemesis of these faculty members. Depending on how we spell out the details, the relationship might well be mutual.

The third Aristotelian category is the friendship of character, where ‘character’ means not ‘moral character’ but one’s personality and core projects (see NE VIII.3, 1156b6-24 and NE VIII.4, 1157a31-36). Two people are friends of character because of who they are, not just because of what each can do for the other. Moreover, each is a good person, and so each is a good for the other person. They wish the good for each other for the sake of the other. Friendship of character is complete in that this kind of friendship subsumes the other kinds of friendship. Friends of character approve of and applaud the character of their friend. As a consequence, friends of character, like friends of pleasure, enjoy each other’s company. Moreover, friends of character, like friends of interests, support and approve of each other in their pursuit of their respective interests – at least for the most part. Often friends of character share many interests, but they need not share all of even their important interests. But when they don’t, those interests are typically compatible with each other, such that each can still approve of the other’s pursuits. If the interests aren’t compatible but yet important to one of the parties, the friendship will be strained and might even dissolve.

We will call the corresponding category an ‘enemy of character.’ The antagonist in an enemyship of character disapproves of his target because of the kind of person he is and as a consequence he will work to frustrate his target’s interests, especially his important interests. The antagonist will oppose his target’s interests in general, regardless of their content. In fact, the antagonist need not disapprove of his target’s interests in and of themselves. Iago’s antagonism of character leads him to foil his target Othello’s interests in a loving marriage and in being regarded as a rational and effective general. But Iago does so not because he condemns the institution of marriage or wants to sabotage Cyprus’s safety. He does so because he loathes Othello and who he is. Since the antagonist need not condemn his target’s interests but yet must act to foil those interests, enemyship of character is disanalogous to the friendship of character. The friends of character approve of each other’s interests and hope to help the other in satisfying these interests. By contrast, the antagonist in an enemyship of character need not disapprove of his target’s interests, though he must work to frustrate those interests. Note that, if one does not work to frustrate the target’s interests, the relationship is not yet enemyship of character. One might intensely dislike a person and yet endorse and maybe even seek to further that person’s interests. In such a case, it seems wrong to call that person your enemy of character, or even an enemy of irritation. At most, we have an irritant.

Nevertheless, it is common for an antagonist in an enemyship of character to disapprove of the target’s interests in and of themselves. Despicable people tend to want despicable things. And of course if I deem someone an enemy because of her character, I will usually find her unpleasant. She will usually be someone I wish I could avoid. But this need not be the case. The person whose character I condemn could still be a fascinating and exciting dinner companion. Thus enemyship of character, unlike friendship of character, does not require completeness, since it does not subsume the other kinds of enemyship.

The distinction between enemies of interest and of character is a tricky one, for both are directed at an individual human being. Interests don’t float free of the person who has them. The relata of enemyship, just like friendship, are always people, not the contents of their volitional states. The difference comes out when we consider the conditions under which those states persist. Enemyship of interests dissolves when those interests change: if one brother had become an abolitionist and deserted the Confederate Army, the enemyship would have been without a basis. Not so with enemies of character: such enemyship can go on, even when the antagonist finds nothing objectionable in the target’s interests.

The limiting case of enemyship of character is the archenemy. Archenemies are enemies of character, but raised to a mythical level. An archenemy is your primary enemy, and in this case, the relation must be reciprocated. You need not differ in every way; just like Sherlock Holmes, Professor Moriarty had a keen analytical mind. But you do differ with respect to most of your important interests. The pursuit of the other’s defeat is relentless, ending only in death. Such relationships are familiar from comic books and superhero stories. In real life, however, almost none of us has a Professor Moriarty against whom to struggle over decades.

With our categories in place, we can ask one last question before moving on to normative issues. There has been much debate over whether Aristotle thinks that only friendships of character are ‘true’ or paradigmatic friendships, or whether all three kinds are equally entitled to the name. We remain neutral on that issue. Still, it is worth asking whether any of our three kinds of enemyship deserves the mantle of ‘true’ enemyship. What answer we give to that question turns on what criterion we choose.

One reason to think that friendships of character have a better claim to being ‘true’ friends than those of utility or pleasure is that friendship of character can withstand the vicissitudes of circumstance. The other kinds of friendship come and go depending on the interests and desires of the people in question. When friendship or enemyship dies with the changing of one’s interests or (dis)pleasures, it seems hard to maintain that the relationship was really built around the person. But when the friendship or enemyship is directed at the character, then the relationship seems to have the right kind of foundation.

If, on the other hand, we want ‘true’ to mean ‘most common,’ we should probably settle on enemyship of interests. Though we might, with self-conscious hyperbole, call someone our ‘archenemy,’ we typically find ourselves at odds with other humans, not because of a conflict of character, but because our interests conflict. Lacking such a conflict of interests, our tendency is to just avoid the person we find loathsome. And if we sought to interact with this person when we didn’t have to, we would be irrational. Moreover, it is probably rare that our interests conflict because we condemn the character of the other person. Typically, conflict arises because of the content of those interests. If our target were to align her interests with ours, the grounds of the enemyship would vanish. There may still, of course, be lingering ill-will, and we might wonder what kind of person our target can be, who even for a time favored, say, the Confederacy over the Union. But to the degree we are rational and capable of forgiveness, we will abandon our hostile approach to our target.

3.1. Should We Have Enemies of Character?

Matters are different with enemies of character. We shall argue that no one should have such an enemy, where the ‘should’ here is the should of practical rationality, not morality. Being the antagonist in such a relationship is a threat to one’s own ability to achieve one’s ends. We shall argue that, even in cases where it is clearly morally permissible to anoint another your enemy of character, it is nevertheless unwise.

Recall that what is distinctive about such enemies is that the antagonist takes the target’s character to be worthy of condemnation. The opposition to the target’s interests is parasitic on opposition to the target’s character. In real life, of course, there will be many cases where the antagonist also condemns these interests because of their intrinsic nature, and not merely because they happen to belong to the target. But here, we shall focus on the case of ‘pure’ enemies of character, bracketing the question whether the target’s interests are odious in their own right.

Our first question, then, is whether, and under what conditions, it is morally permissible, or even mandatory, to have an enemy of character. Two notes of caution are relevant at this point. Some of us have a tendency to see the world in Manichean terms, a tendency that should in general be resisted. And if we suspect that our urge to anoint another our enemy is driven by this tendency, we would be wise to hesitate. A second note of caution arises from an epistemic asymmetry between friendship and enemyship: in the case of friends of character, we typically know a good deal about each other, and do our best to stay informed. If my friend’s character changes, I may try to change along with her (as we both, say, give up drinking and commit to our studies); or I might worry that our friendship will not survive. With enemies of character, we are typically much less knowledgeable about any changes that might be occurring; I am likely to be doing my best to minimize contact of any kind with my target. That means that, even if my target’s character were to change, I would not be well-positioned to know about it. In practice, then, there is a risk of preserving an enemyship of character even after the target’s character has altered. The nature of enemyship makes this risk much greater than its counterpart in friendship.

Setting aside these epistemic issues, we should distinguish two very different cases. Imagine Miranda, a despicable person whose chief goal in life seems to be vexing everyone who has the misfortune to come into her orbit, even when doing so is retrograde to her own interests. Miranda’s default position is disinterested malevolence. The natural question one has to ask before anointing her an enemy of character—however repulsive that character might be—is how she got to be that way. In this case, let’s stipulate that Miranda is suffering from an undiagnosed brain tumor and that, were this tumor not present, she would be a perfectly ordinary person. Since Miranda’s behavior is entirely driven by her tumor, we think it would be morally wrong to anoint her an enemy of character. For she would fail condition iii, the condition of minimal autonomy. She is simply another danger among others, like cyclones and food poisoning.

We need our targets to be autonomous in part because they need to be moral agents, responsible for what they do, and not just moral patients who serve as mere conduits. If they truly lack agency, then we are justified in trying to stop them, but not in holding them morally responsible. So, in this case, moral considerations would prevent us from making Miranda our target in an enemyship of character, though of course that makes her no less dangerous, and no less in need of opposition wherever possible. Now, whether it is psychologically possible for us to avoid taking up a negative reactive attitude toward her is a different question. But we think it’s fairly clear that, while one ought to oppose Miranda, one ought equally to hope that she receives treatment, if possible; Miranda is a proper object of pity rather than enemyship.

A very different case is provided by Lou. To make matters as stark as possible, let’s work with as uncontroversial a case as we can. Lou’s behavior, we can stipulate, has no pathological explanation. Yet in his own way, Lou is just as unpleasant as Miranda. He blames everyone else for his own wrongs, and when confronted with his misdeeds, shows absolutely no remorse, and indeed, intends to repeat them, if he can. Now, many people might deserve to be targets of an enemyship of character who fall short of Lou. But surely if anyone does, it’s Lou. Isn’t one not just morally permitted but positively obligated to make Lou an enemy of character?

We grant that doing so is morally permissible. We hesitate, though, to say that it is obligatory. For there is a competing consideration to bear in mind: one’s own practical rationality. Since, in this case, our target is our target because of his character, we aim to thwart the target’s interests regardless of the content of those interests. Here we encounter the familiar situation of cutting off one’s nose to spite one’s face. If I am against whatever Lou wants to do just because Lou wants to do it, I stand to lose out on opportunities to advance my own interests, including productive interests Lou happens to share. Becoming Lou’s antagonist in this way can imperil my ability to achieve my ends. Such a relation can lead a person to act against her own interests, both apparent and all-things-considered.

Someone might object that, if I do indeed know that Lou’s character is worthy of condemnation, then I automatically have a reason to try to frustrate Lou’s projects. For I know these projects flow from a vicious character. And even if I cannot see anything wrong in, say, Lou’s proposal to limit the library’s weekend hours in the summer, I know that he has something up his sleeve: he wouldn’t be proposing this unless it furthered his malevolent goals.

Knowing that someone’s character is vicious does indeed provide reason to be suspicious. If we found ourselves in a foxhole next to Lou, we would certainly have to wonder whether we’d chosen the right side. But if we are practically rational agents, we should acknowledge that these reasons are defeasible by considerations intrinsic to the project in question. If restricting summer hours is itself a good thing to do, the mere fact that Lou has proposed it should not prevent us from endorsing it.

Yet there might be further reasons, beyond the intrinsic merits of Lou’s proposal, to try to thwart him. If he succeeds, even in achieving a morally neutral goal, he might be better positioned to carry out more nefarious actions in the future. Indeed, his overall plan might be to build up a reservoir of good will, which he can draw on later to further his other, less savory goals. We grant that such situations are possible. The caveat is that opposing Lou’s idea of restricting summer hours merely because it is Lou’s risks diminishing one’s own standing. One is liable to appear unreasonable to the rest of the committee; that might well deplete our own stock of good will among our fellows, and make it correspondingly more difficult to oppose Lou when he rolls out a more sinister proposal in the future. How to deal with a thoroughly bad piece of work such as Lou is rarely straightforward.

To sum up: the real problem with having enemies of character is not that entering into the role of antagonist in such a relationship would itself be a moral failing, though it sometimes is, as when the target is not a moral agent. Instead, the problem with having enemies of character is that it risks undermining our own ability to pursue rational courses of action.