Agent-Relative Deontological Thresholds

Abstract

Personal relationships have an inherent normativity, and this normativity consists in the possibility of deontic change: personal relationships can change the deontic status of acts. Such change usually takes the form from optional to required. In other words, we can have a duty to do something for a friend that would be optional to do for a stranger. For example, while it might be supererogatory and praiseworthy (and in this sense optional) to spend a Sunday helping a stranger who desperately needs help moving house, it might very well be obligatory to do the same for a friend who desperately needs help.

Other forms of deontic change are possible, even though they can take more controversial forms. Perhaps relationships can change the deontic status from impermissible to optional or even required. In such cases, it would be permissible or even obligatory to do things for the sake of a loved one that would be impermissible when done for the sake of a stranger. Are such cases possible? Can we be permitted or required to tell lies, break promises, or do even worse things to benefit our loved ones?

Many people’s moral intuitions seem to support this possibility. Suppose that you know that your best friend has cheated on their spouse. You also know that it was a single event that your friend regrets deeply, that your friend and their spouse love each other dearly, that the relationship is going very well otherwise, and that the partner would never forgive adultery. And suppose that you can save their marriage by lying to the partner. Intuitions might differ, but I think that many would agree that in such a case, lying would be morally permissible or perhaps even required.

Nevertheless, some accounts of friendship seem to deny this. Take for example an Aristotelian account of virtue friendship. If telling a lie or breaking a promise is not virtuous, then it should always be impermissible, regardless of whether we can benefit a friend by doing so (and a virtuous friend might not even want us to do these things, even if they would benefit them). And perhaps there are good reasons to be wary about the possibility of such deontic change, since we do not want to allow for the possibility of such change in deontic status across the board. Even those who accept that an agent can be required to lie in order to benefit a friend need not hold that an agent can also be required to kill someone to provide a good job opportunity to their friend, or that the agent can be required to embezzle money to pay for a nice trip to Paris for their children.

However, this worry represents a questionable kind of slippery-slope reasoning. Most deontologists allow for circumstances that make otherwise impermissible acts permissible, but this does not mean that they believe that everything becomes permissible at some point. If deontologists accept that things that are usually impermissible can become permissible under certain circumstances, then they should at least be open to the possibility that relationships are among the relevant circumstances when this happens. But of course, to defend this idea, two important questions need to be addressed. First, we need to explain why relationships can change the deontic status of acts in such ways. Second, we need to explain why they can do so only in certain cases, but not across the board.

In this paper, I will argue that threshold deontology offers a unified explanation to both questions. Threshold deontology is the idea that morality includes so-called restrictions, which means that some things are impermissible regardless of their consequences, and that these restrictions have thresholds, so it can become permissible or required to do things that are usually impermissible if enough is at stake (Ellis 1992; Alexander 2000; Rosenthal 2018; Hanna 2018). Threshold deontologists usually assume that deontological thresholds do not vary across agents. Thus, once a deontological threshold is reached, all agents are equally justified in violating the respective restriction. However, if we accept that deontological thresholds can be agent-relative, then this explains why some people can be permitted or even required to do things that are impermissible for others (and why this can happen only in certain cases).

Before I elaborate on this, let me motivate why we should consider the possibility that relationships can change the deontic status of acts from impermissible to optional or required in the first place.

1. Relationships and Deontic Status

It is by no means obvious that relationships can make otherwise impermissible acts permissible or even required. Consider a case that has gained some prominence in the literature, namely the Death in Brunswick case that has been introduced to the philosophical discussion by Dean Cocking and Jeannette Kennett:

Death In Brunswick: Carl, the main character of the film Death in Brunswick, is no saint. Weak, vain, and disorganized, he is a severe disappointment to his mother. He drinks too much, and he works as a cook at a seedy nightclub in Brunswick where he falls foul of the owners by falling in love with a young barmaid. One night, Mustapha, his drug-dealing kitchen hand is badly beaten up in the back alley by the nightclub heavies. Carl is warned to keep his mouth shut; Mustapha is told that Carl is responsible for the beating. So late that night, Mustapha staggers into the kitchen and lunges at Carl, who is holding a long-pronged fork. Mustapha impales himself on the fork and dies. In a panic, Carl calls his best friend Dave, an easy-going family man. Against the protests of his wife, June, Dave dresses and drives to the nightclub to see what is up. His initial response when shown the body is that the police must be called. Carl begs him not to, saying that he could not cope with going to jail. Faced with Carl’s fear, Dave takes charge and helps Carl move the body. They take it to the cemetery where Dave works, he breaks into a coffin in an open grave, stamps on the putrefying corpse inside to make room for Mustapha, and re-closes the coffin. Later, they deny all knowledge of Mustapha’s disappearance to his distressed widow and son. (Cocking/Kennett 2000, 279–80)

How should we understand Dave’s actions? There are several possibilities. First, one could hold that what Dave does is wrong, plain and simple. Dave ought to not help Carl: it is morally wrong, and it is also against the requirements of friendship because true friends help each other do the right thing, become more virtuous, etc. Second, one could argue that what Dave does is morally wrong, but that he proves to be a good friend by helping Carl. According to this view, friendships are not morally innocuous: they present a ‘moral danger’ (Cocking/Kennett 2000) because they can bring us to do morally wrong things out of friendship. One could even argue that friendships involve relationship-dependent nonmoral duties that can conflict with moral duties. This would mean that Dave does what he ought to do as a friend, even though he ought not to do it from a moral point of view.¹ A third option is that Dave has an indirect moral duty to help Carl. For example, Daniel Koltonski (2016) argues that Dave ought to let Carl decide what to do in the situation and defer to Carl’s judgment, and if Carl decides that they ought to move the body together, then Dave ought to follow Carl’s judgment. And finally, one could argue that Dave ought to help Carl, full stop: as a friend, Dave has a moral duty to help Carl.

Which of these interpretations should we accept? The first seems overly moralistic. Cocking and Kennett plausibly point out that we would hardly call Dave a bad friend. Their own view is more plausible, but it faces (at least) two problems. First, it implies a deep tension between morality and personal relationships. The view suggests that friendships, love relationships, etc. are only contingently justified by morality. It is merely a lucky coincidence that the demands of these relationships are in line with moral demands, but things could be otherwise, and relationship-dependent duties and moral duties could conflict across the board. Second, the phenomenology supports a moral interpretation of Dave’s obligation.² From Dave’s perspective, helping his friend presumably feels like a moral obligation. He experiences the situation as a moral dilemma, that is, as a clash between two moral obligations. In addition, failing to help Carl would result in reactive attitudes that are characteristic of the violation of moral obligations, such as feelings of guilt and shame on Dave’s side or feelings of resentment on Carl’s. The moral phenomenology also speaks against the third interpretation, according to which Dave has a moral obligation to defer his judgment to Carl. Suppose that Dave does not help Carl, Carl ends up in jail, and Dave experiences feelings of regret. What would he regret in this situation? It seems plausible to think that he regrets not having helped Carl, rather than not having deferred his judgment to Carl. Furthermore, the obligation has two features that are often taken to be characteristic of moral obligations. First, it is welfarist in nature in the sense that its content is to promote the well-being of another person (in this case, Carl). Second, it is categorical in the sense that it does not depend on whether Dave wants to help Carl or not—as the case is described, Dave is very reluctant to help Carl and does not want to help him at first.

Taken together, these considerations might not be a knockdown argument for understanding Dave’s action as the fulfillment of a moral obligation, but they suggest that there are at least good reasons to think that Dave is morally required to help his friend Carl. However, it also seems plausible to think that other people who are not close friends with Carl are not permitted to help Carl avoid jail, just as it would be wrong for Dave to do these things to benefit a stranger.

If this is roughly on the right track, then this means that relationships can change the deontic status of acts not only from optional to required, but also from impermissible to required (even if this involves harm to a third party, in this case, Mustapha’s family). This raises the two questions that I mentioned in the beginning: why can personal relationships turn impermissible actions into morally required ones, and why can they do so only in certain cases?

2. The Change in Deontic Status

A good starting point to explain the change in deontic status from impermissible to required is to consider accounts that explain a deontic change from optional to required and examine whether they also apply to the cases that interest us here. One prominent approach holds that relationships (either directly or indirectly) intensify the strength of reasons, in particular the reason to benefit a loved one (Lord 2016; Lazar 2016; Löschke 2018). This intensifying of reasons then makes otherwise optional acts required. For example, a person might have reason to benefit a stranger as well as a loved one (say, she has a reason to help her friend move house and a reason to help a stranger move house), but the reason to benefit the loved one is stronger than the reason to benefit the stranger in virtue of the relationship. This has an impact on what the person ought to do, all things considered: because it is weaker, the reason to benefit the stranger is more easily outweighed by countervailing reasons (such as the cost to the agent, prima facie options to pursue personal projects, or something along these lines) than the (stronger) reason to benefit the loved one. There are different accounts of how moral reasons lead to moral obligations, but an intuitively plausible idea is that this has something to do with the strength of the moral reasons and whether they are outweighed by countervailing reasons. If relationships intensify the strength of moral reasons to help a loved one, then this can explain why benefitting a stranger can be optional even though benefitting a loved one is obligatory.

Does this explanation extend to cases where the deontic status of an act changes from impermissible to required? The friendship between Dave and Carl would then intensify Dave’s reason to benefit Carl to the extent that the reason outweighs countervailing reasons to not help Carl, including moral reasons that speak against doing so. This account would provide an elegant and unified explanation of how relationships can change the deontic status of acts. However, it does not succeed.

The problem is that the account might explain why relationships change the deontic status from impermissible to required and thereby answer the first of the two aforementioned questions, but it does not explain why they only do so in certain cases, since it does not rule out cases of nepotism that intuitively appear wrong. If relationships can generate a duty in Dave’s case by intensifying Dave’s reason to benefit Carl, then we should expect the same in other cases in which an agent can benefit a loved one. Suppose that a high-ranking politician can benefit his children by providing them lucrative government jobs for which the children are clearly unqualified. And suppose, furthermore, that these jobs would benefit the children significantly. Such cases of nepotism seem morally wrong, but this is difficult to explain based on the account under discussion. After all, according to that view, the politician’s reason to benefit his children is intensified due to their relationship, and the reasons that speak against providing his children these lucrative jobs should be weaker than the reasons that speak against Dave helping Carl (it is arguably morally worse to do the things that Dave does than to provide one’s children with lucrative jobs for which they are not qualified). Hence, the account under discussion seems to imply that the politician is morally required to provide his children with jobs for which they are not qualified. This is certainly a wrong result. While it would be interesting to think about the deeper reason why the intensifier account can explain the deontic change from optional to required but not from impermissible to required, I do not have the space to discuss this matter here.³ Instead, I want to propose an alternative explanation of the deontic change from impermissible to required. This explanation relies not on the idea that relationships intensify moral reasons, but on the idea of deontological thresholds.

Threshold deontology is the view according to which so-called deontological restrictions are an important part of morality, where these restrictions are not absolute but have thresholds. Deontological restrictions are things that we are not permitted to do even if doing them would result in a better outcome, such as lying, stealing, torturing people, or killing innocents. To say that these restrictions have thresholds is to say that at some point, agents are permitted or even required to violate them if enough is at stake—if, in other words, violating the restriction is the only way to avoid a moral catastrophe. So, for example, it is impermissible to torture a person even if torturing one person would prevent the torturing of two other persons, but if torturing one person is the only way to prevent the death of thousands of people, then torturing the person might become permissible or even required. Threshold deontology is not without criticism,⁴ but it is certainly an intuitively compelling view. And it also helps to explain both why relationships can change the deontic status of acts from impermissible to required, and why they can only do so in certain cases.

To see this, note that threshold deontology does not merely provide a weighing explanation to determine the moral status of acts. The threshold deontological idea is not that restrictions are very weighty reasons that are at some point outweighed by consequentialist considerations—or at least that is not the best account of deontological thresholds (for example, it is not clear why restrictions should be so weighty that they can only be outweighed in the face of a moral catastrophe). Threshold deontology should rather be understood in a non-weighing way: restrictions hold independently of considerations about consequences, but at some point and in certain cases, considerations about consequences become relevant to determining the deontic status of an act.⁵ This non-weighing explanation of the change of deontic status from impermissible to permissible/required is an important difference from the intensifying view, which, as we have seen, has problems explaining why this deontic change only happens in certain cases.

The idea of deontological thresholds promises a natural explanation of why relationships can change the deontic status of an act from impermissible to permissible or even required, and why they can only do so in certain cases. After all, the whole idea of deontological thresholds is that in specific circumstances, things that are otherwise impermissible become permissible or required, but only in these specific circumstances. Thus, Dave can be required to do things that would otherwise be impermissible if and because this is the only way to avoid a catastrophe, but he would not be permitted to do anything to benefit Carl. If Dave cannot prevent a catastrophe, then he is required to respect deontological restrictions.

But of course, this leads to a further question. Deontological thresholds are usually understood as agent-neutral in the sense that whether a threshold is reached does not depend on the identity of any particular agent. This also means that every agent is equally permitted or required to violate a restriction if the threshold is reached. Why is Dave then permitted to violate the restriction, but not strangers? I suggest that this is because deontological thresholds can be agent-relative.

3. Agent-Relative Deontological Thresholds

The idea is that what counts as a moral catastrophe can be relative to agents, and it can depend on (among other things) a person’s relationships. Things can be a moral catastrophe, relative to Dave, that would not be a moral catastrophe relative to others. Thus, the friendship between Dave and Carl sets an agent-relative deontological threshold that makes it permissible or required for Dave to do things that would be impermissible for others who do not have the same kind of relationship to Carl, just as these things would be impermissible for Dave if he did them to benefit someone who was not his friend. But how can relationships play this role? To answer this question, let me start with some assumptions that I take to be independently plausible.

First, ‘moral catastrophe’ is something of a technical term. Threshold deontologists use the framework not only to explain why it can be permissible to torture someone or kill an innocent person under very specific circumstances, but also to explain why it can be permissible to break a promise or tell a lie. In the latter cases, the deontological threshold is much lower than the threshold of the restriction against torturing people—breaking a promise is not only permissible when human lives are at stake. This suggests that what counts as a catastrophe that makes it permissible or required to violate a restriction is relative to the restriction at stake. ‘Moral catastrophe’ should thus be understood as a technical term that stands for an outcome that is so objectionable from a moral perspective that it becomes permissible or required to violate a restriction. This, of course, means that the term cannot serve to justify the violation of a deontological restriction, but rather signals that the violation of the restriction is justified. This might appear unsatisfactory, but the justification of threshold deontology must wait for another day.

Second, I assume that it is just if virtuous persons fare well and that it is unjust if virtuous persons suffer. Similarly, it is just if vicious persons suffer and unjust if vicious persons fare well. Of course, some ancient philosophers would argue that it is conceptually impossible that vicious persons fare well or that virtuous persons suffer, since a virtuous life is a good life, but most contemporaries will find such a view difficult to accept (and for good reason). It is certainly intuitively plausible that there is something objectionable from a moral point of view if a deeply immoral person lives a happy life or if a good person faces hardships that she does not deserve. If this is true, then this also suggests that, at least in principle, situations in which virtuous people suffer can be moral catastrophes in the aforementioned technical sense of the term and can therefore justify the violation of a deontological restriction.⁶

The third assumption that I take to be independently plausible is that people do not form intimate relationships with people they consider vicious. We are, of course, aware that our friends have flaws and weaknesses, but we are usually convinced that our friends are good persons, all things considered, and that their flaws and weaknesses do not define who they really are. Of course, there are possible counterexamples to this claim: mobsters engage in deeply immoral lifestyles, but we might want to allow for the possibility of genuine friendships between them. However, the assumption is not that friendships between vicious people are impossible (although that may be the case),⁷ but that it is not possible to form a friendship with someone who one considers to be vicious. And mobsters (I assume) find ways to rationalize their behavior and justify their lifestyles. Thus, even if they know that they engage in immoral conduct, mobsters (I assume) also view their mobster friends as good persons, all things considered, and if that is the case, then friendships between mobsters are not counterexamples to this third assumption.

Together, these assumptions explain why deontological thresholds can be agent-relative. Carl and Dave are friends, and this means that even though Dave is aware of Carl’s flaws, he considers Dave a good person, all things considered—otherwise he would not have formed and maintained a friendship with Carl. Due to their friendship, Dave has intimate knowledge about Carl’s overall character and has experienced the virtuous sides of Carl’s character in his own life (as well as Carl’s vices). He can therefore judge whether Carl deserves the very bad consequences that could be avoided by violating a deontological restriction (such as lying to the widow). And in the example, it seems that Dave can in fact reasonably judge that Carl does not deserve these bad consequences. He knows that Carl did not intend to harm Mustapha, let alone kill him, and that it was an unfortunate accident. Since Dave considers Carl an overall virtuous person, he can reasonably judge that Carl would suffer undeserved and therefore unjust harm if he were to go to prison because of this accident. After all, this one moment does not reflect Carl’s character, it does not reflect who Carl really is, and Carl does not deserve the lifelong repercussions that would result from this single incident. In short, as a friend, Dave can reasonably come to the conclusion that suffering such considerable harm would be something that Carl does not deserve as the (overall and fairly) virtuous person that he is. If Carl can reasonably conclude that Dave does not deserve these consequences, then he can also reasonably judge that it would be a moral catastrophe if Dave suffered them.

The result is, then, that Dave can reasonably judge that he is justified in violating a deontological restriction to prevent this moral catastrophe. Because he can reasonably judge that he is justified in violating a deontological restriction, Dave is justified in doing so. Thus, Dave’s obligation to help Carl can be given a threshold deontological explanation: he is justified in violating a deontological restriction because this helps to avoid a moral catastrophe (in the technical sense of the term as specified above). And this threshold is agent-relative. It exists because of the friendship between Carl and Dave, so other people would not be justified in violating the restriction.

The friendship between Carl and Dave plays an epistemic role in making the deontological threshold agent-relative. Because they are friends, they have greater knowledge of each other’s characters than strangers do. Dave knows the virtuous parts of Carl’s character because he has experienced them over the course of their friendship, and he therefore also knows that it would be inappropriate to reduce the assessment of Dave’s overall character to this specific situation when assessing whether Carl deserves the harm that he would face without Dave’s help. Other people who do not have the same access to the virtuous parts of Carl’s character cannot make the same assessment of the situation. As a result, they can reasonably judge that Carl does deserve considerable harm because of what he has done. Strangers can base their assessment solely on the individual situation (here: Carl killing Mustapha)—in fact, they have no other basis for assessing the situation. It would be inappropriate for them to simply assume that Carl has virtuous character traits and is a good person all things considered; hence, they are in no position to make a valid overall evaluation of Carl’s character. Dave, by contrast, can and ought to make a comprehensive assessment that takes information from their common history into account. Dave’s assessment of Carl must be diachronic to be reasonable; by contrast, the assessment of third parties must be synchronic and focus solely on the individual situation. This explains why the deontological threshold in this situation is agent-relative: it depends on Dave’s epistemic situation and his common history with Carl, which third parties do not share.

Some readers might think that there is an obvious objection to this view. If the friendship between Dave and Carl is only epistemically relevant for creating the agent-relative deontological threshold, then it is not clear whether the agent-relative threshold depends on the friendship between Dave and Carl after all. In fact, it might not even be agent-relative. One might argue that, in principle, every person might be able to gain knowledge about Dave’s character, not only his friends, and that every person who knows about Dave’s character is in a similar position to judge whether Dave deserves the harm that he is about to face. Thus, every person could be equally permitted or required to violate the deontological threshold. Suppose, for example, that we are not talking about Dave, but about a well-known philanthropist whose virtuous character is widely known. In this case, everybody seems to be permitted or required to prevent harm that the philanthropist does not deserve and violate a deontological restriction, not only the philanthropist’s friends.

To answer this possible objection, the first thing to note is that it is by no means obvious that we can have the same knowledge about a stranger’s character that we can have about a friend’s character. A relationship makes it possible to witness someone’s behavior in a wide range of situations, to get first-hand rather than second-hand knowledge about her actions, etc. These things are not possible with regard to a stranger, even if we read in the newspaper about their generous contributions to charity (or something along these lines). Hence, relationships make an important difference regarding the evaluations that we can make of people’s characters.

And yet, the objector might insist that a trustworthy person could inform us about the virtuous character of their friend, and in such a case, we might have all the relevant information available: the trustworthy person has first-hand knowledge about her friend’s character, and we can trust her in her assessment. We can therefore reasonably judge that her friend has an overall virtuous character. Or imagine a Truman Show scenario in which it is possible to observe a person’s behavior in all situations, even in situations in which that person believes they are alone and unobserved. In such a scenario, we would clearly have as much (or even more) information about the person’s overall character as we have about the characters of our friends. A personal relationship does not seem to be necessary to make an appropriate evaluation of their character.

There is, however, an important difference between friendships, on the one hand, and second-hand knowledge of a person’s character and even Truman Show cases, on the other. Epistemologists distinguish between propositional knowledge—knowledge that something is the case—and experiential knowledge or knowledge by acquaintance.⁸ It is notoriously difficult to get a clear understanding of what experiential knowledge amounts to, but the distinction is intuitively plausible: it makes a difference whether you hear that something is the case, or whether you experience that thing directly. Take the case of a religious experience. There is a difference between hearing that God exists, reading the Scriptures, and being convinced by some philosopher’s proof of God’s existence, on the one hand, and having an actual religious experience, on the other hand. A religious experience provides the person who has had it with a certainty about God’s existence that merely reading about God or engaging with proofs of God’s existence does not. A person who has had a religious experience has more reason to believe in God, since she has (from her perspective) undeniable evidence that God exists. In other words, a person who has had a religious experience can more reasonably come to the conclusion that God exists than a person who did not have such an experience.

Unlike propositional knowledge, knowledge by acquaintance does not need to be conceptualized—a religious experience can have undeniable force even if the person does not conceptualize the experience as a religious one while she is having it and only comes to understand later that she has had a religious experience. This means that knowledge by acquaintance is in an important sense not only more certain, but also more fundamental than conceptual knowledge: at least in some cases, we first have an experience and conceptualize it in a certain way afterward.

The distinction between knowledge by acquaintance and propositional knowledge can help to explain the possibility of agent-relative deontological thresholds. It is true that we can learn about any person’s virtuous character by hearing or reading reports of their behavior, and in Truman Show cases, we might even be able to witness their virtuous character. But this is merely propositional knowledge—we lack experiential knowledge of the virtue of strangers, since their virtue is not part of our lives. We are not acquainted with their virtue. This is different in the case of our friends. Over the course of a friendship, friends treat each other in ways that are manifestations of the virtuous parts of their characters. Friends experience each other’s virtue first-hand, and they do so in a nonconceptualized way. They are the beneficiaries of each other’s kindness, generosity, or courage, but they do not necessarily experience an act of generosity as an act of generosity; what they experience is the receipt of a benefit (even though they might later conceptualize the act as an act of generosity). As experiential knowledge, this kind of knowledge that friends have of each other’s character is more fundamental than the kind of knowledge that they can have of the character of strangers, and it also has a higher degree of certainty.

The experiential knowledge of our friends’ virtuous characters makes an important difference. Evaluating whether a deontological threshold is met is at the same time very difficult and very important. If we err on the side of deontology, this means that we fail to prevent a moral catastrophe; if we err on the other side, we violate a deontological restriction and therefore commit a serious moral wrong. Thus, the bar to evaluate whether one is justified in violating a restriction is very high, and the standards for assessment are quite demanding. Due to its fundamental nature and its higher degree of certainty, experiential knowledge contributes to a reasonable assessment that the deontological threshold is met, whereas mere propositional knowledge hardly ever allows such a verdict. If there is an important difference between propositional knowledge and knowledge by acquaintance, and if knowledge by acquaintance has an epistemic advantage over propositional knowledge, then there is a sense in which relationships do make a difference regarding the evaluation of our friends’ characters that we can reasonably make.

I have argued that relationships play an epistemic role in the explanation of the agent-relative deontological threshold: we know that our friends have a virtuous character and that they therefore do not deserve the considerable harm that they might face if we do not violate a deontological restriction. But now one might worry that the actual explanation is that our friends’ virtuous character is part of our lives, whereas the virtuous character of strangers is not. But these are not contradictory explanations. Because the virtuous character of our friends is part of our lives, we have not only propositional knowledge of our friends but also experiential knowledge: we are acquainted with the virtuous character of our friends in a way that we are not acquainted with the virtuous character of strangers.

Let me sum up. For those who do think that Dave is morally required to help Carl avoid jail, the idea of agent-relative thresholds offers a promising explanation of why the obligation exists. And as I will argue now, the idea of agent-relative deontological thresholds also explains why such an obligation exists only in certain cases.

Recall the case of the high-ranking politician who provides his children with elite jobs for which the children are clearly not qualified. Such cases of nepotism should emerge as impermissible in the agent-relative threshold deontological picture. And in fact they do. The politician cannot reasonably judge that it would be a moral catastrophe if his children do not get jobs for which they are not qualified, even if getting those jobs would benefit the children significantly. As a result, the politician cannot reasonably judge that he would be justified in violating the relevant deontological restriction, and it is therefore impermissible for him to do so. Thus, the agent-relative deontological threshold account explains why relationships can change the deontic status of acts from impermissible to permissible or required only in certain cases.

Here is another case where the account delivers the right result. Suppose that Carl did not kill Mustapha by accident but deliberately, with the intention of making him suffer as much as possible. In this case, Carl would certainly deserve the harm that would result from him going to jail. Some cases of wrongdoing are so severe that they ought to take priority in evaluating a person’s overall character. Dave certainly could not reasonably judge that Carl was a virtuous person all things considered if he had deliberately and cruelly killed a person during a moment of weakness. In such a case, the deontological threshold would not be met, and Dave therefore ought to call the police rather than help Carl cover up his wrongdoing. Thus, the agent-relative threshold deontological framework delivers the intuitively right result in such cases as well.

4. Conclusion

Personal relationships can change the deontic status of acts in various ways. One—admittedly controversial—form of deontic change is from impermissible to permissible or required: we can be permitted or required to do things for our friends that would be impermissible when we do them for strangers. For those who believe that such deontic change can happen, the idea of agent-relative deontological thresholds offers a promising account that explains why such deontic change can happen as well as why it only happens in certain situations. This is, admittedly, not a knockdown argument that relationships do play this role and do change the deontic status in this way. However, it is an explanation of a phenomenon that many would consider an important part of their moral lives, and it relies on assumptions that are independently plausible: that deontological restrictions have thresholds; that there are cases in which virtuous people do not deserve to suffer considerable harm; that in many cases, we can reasonably judge that our friends have an overall virtuous character; and that there is an important difference between mere propositional knowledge and experiential knowledge, with the latter having a kind of certainty and fundamentality that the former lacks.

FUNDING

Work on this paper was supported by the Swiss National Science Foundation (SNF), grant number PP00P1_213975/1 (project title: Value-Based Deontological Thresholds).

REFERENCES

Footnotes