Ethic Theory Moral Prac (2011) 14:503–516 DOI 10.1007/s10677-010-9257-4

Open Duties Alexander Jech

Accepted: 8 December 2010 / Published online: 5 January 2011 # Springer Science+Business Media B.V. 2011

Keywords Duty . Deontology . Underdetermination

There is a feature of the deontological landscape that, although a pervasive feature of our lives, has hitherto escaped notice. None of the categorizations of duties of which I know succeeds in capturing a common and important form of duty, which I will call an “open duty.” Consider the categorization often associated with Kant, which distinguishes between perfect and imperfect duties.1 A perfect duty is often conceived of as an absolute prohibition or obligation with regard to a specific kind of action; for example: do not murder, do not lie, do not break your promises. An imperfect duty however consists in an obligation to promote a certain end or to be responsive to certain kinds of concerns: help the poor, educate your children, develop your talents.2 These are not thought to be as strict or precise as perfect duties because of a certain lack of specificity and of compossibility, and the latitude given to choice in determining how to fulfill them.3 E.g., should that last dollar go to your child’s college fund, or to famine relief, or to that speed-reading class you have heard about? As useful as this distinction can be, it cannot account for open duties, which do not stand as a third kind of duty alongside these two, but cut across them instead. This is because the distinction between perfect and imperfect duties relates to the kind of end or action enjoined by the duty, whereas the difference between an open duty and a closed duty rests not on the side of 1

A distinction historically connected to the Enlightenment Natural Law tradition. According to Andrew Blom’s comments on an earlier version of this paper, this distinction goes back at least as far as Grotius’ notion of “imperfect rights.”

2

Some philosophers admit only perfect duties as real duties and consider imperfect duties supererogatory; I do not consider this very plausible, but the point is tangential to the discussion here.

3

To be sure, this distinction can be a bit wobbly: imperfect duties can be completely obligatory in particular instances, and it appears that few today think that perfect duties are completely exceptionless, so it may be unclear what exactly the distinction amounts to. Perhaps we should say that while perfect duties enjoin rules, imperfect duties enjoin ends. A. Jech (*) Program in Political Philosophy, Policy and Law, University of Virginia, Charlottesville, VA, USA e-mail: [email protected]

504

A. Jech

the end or action, but on that of the agent who is enjoined to act. A closed duty belongs to one or more persons, all of whom are required to act to fulfill the duty. An open duty always belongs to more than one person, not all of whose performance of the action is required to fulfill the duty. 1 The Need for “Open Duties” Consider two situations: one is hypothetical, the other historical. In the first of these, an elderly man slips and falls at a bus stop where several other persons are standing waiting and struggles to get back to his feet. Any one of those standing at the stop might help him, and one of them does help him up. Are the others blameworthy for failing to do so? If no one had helped, then surely they all would have been blameworthy. In the actual event, most of them did not help. Yet they are not blameworthy. Further, if the action was required, then why was it necessary for someone to volunteer to perform it? Second, consider Hurricane Katrina, a Category 3 storm whose second landfall struck as far east as Florida and as far west as Texas in late August 2005. The hurricane created severe and widespread devastation throughout the Gulf Coast. Many of those surviving the storm required urgent medical assistance. In light of this need, it was morally required that American medical professionals (at least) travel to the region to provide aid. The importance and urgency of the need made it imperative for them to provide this service. Yet it would have been impossible to have pointed to any specific individual and say that he or she must provide aid. The situations described above are quite ordinary, but they raise questions that are difficult to resolve using traditional deontological equipment, questions about the conditions for blameworthiness in failing to perform one’s duty, about the relationship between volunteering and doing what is required, and about our ability to assign duties to particular individuals. The first situation could easily be resolved if there were only one bystander to the elderly man’s fall. Everything is straightforward: when the elderly man cannot get to his feet, helping him is morally required, and if the man fails to perform the action, he is blameworthy. He cannot do nothing and remain free from blame. There is no question about his needing to volunteer for the duty. And there is no difficulty determining the specific individual to whom the duty belongs. But a difficulty arises from increasing the number of bystanders at the bus stop with the elderly man that ordinary deontic categories cannot capture. Whatever complications, caveats, and conditions we may build into these concepts, there is never any difficulty in discerning to whom the duty belongs, because we are dealing with one agent. However, we find it hard to assign duties to members of such a group. To whom do these duties belong? If we assign it to everyone involved, how is it that in some circumstances they can knowingly leave it undone yet go without blame? This problem in assigning the duty is the distinctive feature of an open duty. The first task is discerning the nature of what we see: how could there be duties falling upon unspecified agents? What would such duties have to be, to have this nature? In Constructions of Reason, Onora O’Neill characterizes some imperfect duties as duties that possess unspecified recipients, duties that therefore do not yield corresponding rights (O’Neill 1989, 189).4 Her idea is suggestive: could there be duties with unspecified agents, 4

In fact, she appears to suggest that this quality is fundamental to imperfect duties, but this cannot be right, if even one’s duties to one’s children and oneself can be imperfect (consider the duty to educate one’s children). The fundamental feature of an imperfect duty, I think, is the importance of some set of ends, the need to weigh each of these against each other, and the variety of means by which someone might promote them: it is these features that give imperfect duties their peculiar cast.

Open Duties

505

which likewise do not yield corresponding rights because it is impossible to specify from whom to extract them? Just as the proper recipient of an imperfect duty to assist the hungry is, in a certain sense, all of the hungry, but not any one of hungry person specifically, so too might there be duties that, while possessed by a whole class of persons, are not the obligation of any specific person out of the group? In certain cases, the nature of imperfect duties allows a degree of latitude about “to whom.” Why is this? It may be that when an ethically significant end is larger than my own powers to promote, then the duty must be vague as to recipients. If I could relieve hunger as easily as I refrain from murder, or if there were only one starving person who required assistance, then no doubt there would be no ambiguity about to whom the duty is owed. So if there are ends that belong to a group of persons, but whose performance requires fewer than all of them to act, do we then have duties whose agents are unspecified? The example above appears to show such a duty, a duty that is “open” with respect to which one of those possessing the duty should perform it. There is something paradoxical about the whole idea of a duty without a specified agent, a paradox which this paper will attempt to resolve. More than a few readers will have some doubt about the reality of this paradox and suspect that, suitably deployed, the concept of imperfect duties will do the work needed here. I must ask these readers to wait until §4 for my rebuttal.

2 The Structure of Open Duties An open duty is expressed by saying that Some are to perform action A, or, more rarely, Some are not to perform action A.5 The first kind is frequently manifested in disasters—for example, in the wake of Hurricane Katrina, we should have said that Some medical professionals are to provide aid to the victims of Katrina—and also in a variety of everyday contexts, e.g., in any household of more than one person, Someone is to take out the trash. The second kind of open duty—where Some are not to do something or another—seems to be most commonly found in situations like a tragedy of the commons, where a resource is not initially scarce but would be exhausted or lose value dramatically if everyone were to utilize it to the full.6 Here, Some are to refrain from using whatever resource is in question. This phrasing—asserting that open duties can be expressed by saying Some are to A—is sure to recall to some readers what Peter Geach said about the deontic logic of statements I have simplified this here. Note that the weakness in the “some” could also appear with regard to the action to be performed if there are several actions that would achieve the end, so the stricter rendering would be: Some are to perform one of actions or sequences of actions A1, A2,…An to promote end E. This would mean that it was imperative only that some one of those actions be performed. I will restrict my analysis to the limit case in which only one action can achieve the end because the issue raised by the underdetermination of means by an end is a problem of practical reason that arises for many closed duties as well. 6 The original example of the tragedy of the commons included an open pasture shared by a number of shepherds. There is just enough grass for all the shepherds to graze all of their present sheep. If more sheep are added, then the grass will eventually become exhausted. The problem is that for any given shepherd the cost of lesser grass from adding a sheep is very small, and spread amongst all the other shepherds, whereas the added revenue from wool, etc., is large, and retained only by himself. If the shepherds equally and unintentionally put themselves in a state where there are too many sheep for the pasture to support, then an open duty, some are not to pasture their sheep here (or, some sheep are not to be pastured here), will be laid upon them. (I say “equally and unintentionally” because otherwise it the duty to remove sheep falls upon whichever shepherds increased the number of sheep beyond the maximal number, not on the shepherds as a group). The solution to such a problem is typically institutional: either collective action—forcing the shepherds to use the commons responsibly—or privatization—preventing the shepherds from socializing the costs of over-grazing. Open duties, as I will explain below, frequently require institutional solutions. 5

506

A. Jech

like “Someone ought to shoot the Shah of Persia!” (Geach 1982). Geach points out that someone who says this “is not likely to mean that the predicable ‘— ought to shoot the Shah of Persia’ is true of somebody: but rather, that the predicable ‘— ought to be shot (by someone or other)’ is true of the Shah” (Geach 1982, 4). The differences between this case and a case like that of the elderly man who slips in front of several bystanders should be quickly apparent. The most striking difference is this: If it is true that the Shah ought to be shot, but not true of any individual that this individual ought to shoot him, then no individual will be blameworthy if no man shoots the Shah, whereas if none of the bystanders helps the fallen man, they will all be blameworthy. This is because in Geach’s scenario, there is reason for the Shah to be shot, but no one with reason to shoot him, whereas with an open duty all members of the closure class have reason to perform the action. Geach’s scenario would only illustrate an open duty if there were several individuals whose duty was to shoot the Shah, rather than none, and the question was which of them should take the shot. The problem is that, when multiple agents have reason to perform the action, and ought to perform it for that reason, but the action requires only one of them to perform it, we cannot say that any one of those who should perform the action is required to perform it. Open duties’ chief characteristics are their teleological orientation and their lack of specificity. There is a gap between the action by which some end could be achieved and the agents who are to perform the action. So let us characterize these duties as occurring when the four following conditions are met7: 1. End E is to be achieved 2. Action A and only A would achieve E 3. There is a set of persons, C, which consists in all those persons P1, P2,…Pn who could adequately perform A or contribute to performing A 4. Some but not all of the persons in C are required for the successful performance of A Call the set of individuals C the closure class and E the ethical imperative. From conditions 1 and 2, we can conclude that: 5. A is to be performed Conjoining this to condition 4: 6. Some but not necessarily all of the closure class are to perform A The general principle behind open duties is that if there is an end that it is ethically imperative to achieve, and there are persons who can perform the action that would achieve it, then it is a duty for those persons to perform it. This principle yields open duties if we accept that there are circumstances where a class of persons could perform this action that requires only a proper subset of them. The closure class can be anything from a chance congregation of individuals brought together by circumstance, such as witnesses to an accident, to a group with more substantial unity, such as a political community. The members are linked together by their common relation to the ethical imperative, viz. their being in a position to achieve it, and not necessarily by any direct relation that the members hold to each other. The bystanders at the bus stop, for example, lack any firm relationship with each other. Each appears in the closure class because he or she is equally capable of coming to the elderly man’s assistance, and none has any tie to the man significant enough to require his or her special attention. Various factors can complicate determining the closure class. For example, if my car breaks down by the roadside, and there are several passing drivers who can help me, does it 7

As mentioned in footnote 7 above, 2 is a simplification, but this does not affect anything of substance.

Open Duties

507

make a difference if one of them is a trained mechanic, and the others merely amateurs? If a child is drowning in a pond, does it make a difference if the child is one I know? What if it is my daughter? Or suppose four friends are out at lunch together and one of them receives a call from the hospital and must travel there as quickly as possible, but does not own a car. Does it make a difference if one friend has nothing else to do while another friend has a full schedule? In short, there are at least three important factors to consider here: the skill or ability to perform the duty, preexisting moral ties to the end to be achieved, and the ease of performing the duty. Each of these factors can come into play both in determining the closure class and in determining the relative obligation that each member of it possesses towards the duty itself. One’s degree of obligation depends upon the degree to which one excels in these factors. The interplay of these factors can greatly complicate how we evaluate any situation involving an open duty. In the most general sense, the closure class includes just those persons who are capable of performing the action without violating competing moral obligations of at least equal importance. However, it is wise to distinguish a narrower closure class from this extended or wider closure class. The narrow closure class is determined by the three factors mentioned above: ability or skill to perform the task, antecedent moral ties to the end to be achieved, and expected ease of execution. Call these the qualifying factors. Those who qualify most in these factors relative to the other members of the closure class fill the narrow closure class. The remainder fill the wider closure class: they are the amateur mechanics, the strangers standing by, or those who would have difficulty performing the duty. Determining the relative closure class of class members is important for moral evaluation, as we will see, although those in both classes may volunteer to perform the duty. The same considerations also illustrate how those in a closure class may attempt to collectively deliberate about an open duty. Whereas when an individual deliberates about whether to perform an imperfect duty, he must determine what place this act has within his whole nexus of ends in order to determine how much, when, and how to promote the end in question, when a group of persons deliberates about an open duty it must additionally determine who should perform the duty. This may prove impossible if all are equally qualified. However, in some circumstances the varying degrees to which individuals are more “qualified” may allow a kind of deliberation—although even so it will likely not be weighty enough to settle the matter. The ethical imperative could be as mundane as taking out the trash or as urgent as saving a life. The first might seem too thin to support any duty, but it is sensible to say that it is an end of some importance for a household that the trash be disposed of if the living quarters are to remain livable, implying a duty that can be discharged by any member of the household but must be discharged by one or more of them. Indeed, anyone who has lived with roommates should recognize how important it is to determine the bearers of open duties of this kind. Although these duties require that some ends possess a kind of importance independent of the agents in the closure class, the duties themselves are not independent of those in the closure class, floating about who-knows-where in the moral landscape. They are thus (again) distinct from the “ought-to-be-ness” that Geach criticized (cf. especially Geach 1982, 1–4). For it is the end and not the duty whose existence and moral importance is independent of agents obligated to achieve it. To avoid the unfortunate idea that the duties themselves are impersonal or independent of the persons in the closure class, we must distinguish two matters here: first, the

508

A. Jech

importance of what is to be achieved, and second, the special features of an open duty regarding agents. The ends grounding open duties are, in some sense, important. It does not matter how we spell out “importance”; the point is simply that it matters that the end be achieved. So, imagine a possible world within which, as in many disaster movies, a large asteroid is headed for the earth. The asteroid is too large to be stopped before striking the earth and wiping out most life on the planet. Now this event would matter to us, if any event ever would. The continuation of life on earth is important to us, and it would be important to preserve it, if anyone could achieve this end. However, in this possible world no one has the duty to stop the event because no one can do anything to stop it. Anyone able to stop it would be obligated to do so, but there is no such person. Some matters, then, are important whether or not anyone makes them their purpose, and whether or not anyone has a duty to see to the matter. Now, although some matters have importance regardless of whether someone has a duty to see to them, the same is not true for duties, even open duties. Open duties require there being some persons for whom the duty exists, just more than are necessary for its performance. Return to a possible world in which an asteroid is headed for the earth, but imagine instead that we know how to stop it. The mission to stop the asteroid requires a certain number of astronauts, but only a few—far fewer than the number available. In fact, sending more than this number would hinder the mission. Because of the great importance of this mission, a duty rests upon all these astronauts to stop the asteroid, but not on any astronaut in particular. This duty does not float free of the astronauts even though the importance of its purpose is independent of their agency—the duty exists, only belonging to all of them. If the asteroid were headed for the earth a hundred years earlier, before any astronauts existed, then no open duty would exist. Life on earth would remain important, but because there would not be any means by which to preserve it, there would not be any duty to do so. Thus although there might be a kind of impersonal importance that is independent of persons (or not, as ones philosophical account demands), there are no impersonal duties. As I said before, I call these “open duties” because their fulfillment is “open” to any member of the closure class actually entering into and appropriating them. The weakness and generality of the proposition stating the conditions for discharging the actual duty—Some of P1, P2,…Pn are to perform A—arises from the underdetermination of who should perform the duty. Similar cases arise in understanding functional explanations in science. For example, Hempel analyzed functional explanations as deficient because of the weakness of a key premise: All vertebrates circulate blood and A heart may circulate blood, yet it is illegitimate to infer All vertebrates possess hearts, since other means of circulating blood are possible (Hempel 1965). The second premise is insufficient to guarantee a determinate conclusion. Thus, what one can infer is that some one of these means, call them M1, M2,… Mn, must be present in all vertebrates. In the case of the open duty, we have an ethical imperative whose performance may be achieved by some subset of a class of moral agents. Who is to perform A, given that any number of members of the closure class could perform it? This question is unanswerable until the duty is actually closed by some one or more members of the closure class. The circumstances created by Hurricane Katrina illustrate this problem.8 It was imperative that some doctors travel to the Gulf Coast to give aid, but it was not imperative 8

It also illustrates another issue: the members of a closure class may vary over time. At different times, different individuals may be in a position to perform the duty, and may drift in and out of the closure class, or from wider to narrower subsets of the class, as their circumstances very. Thus, with respect to possession, open duties are synchronically closed, and diachronically open; with respect to fulfillment, they are simply open.

Open Duties

509

that all doctors travel there. Rather, an even greater disaster would have ensued if all had done so. Nor was there some property P, sufficient either to exclude or include a doctor, such that we could determine exactly which doctors ought to provide aid. (Presumably, there were some doctors whose specific circumstances and qualifying factors demanded either that they go or that they refrain from going, but we aren’t concerned with these: we are concerned with the great mass of doctors who are able to provide aid and without any special circumstances preventing them from or requiring them to do so.) More doctors could give aid than were needed to give aid. (Or, if this assumption is false here, then imagine some disaster for which the condition holds.) A general duty fell upon doctors in the United States such that some of them were obligated to travel to the Gulf Coast to give aid. But which among them? Somehow the common duty must be assigned to specific individuals, and the rest relieved of the duty. How should we characterize this moment of closure?9

3 The Moment of Closure Since the duty falls on all members, how does one person discharge it? Somehow open duties must be discharged by the agents who perform them. If the duty is discharged, then there must be some moment during which the class of agents with an obligation to perform it becomes completely specified, or closed. To distinguish these two states I will call the duty as it falls upon the whole closure class the “unspecified open duty” and the duty that has been closed by particular members the “specified open duty.”10 Therefore, the question is how an unspecified duty becomes a specified one. This process can be thought of as analogous to the process by which goods held in common come to be the private possessions of individuals within a group, a process which, depending upon whether a central authority or an individual member directs the activity, may be called either “distribution” or “appropriation.” From the perspective of the person performing the duty, distribution is non-voluntary, appropriation voluntary. The nonvoluntary distribution of open duties will concern us later in the paper. In the voluntary case—which is primary and paradigmatic—at time t1, persons P1, P2,… Pn each possess a certain relation to the action, A, for which reason they fall within the closure class. Suppose that only one member of the closure class is required to perform the duty. Then suppose that at time t2, P1 chooses to close the duty with a good prospect of success; at t2, then, there is no longer an ethically significant relation between P2…Pn and A.11 The duty has become specified and no longer belongs to the closure class as a whole. Rather it has become the sole property of the person who has closed it.

9

Cf. Robert Nozick’s (Nozick 1974) discussion in Anarchy, State, and Utopia (137–142) concerning the puzzle about the right that all possess in the state of nature to punish. The differences between the right to punish and the right to extract compensation might be explained by considering the right of punishment to be an open duty. Just as the most effective way to solve the difficulties that arise from open punishment is through institutions of justice, so too (I will argue below) the most effective way to see to the performance of many open duties is through the formation of institutions designed to carry out those duties. 10 I do not wish to use the phrase “closed open duty” because this seems confusing, and because it suggests we ought to use the unwieldy phrase “open open duties” to refer to the opposite class. 11 If the closing agent’s prospect of success is not sufficiently good—to be determined, I think, through some weighing involving both the importance of the end and the probability of success—then the remaining members of the closure class ought to stand by to see if the duty is in fact discharged.

510

A. Jech

What happens during this moment of closure? Consider the analogy of someone taking an item from a common store. When does the common good become a private good? Suppose we focus on goods held in common within a family. The food stored in the pantry and kitchen is, with a few exceptions, held in common. However, individual members of the family, at least those of a certain age or maturity, may appropriate these goods to themselves as they see fit. How does this occur? In two ways: someone may decide to consume something, and take it. Alternatively, someone may tell the rest of the family that something belongs to her specifically, by saying something like, “Don’t eat the tortilla chips: I’m saving them for a party.” Likewise the unspecified duty becomes specified when someone decides to perform the duty, and either does so, or communicates his appropriation to others by some sign. The obscurity of closure varies tremendously. One is tempted to say that an intention to perform an action achieving the imperative end is sufficient, but this does not seem accurate. If, for example, one of the four friends at lunch discussed earlier announces her intention, saying, “I will drive you,” then she has closed the group’s duty and it belongs to her alone. However, appropriating the duty is not the same as intending to perform it. Someone standing 20 feet away from the water could intend to save a drowning child, but if someone else, standing closer, rushes in to save the child (performs the duty), then the first person will not have successfully closed it. She will have intended to close it. She did this by trying to perform the action. What then constitutes actually closing the duty? At what point did the person who did in fact rescue the child relieve the others of fulfilling it? Not, as we have seen, when one decides to do so. Not the moment, either, when one begins to move towards the child. The person who only attempts to close the duty also moved towards the child. But certainly when someone in fact rescues the child; and we can draw back the moment of closure to the moment when one began to actually perform the action of rescuing the child, with good prospect of success, as opposed to performing some action preparatory and subordinate to this one, such as moving into position to rescue the child.12 One closes the duty, then, if one begins to engage in whatever activity is specifically called for by the duty, or else makes clear by certain signs that one has chosen to take on the duty and there are reasonable odds of completing the task. If some task were both extremely important and extremely difficult to achieve, closure might not occur until it was actually accomplished. One does not slay the dragon by riding up to his lair, but by cutting off his head.

4 Open Duties and Imperfect Duties In this section I will examine the relationship between open duties and imperfect duties. Two matters need to be addressed: First, are open duties really distinct from imperfect duties? Second, what is the initial moral relation between the members of a closure class and the action to be performed? I argue that this relation is partially that of imperfect duty. As to the first, it might be objected by some that open duties are nothing more than unusual imperfect duties. An imperfect duty consists in the duty to promote a certain end, whose performance we possess some latitude in determining “how” and “when” and “how 12

The prospect of success is an important ingredient in this; the moment of closure does not come at the same moment when a lifeguard begins swims out to save someone as when an amateur swimmer does so.

Open Duties

511

much.” We may see to the promotion of one end (e.g., famine relief) or of another (e.g., the care of my children), and we may promote these different ends in a variety of ways and at different times. It is our duty to always and everywhere refrain from intentionally killing the innocent, but it is not our duty to always and everywhere give alms to the poor. We must give, but how much, to whom, and when, are all determined by weighing the varying needs of the ends towards which we have imperfect duties. There is a significant overlap between open duties and imperfect duties in the kinds of duties falling into both categories, such as acts of beneficence. The objection then is this: Situations in which we must perform an imperfect duty are often those in which another person needs our help. The same can be said of “open” duties. Therefore, are not those situations that are said to generate open duties just situations in which there is a plurality of individuals each with an imperfect duty to promote some end by the performance of some given action, complicated by the plurality of individuals who might perform the action? Is there a real mark by which an open duty can be distinguished from an imperfect duty? Suppose we admit all that was said in the objection regarding the kinds of situations that give rise to open duties: 1. These situations include some end that each member of the closure class has a duty to promote 2. Were only the minimum number of persons necessary to perform the duty present then there would only be an imperfect duty 3. The result of full deliberation on the part of those persons mentioned just above would be that it is morally necessary to act in accordance with this particular imperfect duty (as is occasionally the case, as when we meet someone in desperate need, à la the Parable of the Good Samaritan) The conjunction of these three conditions leaves only one distinguishing mark for the open duty, which is the peculiar underdetermination of agents. Yet if this is allowed, then it is allowed that open duties exist, for it is nothing but this underdetermination of who is to perform the duty that distinguishes an open duty from other duties. All other distinctive features of an open duty seem to flow from this first property. The truth is that open duties are not a category of duties alongside perfect and imperfect duties (if this is the categorization we prefer), but rather, cut across these categories. Even perfect duties sometimes become open: suppose that three people promise each other to help whichever of them should ever be in need, and one of them does come into trouble, but the help of only one of the two others is required: the same situation will arise. The objection that many of the circumstances appropriate to an open duty are the same as those appropriate to an imperfect duty is invalid since these similarities are based on the common teleological orientation of the duties and do not extend to the decisive criterion of the open duty, which is the lack of specificity about who is to perform the duty. A related response to open duties is to suggest that they are, if not imperfect duties, then duties to see to something. These duties are sometimes referred to as “responsibilities” (cf. Goodin 1986). Such duties often occur in hierarchal settings where a manager, for example, is charged with seeing to it that his department accomplishes some task. He might, for example, be told to see to it that inventory is checked according to a set schedule. Now he could accomplish this by doing it himself, but usually he will assign it to other employees. This latitude would not be available if he were simply told to do it himself. Although there are several similarities between open duties and responsibilities—e.g., both are centered on

512

A. Jech

states of affairs and both provide a degree of latitude to a person—this attempt to equate them falters. It is not clear that the minimal involvement required in an open duty is compatible with a responsibility: parents will not escape blame if, despite being perfectly able and willing to do so themselves if necessary, they nonetheless shuffle off the responsibility of raising their children onto the grandparents; but this is generally permissible for an open duty. Moreover, if the difference here is just a matter of distinguishing between a narrow and a wider closure class, then we will need to explain the special properties that open duties possess in virtue of the initial lack of specificity about who is to perform the duty. So, open duties cannot be assimilated to responsibilities.

5 The General Duty to Close Open Duties In addition to whatever other moral reasons someone has to close an open duty, there also appears to be an imperfect duty to take on one’s fair share of open duties. Now we have a general duty of giving to the poor, but the duty does not include a concrete specification as to how much we must give, to whom we must give, at which times we must give, and so on; deliberation fills in this gap with definite decisions based upon circumstances. There are times when deliberation is indifferent to the exact time when we promote the end of an imperfect duty. This is the general case for the duty to close open duties and to relieve others of performing them, because often deliberation cannot determine whether or not closing the duty is necessary in a given instance. What I can consider is my activity of closing these duties and relieving others of them, considered over some suitably long period of time (sometimes relative to others with whom I enter into closure classes). Although we cannot discover which ones we must close, we can look back over our lives and determine whether we are fulfilling this duty, and if we never, or relatively infrequently, close such duties it will follow that we are blameworthy, by allowing the weight of these duties to fall disproportionately upon others and not doing our fair share. This matter stems from our relationship with other human beings as fellows with whom we share in the performance of duties. It is separate from the ordinary blameworthiness conditions for open duties—to be treated in the next section—that are derived from a specific ethical imperative. Each individual appears to possess a single imperfect duty to close open duties and relieve others from performing them, in addition to several imperfect duties related to the different closure classes of which he is a member. (For example, he might have a duty to close open duties within his family and a separate duty to close open duties in other associations, and so on). First, allowing suitable weights to other ends and duties, someone needs to discharge a suitable number of open duties over his lifetime. A general failure to close open duties is blameworthy. Second, even if in general someone has closed his fair share of open duties overall, he also owes it to the members of each of his persistent closure classes to close his fair share of the open duties falling upon the class. Thus, someone might escape from the open duties falling upon him at home by volunteering for more and more administrative tasks in his department at work. This person has discharged his general responsibility to adopt open duties but has been unfair to his wife, upon whom the family duties have presumably fallen. Of course, it is possible that one might never get obligated to perform open duties falling upon certain closure classes: someone else might keep discharging the duties first. E.g., if a member of ones department at work enjoys administrative tasks and therefore leaves few

Open Duties

513

opportunities for others to take them on, or if one spouse enjoys domestic responsibilities and takes these on by choice. Fairness applies only when both find these duties burdensome or desirable. However, most cases with disproportionate distributions of duty-fulfillment involve moral failure.

6 The Conditions of Blameworthiness The most striking aspects of open duties are made clear in moral failure. It is here that their distinguishing features are most evident. So what are the conditions for blameworthiness; i.e., those that justify one in imputing blame to someone for failure to act? Let us begin with an observation. We frequently experience the openness of open duties as a powerful temptation to avoid performing the duty, which social psychologists call “the bystander effect.” This can usually be exposed by considering what would happen if the closure class were exactly the required size, i.e., by closing the duty. When someone standing in a crowd has fallen and needs help, it is very rare for someone to step forward to offer assistance. However, alone, any one of those persons in the crowd would do so. Frequently, too, a person will “spread the blame” to escape censure if he is confronted by his failure to act by protesting, “No one else did it—why aren’t you blaming them?” We should not allow this pervasive mode of moral rationalization to obscure the true nature of open duties, however. Return to the bystanders at the bus stop. Suppose that none of the bystanders assist the fallen man; suppose too that each of them recognizes both the need and the fact that it is not being met. How can we tell that the bystanders failed in their duties? Is it not in their each being capable of performing the act, knowing both that it needed to be done and that it was not being done, while failing to do anything? Let us apply this more generally. Then we shall wish to say that all of those who could have performed this duty, knew that it needed to be done, and knew it was not being done, were blameworthy for failing to do it. Therefore, the conditions for blame are as follows; suppose: B1 B2 B3

There is an open duty D falling upon the closure class, C (P1, P2,…Pn) D has not been performed Then All members, P1, P2,…Pn, of C such that (a) P is aware of D, (b) P is aware of being able to perform D, and (c) P is aware that D has not been performed, are, to the degree that their qualifying factors qualify them to perform D, blameworthy

Call the external conditions specified in B1 and B2 and the three personal conditions specified in B3a–c the blameworthiness conditions. The effect of the qualifying factors upon degree of blameworthiness means that the narrow closure class will be more blameworthy when they meet the blameworthiness conditions than when those in the wider closure class do so. Thus, in the wake of Hurricane Katrina, we can use these factors to distinguish the moral status of American doctors from European doctors who might be able to help those in need.13 If the European doctors came because the American doctors were unwilling or grudging in their willingness to help, then indeed the American doctors would be extremely blameworthy, because helping is far easier for them and because they possess stronger moral ties to their countrymen. On the other hand, if it were the Europeans who were unwilling to help then they might be subject to blame, but to a lower degree than the 13

A question raised in correspondence by Michael DePaul.

514

A. Jech

American doctors. The issue of unwilling members of a closure class draws attention to one kind of nonvoluntary closure: When it is clear to P that blameworthiness will apply to P if P does not close the duty in question, then the closure would no longer be voluntary. (I.e., once it was clear that some need exists that wasn’t going to be met because it had been shirked by the other members of the closure class.) All non-voluntary closures are, from the perspective of the agent, completely obligatory, the decision of the agent being forced by the evasion of the others. The general duty to discharge open duties encompasses a separate kind of moral failure: the failure to perform any open duties, or to perform very few of them. A person might have a dozen open duties fall upon him, not perform any of them, but escape direct blame because others always performed the action. However, per §5, he is still blameworthy because he has never undertaken to close an open duty. Though a person is not required to perform any given open duty, he may still be blamed if he never or rarely performs one of the open duties, since he has a general duty to close such duties as are compatible with his other activities, giving proper weights to each of these. There are conditions for blame that can apply after the duty is closed. These conditions will apply when some agent has closed a duty, but performed it negligently, or abandoned it. Suppose that someone stops alongside the road to help a traveler with a blown tire, but then performs the job so shoddily and inattentively that he ruins the spare tire. He then offers to drive the traveler to a gas station where he can arrange to have his car towed, but takes his time on the way and stops for a leisurely breakfast. The recipient of the action can justly blame this “Good Samaritan,” and it will do no good for him to say, “I didn’t even have to stop to help you, and here you are complaining!” The unhelpful mechanic’s excuse is evidently inadequate. Alternatively, had the mechanic simply driven away after ruining the spare tire, he would be blameworthy for abandoning the duty after appropriating it. Because abandonment frequently imposes further costs upon both the recipient of aid and the other members of the closure class, it appears blameworthy in ways different from a simple failure to act, and indeed exceeds it. However, we will leave this tangent aside and move on. So add the following as the second set of blameworthiness conditions: 1. There is an open duty falling upon P1, P2,…Pn (the closure class) 2. P has closed the duty 3. P has performed the duty negligently or has abandoned it With this new conceptual equipment in hand, it is easier to analyze those puzzling scenarios given above. In the case of the bystanders at the bus, there is an ethical imperative, which is that the elderly man needs to be helped to his feet; this end can be achieved by stepping over to him and offering one’s hand to him; any of the bystanders can achieve this goal, and so all fall into the closure class to whom the duty belongs. The duty closes as soon as one of the bystanders acts with a reasonable expectation of success at achieving the imperative. When one person does step forward, we see why no one else is blameworthy, and we understand under what conditions each would have been blameworthy had different conditions obtained, and how one of them could discharge a duty belonging to the others.

7 Institutionalized Closure of Open Duties Open duties motivate both the creation of institutions and the formation of offices within them. We should begin with their role in the genesis of institutions. Why do we establish

Open Duties

515

fire departments and soup kitchens and other institutions like these? One reason is that we cannot always be sure that when a building is on fire or someone is hungry that there will be someone present to help. However, even if we abstract from this, we still require institutions. Someone who faces hunger while surrounded by hundreds of other individuals will have many sources from which to receive food, but he is unlikely to be regularly fed. Some days he may receive much, others, very little. To avoid this undesirable feast-or-famine situation, the community could establish an institution that would see to the performance of this duty, narrowly specifying the duty to fall just upon those who work within the soup kitchen. Similarly, for fires—even if our community were filled with people with sufficient training to help and if fires broke out only in areas where people were present, it would be more effective to employ an organized department devoted to this task rather than relying upon an unorganized network of individuals to perform the same. In this way, the duty to help those trapped in fires would become closed upon those who are members of the fire department. We see the same logic at work in a household that must determine who is to take out the trash. Although the family can allow anyone who notices the trash to be full to perform this task, this is rarely as desirable as clearly designating who should do it, and when. Open duties fall upon groups having more members than are necessary to perform the task. Some subset of the group must be specified despite there being no reason to greatly prefer one subset of the group to another. With regard to some open duties these conditions lead to ineffective performance, as it will generally be unclear what each individual ought to do, and to unfair distribution of duties, as some individuals exploit the situation to act as “free-riders.” With respect to the issue of efficiency, is it more effective for each citizen to help put out such fires as he notices or to establish a fire department? For each citizen to help those hungry he meets or to establish a food bank and a soup kitchen? We also see that it may be very unclear what exactly my duty in a given situation is, and that it is time-consuming to determine which member of a closure class should act. Very often, it may turn out that too few members of the closure class close the duty, out of ignorance of what others are doing. This lack of information and clarity can also be exploited as an excuse by free riders who do not do their fair share to close various open duties or to contribute in other ways. Thus, to provide for the more effective performance of open duties, greater clarity and publicity about to whom duties belong, and to ensure fairness of distribution of duties, a community or association may choose to establish institutions to close open duties. Thus, communities establish fire departments, police departments, food banks, and the like, which convert an open duty into a conditional duty: i.e., it is now your duty to save those in a burning building if you are a firefighter. The institution exists specifically to perform the action and, by determining which subset within the community is to be taken as having duties to perform it, makes the performance more consistent and regular. This does however create a new open duty to join the institution in question, a duty whose fulfillment however is less urgent and immediate. This transformation may explain why open duties are often difficult to recognize. This new open duty may be closed either voluntarily or nonvoluntarily—for example, one could assign membership in institutions by birth or conscription by lottery. Institutions also tend to create open duties for their members, for example duties concerned with ensuring its continued survival and flourishing. Consider the formation of a family or a household, which generates a variety of duties that are not specified with respect to who is to perform them: Who is to take out the trash? Who is to clean the house? Maintain it? Educate the children? These things must be done but someone must decide

516

A. Jech

who it is to be. Therefore, it is plausible that any association or society generates open duties, inasmuch as various roles and functions related to the achievement of its goals and maintenance must be taken on where there is a plurality of members sufficiently fit to adopt the role(s), e.g., jury duty. These two processes, then, the formation of institutions and the formalization or semi-formalization of rules determining how open duties are to be closed, largely explain why the specific character of open duties so often goes unrecognized. The institutions and rules capture the moral force of these duties, while these same structures elide their distinctive features. The result is that, although open duties are indeed pervasive, they principally appear nakedly only in unexpected crises and chance events whose transitory nature conceals their ubiquity. Acknowledgements The final form and content of this paper has benefited from the comments of many readers and listeners, but especially from those offered by Karl Ameriks, Robert Audi, Andrew Blom, Michael DePaul, Jon Garthoff, Roger Knights, and two anonymous referees at Ethical Theory and Moral Practice.

References Geach P (1982) Whatever happened to deontic logic? Philosophia 11:1–12 Goodin RE (1986) Responsibilities. Philos Q 36:50–56 Hempel CG (1965) The logic of functional analysis. In: Aspects of scientific explanation. Free Press, New York, pp 271–307 Nozick R (1974) Anarchy, state, and utopia. Basic Books, New York O’Neill O (1989) Children’s rights and children’s lives. In: Constructions of reason. Cambridge University Press, New York, pp 187–205