To have an Obligation is to have Sufficient Reason to Act

Obligations of the type in question here are genuine in the sense adumbrated in the previous chapter: if one has a genuine obligation to do a certain thing, one then has sufficient reason to do that thing. This brings in the complex notion of having sufficient reason.

Breaking this down, I start with some amplification of the point that one who has an obligation of the type at issue has reason to act in such a way that the obligation is fulfilled. This is not to say that if you have reason to act in a certain way you have an obligation to act in this way. One may have reason to act in a certain way without having an obligation so to act, but not vice versa.

(p. 28 ) The notion of having reason to act, as I am construing it here, is a relatively broad one. Its breadth may be demonstrated by reference to the following schema.

Person X has reason to perform action A if and only if, there is some consideration C such that C speaks in favour of X's performing A.

Thus suppose, first, that some act Sue might perform would be an act of charity or kindness. I take it that its being of this kind speaks in favour of doing it. If so, Sue has reason to perform that action, by virtue of A's character. Suppose, second, that Sue decided two weeks ago to perform some action, and has not changed her mind. I take it that this consideration speaks in favour of her performing that action today, though in a different way. If so, Sue has reason to perform the act she decided upon. Her decision speaks in favour of the act, I take it, by virtue of what a decision is, not by virtue of the character of the act or of the consequences that are likely to flow from it, given the way it is in itself. One may, then, have reason to act in a certain way by virtue of the character or consequences of that act. One may also have reason to act in a certain way by virtue of other considerations.

I have deliberately couched this point about obligation in terms of ‘having reason to act’ as opposed to ‘having a reason to act’. Philosophers tend to think of a reason for acting in a certain way as something that is securely attached to the nature or consequences of the act itself. ⁵ This may accord with the way people tend to talk of reasons, and marks, of course, an important type of consideration.

If someone asks why you are doing something, he may well have such a consideration in mind. Thus suppose someone asks Jane why she is going to vote for a particular candidate in today's election. Should she reply, ‘Because I decided to do so some while back,’ this would quite likely provoke the response ‘Yes, but why did you so decide?’ What the questioner wants is a statement about what is good about Jane's voting for that candidate, either in itself or in terms of its consequences.

Nonetheless, it is hard not to think of one's prior decision, in and of itself, as in some way speaking in favour of one's performing the act decided upon. Suppose Mike decides to post his tax return this afternoon. Night falls, (p. 29 ) however, and he has not posted the return. He had become absorbed in his work and it was too late to post the return when he stopped for a rest. Remembering his decision, he is likely to feel things have gone off course. With this implication he may say to himself, ‘Oh, I meant to post my tax return this afternoon!’ This would not be the case if, in his estimation, his decision in no way spoke in favour of his acting as he had decided. Mike's likely reaction to his failure to do what he decided to do suggests that a decision does not only give one reason (in my sense) to do what one decided to, do. It also gives one what I refer to as sufficient reason to do it. ⁶

I say that

X has sufficient reason for performing A if and only if a consideration C that speaks in favour of X's doing A is such that, all else being equal, rationality requires that X do A, given C.

I shall regard an unqualified reference to what one ‘ought’ to do, or ‘should’ do in a given situation as equivalent to a reference to what is rationally required of one in the broad sense at issue here. I distinguish, therefore, what one ‘ought’ to do from what one is obligated to do. This accords with common assumptions. My unqualified ‘oughts’ and ‘shoulds’ should be understood accordingly in what follows.

In my terminology, it is common ground between myself and others who write on problems of political obligation that what is at issue in our discussions is a circumstance—call it ‘obligation’ or not—such that someone in that circumstance has sufficient reason, in the sense just characterized, to act in relevant ways. In other words, a positive solution to the problem at hand will demonstrate at least this: certain people in certain contexts have sufficient reason to support and comply with certain political institutions.

To say that one with an obligation has sufficient reason to act in a certain way is not to say that to have sufficient reason to act in some way is, in and of itself, to have an obligation. This is only a partial characterization of obligation. (p. 30 ) Again, to say that to have an obligation is to have sufficient reason to act in a certain way leaves open the possibility that there are radically different kinds of obligation, as I believe there are. ⁸ The same goes for the following further partial characterization of obligations.

Obligations are Independent of Personal Inclinations and Self‐Interest

A salient aspect of obligations is that they may run contrary to the particular personal feelings, urges, and (more mildly) inclinations of the person with the obligation—contrary to that person's inclinations, for short. This shows that obligations are independent of inclinations, in the sense that one can have an obligation to do a certain thing without at all being inclined to do it, and vice versa. ⁹

Obligations are in the same sense independent of the obligated person's self‐interest narrowly construed. That is, they do not necessarily correspond to what is good for a person in terms of his health, wealth, contentment, and so on. ¹⁰ When I refer to a person's ‘self‐interest’ in what follows this is how it should be understood.

The latter point connects with a well‐known contrast made by H. L. A. Hart between being obligated, on the one hand, and being obliged, on the other. ¹¹ One can properly say, ‘I was obliged to do it’ without implying that one had an obligation to do it. One's ‘being obliged’ might be a matter of self‐interest. It could be a matter of obligation, but it need not. In illustration of this, consider the case of residents of an imperator's territory where agents of the imperator are known to track down anyone who contravenes the imperator's directives and punish them severely. One might appropriately say that these residents are obliged to obey the imperator's directives. Whatever precisely this means, one can say it without implying that these residents have any obligation to obey. Many actual residents, including the members of liberal, democratic societies are obliged in this sense to follow an imperator's directives or an important subset of such directives. Some may be more obliged (in this sense) than others.

If one conforms to this distinction, it appears that there will not be specifically prudential obligations. Prudence may oblige one to do something, but will not (p. 31 ) in and of itself support an obligation to do it. That seems to accord with the way the term ‘obligation’ is used.

Sufficiency Versus Conclusiveness

The Practical Significance of Political Obligations Simply as Sufficient Reason

Irrespective of the precise relationship of obligations to inclinations and self‐interest, the simple fact that they provide one with sufficient reason for action already gives obligations a high degree of practical importance. One can see this if one considers a situation where one has a particular obligation to do something and must make an otherwise arbitrary choice. If obligations are (merely) sufficient reasons, rationality requires one to act in accordance with one's obligation.

This is hardly a trivial point. One way to make that clear is by reference to cases like the following. Suppose that, within the confines of one's country, one has to choose whether to drive on the left or on the right. One has no reason to choose one option rather than the other. In particular, one has no expectations, either way, about the likely action of other drivers. Suppose, further, that one takes oneself to have sufficient reason to do what the laws of one's country say one is to do. One then discovers that according to the laws one is to drive on the left. If one is disposed to act rationally, one will then drive on the left. If other drivers are in the same position, so will they. Much harm will thus be avoided. The situation of the drivers before their knowledge of the applicable law is what has come to be known as a coordination problem. ¹⁸ There is little doubt that life in human societies is rife with such problems. ¹⁹ Some are relatively trivial, some are of enormous importance. ²⁰

It is hardly trivial, then, to claim that membership in a political society in and of itself gives one sufficient reason to support its political institutions. Thus the partial account of obligation given in this section already shows clearly the practical significance of a positive answer to the membership problem. The full import of such an answer will of course depend on a better understanding of the kind of obligation at issue.

(p. 34 ) Recalcitrance to the Obligated Person's Will

A further point may be added to these preliminary observations. If we consider two standard contexts in which people talk of obligations, we find that the obligated person is generally not in a position to remove the obligation, prior to fulfilling it, with no more than a mental act of his own.

In the case of a promise, for instance, the promisor cannot simply decide that ‘the promise is off’ and hence free himself of his obligation to fulfil the promise. Rather, as it is often put, he must wait upon release by the promisee. He can of course break the promise if he so decides, but then he has defaulted on an existing obligation. In the case of an obligation not dependent on a promise, such as the obligation to save a drowning child if one can do so without danger to oneself, one might be able to remove the obligation by changing one's circumstances—one might perhaps alert a better swimmer to a drowning child's predicament—but generally speaking one cannot do this simply by a mental act alone.

If obligations generally are recalcitrant to the obligated person's will, then personal decisions do not give rise to obligations, since they can be done away with by a simple change of mind. ²¹ This accords with everyday usage: decisions may be in some sense ‘binding’, yet they are not generally said to give rise to obligations. ²²

Given that obligations are recalcitrant to one's will, it is easy to see why people may feel trapped by them. Decisions give one sufficient reason to act but one can change this by a simple change of mind. Obligations give one sufficient reason to act, and one cannot change this by a simple change of mind. Thus suppose one has to decide between two equally attractive alternatives, A and B. One realizes that one is obligated to choose alternative A. At this point rationality requires one to choose A. Some may find it comforting to know what must be done. Others may find it confining.

Summary of the Discussion So Far

Given the type of obligation with which I am concerned there are several points of contact between my concerns and those of other theorists of political obligation. Perhaps the most important is that if I am obligated to do something I am rationally required to do it, all else being equal. Equivalently, in my terms, one has sufficient reason to do it. That I have an obligation to do something, (p. 35 ) in a given situation, does not close the question of what reason requires me to do. There may be a species of obligation that does close the question. I am not assuming this to be a feature of obligations in general, something on which many theorists concur.

These points are intended to provide no more than a partial description of those obligations with which I am concerned in this book. As will emerge, some so‐called obligations do not all have the features listed. All that do are of considerable practical significance.

Hart's Proposal

The obligations of promises and agreements are generally considered paradigmatic. They have also proved recalcitrant to philosophical explanation. Theories as to how precisely promises obligate have proliferated, without any one being generally judged to be sufficient. ²³ The type of obligation in question has certain quite striking features that were emphasized by the distinguished philosopher of law H. L. A. Hart in a well‐known article. Hart proposed, in fact, that the term ‘obligation’ should be reserved for something with the features in question, contrary to the usage he observed among philosophers.

He wrote:

Most important are the points that (1) that obligations may be voluntarily incurred or created, (2) that they are owed to special persons (who have rights), (3) that they do not arise out of the character of the actions which are obligatory but out of the relationship of the parties. ²⁴

Hart discusses a number of cases that give rise to obligations according to his criteria. ²⁶ What he sees as ‘the most obvious case’ is that of promises. ²⁷ (p. 36 ) Presumably he would be happy to put agreements under this heading also. ²⁸ Hart distinguishes obligations from duties. Obligations as opposed to duties arise only on the basis of transactions or relationships between particular people. One properly speaks of an obligation to keep a promise; one speaks rather of a duty to rescue a stranger. One has an obligation to do what one agreed to do, but a duty to support just institutions in general.

The suggestion that we think of obligation proper along the lines Hart proposed has a long history. Thus, in his Treatise on Obligation, first published in 1791, the great French jurist Robert Joseph Pothier distinguished between two senses of the term ‘obligation’. In the third chapter he writes: ‘The term Obligation has two significations. In its most extensive signification … it is synonymous with the term duty, and comprehends imperfect as well as perfect obligations. We call imperfect obligations, the obligations for which we are accountable to God alone.’ ²⁹ An example he gives is that of obligations or duties of gratitude: ‘He, who has received a signal benefit, is bound to render his benefactor all the services of which he is capable … Yet his benefactor has no right of requiring those services from him … ’. ³⁰ He then turns to the second signification:

The term obligation, in a sense more proper and less extensive, comprehends only perfect obligations, which are called also personal engagements, giving him, with whom they are contracted, the right of requiring performance of them. ³¹

Hart's latter‐day proposal concerning how we should conceive of obligation proper has been taken up by a fair number of those who write on problems of political obligation. ³² Interestingly enough, having made reference to something like Hart's narrow sense of obligation at the outset, and allowed this to be obligation proper, several theorists of political obligation go on to say that they will not—after all—limit their focus to obligation in this sense. ³³ (p. 37 ) Thus they align themselves with authors who do not spend time elaborating a narrower notion. ³⁴

Hart himself proposes what can be taken as a solution to the membership problem in terms of obligations in the narrow sense he favours. His proposal, sometimes referred to as the ‘fair play argument’, has been elaborated in various ways by others. I discuss it later in this book. ³⁵

A relatively narrow use of the term ‘obligation’, where it is tied to something arising from transactions between persons such as those involved in promises and agreements, may indeed be close to the original meaning of the English term. By now, however, everyday usage allows for talk of ‘obligations’ in all of the above contexts. ³⁶ As Hart, in effect, suggests, this broad use may indeed obscure important differences. That is quite likely if there has been an expansion outwards from an original meaning that took account of a particular, relatively distinctive class of things. The outward expansion will then cover things of significantly different types, notwithstanding the presence of similarities to the original class.

In relation to the membership problem, and without any particular type of solution in mind, there are respectable reasons for not reining things in, at the outset, beyond a characterization of obligation such as I have given. This reflects central and relatively indisputable aspects of the prevailing broad use of the term ‘obligation’. It may, indeed, say the most that can be said without moving to a characterization of specific, differing types of obligation.

Caution is particularly in order with respect to certain aspects of Hart's and others' more restrictive conceptions. Hart would not hesitate to characterize his topic in the above quotation—obligation in his narrow sense—as moral obligation, where this is understood as other than legal obligation, and perhaps as having further, positive features as well. ³⁷ For reasons given earlier I prefer not to qualify the obligation at issue in the membership problem in this way.

Again, though Hart specifically says that obligations may be ‘voluntarily incurred or created’, some later writers suppose that obligations, in the proper narrow sense, must be voluntarily incurred. In other words, it is logically impossible to have an obligation—in the sense at issue—that you did not (p. 38 ) incur by some voluntary act. Now the term ‘voluntary’ can be variously interpreted. In a standard interpretation, however, a fully intentional action—as opposed to an automatic response like a knee‐jerk in reaction to a tap on the knee—may yet not be voluntary. Coercion or, indeed, circumstances of strong external pressure, are understood to make an act less than fully voluntary. ³⁸ It is best in the present context not to limit one's focus to obligations incurred voluntarily, in this strong sense. For one thing, though one might expect that obligations resulting from membership in a social group are in some way founded on a relationship between persons, it is less obvious that the relationship will be voluntary in a strong sense.

Leaving aside relationships that are entered unintentionally, one must allow that relationships generally may be entered in circumstances of strong pressure, or where the alternatives are so unattractive as to be unthinkable. This is the situation of many immigrants who come to their new country as refugees from another. They come to the new country because living in the other has become intolerable and the new country is for some reason the most feasible alternative option.

There is a related danger that initially restricting attention to a narrowly specified kind of obligation—particularly if it is qualified as ‘moral’ and required to be based on a strongly voluntary transaction—will prevent a proper understanding of the obligating character of agreements, promises, and whatever else may obligate in a similar way. Though consideration of promises in particular evidently played a large role in prompting the formulation of the restrictions Hart originally proposed, the way in which they obligate—either the ground of the obligation or its nature—may yet have been misunderstood.

Understanding agreements is important in relation to the membership problem since perhaps the most famous solution to it alludes to them. I devote some time to this solution—actual contract theory—in this book. ³⁹ I argue that one central objection to it derives from a misunderstanding of agreements and the type of obligation they involve.

Obligation and Owing

In spite of the reasonableness of concerns about focusing initially on obligations as specified by Hart, or in elaborations of his specification, I shall suggest in (p. 39 ) the next chapter that the most satisfying solution to the membership problem will show that members of political societies have obligations in a sense close to Hart's. ⁴⁰ In this section I specify that sense.

I shall abjure Hart's willingness to qualify the obligations in question as moral, and I shall not take on his assumption concerning the grounds of these obligations in interpersonal relationships or transactions. I shall for now leave that matter open. Finally, I shall assume neither that these obligations can be voluntarily created nor that they cannot. Evidently, I shall not assume that these obligations can only be incurred voluntarily—in any sense of the term.

Taking from Hart's account any qualification about morality, any specification of grounds, and any reference to the possibility of voluntary creation, something extremely significant remains. This comes in Hart's second clause: obligations of the type in question are ‘owed to special persons (who have rights)’. Perhaps ‘special persons’ would better be replaced by ‘specific persons’ in the sense that if an obligation is owed, it has to be owed to someone. There need be nothing ‘special’ about these persons in the sense that they have any particular personal qualities. ‘Specific persons’, however, should not be understood as ruling out all persons as those to whom the obligation is owed. What is key here is the reference to ‘owing’, owing to one or more persons.

One might say, as Hart does, that an obligation is owed. Or one might say, perhaps more perspicuously, that one has an obligation of the kind in question if and only if one owes someone a particular action. Performing that action is the fulfilment of the obligation. The question of how one might come to owe someone an action will be set aside for now. ⁴¹

The person to whom one owes the action is said to have a right to it. Thus rights are ‘correlated’, in a strong sense, with obligations of this type: for X to owe Y action A is for Y to have a right against X to X's performance of A, and vice versa. ⁴²

There is no need to worry about whether we are talking here about obligations proper, as opposed to a central class of obligations—perhaps the original class in terms of a gradual widening of the extension of the term. What is important is not the proper use of a word, but rather the nature of, in a broad sense, a ‘thing’. Namely, an obligation of the type in question.

(p. 40 ) Some authors simply say that an obligation of the type with which they are concerned is an obligation ‘to’ or ‘toward’ a particular person, the correlative right being a right ‘against’ the person with the obligation. ⁴³ This way of putting things is even thinner than the pared‐down version of Hart's account I have proposed so far. It seems to me that by thinning things down this far we have lost something important. For one must at some point ask: what is it to have an ‘obligation to’ another person? What is it to have a right ‘against’ a person? These phrases become open to a variety of interpretations (which they have, indeed, received). Things are pinned down very little by the phrases themselves. I shall therefore not strip any more from Hart's characterization of obligations. I shall maintain within it the notion of ‘owing’. In this way it retains some substance.

In the literature, obligations simply characterized as obligations ‘to’ another person, correlative with rights of that person against the person with the obligation, have become known as ‘directed’ or ‘relational’ obligations. Hoping not to confuse anyone, I shall in what follows use the phrase ‘directed obligation’, and related phrases such as ‘obligation to’, to refer to obligations directed in the more substantial sense I have specified. That is, I shall understand someone to have a directed obligation when and only when he owes another person an act of his own. His obligation is then an obligation to, or towards, that person, who has a correlative right against him to the act that is owed.

One should not think all that can be said about directed obligations has now been said. The notion of a directed obligation may at this point rightly seem puzzling. What is it for someone to owe an action to another person? On what basis can one be properly said to be in this position?

Assuming this idea can be made good, it is reasonable to suppose that directed obligations are obligations according to the partial characterization presented in this chapter. That is, it is reasonable to suppose that one who can plausibly be said to owe another an action will have an obligation according to that characterization. He will have sufficient reason for performing that action, sufficient reason that is independent of his own inclinations or narrow self‐interest and that cannot be eradicated by his own fiat.

It seems that—failing special circumstances—the fact that one owes someone an action must trump one's contrary inclinations and considerations of self‐interest from the point of view of what rationality requires one to do. If I owe you this action, how can I appropriately argue that my own inclinations or considerations of self‐interest, as such, permit me not to give it (p. 41 ) to you? As noted earlier, it is often assumed to be true of genuine obligations in general that they trump inclinations and considerations of self‐interest. It seems, then, that directed obligations fit well with a variety of common understandings about obligation.