Promising and Obligation

Philosophical Perspectives, 23, Ethics, 2009

PROMISING AND OBLIGATION1

Thomas PinkKing’s College London

1. Promising –the communication of an intention to obligate oneself?

Promises to do things are made by promisors to promisees. And theytypically leave the promisors morally obliged to act as promised — an obligationthey have to the promisee, who thereby acquires a moral right to what waspromised to them.

How is this possible? For it is not a trivial fact that promisors owe it morallyto promisees to keep promises. Whether a promisor is morally bound to dowhat he promised — whether breach of the promise would be blameworthywrongdoing — can in a particular case be a substantial moral question, theresolution depending much on debate about factors such as disproportionate orunexpected burdens falling on the promisor, the possibility that the promise wasinduced by coercion or deceit, and so forth. On the other hand it does seem afundamental part of our understanding of promising that absent such factorsit can and normally does obligate the promisor. People who make promiseswould generally acknowledge that they are thereby imposing a moral obligationon themselves to act as promised, such that subsequently breaking the promisewould normally be wrong.

Does the possession, or at least the expression by the promisor of anintention to obligate himself play any role in promising and in the generationof promissory obligation? Many modern authors would suppose so: by itsvery nature, they suppose, the act of promising involves the communication orexpression of an intention to obligate oneself — a communication or expressionthat, if various conditions are met, can then obligate the promisor. Here, forexample, is John Finnis’s account of a promise:

First, what is a promise or undertaking? Being a human practice, engaged inand maintained for diverse practical purposes, promising has its central cases (itsfocal meaning) and its secondary or borderline cases. Centrally, then, a promiseis constituted if and only if (i) A communicates to B his intention to undertake,

390 / Thomas Pink

by that very act of communication (in conjunction with B’s acceptance of it),an obligation to perform a certain action (or to see to it that certain actions areperformed), and (ii) B accepts this undertaking in the interests of himself, or ofA or of some third party C.2

And here is another account by David Owens along similar lines:

But at least this much is true: I am not promising to take you home unless,in saying what I say, I mean to communicate an intention to undertake anobligation to take you home.3

Joseph Raz speaks of promissory obligation as a ‘voluntary obligation’ in thatit is ‘undertaken by acts performed in order to undertake an obligation’4 andhas defended what he calls an ‘obligation conception’ of promising, accordingto which:

To promise is, on this conception, to communicate an intention to undertake bythe very act of communication an obligation to perform a certain action.5

So, on this widely held view, a promise is an action by which someone(the promisor) deliberately communicates or expresses an intention to obligatethemselves to another person (the promisee); and can, subject to the consent oracceptance of the other, by that action bring it about that they are so obligated.

There are two parts to this doctrine. First, there is a claim about the act ofpromising itself: that promising serves to express or communicate an intention toobligate oneself. Now not all promisors need actually possess such an intention.But those promisors that do not are still communicating or suggesting otherwise,and assuming they understand what promising involves, are, by that very fact,being insincere or deceitful. Perhaps such promisors really intend not to actas promised, and so intend or at least prefer if possible to escape any moralobligation, and by promising strive to mislead the promisor as to their trueintentions. Whereas a sincere promisor will always really hold the intention thathis promise expresses; he really will mean to obligate himself. Second, thereis a claim about promissory obligation itself: that its source is to be found inthe promisor’s communication, whether honest or deceitful, of an intention toobligate himself. Promising therefore involves a noteworthy, even remarkable,normative power — a power morally to obligate ourselves voluntarily or at will,just through expressing an intention to become so obligated.

My purpose in this paper is to establish two conclusions. The first is about theintentions of promisors, and, connectedly, about how promisors are understoodby others. It is perfectly true that many promisors may well intend to obligatethemselves to act as promised, and may well frequently be understood both bypromisees and by third parties to possess this intention. But, I shall argue, a

Promising and Obligation / 391

perfectly sincere and competent promisor may still make his promise withoutholding any such intention. A perfectly sincere promisor’s motivation — his goalin making the promise — need have nothing to do with actually incurring suchan obligation, but may be quite different, and be understood to be such by thepromisee. Each party may recognize or be ready to acknowledge that throughthe promise an obligation has been created. Promises generally create obligationsto act as promised; and to have any understanding of what promising involvesis to know this, or at least to know that people generally view promises in theseterms. But the creation of the obligation need not have been the purpose of thepromisor or something desired or hoped for by the promisee, and neither partyto the promise need suppose otherwise. In such a case there need be nothingincompetent, insincere or deceitful about the promisor or his promise.

Secondly, and most importantly, even if a promisor does hold an intentionto obligate himself to act as promised, the expression of such an intention playsno role in the generation of promissory obligations, which originate in quiteanother way. The use of terms such as ‘I promise’ by the promisor may serve toacknowledge the existence of a consequent moral obligation to act as promised,whether or not this obligation was specifically intended. But the source of theobligation does not lie in any such acknowledgement of an obligation, still lessin the expression of an actual intention or purpose that the obligation arise. Thesource of the moral obligation lies instead at a deeper level, in an activity thatdoes not require the use of terms such as ‘I promise’, and that certainly hasnothing to do with any expression of an intention to obligate oneself.

Promising in everyday life contrasts importantly with the use of legalformulae that may have as their sole function the creation of a new legal relation,such as a transfer of rights or a legal obligation. The sole function of a formulasuch as ‘I hereby bequeath’ or ‘I formally contract’ may be simply to effect a legaltransfer, or to impose on oneself a legal obligation; without the use of these setforms, the transfer or the imposition of the legal obligation may not be possible.In which case someone who uses these expressions, and does so with a competentunderstanding of their function, may be said to be expressing an intention toeffect the transfer or to impose on themselves the legal obligation. But theexpression ‘I promise to’, at least outside a specifically legal or institutionalcontext, is not the same. The apparently sincere utterance of these words may,in general, impose an obligation on the speaker. But the production of thatobligation need not exhaust the function of that expression. Indeed promisingmay have other functions, and other things may be done by a promisor thansimply undertaking an obligation — things that might be done even without theuse of expressions such as ‘I promise to’; and it may be these other things donethat help explain the obligations which promising generates. In which case, evenif all parties may foresee the obligation that a promise produces, their goals andintentions may not lie in the production of that obligation, but may rather haveto do with those other things done. Or so I shall now argue.

392 / Thomas Pink

2. What promise-makers intend

There is something strange about the standard account of the intentionexpressed by a promise. It is left entirely to do with the moral status of thepromisor’s future actions. In promising the promisor is supposed, if sincere, tobe concerned with ensuring that he is obliged to do something. For the presenceof such a concern is supposed to be what his promise conveys. Now, as we havealready agreed, people certainly treat promise makers as frequently under a moralobligation to keep their promises. At least they view the deliberate breaking ofa promise as generally wrong, unless the promise was to do something wrong inthe first place, or unless the promise is broken for some fairly good reason. Butthis does not mean that promise makers need have formed an actual intentionto obligate themselves; that it was changing the moral status of their futureactions that was their goal, or that the lack of such an intention involves anyinsincerity.

Suppose, for example, that I am a doctor faced by a worried patient whowants me to be present next week at a crucial operation. The worried patientbegs: ‘Promise me doctor, that you will be there.’ And I reassure her: ‘I promise’.My goal in making this promise is, surely, to give my patient something she hasrequested and badly wants. But precisely for this reason my goal need not be tochange what I am morally obligated to do. For my promisee need not be at allinterested in what duties I am to be under.

Were the matter raised, the patient might well agree that in promising I amobligating myself to her. But I doubt that it is that obligation in her favour thatshe need particularly want to receive. What the deeply worried patient clearlywants is my presence at the operation — and assurance of this presence inadvance. And in making my promise my intention will be to offer the patientprecisely what she wants — not an obligation on my part to be present, but myactual presence, with prior assurance that I will in fact deliver on the offer. Whyshould obligations matter to either of us provided my promise constitutes theoffer and reassurance that the patient wants, and provided I do eventually deliveron my promise — by actually being present at the operation?6

It might be thought that the existence of the obligation must matter tothe patient. For what else will reliably motivate me to act as promised but theexistence of a moral obligation to do so? Without that motivation, how can thepatient be assured that I will be there as promised?

Certainly some philosophers have thought that only a sense of duty — abelief that one is under a moral obligation to deliver — will motivate promise-makers to keep their promises. Hume made this very claim as part of his famousargument for the artificiality or conventional origin of the obligation to keeppromises:

Now ‘tis evident we have no motive leading us to the performance of promises,distinct from a sense of duty. If we thought, that promises had no moral


obligation, we never shou’d feel any inclination to keep them. This is not thecase with the natural virtues.7

But it is simply not obvious that Hume’s claim is true. Why cannot a promisemaker be motivated simply by the thought that they have promised — thatthey have given their word? The content of this thought is not obviouslyabout obligations at all, and its motivating force does not strictly depend onthoughts about obligation. Which is why someone can intelligibly be motivatedto keep a promise even when sceptical about whether they are really boundby any obligation to do so. Suppose, because of a quite unexpected change incircumstances and the resultantly and suddenly huge cost to him of keepinghis word, the doctor wonders whether it might indeed have become strictlypermissible for him not to keep his promise to his patient. He might surely stillbe motivated to keep his promise by the thought that nevertheless, permissibleor not, he had given his word. Such a motivation might appear quixoticto some. But the motivation is not unintelligible, and it does not involvethoughts about moral obligation. I shall return to say something of what itmight involve.

Notice in particular that even if as the doctor I am not at all concerned withthe moral status of my future actions — I am concerned simply to provide honestreassurance to the patient that I will indeed be there — I am plainly not being inany way deceitful or insincere, nor am I being incompetent in my use of language.My utterance of the promise need not be in any way misleading to the patient,for she need not understand me as having the production of a moral obligationas my goal. My sincerity as a promisor she will understand as lying not in thegenuineness of my intention to obligate myself to her, but in the honesty of myoffer to her of my future presence at the operation. And in failing to attributeto me any actual intention or purpose of obligating myself, just as the patientis not failing to understand me — she understands me well enough as seekinghonestly to reassure her — so she is not being deficient in her understanding ofwhat promises are, or what the sincere utterance of a promise implies about apromisor’s state of mind. This doctor-patient exchange is an entirely normal caseof promising. Despite the very possible absence in both parties of any intentionof or concern with obligation, there seems nothing defective or borderlineabout it.

I suspect that with many promises in everyday life, such as the promisemade to the worried patient by the doctor, thoughts about obligation often arisefirst only when or if the promise gets broken; and these thoughts will arise,in particular, in the mind of the promisee who feels wronged by the promisor.But even here the promisee will not simply feel that they have been deprived ofsomething they had a right to. They may also feel hurt and diminished. And thefeeling of hurt that often meets broken promises is something we need to say moreabout. It is certainly not explained simply by referring to some obligation that thepromisor has breached. For one can easily feel that one has been denied a right

394 / Thomas Pink

without feeling especially hurt or personally diminished by that denial. Whereasthere clearly is something rather personal about someone’s breaking their wordto you. And a theory of promising needs to explain what this personal hurt mightinvolve.

What must a promisor really intend? Clearly, they must intend to convey anoffer to the promisee — an offer of the action promised with an assurance orpledge of delivery on that offer that involves but goes beyond the mere expressionof an intention to deliver. But though promises may generally obligate, as we haveseen it need not be the promisor’s actual intention that they should. Even if theobligation’s generation may be foreseen, the purpose of making a promise neednot be to generate that obligation, or to offer the obligation to the promisee.What the promisee is offered first and foremost is not the moral obligation,which in many cases they are unlikely to want, but something quite differentand of much greater interest to most promisees — performance of the actionpromised.

Besides promises to do things or promises to — promises as to performanceor performance promises — there are promises as to fact — factive promisesor promises that.8 ‘I promise you it really happened, right before my veryeyes.’ Now there seems something about promising that which leaves it verylike promising to. But the thing in common is not obviously the expression ofa corresponding intention to obligate oneself. That is, it does not seem thatas promisors to (supposedly) express an intention to obligate themselves to, sopromisors that express an intention to obligate themselves that. For there are noobvious obligations that. Obligations are always obligations to — obligations toperform, to do things or refrain from their doing.

It is tempting to cast around for some obligation to perform that all makersof factive promises distinctively incur through making their promise, so thattheir making of a factive promise that can express their intention to incur thisobligation. But what might this obligation be? Is it an obligation to speak thetruth, or at least to take care that truth is spoken? But in ordinary assertion suchan obligation is plausibly incurred anyway, whether or not a promise as to fact isactually made. Perhaps then it is some further, more demanding obligation. Butthe candidates are not obvious. Certainly there is no single candidate as obviousas the obligation on the maker of a performance promise to perform the actionpromised.

Of course it might be that promising that has nothing much to do withpromising to; it might be that promising that is a peripheral phenomenon, thatsheds no light on promises as to performance. But I shall suggest that that is notso. We shall find that promises that and promises to work in the same way —when considered at a deeper level that is not itself to do with obligations. Andit is only when considered at that deeper level that the genuine connexions thatthere are between performance promises in particular and obligation will reallybecome intelligible.


3. Hume and Locke on promising

Locke thought that we had a natural or pre-social power to make promisesor enter into agreements:

The Promises and Bargains for Truck, etc. between the two Men in the DesertIsland, mentioned by Garcilasso De La Vega, in his History of Peru, or betweena Swiss and an Indian, in the Woods of America, are binding on them, thoughthey are perfectly in a State of Nature, in reference to one another. For Truthand keeping of Faith belongs to Men, as Men, and not as Members of Society.9

Whereas Hume denied that we had any such pre-social power. For Hume, thepossibility of promising must depend on the existence of a social convention orpractice of promising:

I say, first, that a promise is not intelligible naturally, nor antecedent to humanconventions; and that a man, unacquainted with society, could never enterinto any engagements with another, even tho’ they could perceive each other’sthoughts by intuition.10

What is at issue? Plausibly the use of the words ‘I promise’ in making agreementsinvolves and requires what Hume calls ‘human convention’, just as does theuse of any linguistic term. If promising is seen as a practice that, by its verynature, is dependent on the use or availability of certain linguistic expressions,then it might seem as though Hume’s point follows immediately. The possibilityof promising must indeed depend on the existence of a convention or practice ofpromising.

But Hume’s position is not so trivial. For suppose the possibility of promisingdid depend on the availability of a certain kind of linguistic expression.11 It wouldnot follow that the obligations and related rights that fidelity to promises respectswere conventional in basis. For it might still be that these obligations and rightsprecede the practice of making and keeping promises and that they shape anddetermine the form which that practice takes. Whereas Hume’s position was infact the opposite: that these obligations and rights are products of and dependenton the practice of making and keeping promises, and that it is the form whichthat practice has taken which has determined the nature of these obligations andrights.

Hume sees promising and the virtue or obligation of fidelity to promisesas part of what he calls artificial virtue — morality in a form that depends onand derives from a variety of conventions, conventions that have been developednot to assert and protect pre-existing moral standards, but in order to attainquite different ends. Promising is in the same category as justice understood asrespect for property, as allegiance to the state, and as chastity or marital fidelity.

396 / Thomas Pink

Conventions that determine what counts as the making and keeping of promises,that define what constitutes ownership and the constitution of the state and ofpolitical authority — all these have arisen the better to enable individuals tofurther their own self-interest. But in so far as respect for these conventions,just by furthering the self-interest of the generality of individuals, equivalentlyserves the common good, so we come morally to approve of such respect, andto regard it as morally virtuous and obligatory. So conventions that developedto further individuals’ self-interest come to define new forms of moral right andobligation — forms of right and obligation that would not have existed at allwithout those conventions.

Now this sort of claim, at least regarding some obligations and rights, wasnot original to Hume. It formed part of a project that ran through the naturallaw ethics that was widely taught in Protestant universities in Hume’s youth. Forit does seem intuitive that many parts of morality and the rights and obligationsthat compose them might be brought into existence by the development ofvarious forms of social interaction.

Perhaps the point is most intuitive in respect of property. Whether weconceive of humanity as living originally in some earthly paradise or else insome condition of primitive and uncultivated want, it seems unobvious that therewere always property rights and obligations to respect them. Perhaps insteadproperty rights only appeared through a process of human social development.As for property, so for other things too, such as the existence of states andforms of political authority with concomitant rights to command or legislateand obligations to conform to legislation.

So central in particular to the modern or Protestant natural law traditionwas the project of explaining some parts of morality as not natural or originalto human nature but as instead artificial, in Hume’s phrase, or as adventitious,to use the term of his predecessor, the 17th century natural lawyer SamuelPufendorf. These developmental accounts of human morality were linked togeneral theories of human social, economic and political development. For thedevelopment of artificial or adventitious forms of morality went hand in handwith the development of the social institutions and structures that those new,humanly generated forms of morality governed. In the Protestant natural lawtradition we see moral theory linked to highly inventive exercises in historyand social science. And though Hume was not himself a natural lawyer — hespecifically denied what natural law affirms: that moral obligation consists in theexistence of a distinctively practical form of reason in legal or demanding formthat governs us just as possessors of a rational, human nature12 — he otherwiseperpetuates and further develops much of the social scientific and historicalresearch programme of his predecessors.

By what mechanism of social interaction was artificial virtue brought intoexistence? Pufendorf had taken the relevant mechanism to be that of pacta — ofpacts or agreements.13 So, for example, property and our obligation to respect itwas to be explained in terms of agreements to institute and abide by certain rules


of property — agreements that it served each individual’s interests to enter into.But if the adventitious or artificial is the creation of agreement, and agreementsare exchanges of promises, that might make it appear that the obligation to keeppromises must at least be natural. On the other hand, for Pufendorf pacta oragreements need not involve the actual exchange of promises. They can includesomething more like conventions as well as formal promises. For example, it issupposed to be pacta which generate even linguistic meanings. But clearly thebasis of linguistic meanings plausibly lies not in promising, but in the gradualdevelopment of shared habits or conventions of linguistic use.

What Hume clearly perceives as Pufendorf does not, is the importantdifference between a mere convention or mutual adaptation of behaviour andan actual exchange of promises. So Hume replaces Pufendorf’s appeals toundifferentiated pacta. The relevant artificial virtue-creating mechanism of socialinteraction for Hume is definitely convention — a kind of mutual adaptation ofbehaviour in the pursuit of individual self-interest that need never involve anyexchange of promises. That done, Hume is then in a position to explain promisingand the obligation to keep promises as itself, like language, political allegianceor property, a product of convention.

But otherwise Hume follows the same strategy as Pufendorf. For boththinkers, artificial or adventitious morality and the social institutions that comewith it are all creations of human self interest in enlightened form — self interestthat motivates us to make pacts and agreements and to establish and followconventions. Both thinkers conceive of promising and agreement, then, as afeature not of generous giving nor of friendship, but of a common pursuit ofindividual self-interest. Thus Pufendorf:

From what has been said, it is understood how works of humanity or of lovediffer from those which are required from a right properly understood, and are,therefore, directed by actual justice. The former are not caused by nature ofagreements, express or implicit, but are laid upon all men by nature herself onmere grounds of obligation. But whatever things I owed a man from agreementsor covenants, I owe because he has secured a new right against me by my ownconsent. Furthermore, whatever I have done with another man in agreements, Ihave done not so much for his advantage as for my own, while in the duties ofhumanity the very opposite is the case.14

There is much of importance in common between Pufendorf and Hume. Bothopenly differ from Hobbes in seeing an important category of our obligations toothers as arising naturally, independently of any agreement or convention. Bothbelieve in a natural morality of obligations of what Pufendorf terms humanity orlove and Hume terms benevolence. Independently of any social interaction, thiswholly natural part of morality commits us just as human beings to providingvarious kinds of basic help to those in need, and to refraining from variouskinds of assault and harm. It is the establishment of conventions and pacta or

398 / Thomas Pink

agreements which thereafter extends this initially natural morality into new andhumanly created, adventitious forms — and which does so, at least immediately,in the service of self interest rather than benevolence.

And this brings us back to the original dispute between Locke and Hume.Why suppose with Hume that the rights that fidelity to promises protects areproducts of convention, rather than preexisting rights that help shape the formthat our practice of making and keeping promises takes? For many will findsomething deeply intuitive about Locke’s claim that

Truth and keeping of Faith belongs to Men, as Men, and not as Members ofSociety.

So we must examine Hume’s conventionalist theory of promising and promissoryobligation in more detail.

4. The promising convention

The obligation to keep promises is, for Hume, the product of a particularconvention, a promising convention. What is the function of this convention, andhow has it arisen? Hume’s claim is that the convention arises to meet a particularneed — that of enabling and facilitating self-interested exchange.

How can self-interested agents provide others with goods and services withthe assurance of receiving a definite return in exchange? The problem is supposedby Hume not to arise with simultaneous exchanges, since if one party ceasesreciprocating the other can just as immediately stop performing too. But what ifa return can only be made later?

Now as it frequently happens, that these mutual performances cannot be finish’dat the same instant, ‘tis necessary, that one party be contented to remainin uncertainty, and depend on the gratitude of the other for a return ofkindness.15

But why suppose that having received what they wanted, the equally self-interested counterparty will ever make that return? Gratitude won’t reliablymotivate the predominantly self-interested. So, to take his famous example,Hume’s two farmers each have fields in need of harvesting. Neither farmer canharvest his own fields alone; each farmer needs the other’s help to get his harvestin. But neither farmer will help first as neither sees any likelihood that their helpwill be returned. For why should the counterparty see it as in his interest, oncehis fields are harvested, to return the help?

The seasons change; and both of us lose our harvests for want of mutualconfidence and security.16


Self interest may of course get canny, or at least optimistic. We only needthe hopeful thought to spread that favours done without hope of immediateadvantage may yet eventually be returned and bring long term advantage totheir provider. Through the prevalence of such thinking we may enter a cycleof mutually advantageous self-interested exchange. But a problem neverthelessremains. How to mark out interactions where a return is expected, and distinguishthem from ‘the more generous and noble intercourse of friendship and goodoffices’ where return is not in question? And where a definite and specific kindof return is expected, how most securely to motivate and tie the other party intomaking not only some return at all but, moreover, that specific return?

Hence the development of a promising convention. The language of promis-ing marks out the commerce of self-interested exchange. And according to thisconvention whenever someone uses a certain distinctive mode of expressing anintention to do A, such as by saying ‘I promise’, they are thereafter to do A, atthe risk of not being trusted by potential cooperators again if they do not.

In order, therefore, to distinguish those two different sorts of commerce, theinterested and the disinterested, there is a certain form of words invented forthe former, by which we bind ourselves to the performance of any action. Thisform of words constitutes what we call a promise, which is the sanction of theinterested commerce of mankind. When a man says he promises any thing, hein effect expresses a resolution of performing it; and along with that, by makinguse of this form of words, subjects himself to the penalty of never being trustedagain in case of failure.17

By using the expression ‘I promise’, a counterparty becomes motivated by self-interest, thanks to the threatened penalty, to reciprocate in the way promised. Hethen becomes a really credible exchange partner, and both sides can then embarkon a mutually advantageous pattern of exchange:

All of them, by concert, enter into a scheme of actions, calculated for commonbenefit, and agree to be true to their word; nor is there any thing requisite toform this concert or convention, but that every one have a sense of interest inthe faithful fulfilling of engagements, and express that sense to other membersof the society. This immediately causes that interest to operate upon them; andinterest is the first obligation to the performance of promises.18

As benevolent agents capable of concern for others and the common good, wesee that everyone benefits from the existence of the promising convention. Thepromising convention does what needs to be done: to mark out actions expressivenot of gratuitous generosity and friendship, but performed self-interestedlyout of expectation of exchange; and to ensure that that exchange is in factmade. As a result the hoped for non-simultaneous exchanges are facilitated.And so, with the common good in mind, we come to view the keeping of

400 / Thomas Pink

promises as morally obligatory, and come to see fidelity to promises as a moralvirtue:

Afterwards a sentiment of morals concurs with interest, and becomes a newobligation upon mankind. This sentiment of morality, in the performance ofpromises, arises from the same principles as that in the abstinence from theproperty of others.19

5. Criticism of Hume

I have presented the issue of the artificiality of promising as being aboutthe obligations and rights that fidelity to promises conforms to and respects.Are these, as Hume supposes, the product of a convention or practice — apractice that originally served purposes other than that of respecting such rightsand obligations? Or are the obligations and rights prior, and is it these thathave in large part shaped the practice of promising? Obviously one element inpromising — the language by which the concept of promising and the intentionto promise is expressed — is conventional. But the involvement of convention inour capacity to oblige ourselves through promises might go no wider than that.

There is one obvious line of objection to Hume. Take the exchange ofservices, such as those which Hume’s self-interested farmers provide by eachhelping with the harvest of the other. Is the obligation of one to make thepromised return to the labour provided by the other really unintelligible apartfrom an existing practice and convention of making such a return? Hume isadamant that it is so unintelligible:

. . . a promise is not intelligible naturally, nor antecedent to human conventions;and that a man, unacquainted with society, could never enter into any engage-ments with another, even tho’ they could perceive each other’s thoughts byintuition.20

But is there not something immediately intelligible or natural about such anobligation, prior to the establishment of any practice, just as Locke supposed?

To make out the naturalness of the obligation, Hume supposed we wouldhave to appeal to a peculiar supposed natural power that each possesses to imposeobligations on himself at will. And he denied that there was such a power. Butin fact we have no need to posit any such power. What instead we need is a kindof natural right possessed by each — to determine under what conditions he willprovide goods and services to others, at least in cases where he is not obligedto provide those goods and services in any case. One function of promising issimply to acknowledge and respect this right. I promise to you that I will helpyou with your harvest tomorrow if you help me with my harvest today. In sopromising I recognize your right to make your help today conditional on thehelp being returned by me later.


For Hume, the idea of an obligation to reciprocate is only supposed to arisewith the practice of promising. And the practice of promising is only required incases where a simultaneous exchange is not possible and one has simply to trustthe other to perform. A practice then arises to motivate and guarantee the hopedfor reciprocation. And from this practice the idea of an obligation to reciprocateas promised then arises. Prior to the practice there is no obligation, and no rightto expect reciprocation or unjust denial of that right.

But, contrary to this, the idea of an obligation to make a return can surelybe invoked immediately, in relation to any exchange, simultaneous or otherwise.

Suppose Hume’s farmers find themselves in sight of each other, each standingby chance on the other’s property, but separated by a sudden river flood thatneither can ford, and that will diminish only after the sowing season is over.Each is in a position to sow the other’s crops, and there is no one else to do thejob. Even prior to the establishment of any practice of promising, each can makeit clear, by shouting out what he is willing to do, or through otherwise signalingor implying this across the flood, that he will sow the other’s fields, but onlyprovided that the other reciprocates. And so, keeping their eye on the other, eachsows, though only provided he sees the other still involved in sowing too. Witheach in full sight of the other, this looks like a case of simultaneous exchange.In which case it seems that as yet there is no need of some special practice to tieeither farmer into reciprocating. And so on Hume’s theory there is no room yetfor any obligation or sense of such.

But supposing one of the farmers discovers later that despite all appearance,somehow the other (a master illusionist) managed only to feign the effort andso avoided actually sowing any seed at all. Is it not natural to suppose a senseof wrong in the discoverer — that he has not only been deceived but defrauded?It is as with any other exchange. Each can see in themselves and others a rightto determine under what conditions he will do things for the other. Each canrecognize the assertion of such a right, and each can recognize its violation.

To fail to respect that right through fraudulent deception is in fact as muchan invasion of another’s liberty as would one farmer’s actually physically forcingthe other to sow his fields, supposing the farmer could do that. Even supposing nofurther injury were done to the farmer so compelled, he would still view his beingso coerced as an intrusion on his liberty. And would we need a convention to seethe application of such coercion as wrong? Surely if there is anything to Hume’snatural and preconventional virtue of benevolence, forcing the other farmer tosow his fields unrewarded would be contrary to the natural duty towards othersthat both Pufendorf and Hume defend against Hobbes. Why then the need for aconvention to make sense of the wrongfulness of producing the same result by adeception?

Plausibly, natural morality requires us to make certain goods and servicesavailable to others in sufficient need — and to do so whether or not we everagreed to, or whether or not any convention has arisen of so helping others. Butbeyond these duties and granted their extent, it is equally natural to suppose that

402 / Thomas Pink

natural morality affords us a certain liberty. This liberty involves an obligationon others under normal conditions not to coerce us or to force us into actionagainst our will. And it involves our possessing a natural right, within limits, todetermine for ourselves what further goods and services we provide for others,and under what conditions. Now a failure through deceit to deliver a clearlyinsisted on and acknowledged reciprocation to some service given by us is asimmediately recognizable a violation of this right as is the use of outright forceto extract that service from us without return. A function of the practice ofmaking and keeping promises is not to create that right, but to recognize andrespect it.21

So one criticism of Hume’s theory of promising is this. Hume, like Pufendorf,is concerned to occupy a position in moral theory that is not Hobbesian — thatdoes not leave all our obligations towards others the products of agreement orconvention. In particular, there are pre-conventional standards of humanity andbenevolence towards others that preclude various forms of coercive interferencein their lives. But Hume wants at the same time to maintain that the morality offidelity to promises in the context of exchange is wholly artificial, and generatedby practices originally directed simply to the furtherance of self-interest. Yetsurely, it is rights under natural morality that the practice of fidelity to promisesprotects. For if people had no natural right not to be defrauded of their goodsand services, why should they possess any natural right not to be coerced intoproviding them just for the whim and convenience of another? If even whenunaccompanied by any further injury, such coercion is wrong in itself, andnaturally wrong, then so must be the fraud. The coercion and the fraud areboth wrong in the same way — as an intrusion on the agent’s natural liberty todetermine what he does with his own.

Hume takes our existing practice of making and keeping promises to havearisen in its present form to facilitate self-interested exchanges; and he claimsthat this present form has then created and shaped moral obligation in respect ofpromises. But of course if promising had all along been about the facilitation ofself-interested exchange, convention and practice in relation to promising mightwell have taken a different form. The convention might not just have been tokeep promises at the risk of not being trusted again should they be broken. Theconvention might also have been that promise breakers definitely should not betrusted again, at least for some roughly specified interval. Such a conventionwould certainly be a highly effective facilitator of reciprocal exchange. For therewould be the less incentive for any promise breaking, and so people could enterinto and perform first under agreements with the greater confidence of a return.And on the Humean view if such a convention did come into existence, andeveryone really did benefit from the strict adherence to it, it would be adherenceto this convention that would come to count as morally obligatory. And then ofcourse there would be two moral obligations in relation to promises — on thepromise maker to deliver, and on third parties and the promisee never to let thepromise maker get away with failing to play his part.


But this hypothetical convention, no matter how useful a facilitator ofmutually advantageous exchange it might be, has nothing to do with promisingas we ordinarily understand it. Indeed it seems absurdly morally oppressive onpromisees and third parties. On our understanding, in relation to promises onlytheir maker can actually be obligated — to doing what he promised. Othersare at perfect liberty to trust him again should they so choose. But that isbecause what we are dealing with is not a convention serving simply to facilitateexchanges among the self-interested. We are dealing with a practice that assertsand respects preexisting rights. And only the promise breaker has actually donewrong by denying anyone their right — the right to determine the conditionsunder which they provide services and goods to others. Indeed the true basis ofthe promisee’s right to the return promised him — his liberty to determine forhimself what he does for others and on what conditions — precisely involves, byits very nature, a right not after all to insist on that promised return.

In fact the fundamental implausibility of Hume’s theory becomes evenmore evident when we examine social reality, and note how very differentfrom each other are the actual practices involved in promising and property —two parts of morality and social practice that Hume himself repeatedly likensand seeks theoretically to assimilate.22 The institution of property does indeedseem to involve a carefully maintained set of conventions of just the kind thatHume’s theory would require. People do generally and reliably adhere to rulesdetermining ownership of objects; and are reliably penalised in various ways ifthey are ever caught out in breach of those rules. But there is no analogous andequally reliably maintained convention of doing what one has promised with areal risk of being penalised should one ever break one’s word. You promise methat in return for getting into my edited collection, you will supply the paper bymonth end. For no sufficient reason, you break your word — but I exercise myright to forgive the breach and offer an extension to the deadline, which againyou promise to meet. And so it goes. You repeatedly fail to act as promised andI repeatedly fail to penalise you for your failure. In my experience this sort ofthing can go on for several years. There is nothing recognisable in this case asobservance of Hume’s promising convention. The Humean convention of doingwhat one has promised at the risk of being penalised if one does not is in manyareas of life, to a very large degree, a fiction. If the rights of promisees, whichthose promisees have the option of not insisting on, are real, they cannot dependon such a convention’s being generally established and maintained.

But there is now a problem; and this is a problem which faces not only Humebut also the believer in a natural right to impose conditions on one’s servicesand provisions to others. This is the phenomenon of gratuitous promises, madewithout expectation of return, in the context not of mutually self-interestedexchange, but out of simple charity or friendship.

Here I promise to someone in need, or to a friend, that I will help them,without any expectation of return. So I am neither asserting nor recognizinganyone’s right to impose conditions on their services to me or to others. I am

404 / Thomas Pink

simply committing myself to provide such services gratuitously. And do not suchpromises oblige? But if I do wrong in breaking them, this is certainly not becauseI have breached the promisee’s right to determine the conditions under whichthey provide goods and services to me.

Hume, like Pufendorf before him, assumes that the practice of making andkeeping promises is expressive of self-interest; it exists to enable self-interest tobe pursued in a more efficient manner. But many promises are expressions oflove or kindness and sheer fellow-feeling. It is not plausible that such promisesmade out of kindness to friends or those in need do not oblige morally, or aresomehow peripheral to the practice of promising, or are mere echoes of a centralpractice of making and keeping promises in exchange for some return. To make apromise to a great friend, and to make it just out of friendship, is a very plausibleway of obligating oneself to that friend. People who break such promises madeto friends are very intuitively wronging them.

It may of course be that friendship brings with it prior obligations of animperfect sort — obligations to do some things or other for friends.23 But theseinitial obligations are imperfect. I am under no obligation to do any particularthing. Gratuitous promises change this situation. They effect the crystallizationof these imperfect obligations into definite, perfect obligations to do this ratherthan that — to perform precisely the action promised rather than any other. Andthis obviously presents us with a deep difficulty. For the theory of promissoryobligation we have so far developed applies only in the context of exchange,appealing directly it does to a right to determine the conditions under which wemake goods and services available to others.

6. Grotius, transference theory and contract

Some Protestant natural lawyers, such as Grotius, rather than founding prop-erty on promising as Pufendorf did, took property, or at least quasi-ownershiprights over one’s own self and labour, to be part of natural morality. They thenexplained the obligation to keep promises by assimilating the making of a promiseto a form of property transference. On this view, since we naturally have own-ership or quasi-ownership rights over our own labour and powers, by promisingwe can transfer these rights to others. And it is in virtue of transferring theserights to others that we obligate ourselves to them to act as we have promised:

There is the further fact that ownership of property can be transferred by anact of will which is sufficiently manifest, as we have said above. Why then, sincewe have equally right over our actions and over our property, may there not betransferred to a person also . . . the right to do something?24

Can the power to oblige oneself through promising be made sense of in theseterms? Such an account, if defensible, would certainly take in gratuitous promises,


as well as promises involved in exchange. But the obligation attaching to promisesseems not to work in this way.

Suppose I promise to X to do nothing but work all day tomorrow on hisgarden; and then, without revealing this earlier promise, I go on to promiseexactly the same thing to Y. Whether I am making these promises gratuitously,or in each case I am making the promises as parts of some exchange, as areturn for things done for me by each of X and Y, it seems that through sopromising I can obligate myself to each. Each of X and Y gains a right to mylabour tomorrow on their garden. But how can the transfer theory explain this?According to the transfer theory, any ownership right over my labour or libertyin relation to gardening activities tomorrow has already been transferred to X.So what is there left to transfer to Y?

This criticism attaches to any theory that explains the obligation on meas promisor to keep my promise and the right of a promisee to that promisebeing kept in terms of the transference of something peculiar to me in relationto my action — rights, ownership, authority — to the promisee.25 For since Ihave already transferred away the right or ownership or authority over what Ido tomorrow that could have generated an obligation to Y to garden for him,according to the transfer theory Y has only one ground of complaint againstme: that I have fraudulently misrepresented myself as being in a position toobligate myself to him to act as promised. And in fact he has this ground ofcomplaint whether or not I fulfill my promise to him. But that seems absurd.Surely Y would feel himself aggrieved and wronged only if I failed to deliver onmy promise — precisely because it was to doing that that I had obligated andcommitted myself. It is that failure he would want me to make up to him, as thewrong done to him, not some fraudulent misrepresentation as to the availabilityand transfer of a right.

And this reminds us of the lesson of the doctor-patient example withwhich we began. The promisee need not be, in most everyday cases will not be,concerned with receiving rights over the promisor.26 The promisee’s first concernwill in general not be with whether the promisor is managing to put himselfunder an obligation. The promisee’s immediate understanding of a promise is asinvolving an offer — not any ordinary offer, but still an offer; and the offer isof the action promised, not of some moral obligation to perform it. What thepromisee expects, then, is delivery on this offer. And it is only when delivery failsto occur that the promisee will then start thinking about having been wronged —precisely by that non-delivery.

In our example since I cannot keep my promise to both X and Y, only oneof the promisees can obtain performance of the action promised. And it looksas though it is hardly arbitrary which promisee this should be. Other things atleast being equal, I should do the garden of X rather than Y, since it was to Xthat I first promised my help. But of course this does not show that in thereaftermaking the same promise to Y, Y acquired no right to the same service as X.Rather, when I have a like obligation to both X and Y, and it is or becomes

406 / Thomas Pink

impossible to meet both obligations, then other things being equal we do allowpriority — by whom was the right acquired first — to decide which obligationshould be met. But this does not show that the other did not possess the sameright too. For example, Y’s right and my obligation to supply it still exists, andimmediately determines what I should have to do were X, for some reason, torelease me from my promise to him. Obviously I should then have to do Y’sgarden, because my promise to Y had already and directly given Y a right tothat, and there would no longer be any prior commitment to X standing in theway. Again even if X does not release me from my promise to him, Y’s rightand my obligation to supply it still exists to determine how I shall need to makethings up to Y if I do garden for X. What has to be made up to Y is clearly myfailure to do Y’s garden, the action promised. For it was to the performance ofthat very action promised that Y acquired a right, just as did X beforehand.

Y, in regarding himself as having a right to my performance of the actionpromised, is treating the promise as a mode of commitment — a commitmentto deliver on my promise. And commitments and the obligations associatedwith them are quite different from transfers. For while one cannot over-transfer,one certainly can over-commit. And this seems true whether a promise is madegratuitously, or whether it is made in the course of some exchange. I can over-commit by a series of promises made as a favour. But I can also over-commit byentering into multiple agreements to exchange goods and services, dishonestly orrecklessly making a series of offers of return that cannot be honoured all together.

There are cases of promise or agreement where the transfer model mightapply. These are usually cases where the agreement constitutes a transfer of someproperty title, and where once one such agreement has been made in properform, all subsequent agreements by the initial transferor purporting to transferthe same title are by the very fact of the first invalidated. Certain forms ofcontract in a housing market might constitute just such forms of title transfer.Arguably, an exchange of vows in marriage could also be understood take thesame form — a form in which each party transfers themselves to the other,thereby (at least in the absence of any subsequent cancellation of the contract)invalidating any further such exchange of vows either party enters into with athird. Contrast the exchange of vows in actual marriage with a prior promiseto marry, which simply commits the promisor to a future transfer, and whereover-commitment is possible.

7. Gratuitous promises and liberty

We have yet then to arrive at theory of promises that are gratuitous — thatare made not as part of an exchange but, it seems, for nothing in return. Suchpromises may be made out of sheer altruism and generosity or out of friendship.It is important that such promises are very frequently made and they can be asobligatory as any.


It might be tempting to include such promises in our theory thus. What arethe moral implications of the existence of a right to determine, within limits,the conditions under which we make our goods and services available to others?We have viewed such a right as constituting a certain natural liberty: within thelimits set by our obligations to others, we have a right to determine for ourselveswhat we do, and especially how we dispose of our selves and our resources andgoods. The infliction on us of coercion and fraud by others is wrong because anassault on this liberty. But if we have this liberty, can we not voluntarily lay itdown? Can we not obligate ourselves to others at will? If so our ability to obligateand commit ourselves through gratuitous promises is in the end related to ourcapacity to obligate ourselves through promises in the context of exchange. Ineach case the obligations involve the liberty of the individual — whether thesurrender of that liberty by the promisor, or a duty to respect that liberty in apromisee.

This proposed resolution is attractive. But it is, I suggest, misconceived.There may, as we have indeed supposed, be such a liberty. But, as we are nowabout to see, we may not be able to lay it down at will, obligating ourselves justthrough expressing our intention to do so. And secondly, how do we incur theobligation? What is involved in laying down our liberty in these cases, such thatwe are left with an obligation to someone else?

In the context of exchanges we can see what explains the loss of liberty whenthrough promising we lose it. The loss of our liberty may be explained by ourduty to respect that of another, and the conditions he has placed on those ofhis services that we have chosen to accept. But no such explanation is availableto cover gratuitous promises. There the obligations are not incurred to respectany conditions placed on what we have received. Nor can the laying down ofone’s liberty involve any transfer of rights to others, for the reasons we have seen.But if laying down one’s liberty to others does not involve transferring rights tothem, what does it involve when gratuitous promises are in question? Clearly,to lay down one’s liberty involves coming to have an obligation to those others,where one was unbound by any obligation previously. But how then does theobligation arise? No story has been given, and the appeal to liberty here does noexplanatory work, but merely asserts that we have — somehow — a power toobligate ourselves.

Notice that our power to obligate ourselves through gratuitous promisingis not evidently unlimited. It certainly seems not to display any very simplepower to lay down our liberty or obligate ourselves at will. For not all gratuitouspromises obligate the promisor with equal obviousness, even when the promisedservice is accepted by the promisee and even when the promise is accompaniedby a genuine intention on the part of the promisor to obligate himself. And thisis so even when there is no coercion, fraud, or immorality in what is promised,nor any harm to the promisee.

Suppose in the financial district I approach a complete stranger in the street,someone entirely unconnected to me, who is prosperously there employed and

408 / Thomas Pink

in no obvious need of funds. Out of my own whim, and specifically in orderto obligate myself — being so obligated would be interesting, I think, and so,pursuing this interest, I really do intend to bring this state of obligation about —I promise to give him £20 within half an hour, and he accepts the promise of themoney. (Why should he not accept? — he may not really need the money, butevery little helps, and I, as a clear eccentric, might just deliver on my promise.)Am I obliged to deliver the money? If I fail to hand the money over, will Ihave actually wronged the promisee? It seems at least debatable. If the moralobligation does exist at all, it is not obviously a very serious one. Even if thepromisee would be wronged by my failure to deliver on my promise, it is notclear that the wrong would be significant. And it is at least arguable that nobinding obligation exists at all. Yet if I did have the power to obligate myself atwill through promising, then given my genuine intention to obligate myself inthis case, surely my obligation to act as promised would be as clear in this caseas in any?

Compare other examples of gratuitous promising that clearly are obligation-inducing. Consider, for example, a case where I promise a great friend that I willattend the exhibition of his latest artwork. (Note the difference if the friend isnot so great. Perhaps the obligation would be less serious because of that — apoint to which I shall return.) Or consider a case where I promise a favour toa social acquaintance or a work colleague. Or consider a case where I promisea stranger in obvious need that I will help him, with money or in some otherspecific way. In all these cases the promises are similarly gratuitous, but clearly dooblige. Breach would much more plausibly constitute a significant wrong doneto the promisee. Yet the difference between these cases and the first has nothingto do with some supposed power on my part to give up my liberty or to obligatemyself at will — a power which if present at all would be present and operativein all three cases equally.

Certain social contexts such as acquaintance, shared employment, friend-ship, especially serious friendship, or need in a stranger — these make theobligating nature of gratuitous promising particularly clear, and as intelligibleas in any case of exchange. Clearly something else is generating the obligationbeyond a will on my part to obligate myself. And, remember, as the case of thedoctor and his patient showed, even in sincere promising such a will to obligateoneself need not be present anyway.

8. Promises as invitations to trust

I shall now argue that gratuitous promises involve a distinctive way ofrelating oneself to another. And I think this mode is as much present infactive promises or promises that, as in performance promises or promises to.Recognizing and understanding this form of relation will enable us to understandhow gratuitous promises to act generate moral obligations to act as promised.


But it will also uncover what is common to both promises to and promises that,and explain why we see and speak of both cases as equally cases of promising.

The mode of relation involves, first, the promisor making an offer to thepromisee — of information in the case of promising that and of future action orperformance in the case of promising to. In the case of the information providedin promising that, what is offered — the conveyance of information or a truth —is provided simultaneously, if it is to be provided at all. The very act of promisingwill have provided it, and all the promisee or recipient can do is fail to believe.With promising to things are different, and the offer is not yet delivered, and sothere is room for the promisee to refuse what is offered before delivery. If thepromisee accepts, or at least does not refuse, the sincere promisor will thereafterbecome motivationally committed to deliver. That is, he will remain motivatedto deliver what has been offered subject to the promisee’s continued wish toreceive it.

What else is in common to the two cases besides an offer? In addition tothe offer there is an invitation to trust the offeror in relation to his offer andits genuineness. Trust here involves two connected elements. It involves not onlyrelying on the promisor to deliver, but also obtaining comfort and reassurancefrom the promisor in relation to the delivery. More specifically the promisee isinvited to take a twofold trust in the offeror. The promisee is invited to rely on,and to take comfort and reassurance from, both the promisor’s motivation andhis capacity to deliver.

(1) Offer (of information or of action)

+

(2) Invitation to trust in the offeror:

specifically in

- the offeror's motivation to deliver what is offered

- the offeror's capacity

Figure 1. The common essence of promising.

In the case of factive promising that the promisee is invited to trust, first,in the promisor’s motivation to tell the truth — in his honesty. Secondly,the promisee is invited to trust in the promisor’s epistemological capacity orreliability — in his ability to get his assertion right. Notice this invitation to trustis not part of everyday assertion. For whilst in the context of serious conversationassertors are supposed to assert what they really do believe, they do not thereby,just through asserting, actually invite conversational partners to put their trustin the sincerity of the assertor. Still less are conversational partners invited, justthrough the act of assertion, to put their trust in the epistemological capacity orreliability of the assertor.

410 / Thomas Pink

Not only is the addressee not necessarily invited to trust in or rely on theassertor’s sincerity or reliability; there are contexts of serious assertion in whichsuch trust or reliance would be uncalled for or even inappropriate. Thus in adebate or argument, the truth of what is asserted may well be argued for simplyon the basis of the supposedly evident truth of other propositions asserted,and not on the basis of any appeal to the assertor’s own sincerity or epistemiccompetence. Whereas factive promising is quite different. Someone who insists,in the face of scepticism whether something that they witnessed really happened:‘I promise you, it did, I saw it happen with my own eyes’, is exactly inviting trustboth in their own truthful motivation and in their capacity to get what reallyhappened right.

In the case of promising to, the invitation to trust is again twofold. First, asbefore, the promisee is invited to trust in the promisor’s motivation — in thiscase in their continuing motivation to deliver on the offer for as long as thepromisee wishes that they should. Secondly they are again invited to trust inthe promisee’s capacity — in this case in the promisee’s ability to provide whatthey have offered. The second invitation and its object is just as important asthe first. When I promise to do something for you, I am inviting you to trustnot only in my willingness to do it, but in my implied assurance of my actualability to do it. Which is why it seems so dishonest to make promises which, nomatter how motivated one may be to act as promised, one doubts or disbelievesone’s own capacity to deliver. Those who promise that they will pay the rent byFriday may be very motivated to pay as promised should they find the funds(which they may be very motivated to do as well); but they do not promisehonestly if they doubt or disbelieve that by then they will actually have the fundsrequired.

But this account suggests that, far from being something peripheral andirrelevant to any account of promising to, promising that embodies a commonessence and structure found in promising generally. In each case the commonalityis the conjunction of an offer with an invitation to a twofold trust in the offeror —in his motivation and in his capacity to deliver what is offered.

We now see why Hume was wrong to think that only thoughts aboutobligation can motivate us to keep a promise. We can be motivated by a farless philosophical and far less moralistic thought. We can be motivated, as thedoctor was in relation to his patient, by the simple thought that we have givenour word. And now we see what giving one’s word comes to — something that, asI shall now suggest, is commonly a basis of obligation, but which in itself comesto something distinct from mere obligatoriness, and which can intelligibly figureas a motivating thought apart from it. In giving his word, the doctor has invitedthe patient to trust in him; and it is this conception of his relation to the patientwhich can on its own be what is motivating him to deliver on his promise. Andthat invitation to trust chimes with what the patient promisee wanted all along:the receipt not of an obligation in their favour, but of an offer with assurance ofits delivery — an offer in which the promisee can sensibly take trust.


On this account it is the invitation to trust that explains the promissoryobligation in such cases where a gratuitous promise does oblige. The questionwhen do gratuitous promises oblige comes down to this: under what conditionsdoes inviting someone to trust you to do something oblige you to do it? Andrestating the question in these terms really is explanatory. The appeal to aninvitation to trust is not circular, a mere verbal variation on ‘promise’. Forthe idea of trustworthiness gives us independent purchase on when gratuitouspromises give rise to obligations and when they do not, or at least not toobligations that are serious and clear.

For example, the obligation to act as promised most convincingly andcompellingly arises in those cases where the trustworthiness of the promisormost clearly matters to the promisee. And the promisor’s trustworthiness willcertainly matter in a case of need — where the promisee is vulnerable, and needsto entrust themselves to others. Equally it will matter when the promisor isan acquaintance or colleague or friend of the promisee. And in the case of afriend, it will matter more the greater the friendship involved. Which is, otherthings being equal, the greater the friend, the greater the wrong done by breakingpromises made to them; the obligation attached to delivery becomes more seriouswith the friendship. Whereas the promise of £20 to the random stranger in nofinancial need is quite different. Here we saw that the existence of a real andbinding moral obligation to deliver the money to the stranger is very much morearguable. Even if the obligation does exist, it is not obvious that the strangerwould be greatly wronged by non-delivery. Our theory clearly explains why theobligation is plausibly less serious and why even its very existence is at leastdebatable. For in this case it really is debatable how far the trustworthiness ofthe promisor in relation to his offer matters to the promisee. Certainly it is hardto see the trustworthiness of the promisor mattering very much; which is why nomatter how determined and intent the promisor might be to obligate himself, itis so much harder to see a really serious moral obligation here.

Notice, in particular, that the standard account of what promisors intendto communicate — the account given by Finnis, Owens and Raz — doesnothing to make promissory obligation plausible, and does so precisely becausethe interpretation it gives of promising has nothing to do with inviting trust.According to the standard account, when I make a promise, I am deliberatelyexpressing an intention to obligate myself to you. But of course one can dothat — simply by explicitly asserting one’s possession of that very intention. Tothe prosperous stranger in the financial district I could just say, ‘It is my intentionthat I now be under an obligation to you to pay you £20 within half an hour’. Butsuch a linguistic performance taken on its own is hard to recognize as a normalpromise. And the reason why is clear by now. Such a locution has nothing todo with inviting trust. Indeed it would be strictly consistent to combine such alocution which, after all, is about my obligations to you, rather than about mytrustworthiness to you, with an actual disinvitation to trust: ‘I intend to placemyself under an obligation to you to pay you £20, but don’t trust me to deliver

412 / Thomas Pink

on it’. But how could such a locution ever oblige as gratuitous promises do,and what point would making it have? Since in gratuitous promising it is aninvitation to trust which is essential to generating the obligation, it is quite vainto seek to impose a promissory obligation on oneself as promisor while explicitlydiscouraging trust.

Not that trust need actually be given by the promisee for promises to oblige.It is perfectly possible for a promisee to be sceptical about a promisor withoutthat scepticism in any way removing the promisor’s obligation to deliver. This istherefore not a variation on the expectational theory of promissory obligation,where the obligation to keep a promise is explained in terms of an obligationnot to disappoint deliberately raised expectations of delivery.27 The expectationaltheory is not credible, in that promises can perfectly well oblige even when thepromisee remains sceptical of actual delivery, and the promise has done nothingto raise those expectations. So what creates the obligation on the promisor cannotbe the fact that he is actually trusted in by someone. Actual trust does not ofitself obligate in any case. The person trusted might never have intended to receivesuch trust, or they might even have sought to discourage it. What generates theobligation is the fact that the trust, whether actually given or not, was deliberatelyinvited. For deliberately to invite trust, and then to betray that trust for no goodreason, and to do so in circumstances where one’s trustworthiness clearly mattersto the person betrayed, is to show them contempt. And it is that contempt thatexplains the peculiar sense of wrong felt by the victims of broken promises. Asidefrom any inconvenience caused by disappointed expectations — and perhaps thepromisee’s expectations were never very high — there will be not merely a sense ofhaving been wronged, but underlying this sense of having been wronged a feelingof having been trifled with. One has been treated with a degree of contempt; andfor that reason one has been treated wrongly. By not immediately characterizingpromising as mode of voluntary self-obligation, we have managed to explain whythe victims of broken promises should feel wronged, and why the wrong is feltso personally.

Notice that it is perfectly possible effectively to make a promise without usingexpressions such as ‘I promise’. For one can perfectly well invite someone to trustone to do something, or invite someone to trust one in relation to one’s assertionof a fact, without actually using words such as ‘I promise to’ or ‘I promiseyou that’. But where performance promises are in question, and the promiseehas a particular need for reassurance that the future action promised reallywill be performed, a promisor may sometimes go on to use such expressions,or promisees may insist on their use, as a further stage in the expression ofcommitment even once an invitation to trust has already been delivered. Thus apromisor may say: ‘Trust me, I’ll be there — and that’s a promise’. Or a promiseemay on hearing the words, ‘Trust me, I’ll be there’, ask: ‘Was that a promise?’ or‘Is that a promise?’28

This might suggest to some that to utter the words ‘I promise’ really is todo something distinct from and going beyond issuing a mere invitation to trust.


On this view, what these cases show is that to add ‘And that’s a promise’ reallyis to express an intention to obligate oneself; and promisees insist on the use ofsuch expressions even once an invitation to trust has already been made for onereason only — in order to ensure for themselves the receipt of a moral obligationin their favour where none would otherwise have existed. In which case, it mightbe concluded, we should move back to the standard account of promising andpromissory obligation that we began by rejecting — the account of promising asthe expression of an intention to obligate oneself, an expression of intention thatthereby mysteriously brings about the obligation supposedly intended.

But is this reading really plausible? On this view, since inviting trust is onething, but promising is quite another, one could sensibly issue a clear and explicitinvitation to trust while disavowing any promise: ‘Don’t worry, trust me, I reallywill be there — but I’m not promising.’ One’s motive in saying this would beclearly and publicly to avoid an obligation to do it — an obligation that, onthis view, only actually making a promise would incur. But in reality such anexpression is really very odd. We would never, in the ordinary course of things,clearly and emphatically invite someone to trust us that we will do something, andimmediately deny that we are promising to do it. The disavowal of any promisewould surely be heard, rather, as an eccentric withdrawal of the invitation totrust. The addressee would certainly no longer be inclined to place much trust inthe thing being done for as long as they continued to wish it done.

There is of course a way to make sense of the combination of an invitationto trust with an explicit refusal to promise. Suppose we are dealing with somespecial legal context where the utterance of certain words such as ‘I promise’ or ‘Icontract’ creates a legally recognised contract, with rights enforceable under lawattached to it, that might well not otherwise exist. Then, of course, there wouldbe some clear intelligibility to inviting trust while still preferring unambiguouslyto avoid any such legally enforceable relation. Certainly, ‘Trust me, I will do it,but I’m not entering into any contract’ is an entirely intelligible thing to say. Butthen the function in such a context of ‘I contract’ or ‘I promise’ would not beto express any intention to incur a moral obligation. Rather its function wouldsimply be to trigger a purely legal or contractual obligation, the moral force ofwhich might be very debatable, whether this purely legal obligation was intendedor merely foreseen. And use of the phrase ‘I promise’ does not usually have thisspecifically legal function.

Ordinarily, clearly and emphatically to invite someone to trust us that wewill do something just is, in effect, to promise them that we will do it, and willreliably be understood as such. If, after someone’s trust has been clearly andemphatically invited — ‘Trust me, I really will do it’ — the thing is not done,a claim made in justification of not doing it to the effect that ‘I never actuallypromised to’ is unlikely to be heard with any patience. A real moral obligationto deliver comes with the invitation to trust; and one who issues an invitation totrust will be held bound by an obligation to deliver whether or not they actuallyused the word ‘I promise’ and ‘That’s a promise’. They will, in other words, be

414 / Thomas Pink

treated as any promisor who has made a promise; and if they fail to deliver, theywill be regarded as any promise-breaker.

If a person whose trust is being invited already has a moral obligation intheir favour, just by virtue of the fact that their trust has been invited, what arethey seeking in asking ‘Is that a promise?’ or ‘Was that a promise?’ What theyare seeking is not a moral obligation, which they may indeed possess already ifany such obligation arises at all, but rather explicit acknowledgment from thepromisor that through inviting trust a promise was being made and an obligationwas incurred. That is, the function of ‘I promise’ and ‘that’s a promise’ is notto express an intention to obligate oneself — an intention that, as we’ve seen, asincere user of such expressions, such as the doctor who honestly says ‘I promiseI’ll be there’ in trying to comfort and reassure his patient, need not possess.Rather the use of such expressions is explicitly to acknowledge that a seriousinvitation to trust really is being made — serious both in that its making really isintended by the promisor, and really does matter to the promisee, and matter tothe point of obligating the promisor. Moreover emphasis is further given to thefact that the promisor’s offer and his motivation to deliver on it really is subjectto the promisor’s continuing wish to receive it.

For I can trust someone to do something, but to do it without any referenceto my wishes. For example, I can trust them to do it whatever, or to do it subjectto their continuing themselves to think that doing it would be in everyone’sinterest, or subject to a variety of other conditions. But in promising the promiseeis invited not just to trust in an action’s being performed, but in the agent’sperformance of it and motivation to perform being specifically conditional onthe promisee’s wish that that the action be performed. The action is being offeredto the promisee, as something that the promisee can refuse, or accept; but which,if he accepts, will be performed unless and until the promisee ceases to desirethe action’s performance and releases the promisor from his promise. Such acondition is typically understood by both parties in invitations of the form:‘Trust me, I’ll be there’, and is certainly understood in so far as such invitationsto trust are heard as promises. The function of ‘And that’s a promise’ is, then,formally to confirm such an understanding.

To make my invitation, its seriousness and its reference to the promisee’swishes clear, I don’t have to use the words ‘And that’s promise’. But using thosewords or like expressions is one way of making it quite clear what I am doing.And the promisee’s interest in hearing such expressions now becomes obvious.It is not that they are essential to generating an obligation, which may anywayalready have arisen, and which in itself promisees often don’t much want. Whatpromisees seek from promisors, rather, is security and reassurance. And this isthe point of ‘And that’s a promise.’ The promisor’s utterance of ‘And that’s apromise’ makes it clear to all parties that the promisor understands very wellboth what he is doing, and its personal and moral significance to the promisee.It is made very clear that the promisor is giving his word, and that he fullyrealizes that he is doing this and that this matters to the promisee. At least in


the absence of independent evidence of the promisor’s dishonesty or incapacity,this all the more enables the promisee sensibly to put their trust in the promisor,something which the promisee wants to be able to do.

Of course using the phrase ‘And that’s a promise’ may well increase theobligation on the promisor even if one would have arisen anyway. For in so faras it plausibly increases the level of trust invited, so too it adds to the contemptshown to the promisee, and so to the wrong done to the promisee, if for no goodreason the promise is subsequently broken. Hence one can further understandwhy the false impression may have originally arisen that what we have here inthe use of terms such as ‘I promise’ is a quasi-legislative act — an act that quasi-legislates obligations into existence just through the expression of an intentionthat those obligations exist. For certainly the explicit use of terms such as ‘Ipromise’ may increase a promisor’s obligation to the promisee — not by creatinga new obligation with a different basis to any existing before, but rather byreinforcing the original basis of the obligation, which all along lies in the activityof inviting someone’s trust.

Given that explicit use of the expression ‘I promise’ does serve to acknowl-edge the creation of an obligation, suspicious and mistrustful promisees have anevident interest in insisting on its use. For in so far as the expression serves tohighlight one central outcome in many contexts of inviting trust — the creation ofan obligation to deliver on the offer — and thereby adds to that obligation, use ofthat expression binds the promisor to the promisee in a public and unambiguousway. It is not so much that an obligation exists at all; for that obligation mighthave existed without the words. Rather, through the use of this expression, itis now explicitly acknowledged by the promisor what has been done: that thepromisee’s trust has been invited so as to obligate the promisor to deliver on hisoffer. That means that if the promisor has any interest in not breaching or atleast in not being seen to breach an obligation to the promisor, that interest willmotivate him to act as promised. And third parties who are inclined to enforcesuch obligations, or to apply social pressure on those who breach them, will thenbe more sure to act in this particular case. But the source of the obligation isall along the same, and independent of the use of such expressions: it lies in theissuance of an invitation to trust in a context where the trustworthiness of thepromisor is of sufficient significance to the promisee to obligate delivery.

Notice that alongside the use of the expression ‘I promise’ to make a promise,there is an ironic use of it to issue a threat — as in ‘If you trespass on my propertyagain, I promise you I’ll shoot you dead’. Here there is no offer of an actionpromised as some kind of benefit to the promisee which they might choose toaccept; nor is there an obligation on the offeror to perform the action promisedshould the offer be accepted. But of course, one could make the same threat inthe same ironic manner by using the vocabulary of inviting trust: ‘If you trespasson my property again, trust me, I’ll shoot you dead’. And it is plain in eithercase why the invitation is ironic, and why no genuine promise is being made.For the addressee is hardly being invited to take comfort and reassurance from

416 / Thomas Pink

the interlocutor’s motivation and capacity to deliver on the ‘promise’. Rather thereverse is plainly true — it is fear and dread that are being invited.

9. The morality of promising and of promise keeping

We have then two quite distinct bases for the obligation to keep promises.One lies in people’s right to control the conditions under which they providegoods and services to others. Here the basis of the obligation to keep promiseslies in a respect for liberty, and the failure to keep promises to others is akin tocoercion of them. The other foundation for the obligation to keep promises liesin something rather different — a duty to respect an attachment of people toourselves such that our trustworthiness matters to them, whether as those madedependent on others by need, or whether just as our neighbours or our friends.We should employ promises to provide genuine assurance and deliver on it, notto trifle with those to whom assurance matters; and here the failure to keeppromises is wrong because it belittles others, and is form of unkindness to them.So one and the same phenomenon of promising unites two contrasting aspectsof our humanity: the liberty and moral independence of each individual, and thepractical and emotional attachment of individuals to each other.

Each basis of promissory obligation seems to leave the moral obligation tokeep promises an obligation that is, in Pufendorf and Hume’s terms, naturalrather than artificial. That is, the rights on the part of the promisee to deliveryon the promise that are at stake in each case seem pre-conventional. There isno need for some existing practice of promise-making and keeping to exist forone to possess a right to control the conditions under which one does things forothers. Similarly no conventional practice is required to make intelligible people’sgeneral need to be able to trust others, and for that need to be respected by thosewho deliberately invite trust.

This leaves the rights of promisees and the obligation of promisors to respectthem importantly unlike the rights of property holders. For it is very intuitivethat, just as Hume supposed, the right of a property holder to a particular goodthat he owns and the moral obligation of others to respect that right do dependon a developed system of conventions determining ownership — conventionspartly fixed by formal legal statute and partly found in informal custom. In thecase of property, conventions fix ownership rights, and so are presupposed by anymoral obligation to respect a given individual’s ownership of a particular good.But property is not the same as promising, and Hume’s attempt to assimilate thenatural phenomenon of promising to being a supposed analogue of the artificialinstitution of property was mistaken.

But while many promissory rights and obligations do have a natural basis, sothat many are fixed preconventionally, some may still be artificial. For promisingis a practice that is given some degree of legal recognition and enforcement —in the form of legal contract. Here legal statute or legal custom will recognize


certain promises and agreements as contractual, and establish certain legal rightsand obligations in relation them as enforceable through legal process. And hereit is arguable that our moral obligation to respect laws generally will morallyoblige us to respect such legal rights and obligations. Here the immediate basisof our moral obligation to act as promised qua fulfil our contractual obligationmay lie not in any pre-legal, natural obligation to respect the liberty of another,or to deliver on an invitation to trust. For we may not have invited any trust.And perhaps in the contractual exchange the counterparty was not morally atliberty not to do what he has done for us. Here the immediate basis of our moralobligation to respect the contract may indeed consist in the simple existence ofa certain contract, and so in a use of certain words or deeds by each party toconstitute that contract.

So just as there are convention-dependent property rights, equally there areconvention-dependent contractual rights to receipt of goods and services. Butthat does not mean that the rights of promisees in general are convention-dependent, or that even in non-legal contexts, it is convention alone thatdetermines or fixes the rights of promisees and so any moral obligation onthe part of promisors to meet them.

In the case of gratuitous promises there may be some temptation to play-actas if there were a quasi-contractual significance attaching to the use of wordssuch as ‘And that’s a promise’. Children especially may treat whether in invitingtrust they or a parent used words such as ‘I promise’ as if it did signify asort of contract — as if, that is, there really were a special moral obligation tokeep the promise based just on the use of those particular words. And that isbecause children may actually often be involved, along with their adult guardians,in quasi-legal games or institutions of promise-making and promise-keeping,in which the specific use of certain words does engage entirely local ‘familyconventions’ of penalisation or compensation for non-performance. Here thefunction of the words ‘I promise’ may be precisely to engage those structures; andthen it may matter that even if one may have invited trust, one did not actuallyuse those words. But most adult promising is simply not like this, and is notplausibly to be modelled on family games of discipline and cooperation playedwith children. If obligations do arise in gratuitous promises made by adults, theirbasis seems very much to lie in the fact that trust was invited, and not in the useof certain words considered apart from such an invitation. Any adult who clearlyand unmistakably invites trust and then seeks to evade what they have done byinsisting that they never actually used the words ‘I promise’, would be reducingthe ethical to the quasi-legality of a child’s morality of ‘black marks’ and ‘naughtysteps’. If the use of words such as ‘And that’s a promise’ does have any significancein extra-legal contexts, it lies — as we have seen — both in rendering the fact ofan invitation to trust and its exact terms unambiguous, and in deepening the levelof trust invited, and so the wrong done to the promisee if the promise is broken.

We have seen that not all promises are involved in some form of exchangethat invokes the duty to respect the liberty of others. Some promises seem to

418 / Thomas Pink

oblige purely through constituting invitations to trust. On the other hand thereare clear cases of contract and agreement where the opposite seems true; wherethe element of invited trust and the duties flowing from such an invitation seemnot to be significant. If I take out a credit card, I agree and contract to pay offthe debts incurred under it. But does my trustworthiness matter to the creditcompany, and am I even inviting them to trust me to repay? The card companyis not seeking borrowers each of which it can individually trust, but simply wantsto amass a vast credit book that achieves in total an ideal balance of risk andreturn. Too uniformly reliable a pool of card holders, and though there will beno defaults, their borrowings and the income generated on them will be too low.There is nothing, then, in the card company’s relation to me analogous that ofone awaiting the return of a sum loaned to a friend or neighbour to help themget by, and to whom the trustworthiness of that individual borrower does matter,and not simply because they would like the money back. Some might insist that Iam not standing as a genuine promisor to the card company; that in the absenceof an invitation to trust, I agree to repay what I owe but do not really promise.But we could as well admit that we have here a case of promissory obligationrooted in the liberty of the promisee, rather than in their attachment to us.

It would in any case be quite wrong to imagine a rigid distinction between,on the one hand, gratuitous promises that engage a respect for others as thoseto whom our trustworthiness matters and, on the other, contracts or promises inrelation to exchange that engage a respect for their liberty. For exchanges betweenfriends can engage both forms of respect, as can many commercial transactionswhere counterparties in commerce do also meet as at least acquaintances or asneighbours. The rigid distinction between a sphere of benevolence or love and asphere of justice that we find in Pufendorf or Hume is simply not credible.

There is no such rigid distinction between the spheres of justice and of lovebecause it is the same human beings who combine the possession of moral libertyfrom their fellows with an attachment to them as neighbours or as friends or assources of potential help and support; and because many of their dealings witheach other involve both of these characteristics simultaneously. The practice ofpromising in relation to exchange has evolved amongst individuals who also seekassurance and care about each other’s trustworthiness. Hence the developmentof a single practice of making and keeping promises that can invoke two quitedistinct bases of obligation — bases of obligation that can fall apart, but thatare so often found together.

Notes

1. Versions of this paper have been given at the Universities of California atRiverside, CEU Budapest, Chicago, Keio Tokyo, Leuven, London, Princeton,St Andrews, and York.

My thanks for their comments especially to Thomas Baldwin, AnnabelBrett, Daniel Brudney, James Conant, Tim Crane, Wim Decock, Wolfgang


Ertl, Kati Farkas, Peter Graham, John Haldane, Harold James, Jonathan Lear,Michael Martin, Elijah Millgram, Veronique Munoz-Darde, Martha Nussbaum,David Owens, Jon Parkin, John Skorupski, Timothy Stanton, Carlos Steel, EricSchliesser, Martin Stone, Candace Vogler, Gary Watson.

2. Natural Law and Natural Rights, Oxford University Press (1980), pp. 298–299.

3. ‘A simple theory of promising’, Philosophical Review 115: 1 (2006), p. 51.4. The Authority of Law, Oxford University Press (1983), p. 257.5. ‘Promises and obligations’ in P. M. S. Hacker and J. Raz (eds.), Law, Morality

and Society (Oxford University Press, 1979), p. 218.6. That is not to rule out cases where the patient might be particularly concerned

to secure an obligation in her favour. It is just to suggest that there can perfectlywell be other cases where all that is really wanted is the doctor’s presence, andadvance assurance of that presence.

7. A Treatise of Human Nature, L.A. Selby-Bigge (ed.), Oxford University Press(1978). Book III Part II section V ‘Of the obligation of promises’, p. 518.

8. My use of the term ‘promises that’ for promises as to matters of fact is simplya useful abbreviation. It is not meant to deny the fact that one can of courseuse the ‘promise that’ form in communicating what I call promises to: as in ‘Ipromise you that I will do it’.

I have learnt that both in Japanese and in certain US contexts, the use ofone and the same expression, such as ‘I promise’ to effect both factive andperformative promises, is much less common than it is in British English. Forexample, in the US factive promising more commonly involves the use of ‘Iguarantee that’ or ‘I swear that’ or some like expression. For reasons thatwill become evident, this will not matter to my argument. For what is donein promising, either factive or performative, can be done with or without the useof a single set expression such as ‘I promise’.

9. Second Treatise of Government, Peter Laslett (ed.), Cambridge University Press(1960). Chapter 2, ‘Of the state of nature’, para 14, p. 277.

10. Op. cit., p. 516.11. I shall suggest below that the ability to promise and incur promissory obligations

is not in fact dependent on the existence of distinctive linguistic terms such as‘promise’, but can be exercised prior to and independently of them.

12. For more on this central claim in early modern natural law theory, see my‘Natural law and moral obligation from Suarez to Locke’ in Sara Heinamaaand Martina Reuter (eds.) Psychology in Philosophy: from Late Scholasticism toContemporary Thought, Dordrecht: Springer (2008).

13. See his De Iure Naturae et Gentium, Amsterdam (1688).14. On the Law of Nature and Nations, C. H. Oldfather and W. A. Oldfather (trans.),

Oxford: Clarendon Press (1934), Book 3, chapter 4, para 1, pp. 379–80 (Libraryof Congress translation).

15. Op. cit. p. 519.16. Ibid p. 521.17. Ibid. pp. 521–523.18. Ibid. pp. 521–523.19. Ibid. pp. 521–523.20. Ibid, p. 516.

420 / Thomas Pink

21. Notice that neither farmer has had to use terms such as ‘I promise’. The rightto determine under what conditions one provides one’s services and goods toothers preexists the use of such terms as ‘I promise’, and can be understood andasserted without them. It is this preexisting right that helps make the practice ofpromising intelligible. As I suggested earlier, the ability to incur what are in effectpromissory obligations does not depend on linguistic terms such as ‘promise’,but can be exercised prior to and independently of such terms.

22. “This sentiment of morality, in the performance of promises, arises from thesame principles as that in the abstinence from the property of others.” Hume,ibid. pp. 521–523.

23. My thanks to Martha Nussbaum for emphasising this to me.24. Hugo Grotius, On the Law of War and Peace, F. W. Kelsey (trans.), Oxford:

Clarendon Press (1925), para 3, p 329 (Library of Congress translation).25. Such a transfer theory is defended by David Owens (op. cit.). According to him,

promissory obligation involves the transfer of the promisor’s authority over hisaction to the promisee: ‘In promising you a lift, I grant you the authority torequire me to give you a lift. . . . to use Kant’s metaphor, in accepting my promiseyou “take possession of my choice”’ (p. 71).

26. Owens claims (ibid., pp. 71–72) that when the promisee (supposedly) acquiresan authority over the promisor, or ‘takes possession’ of the promisor’s choice,so that the promisor is obliged to perform, this authority ‘is something that theparadigmatic promisee wants for its own sake’. I doubt this is generally true.

27. For a recent version of the expectational theory, see T.M. Scanlon, What WeOwe to Each Other, Harvard University Press (1998), chapter 7.

28. My thanks to Jonathan Lear for emphasizing such cases to me.

Promising and Obligation

Documents