Introduction
There is much philosophical debate about the morality of promising, most of it focused on why we are morally obligated to keep our promises. My aim in this paper is to analyze an aspect of the morality of promising that has not been as thoroughly explored in the literature and to investigate not why we ought to keep promises or when it is permissible to break them, but when it is morally permissible to make promises in the first place, with a special focus on the permissibility of making promises under conditions of uncertainty. (For the purposes of this paper, I presume that we are morally obligated to keep our promises, all else being equal. I will not take a stand here on why this is the case, as my claims about the permissibility of promise-making are neutral regarding what theory we presume about why promise-keeping is morally valuable or required.)
As potential promise-makers, we want guidelines for deciding when it is and is not appropriate to make promises. And as potential recipients of promises, we want to be able to assess whether others have acted appropriately in their promise-making. Such assessment matters in our private lives and personal relationships as well as for our responses to public promises, such as the campaign promises made by candidates for office in democratic political systems. We want our official government representatives to behave in morally upright ways, and whether they do so can affect whether we vote for them and continue to support them. One way in which we can morally assess our politicians is by assessing their campaign promises. We usually focus on whether these promises are ultimately kept or broken. But it is also worth looking into whether such promises should have been made in the first place, as this gives us a more comprehensive assessment of the morality of our politicians’ actions.
Campaign promises are especially interesting because they are often made under conditions of uncertainty; the amount of cooperation required to attain political goals and the pressures politicians are subject to from lobbyists and others make it the case that politicians who promise to Φ usually cannot be certain that they will in fact Φ. And uncertain promises are puzzling on the face of it for two reasons. First, promises typically clue promisees in to what is going to happen in the future, which allows them to plan accordingly; campaign promises in particular convey information about how potential politicians will behave if elected, and voters rely on such information in making decisions about whom to vote for. But how can promises for uncertain ends provide any sort of robust assurance? Second, sincerely promising involves strongly volitionally committing yourself to acting. But how can you commit yourself in good faith to attain an end that you are not certain is attainable?
In this paper, I offer a general account of the conditions under which promise-making is morally permissible, using campaign promises as an interesting concrete illustration to highlight the special challenges raised by uncertainty. My concern is with what conditions must be met—and in particular, how much uncertainty there can be and from what source—for a promise to be morally permissible to make. I first frame my question more clearly and then outline four necessary conditions that must be met for a promise to be morally permissible to make. First, the promise must be for a morally permissible action aiming at a morally permissible end. Second, the making of it must not deceive others. Third, the promise must be made in good faith, which means it must be made with a sufficiently strong level of subjective volitional commitment. Fourth, promise-makers must have a realistic conception of their own abilities, which is especially important in high-stakes political contexts. These conditions will not always be jointly sufficient; there might be additional constraints on permissible promise-making in particular situations.
I then assess whether promises that you are not certain you can keep can meet these conditions—for if they cannot, then most campaign promises are impermissible to make. I argue that uncertain promises can easily meet the first two criteria, since they can be for morally permissible ends and can be nondeceptive. Whether they can meet the last two criteria depends on the source of the promisor's uncertainty. Promises that are uncertain because of external factors can be made in good faith and with a realistic self-assessment. But promises that are uncertain because of internal factors—such as the promisor's doubts about her own strength of will—either fail to be in good faith or are unrealistic. I suggest that many campaign promises are internally uncertain; it follows that they are impermissible to make. Finally, I offer suggestions about what would-be promisors who are internally uncertain might do instead.
1. Framing the Question
I focus on permissibility conditions on promising (which determine when a promise is morally permissible to make) rather than on validity conditions (which determine when an act counts as a promise in the first place). This is because I presume that it is possible to make morally impermissible promises—that is, that a promise to drive the getaway car for a robbery is impermissible to make, but still counts as a promise. (I argue for this claim in more detail in Liberman [Reference Liberman2015: ch. 1]; for arguments to the contrary, see Austin [Reference Austin1962], Shiffrin [Reference Shiffrin, Wallace, Kumar and Freeman2011], and Watson [Reference Watson, Sobel and Wall2009]). However, the main substance of my argument does not turn on this assumption. Those who believe that morally impermissible promises fail to count as genuine promises will be interested in a slightly different version of my question: namely, what conditions must be met—and in particular, how much uncertainty can there be and from what source—for a purported promise to count as a genuine, morally binding promise?
The conditions I propose will presumably be the same according to either version of the question. For example, suppose Benedick promises Beatrice that he will kill Claudio. We can interpret the immorality of the promised end as either entailing that Benedick's promise is morally impermissible to make or as entailing that his misguided attempt at promising fails to generate a valid promise. The same is true of a deceptive, insincere, or unrealistic promise; the deception, insincerity, or lack of realism entails either that making the promise is morally impermissible or that the attempt at promising misfires. While I am focusing on permissibility rather than validity conditions in this paper, independent theoretical considerations that are beyond my present scope will determine which version of the question we should ultimately adopt.
Crucially, in articulating my four conditions on permissible promise-making, I am not attempting to offer criteria that will by themselves always determine whether making a promise is immoral all things considered. Rather, I am offering necessary conditions on when making a promise is morally permissible qua promise. (Those who wish to understand my proposal in terms of validity conditions should think of me as offering necessary conditions on when a promissory offer succeeds at generating a promise.) For there are cases in which the best thing to do all things considered is to make a deceptive promise—say, if a contract killer appears at the door asking about your neighbor's whereabouts, and the only way to make him leave is to promise insincerely that you will call him as soon as the neighbor returns home. In virtue of your deception, the promise is morally problematic to make (or is invalid) under my proposed criteria. But circumstances entail that making (or attempting to make) the promise is still what ought to be done overall. In such cases, the moral reasons to make the promise have to do with avoiding serious harm to another and have nothing to do with the promise itself or the conventions governing permissible promise-making. This distinction is analogous to the distinction between the ‘right kind’ vs. ‘wrong kind’ of normative reasons—e.g., admiring Anne because she is admirable vs. admiring Diego because an evil demon threatened to destroy you if you did not do so (see Gertken and Kiesewetter [Reference Gertken and Kiesewetter2017] for an overview of this literature).
2. Morally Permissible Content
A promise is morally permissible to make only if both the promised action and the means the promisor plans to take to achieve that action are themselves morally permissible. It would be impermissible for your accountant to promise to help you cheat on your taxes or for a political candidate to promise to refuse entry to immigrants of a particular religion because these ends are themselves morally impermissible. This is because promising to Φ incurs a moral obligation to Φ. And morally obligating yourself to do something immoral is perverse—consider someone who joins the army because he believes that he will be morally obligated to follow orders and hopes that he will be given immoral orders that he will then be morally obligated to carry out.
The claim that permissible promises must be for morally permissible ends is not affected by whether promises for immoral ends are better classified as impermissible promises or as impermissible attempts at promising that fail to count as valid promises. For to attempt to incur an obligation to do something immoral is equally problematic; the soldier who joins the military with a perverse desire to be obligated to do unethical things acts immorally even if he is mistaken about what obligations he will in fact incur.
Moreover, promising to Φ usually makes it more likely that you will Φ because we are generally motivated to keep our promises. Promising to attain an immoral end X is an attempt at self-binding that makes it more likely that you will do X. And it is wrong to attempt deliberately to increase the likelihood that you will behave immorally; consider someone who knows that she becomes violent when she drinks too much and overindulges on purpose to increase the chance that she will get into a fight.
3. Lack of Deception
A promise is impermissible to make if making it involves leading others astray in their beliefs. I am presuming that (all else being equal) it is morally impermissible to allow or actively lead people to form false beliefs, especially about matters that are important to them or relevant to their interests. Deception is most likely to occur when the promisors do not intend to do what they promise to do; call this promising insincerely. Whether this counts as deception in a strict sense is not important; what matters is that acting in this way is generally morally impermissible. This is because in most cases, promising to Φ leads the promisee to believe that the promisor is going to Φ. If the promisor does not intend to Φ, the promisee will be likely to form a false belief about the promisor's future actions; a candidate for student council president who insincerely promises to give everyone free pizza to secure the freshman vote will lead the other students to form the false belief that they are going to get free pizza. To avoid deceiving promisees in this way, promisors must actually intend to do what they promise to do.
In claiming that insincere promises typically lead their promisees to form false beliefs, I am simply making a claim about the effect of making promises in most cases; I do not mean to be presuming as expectationalist theorists (such as those of T. M. Scanlon [Reference Scanlon1990 and Reference Scanlon1998:ch. 7] and others) do that the normative force of promising stems from the fact that promises raise expectations in the promisee. If an insincere promise is made without raising the promisee's expectations—say, because the promisor is notoriously untrustworthy, and the promisee is rightly skeptical that he will follow through—this will not in fact deceive the promise, but the promisor may still incur a promissory obligation and may nevertheless be criticizable for attempting to deceive.
4. Promising in Good Faith
In addition to refraining from making insincere and deceptive promises that they have no intention of keeping, promisors must also be sure to have a sufficiently strong commitment to acting as promised. I am speaking of commitment in the subjective attitudinal or volitional sense that involves actually caring about acting (e.g., Alice is committed to working for social justice and so joins grassroots organizations and strives for change), rather than in the normative sense, which involves being such that you ought to act or believe in some way (e.g., philosopher X's theory commits her to implausible conclusions, regardless of whether she recognizes this; see Shpall [Reference Shpall2014] for an account of these two types of commitment). Call a promise that displays an adequate level of volitional commitment one that is made in good faith.
Promising in good faith involves undertaking a robust and counterfactually resilient volitional commitment to acting. For example, suppose you promise me that you will attend my party on Friday, as I am anxious that no one will show up. You assume there are no other parties planned for Friday, and as of now you intend to go. But you recognize that you will probably bail on me if you hear about a cooler party happening elsewhere. Your promise is not in good faith. Because you promised to go, you should be committed to going to my party even if another option you prefer arises, or you no longer feel like going, or your original motivation for making the promise has since lessened or even disappeared.
Promising without a sufficiently strong degree of commitment often deceives the promisee, as the promisee is likely to believe (falsely) that the promisor is robustly committed to acting as promised. But promises that are not in good faith are problematic for further, independent reasons as well. First, promisors who are insufficiently committed are less likely to follow through and are thereby more likely to break their promises and harm their promisees. While you might attend my party, you are less likely to do so than you would be had you made the promise with a stronger degree of commitment. People who promise without good faith act in a way that makes it more likely that two kinds of bad outcome will occur: (1) that a promise will be broken, which might be bad because it harms the promisee, or (assuming keeping promises is intrinsically valuable) bad simply qua instance of promise-breaking, and (2) that the promisor will fail to satisfy her moral obligations. And in general, it is morally impermissible to exacerbate the risk of causing harm to others unnecessarily, causing intrinsically bad things to occur, or failing to satisfy your obligations.
Second, promises that lack good faith display a morally problematic attitude involving a lack of adequate respect and moral consideration for the promisee and her interests. This is so regardless of whether the promisor's lack of commitment in fact deceives the promisee or makes promise-keeping less likely. When you make a promise, you form a special moral relationship with the promisee, incurring a directed obligation that you owe to the promisee to act as you have promised. Making a promise without being deeply committed to satisfying this special directed obligation shows that you do not take the formation of this moral relationship seriously, which is disrespectful to the promisee.
What exactly this failure to respect the promisee consists in will vary depending on what normative theory of promissory obligation you prefer. For example, if what matters in promise-making is inviting the promisee to trust you as Daniel Friedrich and Nicholas Southwood (Reference Friedrich, Southwood and Sheinman2011) suggest, then failing to be sufficiently committed to your promise is to fail to value properly the trust that the promisee places in you. If what matters is instead a transfer of normative authority (e.g., to decide what party you attend tonight) from promisor to promisee as David Owens (Reference Owens2012) argues, then failing to be sufficiently committed to your promise is to fail to respect properly the authority that you have granted to the promisee. And if what matters for promise-making is adhering to valuable social conventions as John Rawls (Reference Rawls1971) claims, then failing to be sufficiently committed to your promise is to fail to take seriously your role in playing fair and helping to maintain a valuable social convention. Other theories of promissory obligation will presumably involve some other kind of lack of proper respect for the promisee. In any case, making an overly casual promise to attend my party is morally problematic because it fails to take my interests as seriously as you should.
5. Realistic Self-Assessment
One can be (irrationally) volitionally committed to something believed to be impossible; the Don Quixote of the musical Man of La Mancha is explicitly committed to dreaming an admittedly impossible dream. Less starkly, one can have an unrealistically optimistic degree of volitional commitment: to commit oneself fully to, say, a career as a professional athlete or movie star in spite of astronomical odds. The volitional commitment involved in permissible promising must not be unrealistic in this sort of way. A promise is permissible to make only if the promisor blamelessly believes that she will be able to carry out the promised action. This constraint requires that the promisor have a reasonable conception of what she is likely to be able to achieve (e.g., it would not be realistic to promise one hundred students that you will grade all of their papers in a single day, as this would require nearly super-human speed and fortitude). It does not require that the promised action is in fact attainable for her (e.g., it would be realistic to promise ten students that you will grade their papers in a day even though receiving unexpected bad news might sap you of the motivation and focus necessary for doing so). Nor does the constraint require that we know ourselves and our own abilities perfectly, or that we reason infallibly about what is realistic for us (e.g., it would be realistic to expect that you will be able to grade ten papers today because you did so yesterday even though you would not have that expectation if you knew yourself well enough to recognize that you tend to be unfocused on rainy days like today).
To make a promise with an unrealistic self-conception involves deliberately incurring a moral obligation to perform an action that you are very unlikely to achieve. This sets yourself up for moral failure in an even more stark way than does making an insincere promise; an insincere promisor might still fulfill her promissory obligation, while an unrealistic promisor almost never will. Like insincere or insufficiently sincere promises, such promises are also likely to be deceptive; the promisee is entitled to believe that the promisor can and will Φ when this has little chance of happening. And even if they are not deceptive, unrealistic promises are like insincere or insufficiently sincere promises in that they display a lack of respect for the promisee and her interests; they deliberately create a directed obligation that likely cannot be fulfilled.
At one extreme are promises that violate the basic presumption that ‘ought implies can’ and that are completely beyond your abilities—say, promising your romantic partner that you will (literally) give them the (actual) moon. But promises can be unrealistic even if they can be achieved, but it is exceptionally unlikely—say, promising your romantic partner that you will never have even a passing sexual thought about another person for the rest of your life. Another way in which a promise can be unrealistic is if it involves an overly optimistic or quixotic self-assessment—say, a US presidential candidate with no political experience promising to renegotiate the US/Iran nuclear deal unilaterally.
Quixotic, overly optimistic commitments might seem permissible (or even admirable) when they concern only an agent's own actions; consider someone who vows to complete a marathon despite recognizing that she might give up if she becomes too tired. We might admire the pluck and dedication of such people and how they attempt to control their own destinies against all odds. However, such undue optimism is not warranted when making promises, which do not concern only your own plans but involve actions whose performance you owe to others. It is disrespectful to promise your running partner that you will finish the marathon together if you believe there is a significant chance that you will give up if you get too worn out before the race ends—at least, assuming that the marathon promise does not involve what David Owens (Reference Owens2008) calls a ‘prophylactic promise’, or a promise the making of which increases the likelihood that the promisor is able to attain the promised action (i.e., the motivation to not let down your running partner causes you to be able to complete the marathon, when you would not have this additional motivation without such a promise). Making a quixotic marathon promise involves knowingly setting yourself up to fail to give your running partner what you owe her. As Jason D'Cruz and Justin Kalef note, ‘there is nothing to reproach in striving or intending to do something that one worries may be beyond one's powers, but the situation is different when one moves beyond this to promising’ (Reference D'Cruz and Kalef2015: 8).
The higher the stakes, the more important it is to ensure that your commitment is grounded in realistic beliefs about your own abilities. It might not matter very much if your running partner does not complete the marathon with you. But it matters a lot more if, say, your business partner does not file essential paperwork by an important deadline. And the stakes are high when it comes to campaign promises, which often concern policies that affect some of the most important things in our lives, including our jobs, finances, national security, rights, and civil liberties. Political promises grounded in implausible or unrealistic self-conceptions—say, a mediocre businessperson who falsely believes himself to be an incredibly good negotiator—can potentially do a lot of damage.
6. Campaign Promises
So far, I have argued that permissible promises must (at least) involve morally permissible content, be nondeceptive, be made in good faith, and involve realistic assessment of the possibilities. In the rest of this paper, I will assess whether these four conditions can be met in conditions of uncertainty. First, though, I motivate my focus on uncertain campaign promises. In this section, I sketch how political campaign pledges are subject to uncertainty and are plausibly construed as promises. In the next section, I explain why appealing to promising to try or promising conditionally when making a promise under conditions of uncertainty will not completely eliminate our need to understand when uncertain unconditional promises to act are permissible.
We often make (or are asked to make) uncertain promises in our everyday lives. Think of a patient imploring a surgeon to promise that the surgery will be successful when the surgeon knows that there is always some risk involved, or of a soldier heading off to war whose parents ask her to promise that she will return. More prosaically, think of promising an editor that you will have a paper finished by next week while worrying that you will not be focused enough to finish it, or promising your mother that you will not get into political debates with your cantankerous uncle at your holiday dinner while knowing that you might not be able to hold yourself back.
Campaign promises are especially likely to be subject to uncertainty. In political systems that involve any sharing or separation of powers, many of the complex goals that politicians promise to attain—such as reducing the federal deficit or pulling the military out of an ongoing conflict—cannot be accomplished unilaterally. Rather, they are tasks that require a great deal of cooperation and compromise from a wide range of people, many of whom have fundamentally different values and priorities, and some of whom are deliberately antagonistic toward each other. Longer-term political goals—such as plans to reduce carbon emissions over a thirty-year period—are even trickier as they also require cooperation from future politicians who have not yet been elected. Unexpected circumstances can also arise (e.g., natural disasters) that make goals that once seemed easily attainable hard to achieve. And even when political goals are attainable, politicians are often faced with massive reelection pressures, which might involve pandering to donors or special interest groups in ways that prevent their promises from being kept. Getting elected in the first place requires a willingness to play the political game, which means that any candidate who successfully gets elected will likely be a good political game player. And good political game players tend to do what it takes to remain in office, even at the cost of abandoning their campaign promises. The temptations faced by politicians to break their word are numerous and deep, and their track records at standing firm against such temptations are not great. If uncertain promises are problematic, then most political promises are in trouble.
I am focusing on the permissibility of making campaign promises in particular because they are a rich example of promising under uncertainty that we have all encountered before and are likely to continue to encounter. But the claims I make are meant to apply to any case of promising under uncertainty, and readers who are unconvinced that campaign pledges are promises can apply my analysis to uncertain promises of other sorts. For this reason, I am abstracting away from many important issues particular to campaign promises that are beyond my present scope, including the question of whether elected officials are trustees or delegates, more general issues about the ethics of electoral representation, and complex and interesting questions about the details of how to cash out the claim that campaign pledges are promises.
I will instead sketch some general reasons to think that campaign pledges can indeed be construed as promises. In many democratic political systems, there is a strong prima facie presumption that campaign pledges are promises: candidates in such systems typically take themselves to be making promises on the campaign trail, and those who fail to follow through are routinely accused by the public and the media of having broken promises. This presumption might not hold in democratic systems with extremely high levels of voter cynicism (perhaps fueled by extremely high levels of corruption). But see, for example, how the website Politifact keeps detailed statistics about what US campaign pledges (construed as promises) have been kept, broken, or compromised on (http://www.politifact.com/truth-o-meter/promises/). If a strong prima facie presumption can be easily accommodated theoretically, we should try to preserve it. Accordingly, I am presuming that at least some campaign pledges are best construed as promises and that citizens are the promisees who accept the candidate's promissory offer by voting for the candidate.
There are multiple ways to cash out this idea; so long as one of them works, we can construe campaign pledges as promises. For example, it could be that the politician makes promissory offers to the electorate as a whole that count as accepted if the politician is elected. On this model, one individual vote for a politician (whether on the basis of one campaign pledge in particular or multiple campaign pledges) does not constitute acceptance of the candidate's promise(s). Rather, all of the voters who vote for the candidate contribute to the collective acceptance of all of the candidate's promises through election. Or it could be that the politician makes promissory offers to each individual citizen, conditional on being elected. Each individual vote counts as an acceptance of the offers; the condition is met and promissory obligations are created only if the candidate is elected. On either model, if voters change their mind at a later date about whether they want the promised actions to occur and wish to release a politician from her promises, they can do so by refraining from reelecting her or by organizing a recall.
The details of how we work out the idea do not matter, as my account of uncertain promises should be compatible with any view of what campaign promises consist in. What is more important is the fact that my analysis of promising under uncertainty will be broadly applicable to other cases, even if it turns out that campaign pledges should not be counted among the class of genuine promises. For example, another common class of uncertain promises are marriage promises of the sort made at traditional weddings; in Liberman (Reference Libermann.d.), I offer an account of when it is permissible to make wedding promises you are not certain you can keep, and when such promises are subject to excusing conditions.
7. Alternatives to Uncertain Promises?
One might worry that promising to act under uncertainty is never appropriate because there are other kinds of promises that we ought to make instead when faced with uncertainty. In the rest of this section, I explain why this is not the case. Although promises to try or conditional promises are sometimes morally appropriate, they will not always be so. Moreover, actual politicians routinely make unconditional uncertain promises to act, which makes it worth investigating whether and when such promises are permissible.
7.1. Promising to try?
Perhaps we should never promise to attain ends we are not certain we can achieve, but we should instead merely promise to try to attain them. It might seem unproblematic for the doctor to promise to try to carry out the surgery successfully, for the child to try to come home from the war, or for the politician to try to reduce the federal deficit even though it does seem problematic for them to promise to attain these uncertain goals. However, there are other cases where promising to try seems much more suspect, such as promising your partner that you will try to be faithful, or try to be home in time for dinner. This is because, as Berislav Marušić argues, promises to try to Φ when Φing is entirely up to us are ‘always wrong, because they hide a possible choice under the veil of our susceptibility to circumstances beyond our control’ (Reference Marušić2016: 2). To promise to try to Φ is to behave as if Φing is not up to you when in fact it is and to falsely represent your Φing as such to the promisee.
Marušić also argues that promises to try to Φ can be problematic in some cases even when Φing is not entirely up to us because it is unclear what successfully trying consists in. This lack of clarity enables promisors to ‘hide a possible choice under the veil of our susceptibility to circumstances beyond our control’, (8) such that ‘someone who merely promises to try to come to your party, without specifying the conditions under which she would fail to come, in effect reserves the right to choose not to come’. (9). This worry is especially salient for campaign promises. Voters want to be able to assess whether such promises have been kept; if the success conditions on a promise to try are not obvious, voters will not be able to engage in such assessment.
This is not to say that promising to try is never permissible. As D'Cruz and Kalef note, ‘responsible promisers typically withdraw from promising to act, and instead promise to try, in circumstances in which they recognize that there is a significant chance that they will not succeed’ (Reference D'Cruz and Kalef2015: 799). Promises to try to succeed in surgery, return from the war, or reduce the federal deficit are presumably promises of this sort—which means that promising to try to act will sometimes be an acceptable alternative to promising to act under uncertainty. All I am presuming here is that promising to try will not always be such an alternative. This will sometimes be the case when Φing is not entirely up to you—that is, when your doubts about your ability to Φ stem from doubts about whether the circumstances will permit Φing, or from what I call external uncertainty. And it will always be the case when Φing is entirely up to you—that is, when your doubts about your ability to Φ stem from doubts about your own strength of will, or from what I call internal uncertainty.
7.2. Promising conditionally?
It is widely assumed that promises are defeasible obligations, meaning that promising to Φ obligates you to do so unless you have a very good excuse, such as a more important conflicting obligation unexpectedly arising, discovering that your promise is impossible to keep, or realizing that it was premised on false information (see Liberman [Reference Liberman2015: ch. 7] for more on the conditions that count as legitimate excuses for promise-breaking). When an excusing condition arises, the promisor is not obligated all-things-considered to act as promised. For example, suppose a political candidate promises to give a speech at a rally, but while driving to the rally encounters an injured cyclist in desperate need of assistance that only she can provide. All things considered, the candidate should break the promise to give a speech and take the cyclist to the hospital instead. Or suppose the candidate is suddenly stricken with laryngitis and loses her voice before the rally. All else is not equal in this case, and she is not morally required to make the speech. One way to understand the defeasibility of promissory obligations is to think of all promises as being conditional in the following way: the agent promises to Φ so long as a legitimate excusing condition does not arise.
Sometimes, the content of a promise is conditional in a further and more particular way. For example, I might promise to take you on a picnic only if the weather is sunny. In this case, my obligation to take you on a picnic no longer holds if one of the typical legitimate excusing conditions arises or if the weather is cloudy. In most cases, cloudy weather would not license my breaking a promise to take you on a picnic—but it does if the promise is conditional on sunny weather.
One might think that when faced with uncertainty, we should not promise to act outright but should instead make an explicitly conditional promise to act, so long as the thing about which we are uncertain does not occur. For example, a politician might promise to refuse campaign contributions from major corporate donors, on the condition that her reelection is viable without this funding. For this promise is not uncertain; she is convinced that she will be able to resist the allure of corporate funding if accepting such funding is not essential for reelection. Conditional promises are sometimes appropriate, such as the promise to take you on a picnic on the condition that the sun is shining. It makes sense to constrain the range of scenarios under which I am obligated to keep my promise to take you on a picnic, especially because a sunny picnic is more pleasant than a cloudy one.
However, this strategy will not succeed across the board. First, it will not always be possible to make a promise that perfectly captures the necessary conditions. For not every case of uncertainty is one in which the promisor can predict the particular circumstances in which the promise will not be kept. Maybe the candidate's ability to refuse corporate funding is uncertain for multiple reasons, some of which she cannot easily predict in advance—which means that there is nothing that her promise can successfully be conditional on. Marušić (Reference Marušić2016: 18–19) frames a similar thought in this way: the conditions on a conditional promise will either rule out the most serious forms of temptation, or they will not. If the conditions exclude all serious sources of temptation, the promise lacks sufficient commitment. If the conditions do not exclude all serious sources of temptation, the promise remains subject to uncertainty. In neither case does the conditional promise get us what we want. (Note that Marušić frames this discussion not in terms of uncertainty, but as part of an argument for the rational permissibility of promising and resolving against the evidence in at least some cases; he argues that appealing to promising to try [or promising conditionally] does not undermine the significance of the cases of promising against the evidence that he discusses.)
Crucially, when the condition on a conditional promise is not met, there is no promissory obligation in place whatsoever. If the sky is cloudy, I am under no promissory obligation whatsoever to take you on a picnic—I do not need to consider taking you on a picnic in my deliberations about what to do, and I do not owe you any reparation or apology if I fail to take you on a picnic. Similarly, if the candidate promises to refuse corporate funding on the condition that reelection is viable without it and this condition is not met, no promissory obligation is in place. The candidate does not have to consider the promise to refuse funding when deliberating about what to do, nor is any sort of apology owed to voters for accepting corporate funds.
This is clearly not the sort of promise that most voters are looking for; it does not assure robustly enough and does not display an appropriate concern for the promisee's interests. Voters generally do not want candidates to be obligated to act as promised only on the condition that doing so is not too difficult. To the contrary, we often assume that campaign promises are meant to lead politicians to action in spite of these sorts of barriers. Conditional promises are not always an adequate substitute for promising to act under uncertainty, especially in political cases where voters seek a more robust form of commitment and assurance. Since political candidates frequently do make unconditional promises to act, and since we cannot completely avoid the need to make such promises by promising to try or promising conditionally instead, we should consider whether uncertain promises can meet the four criteria for permissible promise-making.
8. Morally Permissible Ends and Lack of Deception Under Uncertainty
Uncertain promises can easily meet the first criterion for permissible promise-making so long as they involve only morally permissible content. However, they risk violating the second criterion, which is lack of deception. In most cases, promising to Φ licenses the promisee in forming a belief that you are assuredly going to Φ. And in typical circumstances, the promisee actually will form such a belief, and will structure her future plans around your Φing. But if your Φing is uncertain, then you are not assuredly going to Φ, and the promisee is liable to form a false belief about the certainty of your Φing.
Uncertain promisors can avoid being deceptive in two ways. First, it might be mutually presumed between promisor and promisee that there is uncertainty involved. For example, suppose the government is debating building a highway, and it is uncertain whether its construction will be approved and what route it will take if so. One potential route runs through Old MacDonald's farm. If this route is approved, the government will use eminent domain to claim the land. MacDonald promises his daughter that he will give her the farm to run within the next few years. They both know that the government might seize the land before then. Since his daughter knows that there is some chance that her father will be unable to keep his promise, MacDonald's promise does not deceive her.
If campaign promises are to be nondeceptive in the same way, it must be generally presumed by the populace that all campaign promises are at least somewhat uncertain even if they are framed with certainty for rhetorical purposes. This might be plausible in some political environments. Only very naïve voters assume that all politicians will be able to keep all their promises. Perhaps we share an awareness that this is how the political game is played: that, for example, a presidential candidate's promising to waive student debt does not imply that she will certainly attain this goal, but only that she will remain fully committed to doing so and will achieve this goal if enough other legislators cooperate. If so, then a candidate's uncertain campaign promise to eliminate student debt need not lead voters to believe that debt will certainly be eliminated and will accordingly not be deceptive.
Alternatively, promisors can avoid being deceptive by being explicit about the uncertainty involved in individual cases. If the potential plans for a highway are not common knowledge, MacDonald must clearly inform his daughter that there is a chance that the farm might be seized before he can give it to her, lest she falsely believe that her future ownership of the farm is certain. Likewise, if there is not a general presumption that all campaign promises are at least somewhat uncertain, candidates will have to be explicit about the uncertainty involved in their promise-making. They could do this by highlighting the fact that all campaign promises are uncertain to some extent or by noting that the particular goals they are proposing are uncertain. For example, a presidential candidate might proclaim that he will enforce mass deportations, but explicitly note that the success of this project will depend on the cooperation of the legislature, immigration enforcement agencies, and state and federal officials. A qualified promise of this sort will not lead voters to believe that mass deportations will assuredly occur and will therefore not be deceptive. Of course, candidates tend not to hedge their campaign promises in this way. But if our political system is not one in which campaign promises are generally presumed to be uncertain, then they will have to hedge their uncertain promises accordingly if they are to be nondeceptive.
9. Promising in Good Faith Under Uncertainty
Is it possible to have a sufficiently strong degree of volitional commitment when promising to do something uncertain? Below, I argue it is easy to do so when the source of the uncertainty is external, and the promisor worries that the promise will be impossible to keep for reasons beyond her control. But when the source of the uncertainty is internal to the promisor—such as doubts about the promisor's own strength of will or susceptibility to temptation—promising in good faith is not possible. It follows that the extent to which campaign promises are permissible or problematic will depend on the extent to which they are subject to internal uncertainty.
9.1. Promising in good faith under external uncertainty
Consider the political candidate who promises to give a speech at a rally. Her promise is made in good faith if she is strongly committed to giving the speech. Suppose a legitimate excuse unexpectedly arises—she gets laryngitis or needs to rescue a stranded bicyclist, making keeping the promise impossible or no longer required, respectively. The occurrence of an excusing condition does not undermine or negate the candidate's initial volitional commitment in any way. In general, the strength of a promisor's volitional commitment in making the promise is not altered if an external barrier unexpectedly arises that prevents a promise from being kept or that makes it the case that keeping it is not morally obligatory, all things considered.
Externally uncertain promisors are in the same sort of situation as the candidate, except they foresee that excuses are somewhat likely to arise when they first make their promises. MacDonald's promise is uncertain because he thinks that the government might seize his land, thereby making his promise impossible to keep. But this in no way prevents him from being strongly volitionally committed to giving his daughter the farm at the time that he makes the promise. Ordinary promisors grant that an excuse might unexpectedly arise because life can be unpredictable, although they do not believe that this is likely. Externally uncertain promisors are in a similar position: they grant that an excuse might arise that makes it the case that they are not obligated to keep their promise, and they recognize that this is somewhat likely. This recognition does not alter the fact that such a promise can be made in good faith, and that such a promisor can be fully committed to acting so long as doing so remains possible and morally required. Accordingly, externally uncertain promises can be permissible to make, so long as the promisor is not deceptive about the heightened risk of excusing circumstances that would permissibly lead to the promise not being kept.
9.2. Promising in good faith under internal uncertainty
Some promises are uncertain not because the promisor believes that there is a heightened chance of an excusing condition, but because the promisor believes that there is a significant chance that she will fail to Φ as promised due to inappropriately succumbing to some form of temptation to not Φ. I am not suggesting that miniscule doubts about one's abilities suffice to make a promise internally uncertain; I can permissibly promise to pick you up at the airport even though I recognize that there is an extremely small chance that the stress of driving around a crowded airport at rush hour will cause even a confident and skilled driver like myself to have an emotional breakdown that leaves me unable to drive. But if I am brand-new to city driving and I doubt my ability to navigate a crowded airport, or if I have had breakdowns during stressful driving experiences in the past, my promise is internally uncertain.
How likely my expected failure must be for the promise to count as internally uncertain will vary and will depend on a holistic assessment of how important or significant the promise is. The most obvious factor that affects the importance of a promise is how much it matters to the promisee (at least, within reasonable limits). In general, the more important the promise is to the promisee, the less internal uncertainty is acceptable. Since the promisor cannot always or infallibly know how much the promise matters to the promisee, the determination of the importance threshold may have to involve the promisor's best educated guess.
The importance of the promise to the promisee is determined by a number of factors, including: To what extent is the promisee relying on the promisor to perform the promised action? Has the promisee sunk any costs in expectation of the promise being fulfilled? If the promise is part of a reciprocal bargain, has the promisee already performed her share? To what degree, if any, does the promisee care whether the promise is kept apart from considerations about cost and reliance? How close or morally significant is the relationship between the promisor and the promisee? While the promisee's interests are the primary determinant of importance, promisors can also imbue a promise with extra solemnity or significance by making it in a formal or public setting (e.g., a wedding promise or a campaign pledge). Especially solemn promises will likely permit of less uncertainty than do more casual promises.
In some cases of internally uncertain promising, the promisor believes that she will remain strongly volitionally committed to Φing, yet might akratically fail to Φ. In other cases, the promisor fears that she might be weak-willed in Richard Holton's (Reference Holton2009) sense and that temptation will inappropriately cause her to revise her intention to act as promised. Either option is compatible with my claims about internally uncertain promises. For ease of discussion, I will refer to inappropriate succumbing to temptation as irresolution. Not every failure to keep a promise without an external excusing condition is an irresolute failure. For example, suppose candidate A promises to Φ, then forms the reasonable belief that ~Φing is a much better option. Accordingly, she breaks her promise to Φ. This is not an instance of internal uncertainty, but of updating one's plans after a change in belief. Whether this is what the candidate ought to do all things considered depends on whether ~Φing is in fact a better policy.
In general, though, irresolution is not an excusing condition on promises in the way that impossibility or the presence of a more important conflicting obligation is. The candidate who fails to keep a promise to give a speech because she has laryngitis is off the moral hook. The candidate who fails to keep a promise to give a speech because she is exhausted and decides to take a nap instead is not. To expect that you might be irresolute is not to expect that you might not be morally required all-things-considered to keep the promise, after all. Rather, it is to expect that you will remain morally required to keep the promise, but will impermissibly fail to do so. This presumes that agents are responsible for failures stemming from irresolution—perhaps because they were capable of acting but did not or because they were responsible for their inability to act. Uncertainty about one's own physical or mental capabilities will count as internal or external depending on whether these capabilities are within one's control.
For example, suppose the source of MacDonald's uncertainty about whether he will successfully give his daughter the farm is internal rather than external. MacDonald believes that his daughter would do a better job of running the farm than his son would, and so he promises the farm to her. However, the sexist belief that farming is men's work is prevalent in his community, and his buddies consistently give him a hard time about entrusting the farm to his daughter. MacDonald knows himself well enough to recognize that he is somewhat susceptible to peer pressure; he realizes that there is a good chance that he will succumb to the jeers of his buddies and give the farm to his son. Can MacDonald nevertheless promise in good faith to give the farm to his daughter?
This seems unlikely, as promising to Φ at time t while being uncertain about whether you will be resolute enough to do so is often good evidence that your degree of volitional commitment to Φing is not strong enough right now or that you predict that it will not be strong enough at t. Because it is possible for MacDonald to give his daughter the farm, the fact that he is worried about his ability to stand up to his buddies is a sign that he is insufficiently committed. Were he more strongly committed to giving the farm to his daughter, he would not be so worried that the sexist jeers of his friends might deter him. This is not so with external uncertainty; the fact that MacDonald recognizes that the government might claim his land and prevent him from giving it away does not tell us anything about how strongly committed MacDonald is to giving his daughter the farm.
Campaign promises in particular seem inappropriate to make under internal uncertainty because of their reciprocal nature. Campaign promises function as a bargain: candidate X promises that she will act, voters accept this promise by electing X, and X must act as promised to hold up her end. An internally uncertain candidate predicts that this bargain will remain in place, but that she might impermissibly fail to hold up her end of it. This is not a good faith bargain. Candidates subject to external uncertainty do not violate their bargains in the same way. A candidate who predicts that an obstructionary Congress might prevent her from fulfilling her promise is foreseeing a situation in which she is no longer required to hold up her end of the bargain because an excusing condition has arisen.
I have suggested that internally uncertain promisors cannot be in good faith. However, internally uncertain promises are impermissible to make even if internally uncertain promisors can—albeit irrationally—remain deeply and strongly committed to acting while predicting that they are likely to fail to do so. This is because an internally uncertain promise that meets the good faith condition will violate the realisticness constraint, as I will argue below.
10. Realistic Promising Under Uncertainty
Promises made under external uncertainty can easily meet the realisticness constraint. Promising to Φ when you suspect that you might be prevented from Φing by some outside source does not require forming unrealistically optimistic beliefs about yourself and your abilities. Nor are you setting yourself up for potential moral failure; you are simply making a promise while recognizing that a legitimate excuse is likelier than usual to arise.
Internally uncertain promises are different. Promising to Φ and remaining deeply committed to Φing when you believe that you might be too irresolute to Φ does involve an unrealistic self-assessment that sets you up for potential moral failure. There is something problematic about being strongly committed to, say, remaining faithful to your beloved while simultaneously believing that you are at least somewhat likely to succumb to the temptations of another lover. Holding both attitudes at the same time involves an irrational tension in your self-assessment that violates the realisticness constraint.
If promisors have inappropriately strong levels of commitment, their promises are not in bad faith, but they are unrealistic. For example, Odysseus is very strongly committed to avoiding the sirens now, but has been told that his commitment will waver when he hears their enchanting song in the future. For Odysseus to promise his crew that he will avoid the sirens in the future would be unrealistic and would involve a quixotic overestimation of his capacities. Indeed, he recognizes this and binds himself in a nonpromissory way by literally binding himself to his ship's mast instead.
A common source of overconfidence in our own abilities is the cognitive bias of illusory superiority, also known as the ‘better-than-average effect’: most of us tend to think that we are better than we really are. It therefore seems likely that candidates will believe themselves to be more resolute than they really are. They will accordingly feel certain in their abilities to keep their promises even in cases in which they are in fact susceptible to temptation. However, candidates who fall prey to superiority illusions should know better: it is possible for them to learn about common cognitive biases and to take steps to correct their errors. (Maybe politicians should have staffers give them regular reality checks?) Maintaining a superiority illusion in the face of contrary evidence—especially when the stakes are high—involves an unrealistic conception of oneself. The promises that flow from such illusions will not be internally uncertain, but they will violate the realisticness criterion and will accordingly be impermissible to make. If this is right, then even if an internally uncertain promisor could have a sufficiently strong volitional commitment to acting—contra what I argued above—the promise would nevertheless be impermissible to make because it would be unrealistic.
11. What Should Uncertain Promisors Do?
I have argued that internally uncertain promises are not permissible to make, as they always either fail to be in good faith or are unrealistic. I have also suggested that political promises are subject to a great deal of internal uncertainty. What, then, should uncertain promisors—political and otherwise—do?
The best thing people who are internally uncertain about their ability to keep their promises can do is use strategies to make themselves more certain in their ability to resist temptation. For example, you might ask a colleague to check in with you every day about your progress on the paper you have promised to finish. Or you could ask your mom if you can sit at the kid's table this year to make it easier to keep your promise to avoid fighting with your combative uncle at the holiday dinner. Politicians might engage in such strategies as well. For example, when Carmen Trutanich was campaigning for the position of city attorney of Los Angeles in 2008, he promised that if elected, he would refuse to run for a higher office while serving out his term. He pledged to take out a full-page ad in the L.A. Times proclaiming ‘I AM A LIAR’ if he broke this promise. Trutanich did not keep his promise; he ran for district attorney in 2012, never took out the ad, lost the DA race in a landslide, and was not reelected as city attorney. But Trutanich could have bolstered his resolve by giving the L.A. Times the ad money in 2008 and authorizing them to either run the ad if he ran for higher office or return the money if he completed his term. Candidates who took such preventative measures might be so motivated to avoid embarrassing newspaper ads that they would not succumb to the temptation to break their promises.
Of course, you will not always be able to use external means to make yourself certain of your ability to keep your promise. In such cases, you might be able to make a permissible promise by scaling back the content of what you are promising or by promising conditionally—say, promising your mother that you will not engage your uncle in political discussion unless he starts spouting conspiracy theories. But in other cases, this will not be possible; it will not help the editor if you promise to work hard on your paper or to finish the first section. If you are internally uncertain about whether you will Φ, and you can neither remove this uncertainty through external means nor promise to perform an alternate action that you are more certain about, then your promise to Φ is a morally impermissible one.
This conclusion leaves open the question of whether sometimes the moral considerations in favor of making a promise are so weighty that it is better all things considered to make a promise that lacks good faith or is unrealistic than it would be to refrain from promising. This might be true of campaign promises, which are morally valuable insofar as they predict candidate behavior and exert some pressure for politicians to remain accountable to voters. Were candidates not to make any promises and simply state party platforms instead, they might not be as personally invested in their campaign pledges and might lack incentives to follow through on them. Making an unrealistic or bad faith promise is bad, but being an inconsistent, unpredictable, or fickle political leader is worse. Perhaps the best thing that politicians can do in our current political system, full as it is of temptations that lead to internal uncertainty, is to recognize that their campaign promises are morally problematic to make qua promises—and then to make them anyway.
