1 Introduction

A widely accepted view in western philosophy is that moral obligations entail the ability to fulfill them, often glossed in the slogan that “ought implies can” (Copp 2008; Dahl 1974; Feldman 1986; Fischer 2003; Hare 1965; Howard-Snyder 2006; Moore 1922; Streumer 2003; Van Fraassen 1973). This longstanding principle can be traced from ancient philosophy (Cicero: “promises are not binding when the performance is impossible,” 1856) and modern philosophy (Kant: “duty commands nothing but what we can do,” 1793/1998) to contemporary normative ethics (Vranas: “an agent at a given time has an objective, pro tanto obligation to do only what the agent at that time has the ability and opportunity to do,” 2007: 167), and social political philosophy (O’Neill: “institutions and individuals can have obligations if but only if they have adequate capabilities to fulfill or discharge those obligations,” 2004: 251).

The principle is not only accepted but is also widely applied to debates across several areas of philosophy. For example, it is invoked in debates about whether agents are morally responsible for actions if they could not have done otherwise (Blum 2000; Copp 2008; Frankfurt 1969). It is invoked in debates about determinism and compatibilism in free will (Fischer 2003; van Inwagen 1983; Widerker 1991; Woolfolk et al. 2006). Some have argued that it is relevant to evaluating normative ethical theories that are suspected of ruling that people have certain obligations that they are incapable of fulfilling (Flanagan 1991; Mason 2003). It has also been invoked in discussions about the conceptual possibility of genuine moral dilemmas (Conee 1982; Jacquette 1991).

A central motivation for ought implies can is the assumption that it is intuitively correct (e.g. Moore 1922; Stocker 1971). However an impressive case has been made in both philosophy and cognitive science that ought implies can is not intuitive. In philosophy, several researchers have offered thought experiments and argued that it is intuitively correct to ascribe moral obligations to agents who lack the ability to fulfill them (Graham 2011; King 2014; Mizrahi 2009; Ryan 2003; Sinnott-Armstrong 1984; Spencer 2013). Similarly, in cognitive science, several independent teams of researchers have provided overwhelming scientific evidence that ought implies can is not an intuitive principle of moral psychology (Buckwalter and Turri 2014, 2015; Chituc et al. 2016; Mizrahi 2015b, a; Turri 2017b, though for further discussion also see Cohen 2018; Kissinger-Knox et al. 2018). Across different experimental paradigms, stimuli, and probing methods, these researchers consistently find that subjects systematically ascribe moral requirements in tandem with inability (for reviews, see Buckwalter 2017; Buckwalter and Turri 2019; Henne et al. 2016). Researchers have also provided several hypotheses about why the principle might have appeared intuitive due to conceptual confusion and bias (Buckwalter and Turri 2015; Turri and Blouw 2015). These findings strongly suggest that while ability may be related to moral obligation in important ways, it is not intuitively entailed by it.

The goal of this paper is not to discuss the significance of intuition or experimental observation in rejecting philosophical theories, which has been extensively examined elsewhere (Buckwalter 2014; Sosa 2007; Stich and Tobia 2016; Tobia et al. 2013; Turri 2016; Williamson 2016). Rather, the goal of this paper is to highlight the antecedent theoretical reasons for rejecting ought implies can absent the motivation that it is intuitive. That is, moving beyond counterexamples from individual intuitive thought experiments or from experimental cognitive science, we might instead focus on the principle as a broader theoretical construct and investigate it as a matter of theory choice. Conceptualized that way, we can ask the following question: is there sufficient theoretical motivation to accept this principle of morality, or are there strong antecedent reasons to reject the principle?

This paper presents seven observations across several areas of moral philosophy demonstrating that the theoretical motivation for ought implies can is weak. The principle is in tension with foundational areas of ethical theorizing, including research on apologies, excuses, promises, moral dilemmas, moral language, disability, and moral agency. Across each of these areas, accepting the principle that obligation entails ability creates a theoretical problem that is more easily solved by simply rejecting it rather than accepting it. Taken together, these observations question the theoretical motivation for the principle and strongly suggest that it is false. In what follows, I present seven observations of the theoretical tensions caused by the principle across these areas, and then, on the basis of this body of evidence as whole, argue that it undermines the theoretical motivation for ought implies can.

Before proceeding, three clarificatory points are needed regarding the moral principle under consideration. First, for present purposes ought implies can refers to the principle that moral obligation entails ability. This is separable from the evaluation of other possible relationships between ability and moral obligation, such as pragmatic implicature, and from other possible entailment relationships existing between ability and closely related but distinct moral concepts, such as desert or punishment. Second, it bears repeating that the present investigation involves whether one can ever hold a moral obligation to do something without the ability to do it, not whether one is blameworthy for failing to fulfill obligations when one is unable to. These are importantly separate questions because failing to fulfill a moral obligation does not entail blame (see below for discussion). Third, it also bears emphasizing the sheer strength of the principle under scrutiny. The principle does not say that ability is important for morality, or closely connected to moral obligation in many ways or across many circumstances, but in fact, posits the strongest relation between two concepts: an entailment relation. It is the theoretical motivation for entailment that is being evaluated.

2 Theoretical Motivation in Ethics

A first observation against the theoretical motivation for ought implies can is that it violates the logic of excuses and trivializes a large category of ethically meaningful apologies. We frequently apologize to others when we fail to fulfill our moral obligations and such apologies typically include excuses that feature facts about our inability to act. But the ubiquity of appeals to inability when offering excuses and apologies raises the following puzzle: if moral obligation entails ability, and an agent fails to do something because they are unable to act, then what exactly, in such circumstances does one need to apologize for or be morally excused for doing? Contrary to how it might seem, no obligation can exist in such circumstances if ought implies can is true, and hence, no behavior needs to be excused. If no behavior needs excused then no ethically meaningful apology can be made.

To illustrate this point, consider, for example, a case of a car accident that results in serious injury to others or even death, when a driver otherwise cautiously obeying all rules of the road swerves uncontrollably due to inclement weather. One would have thought that when this happens, one can and definitely should apologize for harming others. But if ought implies can is true, there simply cannot be an obligation not to harm others in these conditions that one needs to apologize for or be excused for doing. If that is the case, then apologizing for harm and death that you cause might serve an important social, legal, or performative function, but it is ethically trivial. Ought implies can renders these apologies trivial because they do not correspond to any actual moral obligation present. The argument goes as follows:

  1. 1.

    Many apologies are made for not doing something that is beyond an ability to do.

  2. 2.

    Ethically meaningful apologies require the existence of a moral obligation to do something.

  3. 3.

    No moral obligation to do something exists beyond an ability to do it.

  4. 4.

    Therefore many apologies are not ethically meaningful.

Ethicists may be willing to grant that some, perhaps even many, of our apologies are not ethically meaningful. But given the ubiquity of apologies and excuses that feature inability, the implications of this argument are noteworthy. According to ought implies can, there cannot be moral obligations to act when one is unable to do so, which means no moral transgression actually occurred. If no moral transgression actually occurred, no ethically meaningful apology can be made. If such apologies cannot be made, a large class of apologies, even those involving serious accidental harm or even death, are not ethically meaningful.

Rejecting ought implies can explains why we so commonly offer excuses and apologies that pertain to inability and ensures that such apologies can be ethically meaningful. Namely, one simple explanation for the facts that we genuinely do apologize for actions in these circumstances, and for why this large subset of apologies is ethically meaningful is that moral obligations can sometimes exist beyond ability. Indeed, a fairly natural reading of these apologies suggests that inability is frequently offered and accepted as a valid excuse for why an obligation in those circumstances was unfulfilled (Turri 2017b). According to this natural reading, if inability can generate an excuse for an unfulfilled moral obligation, then it must be possible for moral obligations to exist beyond ability.

A second observation is that accepting ought implies can risks not only trivializing apologies but also the significance of moral requirements entirely. Specifically, theorists have argued that if obligations were limited by our abilities to fulfill them, one could felicitously discharge moral requirements simply by being placed in situations in which their performance is impossible. This objection is illustrated in a classic example by Sinnott-Armstrong (1984: 252):

Suppose Adams promises at noon to meet Brown at 6:00 p.m. but then goes to a movie at 5:00 p.m. Adams knows that if he goes to the movie, he will not be able to meet Brown on time. But he goes anyway, simply because he wants to see the movie. The theater is 65 minutes from the meeting place, so by 5:00 it is too late for Adams to keep his promise.

If genuine obligations require ability, it is false that Adams is obligated to meet Brown as promised. His obligation to do so is discharged at 5:00 by his inability to meet Brown. The example is meant to illustrate how ought implies can trivializes moral obligation. Morally correct action, in this case, amounts to doing what you please as long as the result renders you unable to fulfill obligations. Of course, it is not impossible to try to defend ought implies can from the charge of trivializing morality. For example, theorists might attempt to concoct complicated new accounts of ability so the fact that Adams cannot reach Brown in time does not end up constituting an appropriate or morally relevant inability to act. However a simpler option is to retain a plausible and straightforward understanding of what it means to lack the ability to do something at a certain time and reject the theory that ought implies can.

A third observation involves promises and the fact that it is possible to make a genuine promise one cannot keep. This observation was concisely made by Julia Driver by considering the following triad of inconsistent propositions (1983: 221):

  1. 1.

    Whenever a person makes a promise to do x, he thereby puts himself under an obligation to do x.

  2. 2.

    If someone is obligated to do x, then he can do x (‘ought’ implies ‘can’).

  3. 3.

    Some people sometimes make promises they cannot keep.

In order to avoid a paradox we must reject one of these propositions. Given that we sometimes make promises we are unable to keep, the simplest solution is to deny the premise that obligation is incompatible with inability. Alternatively, accepting this premise demonstrates just how theoretically costly the ought implies can principle can be. In this case, accepting the principle requires us to accept that promises either do not oblige us to do things, or that it is always possible to keep a promise. These views of promises are highly contentious, though it is not impossible to defend them. For example, it may be possible to develop complex new theories in which making a promise does not actually place one under a moral obligation, or that somehow denies we make genuine or morally significant promises we end up being unable to fulfill. But, the far simpler option is to retain straightforward understandings of what it means to make and fulfill promises and reject the principle that ought implies can.

A fourth observation stems from the structure and affect of moral dilemmas. Physical and temporal constraints often force agents to choose between different courses of action. Nowhere is this more evident than in moral dilemma cases. A common feature of moral dilemmas is that agents must choose between one of two mutually exclusive actions. Imagine, for instance, a lifeguard who sees two drowning swimmers, but due to a vast physical distance between them, can only reach one swimmer in time. The lifeguard must choose which drowning swimmer to rescue at the expense of the other. On the theoretical level, cases like these are often used to evaluate ethical theories (see, for example Mill 1979; Ross 1930). On the personal level, they are perplexing and heart-wrenching. And no matter our choice, we come away from them with a feeling of moral failure that persists long after a decision is made (Williams and Atkinson 1965).

One natural explanation of these features is that moral dilemma cases pit two moral obligations against each other: satisfy one obligation to save one swimmer, at the expense of failing to fulfill the obligation of saving the other swimmer. But if moral obligation requires ability, this explanation is not available. There cannot be a moral obligation to save the other swimmer when it becomes physically impossible to do so. As soon as it becomes impossible, the moral demands to do otherwise vanish. In light of this, many philosophers have wondered whether genuine moral dilemmas could truly exist given the ought implies can principle, and sophisticated theories have been developed about how they could (compare to Lemmon 1962; Marcus 1980; Sinnott-Armstrong 1988; Van Fraassen 1973). Alternatively, though, a very simple explanation of why true moral dilemmas exist, are diagnostic in ethical theorizing, and why they move us so deeply is that moral obligations can sometimes persist despite inability.

A fifth observation comes from a class of obligations knowns as “role oughts”. Role oughts are obligations that are thought to exist in virtue of occupying a certain kind of personal or social role. For example, it might be thought that “teachers ought to explain things clearly” or that “parents ought to take care of their kids” (Feldman 2000: 676). It is also often thought that obligations that exist in virtue of such roles can sometimes persist when ability is absent. For example, it seems felicitous to say that incompetent teachers and incapable parents “may be unable to do what they ought”. In prior research, such observations have been used to argue that there are epistemic obligations to believe certain things in virtue of one’s role as a believer, even if one has no control over assuming the role or belief in question (ibid). The existence of this class of obligations also questions the theoretical motivation for ought implies can. If role oughts extend to the moral domain, as they seem to do with teachers and parents, then it is possible to hold a moral obligation when ability is absent. The existence of this category of obligation adds additional theoretical pressure to reject the principle. If there is a type of moral obligation that can exist without ability then it cannot be true that moral obligation entails ability.

A sixth observation comes from moral discourse and the propriety of certain complaints. For instance, we often blame others when we feel that they have played a role in preventing us from achieving what we judge we ought to. When this happens, it is perfectly felicitous to object to cases of action prevention by appealing to “moral obligation” and “moral oughts”. This is perhaps best illustrated in a classic case by Stocker (1971: 312):

Consider a man who correctly believes that for non- institutional reasons—e.g., simply to prevent needless injury—he ought to pick up the broken glass on a path taken by barefoot children. If some of the children lasso him and tie him to a tree, he can certainly complain that they are making it impossible for him to do what he ought to do (or ought to be doing).

Stocker observes that such statements are inconsistent with ought implies can because they imply the existence of obligations in the absence of ability. If ability limits obligation, then one cannot have an obligation that others are making it impossible to fulfill. In this case, though, sudden arboreal restraints make no difference to what speakers can or cannot say they “ought to do”. Theorists might attempt to find fault with these common and straightforward utterances or devise a complex way to interpret them non-literally. However, a much simpler explanation of these utterances is that ought implies can is false and that this explains why they are felicitous.

A seventh observation is that ought implies can may have troubling discriminatory, and particularly, ableist implications for moral philosophy. One form that ableism can take is to devalue or limit people with disabilities. The ought implies can principle limits the obligations a person can hold relative to their abilities. One aspect of personhood is moral status, and one measure of moral status is a person’s ability to hold moral obligations (for discussion, see Shoemaker 2009). Thus if a person lacks an ability to act, and holding an obligation is a measure of moral status, then according to ought implies can, that person is less of a moral agent then those who do not lack that ability. The result is that a person with a disability that prevents them from acting cannot have the same moral obligations, and subsequently, occupy the same moral status, as others in that same circumstance. A similar worry has been raised by researchers regarding various mental disorders, blameworthiness, and culpable ignorance, leading researchers to conclude that “a successful account of moral responsibility must include a broad range of moral agents, and that means taking seriously the moral agency of those with mental disorders” (Doucet and McChesney 2019: 228). A similar worry may also arise for moral obligations in conditions of oppression more generally. Assuming that these conditions limit what one is able to do, theorists have observed that “it seems implausible and disrespectful to insist that no one is ever responsible for wrongdoing under oppression” (Vargas 2018: 2). Related worries have also been raised for cases of oppression and abuse, namely that such moral assessments are “dangerously patronizing” toward the dignity and rational capacities of human beings (Smith 2008: 390, though see also Hutchison 2018). This is not to imply, of course, that agents are blameworthy or criticizable for actions in such conditions, which are separate questions that require distinct analysis. The present question is why such considerations should unjustly bar individuals from ever being able to hold moral obligations in the first place.

The argument is as follows. The capacity to hold obligations is an important aspect of moral agency and personhood. Ability is limited or absent in conditions of oppression and can be limited in cases of physical or mental disability. Moral obligations cannot exist to the extent that ability is limited. Therefore, the principle potentially limits the moral agency and personhood of the oppressed or disabled. This argument is troubling, for at least two reasons. First, it suggests that obligations can never extend to the oppressed or disabled in many cases as they do to others. Second, if the ability to stand in the moral obligation relation with one another, or occupying moral obligations is blocked, it potentially perpetuates oppression and ableism by falsely suggesting that such agents are not persons with moral standing or moral agency. These are clearly complex questions that require further research to fully address. For present purposes, however, these potentially troubling conclusions can be avoided by granting that it is at least sometimes possible for moral obligation to exist beyond ability and it is possible that the oppressed or disabled can occupy these relationships.

3 Theoretical Argument and Objections

These observations collectively challenge the theoretical motivation for the principle that ought implies can. Namely, if the principle is false, we can resolve several problems across foundational areas of ethical theorizing in an elegant and uniform way. It is possible to have an obligation without an ability to fulfill it, which is what explains why apologizing for things we were unable to do can be ethically meaningful. Inability does not always nullify obligation, which is why we cannot circumvent morality by merely manipulating situations or opportunities. It allows us to make genuine promises we are unable to keep. It straightforwardly explains the existence of, and appropriate affective response to, moral dilemmas. It coheres with a literal reading of straightforward “ought” statements. And, rejecting the principle avoids disrespectful and potentially discriminatory implications about moral agency and personhood.

Given that ought implies can is not a principle that several professional philosophers or laypeople find intuitive, it is sensible to examine whether there is independent theoretical motivation for the principle. Given its widespread application in several philosophical debates, one might have expected there to be strong independent theoretical motivation for the principle. Far from being theoretically motivated, however, the principle is responsible for a long history of supposed puzzles, paradoxes, and in some cases, outright contradictions with fundamental moral notions. The existence of these recurring difficulties is not unlike the recurring symptoms of an underlying disease. To continue the analogy, sophisticated research and lengthy literatures have attempted to isolate and treat the symptoms caused by accepting ought implies can that present within different subareas of ethics individually. The main contribution of this paper is not to impugn any of the attempts to address these difficulties. Rather, the main contribution is to bring together research across broad areas of ethical theorizing to identify a common cause of them, which is easy to miss when conducting specialized research within just one individual research area. The recurrence of these difficulties across ethics undermines the theoretical motivation for the principle and strongly suggests that it is false.

It might be objected that rejecting ought implies can would have at least one negative theoretical consequence, namely, that it would exaggerate the conditions in which one is blameworthy or properly criticizable for failing to do something. Some participants in this debate have observed that the “most plausible argument” for the principle involves the claim that “it is wrong to blame someone for something that they cannot control” (Stern 2004: 46). The basic idea is that agents who fail to fulfill obligations are blameworthy, but agents are not blameworthy when they are unable to act, and thus no moral obligation can be present when they are unable to act (for a parallel argument concerning epistemic norms, see Douven 2006). Prior researchers have noted a similar worry as follows, where “OIC” is short for ought implies can (Brown 1977; Henne et al 2019: 134; see also Mizrahi 2009: 30–31):

  1. 1.

    If OIC is false, then we would blame some agents for failing to do what they cannot do.

  2. 2.

    We don’t blame any agents for failing to do what they cannot do.

  3. 3.

    Therefore, OIC must be true.

However, this line of reasoning does not motivate the principle because it is both controversial and unsound. It is an open question in ethics and action theory whether agents are properly blameworthy or criticizable in conditions where the ability to do otherwise is absent (Fischer and Ravizza 1998; Mckenna 2009; Vargas 2013). Some philosophers have even explicitly challenged this argument with respect to ought implies can by denying the premise that we do not blame agents for doing what they cannot do (Mizrahi 2009). We clearly do blame agents failing to do what they cannot do, in some circumstances at least, as is apparent in cases of self-imposed inability and negligence (Sinnott-Armstrong 1984).

The argument is also unsound because premise one is false. It is initially unclear why blaming agents for failing to do what they cannot do would follow from rejecting ought implies can. Perhaps the thinking is that we would blame agents in such an eventuality because moral obligations would then exist that would go unfulfilled due to inability and blame follows from unfulfilled obligation. But this does not follow because unfulfilled obligation does not entail blameworthiness. We often blame agents when they fail to fulfill obligations, of course, but the two also readily come apart, as is made apparent by the important moral category known as “blameless transgression” (for discussion, see Turri 2016; Turri and Blouw 2015). As discussed above, inability is often offered and accepted as a valid excuse that mitigates or removes blame entirely for failing to fulfill moral obligations that we acknowledge exist in the very same circumstance (see also Buckwalter and Turri 2015; Turri 2017a for experimental evidence). In fact, virtually any moral obligation can be broken both blamefully or blamelessly in various circumstances, and some of the clearest cases of the latter involve unfulfilled obligation due to inability, of various sorts. For these reasons, it does not follow from the fact obligations could sometimes exist without ability that agents are blameworthy for their behavior. This suggests that the sentiment that it would be wrong to blame people for what they cannot do, while perhaps true in many but not all cases, does not motivate the principle that obligation entails ability because inability can excuse blame for unfulfilled obligation.

Another potential objection is that discarding the principle requires rejecting the widely accepted theory that morality is motivational or action guiding (for discussion of action guidance, see Mason 2003; Prichard 1968; Smith 2012). If ought implies can is false, then some moral obligations may be impossible to fulfill. Impossible things, it is sometimes suggested, cannot motivate someone to do something or guide action. Therefore ought implies can is either false or morality is not action guiding. First, it should be noted, of course, that the fact that some obligations are not motivational or action guiding does not entail morality itself is not motivational or action guiding. It may still largely be these things. But more to the point, this concern is undermined by the fact that the impossible can motivate, in at least two senses. In one sense, impossible obligations can motivate when they inspire us to approximate moral ideals (Talbot 2016). According to Talbot, for example, the directive to “love all people as one loves oneself”, while impossible for most agents to achieve, might still guide actions in many ways that successfully promote that ideal (2016: 380–81). In another sense, impossible obligations can motivate us in virtue of our recognition that they can exist in the first place. With respect to moral dilemmas, for example, Ruth Barcan Marcus observes that impossible obligation “motivates us to arrange our lives and institutions with a view to avoiding such conflicts” (1980: 121). Further still, Macrus suggests that this reality could support the existence of the regulative moral principle that “as rational agents with some control of our lives and institutions, we ought to conduct our lives and arrange our institutions so as to minimize predicaments of moral conflict” (121). For these reasons, retaining the principle of action guidance does not motivate ought implies can, and may even count against it.

A final worry is that without the principle there would be nothing to constrain what morality can rightfully demand of us. A related concern is that ought implies can provides the best explanation for why we do not have moral reasons to do outlandish things, such as go back in time to prevent slavery or develop medicine to cure all human diseases (Streumer 2018). In other words, not having the ability to prevent slavery or disease, it might be thought, seems to explain why we do not have reasons to do these things, and in turn, why we should think that moral obligation entails ability. This is an important observation that potentially sheds light on how ability and obligation might sometimes be related. However this concern does not theoretically motivate the principle, either. The principle under consideration is an entailment relation that says that no obligation can exist without ability. But it is entirely consistent with rejecting this principle that ability could sometimes or even frequently limit obligation without being entailed by it. For example, it is possible that lacking the ability to prevent slavery or disease could be what explains why we do not have obligations to do these things, and the fact that obligation does not entail ability, to both be true. There could simply be other counterexamples in which the two come apart.

To see this, suppose for the sake of argument that we do not have any moral reasons to prevent slavery or disease and that our inability to do so explains this. The fact that not having an ability seems relevant to explaining the absence of obligation in a particular circumstance does not justify the conclusion that obligation entails ability, just as for any A and B, the observation that not-B seems related to not-A fails to justify the claim that A entails B. In the present case, this conclusion is not justified, of course because it is possible that inability sometimes or even frequently limits certain moral obligations involving, say, time travel or health miracles without moral obligation going so far as to entail ability (for other challenges and responses to the argument, see Henne et al. 2019; Streumer 2018). And when appealing to an inference to the best explanation, and especially in the context of points raised above, the possibility of weaker relationships between these concepts should be preferred over the strongest possible relationship between them. As it happens, this possibility has been profitably explored by philosophers in related but distinct debates concerning free will and the principle of alternative possibilities, for example, with the theory that moral responsibility requires only that one could have done otherwise, ceteris paribus (Whittle 2016).Footnote 1 Applying this view of moral responsibility to the present discussion of moral obligation, such a theory is well theoretically motivated. Rejecting entailment and allowing for exceptions addresses each of the theoretical problems raised by accepting the principle above while retaining a connection between ability and obligation. This suggests that progress in ethics can be made by adopting a more nuanced account of morality tolerant of inability and by rejecting the poorly motivated theory that ought implies can.