Skip to main content

Weakness of the Will


The primary philosophical topic explored under the rubric "weakness of the will" is roughly what Aristotle called akrasia. This classical Greek term is formed from the alpha privative (basically, a negation sign) and kratos, meaning "strength" or "power." The power at issue is the power to control oneself in the face of actual or anticipated temptation. So akrasia is deficient self-control. Self-control, in this sense, may be understood as constituted primarily by a robust capacity to see to it that one does what one believes to be best on the whole when tempted to do otherwise. The self-controlled person, Aristotle writes, "is in such a state as to master even those [temptations of a certain kind] by which most people are defeated," and the akratic person "is in such a state as to be defeated even by those which most people master" (Nicomachean Ethics 1150a1113).

In Plato's Protagoras, Socrates says that the common view about akratic action is that "many people who know what it is best to do are not willing to do it, though it is in their power, but do something else" (352d). Here he raises (among other issues) the central question in subsequent philosophical discussion of akrasia : Is strict akratic action possible? Strict akratic action may be defined as free, intentional action that is contrary to a conscious belief that the agent has at the time to the effect that it would be best to A (or best not to A )best from the perspective of his own values, desires, beliefs, and the like, as opposed, for example, to a common evaluative perspective that he does not endorse. In this entry, I call beliefs with all the properties just mentioned P beliefs.

A feature of paradigmatic strict akratic actions that is typically taken for granted and rarely made explicit is that the P beliefs with which they conflict are rationally acquired. In virtue of clashing with the agent's rationally acquired P beliefs, akratic actions are subjectively irrational (to some degree, if not without qualification). There is a failure of coherence in the agent of a kind directly relevant to assessments of the agent's rationality. This kind of failure would be exhibited, for example, by a student who freely goes to a party tonight even though he or she has a P belief that it would be best not to go and to study instead.

To some theorists (e.g., R. M. Hare, Socrates, and Gary Watson), the threat that strict akratic action poses to our ability to make sense of human action seems so severe that they deem strict akratic action conceptually or psychologically impossible. Many others, including Donald Davidson, Alfred Mele, David Pears, and Amelie Rorty, try to accommodate strict akratic action in a general theory of human action.

Skepticism about Strict Akratic Action

For the purposes of this entry, it may be assumed (P1 ) that people sometimes act freely and (P2 ) that people sometimes perform intentional actions that are contrary to their P beliefs. Some compulsive hand-washers or crack cocaine addicts may occasionally confirm P2. But acting contrary to one's P belief is not sufficient for acting akratically; one's action must also be free. Some philosophers argue that strict akratic action is impossible because actions contrary to the agent's P beliefs are necessarily unfree.

Assumptions P1 and P2 and the following assertion form a consistent triad: (UF ) All actions contrary to the agent's P belief are unfree. How might a philosopher try to defend UF while granting P1 and P2 ? Here is a sketch of one such defense (Harepresents a similar argument in chapter 5):

Argument A

A1. Having a P belief that it is best to A now is conceptually sufficient for having an intention to A now.

A2. Any agent who intends to A now but does not A now is unable to A now.

A3. Such an agent, being unable to A now, is compelled to performand therefore unfreely performswhatever pertinent intentional action he now performs.

Premise A2 is falsified by simple counterexamples. A professional pitcher who intends to throw a pitch in the strike zone may accidentally miss even though he was able to do what he intended. Of course, the failures in alleged strict akratic actions may be different in important ways, and it may be claimed that A2 simply needs to be revised to capture the difference. One likely suggestion is that in alleged strict akratic actions, the failure involves a change of intentionfor example, a change from intending to study to intending to attend a partywhereas the pitcher's failure does not. Now, either the change of intention is paired with a corresponding change of belief or it is not. If there is a change of belief that matches the change of intentionfor example, a change to believing that it would be better to attend the partythen the agent does not act contrary to his current P belief in executing that intention. But it is assumed that some actions are contrary to their agents' current P beliefs, and the skeptic is supposed to be arguing that all such actions are unfree. So suppose that the change of intention is not paired with a corresponding change of belief and that the agent's P belief persists. Then A1 is false. It is falsified by an agent who had intended in accordance with a P belief but no longer so intends even though the belief persists.

A1 is in dire straits anyway, given P2. Consider compulsive hand-washers or crack cocaine addicts who believe that it is best not to wash their hands now or not to use crack now, but who do so anywayintentionally and unfreely. If A1 is true, they are intentionally washing their hands or using crack while intending not to do so. Although this may be conceptually possiblefor example, perhaps an agent with a split brain may intend not to A while also intending to A and acting on the latter intentionit is a highly implausible hypothesis about representative cases of the kind at issue. A much more plausible hypothesis is that although the troubled agents believe that it would be best not to wash their hands now or not to use crack now, they lack a corresponding intention and instead intend to do what they are doing.

A3 also is problematic. Bob has been dieting and believes it best to order a low-calorie salad for lunch today. Unfortunately, he is tempted by several other items on the menu, including a hamburger, a steak, and a pork sandwich. He orders the steak. Even if Bob was unable to order the salad, we would need an argument that he was compelled to order the steakthat, for example, ordering the burger was not a live option.

Gary Watson offers the following argument for UF :

Argument B

B1. An agent's succumbing to a desire contrary to his P belief cannot be explained by his choosing not to resist nor by his making a culpably insufficient effort to resist.

B2. Only one explanation remains: The agent was unable to resist.

So UF. All actions contrary to the agent's P belief are unfree.

Watson argues that an agent's choosing not to resist cannot explain strict akratic action, for to make such a choice "would be to change" one's P belief (p. 337). For example, "The weak drinker's failure to resist her desire to drink is a failure to implement her choice not to drink. To choose not to implement this choice would be to change her original judgment, and the case would no longer be a case of failure to implement a judgment" (pp. 336337). Watson also contends that an insufficient effort cannot be due to a belief that the effort is not worth the trouble, since the belief that it is worth the trouble is implicit in the violated P belief (p. 338). Nor, he argues, can the insufficient effort be explained by a misjudgment of "the amount of effort required," for misjudgment is "a different fault from weakness of will" (p. 338).

In some alleged instances of strict akratic action, agents believe that it would be best to A, choose accordingly, and then backslide while retaining that belief. In others, agents with the same P belief do not choose accordingly; they do not make the transition from belief to intention. Although Watson has the former kind of case in mind, it is useful to attend to a case of the latter kind. Imagine, if you can, that a drinker, Drew, who has had one shot of bourbon and needs to drive home soon, believes that it would be best to switch now to coffee but neither chooses nor intends to do so and intentionally drinks another bourbon. The reader is not asked to imagine that Drew akratically drinks the second bourbon; it is left open that she drinks it unfreely. If Drew can believe that it would be best not to drink a second bourbon without choosing accordingly, then she can fail "to resist her desire to drink" without there being any failure on her part "to implement her choice not to drink." If she makes no such choice, she does not fail to implement it. And if there is no such failure of implementation, then the reason Watson offers for maintaining that the agent "change[d] her original judgment" is undercut.

A scenario in which a belief-matching choice is made will be discussed shortly. The plausibility of scenarios of the present sort deserves a bit more attention now. Consider the following story. On New Year's Eve, Joe, a smoker, is contemplating kicking the habit. Faced with the practical question of what to do about his smoking, he is deliberating about what it would be best to do about it. He is convinced that it would be best to quit smoking sometime, but he is unsure whether it would be best to quit soon. Joe is under a lot of stress, and he worries that quitting smoking now might drive him over the edge. Eventually, he judges that it would be best to quit by midnight. But he is not yet settled on quitting. Joe tells his partner, Jill, that he has decided that it would be best to stop smoking, beginning tonight. Jill asks, "So is that your New Year's resolution?" Joe sincerely replies, "Not yet; the next hurdle is to decide to quit. If I can do that, I'll have a decent chance of kicking the habit."

This story at least has the appearance of coherence. Seemingly, although Joe decides that it would be best to quit smoking, he may or may not choose (i.e., form the intention) to quit. Watson offers no argument for the incoherence of stories of this kind. (It has not been claimed that Joe is a free agent.)

If Drew can fail to resist her desire for a second bourbon without changing her belief about what it is best to do, what about Lucy, who, like Drew, takes another bourbon despite believing that it would be best to switch now to coffee, but, unlike Drew, chooses to switch now to coffee when she makes her judgment? Watson would say (W1 ) that Lucy's "failure to resist her desire to drink [a second bourbon] is a failure to implement her choice not to drink," (W2a ) that "to choose not to implement this choice [is] to change her original judgment," (W2b ) that to choose not to resist her desire to drink a second bourbon is to change that judgment, and (W3 ) that Lucy's drinking the second bourbon is therefore not a strict akratic action, since it is not contrary to her P belief (pp. 336337). Is W2a or W2b true? Watson offers no argument for either, and some stories in which analogues of both are false certainly seem coherent.

Here is one such story. Alex's friend, Bob, has proposed that they affirm their friendship by becoming blood brothers, since Alex is about to go away to prep school. The ceremony involves the boys' cutting their own right palms with a pocket knife and then shaking hands so that their blood will mingle. Alex is averse to cutting himself, but he carefully weighs his reasons for accepting the proposal against his competing reasons (including his aversion), and he judges that it would be best to accept the proposal and to perform the ceremony at once. He chooses, accordingly, to cut his hand with the knife straightaway. Without considering that he may find the task difficult, he grasps the knife and moves it toward his right palm with the intention of drawing blood. However, as he sees the knife come very close to his skin, he intentionally stops because of his aversion. He chooses not to implement his original choice just now, and he chooses not to resist his aversion further just now. Alex abandons his original choice. But he has not changed his mind about what it is best to do, and he is upset with himself for chickening out. (Soon, Alex resolves to try again, this time without looking. The second attempt succeeds.)

If this story is incoherent, Watson should explain why. If he were to assent to A1, he could appeal to it here: since Alex no longer intends to cut his hand straightaway, it would follow that he no longer believes that it would be best to cut it straightaway. But Watson rejects A1 to accommodate compulsives who act contrary to a P belief.

Explaining Strict Akratic Action

Imagine that although Jack believes that it would be better to study tonight for tomorrow's test than to attend a friend's party, he goes to the party and does not study. To the extent that his belief is sensitive to his motivational states (e.g., his desire to get a decent grade on the test), it has a motivational dimension. That helps explain why strict akratic action is regarded as theoretically perplexing. How, some philosophers wonder, can the motivation that is directly associated with a belief of this kindin this case, Jack's motivation to studybe outstripped by competing motivation, especially when the competing motivation (a desire to have fun tonight) has been taken into account in arriving at the belief?

One answer (defended in Mele 1987) rests partly on the following two theses and on various arguments for those theses.

P beliefs normally are formed at least partly on the basis of our evaluation of the objects of our desires (i.e., the desired items).

The motivational force of our desires does not always match our evaluation of the objects of our desires.

If both theses are true, it should be unsurprising that sometimes, although we believe it better to A than to B, we are more strongly motivated to B than to A. Given how our motivation stacks up, it should also be unsurprising that we B rather than A.

Thesis 1 is a major plank in a standard conception of practical reasoning. In general, when we reason about what to do, we inquire about what it would be best, or better, or good enough, to do, not about what we are most strongly motivated to do. When we ask such questions while having conflicting desires, our answers typically rest significantly on our assessments of the objects of our desireswhich may be out of line with the motivational force of those desires, if thesis 2 is true.

Thesis 2 is confirmed by common experience and thought experiments and has a foundation in empirical studies. Desire-strength is influenced not only by our evaluation of the objects of desires, but also by such factors as the perceived proximity of prospects for desire-satisfaction, the salience of desired objects in perception or in imagination, and the way we attend to desired objects (as Ainslie, Metcalfe and Mischel, and others have observed). Factors such as these need not have a matching effect on assessment of desired objects.

Empirical studies of the role of representations of desired objects in impulsive behavior and delay of gratification (reviewed in Mele 1995) provide ample evidence that our representations of desired objects have two important dimensions, a motivational and an informational one. Our P beliefs may be more sensitive to the informational dimension of our representations than to the motivational dimension, with the result that such beliefs sometimes recommend actions that are out of line with what we are most strongly motivated to do at the time. If so, strict akratic action is a real possibilityprovided that at least some intentional actions that conflict with agents' P beliefs at the time of action are freely performed. To be sure, it has been argued that no such actions can be free, but, as the preceding section indicates, representative arguments for that thesis are unpersuasive.

Unless a desire of ours is irresistible, it is up to us, in some sense, whether we act on it, and it is widely thought that relatively few desires are irresistible. Arguably, in many situations in which we act against our P beliefs, we could have used our resources for self-control in effectively resisting temptation. Normal agents can influence the strength of their desires in a wide variety of ways. For example, they can refuse to focus their attention on the attractive aspects of a tempting course of action and concentrate instead on what is to be accomplished by acting as they judge best. They can attempt to augment their motivation for performing the action judged best by promising themselves rewards for doing so. They can picture a desired item as something unattractivefor example, a wedge of chocolate pie as a wedge of chewing tobaccoor as something that simply is not arousing. Desires normally do not have immutable strengths, and the plasticity of motivational strength is presupposed by standard conceptions of self-control. Occasionally, we act contrary to our P beliefs, and it is implausible that, in all such cases, we are unable to act in accordance with those beliefs.

The key to understanding strict akratic action is a proper appreciation of the point that the motivational force or causal strength of a motivational attitude need not be in line with the agent's evaluation of the object of that attitude. Our P beliefs are based, in significant part, on our assessments of the objects of our desires; and when assessment and motivational force are not aligned, we may believe it better to A than to B while being more strongly motivated to B than to A. If while continuing to have that belief, we freely do B, our action is strictly akratic.

See also Aristotle; Davidson, Donald; Hare, Richard M.; Plato; Power; Socrates.


Ainslie, George. Breakdown of Will. Cambridge, U.K.: Cambridge University Press, 2001.

Aristotle. Nicomachean Ethics. In The Complete Works of Aristotle: The Revised Oxford Translation, edited by Jonathan Barnes. Princeton, NJ: Princeton University Press, 1984.

Charlton, William. Weakness of Will. Oxford: Basil Blackwell, 1988.

Davidson, Donald. Essays on Actions and Events. Oxford: Clarendon Press, 1980.

Dunn, Robert. The Possibility of Weakness of Will. Indianapolis, IN: Hackett, 1987.

Hare, R. M. Freedom and Reason. Oxford: Oxford University Press, 1963.

Mele, Alfred R. Autonomous Agents: From Self-Control to Autonomy. New York: Oxford University Press, 1995.

Mele, Alfred R. Irrationality: An Essay on Akrasia, Self-Deception, and Self-Control. New York: Oxford University Press, 1987.

Metcalfe, Janet, and Walter Mischel. "A Hot/Cool-System Analysis of Delay of Gratification: Dynamics of Willpower." Psychological Review 106 (1999): 319.

Pears, David. Motivated Irrationality. Oxford: Oxford University Press, 1984.

Plato. Protagoras. Translated by C. C. W. Taylor. Oxford: Clarendon Press, 1976.

Rorty, Amelie. "Where Does the Akratic Break Take Place?" Australasian Journal of Philosophy 58 (1980): 333346.

Watson, Gary. "Skepticism about Weakness of the Will." Philosophical Review 86: 316339.

Alfred R. Mele (2005)

Cite this article
Pick a style below, and copy the text for your bibliography.

  • MLA
  • Chicago
  • APA

"Weakness of the Will." Encyclopedia of Philosophy. . 19 Feb. 2019 <>.

"Weakness of the Will." Encyclopedia of Philosophy. . (February 19, 2019).

"Weakness of the Will." Encyclopedia of Philosophy. . Retrieved February 19, 2019 from

Learn more about citation styles

Citation styles gives you the ability to cite reference entries and articles according to common styles from the Modern Language Association (MLA), The Chicago Manual of Style, and the American Psychological Association (APA).

Within the “Cite this article” tool, pick a style to see how all available information looks when formatted according to that style. Then, copy and paste the text into your bibliography or works cited list.

Because each style has its own formatting nuances that evolve over time and not all information is available for every reference entry or article, cannot guarantee each citation it generates. Therefore, it’s best to use citations as a starting point before checking the style against your school or publication’s requirements and the most-recent information available at these sites:

Modern Language Association

The Chicago Manual of Style

American Psychological Association

  • Most online reference entries and articles do not have page numbers. Therefore, that information is unavailable for most content. However, the date of retrieval is often important. Refer to each style’s convention regarding the best way to format page numbers and retrieval dates.
  • In addition to the MLA, Chicago, and APA styles, your school, university, publication, or institution may have its own requirements for citations. Therefore, be sure to refer to those guidelines when editing your bibliography or works cited list.