Skip to content




The problem of akrasia or weakness of will , has a wonderful pedigree. Socrates may be said to have invented it and to have supplied a tentative solution to it, which prompted Aristotle to suggest a few improvements etc. etc. And as things stand today it would be hard to take a degree in Philosophy without encountering this problem at some stage. 

The problem itself may be stated quite simply: despite the fact that I know very well that, for example, smoking is a dangerous practice, that it can lead to an increased risk of heart disease and lung cancer, despite the fact that I have envisaged with awful clarity the shame of confessing to my family that my illness has been self-inflicted, that I have no excuses for the folly of my actions, despite the though of a slow and lingering death traumatized by bitter regret, yet, when I feel like a smoke, all these terrifying and remorseless thoughts vanish at the prospect of this ration minor gratification. Though Socrates was not a smoker, when confronted with Attic variants on this apparently common human failing he remarked: "It would be strange .....if when knowledge was in a man something else could master it and drag it about like a slave." (EN 1145 b 23-25) 

In other words the idea that reason could marshal conclusive arguments in favour of a given course of action (the idea that knowledge could be in a person) and then have the motivational strength of this knowledge overcome by the strength of some occurrent desire presents an unhappy prospect with reference to the project of leading our lives in a rational manner. And if we further suppose that the idea that we can lead a life guided by reason is the presupposition which conditions the possibility of a practical philosophy we can appreciate the force of Socrates' remark: it would be strange, which is to say, scandalous, if Desire could master Reason and drag it about like a slave. 

If the problem of akrasia could not be solved this would constitute a philosophical scandal in that it would put in doubt the possibility of guiding our lives through the exercise of our reason. Before continuing it should be noted at the outset that I am, a la Davidson, equating reasons for action with belief-desire (or 'pro attitude') pairs. And this is because in a situation in which I am trying to make up my mind about what I shall do, a reason for doing x will count as a reason for me only if it consists of an appropriate belief backed by a relevant desire. Thus if I offer as a reason for not smoking the belief that the practice can lead to cancer and premature death this belief will count as a reason for me to act in a certain way only if I do not want to suffer from cancer and thus die prematurely. 

In ordinary speech this fact that my beliefs count as reasons for action because they are backed by relevant desires is usually taken for granted. As Davidsons puts it: "A primary reason consists of a belief and an attitude but it is generally otiose to mention both." (p 82 'Actions, Reasons and Causes' in The Philosophy of Action Ed. A.R. White, O.U.P., Oxford, 1968). In what follows I shall take advantage of this otioseness and talk, as we ordinarily do when we engage in practical deliberations, simply of various desires to do x being met by reasons against doing so. In other words when I cite a reason ('smoking can lead to premature death') it will be assumed that this belief of mine counts as a reason to act for me because I do not want to die prematurely. When I cite a desire as a reason for acting ('I feel like smoking a cigarette') it will be assumed that I have the appropriate beliefs (cigarettes can be smoked, I have a packet in my jacket, etc.) 

Let me now return to the question raised by Socrates. I maintain that there are two sorts of scandals involved in cases of akrasia, the first of which is a logical scandal. The bit of logic that is involved can be illustrated as follows: Imagine a pair of scales. On one side of the scales Reason piles up her weighty arguments against smoking. A shortened life-span, the consequences for loved ones, a loss of self-respect, pain, regret, and death. On the other side Desire offers the prospect of a curling wisp of pleasure. And the logical scandal lies in the fact that all Reason's protestations are outweighed by this single promise of pleasure. How is it possible that these weighty reasons can count for naught when balanced against pleasure? Well, only if . . . and then the traditional solutions are offered. All of these solutions attempt to mitigate the scandal (the logical scandal which in terms of our metaphor amounts to the possibility that something light could outweigh something heavy) by, in effect, saying that somehow, the person did not feel the full weight of the reasons adduced against smoking, that for example they simply 'entertained' them as theoretical considerations and did not truly appreciate their practical import, etc., etc. 

Now all of these explanations about how Reason's arguments could have been 'outweighed' by its lighter rival may be thought of as doing their explaining in terms of the metaphor of the balance. Reason it turns out, is unable to outweigh pleasure because the weight of its arguments is somehow taken away. (Thus the person who is drunk cannot appreciate the weight of reason's arguments) Never for a moment is the central metaphor of the balance which creates the scandal challenged. (It is taken for granted that when opposing motives seek to determine action the one which succeeds does so in virtue of its greater 'weight'. 

Thus it is never doubted that if Reason's argument's were given their full weight, then the scales would tip in favour of Reason and the person would act accordingly. This is the assumption that underpins the attempts to explain akrasia in terms of 'something going wrong' with the reasoning process, something that deprives Reason's arguments of their usual weight. The logic of the balance cannot allow the scandal of something heavy being outweighed by something light and since, by assumption, the considerations produced by reasoning must weigh more heavily with us than the promptings of Desire (or else no practical philosophy would be possible) therefore what has gone wrong must be associated with some kind of tampering with the weights. Suggesting how this tampering can occur is the essence of Aristotle's various solutions to the problem of akrasia. 

As I mentioned before, there are two aspects to scandal (1) there is the more or less logical contradiction inherent in the idea that something heavy could be outweighed by something light, (the saving move here is to explain that the apparently heavy thing is in fact weightless because. . . .) And (2) a related scandal tied in with the more fundamental implication of this failure of Reason to prevail, an implication which reveals a kind of existential scandal concerning the human condition, namely that the imagined prospect of future pains (which Reason summons up through her arguments) can be outweighed by the tingling prospect of those immediate pleasures which Passion offers. This represents anew the scandal of Reason being dragged about like a slave, for unless Reason's warnings about the pains to come can consistently outweigh the promise of pleasure offered up by present desires, the role of Reason as a guide to behaviour, indeed the whole conception of human beings as capable of leading a life, instead of simply being led by the nose, becomes suspect. 

If the second aspect of the scandal is to be avoided Reason must be preserved as the weightier element at all costs, so that if, for example, a case of akrasia is presented in which there is no evidence that the akratic's understanding of the situation is at fault, the dignity of human existence can only be saved by an explanation that places the akratic personality outside the norm. Thus the akratic's desires are supposed to be abnormally strong and this explains his or her odd behaviour - viz; smoking despite the overwhelming weight of evidence warning against this practice. This is a happy solution to the problem since it saves the assumption that, in general, for normal human beings, Reason can overrule the passions: Thus if cases of akrasia are fundamentally pathological, then rational behaviour and hence dignity can be preserved for normal human beings. But - apart from the loss of dignity which might well befall us if Reason lacked the sovereign power to rule the passions consistently - why do we suppose that Reason must be able to rule the passions, that dignity is our birthright (if we are normal)? 

Wherein does the necessity lie? I believe the answer lies in the logic of the metaphors that we use to explain why we make the decisions we make. There are two principal metaphors which we employ, one the previously mentioned metaphor of the scales and the other based on something like a tug-of-war in which the winner wins because they pulled the hardest or were the strongest. (The master/slave metaphor is of this type.) Now in both cases, the logic of the metaphors is such that, necessarily, the heaviest weight will tip the balance, (or the strongest pull will win). So far so good. If this were all that the two metaphors involved they would be quite neutral ways of explaining decision-making. 

Thus I can explain my behaviour by simply noting that I did x because the reasons against it proved to be stronger than the desires for it. Or alternatively: the desires for it weighed more heavily with me than the reasons against it. The difficulties begin when we add in some semi-logical presuppositions which surround these metaphors. For example; it follows, though not necessarily, that if you add more weights to one balance pan it will begin to outweigh the other (not necessarily because everything depends on how much the individual weights weigh). Or, it follows, though not necessarily, that in a tug-of-war, the team which has the most members will be the strongest and therefore win (but again, not necessarily - it depends on the strength of the individuals involved). Now if these considerations are allowed to creep into our talk about decision- making they are quite capable of breeding paradoxes with the very same scandalous flavour as the paradoxes which surround the problem of akrasia. 

Thus the typical akratic - our smoker for example - watches the evidence against their beloved habit mount up, year by year, and hears the logical outrage in the voice of others as they call attention to the numerous arguments against smoking which are ranged against the single consideration of the gratification which smoking involves. The smoker hears the logical pain in their voices, and wonders at his or her own stupidity: can he or she not see that 20 (or so) reasons outweigh or are stronger than one! To explain this lack of common sense they are forced to admit to their own weakness as the only possible explanation since they readily acknowledge that the reason's against smoking far outweigh (20 to 1!) - the reason in favour of it (our desire to smoke). But as we saw above, it does not necessarily follow that 20 items will outweigh one, everything depends on the weights of the items concerned. 

The akratics proper response to all these 'logical' protestations is not to admit to their own 'stupidity' and excuse it as a function of their irrationality, but rather to wonder at the strength of their desire to smoke. In short, there is nothing logically scandalous about the decision the akratic makes once we accept the propriety of the metaphors which we use to describe our decision-making. These metaphors which we use to describe our decision-making. These metaphors suggest that it is the 'weight' or 'strength' of the factors for or against a given line of conduct which alone determine our behaviour. So if we pay attention to the logic of these metaphors it would seem that there is no such thing as the philosophical problem of akrasia. 

To see why let us review this situation: the philosophical problem can only be generated when the metaphor of decision-making is interpreted in terms of a feature of the activity upon which the metaphor is based which is irrelevant to its central logic. Thus the logic of a set of scales is bound up with the idea that this device is able to determine which of two objects or sets of objects is the heaviest. And as we have just seen, the number of objects in each pan is not relevant: it is their weight alone that counts. If this is forgotten a paradox can arise (less outweighing more) and with it a philosophical problem. However, this is not the only way in which the logic of the metaphor can become confused, resulting in a paradox. For example it is perfectly obvious that a set of scales cannot determine which of two objects is the cleanest or the most pliable. It is equally obvious that it can only determine which of two objects is the heaviest if both objects have weight. There is no sense in which an orange is outweighed by the traffic code. Thus when we make our metaphor, when we say that decision-making is a matter of weighing-up the pros and cons of a given line of conduct, the utility of this metaphor will be a function of how closely we stick to the logic of the original activity (weighing things) as we apply it to the factors which we regard as having the power of being able to determine our conduct. If we stick closely to the logic of the original model, the metaphor provides us with an understanding (or at least some grasp) of why we decided to do x rather than y. The factors in favour of x outweigh the factors in favour of y. Now it is a basic implication of the logic of the model that if one object outweighs another then they must both have been of the same type i.e., possessed of a commensurable quality or 'weight' in terms of which they could be compared. And everything will go smoothly as long as the factors (pro and con) in any mental-weighing (our metaphor for decision-making) are of the same type, e.g., desires. No philosophical or logic puzzles can be generated by the fact that my desire to smoke is stronger than my desire to abstain, a desire generated by Reason's arguments. 

However, suppose I deviate from the logic of the original model and, for example, begin weighing lines of conduct motivated by my desires against lines of conduct motivated by my sense of duty. It may seem obvious that my knowledge of my duties should also weigh with me when I am faced with a decision concerning my conduct in which desire and duty conflict. But are the weights of duties and desires commensurable? Do duties weigh with me in exactly the same sense as desires, for example, in terms of the pleasures derived from dutiful behaviour as opposed to the pleasures consequent on the satisfaction of my desires? If they did, then the force of, for example, moral injunctions involving the notion of 'ought' would become difficult to understand. Thus there is very little point in saying that I ought to do my duty rather than follow the path of desire if my understanding of my own capacity to decide between these two paths is derived from the assumption of a passive model of decision-making based on weighing-up those commensurable quantities of pleasure promised by alternative lines of conduct. For on this model of decision-making, which line of conduct I pursue is determined independently by the weights of the factors placed in the balance. 

However, if I take the notion of ought seriously, then I am almost forced to think of the weight which duties have as being a function of my willingness to give them weight. At a stroke, my understanding of decision-making based on the passive model of the weigh scales (in which the 'heaviest' desire independently determines the line of conduct I will pursue) is transformed into an active conception of decision-making in which I have the power to tip the balance, and in which the pleasures involved in dutiful behaviour become irrelevant. Now it follows that if duties are regarded as having a special kind of weight which is a function of my activity (my will) then it follows immediately that the explanation of why, on a given occasion, my desires 'outweighed' my sense of duty can be attributed to either my unwillingness to accord duties their proper weight (i.e. a weight which ought always to be greater than the weight of any conflicting desire) in which case I am an evil person, or, I must regard my dereliction of duty as a function of weakness of will - a constitutional lack which I have no immediate power to rectify. 

It is this latter alternative which characterises the Socrates or Aristotelean explanation of the problem of akrasia. There is something wrong with the akratic. His or her weigh scales are not functioning properly. The advantages of duty, or more appropriately for the Greeks, the life of virtue, simply do not weigh with them as they should. Explanations of where the fault lies abound, but they all are premised on the idea that something has gone wrong with the person's decision-making apparatus. This apparatus operates on the principle that the life of virtue which, as a matter of fact, maximises a person's potential for happiness, will necessarily be preferred (and pursued) over any alternative mode of conduct. More happiness outweighs less happiness: it follows, therefore, that anyone who chooses a mode of conduct which does not maximise their potential for happiness must be confused in some way. Since for the Greeks, this model of decision-making determines what rational decision-making amounts to, there is, by definition, no possibility of irrational decision-making among people whose faculties are unimpaired: i.e., people who fully understand that the advantages of the virtuous life outweigh any alternative modes of conduct. There is, in other words, no possibility of a true existential scandal (the second of the two scandals mentioned above) which would rob human existence of its dignity. 

A normal person who is in possession of their faculties could not decide to live in accordance with a mode of conduct which does not maximise their happiness. - cf.E.N. 1140 b4-6. And this is because the metaphor of the weigh scales (or some logical equivalent such as that of the Master and the Slave where the strong Master necessarily overcome the weak Slave) which dominates the Greek understanding of decision-making does not allow for this possibility. This model of decision-making is passive through and through, so that, for example, Aristotle consistently describes the process of attaining virtue (or some tincture of it) in terms of acquiring good habits, a process which depends primarily upon parents and educators and a sound (or normal) constitution. Through practice (motivated by the praise of parents or peers), a taste for virtue is inculcated which insures that the decision to follow the path of virtue will be a natural function of the direction in which one's character has been developed. The delights associated with behaving virtuously will, as a matter of fact, come to weigh more heavily with the properly conditioned person than any delights associated with vice. 

Let me now review the situation once more: Suppose the metaphor of the weigh scale (or one of its' logical equivalents) determines our understanding of decision-making. This explanation of decision-making will be consistent - i.e. will not create any logical or philosophical puzzles - if the logic of the model is followed in its metaphorical application. If it is not followed - if, for example, attempts are made to compare factors with incommensurable 'weights' (for example, duties and desires), two responses are possible. The first response - the Greek response - is to 'commensurate' the weights and restore the logic of the model. The path of duty, or more properly, the path of virtue, for example, will be shown to be the more desirable way of behaving, more desirable, in fact, than any alternatives. And it is understood that bringing about a situation in which this will be 'in fact' true for any normal person requires appropriate conditioning. Improper conditioning constitutes the explanation of why, on a given occasion, a certain person (the akratic) misread the alternatives and chose the one which promised less happiness. 

The second response - which characterises modern conceptions of agency where decision-making is regarded as involving an act of will - is to furnish the model of decision-making with a new bit of logic. Thus for example a rule may be laid down (a moral rule) such that particular factors (duties) must always be regarded as outweighing others (desires). (This rule perhaps stems from the need to preserve the autonomy (and thus the dignity) of the individual, a consideration which is absent in the Greek situation where conditioning is regarded as the necessary precursor of a person's leading a noble life). If you then actively follow this rule when you make decisions you are good - morally good - you possess a good will. If you disobey this rule you are morally bad or evil. 

How precisely you are supposed to be able to follow or disobey the rule is of course, another problem, the problem of free will. This problem, in terms of the model we have been presenting, seems to turn on the question of how the path of virtue (which through appropriate conditioning could come to seem eminently desirable and thus an alternative which outweighed all competitors) became the path of duty, an alternative whose weight was to be measured in terms of a new quality - moral goodness - which could only outweigh its rival - pleasure or happiness - though our wilful intervention. This problem, however, may be left for some future President to solve.