Back to home page

* 1984 The Institute of Mind and Behavior, Inc. The Journal of Mind and Behaviot Winter, 1984, Volume 5, Number 1 Pages 1-10 ISSN 0271-0137

Desires Don't Cause Actions

J. Michael Russell California State University, Fullerton

The present paper presents the argument that desires do not cause actions because language which portrays an event passively, as having been brought about or made to happen, is incompatible with regarding that event as a genuine action of an agent. This being so, there is something fundamentally muddled about the perspective typically taken by the social sciences. For if we lose the concept of an action, we lose the concept of the person as well, and the social sciences will not be able to tell us about people. And, specifically, it is wrong to say that actions are caused by desires. Typically, action and the desire to act are not sufficiently distinct for them to be causally related. When we can distinguish these, there are yet further difficulties in trying to causally relate a desire to a (subsequent) action, since this relationship will hold where a person desires to go along with his or her desire, and this (second order) desire is not sufficiently distinct from actually going along with that desire for this to be a causal relationship. Moreover, either we think of this (separate) desire as a something in response to which I act, in which case it does not sufficiently explain my action unless we also refer to how I desire to respond to this something, or else this desire "brings my act about" in a way which makes it wrong to call it my act. Finally, considered phenomenologically, desires are not analo­gous to events which cause.

One good way to explain what someone has done is to say what it was that he or she desired to do. Although human desires figure profoundly into our explaining human actions, I think we should not say that such explanations are causal. It is true, of course, that in ordinary usage we do at times speak of a person's desires as causing the individual to do this or that, but this is a confusing and muddled practice which, if we follow it, we at least ought not to take quite seriously.

Part of my reason for maintaining that human actions are not caused by desires is that I think it wrong to regard actions as caused at all. I shall begin with that general thesis. This will be followed by further argument as to why, specifically, desires do not cause actions.

Actions Aren't Caused At All

Human action, qua action, is uncaused. For although I surely am acted upon by the world, I am in that respect not acting. To speak of the motion of

Request for reprints should be sent to J. Michael Russell, Ph.D., Philosophy and Human Services, California State University, Fullerton, California 92634.


my left arm when it rises as my action excludes speaking of it as acted upon. These forms of speech involve contrary and excluding perspectives, different "language games." That is, it is a fact of our language that if we portray something as being my action we cannot in the same "logical breath" portray it as having been caused. This is not merely a fact about the word "cause," but about the whole difference in perspective between my being considered passive versus active in the movement of my body. I emphasize that it is not just the word "cause" but a whole point of view that is miscast: it happens to be a point of view to which most of psychology and the social sciences are committed; to that extent, these disciplines make no sense. (It is possible to do much that is constructive while engaged in an enterprise which makes no sense. One might build a better cannon with the nonsensical objective of sinking four out of three battleships. Psychology is like that.)

Suppose, to begin with, that my arm is raised by wires. To say that my wires raised my arm is to say that I did not raise it. It was not my action. I did not do it. It happened to me. If you believed that I had raised my arm, and then discovered wires pulling on my arm, you would speak differently about what you saw: you would speak of it as something I was undergoing, not something I was doing. For to say that I was undergoing the action is to deny to that extent that I was doing it. We may change our (linguistic) perspective on something which we had called "my action," and come to view it as caused; then we may speak quite loosely of the cause of my action. You may come to find out that my arm was "waving" because of the wires pulling it, puppet style. But if you then speak of the wires as the cause of my action you are not really speaking of it as my action. This is so with the discovery of actual wires or levers, agents which physically and "mechanically" push or pull me, or with any change in perspective which "literally" or "metaphorically" draws on any sort of cause, force, power, agency, nature, push, pull, or pressure, which might act on me. If it is because of this that my arm moves, I do not move my arm.

In so far as you, or wires, or anything acts upon me to move my arm, I do not move it. In such a case, under such a description, it is wrong to speak of this as my action, and wrong to speak of the cause of my action or to speak of what made me act that way. In the case of wires, it makes no difference whether these run to my arm from external sources or whether they are concealed under my flesh. Either way, to say that my arm moves because of the pulling of these wires is to deny, just in so far as I am portrayed as acted upon, that I am acting. It makes no difference whether we deal with literally mechanistic causes or with concepts more or less metaphorical and parasitic on that model. Many philosophers and psychologists do not regard them­selves as having a mechanistic understanding of causality, but contend that with this concept they refer merely to empirically discoverable law-like relationships between events and not to a mechanistic account of these relationships. This sort of Hume-like approach can be found in Skinner, who


wants to talk just about "functional relationships." However, as long as there remains in these terms so much as a trace of the idea that one event brings about the next, there remains the perspective of what happens to or acts upon the second event. It seems to me that this point of view runs throughout psychol­ogy rampantly, and not just in traces. Psychology is fundamentally wedded to thinking in terms of what makes people act the way they do. If that is so, its viewpoint is muddled in a fundamental way.

It may be replied that the preceding argument is based on considering how ordinary language works, and that there is something wrong with the argu­ment since in ordinary language we in fact do speak frequently of the causes of a person's actions. The fact that we talk this way seems to defeat my argument. But actually, linguistic practice bears me out. It is true that we ask "what caused you to act that way?", "were you carried away?", "what made you do it, do you suppose?", "have you been under too much pressure?", and we say "you must have been overwhelmed!", "you were forced to do it!", etc. But the whole point of such locutions is precisely to portray us as victims rather than as agents, passive not active, moved rather than movers. When we talk this way we say that a person's "action" was not a bona fide action. We say that our bodily movement was in effect not our action but was owing to what acted upon us, e.g., some cause. It is by achieving that change in perspective that such expressions are exculpating. There may be a humane or a therapeutic value in portraying something seemingly our action as something we under­went; it may be humane to avoid a language which suggests blame. But the reasons why it is less judgmental to speak of what caused a person to act as he or she did is precisely to suggest that the person did not really act at all. For in so far as we are acted upon, we do not act.

If the preceding argument is sound, I think it has devastating consequences for the prevalent viewpoint of psychology and the social sciences. In effect the argument concludes that if anyone has ever acted, those actions were not caused; but the psychologist persuaded by the argument might be satisfied with the option that, from the psychologist's point of view, perhaps we should just admit that persons never really act. Maybe we can get along without that concept. Psychology is concerned with human behavior; if the concept of action is going to provoke conceptual difficulties, maybe it can just as well be abandoned.

To this I reply that the notion of a person is logically dependent upon the concept of an action (a person's action), so that the psychologist who gives up talking about actions will also have to give up talking about persons. This move will turn out to land psychology in absurdity. It will be obvious on a moment's reflection that the way we in fact think about persons is glaringly dependent on the notion of what person's do. When we tell others about ourselves, we tell them about what we do, what we like to do, how we spend our time, what sort of conduct they can expect from us. Even when we tell

4                        RUSSELL

others about something we have been undergoing, they learn about us when they learn how we act in response to what happens to us. Of course we talk about what people look like, how tall they are, or how heavy or thin. Plainly there are some things we say about persons that do not pertain particularly to what those persons do. But then if we want to say what they are like we talk about how they act.

That there is this conceptual interdependency between "persons" and "actions" may be established on a rudimentary level simply by consulting a dictionary and making a list of words one finds which may be meaningfully predicated of persons. Then remove the words which presuppose the concept of action. What will emerge is that the vast majority of things we want to say about human beings involve tacit or explicit reference to how those beings act. In fact, it will be words rooted in the concept of action—undertaking rather than undergoing—which enable us to have our concept of persons as distin­guished from other sorts of things. To abandon the notion of action would be to render hopelessly chaotic our comprehension of persons. But if this is right, there is something absurd about psychology. For I suppose that psy­chology aspires to help humans understand themselves, and yet that its perspective is incompatible with the concept of human action, hence also with the very conceptual tools with which we identify human beings. That is, if psychology is fundamentally wedded to viewing people as passive rather than as agents, i.e., in terms of what they undergo, then psychology does not tell us about us at all, but, at best, about what acts upon us. The absurdity of psychology is that it cannot address itself to human beings.

Actions Are Not Caused By Desires

Thus, my first reason for claiming that desires do not cause actions is that actions are not caused at all. However, someone might agree with the general drift of the argument, and yet not draw this conclusion. A reply might go like this:

"It is true that we cease talking about something as an action when we adopt toward it the perspective of something which happens to the individual. We do this when we do not attribute movements to the person. Attributing it to the individual is a matter of attributing it to his or her relevant desires, intentions, and beliefs. When a person's movement does accord with having certain relevant desires, intentions, beliefs, in such a way that the movement is attributed to these, that is when we regard the movement as done by this individual rather than as being undergone by him or her. It is just because the person's actions are caused by his or her desires (etc.) that they are actions of the individual. Something can be both an action of a person and also caused when the causes are 'internal to the person' instead of 'forces acting on the person.' This is just what many people suppose: human action is caused by the


desires, etc., of the agents."

My first reply to this objection is that if a movement of mine was "caused" by something called a desire, whatever that might be, then this movement would thereby be something happening to me, not something I really did. The fact that a desire might be conceived of as something "inside" me should make no more difference to this argument than would be made by the case of wires literally operating mechanically on my arm to make it move, which were embedded under my skin rather than pulling me from an "external" source. If my arm is being made to move, then, from that perspective and as long as that perspective is taken, this is not my action.

I think the above reply should be satisfactory. But I want to develop further an argument of a somewhat different sort regarding whether desires might cause actions.

I propose, to begin with, that there is a certain range of cases of desires in which desire and action are not sufficiently distinguishable to be causally related. It does seem plain enough that for two events to be causally related they must be distinguishable events. Whether one is more a "mechanist" or more like Hume in one's analysis of "cause," we need to be able to distinguish cause and effect if we are to claim that they are causally related. Yet in some cases desiring something is indistinguishable from the actions constitutive of that desire. For example, there is an unfaltering and spontaneous sort of reaching with gusto for a beer which is the very paradigm of what it is to desire a beer. Such action shows what a person wants in a very basic way. Such behavior just is what it is to desire a beer. Paradigm examples like this provide the very meaning of such concepts of desire, and make it possible for such terminology to be taught. That is why the examples and the desire are not separate events. The desire is not somehow "hidden behind" the behavior, as philosophers apparently used to believe, nor is it alluded to in some sort of indirect manner. Reaching like that for a beer provides something like the criteria for "desiring a beer"—not just a symptom.

Here it is instructive to compare how one "would not" reach like that unless one wanted it and how one "could not" make all those complex coordinated movements involved in grasping the glass unless one was con­scious of the glass' location. The "could not unless you were conscious" bears resemblance to an analytic proposition, and is not an inference extrapolated from experience. It is, rather, that the very meaning of what it is to be conscious of such things is built from such unambiguous examples as these. And one just would not reach for the beer like that unless one desired a beer, not because it is a foolproof inference, which it is not, but because it is in the ordinary case what that desire is all about. Scepticism rears its head here, as in all too many philosophical discussions, and it is objected that one after all might be just pretending to want a beer, or feel under some sort of coercion. And this is true. But then we would go wrong if we thought that desiring a beer

6                       RUSSELL

was acting like that plus something else, like a special mental event. On the contrary, it is acting like that in the relevant circumstances and minus some­thing like the special circumstances of coercion or pretense.

For my present purposes I do not need here to prove that in reaching thus for a beer there is not any mental "desire-happening" over and above the relevant action. My opinion is that indeed there need not be any such thing. But the argument requires less. Whatever "inner aspects" there might be to this beer desiring, the aspect of the actual behavior in the paradigm case is constitutive of "desires a beer" and is not incidental to it. When action provides such a paradigm for attributing desire, the desire and desire-behavior are not distinct. Yet if we are going to say that one event brings about another, and understand what we are talking about, these will have to be reasonably distinct. Here the desire and desire-behavior do not have the distinctness of entities which a causal account would require.

There is another variety of cases which are not paradigms of what it is to have a certain desire, but where we again find that the desire and the actions pertinent to it are not distinguishable in a way which makes a causal relation­ship between desire and action possible. These are cases where desire locu­tions pertain to what might be called long range patterns of behavior, rather than to any one paradigmatic episode. When we describe someone as desiring fame and fortune, or as wanting to be considerate, or desiring the happiness of his or her family, we make broad sketches of their character, and are, many times, making note of emerging patterns over a prolonged period. Here, again, we are not necessarily saying anything about some mental event or entity which is different from this pattern and which might be related to it causally. Much of what psychotherapists refer to as "unconscious desires" fall within this category of descriptions of broad behavioral patterns, and it is for this reason especially interesting to note that these desires, when undistinguished from the patterns they bring into focus, are not properly causes, though one often finds them portrayed that way.

I have proposed, then, that in cases where desire terminology simply brings into focus a long range behavioral pattern, and also in cases where actions paradigmatically illustrate a certain desire, desire cannot be the cause of that action. My remarks are a weak version of an argument by Melden (1961), in Free Action, which holds quite generally that a desire cannot be an internal event which produces the doing which it is the desire to do, since the concept of desiring anything is logically dependent upon the relevant doings. I have not wished to side with Melden's belief that desires are never sufficiently distinguishable from their related doings to be related to them causally. I am satisfied that it is pretty significant just to have brought it out that some desires are not the sort of items as can be made out to be the causes of the actions to which they pertain.

However, there appear to be many cases in which we can distinguish action


and desire. We can easily imagine a person who does not reach for the beer he or she desires, or who does not reach for it until after deliberating and worrying over how fattening it would be. Here desire and action seem separable, so that with respect to the foregoing objections it would make sense to propose that the desire causes the subsequent action. And yet even for these cases it will be important to remember that the general concept of this desire is not independent of paradigm examples. On the contrary, we would not have any idea of what it is for a person to want a beer if it were not for cases like the straight-forward one of reaching for a beer. The concept of wanting a beer but not reaching for it is logically parasitic on the concept of just reaching for a beer "because one wants to." To the extent that this is so it will be improper to construct causal relations between desire and action. But as this point may not carry the persuasive force it perhaps deserves, let us consider these matters from other directions.

It may be said that the causal theorist is not obliged to say why events follow one-another, nor to think in terms of one event's mechanistically "making another happen"; the task is simply to describe the correlations between events which in fact exist. On this model of causality, it ought in theory to be possible to discover correlations between desire-events and action-events, the descriptions of which would be causal laws (i.e., laws of statistical regularity).

However, I think it can be shown that in order just to describe these factual relationships the theorist will be forced to say that a condition for a correla­tion between beer desiring and beer drinking is that the subjects desire to heed their desires (assuming still, that the desires are distinct from the acts of beer drinking).

After all, it is not true that people always act on their desires, and a causal account (a description of de facto correlations) must identify those classes of cases for which this correlation does not hold. One such class is that where people have overriding desires contrary to heeding their desire for drinking a beer. Formulation of the correlation between beer desiring and beer drinking will state the exclusion of this class (along with other classes of counter-conditions, e.g., where there is no beer, where one is under oath not to drink, etc.). May we say that this exclusion means that the person must desire to heed this desire? Not yet. For, odd as it sounds, perhaps it is conceivable that a person might desire a beer and yet be indifferent as to heeding the desire. But now if we are to exclude this class in our formulation of the correlation, we have in effect said that the people for whom there is a correlation between beer desiring and subsequent beer drinking are a class restricted to those who desire to heed these occurrences of desire. Well, what, then, is this "desire to heed one's desire?" It is a concept which is logically dependent upon paradigm cases of actually heeding one's occurrence of desire, where the desire to heed and the heeding are not distinguishable as separate events. And so we come to a required element in the causal correlation between desire occurrence and

8                                                RUSSELL

subsequent action which is itself a desire (to heed one's desire) which does not lend itself to analysis in causal terms, for the reasons we considered earlier. To restate: in order to propose a causal relationship between desire and action, where these are distinguished, we need to describe the cases which do not fit, among which will be cases which require us to stipulate that the person be one who desires to heed his or her desire; but here this desire is not a cause. Hence, the attempt at a causal analysis remains crucially incomplete and unworkable.

Why should it happen that my desiring a beer, when this is separate from my reaching for the beer, should be followed by my act of reaching? When we are not artificially trying to impose a causal perspective, the obvious answer to this question is, "Because I want to." Whether this will satisfy us as an explanation depends on what we feel needs to be explained, and what we think can count as an explanation. My opinion about much of what in fact counts as explanation of human action, is that we are puzzled when some behavior stands out as unfamiliar, and regard it as "explained" when we see how to describe it in a way which integrates it into a pattern with which we are familiar.

Could the supposedly distinct desire to drink a beer be comprehensibly related to the subsequent drinking unless we suppose that I am conscious of this desire? We might agree that a person could not reach for the beer glass (with all the refined and coordinated movements which characterize reaching for something) unless one were conscious of it. Coordinated and meaningful responses to such things as beer glasses simply count for what it is to be conscious of beer glasses. By the same token, responding appropriately to this "occurrence of the desire" perhaps ought to count just as paradigmatically for consciousness of that event. That is, you could not respond meaningfully to this desire without being conscious of it for the reason that such response would count paradigmatically as consciousness of the desire. Similarly, it is a point of logic that you would not respond to this occurrent desire unless you desired to do so. Your desiring to respond to your desire simply is your responding to it, and it is exactly this which we should use to try to explain what it is to desire to heed your desire. Then this desire cannot be the cause of the heeding. We do not know how to make a connection between desire and action once we have distinguished them unless we make certain moves which exclude the concept of causality. For either the action of reaching for the beer is explained in terms of the desire simply because it counts as a case of desiring a beer, (thus being integrated into a pattern with which we are familiar), or else it is explained as a response to the occurrence of a separable desire, in which case it simply counts as a case of desiring to respond affirmatively to the occurrence of that desire. In at least one of these places causality can get no foothold, since causal explanations require a distinction between the cause and the effect which is lacking when we say the desire simply is the relevant action.

I suppose, then, that in order for it to be comprehensible to us that a desire


distinct from action should be correlated with action, we must suppose that the person is conscious of the desire. For a familiar explanation of this situation would be that the person was conscious of having a desire, and desired to heed it. But some will contend that desires can exist in persons without their being conscious of them. I have agreed with this possibility of unconscious desires earlier, when I suggested that such desires consisted in emergent patterns in a person's behavior. But if this unconscious desire is to be something distinct from this behavior, i.e., where the criteria for having it is not simply that the person behaves in the relevant ways, then it seems to me this is a notion of desire which we do not understand, perhaps cannot understand, and which, in any case, has no connection with the English word "desire." Neither public nor "private criterion" could be relevant for estab­lishing the meaning of this "unconscious desire" terminology, since this is supposed to be a sort of desire distinct both from behavior and from our consciousness of it. Therefore we have not the faintest idea what could be meant by "unconscious desire" in this sense.

We might try comparing the idea of a desire of which we were not conscious but which somehow affected our behavior to the idea of a mental image or episode of "internal speech" which happened without our "seeing," "hearing," or "being conscious of it. That there should be a correlation between such episodes and subsequent behavior would be as inexplicable to us as if you said or wrote "have a beer" and, though I did not see or hear you, and was not conscious of the occurrence of this behavior of yours, I then had a beer. This would not make sense at all.

The case is no better with "mental events" which are not phenomenological analogues with behavioral events (Russell, 1980) such as speaking or seeing. For whatever these "occurrences within the mind" are supposed to be like, we must make sense of how I am to act in response to something of which I am not conscious.

The only other possibility which I can discern is that this occurrent desire is supposed to act on me without my needing to be conscious of it by doing so in a mechanical fashion, or some comparable manner of making me do some­thing. But then, in the first place, I would not correctly be described as acting but only as acted upon. This was the point of the first part of this paper. Secondly, then we would be using the concept of desire so differently from the way we are used to that this would not connect with our concept of desire, so that we could not understand a person who proposed that desires worked this way.

I shall conclude with a few phenomenological remarks on beer drinking. I am seated at a bar with a group of friends. I have just gulped down a beer. To be sure, we shall not say that my desire was "nothing more than" or "merely" my conduct: but it ran through my conduct like a theme, and did not stand out as a separate item different from the eager gusto with which I drank off the contents. It appeared, if at all, as simultaneously the meaning of my conduct

10                                              RUSSELL

and the subject of various anticipatory thoughts and comments, and these were constitutive aspects of my relationship with this glass, no one of them alone essential. I found myself enjoying it thoroughly. Now one of my friends has refilled my glass, and I start to reach for it but I pause as I reflect upon my earlier resolve to get home early tonight. My desire almost inhabits my body as the tension in my posture as I lean toward the bar and yet refrain from reaching out. Desire gets sorted out as an object for me when something goes wrong, when I cannot or for some reason will not simply do what I feel like doing, what I would find myself doing in the absence of obstacles. This desire now is an object which awaits my doing something about it. Indeed what I desire is quite exactly over there on the bar. What can it do to me? It invites me. Well, shall I accept or not? How comforting it would be here to consider myself as actually pulled by this desirable beer glass, as if it were not an invitation but a mechanical force. I think to myself, "Oh why not!", and reach for the glass—as if thought and action were one being, one coordinated act. Neither with the first nor the second glass did my desire appear as a something which made me act. For in the first case the desire did not appear at all; it did not act on me, it inhabited me in my movement, and was the meaning of what I did. In the second case the desire did appear as an object for me, but as such it no longer was me but something in the world which was there for me to do something about. For if and when mental words refer to mental events which appear as objects for my consciousness (which they usually do not), these events have a status exactly like situations and objects in the world around me; they stand in need of interpretation by me and await my doing something about them (Russell, 1978). Once my desire to drink a beer becomes an explicit wish which is an object for my consciousness, there is no more a causal connection between this and what I do than if you had expressed a desire for me to have another glass: I shall do so if I wish.

In the second part of this paper I have tried to show that human actions are not caused by desires. I think that with slight modifications similar arguments could be given to show that they are not caused by intentions or beliefs either. All this has been meant to support the thesis of part one, that actions are not caused at all. I hope I will not be misunderstood to be holding that human actions cannot be explained (or predicted). My view on this (larger) issue is that we explain actions by giving an account of what the agent desired, believed, intended, and in related language of that sort, but that the way this language works is misunderstood when it is cast into a causal perspective.


Melden, A.I. (1961). Free action. New York: Humanities Press.

Russell, J.M. (1978). Sartre, therapy, and expanding the concept of responsibility. The American Journal of Psychoanalysis, XXXVIII(3), 259-269.

Russell, J.M. (1980). How to think about thinking. The journal of Mind and Behavior, i, 45-62.