It’s For Your Own Good: The Obligation to Violate Autonomy
This is the third in a series of essays I wrote as an undergraduate, honours, and then masters student in political theory. I’ve edited it somewhat, but essentially this is the work of a political science honours student, written in 1990. It was also presented as a paper at the 1990 conference of the Australasian Political Studies Association. This version is in memory of Margery Eagle, lecturer in political thought at the University of Tasmania.
Contemporary political philosophy sees respect for personal autonomy as an important basis of a just political system, yet our concern for the wellbeing of those close to us may override our concern for their autonomy. This paper discusses the concepts of paternalism and autonomy in the light of John Steinbeck’s novel Of Mice and Men. (Note that the paper assumes knowledge of its characters and plot. If you haven’t read the book I recommend doing so, but failing that, please read this synopsis before continuing.)
Of Mice and Men (1937) is a powerful novel, in part because it is a deceptive one. For nine-tenths of the book, Steinbeck carries us along with his descriptions of a foreign time, a foreign place, and an intriguing cast of characters. For nine-tenths of the book, we think we’re reading a simple story. Then, in the space of a paragraph, the world Steinbeck has constructed collapses, and from there we coast downhill to the book’s inevitable and tragic conclusion. Lennie has accidentally killed someone, and it is clear what his friend George will and must do.
Steinbeck plants all the prerequisites for the book’s conclusion in that deceptive nine-tenths, so that while we don’t expect what happens, we aren’t surprised by it. Rather, we are surprised by our own intense responses to it.
To me, the greatest surprise was that I thought George was justified in performing what is perhaps the ultimate paternalistic act: the mercy killing of another human being. Perhaps “justified” is not the right word: I couldn’t feel that George was wrong; I wasn’t sure that he was right; I simply thought his actions were inevitable.
Is this a reasonable response? The word “paternalism” has pejorative connotations in everyday language, while in political philosophy it implies a disregard for a person’s autonomy. It would seem difficult to justify paternalism in any circumstances. Yet Of Mice and Men suggests there may be times when paternalism is an intensely moral course of action. The usual feeling that paternalism is wrong does not seem appropriate here. In this paper I have attempted to find a set of conditions which make paternalism more “neutral”—conditions which must exist if we are to make any sense of a reading of Of Mice and Men. By considering the extreme case of Lennie, I hope to discover some conditions that apply to us all.
Paternalism seems to me more a matter of neutrality than right or wrong because it is inevitable, and it seems inevitable because people at times feel obliged to behave paternalistically. We feel such an obligation, to varying degrees, towards those we are close to. These close relationships may be between parents and their children, between spouses, between other physical relatives, or between friends. Since almost everyone is involved in such relationships, paternalism seems inevitable. This in turn calls into question the attainability, or even the desirability, of a society where respect for a person’s autonomy is of paramount importance.
In the first part of this paper I present some definitions of paternalism and autonomy and consider whether Lennie was or was not autonomous. I argue that moral autonomy is not enough to make a person autonomous. I then consider some cases of paternalism towards people who clearly are autonomous—cases which may be seen as inevitable. From there, I argue that paternalism is inevitable in close relationships, attempt to find what levels of paternalism could be considered justified in various cases, and finally draw some conclusions for autonomy.
Defining Paternalism and Autonomy
Gerald Dworkin (1988:107) defines paternalism as “the denial of autonomy [and] the substitution of one person’s judgment for another.” When the paternalist interferes in this way, he justifies his interference “by reasons referring exclusively to the welfare, good, happiness, needs, interests, or values of the person being coerced” (1988:121).
Does paternalism always involve coercion? Although it often does, I think George’s “mercy killing”1 in Of Mice and Men, while not coercion, is also an act of paternalism. Coercion necessarily involves communication between the coercer and the victim, in order for the coercer to convince the victim to act in a certain way (Raz 1982:108). George, in shooting Lennie in the back of the head, clearly does not communicate his intentions to Lennie; he is not interfering with any autonomy Lennie might have by trying to get him to change his behaviour. Rather, George’s interference with Lennie’s autonomy is in removing it altogether.
This doesn’t mean that every murder is a paternalistic act. George’s motives make his actions paternalistic. George kills Lennie for Lennie’s own good. Others may interpret George’s behaviour differently. Perhaps George kills Lennie as punishment, or to save Lennie from the terror of the lynch mob, or even to free himself from the burden of looking after Lennie. Some of these may contribute to George’s behaviour, but I don’t think any are the root cause of that behaviour. Rather, in the end George finally accepts what he has suspected all along: Lennie cannot function in normal society. The “place of their own” will always be a dream. Lennie will stumble from one disaster to the next, surviving only with George’s help. He will never achieve his goal, his ideal of happiness; at the same time, he will be a threat to others’ is now a threat to Lennie’s life. So George does the “merciful” thing and kills Lennie—the ultimate act of paternalism.
Onora O’Neill (1985:271), discussing Kant’s rules for treating human beings as persons, says “the paternalist tries to express beneficence or love by imposing a conception of others’ ends or interests. ... Paternalism towards those who have their own ends is not a form of love.” She suggests that paternalism on the one hand and respect and love on the other are mutually exclusive (1985:265). Yet George, until the final scene, shows Lennie respect and love. He makes Lennie’s ends (getting the “little place of their own”) in part his own, as Kant and O’Neill require. Only when George knows these ends are impossible does he act paternalistically. His actions come out of respect and love for a tragic human being.
Dworkin says paternalism involves the denial of autonomy. This raises further questions about George’s behaviour. Does Lennie have any autonomy to be denied? Can George therefore be paternalistic at all? To help answer these questions we must find what makes a human being autonomous.
John Benson (1983:6) gives a definition which sounds reasonable:
The autonomous man has a mind of his own and a will of his own. He exercises independence in his thinking and in his decisions about practical affairs.
But others give a narrower definition. Arthur Kuflik equates autonomy with moral autonomy. To Kuflik, “what matters, from the stand-point of autonomy, are the principles by which a person lives” (1984:272). Moral autonomy is “a kind of higher-order control over the moral quality of one’s life” (1984:273). Kuflik convincingly argues that this kind of autonomy cannot be alienated: that is, a rational human being can’t abdicate his power of moral decision-making to another person. If that person were to order him to do something morally dubious, the rational human being would at the very least think about the morality of the ordered action. In thinking about it—even if he physically carries the orders out—he is acting as a critically reflective moral agent. He remains morally autonomous.
It could be argued that Lennie demonstrates a degree of moral autonomy. In some ways he is not wholly rational, and at times he needs George to tell him how to act (notably, when to fight others). But Lennie might have (unconsciously) appointed George as a “moral guardian,” with that guardianship subject to review—something which Kuflik allows (1984:291). There are strong signs that Lennie has a sense of right and wrong:
[Lennie] looked down at her, and carefully he removed his hand from over her mouth, and she lay still. ‘I don’t want ta hurt you,’ he said, ‘but George’ll be mad if you yell.’ When she didn’t answer or move he bent closely over her. He lifted her arm and let it drop. For a moment he seemed bewildered. And then he whispered in fright: “I done a bad thing. I done another bad thing.”
Lennie also has a clear goal in life, a dream he earnestly wants to see realised:
[Lennie] laughed delightedly. ‘Go on now, George.’
‘You got it by heart. You can do it yourself.’
‘No, you. I forget some a’ the things. Tell about how it’s gonna be.’
‘OK. Some day—we’re gonna get the jack together and we’re gonna have a little house and a couple of acres an’ a cow and some pigs and ...’
‘An’ live off the fatta the lan’,’ Lennie shouted. ‘An’ have rabbits. Go on, George!’
Lennie is probably not fully morally autonomous. But he has a level of moral autonomy, which means that, if we accept Kuflik’s reasoning, he is somewhat autonomous. To me, this suggests that Kuflik is over-simplifying: there is more to being autonomous than maintaining one’s moral autonomy.
Significant problems arise if we regard moral autonomy as paramount. For example, a war criminal might argue that he was “just following orders” when performing his crimes, and that he inwardly disagreed with those orders. If this defence were allowed, then moral autonomy could be advanced as a defence in many situations. Since we cannot read people’s minds to see if they really are inwardly morally worthy, we would have to give them the benefit of the doubt, and every criminal could go free by pleading that external forces had irresistibly coerced him.
To rely on moral autonomy to define a person’s overall autonomy would also eliminate all major forms of paternalism, because, as Kuflik demonstrates, moral autonomy cannot be alienated. The only way to be paternalistic would be to act as George does, by killing someone “for his own good.” Any other actions would not be paternalistic, because we could argue that the person’s moral autonomy always remains intact.
Kuflik at one stage does hint that there is more to autonomy than one’s thoughts (1984:275, emphasis mine):
To function autonomously ... one must be able to survey alternatives, reach decisions, and make a sincere effort to act as one has decided one ought to do.
This effort, however, must be more than an internal process. It must involve some external demonstration of what might be called “practical” or “physical” autonomy. Otherwise, how can human society judge whether someone is autonomous? If an external demonstration is made, society can judge more easily. For example, a war criminal being tried at Nuremberg might argue that he had to follow orders or face being locked up or shot, and that he had inwardly disagreed with those orders. His defence would be taken more seriously if witnesses could testify that he had secretly smuggled Jews out of the concentration camps to safety.
It seems that overall autonomy involves thought and action. Benson’s definition of autonomy follows this pattern, as does that of Gerald Dworkin (1988:108):
I am defining autonomy as the capacity to reflect upon one’s motivational structure and to make changes in that structure. Thus, autonomy is not simply a reflective capacity but also includes some ability to alter one’s preferences and to make them effective in action.
Autonomy is linked to activity, to making rather than being (Dworkin 1988:112). Joseph Raz (1982:110) gives a similar definition:
(Significantly) autonomous persons are those who can shape their life and determine its course. ... In a word, significantly autonomous agents are part creators of their own moral world.
An objection here might be, “What about a rational person who is locked up against his will, perhaps even unjustly? He has his own goals, his own moral sense, but is prevented from acting on them. Isn’t he still autonomous?” The simple answer is no. Joseph Raz states that “harsh natural conditions can reduce the degree of autonomy of a person to a bare minimum just as effectively as systematic coercive intervention” (1988:112). A person’s autonomy is reduced by outside forces. Otherwise, a person free from outside influence could not be held up as an ideal of autonomy, because anyone maintaining his moral autonomy would be fully autonomous. The denial of autonomy would be a meaningless concept, because even a prisoner would be fully autonomous.
Is Lennie autonomous? Clearly, in the light of these definitions, he is not. He has a “life plan” (he wants a “little place”), but there is no way he alone can work towards its achievement: he needs the help of others, mainly George, to see it realised.2 He has a sense of right and wrong, but his own physical strength results in his doing “bad things” without intending to—he may not want to harm a mouse, but he often does. Although he has a self-directed moral framework, Lennie does not have control over the actual unfolding of his life. Trapped inside his unwieldy mind and body lies a morally worthy spirit, but this doesn’t change the fact that Lennie is not an autonomous person.
But if Lennie is not an autonomous person, how can he be susceptible to paternalism, if paternalism involves the denial of autonomy? Here Dworkin’s definition of paternalism needs some clarification.
The key to acting paternalistically is the belief that the action is for the good of the subject; if there were no more to paternalism, then we could behave paternalistically towards any human being, whether autonomous or not. The denial of autonomy is a significant part of paternalism, but this still leaves a wide range of human beings who can be treated paternalistically, because every person has some autonomy which can be denied.
Gerald Dworkin (1988:111) suggests that, while only some people are autonomous, every person has some autonomy:
Although all (normal) persons have [a capacity for autonomous development] it is not the case that all have an equal capacity. ... Our conception of a moral agent is of a creature who possesses this capacity above some significant threshold.
That is, only those above some “significant threshold” are autonomous.
Setting a fixed threshold of autonomy would be an impossible task, but on a case-by-case basis it should be possible to decide whether a person is clearly autonomous, clearly not autonomous, or borderline. I think it is fair to say that a person who possesses only moral autonomy (such as Lennie) is not autonomous, and hence falls below the threshold. But moral autonomy would seem to be a prerequisite for those above the threshold: for example, a person who has been brain-washed into believing in certain values, yet who can act autonomously in day-to-day practical matters, cannot be considered an “autonomous person.”
Even a person who falls below this threshold and is clearly not autonomous, and perhaps even a person who isn’t normal or rational, can be treated paternalistically, because somewhere he has some kind of autonomy—even if only to a very small degree—which could be denied “for his own good.” This suggests that there is only one kind of human being we cannot treat paternalistically: the non-person. An example would be a human being in a vegetative state whose brain is not functioning. A human paralysed from head to toe would not be a non-person, because he may have a normally functioning mind; and someone in a deep coma would not be a non-person, because he might at some stage come out of the coma and be perfectly normal.
Can Paternalism be Inevitable?
If we can treat any person paternalistically, can this paternalism ever be unavoidable or inevitable? Most people do not take pleasure in being paternalistic, but at times we feel obliged to do so. Why? And who is it permissible to be paternalistic towards?
One line of reasoning might consider the “degree of autonomy” suggested by Dworkin. Perhaps we will be inevitably paternalistic only towards people who aren’t autonomous (that is, who fall below the “significant threshold”). Or perhaps we will be inevitably paternalistic only towards people below some lower threshold. I don’t believe either is the case, because it is possible to imagine examples of inevitable paternalism between autonomous persons.
One example has been given by Dworkin (1988:107): a husband who hides his sleeping pills because he fears his wife’s suicidal tendencies is being paternalistic—he is denying her the autonomous right to decide whether or not to keep living.
A second example is that of a wife pressuring her husband into giving up smoking. Assuming that the husband knows the possible consequences of smoking and has decided that the benefits outweigh the risks, the wife is being paternalistic—she has assessed that smoking is not good for her husband, and is acting upon that belief.
A third example is that of parents pressuring their son into following a particular pat in life—for example, studying medicine—when he favours a different one—say, writing poetry. Peter Weir’s 1989 film Dead Poets Society shows a similar example.
These examples deal with paternalism prompted by important or fundamental matters, rather than trivial matters (such as a father telling his daughter to stop wearing red dresses). An initial conclusion to be drawn from these “fundamental” examples is that paternalism on our part seems inevitable when someone adopts what we see as life-threatening behaviour. This behaviour might be one of three types:
- behaviour that presents direct physical danger to the person who adopts it (for example, the wife’s suicidal behaviour);
- behaviour that presents a long-term but real threat to the person’s health or life (for example, the husband who smokes); or
- behaviour that threatens a person’s chances of having a happy and fulfilling life (such as giving up a secure career to be a poet).
In the case of George and Lennie, it is clear that George feels Lennie has adopted life-threatening behaviour (or, to use my earlier words, George feels that Lennie cannot function in normal society and will never achieve his ideal of happiness). But feeling that somebody has adopted life-threatening behaviour does not automatically entitle or oblige us to behave paternalistically towards that person. If we see a stranger smoking in the street, most of us won’t go up to him and demand he give the habit away. If one of the ranch-hands who’d known Lennie for a day or two had killed him and then claimed to have done it “for Lennie’s own good” or “to put him out of his misery,” we would think either that he was lying or that he had no right to act so paternalistically.
Thus, there is something more to “inevitable paternalism” than preventing self-inflicted damage. I think it is that paternalism is inevitable within close relationships. This is especially true when “life-threatening behaviour” is involved, but also when dealing with less extreme matters.
Close relationships can be entered into voluntarily, such as those between friends or between husband and wife, or involuntarily, such as a child’s relationship with his parents (although after a certain age even this relationship could be considered voluntarily—the child could, if he wished, break off contact with his parents). The critical factor is that when we are in a close relationship with another person, we care about that person. We care about how happy he is, whether he is succeeds or fails in life, and whether he lives or dies. If we stopped caring about him, we could no longer be said to be “close” to him.
The need to maintain a close relationship with a person who is important to us will occasionally oblige us to be paternalistic. Because we care about that person, we will try, as we see it, to “help” him if we feel he is “going wrong” or threatening his own happiness. If we do nothing, we are not acting as a friend should. The words we use when “keeping clear” from someone are often something like “I don’t care what you do”; if we say this about any non-trivial thing the person does, we are indeed showing that we do not care about him. Even if we feel we should do nothing because we agree with what he is doing, we should at least indicate that this is so. To remain silent is to imply we are uninterested.
A corollary of this is that to enter a close relationship, even through a voluntary, autonomous decision, is to invite paternalism towards oneself, and to threaten one’s own autonomy. For, just as we care about the other person and feel we must help him, he will care about us and feel the need to help us. Not only this, but we will also modify our own behaviour and our own autonomy to satisfy the person we care about. To quote O’Neill (1985:265), we “make the other’s ends, whose achievement would constitute his or her happiness, in part [our] own.” George provides a clear demonstration of this—he shares Lennie’s dream of the “little place,” even though at times his heart seems to lie elsewhere:
‘God, you’re a lot of trouble,’ said George. ‘I could get along so easy and so nice if I didn’t have you on my tail. I could live so easy and maybe have a girl.’
Of course, the paternalistic behaviour we will sometimes feel obliged to adopt may cause problems. We may act paternalistically when we really should (according to some neutral code of behaviour) refrain from interfering. This might be the case with the son who wants to be a poet: the parents who pressure him into becoming a doctor might in fact be sentencing him to a miserable life. The problem is that we can only act according to an individual sense of right and wrong. We can’t refer to a “neutral code of behaviour” which tells us when and when not to act paternalistically, because no such code exists. We can only act according to our own sense of what is good for the person we care about, and what is not. But we will still feel obliged to act, even if it may turn out that we were wrong to intervene.
Our own views, if they are at odds with those of the person we care about, may cause ill-feeling if we try to force them upon that person. For this reason, most people try to avoid paternalism over trivial matters, for fear of exhausting the other person’s patience and causing him to break off the close relationship. But over fundamental matters, which are usually of the “life-threatening” kind, our concern for the person we care about will outweigh these considerations, and we will feel obliged to be paternalistic. It may be that this paternalistic behaviour will cause ill-feeling and will result in the end of the close relationship. For example, the son may become a poet anyway, and end up drifting apart from parents he considers do not understand him. But all that this shows is that it is difficult for people who disagree over fundamental things to maintain a close relationship.
These thoughts help explain my initial reactions to Of Mice and Men. As far as we know, George is the only person alive who really cares about Lennie.3 He knows that Lennie is no longer able to lead any reasonable sort of life, so feels obliged to help this person he cares about. That George will act paternalistically seems inevitable. But for any of the other characters to do as George does would seem wrong, because they have little or no closeness to Lennie.
How Much Paternalism?
One question about inevitable paternalism remains. We may feel that a person close to us is going wrong, and that we are obliged to be paternalistic to set him right. But how far do we go in trying to change his behaviour? In other words, what level of paternalism is appropriate? The level of paternalism we adopt can range from simply saying that we think someone is doing the wrong thing, through to a protracted argument, through to physical restraining, through to the extreme of a mercy killing.
Determining an appropriate level of paternalism is, I believe, where the subject’s “degree of autonomy” can play a role. The greater the degree of autonomy a person has, the lower is the level of paternalism we are justified in adopting towards that person. When dealing with an “autonomous” person, we may be right in telling him our opinions of his behaviour, and perhaps in arguing our case more strongly, but we would clearly be wrong to go much further. The wife of the man who smokes, for example, would be over-reacting if she locked away all his money, or locked him up, although she would not be if she chose instead to argue her case strongly on various occasions.
Conversely, when dealing with a person with low levels of autonomy, a higher level of paternalism might be appropriate. This is especially the case when we are dealing with those of limited experience—for example, parents are perfectly justified in keeping medicines out of the reach of small children, to protect them from poisoning themselves. The parents dealing with the son who wants to be a poet may apply this logic—they may feel their son has limited experience of life, and that he’ll thank them for their intervention when he’s older. But as the son becomes older, more experienced, and more autonomous, the parents should feel progressively less justified in adopting anything but a low level of paternalism towards him.
The level of paternalism appropriate in a given situation will also be modified by the seriousness of the situation. When someone is behaving in a manner which directly threatens his life, another person close to him will feel justified in taking physical steps to restrain him. An example is the husband who hides his sleeping pills from his suicidal wife. But even here we would take the subject’s degree of autonomy into account. The husband would feel less sure about actually locking his (autonomous) wife up, or committing her to an asylum.
In extremely serious situations, even complete strangers will feel compelled to act paternalistically towards others, perhaps even to a significant level. They might feel that their compassion for humankind makes them close enough to a person to intervene, if that person’s behaviour is serious enough. Strangers will try to dissuade someone standing on the ledge of a tall building from jumping off. That they may feel a sense of guilt if they don’t try (as in Albert Camus’s novel The Fall) is another indication that we sometimes feel obliged to act paternalistically.
The case of Lennie is obviously a combination of all the extremes mentioned. Lennie’s autonomy is extremely limited. His situation is very serious, and the possible futures he faces are all disastrous. George feels obliged to act paternalistically because he cares deeply about his close friend, and he chooses the highest level of paternalism, that of killing Lennie, because of the extreme circumstances.
We will often feel obliged to be paternalistic towards people close to us for the very reason that we care about them. Since it is self-evident that human beings need close relationships, it seems that some paternalism is inevitable. If we accept that caring and paternalism are linked as I have outlined, then it would even seem that paternalism is desirable, inasmuch as we desire that others care about us.
This does not mean that paternalism cannot be “wrong” (or perhaps “inappropriate” or “uncalled-for”). Paternalism is wrong when someone is gratuitously paternalistic—when he feels no obligation to express his concern for a friend, or to restrain a loved one he thinks is endangering himself. He may be attempting to impose his idea of what’s best for another person merely because he enjoys dominating others. This behaviour contributes to people’s negative perception of paternalism, especially when the matters prompting the gratuitous paternalism are trivial.
The victims (or, to use a more neutral term, objects) of paternalism might also see paternalism as wrong in cases where it is inevitable. The husband who smokes might feel his wife is wrong in trying to get him to stop, no matter what obligation she feels she has. In this case, the husband is confusing his views about her actions, actions which are in fact an inevitable result of her concern for him, with his views about her opinions, opinons which he thinks are wrong (he doesn’t think he should give up smoking). These sorts of misunderstandings often result in paternalism being seen in a negative light. But perhaps this is no bad thing—it means that our natural impulses are to avoid paternalism whenever reasonable, which reduces the occurrence of “gratuitous paternalism.”
Still, paternalism does seem legitimate in certain situations, as Of Mice and Men indicates. The question then is, how can the maintenance of autonomy be of primary concern when the very nature of human relationships makes paternalism and the denial of autonomy inevitable and perhaps even desirable?
The answer is that autonomy cannot be of overriding concern. We cannot ignore autonomy (after all, we need to know how autonomous a person is to know what level of paternalism is justified when dealing with him). But neither should we concentrate on autonomy as an ideal to the exclusion of all others. People value close relationships and a place in society at least as much as being self-directed.
The condemnation which paternalism usually receives in political thought is a result of a reductionism towards autonomy and the self that underestimates the influence other human beings have on everyone’s life. This condemnation usually stems from a concern about legal paternalism. Whether governments should or should not be paternalistic is a separate issue, but by promoting the ideal of autonomy as a defence against this form of paternalism, political philosophers seem to have overlooked the major role paternalism plays in relationships outside the realm of politics.
I suspect that of the two, maintaining autonomy on the one hand, and maintaining human relationships on the other, most of us would think that the latter was more important. People value love and friendship highly. But unfortunately, as Of Mice and Men demonstrates, love and friendship sometimes impose upon us unpleasant and painful obligations.
May-September 1990/March 2007
1. A term suggested by Slater (1974:131).
2. To some degree, we all need the help of other people to see our life plans realised. But Lennie is an extreme case: he needs help in almost all areas of life.
3. The only other person we know of who cared about Lennie, his Aunt Clara, is dead by the time the story starts.
Benson, John. “Who is the Autonomous Man?” Philosophy 58 (1983): 5-17.
Dworkin, Gerald. The Theory and Practice of Autonomy. Cambridge: Cambridge University Press, 1988.
Kuflik, Arthur. “The Inalienability of Autonomy.” Philosophy and Public Affairs 13 (Fall 1984): 271-98.
Lisca, Peter. The Wide World of John Steinbeck. New Brunswick, N.J.: Rutgers University Press, 1958.
O’Neill, Onora. “Between Consenting Adults.” Philosophy and Public Affairs 14 (Summer 1985): 252-77.
Raz, Joseph. “Liberalism, Autonomy, and the Politics of Neutral Concern.” Midwest Studies in Philosophy 7 (1982): 89-120.
Slater, John F. “Steinbeck’s Of Mice and Men (Novel).” In A Study Guide to Steinbeck: A Handbook to his Major Works, ed. Tetsumaro Hayashi. Metuchen, N.J.: The Scarecrow Press, Inc., 1974.
Steinbeck, John. Of Mice and Men. New York: Covici, Friede, Inc., 1937.