Three arguments about paternalism

 

 

Hans O. Melberg

(hmelberg@hotmail.com)

(National Institute for Alcohol and Drug Research)

University of Oslo

Oslo, 23. November 2000

 

Essay for the Ph.D course in Health Economics (Paul Doland)

 

Introduction

A Norwegian newspaper recently reported on a debate about whether it should be legal to parachute off certain popular cliffs (base-jumping). One participant argued that it should be legal because "the individual has the right to liberty as long as it does not directly harm others" (my translation). Against this it was argued that the risk of dying was "extreme" and that the death of an individual in fact did represent "direct harm" to others (e.g. relatives).[1]  The debate illustrates some of the problems involved when discussing paternalistic actions (What is harm? Does risk in itself constitute good grounds for intervention?) and the first purpose of this paper is to identify the key issues involved.[2] I then go on to make three arguments. First, I will present an account of why I disagree with those who want to avoid the debate about ethical issues like paternalism since - they argue - it no more meaningful than debating whether vanilla or chocolate ice cream is best; It is simply a question of taste and "de Gustibus non est disputandum." Second, I want to argue against those who write that all talk about a true self is wrong. Third - and related to the second - I will argue that there is a tendency to confuse philosophical and pragmatic arguments in the debate. For instance, the so-called slippery slope argument can be used to argue against paternalism on a pragmatic level, but the dangers of a slippery slope has no relevance to the question of whether it is possible to talk about a "true self" on the conceptual level.

 

The issues

John S. Mill's argued that "the only purpose for which power can be rightfully exercised over any member of a civilised community, against his will, is to prevent harm to others" (Mill 1859/1993, p. 78). We may call this the harm principle and it implies that paternalism is prohibited - you are never allowed exercise power in order to benefit a person against his own will.[3] At first sight the harm principle seems to severely limit the amount of intervention that is morally acceptable. A closer look, however, reveals that the terms are so ambiguous that it opens up the door for much more intervention than one might believe at first sight.

 

For instance, campaigns against smoking are usually considered paternalistic, but if power is defined narrowly as coercion, then an advertising campaign is not coercive (at least not on directly) and hence not forbidden on Mill's account. Moreover, if liberty is defined as "doing what one desires" (real will theory, positive liberty), then more interventions are acceptable than if one defines liberty as "freedom from interference" (negative liberty). Third, if you accept a theory that distinguish between the "true will" of a person and his "false self" then Mill's principle does not rule out interventions which we commonly label paternalistic (Arguable laws against drugs could fit here). Fourth, most people do not label intervention paternalistic if the person himself consents, but this raises the question of what type of consent that is valid. If we count subsequent and prior consent, we allow many more interventions than if we only accept current consent.[4] And, finally, as illustrated by the debate in the Norwegian newspaper: If "harm to others" is defined widely, then Mill's principle can be used to intervene against base-jumping. In short, the principle comes build in with so many ambiguities that its value as a guide to real policy-making is weak. The real force comes from the definition of the terms in the principle, not the principle itself. Of course, one might object to the principle itself also, but that is not the issue in this paper.[5]

 

I have so far argued that Mill's principle requires us to define a number of very controversial terms in order to give us an answer on real issues (see Table 1). The question then becomes whether it is at all possible to have a meaningful discussion about paternalism. If the definition of paternalism is so dependent on terms that have no value-neutral or uncontroversial meaning, is it not a waste of time to debate the issue? Isn't all you can really say just "I say yes/no because I feel that is the right answer" and that is the end of discussion?

 

Table 1: Some key questions and terms in the debate about paternalism

Rights	Power	Liberty	Self	Consent	Harm
- one or many? - positive or negative? - justification of (utilitiarian, contractual, natural)	- only coercive or also non-coercive? - diminish choice set or just change choice set	- positive or negative	- empirical self (revealed preferences) or also accept "true will" theories? (Not mutually exclusive)	- time aspect: accept only current or also prior and subsequent? - accept hypothetical consent? - accept only rational and informed consent?	- direct vs. indirect - physical vs. non-physical - Is all harm morally significant? (eg. envy?)

 

 

Reasons why meaningful debate is possible

In an article about justice in the legal system I once came across the following argument: "to use justice as an argument amounts to knocking on the table: it is an emotional expression … it is impossible to have a rational discussion with a person who mobilises 'justice' because he does not say anything about which we can argue for or against" (Alf Ross quoted from Melberg 1996, p. 475). I want to argue against this position.[6]

 

First of all it is possible that there are gaps in the information on which we base some of our values. My belief that human nature is good (or bad) may affect the value I attach to individual freedom and if you can give me new information about human nature that I previously had not considered, then I will revise my weighting. There is nothing irrational about this and it illustrates that a debate about values need not always descend into emotional expressions of "yes" and "no." Gaps in the information base creates some scope for meaningful discussion.

 

Second - and slightly more controversially - we may demand that a person's value system should have a certain logical structure. At the very least the person should not have values that are logically inconsistent. If I accept "right wing" intervention like compulsory military service, then the very same arguments used to justify that policy might also force me to accept a "left-wing" intervention like compulsory union membership in a factory. Since there may be logical inconsistencies in a person's value system, there is also scope for rational discussion. We may point out the inconsistencies (or implied inconsistencies) and the other person is then forced to revise his values. More controversially one may require that the values should not be logically self-justifying. This is discussed by Archard (1993, p. 342). In his words:  "The main worry that some have had about allowing that paternalism might be justified by subsequent consent is that the paternalism could create that very consent." For instance, compulsory treatment of drug users can be questionable even if all of them subsequently say they are now happy that they were forced to quit drugs (Why should the subsequent self be more important than the previous self?). Archard himself concludes that not all self-justifying paternalism is morally unacceptable, but that is not the point here. The key is that whether something is self-justifying is not a question of taste and for that reason it is possible to have a meaningful discussion.

 

A third reason why there is scope for meaningful discussion is the distinction between deep and shallow values. Imagine two people who disagree about whether some action is paternalistic or not - for instance about making the use of some drugs illegal. They both express an opinion based on certain values. The question is then whether there is scope for further argument and the answer may be yes if the values are "shallow." That is; when asked why they think the values are valuable they are able to give you a reason and this reason then constitutes the "deeper" value. If they agree on the basic value, one may then work logically to see whether the shallow values really follow from the deep values. For instance, when asked why they support or are against paternalism, they might argue that it is important to respect the autonomy of persons. You may then argue that if autonomy is the real value underlying the argument against paternalism, then this has implications of what we should label paternalism or not (i.e. only those acts which violate your autonomous preferences; see Scoccia (1990) for more on how the "autonomy justification" affects paternalism). This moves the debate to a deeper level and opens up scope for rational discussion. We may, for instance, agree that paternalism is acceptable to the extent it would be chosen by a rational person behind a veil of ignorance about his own position in society (rich/poor, male/female, high/low ability; good/poor judgement), but we may disagree on the implications that follow from this. Agreement on the underlying value opens up the possibility for meaningful discussion since it is not just emotions that are involved when we debate which shallow values that can be derived from a deeper value. And, even if we do not agree on the deeper values, out understanding has improved if we are able to uncover the deeper reasons for our disagreement beyond statements like "I like it" and "I do not like it."

 

Finally there is the fourth reason why I believe it is possible to have a meaningful debate about ethical issues. This is, in short, that a value that is meant to guide behaviour must be such that it does not demand the impossible. As an example, consider Nozick's argument that past infringements of rights should be compensated so the individuals are as well off as they would have been without the intervention. Sometimes this principle may work, but at other times we simply do not have the information necessary to use the principle; For instance, what is the correct compensation for being lobotomised in the 1950s when this was considered the best (paternalistic) course of treatment for certain people in Norway. We do not know what would have happened without the lobotomy and to try to estimate this requires precise knowledge of detailed causal mechanism - information which we do not have. Now, to point out that a value is not operative because it requires information which we do not have is not an irrational exercise like "knocking on the table." On the contrary, to work out the implications and information requirement is a deductive process and it can be criticised based on faulty logical connections.

 

In sum, I do not agree that there is no scope for meaningful debate about the definition of controversial terms like rights, power, liberty, the self and harm. This does not, of course, imply that it is always possible reach neutral and valid definitions of the terms. People may reach different (consistent) value-systems even if they have the same information and knowledge about the implications. Hence, at some point two people who disagree may reach the point where no meaningful discussion is possible. Before that, however, there is scope for meaningful exchange and it would be wrong to strangle the debate at the beginning with the argument that "this is all a question of taste." To do so would be to ignore the scope for discussion opened up by the four arguments above.

 

How important are the four reasons above; Do they carve out enough room for significant discussion, or will we quickly reach the point where no more discussion is possible? I have no good answer to this question. It depends on the extent to which our information base is filled with gaps, the extent to which our value-systems are logically inconsistent and the extent to which there is a common "deep-value" behind our surface values. And even if there is significant space created by my four arguments, there is still the problem of human weakness (such as pride) which may hinder the discussion. That is, even if I point out an informational gap or a logical inconsistency in your argument, you are reluctant to change your opinions because you dislike admitting mistakes. In sum, although I have argued that there is some room for valuable exchange, I do not want to appear too optimistic about the possibilities created by this space.

 

The reasons given above have some implications as to how the debate should be conducted in order to be useful. If we look upon the process as one of exchanging information in order to reach a "reflective equilibrium", then one valuable tool is to present as many examples as possible. We then examine whether there is a gap between our intuitive understanding of the "correct" answer to the example and the answer presented by our explicit principles. If there is a gap something has to give in - either the principle has to be revised or we have to accept that our intuition is wrong. Mill's example of whether it is morally acceptable to stop a man who is running towards an unsafe bridge is a classic illustration of this process. We may start from the position that intervention is never acceptable, but it is very hard to accept that conclusion in Mill's example and hence people either have to revise the principle or reject the intuition that we should stop the running man.[7]

 

I have so far argued against the view that all debate on paternalism always amounts to emotional expressions of likes and dislikes. I now want to turn attack a second argument: the view that all talk about a "real self" is nonsense.

 

False and real selves?

In an article about paternalism and drugs, Douglas N. Husak (1989, p. 365) makes the following argument:

 "… it may be contended that consumers do not want, enjoy, or derive satisfaction from recreational drug use. This strategy will be dismissed by economists who take wants as given and make policy recommendations designed to maximize the satisfaction of existing consumer preferences. And it is almost certain to be rejected by philosophers familiar with the sleight-of-hand that results from positing an alleged 'rational will' and then supposing that a person's 'true wants' can be identified apart from his expressed desires."

 

I think Husak is wrong both in his interpretation of the literature and the more substantial points. Substantially I think it is wrong to equate choice and preferences in the way Husak does (Note: the term 'expressed preference' is used by Husak to mean choice, not verbal expression). Less important, I believe there are plenty of economist and philosophers who disagree that talk about true selves is just a sleight of hand. Here is an obvious counter-example: A person is offered the choice between an apple and an orange. He selects the apple because he believes that the orange is poisoned. If this is a true description, then the inference from "he selected the apple" to "he prefers apples to oranges" is simply wrong. Thus, there is not necessarily always a "sleight of hand" involved when we talk about true selves.

 

The problem can be, and has been, analysed more systematically. For instance, in 1973 Amartya Sen published a paper which attacked how the concept of "revealed preferences" had been used in the literature. One of the main arguments was that the inference from choice to preferences is invalid when the choice was not made using rationality. To simplify slightly we may use the following illustration: If I hit a person out of anger, then it is wrong to treat that action as if it were selected to express my preferences using the process of rational and cold weighting of costs and benefits. And Sen is not the only economist to be interested in the issue. For instance, Hausman (2000, p. 99) has recently written a paper in which he argues that the "notion of 'revealed preferences' is unclear and should be abandoned." Hence, it is wrong to say that economists simply equate expressed choice and true preferences and work from there.

 

The debate is, of course, slightly more complicated. One objection could be that Sen, Hausman (and I!) misinterpret the meaning of preference that the adherents of revealed preference theory use. They would argue that they are simply saying that when the person selected the apple, he really preferred the apple to the alternative - the poisoned orange. To me this is to mix up beliefs and preferences in a way that makes the theory impotent in welfare economics. To evaluate the welfare effects of a policy change (should we provide more apples or oranges?), we need to know what people prefer. And "prefer" is here used in the in the sense of pure preference - not what he would choose if he believed X or Y. Another objection to my argument, could be that there is no alternative to use choices as equal to preferences. There is, it is argued, no alternative wisdom except your actions that allows me to infer what you really want. This, in turn, leads to the argument that if we allow a 'mystical' source to contaminate the inference to preferences, we are led to a paternalistic and totalitarian dictatorship in which some claim to know the true self of all the others. This leads me to my third argument: Why it is wrong to argue that talk about true selves should be abandoned since it leads to totalitarianism.

 

Paternalism and the threat of totalitarianism

In his well know essay "Two concepts of liberty", Isaiah Berlin (1958/1969) argued that:

"Once I take this view [that there is a 'real self' as opposed to an empirical self], I am in a position to ignore the actual wishes of men or societies, to bully, oppress, torture then in the name, and on behalf, of their 'real' selves …" (my summary in [])

 

Similar arguments are easy to find in writings about paternalism i.e. the worry that accepting a theory of a "true" self leads to 'too much' intervention or - plainly speaking - totalitarianism. I shall argue that the spectre of totalitarianism is irrelevant to the validity of the theory of true selves.

 

The argument is simple: Even if I accept a theory of "true" selves on a conceptual level, it does not follow that I accept paternalistic intervention in the real world. It is perfectly consistent to argue both that you sometimes know better than the other person what is good for him (here: his 'real' desires) and at the same time that you should not interfere. This may be so for many reasons. First of all, it is very likely that the interference in itself is counterproductive since people are easily offended if others try to interfere. Moving from the personal level to the state level, you may also believe that it is dangerous to give the government too much power to interfere. Hence, you may argue in favour of a corner solution of no-interference even if you believe that sometimes interference would be good (based on a theory of a true self). As a matter of practical politics you believe intervention will lead to a slippery slope in which the government starts to intervene in areas you do not think is right. The key point here is that the practical corner solution need not imply that you adopt a conceptual corner solution in the debate about true and false selves. In short, Amartya Sen should not stop criticising reveal preference theory simply because some future dictator (maybe a slave of a then defunct (!) economist) may misinterpret and abuse his ideas to create a dictatorship. An idea cannot be wrong because it can be abused and the argument that it can be abused is not relevant to the discussion of whether the idea itself is correct.

 

Finally, it is not true that there is no source of information about the self than expressed choice. As Sen (1999) has recently emphasised, we should to broaden the informational base in welfare economics. For instance, we could simply ask people about their preferences (as is now done in health and environmental economics). Moreover, we should accept the possibility that out actions are not fully consistent. This implies that the action of taking drugs alone is not enough to say that the person really prefers drugs. There may be other actions - for instance seeking treatment - that express contradictory preferences and this makes it necessary to look at more than one area of expression on order to reach a conclusion about what the person really wants. This is, or course, not to argue that current actions say nothing about preferences, only that words and actions at other times and places also matter.

 

Conclusion

Instead of trying to survey the area of paternalism, I have focused on three arguments relevant to the debate. Whether it was possible to have a meaningful discussion about paternalism; Whether the phrase "true self" as opposed to "the self expressed in choice" can be defended conceptually; And, finally, whether the "true-self leads to totalitarianism" argument is relevant to whether we should accept a theory of true selves. Since these were just three fragments relevant to the debate, there is no large conclusion that unifies them.

 

 

 

 

References

Archard, David (1990) Paternalism defined. Analysis, pp. 36-42.

Archard, David (1993) Self-justifying paternalism. The Journal of Value Inquiry 27, pp. 341-352.

Arneson, Richard J. (1989) Paternalism, Utility, and Fairness. Revue Internationale de Philosophie 43(170), pp. 409-437.

Berlin, Isaiah (1958/1969) Two concepts of liberty. In Isaiha Berlin (1969) Four Essays on liberty. Oxford: Oxford University Press.

Hausmann, Daniel (2000) Revealed Preference, Belief, and Game Theory. Economics and Philosophy 16, pp. 99-115.

Husak, Douglas N. (1989) Recreational Drugs and Paternalism. Law and Philosophy 8, pp. 353-381.

Kultgen, John (1992) Consent and the justification of paternalism. The Southern Journal of Philosophy 30(3), pp. 89-113.

Laselva, Samuel V. (1988) 'A Single Truth': Mill on Harm, Paternalism and Good Samaritanism. Political Studies 36, pp. 486-496.

Melberg, Hans O. (1996) Tre årsaker til urettferdigheter ("Three causes of injustice"). Lov og Rett 8 (1996), pp. 471-499.

Mill, John S. (1859/1993) On Liberty. In Geraint Williams (ed.) (1993) Utilitarianism, On Liberty, Considerations on Representative Government. London: Orient Press.

Scoccia, Danny (1990) Paternalism and Respect for Autonomy. Ethics 100, pp. 318-334.

Sen, Amartya (1973/1986)Behaviour and the concept of preferences. In Jon Elster (ed.) (1986) Rational Choice. Cambridge: Cambridge University Press.

Sen, Amartya (1999) The Possibility of Social Choice. American Economic Review 89, pp. 349-378.