From IVR encyclopedie

Jump to: navigation, search

by Roger Shiner

The Problem

A legal system performs many social functions (see, for example (Raz 1979, Chapter 9). Prominent among such functions is the promotion of the welfare and security of citizens and their property. H.L.A. Hart famously distinguished between laws that impose duties by which citizens must abide, and laws that provide facilities that citizens can make use of if they wish (Hart 1994, 27–9). The promotion of welfare and security involves both kinds of law. Criminal law directly imposes duties of action and restraint, as does the law of tort or delict. Contract law, for example, provides for the enforcement of agreements made but not kept. This function at the highest level of generality may be characterized as the use of law to preventor to provide compensation for, harm. `Harm’ may be characterized at this level of generality as `the thwarting, setting back, or defeating of an interest’ (Feinberg 1984, 33).

The law’s exercise of this harm-related function is typically achieved by the deployment of coercive force: the law lists fines or imprisonment as consequences of criminal conviction, and it also provides mechanisms for the enforced payment of damages and restitution. A crucial question for jurisprudence and legal theory, therefore, is the normative question of the parameters for the morally legitimate use of such force. The concept of harm — again, at a high level of generality — is widely held to be an essential part of the correct account of such parameters. The source of this view is John Stuart Mill’s famous statement that `the only purpose for which power can be rightfully exercised over any member of a civilized community, against his will, is to prevent harm to others’ (Mill 1977, 223). This statement became known as `the Harm Principle’. It is a basic tenet of a certain kind of democratic liberalism that the Harm Principle gives the correct answer to our question about the parameters for the morally legitimate use of the law’s coercive force.

However, it is intuitively clear that the matter cannot be so straightforward, if we substitute for `harm’ in the Principle a broad characterization such as `setback to interests’. If A runs over B’s pet rat with his car, B suffers a setback to her interests, but it is not clear that causing B this loss should as such be subject to coercive penalty: further questions remain as to how or in what state of mind such harm was caused. Moreover, many legal systems make simply the possession of small amounts of marijuana a criminal offence; yet it is not clear how simply the possession in itself of a small amount of marijuana constitutes a harm, as opposed, for example, to its dissemination or consumption. Even if the Harm Principle validly comprehends regulation of behaviour likely to result in harm, there are still many questions as to the proper degree of likelihood, for example. If the Harm Principle is a guide to the proper way for the law to exercise its harm-preventing function, it is at best an uncertain guide. Exactly what notion of `harm’ the Principle deploys, and whether that is a defensible notion, are matters for investigation.


As Feinberg pointed out (Feinberg 1984, 31), the basic core of the notion of `harm’, insofar as it is of interest to the law, seems to be that of a `harmed condition’; someone is in the condition of being harmed: a setback to interests has occurred. But clearly there are `harms’ so understood in which the law’s interest will rapidly vanish. A person whose heart failure results from the natural course of disease is in the same harmed condition (in this sense of `harmed condition’) as one whose heart failure results from the ingestion of deliberately administered poison. This point leads to the thought that the notion required for jurisprudential purposes is not so much `harm’ as `wrong’ or unjustified violation of right. Not all harms are wrongs (suppose I put the poison into your drink by accident). So the focus has to be on that subset of harms that are also wrongs.

What, however, is it that turns a harm into a wrong? What do we need to add, to define the relevant subset? Obviously, one element is causation: one important difference between the rock hitting you as a result of a landslide and the rock hitting you as a result of my throwing it lies in the cause of the rock hitting you. But more is needed. Consider the difference between my throwing the rock and you unforeseeably getting into its path, my throwing the rock without paying any attention to who might get hit, and my throwing it intending to hit you. The law’s interest in your `harmed condition’ of being rock-damaged is confined to the latter two cases.

This train of thought leads to what has been called by Duff the `conduct-cause-harm’ or `consequentialist’ model of wrongdoing (Duff 1990, 105–11; Duff 2002, 58). The coercive force of the law is properly used only against those whose conduct causes (or threatens or makes more likely) harm to another — in short, the law should proceed according to the Harm Principle. This model in terms of the Harm Principle raises two different kinds of questions. The first set of questions has to do with the scope of the Principle, and the second set has to do with the model itself.

Scope: Concerns about the scope of the model, even in the case of those who fundamentally accept the model as valid, fall into four chief classes.

i) Moral harm (see also LEGAL MORALISM): on occasion the law has tried to control coercively action that damages, not citizens’ bodies or property, but (as it were) their souls. Laws regulating pornography, for example, have often been phrased in terms of `a tendency to deprave and corrupt’: the underlying though is that harm is done by the dissemination of pornography, harm to those who are exposed to it. The harm is not physical, but moral: those exposed will become worse persons. The liberty-fostering perspective that gives us the Harm Principle objects loudly, of course, to such an interpretation of the Principle (for the latest forceful statement, see Sumner 2004; see also Feinberg 1988, passim). Note, however, that at issue here is the interpretation of the Harm Principle. That the legal moralists are wrong, if they are, is not given just by the Harm Principle on its own, but by the Principle as interpreted by the opponents of the moralists. The moralists take themselves to be respecting the form of the Principle, if not its original author’s intent.

ii) Offence: People take offence at many things, and, subjectively at least, the resulting feelings constitute a `harm’, a `setback to interests’. But should we allow the causation of such feelings to invoke the coercive power of the law? Again, the standard liberal approach is to take the Harm Principle as not including `harm’ in the sense of `offence’. Feinberg usefully distinguishes between `bare-knowledge offence’ and `profound offence’ (Feinberg 1985, 61–9), though these are not exclusive and exhaustive categories. `Bare-knowledge offence’ is felt simply as a result of being aware of the existence of things found offensive — a person would never themselves read or watch pornography, but they know it is out there, they know people do that kind of thing, and they find it offensive. `Profound offence’ would be the offensiveness to a Jew, for example, of anti-Semitic propaganda. The Harm Principle would seem to have to stretch a great deal further to cover the former case, not so far to cover the latter. Words do hurt, just as much as sticks and stones, though differently. But still the stretch is controversial. There is a big difference between giving the target of racial abuse a civil cause of action for defamation, and criminalizing hate propaganda. Analogously, the law of tort regards as actionable not only nuisance but also infliction of emotional distress where the `harm’ is simply psychological effects, although a connection is preserved with more mainstream harms by setting the evidential requirement  high. Nonetheless, the situation is as with (ii). The issue here is not the applicability of the Harm Principle simpliciter, but rather the interpretation of the Principle. Just what does `harm’ mean in the Harm Principle? Can it cover non-physical harms?

iii) Paternalism: Mill in his original presentation of the statement that became known as the Harm Principle also explicitly ruled out paternalism. The passage quoted above continues: `His own good, either physical or moral, is not a sufficient warrant’ — warrant to exercise power over him against his will, that is. But the issue is not so simple, especially in the case of legal paternalism (see Husak 2003). One difficulty is even in plausibly identifying legislation as paternalistic: much supposedly paternalistic legislation is defensible as promoting citizen and community welfare. Moreover, justifications for paternalism exist, especially what might be called autonomy-promoting paternalism. It is not clear that the exclusion of paternalistic legislation from the Harm Principle is an automatic consequence of the Principle, or whether it in fact stems from a range of different moral concerns.

iv) Inchoate offences: All legal systems proscribe, not merely the bringing about of harmful results, but also activities that stand in relations of various sorts to such results without bringing them about. It is against the law, for example, to attempt to murder, as well as to murder; to conspire to defraud, as well as to defraud; to risk injury, as well as to cause injury. It does not seem difficult to find ways of justifying such offences, because of their connection to the bringing about of harmful results and the culpability implicit in such a connection (on attempts, for example, see Duff 1996). Nonetheless, the Harm Principle is here being interpreted, not simply applied.

The `Conduct-cause-harm’ Model. A quite different concern about the role played by the concept of harm in the legitimacy of the law has to do with the `conduct-cause-harm’ model itself. The focus of the concern consists of two related points. Consider, first, the crime of sexual assault. In Canadian law, sexual assault is an assault committed in circumstances of a sexual nature such that the sexual integrity of the victim is violated. Compare that definition with a more conventional one in terms of the occurrence of penetration. The objection can be made to the conventional definition that, in characterizing sexual assault as a harmful result caused by conduct, the definition diverts attention away from the conduct itself, and so leaves out what is most fundamental and striking about such assaults — namely, that the victim’s autonomy is deeply violated, that the victim’s status as a person is repudiated, thast the victim is reduced to a means of gratification (Duff 2001, 24–5). The Canadian definition has the advantage that the expression `sexual integrity’ does not mean simply physical integrity defined by the outer limits of the body. The expression is not simply a disguised way of referring to bodily penetration. Instead, it is intended to cover such actions as groping, grabbing, slapping and the like, although degrees of heinousness are acknowledged. Violating integrity, however, is not an action that can be broken down into `conduct-cause-harm’. The description of the action itself says why the action is repugnant. We are dealing here with a case of, if you like, non-consequential harm. The harm at issue is constitued by the action, not caused by it. The dissemination of pornography can be represented as just such a form of harm (Shiner 1986).

This point can, second, be extended more widely. Consider again the Canadian definition of sexual assault, invoking the idea of `circumstances of a sexual nature’. This reference to `circumstances of a sexual nature’ brings out that the whole matter of the occurrence of harm so as to concern the law is fundamentally social in character. The harm-causing acts with which the law is concerned are socially embedded acts. The individuals who are causing and suffering the harm are also socially embedded. We fail to understand how it is true that harm is being caused unless we attend, not merely to conduct and the results of conduct, but also to the modes, contexts and embedded attitudes of conduct. An assault on a victim may contain precisely the same bodily movements as the caress of a lover. If there is harm done in one case, but not in the other, it is not because some bodily condition results from the one that does not result from the other. It is because different action descriptions are appropriate, drawing for their differences on the background stock of social concepts and norms in which the bodily movements are embedded.

Conclusion. These are large and complex issues, and are not settled here. But this much is clear. Within the liberal democratic tradition, the concept of harm can have a valid use at a high level of generality and abstraction to gesture towards a large area of the law’s concern with the control of human action. To encapsulate this concern in the Harm Principle is permissible. But to do that is not the same as using the Harm Principle mechanically as a concrete guide to the design of the law. It is much too inexact for that. If one is coming from a long way outside the liberal democratic tradition, then to arrive at the point where the Harm Principle seems to get everything right is hard, but worthwhile, work. However, within the liberal democratic tradition, to arrive at the Harm Principle is easy. The hard work lies ahead.


Duff, R. 1990. Intention, Agency and Criminal Liability: Philosophy of Action and the Criminal Law. Oxford: Blackwell.

Duff, R. 1996. Criminal Attempts. Oxford: Clarendon Press.

Duff, R. 2001. `Harms and Wrongs’. Buffalo Criminal Law Review 5:13–45.

Duff, R. 2002. `Rule-Violations and Wrong-Doing’. In Criminal Law Theory: Doctrines of the General Part, eds S. Shute and A. Simester, 47–74. Oxford: Oxford University Press.

Feinberg, J. 1984. The Moral Limits of the Criminal Law. Vol. I, Harm to Others. New York: Oxford University Press.

Feinberg, J. 1985. The Moral Limits of the Criminal Law. Vol. II, Offense to Others. New York: Oxford University Press.

Feinberg, J. 1988. The Moral Limits of the Criminal Law. Vol. IV, Harmless Wrongdoing. New York: Oxford University Press.

Hart, H. 1994. The Concept of Law. 2nd edn. Ed. P. A. Bulloch and J. Raz. Oxford: Clarendon Press.

Husak, D. 2003. `Legal Paternalism’. In Oxford Handbook to Practical Ethics, ed. H. LaFollette, 387–412. Oxford: Clarendon Press.

Mill, J. S. 1977. On Liberty. The Collected Works of John Stuart Mill, Vol. 18. Ed. J. M. Robson. Toronto: University of Toronto Press.

Raz, J. 1979. The Authority of Law. Oxford: Clarendon Press.

Shiner, R. A. 1986. `Pornography and Freedom of Speech’. In Freedom of Speech: Basis and Limits, ed. G. Maher. Archiv für Rechts- und Sozialphilosophie Beiheft 28, 11–28. Stuttgart: Franz Steiner Verlag.

Sumner, L. 2004. The Hateful and the Obscene: Studies in the Limits of Free Expression. Toronto: University of Toronto Press.

Personal tools