The Harm Principle


Morality, as with the law, imposes a limitation of the free conduct and agency of the individual, and comparing moral strictures with those of the law is a useful way to begin. A central question in political philosophy, especially in the Liberal tradition, is to what extent the state, though the offices of the law, should restrict the freedom of peple who make up civil society. The answer given by John Stuart Mill appeals to what has become known as the Harm Principle: “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 1989, ADD REF). While we can all agree that the preventing harm is an important, even overriding concern of the state, Mill’s statement of the Harm Principle prompts a number of futher questions. For instance, how does the state ‘rightfully exercise’ power, how, in practice, are laws to be formulated and enforced, and what penalties should be exacted on those who disobey them? These questions presuppose that we know what harm, or harms, are, for there can be no legislating against harm if one does not know what this is. This issue, unlike those concerned with legislation, marks ground shared between morality and politics: morality is concerned with the prevention of harm, but without enforcement of conformity by state power. The question naturally arises as to just what it is that should make people limit their behaviour, actions, choices, decisions, etc., in order to avoid harming others. I will come back to this matter when we have come to some decisions about what harming and harms are.


One way to state what harm is is simply to give a list of all those acts that count as acts of harming. This would be time-consuming, generate a long list and most probably omit things that should be included. Another approach is to categorise harms into types of kinds of act. This is essentially what Bernard Gert does (Gert 2005: 20-21). His moral system comprises ten moral rules, the first eight of which prohibit a particular kind of harming. For instance, here are his first two rules:


Rule 1: Do not kill

Rule 2: Do not cause pain                                                                                1.


Killing and causing pain are clearly harms and acts which kill and cause pain are clearly examples of harming – indeed, these are paradigm cases of harms and harming: in the normal course of events, no one ever wants to be killed or be in pain. In my previous work on weapons research I have adopted a version of Gert’s moral system and made use of the rules that specifically mentioned kinds of harms, such as those in 1. This is because it seems that the kind of harms caused by weapons research are most obviously physical injuries and death. Here, however, I am going to take a slighty different tack, and ask whether we can give a general account of harm, such that what are mentioned in Gert’s rules are particular kinds of harming. This will help us have a general understanding of what harm is.[i]


Joel Feinberg has considered this matter in his study of the moral basis of criminal law (in Feinberg 1984). Feinberg begins with Mill’s statement of the Harm Principle and then asks, as we did, just what is meant by “harm”. After some careful discussion, he says “only setbacks to interests that are wrongs …are to count as harms in the appropriate sense” (Feinberg 1984: 36). As to interests


…interests …are distinguishable components of a person’s well being: he flourishes or languishes as they flourish or languish. What promotes them is to his advantage, or in his interest; what thwarts them is to his detriment, or against his interest. They can be blocked or defeated by events of an impersonal nature or by plain bad luck. But they can only be ‘invaded’ by human beings…It is only when an interest is thwarted through an invasion [by human beings] that its possessor is harmed in the legal sense..One person harms another …by invading or setting back his interest. (Feinberg 1984: 34, original emphasis)


In this passage Feinberg says that the setting back of an interest of one person by another is harming in the legal sense; I will assume here that this is also true of harm in the moral sense.


On this basis, we can come up with following rule based on this interpretation of the principle :


 Do not invade, and so set back, the interests of others.


One should not invade the interest of another because that is to harm them, and moral persons should voluntarily limit their freedom to act by foregoing any action that would harm another person (or moral subject).[ii] So what we have from Feinberg is a single sense of harm, namely the setting back of an interest, which lets us formulate the principle. But does this really help? We began this section by stating Mill’s Harm Principle as the starting point for developing a moral system, looked for a way to characterise harm and harming, and now we have an account thereof in terms of interests. Does this not substitute one poorly understood idea for another one, one which seems somewhat abstract and artificial? Feinberg has however provided some helpful clarification: interests are about all aspects of our well-being, we flourish or do well when they flourish, we do badly and languish when they do. The kinds of harms Gert refers to in the two rules are clearly set-backs to interests. Things go badly for us when we are in pain, and that is certainly true if we are killed. The other rules of his system mention being disabled, lied to, cheated, deprived of freedom and pleasure, and clearly these too are set-backs to interests. Interests are what are important to us, essential even, and as Feinberg says, we flourish when they do, and we do poorly when they do.


But the moral principle just formulated should not be obeyed in all circumstances, come what may. Consider this example: would not aiding a runaway slave be a set-back to the interest of another, namely the legal owner of the slave, and as such is it not proscribed by the principle, and if that is the case, is there not something wrong with the principle? Slave owning was certainly an interest of those who lived in the South before the Civil War, as it has been to most slave owners in history. Slaves do upaid work and hence can be used for economic benefit. What this establishes is that the principle formulated above may be sufficient to fix the class of harms, but it is not yet a moral principle, for it cannot be morally permissible to support a practice such as slavery. In the paragraph after the passage quoted above, Feinberg says “One person wrongs another when his indefensible (unjustifiable and inexcusable conduct) violates the other’s right…” (Feinberg 1984: 34). Granted that we accept that no has a right to own slaves, then aiding runaway slaves or otherwise trying to undermine the institution of slavery is not to wrong slave owners: it is not to violate their rights because they do not have a right to own slaves. An interest in slave-owning is indefensible. This suggests:


HP: Do not harm others by invading, and so setting back, their rightful interests.


It is thus only the ‘rightful’ interests of others that are to be protected by morality; it is only harms to these interests that are also wrongs which morality forbids.


Setting back an interest that someone does not rightfully possess does not, according to HP, thereby wrong the person. Setting back an interest that the person does righfully possess is wrong but it may turn out to be justified. For instance, causing a little pain to a person, and so setting back her interest in not being in pain, by giving her a flu shot may prevent much more pain, associated with getting the flu, and this (should) count as a justifiction. The original judgement of wrongdoing will be withdrawn. What distinguishes these two classes of interests? An activity like slaveowning is harmful, but it is never justifiable.[iii] This suggests the following criterion for inclusion in the class of interests that are righfully held: either they are not harmful at all, or if they are, then they can be justified and hence not classed as wrongdoing. Rape, torture of sentient creatures, murder and racial and sexual discrimination are harmful, they are never justifiable and they are always wrong.[iv] Justification will now be discussed in more detail.



[i] I personally find Gert’s system very appealing, which might perhaps have led to my not recognising that others might want more to be said about what harm is. I take the opportunity to do so here. However, the following remarks about Feinberg’s views are intended to underpin Gert’s system, not replace it.

[ii] Many, myself included, think animals as well as humans should be ‘protected’ by morality, even though they are not moral agents.

[iii] The rules and strictures against harming that are warranted by HP are not be understood as absolute prohibitions – this is implied by the possiblity of justification and the withdrawl of charges of wrongdoing. However, interests that belong to the category of those that are not rightfully held generate a class of rules that are absolute: it is never justifiable to own slaves, to rape, to torture, etc.

[iv] These choices can be disputed; for example, some unfortunately believe that torture can be justified. Such questions would require separate discussion and would need to be decided on a case by case basis. One would need to look in detail at what counts as an acceptable justification.