Action, omission, and the stringency of duties.

• Author: Kamm, Frances M.
• Position: Symposium: Act & Crime

In this Article I examine the way in which Michael Moore in Act and Crime(1) distinguishes between action and omission, and his reasons for thinking that negative duties are more stringent than positive duties. In Part 1, I begin by summarizing points of his argument, including his description of views alternative to his own. I then describe my basic points of agreement and disagreement with Moore in Part II. Finally, in Part III I elaborate on some finer contentious points and on positions we both share. My discussion focuses on the first three chapters of Moore's book since these chapters discuss issues relevant to my concerns.(2)

1. MOORE ON ACTION/OMISSION AND THE STRINGENCY OF DUTIES

   For Moore an act is a willed event, in particular a willed movement. An omission is a willed nonevent, hence a nonact of a particular sort, a nonmovement of a particular sort. In formulating this first thesis, Moore reviews and rejects at least four other ways of formulating the act/omission distinction: (1) a grammatical test; (2) a moral test, that is, if it is not wrong, it is an omission; (3) the Relativized Baseline Thesis, that is, an act is what makes the world worse than some baseline condition, and an omission only keeps or returns the world to this condition, where the baseline for different agents differs; and (4) a view (developed by Leo Katz)(3) that a person's act is in question when an event would not have occurred if she had not existed, and a person's omission is in question if an event would have occurred even if she had not existed.(4)

   Moore's second thesis is that negative duties require us not to do acts which cause states of affairs to be worse, and positive duties require us not to omit acts that cause states of affairs to be better.(5) One can only make the world worse by an act, for it alone can cause harm.(6) An omission (of an act) which fails to make the world better therefore cannot cause harm.(7) Moore argues that this thesis depends on an analysis of causation which, he says, is stronger than the necessary condition analysis of causation, for omission of an act (or what is done while omitting to do the act) can be a necessary condition for harm even if it cannot cause the harm.(8)

   Moore's third thesis is that negative duties are more stringent than positive duties because not making the world worse is a more stringent duty than making it better.(9) One piece of Moore's evidence of this difference in stringency can be referred to as a defeasibility test--it takes more to defeat the negative duty than the positive duty. For example, he says, I may leave off saving one baby to save two, but I may not pull away a rope saving one baby in order to save two ("The Dying Baby Cases").(10) Two lives are enough to override saving one life, but not enough to justify killing one life.(11) A second piece of evidence is his intuitive sense, in cases originally presented by James Rachels,(12) that it is worse to kill someone maliciously than to maliciously refrain from the slight effort necessary to save a life.(13)

2. WHY NEGATIVE DUTIES ARE MORE STRINGENT THAN POSITIVE DUTIES

   Basically, I agree with Moore's characterization of the act/ omission distinction: that an act is a willed movement, an omission is a nonact of a particular sort, and that omissions do not cause events.(14) I also agree that negative duties are more stringent than positive duties, at least when all factors besides act/omission are held constant (for example, where motive, intent, and effort are the same in a case involving a duty to avoid an act and in one involving a duty to perform an act). I believe it is even likely that negative duties are more stringent than positive duties when a good motive and noble intention accompany the failure to perform the negative duty and a bad motive and ignoble intention accompany the failure to perform the positive duty. For example, there is a more stringent duty (1) to avoid the act of intentionally rushing to the hospital to save five friends when one foresees that one will run over someone en route than (2) to perform the act of saving a life when one's only reason for not doing so is a desire, propelled by personal gain, to have the person die.

   The negative duty to avoid acting is also more stringent than the positive duty to aid, in the sense that more effort could be required to perform the duty and more loss could be imposed for failure to perform. This is consistent with the fact that not performing a positive duty exhibits a worse character and is in some sense morally more objectionable than not performing the negative duty. This consistency implies that a difference in stringency is an indication of some moral difference besides moral objectionableness.

   I disagree with Moore's proposed explanation of the moral significance of the negative-positive distinction. He says that negative duties are more stringent than the positive duties, and argues that it is because we have a stronger duty not to make the world worse than we have to make it better.(15) But there are cases in which a negative duty is more stringent than a comparable positive duty, even though failing to perform the negative duty would make the world better. For example, I must not euthanize someone against her will even though death would be in her interest. Philippa Foot supports this view by saying that the duty of justice overrides the duty of charity.(16)

   The negative duties are, I believe, less concerned with not making things worse than with not interfering with that to which others are entitled, and showing respect for a certain sort of inviolability of the person. In particular, the moral duties not to interfere with what someone else has independently of us is more stringent than the positive duty to provide someone with more than she would have independently of our aid. Note that the duty not to interfere with what the person has independently of you holds even if what she has is retained only via the help of others. This thesis--call it the Entitlement Thesis--emphasizes that relative to a particular person, you are entitled only to what you have independently of the other person, and she is entitled to the efforts which she could make on your behalf.

   However, even the Entitlement Thesis seems not to go deep enough for purposes of explaining the distinction between negative and positive duties. Within the class of aiding behaviors which we compare with negative duties not to interfere, there are: (1) acts that we decide we have no duty to perform, (2) acts that we decide we have a duty to perform even without a contractual obligation, and (3) acts that we decide we have a duty to perform due to contractual obligations. The view of the world represented by the Entitlement Thesis may help explain why we decide certain acts fall into class (1): relative to some person, we are entitled to what we have independently of that person, and therefore she cannot claim it for use on her behalf. But the claim that negative duties are stricter than positive duties need not be merely the claim that there are no comparable positive duties. It may also be that the duties in categories (2) and (3) are not as stringent, even though the presence of a duty implies that we are not fully entitled relative to another to what we have independently of her, and may be subject to a claim for efforts on her behalf. Although we are not entitled to keep from her what we have independently of her, it may be that denying her our efforts is less bad than interfering with what she has independently of us.

3. THE ENTITLEMENT THESIS AND TESTS FOR STRINGENCY OF DUTIES

   These are the basic points of agreement and disagreement between Moore and myself. Now I shall proceed to embellish on these positions and explore the finer points of the issues involved.

   1. The Entitlement Thesis

      The Entitlement Thesis that I have proposed as a partial account for the distinction between negative and positive duties can help buttress the moral significance of the Relativized Baseline View which Moore criticizes as a way of drawing an act/omission distinction.(17) I agree that the Relativized Baseline View is not necessary to drawing the act/omission distinction; nonetheless, in conjunction with the Entitlement Thesis, it can help explain (despite Moore's denials) why terminating life support is often more like letting die than it is like killing, even though terminating aid involves an act. The person who has begun aid and terminates it, like the person who refuses to begin it, deprives another only of what she would get via aid, not something she has independently of the person who will assist or terminate aid. But an intruder who terminates aid that someone else has begun kills a person who is independent of her; she deprives that person of more than she would provide. This difference accounts for why the baseline for a person terminating aid she is providing is the condition the person was in before she began aid, but the baseline for the intruder is how she finds a person when that person is already receiving aid. Furthermore, contrary to what Moore argues, the Relativized Baseline View need not imply that we can kill someone we have saved whenever we want. For example, if we cease aiding at t, but the person would have continued living until [t.sub.2], then if we kill the person between [t.sub.1] and [t.sub.2] the person would lose time alive that we had not provided. She would lose more than we had provided and our relationship to her becomes that of any intruder.(18)

      My suggestion here is that there is a nonmoral criterion for determining when a termination of aid is a killing. That is, we do not call the termination of aid a killing merely because we think it violates a moral right. Essentially, removing a defense against a potential cause of death--whether the cause is one which had already threatened the person or is entirely new--is a killing if the...