Print This Post Print This Post

Chapter 3, “Intention, Harm and the Possibility of a Unified Theory” is focused on responding to Warren Quinn’s attempt to provide and ground accounts of the Doctrine of Doing and Allowing (DDA) and the Doctrine of Double Effect (DDE) in two very interesting essays that were published as Chapters 7 and 8 of Morality and Action (CUP, 1993).

This chapter is split into three sections, which I will now summarise in turn, raising issues that invite further discussion as I go along.

The first section concerns the DDA, i.e. the doctrine that there is a morally significant difference between doing and allowing harm. Quinn provides a particular account of the doctrine (discussed in section (a)), as well as an account of what grounds the morally significant difference between doing and allowing (discussed in section (b)). Kamm begins by criticising this “revisionary” account, then examines the proffered foundation for the account.

(a) According to Quinn’s revisionary account of DDA, we need to distinguish between “positive agency” and “negative agency”, the first of which we take to be what is involved in doing, and the second of which we take to be what is involved in allowing. This is revisionary because “positive agency” (doing) is meant to cover cases where the movement of an object is presently under one’s control even when one is omitting to act. In essence: some cases of doing nothing are cases of doing, while other cases of doing nothing are cases of allowing.

I notice that Quinn is particularly interested in the “powers of control” that people exercise over objects for periods of time (see his p. 160), but Kamm focuses in her discussion on foreseeing and intending “movement”. Of course, Quinn also discusses movement, but I take his basic concern to be with being in control of other objects (and people). Here is a charitably described example: When I get in my car to drive from my office to my home, I am in control of the car for the period of time it takes for me to get home. Suppose I like to take my hands off the steering wheel from time to time, just for fun (but never for long enough to make it the case that I’ve lost control of the car). I see an injured person a fair distance away, on the road in front of the car, at some point in time when I have already taken my hands off the wheel. I foresee that if I don’t put my hands back on the wheel and swerve, I’ll hit this person. It seems one might well want to say that my doing nothing in this case amounts to a doing, rather than a mere allowing. Perhaps this case is very different, in a morally relevant way, from a case where one allows a stranger to drown as one walks past a pond.

The first of Kamm’s criticisms concerns cases involving movement (without mention of control), and the second concerns cases where one begins life by imposing a burden. Neither of these are as important (or as perspicuously relevant) criticisms as the third criticism, which concerns implications for the Trolley Case. Kamm argues that on Quinn’s account of DDA, it follows that if a trolley were headed toward one person and I redirected it toward five people then this would be no worse than if I let a trolley hit five people instead of redirecting it to hit one person. This is because, on Quinn’s account, one is in control of the trolley in such cases, and so whatever choice one makes one is doing something, rather than allowing something to happen.

To be fair to Quinn, I think he could just respond to this criticism by saying that there may still be a morally significant difference here (such that switching from one to five is much worse than switching from five to one), but that it won’t be captured by DDA. Of course, Kamm might still be entitled to be surprised that this is not the kind of case where DDA is useful (I guess this all depends on what one takes the purpose of DDA to be).

A fourth criticism of Quinn’s account of DDA provided by Kamm is that it seems to follow from this account that when an agent omits to do something, the intention of the agent can determine whether or not his omission is impermissible, and it does this in a counterintuitive fashion. To see this we need consider three rescue cases (the first two of which are from Quinn). In Rescue Case I a car is being driven to hospital with five people in dire need of medical care in the back and the agent foresees that one other person will be run over by the car if he does not stop. Quinn thinks it would be impermissible to omit to brake in this case, because the agent intentionally controls the car. In Rescue Case II five people need to be looked after in the back of a car if they are to be prevented from dieing, but the agent foresees that by looking after the five people the car will continue on and hit one person. Here the agent does not intend for the car to continue on, and Quinn thinks this absence of an intention makes it permissible to not brake in this case. In Rescue Variant I, provided by Kamm herself, the agent intends that the car continue on while foreseeing that it will hit one person, and because of this intention the agent goes to the back of the car to help the five (the agent would not do this otherwise), thus losing the option of braking. Kamm contends, quite plausibly, that it is permissible for the agent to go back to help the five in Rescue Case Variant I (she also thinks Quinn would in fact agree about this), and then she points out that it seems to follow from this that the presence or absence of an intention for the car to go on can not legitimately play a role in explaining why it is impermissible to save the five in Rescue Case I, but permissible to save the five in Rescue Case II.

It seems that the third and fourth criticism should be considered together (although Kamm doesn’t make this explicit), and that when we do consider them together it seems natural to conclude that Kamm has shown that Quinn’s version of DDA is not going to be able to do the moral work he wanted it to do. However, an important question we can ask is: Given that Kamm has established that Quinn’s DDA can not be used to point to deep moral differences, why should we think that the distinction between doing and allowing is morally significant? Kamm has shown that Quinn’s DDA is not morally useful, and she has described it as “revisionary”, but I am not sure that it is revisionary (she does not seem to have an argument for this claim). Perhaps Quinn has drawn the distinction between doing and allowing just the right way (as a piece of conceptual analysis), but, surprisingly, it follows from his analysis that the distinction is not morally useful, contrary to his original aims.

(b) Quinn thinks DDA marks out a distinction that is morally significant, and that what makes the distinction morally significant is the fact that individuals have special authority to control the direction of their own lives and therefore have negative rights not to be interfered with. Kamm is right to point out here that it is just not clear how these special authority and negative rights claims are meant to do the grounding work that Quinn wants them to do, in relation to DDA.

Kamm herself thinks that whenever one intends some event to occur that interferes with another person one has an attitude that is such that one is failing to respect the other person’s special authority over their own life, but that it can’t be the case that one has an attitude inconsistent with respecting another person whenever one merely foresees that an event one intends to bring about will interfere with someone. I am not sure that this is so (i.e. that if one is inclined to think that a failure of respect always occurs in the first case, one shouldn’t also think it always occurs in the second case). Still, her main point here is that it is an intending of an event that one foresees will interfere with someone, rather than an intending to interfere with someone that plays the key role in Quinn’s account of DDA, and it is just not clear how the first of these is meant to always go along with disrespecting another person’s authority over their own life, or with interfering with another person’s negative rights.

One further point that Kamm makes in this section is worth mentioning. The connection between negative rights and having authority over oneself can not be quite as straightforward as Quinn thinks it is, because, for all he has said, we could have a positive duty to always bring about the greater general good and this duty could make us mere “moral cells” (Quinn’s term for describing what we are like if we lose individual control over our lives) even when our negative rights are not being interfered with. Kamm thinks that is surprising that Quinn never discusses the issue of a prerogative to not always sacrifice ourselves for the greater good.

Kamm begins this section with a discussion of Quinn’s account of the DDE (which she again regards as revisionary), and of his reasons for adopting his particular account. Quinn asks us to consider a case where we intend to chop up a healthy person for a medical experiment. Now it might be true to say that if (per impossible) all the bits of the body managed to recombine after the experiment so that the individual survived we would be happy with that result. This might be thought to show that we don’t really intend the death of the person in such a case, but if this is so the DDE will be essentially impotent, since every case we consider might be able to be correctly described as a case of foreseeing a bad event rather than a case of intending a bad event. To avoid this result, Quinn adds to the traditional analysis of DDE the condition that one not intend the involvement of a person in a way that will foreseeably cause that person harm. He attempt to justify the addition of this condition on the grounds that every person has a special authority over their own self such that they can not be used by others without permission.

Kamm says she finds Quinn’s DDE “sensible”, but she believes it is very different than the traditional version of DDE supported by Thomas Nagel and others, because according to Quinn’s DDE it is no worse to intend harm than to intend to do something very minor that interferes with a person when one foresees that it will lead to harm. Kamm notes that if one accepts Quinn’s DDE and still has the traditional Nagel intuition one will need to say more to explain why sometimes, but not always, an intended event and its foreseen harm are so close together that they can not be pulled apart for the purpose of making a sound judgement about whether one option is worse than another option.

Kamm is concerned that a certain kind of case will reveal that Quinn’s DDE is too much of a blunt instrument. She calls one such case Scarce Resources Variant I. In this case, a doctor has scarce resources and he knows that if he gives his resources to five patients (rather than one) they will be cured of one disease but soon develop another disease (and the one will die). The doctor foresees that when they suffer from this second disease the five will require organ transplants to survive, and that the organs of one person will be enough for the five to share and thus survive. It is pointless for the doctor to help the five unless he counts on the one dieing (because he needs the organs of the one to keep the five alive later). The doctor omits to help the one and he clearly involves the one in a plan that is very bad for him, contrary to the desires of the one.

Kamm thinks that it is permissible for the doctor to act in this case, and that this reveals that there is a difference between (a) acting or omitting in a way where one deliberately involves someone in order to further one’s purposes, and (b) acting or omitting only because someone’s involvement will help one further one’s purposes. This is an important distinction, assuming it can stand the test of scrutiny. I have my doubts. Consider the case just described. Assume that all the patients are all equally healthy, all of equal age, all equally deserving of help, and all suffering from the same illness. Now, it seems to me that in making a choice about which patient to save, the doctor is behaving in such a way that (a) is a correct description of his behaviour, and that this is true even if he uses a very fair method (i.e. throws a die) to decide which of the five patients will die. What about if Kamm were to simply stipulate that one of the patients is less healthy, or much older, or less equally deserving of help, than the other five patients? It still seems to me that the doctor is making a choice that deliberately involves the one person. After all, he could have chosen (unwisely) to let the least attractive patient (who is not the oldest patient, as it happens) die, or used any one of an indefinitely large number of other criteria to choose a patient to let die. Which criterion he uses is up to him. Of course, we could again change the case so that one patient will simply die before the doctor needs to make a choice, but such a case won’t help Kamm make her point. Perhaps I am overlooking something here?

I am more happy with some further comments Kamm provides concerning Quinn’s account of DDE (I am not discussing all of the comments she makes, but focusing on what I take to be the most important points). She says that Quinn diverges from traditional defenders of DDE in ceasing to view DDE as a principle that derives the impermissibility of an act from the intention of an agent. In the basic Scarce Resources Case (where all we stipulate is that there are only enough resources to save five, and that a doctor must choose between saving five and saving one), a single patient loses the right to be aided by a doctor if the doctor can aid five instead. It seems to follow from this, Kamm says, that even if a bad doctor decides to aid five only because she intends that one particular patient dies, the doctor’s act is still permissible. Kamm and Quinn are both happy with this (although Quinn did not emphasize that this was a way in which he diverged from traditional defenders of DDE, Kamm notes), and it is hard to disagree that this case reveals that the traditional use of DDE was too simplistic.

Once one has seen that what is really crucial in cases like Scarce Resources is the issue of determining when one person’s rights yield to other people’s rights, then there will be many cases when the intentions of an agent (such as a doctor) will be irrelevant to deciding whether or not an act is permissible, although they will not be irrelevant to determining whether or not an agent acts badly or for the wrong reasons. To bring this out, Kamm describes a case like Scarce Resources where the only reason that motivates a doctor to save five rather than one is his hope that he will win a prize that will follow from him being able to develop a fast cure for the five by researching matters involving the death of the one. Kamm thinks Quinn should agree that in this case the rights of the one still yield to the rights of the five.

I am a little unsure as to what Kamm expects the reader to conclude at the end of reading section II, but I think one is meant to conclude that Quinn’s account of DDE is an improvement on the traditional account, and that although it is not false (and this is one respect in which it is unlike the traditional DDE), it is far less useful in helping answer substantive moral questions than we might have initially thought it would be. By doing this, she is preparing us for the Doctrine of Triple Effect, which will be discussed in the next chapter.

Kamm now considers the relationship between DDA and DDE. The first observation she makes is that when we follow Quinn’s accounts of both of these, all cases where an agent intends the events that bring about the harmful involvement of other persons and where the DDE returns a negative verdict will also be cases where the DDA returns a negative verdict. This is because in all of the cases of intending an event an agent will either be doing something or doing nothing (to use my own way of putting the matter), and not merely foreseeing a consequence of acting in a certain way. On standard views DDA and DDE cut across each other (allowing for four distinct logical possibilities), and it is surprising that Quinn falsely claimed that this was true of his own accounts of DDA and DDE as well (Kamm provides a direct quotation where Quinn makes this claim).

In the remainder of the chapter, Kamm provides her own account of what unifies DDA and DDE. She claims this account is consistent with Quinn’s account. I will finish by briefly summarising the account, ignoring some interesting but also subtle details (some of these details are provided at least partly in order to explain why Quinn might have been right to think various things that he thought, despite not quite seeing why he was right).

The basic idea of Kamm’s “unified theory” is that both DDA and DDE have as their basic foundation a criterion of respect concerning the special authority that all agents have over their own selves. Kamm asks us to consider the following options (summarised in a rather confusing diagram):

(1) Harming without Intention
(2) Harming with Intention
(3) Harming with merely Foreseeing
(4) Not Aiding without Intention of Harm
(5) Not Aiding with Intention of Harm
(6) Not Aiding with merely Foreseeing Harm

In all cases of harming another person (1-3) there is said to be a prima facie ground of disrespect for the authority the victim has over him or herself, since the agent interferes with claims to non-interference provided by such authority.

In all cases where someone does not aid another person and also intends harm or involvement leading to harm (5) there is said to be a lack of respect for authority over the self because there is a preference present for a state of affairs that involves interference with, or harmful involvement of, someone without his or her consent.

In all cases of not aiding where there is no intended harm or involvement leading to harm (4,6), there is said to be no lack of respect for authority over the self.

In all cases where someone harms another and also intends harm or involvement leading to harm (2) the agent is said to be doubly guilty of disrespect for authority over the self [note: given the prima facie claim above, Kamm should probably have qualified the claim here as well]. This is said to explain why this form of harming is worse than harming without an intention to harm (1).

In addition to possibly addressing various concerns and questions raised earlier, all that there is now left for us to do is to ask ourselves to what extent these distinctions are useful, and to what extent the claims that accompany them are correct.


  1. 1. Posted by S. Matthew Liao | July 21, 2007 3:08 pm

  2. 2. Posted by Jeff Huggins | July 21, 2007 6:34 pm

    I found Kamm’s Chapter 3 very interesting, and the commentary very helpful.

    At the end of his helpful commentary, Daniel writes, ‘In addition to possibly addressing various concerns and questions raised earlier, all that there is now left for us to do is to ask ourselves to what extent these distinctions are useful, and to what extent the claims that accompany them are correct.’ In my comments, I’ll try to address the middle phrase in this statement, that is, to ask ‘to what extent these distinctions are useful.’

    Regarding a ‘special authority to control … [our] own lives’ as individuals, and a ‘right not to be interfered with’ by others: This ‘special authority’ and this ‘right’ do not really exist in an absolute sense, in most senses anyhow, as far as current science can tell and as far as we can discern from observation, of course. To the degree they exist as ‘rights’, they do so because we (humans) grant them to each other because of certain dynamics in our evolved ‘human nature’ (using the term as a shorthand) and because of the benefits of doing so. From a strictly scientific standpoint, it can be said that, once a human is born, the only ‘right’ she/he has is to try to survive, and the only obligation (or more accurately, ‘sure thing’) is to eventually die, at least in a biological sense.

    That said, because of the nature of humans and of human life, we humans do not want to be interfered with, usually. We want some sense of security and predictability and control. Because we have ‘mirror neurons’ and, more generally, human empathy (to differing degrees among individuals), we can understand that other humans also want a similar sense of security, predictability, and control. Thus, we grant the ‘right not to be interfered with’ to each other, intuitively, and sometimes also more formally (i.e., in ethical or legal codes). Of course, we are normally much more sensitive when our own ‘right not to be interfered with’ is violated than when we violate the same ‘right not to be interfered with’ of others. None of this is new, of course. But, it helps (I believe) to keep this in mind when considering the cases and questions posed as well as when considering the topics below.

    Regarding doing harm, allowing harm, and omitting to act: Doing harm requires action, and the reach and scope of human action are limited (or at least have been limited throughout most of history). We can only punch someone in the nose if we can reach him with our fist, and if we move our fist. Today, of course, with more powerful weapons, as well as with buttons and wires and wireless, we can initiate much greater harm at much greater distance. But, doing harm still requires action. On the other hand, allowing harm to occur without stopping it, i.e., omitting to act, is a much less limited and boundaried concept. In other words, harms are happening nearly everywhere, all the time. While the scope of personal action begins small (with the person) and extends outward from the person (but still within limits), the scope of all harms that might be addressed is as large as the universe, or at least as large as Earth. Individual humans, of course, cannot be everywhere at once. We cannot put an end to all harms, especially as individuals. We have limited resources, limited energies, and limited reach. Relatively speaking, we ‘omit’ many more actions than we commit. And, to a degree, our intuitive ‘right not to be interfered with’ includes an intuitive right ‘not to be interfered with’ by any absolute obligation to assist others, leaving that choice (to assist or not to assist) to the individual. That said, different cultures have different cultural expectations about when assistance is encouraged and when lack of assistance is blameworthy.

    All this is to say that there is a substantial difference between doing harm and allowing harm, I believe, but the nature and degree of that difference depend on particulars (especially of the situation). Furthermore, the differences stem from, and are grounded in, ‘human nature’ (shorthand term) and the requirements and conditions of human life. And, none of this is meant to excuse inaction or indifference in situations where action is clearly (intuitively and logically) called for. Most of us probably know what Dante said.

    A great (and short) book to read that sheds light on this subject is Robert Axelrod’s The Evolution Of Cooperation.

    Regarding intending and foreseeing: There is a difference, of course, between intending and foreseeing, but (it seems to me that) the difference is often less, and perhaps more subtle, than that between causing harm and allowing harm. Certainly, intending harm is not good, unless of course the harm being intended is justified for some compelling reason. But, when it comes to foreseeing (though not intending) harm, it seems to me that many more particulars of a situation weigh in on the question of whether a harm that is foreseen (or the action that would bring about the foreseen harm) is permissible. Perhaps that much is not new. That said, as before, the understanding that can guide us in determining what the differences are, and why they exist, and what their nature is, and the particulars that bring them about, is grounded in an understanding of ‘human nature’ (shorthand again) and in an understanding of the ‘effective’ function of morality itself.

    Regarding making a choice (or taking an action) that is permissible (in terms of the action itself and its direct consequences) but that is made/taken for wrong reasons: In this situation, the action (narrowly considered) and its direct consequences (narrowly considered) may be completely permissible, but that does not mean, in my view, that motivations don’t matter or that the motivations should not enter into the public equation regarding whether/how to respond to the agent and how to teach citizens (young and old) regarding the importance of motivations. One might ask, why? Is this view just pulled from thin air? No. The answer in part involves what has already been mentioned, i.e., ‘human nature’ and an understanding of the ‘effective’ function of morality itself. And it also involves the facts that humans are social beings, who learn from one another (and who often mimic one another), and that social-moral dilemmas, choices, and actions don’t usually occur in a vacuum, isolated from society and from the eyes or influence-able minds of others in the society.

  3. 3. Posted by Daniel Star | July 24, 2007 12:24 pm

    I should note that I made a small mistake in the paragraph that begins “Kamm thinks that it is permissible for the doctor to act in this case”. A sentence in this paragraph contains the phrase “he uses a very fair method (i.e. throws a die) to decide which of the five patients will die”. The number mentioned here should be six, rather than five, because the case Kamm describes is one where the doctor has one patient on one side and five on the other side (the total number of patients thus being six). I am not sure that the argument against Kamm that I provide in this paragraph works, but if it doesn’t, I’d like to know why. Does anyone have any comments on this argument?

  4. 4. Posted by S. Matthew Liao | July 24, 2007 2:31 pm

    I have some comments regarding Kamm’s treatment of Quinn’s account of DDA.

    1. p. 79: Kamm says

    [Quinn’s] revision of the DDA raises several problems. First, there is an oddity in focusing on omitting to act because one intends the movement of an object; this sort of intention implies that there is an important moral distinction between (i) a case in which someone refuses to remove water from a tub when sees aht this is the way to save a child from drowing, because he intends that the water remain at a certain high level, and (a) a case in which someone omits to aid the child whom he sees is about to be drowned if water rises because he intends the water level to move upward. Drawing a moral distinction here seems odd, because, in both cases, one intends some state of the world—whether it is a moving object or not, the action of an object of its in action—that causes a death.

    It seems that on Quinn’s notion of positive agency – as Daniel has very nicely described – both would qualify as “doings” since both are within one’s control.

    2. p. 79: Kamm says

    cases in which someone begins his life in a position which involves his imposing on someone (so that there is no movement toward imposition) should, I believe, be treated like cases involving someone’s doing harm (perhaps unknowingly), even though these cases involve no action. Quinn’s revision of the DDA, with the focus on the action of an agent or object, has no way of capturing this intuition.

    I’m not sure about this. How is this individual ‘doing’ harm? Perhaps this individual is an innocent threat. If so, to describe this individual as ‘doing’ something, when the individual has no control over the action seems odd.

    3. p. 79: Kamm argues that Quinn’s DDA gives the wrong solution to the Trolley Case. In particular it implies that if a trolley were headed towards one person and we redirected it toward five, we would be doing something no worse than if we let a trolley his five rather redirect is towards, because both are forms of positive agency.

    Reinforcing a point Daniel has made, in Quinn’s article on DDA when he discusses his solution to the Trolley Case (p. 305), Quinn does say that “And that is why he may pick the alternative that does less harm,” which suggests that he thinks that there is a distinction to be made here, but it is not made by DDA. For example, it may be motivated by the idea that that one should do the less harm, given that both are forms of positive agency.

    4. p, 80: Kamm says,

    consider a variant on Quinn’s Rescue cases. Suppose we go to the back of the car to provide the five with aid they need to survive. However, we do this only because we intend that we not be available to prevent the car continuing on the road because we intend that it continue. . . .

    Kamm argues that the fact that we intend the car to continue and hurt the one person on the road does not imply that it is impermissible for us to go and provide needed aid to the five.

    I think this move may be begging the question against DDE. A defender of DDE would deny that this is permissible, given that one is intending to harm the one. Notice that this case is just like the Munitions Grief Case.

    5. pp. 81-82: Kamm gives the impressions that Quinn does not defend his idea that “A person is constituted by his body and mind . . . . For that very reason, it is fitting that he have primary say over what may be done to them. . . because any arrangement that denied him that say would be a grave indignity.”

    I think Kamm is right that Quinn could say more about this idea. But it is important to note that Quinn does defend it by considering and rejecting alternative conceptions of persons where persons do not have such strong “predecedence of negative rights.” See for example his article on DDA (pp. 305-312).

  5. 5. Posted by S. Matthew Liao | July 24, 2007 3:30 pm

    Some comments on Kamm’s treatment of Quinn’s account of DDE.

    6. p. 84: Kamm says that in Scarce Resources Variant I, the doctor does not intend the single patient’s death, even though she only saves the five because that death will occur.

    Again, I think a defender of DDE would disagree. A defender of DDE would say that the doctor does intend the patient’s death, given that she knew that she could have saved the one patient but she deliberately allowed the patient to die so that she could then save the five. And, against Kamm, a defender of DDE would deny that it is permissible for the doctor to save the five on such a ground.

    Daniel: I’m not sure what you mean by “(a) is a correct description of his behaviour.” I think there is a dispute here about what the doctor’s behaviour is. According to Kamm, the doctor does not intend that the one patient die. According to a defender of DDE, the doctor does intend that the one patient die. Had the doctor tossed a six-sided die and decided to save the five, I think a defender of DDE may have less objections to such a procedure.

    7. p. 85: Kamm notes that Quinn says that intention does not matter in supererogatory cases. For example, suppose that a certain form of life-saving aid is supererogatory, but the only reason the agent refuses to give it is that he intends that someone die. According to Quinn, this alone does not make it impermissible for him not to aid. Kamm argues that this shows that Quinn should agree that “if a bad doctor decides to aid the five only because she intends the death of the one patient as an end, her not-aiding the one is still permissible”

    Against Kamm (and Daniel), I think this inference does not follow. Intention might not matter in supererogatory cases because agents in such cases have prerogatives not to act. But the case Kamm has described is different. Agents in such cases do not have prerogatives not to act. So a DDE defender could argue that in such cases intentions do matter.

  6. 6. Posted by Daniel Star | July 24, 2007 11:03 pm

    I think Matthew’s comments very nicely supplement my own summary, since most of them concern parts of the text I decided not to discuss in the summary. When I first read the chapter, I shared Matthew’s puzzlement with respect to the parts of the text discussed in his points 1 and 2, for instance, and so I decided that I’d only mention the relevant arguments in passing (given that I had plenty to write about already), and focus on what I took to be stronger arguments. That being said, I am all for completeness at the end of the day. I agree with most of what Matthew says above.

    Now I would like to comment on Matthew’s 6 and 7. With respect to point 6, I, like Matthew, think that the doctor does intend the patient’s death in Scarce Resources Variant I. The section of my text that Matthew quotes from is a section where I am just trying to suggest that it is in fact correct to describe the doctor’s behaviour in such cases as falling under Kamm’s (a) acting or omitting in a way where one deliberately involves someone in order to further one’s purposes, rather than only under Kamm’s (b) acting or omitting only because someone’s involvement will help one further one’s purposes. And I think this is true even if one roles a fair die to decide who will live and who will die. I agree with Matthew that “there is a dispute here about what the doctor’s behaviour is”, so I don’t think he really disagrees with me about this. Matthew also says “Had the doctor tossed a six-sided die and decided to save the five, I think a defender of DDE may have less objections to such a procedure.” Well, I don’t see how this makes a difference to the doctor’s intentions in way that is helpful to the defender of DDE (qua defender of DDE). Say the die reads two, and the doctor previously committed himself to patient two dieing in such a case, then I still think he forms an intention to let patient two die after rolling the die. Nothing is forcing him to stick to his original procedure, so he is still making a choice as to who will die. A fair and just choice, perhaps, but still a choice.

    With respect to point 7, I’m just not sure why my name is in parenthesis next to Kamm’s. Again, I skipped discussion of the argument concerning cases involving the supererogatory precisely because I suspected that the argument wasn’t very helpful.

  7. 7. Posted by S. Matthew Liao | July 25, 2007 12:24 am

    Daniel, thanks very much for the clarifications! That was very helpful.

    I took you to mean that the doctor who rolls the die would allow the die to decide whom to kill. So if the die didn’t fall on the number of the one patient, then the doctor wouldn’t kill the one patient. I took that to mean that although the doctor wants to kill the one patient, the doctor is willing to let some external, fair procedure determine his action. On that idea, suppose the die did fall on the number of the one patient and the doctor kills the one patient as a result, I allege that defenders of DDE may have somewhat of a lesser complaint.

    But perhaps you mean that the doctor will kill the one patient regardless of what happens with the die. If so, then I agree that rolling the die doesn’t make much difference.

    Although you don’t discuss the supererogation case explicitly, I put you in the parenthesis in 7 because you mentioned the inference Kamm drew from the case and you also seem to endorse it:

    It seems to follow from this, Kamm says, that even if a bad doctor decides to aid five only because she intends that one particular patient dies, the doctor’s act is still permissible. Kamm and Quinn are both happy with this (although Quinn did not emphasize that this was a way in which he diverged from traditional defenders of DDE, Kamm notes), and it is hard to disagree that this case reveals that the traditional use of DDE was too simplistic.

  8. 8. Posted by John Alexander | July 25, 2007 1:12 am

    This has little to do with the chapter, but…

    In the trolley problem it has been argued that it is permissible to throw the switch even though one person will be killed to save the lives of five people, but it is not permissible to throw another person in front of the trolley to stop it from killing five people. However, if a body will stop the trolley then we should not throw the switch and divert it onto another track killing one person. If a body will stop the trolley then if the trolley continues on its original course it will kill only the first of the five people and that body will stop the trolley before it kills the other four. There is no reason for the innocent bystander to get involved if we grant the bystander the same knowledge in both cases. There is no moral problem if we know that a body will stop the trolley.

    But I am missing something I am sure.

  9. 9. Posted by John Alexander | July 25, 2007 1:33 am

    I should have made it clear that there also would be reason for the bystander to throw the person in front of the trolley to stop it. It will stop when it hits the 1st worker.

  10. 10. Posted by Daniel Star | July 25, 2007 1:14 pm

    Matthew, I’m sorry I didn’t express myself more clearly in the first place. What I had in mind was a situation where there are six people, all of whom are equally unhealthy, and where there are only sufficient resources to save five people. I was imagining that the doctor might decide who to fail to save by throwing a die. He decides he will take a one on the die as indicating that he can leave the first person to die, a two as indicating that he can leave the second person to die, and so on. Or the doctor might have a policy in such situation to not save the person whose name is last in a purely alphabetical ranking of names. There are any number of other methods the doctor could use. Some of theses methods would strike us as fair, some as unfair, and we may be unsure what to think about the fairness or otherwise of some of them. My intuition is that, no matter which method is used, the doctor is still “acting or omitting in a way where one deliberately involves someone in order to further one’s purposes”. Even if the method chosen is a very fair method, there is no getting away from the fact that one is intentionally singling out one particular person for death, in order to save resources to save the five others, rather than “acting or omitting only because someone’s involvement will help one further one’s purposes” (where, on Kamm’s picture, this excludes the possibility of “acting or omitting in a way where one deliberately involves someone in order to further one’s purposes”). Do you have a different intuition than mine about the kind of case I had in mind, or were you just thinking of different kinds of cases than the one I had in mind?

    I think the other disagreement (the one concerning parentheses) follows from slightly different readings of the first paragraph at the top of page 85 of Intricate Ethics. I thought Kamm could make her point by just using the basic Scarce Resources Case, without depending on the variation in the case where the life-saving aid being considered is supererogatory. But re-reading the paragraph, I see why you thought I was agreeing with her, because it does seem like she is talking about the supererogatory version of the case throughout the paragraph (I previously thought she was only doing so early on in the paragraph).

  11. 11. Posted by S. Matthew Liao | July 25, 2007 3:42 pm

    Daniel, On the case you have mind, wouldn’t a defender of DDE say that when the doctor rolls the die, although he foresees that a particular individual will will die, he is not intending that the individual dies? So if ‘deliberately’ in (a) means that one intends, then a defender of DDE could say that the doctor is not involving the individual ‘deliberately.’

  12. 12. Posted by John Alexander | July 25, 2007 6:03 pm

    Is part of the problem the possibility, that many of us (may) have, that the distinction between ‘doing’ and ‘allowing’ may be unwarranted as far as determining the moral permissibility of outcomes are concerned? It seems to me that when we are faced with a choice between ‘doing’ A or ‘allowing’ B, the choice is really one that involves us intentionally doing one or the other so that the outcome is intended in its entirety including all foreseen harm. DDE seems to allow us to separate out one aspect of the outcome as somehow escaping our intended choices because it can be suggested that the harm is merely ‘foreseen’ and not required as part of the outcome.

    To give a simply example: suppose Tom is walking by the river and sees a baby drowning. He can either choose to save the baby by bending over and picking it up or he can choose to let the baby drown. If he bends over he is ‘doing’ an action, but if he lets the baby to drown he is ‘allowing’ something to happen that would happen if he did not intervene. Now, if he does not pick up the baby, does he not intend the baby to drown if he can foresee that it will drown if he does not act to pick it up? It seems to me that if Tom can foresee that the baby will drown if he does not intervene then he causes the baby to drown because he could have saved it had he chosen to do so. But, he choose not to and I think most of us would think he is causally responsible for the baby drowning because he could have affected which outcome would occur based on what he choose to do.

    The same analysis can be applied to the trolley case where the bystander (Mary) is at the switch. The bystander may not be causally responsible for the trolley running amok, but she is causally responsible for which outcome occurs because the outcome is a direct result of the choice she makes. Because the outcome must include all the elements that make it up, including the death of the one person on the siding (should she choose to throw the switch), then she intentionally kills that one person if she throws the switch. The death of the one person is an element of the outcome and is therefore intended.

    I suspect that the same analysis can be applied to the doctor’s case.

  13. 13. Posted by Helen Frowe | July 25, 2007 9:00 pm


    I’m not sure that we ought to accept either of the two claims that you make above. You claim that Mary is causally responsible for the outcome of the trolley problem irrespective of which outcome results. So, Mary causes the death of the one if she switches, and she causes the death of the five if she doesn’t. But it’s hard to see how a person can be causally responsible for an outcome that would have resulted from an identical sequence of events had she been absent. There’s something odd about labelling the killing of the five a result of a choice that Mary made, when the killing of the five would have resulted whether or not Mary ever existed (we can suppose). Choosing not to prevent X is not the same as causing X to happen. In contrast, the one would not have been killed in Mary’s absence – if she flicks the switch, Mary intervenes in the causal process in a way that causes the one’s death. (Kamm has more on this in C&A)

    This takes us to your second point – that since the death of the one results if Mary flicks the switch, the death of the one must be intended. It just doesn’t follow from the fact that something was an upshot of an action that it was intended. We can assume that Mary would have been only too happy if the one had miraculously leapt out of the trolley’s path. As Thomson says, if the one survives the trolley hit, Mary does not proceed to beat the one to death with an iron bar. The reason that she does not do this is that Mary neither intends nor requires the death of the one.

    Your comments about the drowning child case seem to conflate our desire to attribute moral blame to Tom (and, indeed moral responsibility, since one can be responsible for failing to save) with Tom’s being causally responsible for the baby’s death. We ought, I think, to resist this temptation.

  14. 14. Posted by Daniel Star | July 25, 2007 9:59 pm

    Matthew, I agree that the defender of DDE could say that at the time the Doctor rolls the die he is is not intending the death of anyone in particular, merely foreseeing the death of someone. But what about after the die is rolled and he then settles on a particular individual? I don’t think the defender of DDE can say, well, it was the die that chose to focus on this individual, so I am not intending that anyone in particular die. That would be to act in bad faith (i.e. to pretend one is not responsible for something one is responsible for). I should also say that the quotations I used in setting up the distinction I was focusing on at this point in the summary are from page 84 of Kamm’s book, and that she is commenting on Quinn’s version of DDE at this point, not the traditional DDE.

  15. 15. Posted by Jeff Huggins | July 26, 2007 12:12 am

    I have a question for someone/anyone. The discussion in this particular chapter of Kamm’s, and in this thread, is fascinating and helpful in ways, but, in other ways, it leaves me with the question of what is being sought, in philosophical terms, when we consider the more complex, rare, and contrived cases (as opposed to the more common and readily imaginable cases)? Put another way, when one considers a case that is fairly (or highly) complex; that represents a situation that is very rare; that might rely on specifics that are not stated in the case itself but that a respondent must imagine or assume in order to come to (or ‘land on’) one conclusion or another; where respondents’ intuitions lead to results of 53% ‘permissible’ and 47% ‘not permissible’; and where the discernment of intricate ethical principles is tied very closely to semantics (and perhaps also to views that are culturally dependent); in those types of situations, when one eventually arrives at an answer (e.g., an intricate principle), what is that answer claimed to represent? Are we here searching for human ‘universals’, or for ethical universals that are claimed to be independent of human body-brain-minds, or for principles that we think should be voluntarily agreed to in order to improve the world because we believe that they would be helpful in doing so?

    To be clear, I think it is very helpful, of course, to understand and explore intuitive human responses to the more straightforward cases, as Kamm does, as other philosophers do, and as scientists also do. But, when it comes to the more rare and ‘intricate’ cases, and to the intricate principles that might be derived from those cases, I’m not so sure what the goal is (of those mentioned above, or perhaps another not mentioned)? I can also see, easily, some important tangential benefits of considering the highly intricate cases: For example, the intricate discussions can sometimes help shed light on our understanding of the more common and simple cases and the more common approximate principles. But, when it comes to deriving highly intricate principles from the mixed results associated with highly intricate cases, what are those intricate principles claimed to be? What do they represent? Great Thanks (in advance) to anyone who can help me out here.

  16. 16. Posted by John Alexander | July 26, 2007 2:06 pm

    Thank you very much for your comments. A few replies:
    1) You wrote: “But it’s hard to see how a person can be causally responsible for an outcome that would have resulted from an identical sequence of events had she been absent.” I do not believe that they are identical for the simple fact that she is present. It is the fact that she is present and can affect which outcome will occur that makes her part of the causal sequencing of events. Her presence at the switch makes her part of the event. Someone will die as a result of her being at the switch.
    2) Further on you write, “It just doesn’t follow from the fact that something was an upshot of an action that it was intended.” I suppose this depends on how narrowly we want to define the range of what is intended. I seems to me that if I know that an action that I will perform will result in a,b,c, and d, then I intend a,b,c, and d to occur because these are the necessary (not in any modal sense here) components that make up the event that I will bring about by my action. I may not want d to occur; d may not be required to occur in order for me to get the results I want from the action I perform. But, and this is where we disagree, that fact that it is not desired nor required does not mean that it is not intended if I know, or have a reasonable belief, that d will occur as a result of my action. Another way to look at this is that Mary knows that eventually the trolley will stop. In order for it to stop it has to travel a certain distance with resistance sufficient to bring it to a stop. We can assume, due to the way the problem is set up, that the trolley is going to stop at some point past where the one person is standing. When (and if) she throws the switch she has to intend that the trolley will travel this distance necessary for it to stop. Therefore, she has to intend the entire sequence of events that make up the larger event relative to the trolley stopping.
    3) I still think that Tom is causally responsible for which outcome regarding the baby drowning occurs as a result of his decision for the same reason that Mary is responsible for whichever outcome she chooses. It seems to me that we do attach moral responsibility to those actions that we cause. I think that deontologists would have to accept this. Is not one of their main points in the theory of punishment that we can only justly punish someone IFF that person did what he is accused of doing?

    Jeff: I appreciate your concern. I might even state it a bit stronger (if I understand your concern correctly). Is it the case that some of the examples being discussed are so contrived that they beg the question in favor of the outcome desired, or are they so unrealistic as to have no practical import. I worry that given the complexity and unusualness of some of the issues being raised that it is not going to be possible to make a decision regarding how one ought to act in concrete situations if one has to think of all possible ramifications and possible implications of what would occur if we accepted a particular moral principle such as DDE, DDA, or DTE. But wait a minute, why do we need moral principles? Why not simply rely on the intuitions that we have regarding each case? If intuitions have epistemic relevance in determining how we should act, then having an intuition on how we should act should be sufficient to justify how we act. But this requires us to explain the divergence of intuitions regarding particular cases.

    One of the things that I like about the original trolley problem is that is reflects the general structure of many of the moral issues (dilemmas) that we face in real life. As a manager I had to make decisions that affected people and organizations. For example, do I layoff 20% of my workforce or allow the business to ‘go under?” I can understand this type of problem because it is one I faced. I do not understand the relevance of some of the other examples of the trolley problem, i.e., throwing a person in front of the trolley to stop it. Outside of the fact that it is unrealistic, I do not see how it arises (for reason mentioned in an earlier comment). I realize that this may not be a concern for some philosophers, but how is this type of ‘rarefied’ analysis going to be instructive to the general population concerning how we should act?

  17. 17. Posted by Jeff Huggins | July 26, 2007 4:25 pm

    John: Thanks for your response. I agree with much of it, and it seems that you and I share at least some of the same concerns.

    That said, my view is that there is a very informed and grounded ‘middle ground’ between A) highly intricate principles based (sometimes) on highly intricate cases, and B) the idea that each individual’s own intuitions are sufficient to determine the ‘moral’ answer to all moral dilemmas, and that consistency, principles, and respect for others can be discarded. When I say ‘informed and grounded’, I mean informed by, and grounded in, a (still-in-progress) scientific understanding of human behavior, human social-moral dynamics, and related matters, including why they exist, what they are ‘effectively’ for, how they work, and so forth. When I say ‘middle ground’, I don’t mean that we should choose or respect this ground because it is ‘middle’ between highly intricate principles and a notion that ‘all intuitions are right.’ Instead, we should understand it, I believe, because it is informed and grounded.

    An understanding of human morality, in my view, does not (usually) lead to highly intricate, precise, specific ‘principles’ in the way people often use that term. Nor does it lead, in most cases, to specific, precise written rules that are seen as defining behaviors that are always right or always wrong, no matter what the context, even when they (the rules) conflict with each other in some instances. Instead, a helpful understanding of human morality (including deriving central ‘oughts’) is, I believe, grounded in the following elements, among others:

    • An understanding of foundational human social-moral dynamics, tendencies, and patterns (notice I did not use the word ‘principles’ at this point);

    • An understanding of the ‘effective’ function of human morality itself;

    • An understanding of the human social-moral dynamics, tendencies, and patterns that are best suited to the accomplishment of the ‘effective’ function of morality. What I mean by this point is, for example, among other things, that human ‘reciprocity’ is a rather basic human social-moral dynamic and that much is known about the specific subsets of reciprocal dynamics (i.e., behaviors) that lead to positive outcomes, i.e., to outcomes that are aligned with the ‘effective’ function of morality itself and that contribute to it. (For example, see Axelrod’s The Evolution Of Cooperation.) And, …

    • An understanding of all these things in the context of human abilities, limits, and imperfections.

    As mentioned in the earlier post, I am not quite sure what is being sought, philosophically speaking, in attempting to define highly intricate principles based on intricate thinking applied to (often) mixed intuitive responses to cases involving complex and un-ordinary situations. I’m hoping that people will ‘help me out’ with their views on that question as posed in my earlier post.

  18. 18. Posted by Jussi Suikkanen | July 26, 2007 4:33 pm


    what’s your view on causation then? I mean if I remember D. Lewis right he held something like that event A is a cause of event B iff in the closest possible world in which A does not take place B does not take place either. Something like that seems like an appealing view about causation. In the trolley case you are talking about, this condition is not satisfied. In the closest possible world you are not present at the switch the same things happen – so you cannot be the cause or part of the causal chain that is responsible causally for the hit. If you reject this view of causation, you need to give an alternative picture with which we should think about causation.

  19. 19. Posted by John Alexander | July 26, 2007 6:17 pm

    I agree that if we accept Lewis then my ascribing a causal relationship to Mary and the death of five people is wrong. But why should I accept Lewis (or possible world semantics for that matter)? In my view, A is the cause of B if A’s actions can affect whether B occurs or not. Certainly in the trolley case, Mary’s actions can affect (determine in this world) whether the five people are killed or not. She can cause the trolley to continue on its original course by not throwing the switch or save the five people by throwing the switch. She determines which is going to happen. I would maintain that her decisions are actions; mental actions. If we can call the omission of doing something an act (act of omission) as well as doing something an act (act of commission), I see no reason not to call her relationship affecting which outcome will occur as a result of her physical actions a causal one in so far as she determines thru her mental actions which physical action will occur.

  20. 20. Posted by Jussi Suikkanen | July 26, 2007 7:33 pm

    Because Lewis was right about pretty much everything :-). And you get the same result with pretty much any plausible theory of causation (try Mackie’s INUS condition for instance). I worry that this:
    “In my view, A is the cause of B if A’s actions can affect whether B occurs or not.”
    profilitates causes way too much. I mean the actions of *so* many things can affect whether something occurs (depending of course your reading of ‘can’ here). For one this cannot distinguish between causal enablers and causes. The actions of wind can affect whether the match I strike catches fire but that doesn’t make the wind cause of the fire.

  21. 21. Posted by Fiona Woollard | July 27, 2007 8:26 am

    Dear Jussi,

    I had always taken in both Lewis’ counterfactual theory and Mackie’s INUS theory do leave an agent’s behaviour as the cause of those things he allows. For Lewis, if Tom had not behaved as he did (i.e. refrained from picking up the baby) then the baby would not have died. So Tom’s failure to pick up the baby is a cause of the baby’s death. Similarly, Tom’s failure to pick up the baby is an INUS condition of its death. Mackie explicitly notes that absences can be causes – he picks out the failure of water sprinklers to go off as the cause of a warehouse burning down (or something like that).

    However, I agree that on these accounts Tom’s presence is not a cause of the baby’s death. Thus we might want to say that Tom’s behaviour – his failure to pick up the baby – is a cause of the death, but that Tom’s presence is not a cause of the death and thus Tom does not count as a cause of the death. The theories you mention are both theories of event causation. There are questions about how we move from an account of event causation to one of agent causation.

    I think your point about causal enablers is very important. My personal response is to distinguish a narrow notion of cause, which distinguishes real causes from mere conditions (enablers, absence of disablers etc.), and a wider notion that does not make this distinction, but basically embraces anything that is causally relevant. However, unlike you, I see both Mackie and Lewis as giving accounts of the wider notion.

  22. 22. Posted by Jussi Suikkanen | July 27, 2007 9:38 am


    that’s well and good. Maybe, I’m not sure about this:

    “if Tom had not behaved as he did (i.e. refrained from picking up the baby) then the baby would not have died.”

    At least it’s not something I’ve thought would hold. The assessment of such counterfactuals is very tricky. It’s not clear that the claims are even truth-apt. In any case, remember that we have to look at the closest worlds. So, now take the actual behaviour of Tom and call it b1. Call the action of saving the baby b2. Now, the question is, are the possible worlds in which Tom’s behaviour is closer to b1 than b2 and where the baby drowns? It does seem likely that there well could be.

    Refraining to pick up the child I take it means intending to let the baby drown. It does seem possibly that there are close worlds where he lacks this intention and where he does not save the baby and where she dies. So, I think there is a case for claiming that the refraining to save the baby does not cause the baby to die on the counterfactual account. This is even clearer if we described Tom’s actions in merely behavioural terms which did not include intentional terms like refraining. For any actual behaviour Tom takes, there are close worlds in which he behaves in others ways and in which the baby does not get saved.

    Similarly, it’s not clear whether his intentended refrainment is a non-redundant part of a sufficient condition. You are right that he allows absenses to be causes.

    I agree about the last point to a degree. This distinction seems to work against both Mackie and Lewis if are talking about the real causes. And, there are so many other views about the narrow sense of causation that all have problems.

    I do think though that even under the wider notion of causation we have to draw the lines somewhere. So, you might think with Lewis that every event has a causal history that includes a vast network of events that are causally dependent on one another. I agree that many enablers and absences of disablers go into this. But, there comes a point where we begin to stretch even the wider notion. How many things can we imagine that disable the striking of a match ability to set the match on fire? Infinitely many I would say. Is the absence of each one of them really causally relevant in the wider sense of cause? I’m not sure.

  23. 23. Posted by John Alexander | July 27, 2007 1:20 pm

    Good morning Fiona and Jussi (and anyone else who is reading this)
    What an interesting discussion. Obviously I am leaning (falling) in the direction of what Fiona said and I thank her for presenting an analysis of causation so much more clearly then I was able to.

    Let us consider the idea that A causes B and that if A does not occur B will not occur. Late last night as I was thinking about what Jussi had said (and had not yet read Fiona’s comments) the following thoughts came to mind:
    1) Applied to the trolley case (we can make the appropriate substitutions in the drowning baby case) A equals the trolley running amok along a certain line and B equals what will happen to five people that are on the line the trolley will travel if it is not diverted, or stopped. In this scenario B, the death of five people, will not occur unless A occurs. But this scenario describes a world where Mary is not at the switch deciding what to do so applying it to the world where Mary is at the switch deciding what to do as explaining what happens to B is not an accurate description of the causal series in the world where Mary is at the switch deciding what to do. In the world where Mary is at the switch deciding what to do we have A causes B which causes C. (It is important that we include in the causal sequencing of events that Mary is deciding what to do. It might be the case that Mary is at the switch but is unaware of A, and will remain unaware of A until it is too late. In that case A would cause C.) In the world where Mary is at the switch and deciding what to do regarding A, A is the trolley running amok, B is Mary at the switch and deciding what to do regarding A, and C is the outcome of what she decides to do. In this world, because Mary must make a decision that affects the outcome of C, B is the cause of C although the trolley will be also be involved in C.
    2) Imagine that John is having an argument with his wife and during the argument he picks up a gun and shots and kills her. Now, on Jussi’s account, as I understand it (A causes B) Mary did not kill John, it was the bullet from the gun that killed John. Now, of course, in an important sense this is true, but most of us would claim that it was John that killed his wife because he is the causal factor in the bullet being sent on its way towards her. How else could we hold John accountable for his action?

    I think what is happening (and I am not an expert in the theory of causation. Really?!!!!) is that we are treating causation as a linear series of events; A causes B which causes C which causes D, etc. I think that we should look at the series in the trolley case and the drowning baby case as examples of non-linear series where some of the causal events are above the line (I do not know how to represent this).

    My account of what causes the death of someone in these cases (including the drowning baby case) seems to apply to many of the examples that Kamm has so far utilized to analyze and explain DDA, DDE, and DTE. If this is so, and we can attach intentionality (deciding what to do) as a causal factor in which outcome occurs then it seems to me that DDA, DDE, and DTE are not required in order to explain the moral implications of what occurs. But, there is alot of book left to read so I will withhold final judgment.

  24. 24. Posted by John Alexander | July 27, 2007 3:57 pm

    Sometimes I am dumber then a bag of dirt. Obviously I changed my example of John killing his wife from Mary killing John, but I did not correct it throughout the example. Maybe, someday, there will be a ‘stupid’ check on computers, but until then I apologize for being sloppy.

Post a comment

Name: (required)

Email Address: (required) (will not be published)



(Spamcheck Enabled)

This work is licensed under a Attribution-NonCommercial-NoDerivs 3.0.