Print This Post Print This Post

Chapter Six presents Kamm’s response to Peter Unger’s Living High and Letting Die. According to Kamm, “Unger has tried to show that relying on intuitive judgments in cases is a worthless methodology for finding principles.” (p. 190). Therefore, for him, “when there is a conflict between the theses supported by general reflection (e.g., reduce suffering) and judgments about particular cases, we should stick with the results of general reflection, for our intuitions about cases are unreliable and manipulable by morally irrelevant factors.” (p.192).

In particular, Unger emphasizes two morally irrelevant factors that, he argues, motivate our intuitions. One is separation between persons, which Kamm calls “Unger-separation” and that she says must designate the relation between persons who “do not share the same or similar properties” (p. 200). Unger believes that we tend to view it as impermissible to sacrifice someone for the sake of other people from whom she is Unger-separate (e.g. they are tied to tracks but he isn’t: he stands on a footbridge). A second factor is “protophysics,” which means that the agent starts some volition. Unger believes that we tend to view it as impermissible to sacrifice someone for the sake of other people when the only way to do so involves protophysics (e.g. we must push a person standing on a footbridge, thus starting a volition). Unger adds, and Kamm agrees, that neither Unger-separation nor protophysics have real moral significance. Unger’s conclusion is that our intuitions, which, he believes, are guided by these two factors, do not really track what we should do.

Kamm’s Unger also has a view on substance: morally we must, quite generally, impose property-loss or death on ourselves or on others in order to prevent even greater property-loss or death. There is no place for partiality. Nor is there place for the distinctions between “harming someone so that he suffers mortally and not-aiding him when he will suffer mortally,” and between “harming someone by re-directing a threat towards him and harming him by using him to stop a threat.” (p. 191). Unger substantiates both the methodological conclusion and the substantive one by taking up the thought experiments that have been thought to substantiate these distinctions and constructing very similar thought experiments that tend to elicit different judgments. The differences between these very similar thought experiments stem, he argues, from differences in terms of Unger-separation and/or protophysics, which lack moral significance.

[Nir’s question 1: Is Kamm’s reconstruction of Unger (here and elsewhere) accurate? In particular, doesn’t Unger also point to the unimportance of further influences on our intuition, such as brute geographical proximity to the victim (in aid cases), and the number, the order, and the frame of the options that cases include?]

Now let us turn to Kamm’s many arguments against Unger, in respective chapter sections.

I. Introduction + II. Unger’s Ethical Method:

Kamm here recounts Unger’s own views, but in passing she poses two challenges:

1. Occasionally Unger himself slips and takes our intuitive judgments at face value, using strong words against people who would act in line with his own general moral views (p. 192).

2. “his method for finding the correct values is that of reflecting directly on general claims, such as the claim that mortal loss should be minimized. This too seems like an intuitive judgment, even if not a case-based one.” (p. 193).

III. Property Losses and the Duty to Aid:

Kamm here summarizes the claims she made in a longer piece on Unger and aid to distant others. Following Singer, Unger had mocked the idea that brute geographical proximity between the person in need of help and the agent can make a real moral difference. Kamm responds (on p. 194):

1. What really makes a difference, intuitively, is the geographical proximity, not between the person in need of help and the agent, but between that person and the tool that the agent could use to help that person, e.g., a boat that a distant owner could direct via remote control toward a person drowning off of the boat. Kamm says that, intuitively, the agent must offer help–even if this will destroy her property irreparably.

2. Even if proximity does not always make intuitive difference, for a factor to be morally relevant, it is enough if sometimes it makes a real difference.

3. Unger criticizes nonconsequentialists for allegedly telling us to prioritize the geographically near. Kamm responds that, for some nonconsequentialists, herself included, “supererogatory acts [such as helping the distant, which] we are not duty-bound to give–may be done instead of what is strictly our duty.”

[Nir’s Question 2: This section seems to answer Unger only superficially. Unger says that our intuitions vary in accordance with a morally irrelevant factor: geographical proximity. Instead of showing that they don’t, or that this factor is morally relevant, Kamm only points out that they vary in accordance with geographical proximity of a different type than the one that Unger cites. Surely brute geographical proximity to a tool is no less irrelevant than brute geographical proximity to the agent. If what determined nonconsequentialist intuition were not proximity to the agent (or Unger-separation, or protophysics) but instead the density of the victim’s pigments, or what the people whose intuitions we survey had for breakfast, then Unger’s main conclusion would stand. It would clearly remain the case that intuition, and any theories that it alone “substantiates,” lack authority. Now Kamm seems to assume that proximity between a person in need of help and the tool that she needs does have real moral significance. That geographical proximity is unlike pigments. But she here offers no justification to support that assumption (except her related intuitions, which obviously cannot justify their own justification). Kamm may believe that such a justification could in principle be given. But that seems overly optimistic. Brute geographical proximity is something that usually we feel lacks moral relevance; it links closely with no central moral concept; most nonconsequentialists accept its irrelevance too.]

[Nir’s Question 3: Kamm’s views on permissible harm to property, which surface here, contrast with her views on permissible harm to the body, which chapter 5 and section IV of the present chapter expound. The permissibility of harm to property depends on geographical proximity of sorts, whereas that of harm to the body seems not to (see e.g. the Yard Case, on p. 202). However, damage to property can deprive the person of livelihood and cause her similar harm as damage to her body does. Wherefore this contrast? Elsewhere Kamm points out that “someone may be inviolable with respect to one thing… but not another thing…” (p. 255). But that does not clarify why inviolability should discriminate between, specifically, harmful damage to someone’s property and harmful damage to her body. Does Kamm elsewhere justify her different treatment of property and the body?]

IV. The duty to harm oneself and others:

Kamm summarizes Unger’s views on harming one’s own, and other people’s, bodies while preventing greater evils. She focuses on a particular case from Unger (“Switches and Skates,” quoted on pp. 196-7). In that case, if we do nothing to stop a hurtling trolley (a), then it will kill 6 innocents trapped on the tracks (we kill 6). If we operate switch (b), the trolley shifts to another line on which only 3 are trapped and it crashes only them (we kill 3 and save 6). If we operate switch (c), another trolley is released, and it rolls down, and, because 2 very fat people are trapped on board of that other trolley, when it collides with the first trolley both trolleys foreseeably go off track into an empty area, and only the 2 fat passengers die (we kill 2 and save 6). Finally, if we operate switch (d), we activate the remote controlled skates that a nearby fat man wears, and he rolls down, crashes onto the hurtling trolley and dies, but he also stops it (we kill 1 and save 6). Unger says that in this case the intuition is that it’s permissible to choose (d). However, he notes, the fat man on the skates is only superficially different from the fat man in the classical Footbridge case–where most people say that it is impermissible to push one fat man to be crushed by the trolley as a means to saving the six. Unger accepts that the cases somewhat differ, for example, here there are many options, whereas in the classical fat man on a Footbridge case there are just two options. Furthermore, these differences causally explain why we have such different judgments on the two cases. But surely factors like the number of alternatives are not truly morally relevant to what we should do in a case. [Compare: Sidney Morgenbesser walks into a restaurant, has dinner, and then asks the waitress what they have for dessert. She says apple pie and blueberry pie. Sidney Morgenbesser says he’ll have the apple pie. She comes back in a moment and says that they also have cherry pie. So Sidney Morgenbesser says “In that case, I’ll have the blueberry pie.”]

Kamm’s main responses:

1. She denies that intuitively it’s permissible to choose (d). In fact, the buck stops earlier: her intuition is that even (c) is impermissible (pp. 198f.)

[Nir’s Question 4: My own intuitions side with Unger’s. What to do in the event of disagreement between different people’s intuitions? Are we at an impasse? Kamm writes: “I have made reference to “my intuitions” rather than made predictions about what others’ intuitions will be. This is because I believe that it is primarily through an individual generating her own intuition judgments… that we can make progress by a method employing intuitive judgments.” Are other people’s intuitions simply irrelevant, according to Kamm, or just weighted less than one’s own intuitions? If one’s intuitions are known to be biased in the context, do others’ intuitions then get equal weight? Besides, is any of this convincing? What, if anything, gives each of us the “right” to privilege her own judgments?]

2. Insofar as the case succeeds in tempting us to say that (d) is permissible, that’s possibly because it has a “sorites structure: Creating a series with (supposedly) slight differences between each member of the series gets us to equate the two extremes of the series, which we know to be distinguishable.” (p. 200).

[Nir’s Question 5: I wish that Kamm developed this important point further. Some of Unger’s cases elicit new intuition only by tempting us to make clearly invalid, e.g. sorites, moves. Such cases do not show that all intuitions are worthless, just that some, induced by Unger in these contrived cases, are worthless. Compare: the fact that when you are sober you see one finger and when you are drunk you see two fingers doesn’t show that your vision is worthless; the most it can show is that it is worthless when you are drunk.]

3. Unger’s determinate error theory, according to which our intuitions respond to Unger-separation and protophysics alone, such that in the absence of these factors we tend to think that people are “fair game,” is false. Unger-separation, protophysics and their combination are neither necessary nor sufficient for inducing the intuition that some concrete way of harming someone to save unrelated others from a greater evil is impermissible (therefore, that intuition must have some other source–potentially, our covert endorsement of Kamm’s ethical principle). Specifically,

a. As already mentioned, for Kamm, in Switches and Skates, (c) is impermissible–although it involves no Unger-separation: the 2 on the other trolley are similar to the 6 to be saved [presumably in that the all are trapped along the same track system] (p. 202).

b. In the absence of Unger-separation and protophysics, we still do not feel that people are simply “fair game.” As several hypothetical cases demonstrate, there remain some things that it’s intuitively impermissible to do to such people (p. 201). For example, consider the Push-Three 2 Case: the only way to stop a trolley that’s heading toward 6 and that we could have diverted onto a track with 3 is to throw these 3 people in its path, before it reaches the 6: a technical problem disallows us simply to divert the trolley so that its running over the 3 would save the 6. Although the 3 are not Unger-separate from the 6 (since all are trapped on the same track system), intuitively, throwing the 3 in the trolley’s path remains impermissible. The absence of Unger-separation does not make them fair game.

[Nir’s Question 6: Regarding this case I have very weak and conflicted intuitions. Part of me wants to say that it’s OK to throw the 3 in the trolley’s path, and part of me wants to say that that’s not OK. Where does that leave me? Some anti-consequentialists have argued that even weak and conflicted intuitions are sufficient to defeat consequentialism. For example, Bernard Williams said that the mere fact that it’s unclear to us that Jim must kill the Indian undermines utilitarianism, because it indicates the presence of agent-relative reasons for action–which utilitarianism denies. Is that also Kamm’s stance, or must she rely only on cases on which intuitions are clear?]

c. The Second Trolley Redirection Case (p. 202) shows that the absence of both Unger-separation and protophysics does not always prompt intuitive permissibility. The Yard Case and the Lazy Susan V Case (pp. 202-3) show that Unger-separation, protophysics, and their combination do not always prompt intuitive impermissibility.

[Nir’s Question 7: Must Unger claim that Unger-separation and protophysics are sufficient/necessary individually/together for permissible harm? Some biases cloud our thinking in weaker fashions, e.g. by somewhat increasing the probability that we will give certain answers. For example, people tend to judge their own nation’s wars as just more than they do other nations’ wars. But it’s NOT the case that a war’s being one’s own nation’s is psychologically necessary or sufficient for seeing it as just.]

d. In Unger’s use, even family resemblance between members of a set would rule out their Unger-separation. They could count as similar and not Unger-separate. But how can family resemblance be thought to be of great moral significance? (my interpretation of Kamm’s p. 199).

[Nir’s Question 8: But Unger is not saying that Unger-separation has real moral significance. He denies that. He says only that commonplace intuitions are influenced by it. So perhaps I misunderstand Kamm here. Perhaps she really argues that Unger-separation is too OBVIOUSLY lacking in moral significance (because even family resemblance rules it out); so obvious is its insignificance that it is unreasonable or uncharitable of him to assume, as he does, that intelligent ordinary folks are influenced by Unger-separation and in that sense ascribe it real significance.]

4. Kamm suggests that the principle of permissible harm that she had developed in chapter 5 is what generates our intuitions on all of these cases (pp. 203ff.).

a. On Kamm’s principle of permissible harm, introducing an evil, e.g. causing someone to suffer a harm that she would not incur otherwise, is impermissible even if introducing it would be an efficient means to preventing a greater evil, e.g. if it would save another from a greater harm. However, if introducing a harm is only the effect, the non-causal flip-side, the condition of action, and so forth of preventing a greater evil (not strictly speaking a means to preventing it), then its introduction becomes permissible. That principle explains, for example, why in Switches and Skates, alternatives (c) and (d) are, on Kamm’s intuitions [but not on Unger’s and my own!], impermissible. Both options involve introducing harm to someone who would not otherwise be harmed–the 2 on the other trolley or the 1 on the skates; and they introduce that harm as a means necessary to saving the 6. By contrast, in the classical Trolley case, redirecting the trolley onto another track where 3 who are trapped on the tracks will be run over has killing the 3 only as its effect or non-causal flipside. That’s why redirecting it is permissible.

[Nir’s Question 9: As it is phrased in chapter 6, Kamm’s principle of permissible harm forbids even slapping one person as a sure means to saving another from death; and even mortally stabbing one person as a sure means to saving a billion others from death. Does she mean that?]

[Nir’s Question 10: If the distinction between causal means on the one hand and non-causal flip-side on the other has little moral significance, then our intuitions turn out only to be worthless in a different way than Unger took them to be. They do not turn out to be authoritative and to underwrite nonconsequentialism. So a lot hangs on the question whether this distinction has real moral significance, in the sense of justifying our intuitions, and not only in the sense of motivating them. By analogy, had Kamm shown only that what drives our intuitions is what we had for breakfast, she would refute Unger’s diagnosis of what drives them; but Unger would be right on the important question: on whether our intuitions are worthless or they should give pause to consequentialists. Kamm does suggest the following justification of these intuitions–both here (p. 205) and elsewhere (pp. 28-9, 253-6, 416-7 and earlier publications): our right not to be violated, not even as a means, gives us a high status, higher than the one that a right to be saved would give us. However, this “justification” doesn’t show that the right not to be violated as a means imparts a higher status on us than would a right (which for Kamm we lack) not to be sacrificed as an effect, as a condition of action, as a non-causal flipside and so forth. As far as I can see, Kamm nowhere shows that. So her status-based “justification” seems inadequate. It does not even begin to justify a mainstay of her recent theory: the alleged moral significance of the distinction between means on the one hand and effects, conditions of action, non-causal flipsides and so forth on the other.]

5. “Sometimes, contexts have intrinsic moral significance, and so we cannot draw implications from them about what to do in other contexts.” (p. 198). Kamm here probably suggests that the existence of many alternatives in Switches and Skates may affect what we should do in that context. In other words, Kamm disputes the point of Sidney Morgenbesser’s joke: the independence of irrelevant alternatives.

6. Kamm addresses Unger’s views on harms directly done by an intervening agent (pp. 207-9). Unger argues that the intuitive distinction between such harms and harms that we do directly is unjustified. He also argues that our intuitions about this respond to morally insignificant Unger-separation and protophysics. But Kamm offers a battery of cases to counteract his position. These cases also brilliantly elicit judgments of permissibility in the presence of Unger-separation and protophysics (which Unger claims tend to generate judgments of impermissibility), and ones of impermissibility in their absence. For example, in Kamm’s Lesser-Loss-Card Case 2 (p. 208), if I pick card (a), an evil person will send fewer henchmen to kill Africans than he otherwise would, killing 20 instead of 50, but he will also send the rest of his henchmen to chop off the foot of an Asian. So the Asian is similarly threatened by this evil person and therefore [I assume] is not Unger-separate from the Africans. If instead I pick card (b), the evil person will first send a henchman to kill (not just to maim) the Asian, and this will definitely calm the evil person, so he will not send his henchmen to Africa. Kamm says that intuitively it is impermissible to pick card (b), although the Asian is not Unger-separate from the Africans, and, we may add, although that involves no additional protophysics in comparison with picking card (a). Her own account of the impermissibility of picking card (b) is that it would introduce harm to the Asian (the added harm from being killed, and not just losing a foot), which is a necessary means to saving the Africans.

7. Kamm also addresses Unger’s views on tolerating and imposing what she calls nonphysical harm on oneself [by which she probably means nonmortal loss] (pp. 209-14).

a. Unger says that we have a duty to suffer harm in order to prevent greater harm and especially mortal loss to others. He suggests that intuitions to the contrary stem from illegitimate attention to Unger-separation. Kamm argues, on the basis of further hypothetical cases, that the basis for our judgments is different (pp. 210-3).

b. Unger stipulates that our duty to provide for our dependents, e.g. our children, may override our Ungerian duty to suffer property loss to lessen distant others’ mortal loss. Kamm exposes the inconsistency of this stipulation with Unger’s general views (pp. 213-4). First, given Unger’s general views on the overriding importance of minimizing harm impartially, he probably should have said that we have no right to create children and to save money for their education: we should instead maximize the resources that we can and do donate to Oxfam. Moreover, if one were morally permitted to fail to save distant others for the sake of one’s own child’s education, then it would follow, on Unger’s further view about the similarity of not-saving and killing, that one is permitted to kill those distant others for the sake of one’s child’s education–a counter-intuitive result that runs against the spirit of Unger’s book.

8. Kamm also addresses Unger’s views on imposing mortal loss on oneself (pp. 214-6). Unger believes that we have a duty to suffer even mortal loss to prevent mortal loss to two strangers. Kamm points out that in this context Unger fails to discuss a case that elicits relevant intuitions directly. He presents only very indirect evidence, with merely contestable relevance:

a. Unger presents cases that involve sacrificing one’s property, not one’s life. Kamm’s response implies that a duty to lose property to save others’ lives is compatible with the absence of a duty to lose life itself for their sakes.

b. Unger also presents cases where it seems permissible to redirect a trolley to kill 1 person and save 6, and where it doesn’t seem to matter whether the 1 happens to be oneself. In response, Kamm notes that it being permissible to kill oneself for others’ sakes would not establish that doing so is obligatory. She also charitably generates cases where killing someone to save others does feel conditionally obligatory. But she answers that such cases would not suggest that killing the 1 (oneself, say) is unconditionally obligatory; for it is intuitively only permissible (not obligatory) to sacrifice one’s back from breaking to save two other people’s lives. To sacrifice one’s life must not turn out to be more stringently required than to sacrifice one’s back. Finally, she argues that it is intuitively permissible for the 1 person in the classical Trolley case to defend himself against our permissible redirection of the trolley by redirecting it back toward the 6–suggesting that all of us, presumably including the agent, have the prerogative to avoid dying for others’ sakes.

9. Finally, Kamm criticizes Unger’s attempt to explain how someone who accepts Unger’s unusual normative views can live among, and agree with, other folks, who have very different views. She focuses on Unger’s views that we must send a lot of money to Oxfam and that we may permissibly steal in order to do so (pp. 216-20).

a. Kamm’s Unger offers two responses. First, in some contexts, we evaluate actions on the basis of the agent’s subjective beliefs and intentions, and not on the basis of whether he acts in objectively correct ways. (This is Kamm’s phrasing. Unger puts his response differently). Usually a person lacks enough information to conclude with ample confidence that stealing money would minimize harm, and so subjectively the right thing to do is not to steal for Oxfam, and we may praise a non-thief for that. Kamm responds that using notions of (im)permissibility in a subjective sense would not produce real agreement with other folks, because they use these notions in an objective sense (p. 217). Unger also responds that in some contexts, we evaluate actions on the basis of prevalent social norms, and not on the basis of true morality. Someone who says that, e.g., stealing to help Oxfam is impermissible may be using “impermissible” for socially unacceptable. Kamm similarly answers that using notions of (im)permissibility in a social norms sense would not produce real agreement with other folks, since they use these notions in a true moral sense (Ibid.). She adds that someone can act in ways that are objectively wrong according to Unger (e.g. fail to steal for Oxfam) and contrary to this person’s own Ungerite values, and so, Unger must accept, subjectively wrong, but which are right according to social norms. Then, it is unclear whether that person would be said to be doing wrong (subjectively) or right (socially) (p. 218). Kamm thinks that such complications are especially acute in cases of comparative judgment, e.g. the common folks’ judgment that it is worse to steal money in order to give it to Oxfam than it is to fail to give money to Oxfam (pp. 218-9). [Is that really commonsense morality’s approach to Oxfam-serving Robin Hoods?]

b. Kamm concludes by considering, and rejecting, a very different alternate strategy for reconciling some of Unger’s views with ordinary nonconsequentialist moral views. The suggestion is that Unger’s own support for harming people to prevent greater harm to others is significant in a way that Unger did not envisage: his support constitutes implicit consent to sacrificing Unger himself to prevent greater harm to others. According to ordinary morality, a competent adult’s consent to be sacrificed may legitimize sacrificing her. Therefore, ordinary morality might be thought to permit sacrificing Unger on account of his own philosophy. However, Kamm notes (with disappointment?) that being an Ungerite cannot simply be taken as consent to be sacrificed (pp. 219-20). On this (reassuring?) note ends Kamm’s response to the Ungerites.


  1. 1. Posted by S. Matthew Liao | August 11, 2007 3:32 pm

    Nir, thanks alot for your helpful summary and points for consideration.

    To kick off the discussions, on your question 9, wouldn’t Kamm respond that she is a threshold deontologist? That is, she believes that constraints on harming persons have thresholds beyond which they may be overriden (pp. 169-170). For example, she thinks that one may permissibly take another person’s car without his permission in order to rush someone to the hospital in a grave emergency (p. 231). Similarly, she could say, in response to you, that slapping one person as a sure means to saving another from death, or mortally stabbing one person as a sure means to saving a billion others from death, are permissible, because the constraints on harming do not apply in these cases.

    I think you are right in question 2 that Kamm has yet to justify her view that proximity between a person in need of help and the tool that she needs have real moral significance, and I think Kamm will say more about this matter in Chapters 11 and 12.

  2. 2. Posted by Jeff Huggins | August 11, 2007 11:29 pm

    Nir, thanks very much for your very helpful summary and comments.

    I’ll provide my views on several of the questions you posed as well as make a few more general comments. First, the general.

    (Note: I have not read Unger, so my comments on the chapter and on “Unger” rely on Kamm’s views and statements about Unger’s views and arguments, i.e., they rely on “Kamm’s Unger.”)

    It seems to me that this chapter not only reveals flaws in some of Unger’s arguments, but it also shines additional light (continuing from earlier chapters) on some of the problems with Kamm’s own approach (and intricacies), where they exist, in my view.

    Just as one example: Humans need at least some degree of (and sense of) ‘stability’ and ‘security’ and ‘predictability’ and ‘common expectation’ in order to get up in the morning, function, survive, (dare to) have kids, and so forth. We humans know that some types of events can intrude on these, and the world is unpredictable and insecure in some ways, but we rely on those ways in which it is, at least usually, reasonably secure and predictable. Thus, for example, although we wouldn’t like it, we can at least comprehend and understand that, if we are standing on a track (even one that we think is inactive), and if a trolley comes our way, and if we can’t get off, then we may get killed. On the other hand, we can’t comprehend (or comfortably accept) a world in which we might get thrown in front of a trolley without warning, at any instant, while we are standing safely to the side of the track, or on a footbridge above the track. Nor can we comprehend (or comfortably accept) a world in which we might walk into a hospital with a minor ailment and find that our vital organs are being taken to save five other lives.

    Now, if Kamm is presenting Unger’s views accurately, it seems to me that his views try to create some sort of ‘perfect’ (yet mechanistic) world that completely ignores (in at least several senses) what it is to be human and how human life functions. According to Kamm, Unger would have me sacrifice my life right now, and perhaps have my children sell my various pieces (at least the worthwhile ones) in order to save three (or even two) starving adults elsewhere in the world. And other such things. And (if my understanding is correct), he would permit, and even encourage, pushing a person in front of a trolley to save six, or five, or four, or etc. people down the track. At least some (and perhaps many) of the things he suggests would largely undermine the ‘stability’ and ‘predictability’ (as fragile as they already are) necessary for humans to function as biological and psychological beings. So, in this sense, his theory (if I understand it correctly) is ‘ideal’ (but only in some senses) but highly unreal.

    On the other hand, Kamm’s approach, terminologies, intricacies, and explanations, while they result (in many cases) in much more sensible decisions (relative to Unger’s, and in general), are not (sometimes) based on how humans work, and why, in my view. It seems to me that, while some of the intricate principles result in choices that do serve to preserve the necessary (for human life) sense of stability that I mention here, the explanations and terms used do not reflect the root origins and mechanisms of what’s going on, in my view. That is, the intricate principles often serve to preserve these necessities and realities (because intuitive answers to the more clear cases lead in that direction), but the terms and logical forms used in the principles themselves seem to be based on other concepts and other explanations.

    Another example has to do with the role of children in a parent’s moral decision-making, a subject on which there is modest discussion in the chapter. The discussion, tangentially, seemed to raise questions about whether people should have children or ought to avoid having children (in order to better serve people already living) and whether a parent should give up life and property even though her/his children would then be parentless and in a bad situation. That discussion, although brief, seemed to risk overlooking two very fundamental factors: that the reason we humans have social-moral faculties and dynamics at all is that these faculties and dynamics contribute (substantially) to our survival from generation to generation to generation; and that human morality itself won’t even exist 200 years from now unless at least some of us have children today.

    On another note, a couple of Kamm’s comments in this chapter, as well as a helpful comment by Matthew in his post above, bring to mind another observation and (to some degree) concern. As the principles get more intricate, and the cases get more intricate and varied, so many tools and degrees of flexibility have been introduced that, it seems to me, almost any answer can be explained (at least to the more intricate variations) if one wants to justify it. For example, we have (within a non-consequentialist framework) greater goods often justifying lesser evils. On the other hand, we have rights to not be harmed. On the other hand, if Matthew is correct in pointing out that Kamm can be considered a “threshold deontologist”—that is, that “constraints on harming persons have thresholds beyond which they may be overridden”—then our rights to not be harmed can be overridden if the goods are good enough. Then we have Kamm’s comment that “Sometimes, contexts have intrinsic moral significance, and so we cannot draw implications from them about what to do in other contexts.” And, although the more simple and common cases often draw out more clear intuitive responses, the more complex or rare cases often lead to very muddy (inconclusive) results, such as in the earlier Munitions Grief Case and in the current poll regarding The Loop Case. (Last I checked, of 10 responses to The Loop Case, 50% responded “yes” and 50% “no”.) So, my observation is this: It seems to me that Kamm’s approach is about at the point (or perhaps past it?) where it has given itself so many tools, avenues, and degrees of flexibility (nearly everything except a trump card) that it can address and explain nearly any set of results to nearly any case, after the fact. But, it doesn’t seem to do so, in some cases at least (based on what I can tell so far, not having read her earlier works), based on how humans actually ‘work’ from a psychological-biological standpoint, nor based on any explicit view or acknowledgment of the ‘effective’ function of human social-moral dynamics (and morality) in the overall scheme of life. If these avenues of understanding were more explicitly considered and incorporated, I think that the principles, explanations, and terminologies might benefit. (None of this, of course, is to suggest that intuitions are not important. In my view, intuitions at the micro (case) and macro (general value) levels are both very important inputs, and Kamm’s exploration of intuitions is very helpful and thought-provoking.)

    (Nir, since I realize that my post here is already long, I’ll leave my more specific thoughts regarding your great questions for a future post, perhaps tomorrow or early next week.)

  3. 3. Posted by David Wasserman | August 13, 2007 4:45 pm

    Just a quick comment on Nir’s Q10, which raises a critical challenge to Kamm’s critique of Unger. It would hardly be a victory for nonconseuqentialists if Kamm succeeded in showing that DPP tracked our intutions better than Unger-grouping and protophysics. The notion of a shared problem, which underlies Unger-grouping, may be vague and easily manipulated, but it may also provide the intuitive material for a more defensible account of shared vulnerability as a source of constraints and obligations. Such an account would obviously not yield the same judgments as our unreconstructed intuitions, and could reject the judgments yileded by those intuitions as unrefined or overgeneralized. Similarly, protophysics contains a number of “prinicples” involving degree of activity (e.g. diverting vs. initiating) or extent of choice (as resticted or enlarged by the physical possibilities for redistributing harm)that may bear on the agent’s responsibility for harm in a plausible NC account.

    In tonctrast, the DPP’s distinction between causal means and non-causal flip sides does not seem a promising basis for a fuller NC account, for fleshing out the distinction between substitution and subordination. As I suggested in an earlier post, the premium it places on the temporal priority of (some aspect of) the greater good over the lesser evil seems harder to defend than to explain away in terms of the the “endowment effect” — the overgeneralization of the priority accorded vested interests and rights to whatever good happens to have become, however recently, the status quo.

  4. 4. Posted by Jeff Huggins | August 13, 2007 6:21 pm

    David writes, “The notion of a shared problem, which underlies Unger-grouping, may be vague and easily manipulated, but it may also provide the intuitive material for a more defensible account of shared vulnerability as a source of constraints and obligations.”

    I’m not writing here to defend Kamm’s (sometimes) intricacies or Unger’s logic or conclusions, but I agree with what I think David is pointing out in some senses. Because we humans are social and interdependent beings, ours is a ‘shared problem’. And, this shared problem, along with (as David puts it) our ‘shared vulnerability’, are indeed foundationally important dynamics that lead to what are here called ‘constraints’ and obligations.

    As I’ve mentioned in other posts, humans need a basic sense of security, stability, and predictability (however imperfect and subject to certain types of events) in order to function and to “keep on keepin’ on” (as Bob Dylan sings). (See my post #2 above.) If we humans were to accept and contribute to a world in which we could be pushed in front of any big moving object, at any time, by anybody, and that such action (on the pusher’s part) would be encouraged or deemed OK by the community; or a world in which we could have five vital organs ripped from us by a trained medical professional, upon entering the hospital for a minor ailment, and have that professional’s action encouraged or judged OK by the community; then we realize (at least intuitively if not also consciously) that these senses of security and stability that we need would be substantially undermined. Choosing to solve our ‘shared problems’ in such a way would make us all feel very ‘vulnerable’ to a degree that far surpasses (and adds to) our current ‘shared vulnerabilities.’ Our higher moral values call us, intuitively and rationally, to want to protect and save life, but our intuitions also (very roughly and imprecisely) tell us that we must do so in ways that also preserve the basic level of security necessary for the functioning of life. When the situations (or cases) we face allow solutions that respect both of these goals, great. When the situations (or cases) we face force us to choose between one goal or the other, we naturally have problems: Shall I save these specific five (for example) lives in a way that would undermine the sense of security humans (including me) find necessary to live life in general? Or, shall I have a basic respect for the latter consideration, and live life in a way that preserves at least some sense of security and stability, and allow nature to take its course with the five people I might otherwise save? In many cases, solutions allow both goals to be achieved. And, of course, numbers, degrees, and thresholds do matter in many cases.

    Understanding this notion, along with other notions (such as the dynamics of human reciprocity, human empathy, etc.), provides, in my view, a more defensible account of what we are here calling ‘constraints’ and obligations. And, understanding this notion (as well as others) in the broader context of the ‘effective’ function of morality itself leads to a grounded overall framework (paradigm?) that sheds much light on the whole matter, I believe.

  5. 5. Posted by Nir Eyal | August 14, 2007 4:05 pm

    Dear Matthew, Jeff and David,

    Many thanks for your thoughtful comments on my post regarding chapter 6.

    Matthew, you are of course right that Kamm “is a threshold deontologist.” I was puzzled only because “As it is phrased in chapter 6,” Kamm’s principle of permissible harm would seem to suggest otherwise. As I recall, the phrasing there is NOT of the form: “There is a defeasible presumption/reason against A-ing”. It is of the form “A-ing is impermissible.” That said, I am traveling and I do not have Kamm’s book in front of me, and so I apologize if this is incorrect.

    Your suggestion inspires interesting questions that go beyond the phrasing of the principle. How exactly does the good that may override the threshold enter fit in, within Kamm’s scheme? Doesn’t the principle already address the existence of goods to be gained from doing evil, leaving no place for goods that may override it? The principle already relates to doing evil as a necessary means to doing (A CERTAIN DEGREE OF) GOOD. In particular, there are two possible interpretations of how harming someone as a necessary means to doing good would be problematic (and other things equal, impermissible):

    1. To harm a person is evil, and only some causal relations can excuse that evil. The fact that that evil is a means to generating good does not excuse it at all. At the same time, harming a person as the mere side effect/non-causal flip side/condition of action/… for generating a good can excuse that evil. [This interpretation excludes the good to be gained from consideration. Any utilities to be gained are “irrelevant utilities.” Therefore, the challenge for Kamm would be as follows: How can this otherwise irrelevant good suddenly become fully relevant—so much so that it makes transgressions permissible on balance—once the level of good to be gained passes the threshold?]

    2. To harm a person as a necessary means to an end IS the relevant evil. Doing so treats the person as a mere means, and that’s evil. At the same time, harming a person as the side effect/non-causal flip side/condition of action/… of generating a good is not evil. [This interpretation sees the good to be gained as part of the problem. That harming someone is a necessary means to a great good only makes things worse. Therefore, the challenge for Kamm would be as follows: How can the fact that a LOT of good stands to be gained make the act more—rather than less—permissible?]

    Jeff, you criticize Unger and perhaps Kamm for failing to notice that “humans need at least some degree of (and sense of) ‘stability’ and ‘security’ and ‘predictability’ and ‘common expectation’…” For you, this anthropological fact (let us assume that it is a fact) must touch on the questions that Unger and Kamm canvass. You suggest that this fact may explain why it’s not OK to kill people abruptly to use their organs to save several other people, to suddenly push an innocent bystander to his death to stop a hurtling trolley, and so forth. However, can this fact really serve as the explanation? Some instances of harming someone as a means do NOT ruin our collective sense of predictability. A single act usually has little if any influence on overall stability and the collective sense of stability, certainly when the act is done secretly. Surely these instances of killing innocents are every bit as problematic as others. In fact, harming someone as a means may enhance the social sense of predictability. Suppose that up until now, every day at 12 Joe killed an innocent as a means, but that Joe abruptly changes his ways, and so, if you do NOT follow his past actions, your omission will undermine the social sense of predictability. Indeed, minimizing harm by sacrificing someone as a means can coincide with saving predictability. For example, in the footbridge/fat man example, imagine that the 6 who are trapped on the tracks have been trapped there only for 2 seconds, and unpredictably so: due to an accident that no one could have predicted in advance. Therefore, pushing the fat man on the tracks to be crushed by the trolley as a means to saving them would protect our collective sense of predictability more than it would harm it. It is true that pushing him would impose 1 unpredictable death, but it would prevent 6 unpredictable deaths. Does that make it intuitively OK to push the fat man? A further problem is that in the abstract it seems more urgent to save lives than to save predictability and the sense of predictability, certainly given that the impact of individual actions on (the collective sense of) predictability is small. Finally, potential harm to everyone’s sense of predictability is one that can easily enter Kamm- and Unger-style calculations—as potential harm: “Pushing this man will save 6, but it will sacrifice 1 as a means, and it will have the side-effect of slightly increasing the chance of harm to everyone’s sense of predictability.” Kamm’s and Unger’s stylized examples legitimately abstract from many potential side-effects in the interests of simplicity.

    Moving to another issue, you seem to support “human morality.” The latter can potentially designate different candidate constraints on normative theories. It may hold that normative theories must take into account, respectively, what people need, what they identify with, what they could bring themselves to do, what they could calculate, what they usually regard as cruel and incomprehensible, and much else. Kamm and Unger may respond that the onus is on you to sharpen this general suggestion and then to defend its sharpened version. But I would be skeptical. Many candidate versions represent implausible constraints on normative theories: nearly as implausible as those represented by Bay Area morality (which for Bay Area residents rules out normative theories that clash with their needs, their identities, their calculation abilities, and so forth); and by human chemistry (which for human beings rules out chemistry theories that clash with our needs, our identities, and so forth).

    David, it seems like you and I are in deep agreement on some of these issues. It is not ESPECIALLY dignifying to be the sort of being that agents must not sacrifice as a necessary means: not MORE dignifying than it is to be the sort of being that agents must not sacrifice as a necessary condition of action, side effect, non-causal flip side, and so forth.

    Thanks again, everyone, for your very thought-provoking points.


  6. 6. Posted by Jeff Huggins | August 14, 2007 8:39 pm

    Nir, thanks very much for your comments. To clarify, and in response, …

    You ask, “can this fact really serve as the explanation?” My answer is that it serves as an important part of the explanation, also in the context of what I’ll refer to here as the broader ‘effective’ function of morality itself.

    You mention that “Some instances of harming someone as a means do NOT ruin our collective sense of predictability. A single act usually has little if any influence on overall stability and the collective sense of stability, certainly when the act is done secretly.” My abbreviated response (including just a couple aspects of the response) is as follows: First, your comment raises the question (a good one) of how we are to consider cases when they are posed, and what will be done with the principles derived from them? In other words, when considering a case, are we supposed to imagine that these (cases) are one-time instances, and that nobody is watching, and that our decisions and actions will not (and are not meant to) influence or serve as example to others, and that the decisions (and resulting principles) are not meant to apply to similar situations (or that such similar situations will only very rarely arise), and etc.? Are we supposed to assume that these are one-time, isolated, instances-in-a-vacuum, divorced of all social implications? If so, is it being proposed (by Kamm or others) that the derived principles also only apply to identical instances if they are one-time, isolated, instances-in-a-vacuum? Given that we (humans) are a social species, and given how we work, and given how societies (communities) work, and given that the derived principles are (at least presumably) meant to represent or serve as broadly-applicable principles, we should not make the ‘isolated, one-time, instance-in-a-vacuum’ assumption when we consider and address the cases, in my view. Put another way, the question of ‘what if everyone did this in similar instances?’ combined with the assumption that somewhat similar instances will be not-too-uncommon, is (or at least should be) relevant. And, whether or not we want these types of factors to be relevant, they are relevant to our intuitions, I believe.

    Second, your point assumes something as a pre-existing ‘given’ that my answer is meant to explain, at least in part. In other words, your point itself includes the phrase “as a means”. To resolve the issue, we must go deeper and ask, WHY (what is the reason) that we conclude (to a degree) that people should not be used “as a means” in most of these types of situations? I would argue that part (not all) of the reason is the very reason I am suggesting. (Please let me know if I’m not communicating this point clearly. If so, I’ll try again.)

    In response to some of your further points (a bit farther along in your paragraph), I’m not suggesting that the human need for at least some degree of (and sense of) stability, security, predictability, and ‘common expectation’ is the only factor in considering moral cases where one of the solutions might undermine these. Instead, I’m suggesting that it’s an important factor, to be considered along with several other factors. Clearly, there are many instances where ‘stability’ and ‘predictability’ can be created (in one sense at least) through actions that are clearly harmful and wrong. Stability and perceived security and predictability are not the highest values, of course, and I haven’t been arguing that they are. Instead, I suggest that they are very important considerations and that the human need for them explains (in part) some of the underlying dynamics of things that we are choosing (here) to call ‘constraints’ and ‘rights’ and so forth.

    You write that, “A further problem is that in the abstract it seems more urgent to save lives than to save predictability and the sense of predictability, certainly given that the impact of individual actions on (the collective sense of) predictability is small.” My abbreviated response (i.e., some of the considerations): If we should choose from the differing (but related) words I used in my point, rather than using the word ‘predictability’, the better single phrase is probably ‘sense of basic security’ and the things (including basic social dynamics) we feel we can trust that create that sense. Also, ideally (and in most normal cases), situations have solutions that allow us to save lives in a way that also preserves common expectations and a basic sense of security. Usually, we don’t have to use someone else, against his/her will (or without even asking), to try to save someone nearby, at least in common situations that most people experience. It is in cases that, by their nature and set-up, force us into choosing between the two, that we face a dilemma and (as the intuitive results show) either get mixed results or choose on the side of preserving our sense of security, unless some sort of large threshold is met. If we didn’t value (and need) the sense of security that I’m talking about, then many of Unger’s conclusions would be more correct than Kamm’s. Indeed, the sense of security that I’m talking about is one of the main factors that supports/explains many NC conclusions.

    Regarding your comment that “the impact of individual actions on (the collective sense of) predictability is small”, I think my earlier comments address that. I would just add one point on that subject: The intuitions that are used to derive the more general principles, and the more general principles themselves, are meant to reflect and/or apply to broader ‘like’ situations.

    You write, (as an example of your next point), “ ‘Pushing this man will save 6, but it will sacrifice 1 as a means, and it will have the side-effect of slightly increasing the chance of harm to everyone’s sense of predictability.’ Kamm’s and Unger’s stylized examples legitimately abstract from many potential side-effects in the interests of simplicity.”

    My response is severalfold: First, you use the term ‘means’ (along with the implications it carries) and you mention the “side-effect of slightly increasing the chance of harm to everyone’s sense of predictability” in the same description. That’s fine: It would make sure that people get the point. But, the one is contained in the implications of the other. In fact, as I am suggesting, the latter point is a not-insubstantial part of the explanation of the reason that we don’t like being used as a ‘means.’ So, there is some redundancy in the way you word it. Indeed, the way you word it (as an example, not that you would propose to word it that way) demonstrates one of my concerns: By simply agreeing that humans should not be used as a ‘means’ (in most instances), and by leaving it at that, and by using that term, without (perhaps) a rooted understanding of why we don’t want to be used as means, we run the risk of missing some important points and missing out on understanding the underlying reasons.

    Also, it is not at all clear (to me anyhow) that consequentialists or NCs explicitly consider these types of factors. The notion of (as you put it) “abstract[ing] from many potential side-effects in the interests of simplicity” is OK if we all know that these important side-effects are there, and if we consider them. If they are completely missed, intentionally left out, or misunderstood, that’s another matter.

    Also, the explicit (or conscious) idea that pushing someone off a bridge to stop a trolley leads to a ‘side-effect’ of reducing everyone’s sense of security (if it is seen or conveyed as an OK or right action) is only one part of the equation, i.e., the part that has to do with our conscious rational comparison of consequences and other considerations. The other part is that our strong intuitive hesitancy to make that choice and take that action (unless the benefits are immense, and unless the conditions are just right and nearly require such action) is itself an intuitive (you could say subconscious, instinctual) outcome within us that results from the fact that we have evolved in a way that wants the sense of ‘security’ in life that would be undermined if we generally acted in such a way. In other words, we have been shaped to not act in that way (i.e., to push each other off bridges). We just find it difficult to explain to each other exactly why. And we name it with different terminologies.

    Moving to the other issue you mention, I’m not sure what you mean when you say that I “seem to support ‘human morality’”? I certainly don’t think, of course, that humans always behave morally. You write that “Kamm and Unger may respond that the onus is on you to sharpen this general suggestion and then to defend its sharpened version. But I would be skeptical.” It’s not clear to me, yet, what you would be skeptical about. But, I’m always happy to present and defend my views and theories. If you like, you can read them by visiting my Web site (which you can reach by clicking on my name above): There on the Web site, on the page called ‘ADDITIONAL MATERIAL’, I’ve posted several papers. The best way to gain an overview, if you like, is to first read the one titled “A Framework and Paradigm Of Morality” and then to also read the first five or six pages of the one titled “Portions of the Supporting Argument in Additional Forms.” The third one to read would be “The Nature of the Relationship Between Is and Ought”, but the other two mentioned here should come first. I’m very interested in comments on those, which you can send to me via my Web site, if you have a chance to read the papers. Or, of course, I’m happy to answer specific questions here.

    Thanks again, very much, for your great comments and questions.

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.