Wednesday, March 7, 2012

Virtue Ethics & the Principle of Double Effect

In my last post I offered an ethical analysis of Sophie's Choice from the perspective of virtue ethics (VE). The exercise was fruitfull indeed to my own thinking and I received many great comments (in the ComBox). I wanted to do these comments justice in the form of another blog post.

What is the Principle of Double Effect (PDE)? It has been formulate by Joseph Mangan as follows:

(1) That the action in itself from its very object be good or at least indifferent.
(2) That the good effect and not the evil effect be intended.
(3) That the good effect be not produced by means of the evil effect.
(4) That there be a proportionately grave reason for permitting the evil effect.

From PDE there arises a prima facie conflict between VE portrayed in my blog post and PDE. It seems that if PDE tracks moral obligation then there will be occasions where you can separate moral obligation from acts of virtue. That is, given that to act virtuously is to act for what is most ordered to your happiness (through participation in that which is ordered to the happiness of humanity as such) it seems that there will be occasions that arise where one is obliged not to do what is virtuous. On one hand, if Sophie ought to have applied the PDE, and in so doing she was obliged to choose one of her children over another, this would have prevented her from doing a virtuous action (i.e., not making a choice as to which child was to be murdered, as consistent with VE) for sake of what she was morally obliged to do. Of course, a requirement for PDE is that the act done must itself be morally good or at least indifferent. On one interpretation of the application of the PDE to Sophie’s situation, choosing to save the life of her son would be morally good, but with the bad effect of her daughter dying.

So, how should this issue be resolved? I think that the answer is to reflect on the notion of moral obligation in VE. In VE, one conducts moral reasoning based on an admixture of particular conditions given in the context and universal moral principles. The latter principles are universally and a priori obligatory precisely because they are inherently contrary to human perfection, nobility and happiness. However, the context that the moral agent finds herself in is contingent and unique. Because of this, oftentimes a moral agent has several options to pursue (virtuous actions to enact) and settling on one of them will be the result of a rational choice. For example, even when following those universally morally obligatory principles, in Sophie’s situation, she still could have: applied the PDE, chosen her son, chosen her daughter, chosen neither, tried to escape, tried to appeal to the Nazi’s sense of justice, attacked him, try to confuse or distract the Nazi, etc., or some combination. Because of (5) the fact that the context determines the possible goals that can be accomplished and rationally chosen; and, (6) since there are not always universally obliging principles that will restrict a moral agent's options to one only; it follows that (7) in many cases there is no moral obligation to do one action only, and there are a variety of options open for rational choice.

However, even though there may be no moral obligation, there is still a question of moral excellence and virtue. In these cases, a VE would say that there is a “right thing to do,” but that it does arise in the same way as a moral obligation. How the morally excellent thing is determined is famously put in the Aristotelian dictum, “The virtuous thing to be done is that which the virtuous person would to do!” Since the virtuous person is an excellent practical reasoner, and has a habit of making choices that are morally excellent, she is most capable in choosing that action that hits the golden mean requisite for virtuous action (which falls between two vices). Yet, two things must be kept in mind: First, that what a given virtuous person chooses to do may in principle be different than what a different virtuous person chooses to do. Second, that one is universally obliged to attempt virtuous activity.

Given this background, if PDE is not morally obligatory (a question I address below), Sophie still should do what a virtuous person would do. However, given the moral context, she may have some options regarding what it is that would be most excellent to do. Given that we are learning about Sophie’s situation through a limited and fictitious narrative, and so are not privy to all of the concrete elements of her situation (as well as the various things that she is capable of), we must settle for a ceteris paribus consideration of it. Even the addition of one more concrete factor could change what the most morally excellent or virtuous thing would be. However, given the foreseeable options that she was faced with, in doing the morally excellent thing, I argued that she would give up both children (of course, I am not claiming for myself the prestigious designation as "the virtuous person"). For example, it would have been rash (the vice which is excess of confidence; the virtue which is courage) for her to attack the Nazi, or to try to escape, etc. Although this would have been the most morally excellent thing to do, it was not obligatory, because of the contingency and indeterminacy of her epistemic moral situation. Therefore, assuming that (7) applying the PDE is not morally obligatory (it ought to be applied and followed in every occasion in which it can be applied and followed), and (8) that applying and following the PDE is not the most morally excellent or virtuous thing to do in her situation, it is reasonable that the morally excellent thing for Sophie to do would be not to choose even though she is not morally obliged to do so.

But what about PDE? We have three options: either (9) PDE is a universal moral principle, such that in every moral situation in which can be used it should be used, or (10) PDE is not a universal moral principle but one that in specific contexts would be a principle that a virtuous person would utilize. Finally, there is the option that (11) PDE is false as a moral principle.

Whereas a VE need not hold ‘(11)’, she also need not hold ‘(9)’.

In fact, I think that there is good reason to deny ‘(9)’. This is because one should have a worry about PDE that it sometimes falls on the side of a consequentialist calculus to the exclusion of supporting the larger goal of morally virtue. We need to think about what the motivation is for utilizing the PDE. Rather than moral excellence in personal virtue as its target, it seems to be motivated by maximizing good and/or minimizing evils without falling afoul of other moral obligations. Now, of course, VE is interested in obtaining good results, even within the moral sphere, but in light of what is most noble and good in human nature. It needs to noted that in principle it is possible for one to become vicious (indirectly) even though one is applying PDE. One may even imagine a scenario where one is obliged to commit a vicious action as an ill effect of some good action. There is the further factor that since one is trying to obtain some particular effect among others, their choice is voluntary (more about this below). One can extrapolate even further and envision an unlucky chap who is morally obliged to commit, due to PDE, so many indirectly vicious actions of one kind that he habituates a vice. In the end, you become the things that you voluntarily do, even if you are responsible for them as a result of something else that you intend. Thus, it appears that ‘(9)’ is false.

(Of course, if ‘(9)’ is false, the use of PDE needs to be rethought if it is to serve a VE. It is to be noted that PDE is derived from a fairly minor point made by Aquinas in of his Summa theologiae and is arguably not a central feature of his moral teaching.)

Now, regarding ‘(10)’, some may argue that in this case applying the PDE is the most virtuous thing to do in Sophie's case. Some may say that it is noble to make the difficult choice of sacrificing a child for the sake of preserving the life of another. However, if it can be indicated that in applying PDE Sophie would be utilizing consequentialist moral reasoning such that an evil action is willed for the sake of a good result, then it will be shown to be inconsistent with what is virtuous. On this ground, we can exclude an application of the PDE as being the most virtuous choice in her situation. I am assuming her that pace Mill VE is inconsistent with this central consequentialist tenet. Compare the two following propositions in relation to Sophie’s’ choice:

(12) Sophie sacrificed her daughter in order to save her son.

(13) Sophie saved her son from death with the unintended bad result that her daughter would be murdered.

It looks to me that Sophie’s choice is as much in line with ‘(12)’ as it is with ‘(13)’.

For one thing, ‘(12)’ expresses a consequentialist means-ends reasoning. The reason why Sophie would be using her daughter’s death as a means of saving her son is because in the concrete situation of her making her choice, she can be said to have intended the death of her daughter--insofar as she could have made it not occur--as a trade-off for the life of her son. When her daughter was being taken away to her death, one can imagine the little girl thinking, “My mother chose for me to die when I could have lived!” As Aristotle suggests, despite strong external forces influencing a person’s decision, there is no reason that an action cannot be voluntary if it follows internally and from rational choice. Although I am sympathetic to the notion of material cooperation as weaker than intentional cooperation, in Sophie’s case although she does not want the evil, through rational choice, it is not devoid of volition. In fact, once she commits to saving one child over another, immediately a utilitarian process of reasoning seems to take over, leaving us with the question of why Sophie chose her son over her daughter. Once this line of thought is engage, she can be said to cooperate with and take possession of the evil in more “formal-ish” and means-end way; i.e., more formal, then if she refused to make the choice. Perhaps this in part explains the extreme guilt that she felt that resulted in her suicide, after she was released from the prison camp. This being said, it seems to me that the application of the PDE in this case would be tantamount to consequentialist reasoning and far from being the most excellent or virtuous thing to do.

For another thing, as Aquinas suggests, sometimes the final cause is contained in the formal determination of an action and becomes included in the object of the moral act. The example that he gives in the beginning of ST is that a person who steals in order to commit adultery is more formally an adulterer than a thief. If this is so, then, although from one perspective her primary purpose is saving her son when sacrificing her daughter, it is also true that she is choosing to sacrifice her daughter for the sake of saving her son. Although I recognize the difference between material and formal cooperation in evil, I still think that there is enough doubt regarding the application of PDE in this situation as the most excellent and noble and virtuous thing to do.

If the end of our action is sometimes included in the moral object that defines our action, given the fact that moral contexts are concrete, it may be the case that PDE doesn't apply in Sophie's situation at all. As mentioned above, PDE requires that the action done must itself must be either morally good or neutral. It may be too naive to suggest Sophie's choice to save her son does not also include the further notion that it is at the expense of saving her daughter. I think that applications of PDE can easily lapse into an abstractivism of convenience when identifying the act that is good or neutral that can serve as the primary intention for our actions while the more concrete and morally relevant factors fall into place as the secondary effect.

I maintain that PDE is an available tool for determining what is morally excellent in specific contexts, but that it does not engender moral obligation. However, I believe that PDE should not be used in Sophie’s situation and instead she should opt for non-cooperation. As well as not rejecting PDE out of hand, I also do not want to suggest that it is always vicious to act in cooperation with evil, when taken materially.

Ultimately, we need to look at this debate as a different conception of what the fundamental goal of ethical analysis and ethical behavior is. A virtue ethicist exclusively has as her goal the habitual perfecting of one self in the image of the good and nobility of humanity and not the more casuistic goal of avoiding doing wrong things as determined by obligation. This is the case even though the notion of moral obligation has its place in VE as principles that are universally contrary to human good and happiness (such as adultery, murder, etc.).


  1. Thank you for these posts, Paul!

    A few initial thoughts:

    First, the way you construe Sophie's choice, namely, "Sophie would be utilizing consequentialist moral reasoning such that an evil action is willed for the sake of a good result, then it will be shown to be inconsistent with what is virtuous. On this ground, we can exclude an application of the PDE as being the most virtuous choice in her situation."

    But a proper understanding of the PDE demands that the evil *not* be directly willed as a means for the sake of a good end. You seem to be saying that she is *willing* the death of one child as a means to save the life of another. If that were the case, then this scenario violates Mangan's third criterion, that the evil *not* be the means to obtain the good. (Incidentally, even if Mangan wrote it with this language, principle number two, "That the good effect and not the evil effect be intended" is imprecise. It should read that the evil effect may not be directly willed [whether as an intended end or as a chosen means]. This is a very important distinction! And it's based on Thomas' identification that the will intends the end and chooses the means. Moral evil arises if either act of will is directly ordered to something evil.)

    Second, in Sophie's case, it is not necessary, it seems to me, that she must directly will the death of the one child. In this particular case, it may be that she foresees that as a possible evil effect, but it is not a necessary effect (since circumstances or persons may intervene preventing the murder of that child intended by the soldier).

    Third, the way that the PDE historically arises in the history of Catholic moral thought is not as a way to maximize good and minimize evil, but rather - as in the example from Aquinas' "Summa" you cite - in cases where (virtuous, good) action seems obligatory though attended by foreseeable evil effects. The examples of such situations mentioned by Aquinas include the well-known capital punishment one and the less well-known gangrenous limb one. The latter is less controversial in our present context. It would be virtuous (good, prudent, just...) to preserve bodily life in the face of lethal gangrene, if you were able to do so without directly willing evil (whether by intention or choice). Amputating the limb is a foreseeable side-effect (and one that in this case, unlike Sophie's, is unavoidable) that, as long as it's not directly willed, is understood as morally permissible and, I'd argue *virtuous*.

    On that note, applying the PDE does not necessarily entail a moral vision that allows morally licit actions having nothing at all to do with virtue (despite the fact that PDE has been abusively invoked to justify vicious actions). I think a cautious and correct application of PDE can be not only morally licit, but even virtuous.

    These are just some rambling comments...

    1. Mike, thanks for your comments.

      Regarding your first point, what I was suggesting was that in this unique case (I'm not saying that this is always the case), and why it makes it perhaps more difficult from other ethical dilemmas regarding the application of PDE, is that even on the most earnest application of PDE, one cannot avoid nevertheless transgressing Mangan's third criterion. This is why I say that '(13)' cannot be extricated from '(12)'. My worry about absolute applications of PDE is that it benefits from a convenient abstractionism that in this case seems naive and occluding one from what's virtuous.

      Regarding your second point, it would be a prudent assessment that the one to be chosen would in fact die and we depend on such prudent judgment in sizing up morally relevant situations. In consenting to her daughter being taken, she consented to her being taken away as a means to sending her to her death. To claim otherwise, it seems to me, would be like saying that when someone puts cyanide in someone's drink they are not consenting to killing them because they may not drink it, or it may not actually kill them.

      Regarding the third point, it seems that you are assuming precisely what I think a VE is seeking to deny: namely, that one has an obligation (in the way I define it in the post; as a priori holding) to keep one's offspring alive at all costs (or something like that). Although for practical purposes is rather a strong commitment that we as parents have and admits of few exceptions, there are occasions when we need not follow this, for example, a pioneer father giving his youngish son a rifle and the charge to help him protect his homestead from bloodthirsty intruders. So I see no obligation there. Although it would be virtuous to save one child, the virtue of the action is undermined by sacrificing the other one, and participating in the evil perpetuated by the Nazi.

      Hey brother, thanks for reading and commenting on the post!

  2. Okay, I just read through the rest of the post more carefully and I might have to modify some of my initial comments (if/when I can slip in some time to do so...)

  3. I love your blog - I've never so learned so much about one subject from just one blog before. Have to look up a lot of it on Wikipedia, but it's sinking in.

  4. This is a minor point and I realize it distracts a little from your main points, but as for attempting to escape, I think that tended to make the Nazis take revenge on whole groups of innocent people (that's what was happening to St. Maximilian Kolbe's group for instance).

    Could you address what Thomism would say about the whole innocent group's moral obligation (if they had one) to the rest of the group?

  5. Paul:

    This regards the question whether Sophie intended her daughter's death.

    Imagine that the Nazi instead gave her a choice of which of the three of them he should take away, and she volunteered herself. Clearly, there would be nothing morally problematic about that: that would be an act of moral heroism, of great charity.

    But it is just as wrong to intend your own death as the death of any other innocent. So if handing someone over the Nazi is tantamount to intending death for that someone, then it is wrong to hand oneself over. But it's not wrong to hand oneself over. (The case of St Maximilian Kolbe is a case in point.) What one intends is not death for oneself, but something else, maybe just as that one be taken away by the Nazis.

    The same is true when one hands one's child over.

    Now, back to the virtue point. Imagine two scenarios. Sophie1 hates her daughter and loves her son. She hands her daughter over because she hates her. Here, there is vice in play. Sophie2, however, genuinely loves each of her children as herself. In that case, handing her daughter over is much like handing herself over: her daughter's pain will be her own pain. And hence this is an act of virtue.

    So whether the action is virtuous or not depends on further questions about mental attitudes of the agent.

    1. Alex (sorry about the delay),

      First, I think that handing yourself over is different from handing over one of your children, because of the nature of courage. The virtue of courage involves risking your own life for the sake of another or some noble cause, etc. I can't connect up sacrificing someone else with a virtue. One may say that you are performing the virtue of being a good parent by sacrificing one of your children for the sake of the other, but that doesn't seem to work.

      I think that it is premature to conclude in your Sophie1 & Sophie2 case that the former is vicious and the latter is virtuous. I think that your analysis is correct if you apply Sophie1 and Sophie to your earlier example about only being able to save one of your children from drowning. If your choice to save one of your children because you hate the other one, that would be vicious, I agree. And, I also agree that it is virtuous to be willing to give up one of your children for the sake of saving the other one in the drowning case. But I see the Sophie's Choice example differently.

      It seems to me that there are at least two ways of failing to act in a way that accords with the nobility of the idealized human being (the so-called eudaimonian measuring stick for our actions by which we participate more or less in happiness). (1) By failing to perform an action that is sufficiently proportionate to the ideal for eudaimonia. (2) By intending and doing an action that is positively contrary to that which is consistent with the idea of human eudaimonia. I think that Sophie fails to participate in eudamonia in sense (1) only when she allows both children to be taken away. I see that if Sophie were to choose one of her children, she would fall short of eudaimonia in according to sense (2) because she would be practically uniting her intention with the originally evil intention of the SS officer in such a way not to easily to be disentangled. I think it is for this reason that many of us have an intuition in this scenario that Sophie is doing something wrong when she allows one of her children to be taken away.

      It seems to me that given the possibility of either (1) or (2), out of desire for eudaimonia, and something approximating to it, one should opt for (1) and not (2).

  6. Paul,

    I wonder if you are expecting something more from PDE than it was ever intended to provide. As I understand PDE, it is a principle for determining permissibility, not obligation, and even less for determining "the most excellent action" that can be performed. When Aquinas introduces it for self-defense, he means only to say that to resist an attacker with a force that may be lethal is permissible, not that you are obliged to defend yourself this way. You may be a pacifist, who declines the permission that PDE gives you to use potentially lethal force in defending yourself. Thus it is no difficulty for PDE at all that an action that it declares to be permissible is not the most virtuous action that could be performed.

    I do think that the PDE shows that Sophie need not intend the death of her daughter and that her choice is therefore permissible. And I think that Alexander Pruss supports this conclusion well with his argument.



    1. Hi John,

      My apologies for the delay.

      I agree that it is somewhat strange for me to bring PDE into the context of discussion over obligation and determining excellent action, but I was trying to make sense of it within a unified eudaimonian ethics. It does seem to me, though, that permissibility and obligation are logicall related: if x is permissible, it is not obligatory; if x is obligatory then it is not permissible, etc. I was trying to tie PDE in with VE by assuming that x is moral when x is that which is most inline with eudaimonia. Having assumed this, I was trying to see how PDE fits into to that model. Is it something like prudence itself: something obligatory on our actions; something that we should seek to live fully inline with? Or, is it something that in specific cases should be used because it would yield the most excellent action? On VE, a morally virtuous person should not do x if x is not in someway related to excellence in action. Thus, the question arises why someone would be motivated to apply PDE if its goals are independent of this ultimate goal.

      I hope that my thoughts have been made clearer on this point and thank you for your point.

  7. Ethics is the activity of man directed to secure the inner perfection of his own personality.