Appendix 3: The principle of double effect

In the older treatises on moral theology, one finds a treatment of a difficult type of moral problem in which there might seem to be but really is not a conflict of moral responsibilities. For instance, a pregnant woman is diagnosed as having cancer of the uterus. Treatment of the disease is likely to result in the death of her child; nontreatment, in the spread of the disease and the woman’s death after the delivery of the child. At first glance, both treatment and nontreatment seem justifiable in view of the goods sought and unjustifiable in view of the harms expected. Cases of this type were resolved by clarification of the moral act rather than by clarification of the relevant norm.

The clarification of the moral act began by noticing that in cases of this sort, the same act has two effects, one good and the other bad.¹⁴ The clarification, called “the principle of double effect,” usually was summarized along the following lines:

One may perform an act having two effects, one good and the other bad, if four conditions are fulfilled simultaneously:

1) The act must not be wrong in itself, even apart from consideration of the bad effect. (Thus the principle was not used to deal with the good and the bad effects of an act admittedly excluded by an absolute norm.)

2) The agent’s intention must be right. (Thus if one’s precise purpose is to destroy, damage, or impede some basic human good, the deed carrying out this purpose could not be justified by the principle.)

3) The evil effect must not be a means to the good effect. (Thus if one chooses to destroy, damage, or impede some basic human good, although one chooses this for the sake of a good one might otherwise rightly pursue, the deed carrying out this choice could not be justified by the principle.)

4) There must be a proportionately grave reason to justify the act. (Thus, even if all the other conditions were fulfilled, one still might be obliged by the moral significance of the expected bad effect to abstain from the action.)

A moralist working in the classical framework would have applied this principle to the case of the cancerous uterus as follows. The treatment of the cancer itself is not a bad act. If the woman were not pregnant, no doubt it would be obligatory. The purpose is not to damage or kill the unborn child. The operation is unlike one done to get rid of an unwanted child; if possible, the child will be saved. Moreover, whatever harm comes to the child is not a means to the good sought by the treatment. If the woman were not pregnant or if the child were miraculously preserved from harm, the treatment of the cancer would be every bit as effective. The case is unlike one in which the child is killed in order to lessen the load it is putting on the mother’s system. Finally, there is a very serious reason for going on with the treatment despite its bad effect; the mother’s life is just as much at stake as the child’s. Thus the case is unlike one in which a pregnant woman has a medically indicated but not urgently necessary hysterectomy (removal of uterus) despite her pregnant condition, thus bringing about the death of the child merely to do something which could be done later.

As formulated, the principle of double effect gives rise to three sorts of difficulties.

First, what is to count as “the act” mentioned in the first condition? Are abortionists killing babies? Or is the act something else? For example, does an abortionist who injects saline into a pregnant uterus precisely do the act of injecting saline, with the death of the baby as an effect of the act? Again, if a soldier in battle takes dead aim at an enemy soldier, is the act one of killing?

Second, what is to count as “a means” mentioned in the third condition? Is every cause of a desired effect a means to the end sought? For instance, if the effect desired by the soldier is his own safety, and if the death of the enemy soldier in fact secures this, then is the enemy’s death a means to the good end?

Third, what is to count as a “proportionately grave reason” mentioned in the fourth condition? Does the physical proximity or probability of the bad effect have something to do with this? Is the justification in the cancerous uterus case simply that one life is as good as another? Should one take into account the probability that the child could live many more years than the mother? Or can one consider here factors such as the woman’s responsibilities for other children?

It seems to me there are two sources of these difficulties. First, the older moral theologians started out by thinking of human acts in a commonsense way, as chunks of behavior having some moral significance because of their inherent characteristics and their being done on purpose. If one takes this view, one literally never knows exactly what anyone is doing, and so one will not be able to deal with precision with difficult cases of the sort for which the principle of double effect was designed.

Second, the classical moralists also lacked an explicit understanding of the modes of responsibility. Hence, they could not define “proportionately grave reason” simply in moral terms, and they sometimes talked in ways which have provided an opening for proportionalism.

Nevertheless, it is worth noting that many of the older moralists, in talking about what is proportionately grave, suggested that its meaning at least included the following consideration. In doing something which brings about unintentional bad consequences, one still might be acting irreverently with religious things, unfairly toward other people, or recklessly with one’s own well-being. If so, one’s reason would not be proportionately grave. This consideration clearly avoids anything like proportionalism, since it brings into play other relevant moral norms rather than proposing to weigh and balance the good and bad effects considered prior to moral specification—that is, taken simply as basic human goods.