Arcane principle behind traditional teaching on gynaecology and ethics is bad philosophy

James Mackey reminded readers recently, on these pages, of a distinction taught to generations of clerical and medical students in Ireland, between causing the death of a foetus "directly and intentionally" and doing so indirectly. This distinction is borrowed from medieval theology, and is called the Principle of Double Effect (PDE for short).

Thomas Aquinas (1225-74) argued that, although God commanded: "Thou shalt not kill", killing another person in self-defence is permitted in some cases, and the state may even kill heretics because they "corrupt the faith". Since then, PDE has been applied to a wide range of ethical problems, including some that occur in medical practice.

Unfortunately, PDE is poor philosophy; it is more likely to confuse than to resolve ethical dilemmas. If someone shoots Joe dead, they might try to avoid responsibility for their action by contracting the scope of the action and describing it as follows: “I merely moved my index finger; I never intended Joe’s death.” It seems obvious, however, that a moral or legal evaluation of an action requires an initial description that includes all its relevant foreseeable effects.

Many human actions include an indefinitely large number of consequences, some of which occur long after the initial action. An agent may wish that some result would not occur, or may mentally select a particular result as their primary objective.

But such private thoughts are irrelevant to the morality of their action. Pascal, in his Provincial Letters, famously mocked the "method of directing the intention, which consists in selecting something permissible as the objective of one's action". All the foreseeable effects of any voluntary action are included in its intentional description.

PDE permits an action that has multiple effects, of which at least one is usually prohibited, if the morally objectionable effect is not “intended” and if the benefits achieved are proportional to any harm caused. Assuming that we can subdivide a complex action into discrete parts, each of which is evaluated separately, PDE also prohibits doing something morally evil as a means to a good end.

Despite its intuitive appeal, however, it has proved impossible since the middle ages to provide a plausible version of PDE, and most philosophers reject it today. Why?

There are good reasons not to begin any ethical assessment with unconditional prohibitions that create the moral dilemmas to which PDE is applied. Rather than say: it is never permissible to cause the death of a human being, one should say: it is almost always impermissible, but there are exceptions, such as killing someone in self-defence.

Another fundamental objection is that Aquinas assumed that the moral quality of an action is determined by the intention of the agent, so that changing one’s intention could change an immoral action into one that is permitted. That confuses the morality of the agent with the permissibility of their action. Good agents may perform bad actions, and vice versa. Intentions are relevant when deciding if someone is responsible for their action, but the permissibility of an action also depends on other factors, including its consequences.

Third, the distinction between “intended” effects and those merely foreseen is arbitrary or artificial, and cannot support a rationally defensible distinction between morally permissible and impermissible actions. It is not possible, simply by having different thoughts, to change the morality of an action.

Finally, proponents of PDE assume that one can evaluate discrete features of a complex action that has multiple effects as if each were a standalone action, and then discuss whether one feature is a means to an end. For example, they say: it is not permissible to abort a foetus if that is necessary to save a mother’s life, because that involves “directly” or “intentionally” doing something impermissible. In contrast, they concede that it may be morally permissible to remove a pregnant woman’s womb for medical reasons, even if the death of a foetus results. That analysis assumes a moral distinction between two medical interventions because of the causal sequence in which multiple effects are related to each other. But when medical necessity requires a particular action, the sequence in which the effects occur is morally irrelevant. Either option, if medically required, is an exception to a general rule against taking human life.

PDE is an unreliable guide when legislating for medical dilemmas.

Desmond M Clarke is emeritus professor of philosophy at UCC.