RSS Subscription Logo
Catholic Medical Quarterly Volume 62 (1) February 2012, p20-26

Double Effect in Beauchamp and Childress

By Joseph Shaw

Introduction

In this paper I will examine and criticise the discussion of the ‘Rule of Double Effect’, better known as the Principle (or Doctrine) of Double Effect (PDE), in the widely used textbook, ‘Principles of Biomedical Ethics’ by Tom Beauchamp and James Childress.[1] The authors address it in a short section in their chapter on ‘Maleficence’ (pp162-167).
My own initial characterisation of the PDE would be this: that it focuses attention on the intention of the agent, as being the key to determining whether he is breaking a moral rule such as ‘do not kill’. If he acts with the intention of bringing about a death, then he is breaking the rule; if not, the action does not break the rule, though it may be wrong on the grounds of negligence or recklessness. The same approach can be taken to other moral rules. The ‘double effect’ (i.e. two effects) of the agent’s action are its intended effect and the merely foreseen effect it has: when a death is foreseen, as a result of a proposed action, but not intended, then if there are strong enough reasons to do the action it can be permissible.

1. The moral relevance of intention

Scales of JusticeBeauchamp and Childress treat the PDE as among a group of related distinctions which are ‘outmoded and untenable’.[2] The accusation of being ‘outmoded’ is surprising since interest in the principle has enormously increased in the last half-century, but no argument from philosophical fashion is given in the following text, so this should perhaps be regarded as mere rhetoric. The two main arguments which are given against the PDE focus on the moral relevance of intention, and about the definition of intention, respectively.

The first argument depends on a comparison between two pairs of cases.[3] The first pair is the hysterectomy of a pregnant cancer patient, and the removal of a fallopian tube containing an ectopic pregnancy. The second pair are abortion for a mother with a weak heart (where the continuing pregnancy ‘will probably result in her death’), and the craniotomy of a fetus in the process of being born (where the mother otherwise ‘will die’). Beauchamp and Childress claim that ‘according to proponents’ of the PDE the first two actions are permissible, and the second two are not.

No attempt is made by Beauchamp and Childress to distinguish the conceptual structure created by the PDE and the specific moral content given by Catholic teaching. While I am personally equally happy to defend both, it would simplify matters considerably if it were possible to defend them one at a time. The PDE is in principle separate from the question of the moral status of the unborn child, or the sacredness of human life. It applies to rules, and could be used for rules with quite different contents. It is particularly unhelpful for Beauchamp and Childress to use only examples involving unborn life, which appears to link the PDE with the issue of abortion, and to suggest that the coherence of the PDE depends on the defence of a presumed interpretation of the Church’s teaching on craniotomy, which is one of the most controversial and conceptually complex issues in Catholic medical ethics.[4] Their discussion slides from characterising the PDE in terms of ‘Roman Catholic teaching’ to ‘proponents of the [P]DE’ as if there were no difference.[5]

Beauchamp and Childress’s argument at this point is that ‘it is not likely that a morally relevant difference can be established’ between the first pair of cases and the second pair. In neither cases does the agent want or desire the death of the fetus, and the descriptions of the acts in these cases do not indicate morally relevant differences between intending, on the one hand, and foreseeing but not intending, on the other.[6]

It is necessary to remind oneself here that the PDE encapsulates the claim that whether a harm is intended or merely foreseen is of great moral significance. Beauchamp and Childress appear to be demanding that the presence or absence of an intention to harm be an indication of the presence or absence of a separate, morally relevant, feature of the cases. This argument simply begs the question against the claim that intention is morally relevant in itself. There is no argument that intention is not morally relevant, just an implicit assertion.

The only way of arguing in favour of the moral relevance of intention is by appealing to moral intuitions, and such arguments are not difficult to make. If you discover that the driver who scratched your car did so by intention, and not, as you had assumed, accidentally, negligently, recklessly, or perhaps foreseeably, but nevertheless for some other reason, the moral seriousness of the act is greatly increased. Similarly, if you hear that what you thought was a deliberate murder was in fact done accidentally, negligently and so on, or that the agent foreseeably, but without intention, brought about the death in performing an action to save several other lives (exploding a terrorist’s bomb where the fewest possible people were present, for example), then the moral gravity of the action is greatly reduced; it may even turn out to be a good action. Beauchamp and Childress do not make any such investigation of moral intuitions, however, and merely assume that intention is morally irrelevant.

2. The concept of intention

They go on to discuss the concept of intention itself in more detail. As they put it,

One of the few widely shared views [on the PDE] is that intentional actions require that an agent have a plan… For an action to be intentional, it must correspond to the agent’s plan for its performance.[7]

They then give an example of what this means (from Alvin Goodman), a man sticking his arm out of his car window in order to signal, knowing, but not intending, that his hand will get wet, since it is raining. Getting wet was not part of the plan.

Beauchamp and Childress explain that the defender of the PDE needs the ‘narrow’ understanding of intention this implies, and distinguishes between

effects that are desired and wanted and effects that are foreseen but not desired or wanted. …the latter effects [are] foreseen, but not intended.[8]

It is surprising to read, then, in the next paragraph, that the distinction they have just expounded should be rejected. Since, they say, non-intended upshots of action which are foreseen by the agent can be said to be ‘tolerated’ by the agent, the agent in this sense ‘includes them as a part of his or her plan of intentional action.’ They then suggest that such upshots are ‘willed’, and on the strength of this propose a conception of intention which does not distinguish between upshots which are intended and those which are merely foreseen:

…intentional actions and intended effects include any action and any effect specifically willed in accordance with a plan, including tolerated as well as wanted effects.[9]

(The word ‘specifically’ was added to the text in the Fifth Edition.)[10] They go on:

Under this conception of intentional acts and intended effects, the distinction between what agents intend and what they merely foresee in a planned action is not viable.[11]

The next step is to apply the broken-backed version of the PDE which makes use of this broadened notion of intention to the hysterectomy case, and find, unsurprisingly, that it fails to work.[12]

The real argument here, accordingly, is against the coherence of the distinction between intention and foresight. It appears to consist in the claim that there is something unsatisfactory about saying that upshots which an agent takes into account are not ‘willed’ or intended. Beauchamp and Childress don’t want to limit their point to those upshots which form what we might call a working part of an agent’s plan, as might be suggested by the phrase ‘specifically willed in accordance with a plan’. On the contrary, they give the example of a person who wants to turn on a light, noting with irritation that the same switch activates a noisy fan, something he ‘wants to avoid’: in this case he ‘intends the effect’ even though he ‘did not intend it as the goal of the action’.[13] The sound of the noisy fan does not appear to be part of the agent’s plan at all.

The last quoted sentence was added to the Fifth Edition, to clarify the rather ambiguous formulation ‘it would be conceptually mistaken to say that he unintentionally brought about the obnoxious sound by flipping the switch.’ He certainly did bring it about by an intentional action, just as I might unintentionally knock over a table by the intentional action of jumping to catch a ball. Beauchamp and Childress are, however, committed to the much stronger claim that it was not simply the result of an intentional, that is, a deliberate, action, but was intended as an upshot.

It is actually very easy to give examples of agents who clearly don’t intend all foreseen upshots, as the word ‘intention’ is normally understood. A classic one is stammering Peter, who wishes to defend his dead father’s honour, but foresees, with dread, that if he does so he will stammer and make a fool of himself. The suggestion that he intends to stammer is ludicrous.

Equally ludicrous is the suggestion that a sniper intends to reveal his position by shooting at a target. That is the last thing he intends, although he foresees it. Again, to say that the occupants of a stricken submarine intend to use up their remaining oxygen, by breathing, suggests that they are suicidal, because to say an upshot is intended implies the agent is aiming at it, trying to bring it about, as an end or a means, quite the opposite of struggling against it. Since it is quite routine for agents to struggle against upshots which they are knowingly bringing about (trying not to make a sound when returning home late, trying not to make a mess eating spaghetti) it clearly won’t do to say that all foreseen upshots are intended.

We should go further, however, and acknowledge that not all desired upshots are intended. The classic example of this is the censor who is told to read naughty books. The censor may have a desire to read them anyway: a desire he has no thought of satisfying in normal circumstances. The fact that he is subject to this temptation of the flesh does not allow us to conclude that, when ordered to read the books as part of his job, he is doing so with the intention of satisfying his impure desire. His intention may simply be to do what he is told, and carry out his job.

Indeed, this point is of the greatest importance. The Catholic Church and the Hippocratic tradition alike regard it as wrong for a doctor to kill his patients. This is a moral rule which protects the value of life in all circumstances, even though in some circumstances the end of life may reasonably be regarded as no bad thing. A doctor may legitimately be relieved when a patient dies, and even desire it. He may also, legitimately, in certain circumstances give palliative care which he knows is likely to hasten death. These observations will leave us with an insoluble tangle unless we can make use of the concept of intention: the rule against killing is understood as prohibiting the bringing about of a death by intention, and an agent does not intend what he foresees, or causes, or even desires, but what he is aiming at. In more technical language, an agent intends those upshots the expectation of which motivated him to do the action. In other words, when you ask ‘Why did you give this injection?’, the reasons (honestly) given are the intentions.[14]

To accept that there is a distinction between intention and foresight is not in itself to accept that this distinction is morally important; still less is it to accept specific moral principles which make use of it. Jonathan Bennet, no friend of the PDE, considers a claim very similar to that of Beauchamp and Childress, that agents accept the side-effects of their actions and so cannot disown them:

That remark belongs to morality rather than meaning analysis. I urge that the two be kept apart. In … maintaining the intended/foreseen distinction, we do not give moral significance to the latter. We merely ensure that those who do give it significance will get their day in court and will be allowed to speak clearly.[15]

Possible arguments, from intuition, to show that the intention/foresight distinction is not only used in common speech but actually has moral significance have already been indicated, in Section 1. In the final part of this paper I will address the puzzling question which remains: why do Beauchamp and Childress not wish to give the supporters of the PDE ‘their day in court’, as Bennett puts it?

3. The PDE as a principle of rationality

Beauchamp and Childress’ model of decision making is illustrated by an appeal to John Rawls’ notion of ‘Reflective Equilibrium’.

Equilibrium occurs after one evaluates the strengths and weaknesses of all plausible moral judgments, principles, and relevant background theories, incorporating as wide a variety of kinds and levels of legitimate beliefs as possible. To be included are beliefs about particular cases, about rules and principles, about virtue and character, about consquentialist and nonconsequentialist forms of justification, about the moral status of fetuses and experimental animals, about the role of moral sentiments, and so forth.[16]

One difficulty with this appeal is that while Rawls wanted to establish an intuitively attractive moral theory, whose principles would then establish the rules of justice for society, Beauchamp and Childress appear to be suggesting the use of reflective equilibrium to settle highly specific issues, such as a hospital’s policy on distributing organs for transplant, and even specific cases an individual doctor might encounter. Saying that, when faced with an urgent moral dilemma, an agent should go back to basics and construct a moral theory, is hardly helpful advice.

Another difficulty is more fundamental: by what criteria are we supposed to evaluate all these judgments, principles, and theories? All the possible criteria for making moral judgements have been put into the weighing scales; what criteria are left to constitute the scales themselves?

There are, of course, ways for the proponents of one moral tradition to criticise another, both by seeking to demonstrate incoherence in their opponents’ positions and by appeal to shared intuitions and moral practices. It is useless to pretend, however, that such criticisms are able to rise above both traditions and assess them from a third, objective and unobjectional, point of view.

Moral theories are at bottom theories of rationality. They tell us what features of a situation generate reasons for action, and how an assessment of those reasons should translate into action. Utilitarianism and Kantianism, for example, have fundamentally opposed ideas of how moral reasoning works, despite often coming to the same conclusion in specific cases. It is extremely difficult to see how such ‘background theories’ could be weighed up against each other in a process of reflective equilibrium. What theory of how to assess reasons for action could assess different theories of assessing reasons for action?

Beauchamp and Childress ignore this kind of problem; they are in fact most comfortable weighing up not theories but rules. They are committed to the view that clashes between rules, moral dilemmas, are inevitable; the only theoretical alterative, they suggest, is an appeal to a ‘supreme value’, an appeal which they attribute to Utilitarianism and Kantianism, and reject. Rather than this, they advise their readers to weigh things up and come to a conclusion which, they concede, may not be objectively preferable to all the others.[17]

The PDE is an alternative response to moral dilemmas: essentially it is a system in which instead of weighing conflicting considerations against each other, the most important ones can be respected. To illustrate, consider two simple dilemmas.

First, a stricken aircraft is in danger of crashing into a densely populated city. The pilot, Adam, could steer the aircraft into a thinly populated area outside the city, where he will be directly causally responsible for a small number of deaths.

 Second, a blackmailer threatens to kill two innocent people if an agent, Beatrice, refuses to kill one.

Beauchamp and Childress might suggest here that in both cases there is a dilemma caused by the clash between the rule ‘minimise harms’ (or, in their jargon, ‘beneficence’), and the rule ‘respect the sanctity of life’ (‘nonmaleficence’). The PDE, on the other hand, would allow Adam and Beatrice to act without breaking either rule. Adam can minimise harms without intending to kill; Beatrice can refuse to kill the one victim without intending to deaths of the other two.

These conclusions are certainly not uncontroversial, but this approach must be acknowledged as a theoretical possibility. If it is a theoretical possibility, however, then the method of weighing every moral consideration against every other has two problems.

First, it must be conceded that Beauchamp and Childress are not, after all, weighing up all moral considerations impartially. It is impossible to weigh up ‘weighing things up’ against ‘using the PDE’, so the PDE is simply excluded.

Secondly, weighing things up can no longer be presented as the only thing we can do in the face of apparently conflicting rules (or: the only thing we can do, having rejected the appeal to a ‘supreme principle’). If it is not the only option, then we will need an argument if we are to accept the method of weighing things up against the PDE. The PDE can, after all, claim with some plausibility to express a large part of everyday moral thinking, and it has the backing of an intellectual tradition going back at least eight centuries.

In short, even the possibility of the PDE being a viable alternative to Beauchamp and Childress’ favoured approach causes difficulties for the latter.

Conclusion

Our authors adopt an approach to moral decision-making which may at first glance appear little more than common sense: identify the morally relevant features of a situation, and the moral rules and theories of justification which pertain to them, and weigh it all up. The apparent obviousness of their approach, indeed, is the major factor in its favour. What else can you do, one might ask, when faced with conflicting considerations, and in particular when there is disagreement about moral theories?

This obviousness, however, dissipates when one recognises that they are proposing a theory of rationality which is opposed to that of any of the standard moral theories. It is still harder to sustain when one recognises that there is a well-established way of dealing with conflicting moral considerations, the Principle of Double Effect, which is arbitrarily excluded from their moral thinking.

Beauchamp and Childress may say, in their defence, that they have given arguments against the PDE. These I dealt with in the first two sections of this paper: that the distinction between the intended and the foreseen is not morally relevant, and furthermore that this distinction is not coherent. Both arguments are, in fact, vulnerable to the kinds of intuitively appealing examples which have been the staple of the academic discussion of the concept of intention for decades, not to say centuries.
 

Dr Joseph Shaw is a Lecturer at St Benet’s Hall, Oxford University
 

References

  • [1] Tom L. Beauchamp and James F. Childress Principles of Biomedical Ethics (New York and Oxford: Oxford University Press, 2009), Sixth Edition (hereafter Beauchamp and Childress (2009)). I shall also refer to the Fifth Edition (Beauchamp and Childress (2001)) and the Fourth Edition (Beauchamp and Childress (1994)).
  • [2] Beauchamp and Childress (2009) p155. Beauchamp and Childress (1994) expresses it more cautiously: ‘We will argue that these distinctions are all untenable.’ (p196).
  • [3] Beauchamp and Childress (2009) p163.
  • [4] It is an issue I address in my ‘Killing in the Catholic Tradition—I: Craniotomy’ The Downside Review Vol 123, No. 432, July 2005, pp2180-204.
  • [5] Beauchamp and Childress (2009) p163.
  • [6] Ibid.
  • [7] Ibid.
  • [8] Op. cit. p164.
  • [9] Ibid.
  • [10] See Beauchamp and Childress (1994) p209.
  • [11] Beauchamp and Childress (2009) p164.
  • [12] Op. cit. p165.
  • [13] Op. cit. p164.
  • [14] For more on the definition of intention, see Joseph Shaw ‘Intention in Ethics’ in Canadian Journal of Philosophy Vol 26 no.23, pp187-223.
  • [15] Jonathan Bennett ‘The Act Itself’ (Oxford: Clarendon Press, 1995) p203.
  • [16] Beauchamp and Childress (2009) p383. This section of the book is new to the Sixth Edition; the chapter and indeed the whole of ‘Part III: Theory and Method’, was new to the Fifth Edition. It seems to be an attempt to argue more systematically for the kind of view already expressed in Chapter 1, for example in the discussion of moral dilemmas, which I refer to below.
  • [17] Op. cit. p12.