1 [1993] AC 789.

2 Ibid, at 887. In this paper the term ‘doctor’ is taken to include any other medical practitioner engaging in the relevant conduct.

3 See, eg, Keown, J ‘Restoring Moral and Intellectual Shape to the Law after Bland’ (1997) 113 LQR at 481 Google Scholar, and, more recently, Coggon, J ‘Ignoring the moral and intellectual shape of the law’ (2007) 27 LS at 110 Google Scholar.

4 Above n 1, at 898.

5 Huxtable, R Euthanasia, Ethics and the Law: From Conflict to Compromise (Abingdon: Routledge-Cavendish, 2007)Google Scholar p 125. The fundamental premise of Huxtable's book is echoed in the title, and it is this premise that I aim to put into question in this paper.

6 Moore, M Act and Crime: The Philosophy of Action and its Implication for Criminal Law (Oxford: Oxford University Press, 1993)Google Scholar p 27.

7 Ashworth, A Principles of Criminal Law (Oxford: Oxford University Press, 4th edn, 2003)Google Scholar p 113.

8 Huxtable, above n 5, p 117.

9 Ibid, p 117.

10 This was the evidence given by Professor Sheila McLean in support of Lord Joffe's Bill in Assisted Dying for the Terminally Ill (HL) 2005: Evidence, at 14.

11 Huxtable, above n 5, p 120 (emphasis removed).

12 Ashworth, A ‘the scope of liability for omissions’ (1989) 105 LQR 424 Google Scholar at 437.

13 Ibid, at 437.

14 Magnusson, R ‘the future of the euthanasia debate in Australia’ (1996) 20(4) Melb Univ L Rev 1108 Google Scholar at 1118.

15 There have been other attempts to defend the acts/omission distinction. See, eg, Stauch, M ‘Causal authorship and the equality principle: a defence of the acts/omissions distinction in euthanasia’ (2000) 26 J Med Ethics 254 CrossRefGoogle Scholar; McLachlan, H ‘the ethics of killing and letting die: active and passive euthanasia’ (2008) 34 J Med Ethics 636 CrossRefGoogle Scholar. However, Stauch's defence focuses exclusively on its moral relevance and introduces ethical notions such as his equality principle which, in my view, arguably only cloud the distinction in more controversy. Further, as the title of his paper implies, he seems to accept that the distinction is merely between different kinds of euthanasia, whereas I claim here that omissions, when lawful, are not euthanasia at all. What is required, to defend the distinction, is simply a clarification of the concept of omissions and its relationship to causality and the existence of accepted moral and legal duties. The question of the moral relevance of the distinction between acts and omissions will then take care of itself. McLachlan's defence consists of denying that ‘omissions can be causes’. This is mistaken, and his defence will be examined at various points in this paper.

16 This objection has been developed by D Brock ‘Voluntary active euthanasia’ in Beauchamp, Tl and Walters, L (eds) Contemporary Issues in Bioethics (Wadsworth Publishing, 5th edn, 1999)Google Scholar pp 296–305 and Huxtable, above n 5, p 120.

17 Above n 1, at 895. Lord Mustill's doubts about the rule are echoed by Thomas J in Auckland Area Health Board v Attorney-General [1993] 1 NZLR 235 at 249, where he notes: ‘... framing the issue in these terms begs the question: the discontinuance of the ventilator is the cause of death if it is not legally acceptable but not the cause of death if it is!’.

18 The killing/letting die distinction is not, of course, the same as the act/omission distinction. First, there are many instances of the latter that do not involve any killing or letting die at all – eg where I forget (omit) to put the food back in the fridge. Secondly, even where killing and letting die are concerned, it is possible to argue that one may let die not only by omission, but by certain acts, though for reasons we shall see, the law will classify the broader conduct, which may involve acts, as an omission. Third, as we shall see, it is possible to kill by omission, and not merely by act. However, in many cases, the two distinctions do map onto one another: when withholding treatment, thereby letting a patient die, I let the patient die by omission. Further, if the broader classification of omission adopted by the courts is accepted, the two distinctions effectively map onto one another, at least where lawful conduct is concerned. Therefore, it is important not to make too much of this qualification.

19 Brock, above n 16, p 298.

20 Bertha Manninen claims that it is an excellent illustration of the difficulty in the law in this area. See Manninen, B ‘a case for justified non-voluntary active euthanasia: exploring the ethics of the Groningen Protocol’ (2006) 32 J Med Ethics 643 CrossRefGoogle Scholar at 646–647.

21 Nothing in this view is taken to imply, of course, that there are no contexts in which the doctor is a layperson – where, for instance, the doctor is not acting within the scope of his or her employment, as that question might need to be determined by a court in certain circumstances, or is otherwise acting unlawfully. The objection advanced by Brock and others to the distinction between the greedy son/ interloper cases and the doctor cases is that they each ‘do the same thing’even when the doctor is acting lawfully , and it is this claim that I deny.

22 Above n 1, at 866. For further analysis of Lord Goff's suggestion, see section 2(c).

23 Ibid, at 895.

24 Some critics have denied that omissions are causes, or have consequences. McLachlan defends such a view, criticised by Gerrard, E and Wilkinson, S in their paper ‘Passive euthanasia’ (2005) 31 J Med Ethics 64 Google Scholar, in McLachlan, above n 15, at 637. Michael Moore has also defended this view in his recent Causation and Responsibility (Oxford: Oxford University Press, 2009)Google ScholarPubMed, p 444f. The mistake in such a view consists of correctly noticing that omissions are not causes in the sense in which acts are, but then incorrectly concluding that it makes no sense to describe them as causes at all.

25 Rachels, J ‘Killing and letting die’ in Encyclopaedia of Ethics vol 2 (London: Routledge, 2nd edn, 2001)Google Scholar p 947 at p 949.

26 See also Jonathan Glover, who refers to an omission as ‘a positive act of withholding the medicine’ in Causing Death and Saving Lives (London: Penguin, 1977)Google Scholar ch 3. James Rachels' view was first presented in his influential ‘Active and passive euthanasia’ (1975) 292 New England Journal of Medicine 78. John Harris shares the same view: ‘it is natural to think of an omission as something that the man does’ Violence and Responsibility (London: Routledge and Kegan Paul, 1980)Google ScholarPubMed p 29. This view originates with Bentham. But a failure to act is no more a kind of act than is doubt a kind of certainty, or ignorance a kind of knowledge.

27 The absurdity of such a response is made perspicuous if we add the words ‘namely nothing’ after the question, as follows: ‘why did you stand by and do something, namely nothing?’.

28 The following discussion of our concept of causation is indebted to the account of Hacker, Peter in Human Nature: The Categorial Framework (Oxford: Blackwell Publishing, 2007)CrossRefGoogle Scholar ch 3.

29 Lowe, S ‘the right to refuse treatment is not a right to be killed’ (1997) 23 J Med Ethics 154 CrossRefGoogle ScholarPubMed at 156.

30 Hacker, above n 28, pp 86–87.

31 Rundle, B Why there is Something Rather than Nothing (Oxford: Oxford University Press, 2004)CrossRefGoogle Scholar p 73.

32 See Hart, Hla and Honoré, A Causation in the Law (Oxford: Clarenden Press, 1959)Google Scholar p 32. It has been objected that to pick out a condition as a cause in this way is to confuse two separate issues: what the cause is, with what strikes us to be the cause ( Kuhse, H The Sanctity of Life Doctrine in Medicine: A Critique (Oxford: Oxford University Press, 1987)Google Scholar p 56). This is mistaken. In singling out the salient condition, we are not committed to the claim that it is the cause because it strikes us as such. Rather, we are claiming that, of all the conditions that exist, it was the height of the building that made the difference – it was not, for instance, any vibrations caused by passing traffic or bad weather. Those conditions also exist, but they exist in other cases where other buildings are concerned but do not cause those other buildings to topple and, for that reason, are not the cause of the collapse in this case.

33 Hart and Honoré, above n 32, p 15.

34 Ibid, p 15.

35 Hugh McLachlan is therefore mistaken in denying that omissions can genuinely be causative. At one point in his paper, the awkwardness of this extreme position becomes apparent, when he is forced to concede that ‘omissions can indirectly lead to deaths’ (above n 15, p 637). McLachlan does not explain how omissions can lead to, but not cause, deaths. For detailed and incisive criticism of McLachlan's position, see Coggon, J ‘on acts, omissions, and responsibility’ (2008) 34 J Med Ethics 576 CrossRefGoogle Scholar. For McLachlan's response, see his ‘To kill is not the same as to let die: a reply to Coggon’ (2009) 35 J Med Ethics 456 CrossRefGoogle ScholarPubMed.

36 He claims that they are best characterised as ‘the absence of preventative anti-causes’, of which there might be a ‘limitless’ amount. See McLachlan, above n 15, at 637. However, for reasons we will see, it is not necessary to endorse this obscure and cumbersome characterisation.

37 Ibid, at 637. McLachlan should, in my opinion, have engaged with the work of Hart and Honoré on these issues. We will see below that, although Hart and Honoré's account of causation seems to be out of favour, common criticisms of their account are not cogent.

38 This example is mine, not McLachlan's.

39 It is worth noting that there are limited circumstances where an omission has explanatory value even though there is no expectation that an agent would have acted in those circumstances. Thus, the guest of an important host might refrain from advising the host that the host has left the milk out of the fridge after making the tea. Here, perhaps because the guest does not want to cause embarrassment to the host or adversely affect what may be an important meeting, the guest may refrain from saying anything, and we can say that the reason the milk turned sour is that the host left it out of the fridge and the guest did not want to say anything for fear of embarrassing the host. But this example does not contradict the account given here, for the countervailing circumstances of not wanting to embarrass the host offsets the expectation that the agent would otherwise act and in that sense is logically tied up with it, as an ‘excuse’.

40 Hart and Honoré do not explicitly connect this notion of cause as used to refer to the absence of a condition resulting in a consequence with the use of ‘reason’in this context as I have done.

41 See Kuhse, above n 32, pp 56–58.

42 Notice that it follows from this account that omissions can kill. At time t1 , for the reasons given, the omission would have been the relevant explanatory factor and, from the point of view of assessing liability, would therefore amount to ‘killing’. It is therefore quite mistaken to attempt to defend the moral relevance of the act/omission distinction by arguing that acts can be causes but that omissions cannot be, as McLachlan does.

43 It is doubtless true that, in the context in which life is already being artificially prolonged by life-prolonging measures, it would have explanatory force in some contexts to say that the reason the patient died is that the doctor decided no longer to provide any life-sustaining measures, but in the context of seeking explanations of legal liability, where the conduct is lawful, such a statement would have no explanatory value, for it does not distinguish this case from other cases where the measures are lawfully withheld.

44 A related criticism of Hart and Honoré, endorsed by Kuhse, has been made by John Harris. Harris claims that Hart and Honoré's account is tied excessively to the notions of standard practice and normal conditions, and so cannot explain causation in contexts where what is standard or normal itself has disastrous consequences. Harris gives the following example. Every year, just like clock-work, the poor and jobless, aged and infirm, suffer terribly and many of them die. What is the cause? The myopic view is that they die because they are poor and jobless, aged and infirm. This distinguishes them from those who do not suffer or do not die. But the World Moral Authority may identify neglect of other members of society or of the government as the cause, and other features as mere conditions. And ‘surely the World Moral Authority's causal explanation is not upset by the discovery that this society normally neglects its weakest members’: Harris, above n 26, p 41. Harris' criticism is, however, based on a misunderstanding. The relevant inquiry, on the Hart and Honoré account, is: is there a society where the same neglect of the poor, jobless and infirm occurs, but where the poor, jobless and infirm do not die? If there is, then the neglect would be common to both those societies and would, in that event, be a mere condition, rather than a cause. However, if there is not, then the neglect can be seen as more than a mere condition – it is genuinely causal. Contrary to Harris' view, the World Moral Authority's explanation is accommodated by the Hart and Honoré account. Hart and Honoré are not committed to the view that if something standardly, normally, or always happens, it cannot be a cause. Clearly, catching the 'flu virus standardly or normally causes one to develop the symptoms of 'flu. Rather, they are merely claiming that, in some contexts, it is necessary to look at how the particular case contrasts with what normally happens in order to identify the cause – normally the signalman lifts the lever and the train does not crash. This time, however, he failed to lift the lever and the train crashed.

45 R v Gibbins and Proctor (1918) 13 Cr App R 134.

46 Above n 1, at 893.

47 Above n 45.

48 For this reason, Coggon, in his incisive criticism of McLachlan's paper (above n 35), is nonetheless mistaken to claim that the ‘question is about responsibility, not about omissions’ (at 578). This begs the question of causation: we do not normally hold people responsible for things they have not caused. True, as argued here, whether an omission is causative will depend on whether there is an expectation that a person will act and on whether they failed to act, thus leading to a consequence that would not have eventuated had they acted. But that does not mean that the question is not about omissions at all – the two issues of responsibility and omission are logically interconnected. An analysis of what ‘cause’ means and its relationship to responsibility therefore cannot be bypassed, and I have attempted to provide that analysis here.

49 Above n 45.

50 See the remark by Thomas J from Auckland Area Health Board v Attorney-General [1993] 1 NZLR 235, at 249, cited above at 17.

51 This example is from Hart and Honoré.

52 I discuss the apparent inconsistency of regarding the conduct an omission in the case of a doctor but an act in the case of an interloper in the following section.

53 Lowe, above n 29, at 156.

54 Some think, of course, that withholding is nevertheless a species of euthanasia, but not on the basis that withholding is an act (though, as we saw above, some critics like Rachels and Glover may mistakenly construe omissions as negative acts).

55 I say ‘arguably’ here because, for reasons we shall come to in section 2(d), the question of causation could be decided independently of the act/omission distinction.

56 The Law Lords of course took themselves only to be making a declaration of lawfulness in the circumstances before them of Anthony Bland, and not to be changing the law – fresh declarations being required in future cases. However, this does not affect the point being made here, for in order to declare the conduct lawful on that one occasion, the concept of omission was extended, for reasons we shall come to. Furthermore, the claim that the case simply declared the law as it currently stood and did not set a precedent is arguably naïve, as subsequent examples of withdrawal where declarations have not been sought, in spite of the ruling, have shown.

57 See, in particular, Ashworth, who recognises that ‘although there are clear cases of omission and some clear cases of act, there are many ambiguous cases’, but who concluded that those ambiguous cases were being used as a cloak for avoiding the moral issues (Ashworth, above n 7, p 113). One way of dealing with the moral issues is precisely to clarify the law and remove the ambiguity at law. Where the case is genuinely borderline and reasonable minds can disagree, a decision may need to be made, and in that case it is more correct to say that shirking the need to make the decision is avoiding the moral (and legal) issues.

58 Above n 1, at 866 (emphasis added).

59 Ibid, at 866.

60 See, eg, Price, D ‘Euthanasia, pain relief and double effect’ (1992) 17 LS 323 Google Scholar at 340; Huxtable, above n 5, pp 119–120.

61 Above n 1, at 875.

62 Ibid, at 881.

63 Ibid, at 881 (emphasis added).

64 Ibid, at 882 (emphasis added). It may seem that Lord Browne-Wilkinson is seeking only to deny that the conduct is positive , not that it is an act. However, he is actually seeking to deny that it is an act, as the following wording shows: ‘even if, contrary to my view , the removal of the tube is to be classified as an act ...’: ibid, at 882 (emphasis added).

65 Ibid, at 882.

66 Ashworth, above n 12, at 437.

67 Ibid, at 437. See also Huxtable, above n 5, p 140, who cites this passage in support of his own argument for the same conclusion.

68 Moore, above n 6, p 27.

69 This is so notwithstanding that their Lordships considered themselves merely to be declaring the lawfulness of the conduct before them rather than changing the law. On this point, see above n 56.

70 James Rachels' arguments for its moral irrelevance stem, as we have seen, from this error. It is in fact a common mistake for critics to concede that a logical distinction can be drawn, but dismiss that logical distinction as morally irrelevant. A logical distinction can also be drawn between murder and manslaughter, but it is precisely the considerations that ground the logical distinction that are also relevant morally.

71 For more on the moral relevance of the distinction, see McGee, A ‘Finding a way through the ethical and legal maze: withdrawal of medical treatment and euthanasia’ (2005) 13(3) Med LR 357 Google Scholar at 382–383.

72 Poultry Act 1968 (NZ), s 2. The example comes from Pearce, Dc and Geddes, Rs Statutory Interpretation in Australia (Butterworths Australia, 5th edn, 2001)Google Scholar p 116.

73 I stop short of asserting that ‘deeming’cannot be used to describe what the Law Lords were doing, for of course the word ‘deeming’ is itself flexible. I am merely recommending that it should not be used for the reasons given, for it is liable to invite misplaced objections to the decision.

74 As we shall see, however, the decision could have been justified even if the conduct was classified as an act. The point to note here, though, is that classifying it as an act does not remove the arguments for classifying it as an omission. The arguments for both conclusions were cogent, which is why a conceptual decision had to be made.

75 Above n 1, at 866.

76 Thus, explanatory causes are a broader genus of which agent causes and non-agent causes are species. It is certainly true that an agent cause is causative regardless of whether it is explanatorily significant, whereas an omission is, I have argued, only causative if it is explanatorily significant. However agent causes, too, are only causes at law if they are explanatorily significant for the purposes of establishing legal liability. And omissions only have causal consequences where the conduct omitted is expected to occur, so that its non-occurrence is a reason for why a consequence that would not normally eventuate actually eventuated on this occasion (hence, the food went stale because he forgot to put it back in the fridge, etc).

77 McGee, above n 71.