Yasui, Yuichi --- "Reconsidering the But For Test in the Criminal Law" [2008] NZLawStuJl 9; (2008) 1 NZLSJ 543

Last Updated: 14 January 2013

RECONSIDERING THE BUT FOR TEST IN THE CRIMINAL LAW

YUICHI YASUI^∗

Introduction

The first step in any consideration of causation problems in criminal law is said to be the application of the but for test: would the consequence defined in the relevant provision have occurred but for the defendant’s conduct? Its application appears so simple that it seems little argument has been made in criminal law in New Zealand over the way it should be applied. It is true that in most cases the result would be the same irrespective of the way it is applied: the identical conclusion would be reached automatically. However, as discussed below, there may be cases the conclusion of which differs according to the way it is applied. A workable, reliable test is the one which provides an objective standard according to which the same conclusion is to be reached in relation to the same matter whoever applies it.

There are two issues which may need attention regarding the application of the but for test:

(1) What conduct is the but for test interested in?

(2) What consequence should the but for test be concerned with?

The first question is not so much one of how the but for test should be applied as of what conduct the but for test should be applied to. It is nevertheless of practical importance and thus will be discussed below. The second question is involved with how the consequence in question should be viewed, in concrete terms or in the abstract.

Besides these issues, there has been argument as to whether there is any

^∗ _{LLB, Osaka University; LLM Waseda University. Studying towards GCertLaw, Victoria}

University of Wellington. The author would like to acknowledge Rebecca Thomson for her information and advice and Takayuki Kato for his encouragement and inspiration.

need to modify the but for test or the way it is applied to accommodate some unusual, difficult circumstances. The most troublesome are the cases of alternative concurrence discussed below.

Lastly, there is a more fundamental problem; that is, is there any need to adhere to the but for test or the enquiry for a factual nexus between the defendant’s conduct and the result in question? The discussion which follows briefly surveys these problems in the context of criminal law.

A. The conduct and the consequence to which the but for test is applied

Some crimes (that is, result crimes) require as actus reus not only conduct but also a particular consequence and causation. Generally, causation in criminal law is to be considered from two distinct perspectives: whether there is a factual cause; and, if there is, whether there is a legal cause.

The first issue is determined by applying the but for test, the test to narrow down the object of consideration: it excludes from consideration the conduct which requires no further causal enquiry. Its application is usually easy, entailing no policy decision. If the consequence would not have occurred but for the conduct in question, there is factual causation between the conduct and the consequence; if the consequence would have occurred, there is not.

At this stage two questions may arise: (1) What conduct is the but for test interested in? and (2) What consequence should the but for test be concerned with?

1. What conduct is the but for test interested in?

When causation is comprised in constituent elements of actus reus, what we need to determine first is whether there is an act (or omission) which falls within the relevant actus reus; that is, whether there is an act (or omission) prohibited by the relevant provision. The next issue we

will consider is whether the consequence specified in the actus reus did occur. The problem of causation comes third for the practical purpose because causation is a link connecting the prohibited act (or omission) and the consequence, and thus these two things have to be confirmed in advance.

To take an example from Principles of Criminal Law,¹ suppose that D invites V for lunch, whose car is hit by a truck on her way to their rendezvous, the brakes of which have failed. In such a case, the criminal law will look first at the driver of the truck or its manufacturers concerned to determine whether the driver’s or manufacturers’ conduct falls within the relevant actus reus. Only after confirming the relevant consequence will causal enquiry be instituted, that is, whether there is factual causation as between the driver’s or manufacturers’ conduct and the consequence. The question of whether there is factual causation between D’s invitation and the consequence does not arise unless some evidence exists suggesting, for example, that D knew such an accident would happen and that the act of invitation constitutes the relevant actus reus. In the absence of such exceptional conditions, on those facts above, there is no prohibited act on the part of D from which causation flows which needs to be considered.²

^{Occasionally, the but for test is applied not only to the conduct falling}

within the actus reus but also to the conduct obviously not falling within the actus reus and the result of its application has been demonstrated.³ This, however, would seem confusing. Since the but for test is the one to investigate whether the conduct in question constitutes the relevant crime, there is no practical purpose in the application of the but for test to those acts or omissions which are not prohibited by the law. Causation is one constituent element of actus reus. It is a link connecting the prohibited conduct and the consequence; accordingly, no problem of causation will arise, at least for the purpose of assessing criminal liability, in the absence of some

¹ AP Simester and WJ Brookbanks, Principles of Criminal Law (3rd ed, Wellington, Brookers, 2007) 57.

² See Peter Cane, Responsibility in Law and Morality (Hart Publishing, 2002) 120. See also Jonathan Herring, Criminal Law Text, Cases, and Materials (2nd ed, Oxford University Press, 2006) 102.

³ See, for example, Simester and Brookbanks, above n1, 57, 76.

prohibited conduct.⁴

2. What consequence should the but for test be concerned with?

Another question which arises is how the but for test ought to be applied; in other words, to what consequence ought the but for test look.

There are two possible approaches. One approach, taking the consequence in general terms, applies the but for test in the following way: would a consequence defined in the relevant actus reus have occurred (probably in any way) but for the defendant’s conduct?⁵ The other, on the other hand, taking the consequence in concrete terms, applies the but for test in the following way: would the specific consequence which did occur have still occurred at the same time in the same way but for the defendant’s conduct?⁶

Again, to borrow an example from Principles of Criminal Law,⁷ suppose that D sets fire to V’s house, razing it. It turns out that there was a faulty electrical circuit in the house about to overheat and cause a similar fire. According to the first approach, factual causation cannot be established between D’s setting the fire and the destruction of V’s house. This is because V’s house would have been burned down by the faulty electrical circuit but for D’s conduct.

⁴ Glanville Williams, Textbook of Criminal Law (2nd ed, London, Stevens, 1987) states that “[s]urely the notion of but-for causation is ridiculously wide” but that when the but-for causal enquiry is made “one starts with the defendant who is charged” (p 379).

⁵ See, for example, Simester and Brookbanks, above n1, 58; and R Wright, “Causation, Responsibility, Risk, Probability, Naked Statistics and Proof” (1998) 73 Iowa LR 1022. Don Stuart, in his book Canadian Criminal Law (5th ed, Ontario, Carswell, 2007) at p 150, describes this way of application as “mechanical”.

⁶ For example, Herring, above n2, 102; Richard Card, Sir Rupert Cross and Philip Asterley Jones, Criminal Law (17th ed, Oxford University Press, 2006) 69; Michael Allen, Textbook on Criminal Law (9th ed, Oxford University Press, 2007) 34; and David Ormerod, Smith and HoganCriminal Law (10th ed, London, Butterworths, 2002) 43, who indicates this approach saying “[the but for principle] is that D’s act cannot be the

cause of an event if the event would have occurred in precisely the same way” but for the act (emphasis mine).

The second approach, on the other hand, would find factual causation at least if it can be proved that the house would not have been razed as and when it was but for D’s act. This approach has commonly been espoused by academics in Japan,⁸ whereas in New Zealand the but for test is relied upon as a first step to solve the problem of causation, though as discussed below the function of the but for test differs between the two.

The latter approach in Japan will generally go as far as to adopt the principle which does not allow one to give consideration to what did not actually happen.⁹ In the above case where D sets fire to V’s house, even if the faulty electrical circuit might have caused a similar fire, which might have razed the house in the same way at the same time,¹⁰ but for causation, according to that principle, would be established. This is because that principle does not take into account the fire which the faulty electrical circuit would have caused. It is only one hypothesis which did not actually happen. If one can consider any hypothesis event in applying the but for test, its application will become rather arbitrary¹¹ and the but for test will not function as an objective standard.

The difference between those two approaches seems to derive from their slants on the function of the but for test. Generally speaking, the idea underlying the second approach is that “[s]omething cannot be a legal cause unless it is a factual cause”.¹² In other words, the but for test tells us whether the conduct was a necessary condition of the consequence in question.¹³ It follows that if the conduct is found to

⁸ For example, Minoru Oya, Keihokogisoron (3rd ed, Tokyo, Seibundo, 2007) 221.

⁹ This is the principle of disallowance of adding any hypothetical event which did not actually happen, the principle which is commonly accepted in Japan. However, this principle needs qualifying when the but for test is applied to an omission. In the case of omission, “but for the omission in question” means “if a legal duty imposed on the defendant had been carried out”, and thus the application of the but for test involves the consideration of a hypothetical event, that is, the performance of the relevant legal duty not actually done.

¹⁰ Though this sort of proof will be extremely hard, if not impossible.

¹¹ Mitsuo Okano, “‘Joukenkankei’ Sonpinohandan” (1981) The Law School 29, 41.

¹² Herring, above n2, 102.

¹³ Cane, above n2, 120.

have no factual connection with the consequence, no question of legal causation will arise. This may result in cases contrary to common sense. Therefore, the but for test needs to be applied strictly in order not to allow wrongful conduct to slip easily through the meshes set up by that test.

By contrast, the first approach does not regard the but for test as a necessary requirement, nor does it consider that test to be concerned with factual causation. Rather, it views the test as a formula which merely expresses a certain sort of relationship between the conduct and the consequence.¹⁴ As such no problem will arise if the but for test is applied in a relaxed way. This view has an advantage in that, in applying the but for test, it will never encounter troublesome problems discussed below which may stump the other approach.¹⁵ However, if the factual relationship between the conduct and the consequence, which is to be made out by applying the but for test “is actually not a species of causation at all”¹⁶ and not a necessary requirement for showing causation, why is the but for test employed in considering a problem of causation? There seems to be no cogent reason advanced.

The purpose of causal enquiry is to determine whether the very consequence that has been brought about can properly be attributed to the defendant’s conduct, and the purpose of applying the but for test is to determine whether there is a connection as a matter of fact between that consequence and the conduct; accordingly, what needs to be brought into question should not be a consequence in general or abstract terms defined in the relevant provision but the specific consequence that has actually occurred.

B. The need for modification or substitution

1. Application to an omission

¹⁴ Simester and Brookbanks, above n1, 58.

¹⁵ This view, however, will have to deal with such problems at the next stage, that is, where enquiry for legal causation is made.

As previously mentioned,¹⁷ the application of the but for test to an omission entails consideration of a hypothetical event which did not occur. This is also the case with the second, rigorous approach above. Since “but for the omission” means “if a legal duty imposed on the defendant had been performed”, it is inevitable that the execution of the relevant legal duty not actually carried out should be hypothetically considered.

2. Application to the unlawful act of driving without a licence

Further, there seems to be at least one case in which the but for test itself needs qualifying. This may occur where the actus reus comprises conduct and those circumstances which bear on the quality of the conduct.¹⁸ The argument which has recently been made is of manslaughter. In R v Hawkins,¹⁹ for example, D caused the death of V while driving a motor vehicle while in the throes of an epileptic seizure and was charged with unlawful act manslaughter.²⁰ D had driven without a driver licence when she was subject to epileptic fits. Goddard J held that, even if D had had a licence, the accident would have occurred; that the unlawful act of driving without a licence was thus not causative of V’s death; and that D’s susceptibility to epilepsy did not transform that act into a causative factor.

The provision for unlawful act manslaughter, which is designed to prohibit strictly an unlawful act likely to do harm to others, imposes on the accused criminal liability for the death where the accused prosecuted such an act and someone’s death ensued. It condemns unlawful acts, not lawful acts, when they have brought about death. It follows that, insofar as that provision is concerned, what counts as to causation is the fact that death occurred because the conduct in question was unlawful. The aim of forbidding unlawful activities to protect human lives can well be achieved by requiring causation between death and

¹⁷ See above n9.

¹⁸ For a more detailed discussion on this matter, see Kevin Dawkins and Margaret Briggs, “Criminal Law” [2003] NZ Law Rev 570-576 (Dawkins).

¹⁹ R v Hawkins 21/2/01, Goddard J, HC Napier T18-00.

²⁰ Section 160(2)(a) Crimes Act 1961.

unlawfulness of the accused’s conduct.²¹ If so, the but for test can be modified as follows: would the consequence (that is, death) have occurred but for the unlawfulness²² of the accused’s conduct, or the fact(s) laying foundation for the unlawfulness of it? As in the case of omissions discussed above, this application of the but for test entails the consideration of hypothetical conditions which did not happen or exist.

The act of driving is lawful when the driver holds a licence. In the case of driving without a licence, therefore, the application of that modified test involves hypothetical consideration: would the victim have died if the defendant had had a licence when driving? Where, as in Hawkins, D is charged with unlawful act manslaughter when D, driving a motor vehicle while in the throes of an epileptic seizure, killed V, if the accident was brought about owing to that medical condition, it can be said that V’s death would have occurred had D been driving with a licence. Unlawfulness of D’s act of driving was causally irrelevant to the death.

Even if the but for test is applied without modification and but for causation is established, legal causation could be denied if the way of applying the test for legal causation (i.e. “operating and substantial” test²³) is adjusted so that it will be applied to unlawfulness, or those facts which render the conduct unlawful. In the above case of driving without a licence, the normal test will ask if the act of D’s driving without a licence was an “operating and substantial” cause of V’s death; whereas the adjusted test would ask if the absence of a licence while D was driving was an “operating and substantial” cause of V’s death. The answer to the former question is probably yes. This is because, as was held in Hawkins, D’s election to drive when suffering from a medical condition that could adversely affect D’s ability to drive safely may be said to be an operating and substantial cause of V’s death, but the act of D’s driving itself also made an operating and substantial contribution

²¹ Unlawfulness of the conduct in question turns on whether there are facts rendering the conduct unlawful.

²² Not the unlawful conduct itself.

²³ Herring, above n2, 103; R v Myatt [1991] 1 NZLR 674 (CA) 682-683.

to the death. What would have made no difference was whether or not D had a licence when driving. Thus, the absence of a licence cannot be regarded as an “operating and substantial” cause. It is causally immaterial, and the answer to the latter question is no.

3. Application to the case of alternative concurrence²⁴

Where simultaneous acts appear to be contributing to the result, mechanical application of the but for test may lead to a conclusion contrary to our common sense.²⁵ A classic example is as follows:

Case 1 D1 added a lethal dose of poisonous drugs to V’s cup of coffee. As it happened, D2, without conspiring with D1, also added a lethal amount of poison to the same cup. V, unknowingly having the toxic coffee, immediately

died.

Another hypothesised scenario may be compared to this:

Case 2. The same as above, except that both the poison D1 added and the poison D2 added were half the lethal dose.

Another scenario which may be contrasted with both Case 1 and Case

2, is as follows:

Case 3 The same as Case 1, except that the poison D2 added was half the lethal dose.

Case 2 is the easiest to tackle. Because both the poison D1 added and the poison D2 added were half the lethal dose, it can be said that but for adding the poison by either of them, V would not have been killed. It follows that but for causation is established both between D1’s act and V’s death and between D2’s act and V’s death.

²⁴ For a general discussion on this point, see Seiji Saito, “Iwayuru ‘Takuitsutekinakyogo’

womegutte” (1980) The Law School 24, 96-102.

²⁵ This point is often made in tort law. See for example March v E [1991] HCA 12; (1991) 171 CLR 506 (HCA), 516 (per Mason CJ).

Case 1, by contrast, has been considered to be the toughest to deal with. The mechanical, relaxed application of the but for test will surely negate but for causation both between D1’s act and V’s death and between D2’s act and V’s death. On the other hand, according to the view which looks at the particular consequence that did occur, but for causation can be established both between D1’s act and V’s death and between D2’s act and V’s death if, and only if, it is proved that the poison they added combined together and, even slightly, hastened V’s death accordingly.

More troublesome are cases in which the prosecution failed to prove beyond reasonable doubt which poison actually had the fatal effect on V or the prosecution successfully proved beyond reasonable doubt that no difference would have been made in when and how V died had only either D1 or D2 participated. When considering these cases, one needs to bear in mind the following points:

(1) The defendant ought to incur criminal liability for his or her own conduct; but

(2) The prosecution must prove causation beyond reasonable doubt. One policy reason underlying this principle is the notion that the defendant ought not to incur criminal liability for another person’s conduct unrelated to him or her.²⁶

In light of (2), there is no way but to acknowledge that but for causation cannot be found either between D1’s act and V’s death or between D2’s act and V’s death where the prosecution failed to prove which poison actually had the fatal effect on V. This conclusion must be accepted in New Zealand if one recognises that “[n]o concept is more important to the criminal law than the requirement of proof of guilt beyond reasonable doubt”.²⁷

Then what if the prosecution successfully proved beyond reasonable doubt that no difference would have been made in when and how V

²⁶ This means that there is no complicity between the defendant and the other person.

²⁷ Kevin Dawkins and Margaret Briggs, “Criminal Law” [2007] NZ Law Review 137.

died had only either D1 or D2 participated?²⁸ The same question will arise where, for example, D3 and D4, without conspiring, fired at V and the bullets happened to enter the heart of V at the same time, who was killed instantly (Case 4). If the but for test is applied to D1 and D2 individually or D3 and D4 individually, but for causation cannot be established between any of the defendants and V’s death because but for the conduct of D1 or D2, or of D3 or D4, V would have been killed at the same time in the same way.

On reflection, though, it is because D2 or D4 did an act sufficient to kill V that but for causation, by applying the but for test, cannot be established between D1’s or D3’s conduct and V’s death, and vice versa. This indicates that there can be factual causation found between the act of at least either of them in each case and V’s death.²⁹

Here lies a quandary. The point (1) mentioned previously that the defendant ought to incur criminal liability for his or her conduct would lead to the conclusion that at least D1 or D2 in Case 1 and D3 or D4 in Case 4 ought to bear responsibility for V’s death because there is factual causation between the act of either of them and V’s death.

However, there is no way to choose which one should incur the responsibility. Proposition (1) needs to be read in conjunction with the principle (2); that is, it is only when the prosecution have proved causation beyond reasonable doubt that the defendant may be held criminally liable for his or her conduct.

Although there may be factual causation between the act of either of the defendants and V’s death, the prosecution cannot prove beyond reasonable doubt that one act, not the other, is of causal relevance to the result. It follows that factual causation has to be negated both between D1’s conduct and V’s death and between D2’s conduct and V’s death in Case 1 and both between D3’s shooting and V’s death and

²⁸ For the criticism that such a phenomenon can never physically happen, see Yasushi

Ito, Mitsumasa Matsuike, Koichi Kawaguti, and Rikizo Kuzuhara, Keihokyokasho Souron

(Jo) (Tokyo, Saganoshoin, 1992) 157-158 (Mitsumasa Matsuike).

²⁹ Leaving aside the matter as to whether to call this but for causation.

D4’s shooting and V’s death in Case 4.

This conclusion, however, has been regarded as absurd, generally for the following two reasons.³⁰ Firstly, each of the defendants did a dangerous act which was sufficient to kill V. Secondly, imbalance arises between Case 1 and Case 2, where, though D1 and D2 prosecuted a less dangerous act than in Case 1, but for causation is established between D1’s poisoning the coffee and V’s death and between D2’s poisoning the coffee and V’s death.

For those reasons, it has been argued that the but for test should be modified to cope with that problem. The modified version of the but for test is as follows: where there are some acts or omissions, any one of which is sufficient to cause the particular consequence in question, but for causation is established if, but for all of them, that consequence would not have occurred.³¹ If this modified but for test is applied, since V’s death would not have occurred but for the poisoning of the coffee by D1 and D2 in Case 1, or the shooting by D3 and D4 in Case 4, but for causation can be established in relation to any one of the defendants.

However, this view has met with the sharp criticism that no theoretical ground is shown which enables all the relevant acts or omissions to be seen in toto. Unlike where the defendants are joint principals, where they are not, they acted or omitted independently, and therefore their acts or omissions cannot be viewed as a whole.³² In the above cases, D1 is not associated with D2, and thus the poisoning of the coffee by D1 and by D2 cannot be taken collectively; D3 is not associated with D4, and thus the shooting by D3 and by D4 cannot be taken collectively, with the result that there can be no but for causation established.

To the argument that this conclusion would create imbalance between

Case 1 and Case 2, the following counter-argument is possible. For the

³⁰ See for example Oya, above n8, 221.

³¹ For example, Oya, above n8, 222; and Masahide Maeda, Criminal Law: The General Part

(4th ed, Tokyo, University of Tokyo Press, 2004) 221.

³² For example, Saku Machino, Keihosoronkogian I (2nd ed, Tokyo, Shinzansha, 1998) 157.

purpose of constructive discussion, it is useful to focus on the conduct of D2. In Case 1, both D1 and D2 added a lethal dose of poison. In Case 2, it is true that D2 added only half the lethal dose of poison, but D1 also added only half the lethal dose; that is, circumstances unrelated to D2 differ between Case 1 and Case 2.³³ This difference defies meaningful comparison.

It is between the conduct of D2 in Case 1 and of D2 in Case 3 that the right balance should be achieved. Since conditions unrelated to D2 do not differ between Case 1 and Case 3, it can well be said that imbalance would arise if D2’s criminal responsibility was lighter in Case 1 than in Case 3. In Case 3, but for causation can be established between D1’s poisoning the coffee and V’s death because but for D1’s poisoning the coffee V would not have been killed. However, as between D2’s poisoning the coffee and V’s death, unless the prosecution proved beyond reasonable doubt that the poison which D1 and D2 added combined together and, even slightly, hastened V’s death accordingly, but for causation cannot be established.³⁴ In light of this conclusion, it cannot be said that there is imbalance between Case 1 and Case 3 in D2’s criminal responsibility because it is not lighter in Case 1 than in Case 3.

The problem is that no one will incur liability for V’s death in Case 1 and in Case 4³⁵ if the unmodified but for test is maintained, though all the defendants did a dangerous act which was sufficient to kill V.

To avoid this problem, it is necessary to replace the but for test³⁶ or relinquish enquiry for factual causation itself.³⁷ The former would be

³³ Atsushi Yamaguchi, Keihosoron (revised ed, Tokyo, Yuhikaku, 2005) 50 at n19.

³⁴ Though it might be argued that the but for test should be modified here as well as in Case 1 and in Case 4. However, such a modification is unnecessary in Case 3 because there is at least one defendant (ie D1) who will bear criminal liability for V’s death.

³⁵ Though all the defendants will incur liability for attempted murder.

³⁶ See, for example, Wright, above n5, 1018-1042; and Keiichi Yamanaka, Keihosoron I

(Tokyo, Seibundo, 2000) 251-253.

³⁷ Smith and Hogan, above n6, state that the principle of but for causation may have exceptions and that “[i]t seems safe to assume that [D1 and D2 in Case 1, and D3 and D4 in Case 4] will be held to have caused” the death of V (n14 at 43), but no theoretical basis is shown.

impracticable because New Zealand law appears to be more interested in argument based upon common sense than argument based upon theory. The practicable approach would thus be the latter.

Unlike civil law countries such as Japan, in New Zealand the test for legal causation is not logically premised on factual causation. The general principle in Japan is that causation in law will only be established if, on the facts based on which factual causation has been found, according to the experience of ordinary people, the consequence could be regarded as natural, or not unusual.³⁸ For example, where D3 and D4, without conspiring, fired at V’s heart and only D4’s bullet hit V, who was killed instantly, there is no factual causation between D3’s shot and V’s death, and the question of legal causation will never arise between them. No answer can in theory be provided as to whether there is legal causation where there is no factual causation proved, upon which legal causation could be established. An absurd conclusion would be reached if it was possible to make enquiry for legal causation without establishing factual causation. In the above case, since, according to the experience of ordinary people, it is not unusual for a person covered with a rifle to be killed if the offender aimed at the person’s heart and shot, legal causation could be established between D3’s shot and V’s death, if enquiry for legal causation did not have to be premised on those facts on which factual causation has been shown. Notice may need to be taken of the fact that the Japanese test for legal causation has been derived from German law and has theoretical foundations.³⁹ The question of causation arises at the stage of Tatbestand,⁴⁰ the concept

³⁸ For further details, see for example Humiaki Uchida, “Sotoingakankeisetsuniokerusotoseino

‘Handankijun’ to ‘Sotosei’” (1981) The Law School 29, 15-23.

³⁹ For a causal theory of German law, see Hart and Honoré, Causation in the law (2nd ed, Oxford, Clarendon Press, 1985) Chapter XVII.

⁴⁰ The difference between the criminal law systems of New Zealand and Germany defies the comprehensive translation of the concept Tatbestand into English. Roughly speaking, in German criminal law, determination of whether an act or omission constitutes a relevant crime involves three stage enquiries: (1) whether the act or omission falls within the definition of the offence; (2) whether the act or omission can be justified; and (3) whether the defendant can be excused. Tatbestand covers the question (1) and thus could be translated into the “definition of an offence” (see George P Fletcher, “Criminal Theory in the Twentieth Century” 2 Theoretical Inq L 265,

237). Tatbestand, however, could mean more. An offence is codified to prohibit certain conduct which should be viewed as illegal; accordingly, an act or omission which has

which has been derived from German law and is unknown to New

Zealand criminal law.

In New Zealand, enquiry for legal causation, whether or not it is based on the facts on which factual causation has been established, would lead to the identical conclusion. In the above example, since D3’s shot did not make an “operating and substantial”⁴¹ contribution to V’s death, legal causation could not be established between D3’s shot and V’s death even if the application of the but for test was skipped.

C. Relinquishment of the but for test

The test for legal causation (one standard being an “operating and substantial” test) in New Zealand can be said to cover the test for factual causation. Apart from the cases of alternative concurrence (i.e. where there are some acts or omissions any one of which is sufficient to cause the particular consequence in question), where the defendant’s conduct can be said to have made an operating and substantial contribution to the consequence in question, it can also be said that but for it the consequence would not have occurred when and as it did. The but for test is also unnecessary in those cases involving intervening events or acts in which the above test does not fit. Relevant principles such as a “reasonable foreseeability” test⁴² and a “free, deliberate and informed” test⁴³ may be said to cover enquiry for factual causation. This is because an intervening event or act which is reasonably foreseeable or not “free, deliberate and informed” may properly be

passed through the first stage enquiry and falls within the relevant definition will at least be deemed as prima facie illegal. Therefore, although it could mean more, Tatbestand may be said to denote constituent elements, mental and physical, of the offence, the satisfaction of which indicates that at least the conduct in question is prima facie illegal. This function which Tatbestand serves is reflected in the way in which causation should be examined. For a general discussion on this point, see Noriyuki Nishida, “Koseiyokennogainen” in Noriyuki Nishida and Atsushi Yamaguchi (ed), Keihonosoten, (3rd ed, Tokyo Yuhikaku, 2000 ) 14-15.

^{41 Herring, above n2, 103; R v Myatt [1991] 1 NZLR 674 (CA), 682-683.}

⁴² Simester and Brookbanks, above n1, 61; Robertson (ed), Adams on Criminal Law

(Wellington, Brookers, 1992) CA158.10 (looseleaf).

⁴³ R v Pagett [1983] EWCA Crim 1; (1983) 76 Cr App R 279, 288 (per Sir Robert Goff LJ, citing Hart and

Honoré (1st ed), above n39).

viewed as depending on a defendant’s initial conduct.⁴⁴ The dependence of the intervening event or act upon the defendant’s initial conduct demonstrates the existence of but for causation between the defendant’s conduct and the result: but for the defendant’s initial conduct, the result would not have been brought about by the intervening event or act at the same time in the same way. Therefore, except for those anomalous cases above, legal causation can rightly be considered even if enquiry for factual causation has been skipped.

Further, since the test for legal causation is not logically premised on factual causation, enquiry for factual causation can be omitted theoretically as well. There is thus no need to stick to enquiry for factual causation.

Enquiry for legal causation involves a value judgment, the question being whether the consequence can fairly be “imputable” to the defendant.⁴⁵ Such cases as Case 1 and Case 4 above could well be dealt with at the stage of legally causal enquiry which may entail evaluative considerations. Even if a factually causal enquiry has been omitted, these cases could rightly be considered. Here, the question is which of the two alternatives is more just: to hold each defendant liable for V’s death, or to hold neither of the defendants liable for the death.⁴⁶ Policy argument is more suitable at this stage about, say, what objects are to be achieved by the penal code. If, for example, one thinks criminal law exists to prevent particular consequences defined in the relevant provisions from occurring by prohibiting activities provided explicitly or implicitly in them, it would be futile, for this purpose, to blame a

⁴⁴ Simester and Brookbanks, above n1, state, in the context of an intervening act by a

third party, that “in law, [the intervention which is not “free, deliberate and informed”]

has the status of being itself a consequence of D’s wrongdoing” (at 63).

⁴⁵ Williams, above n4, 381.

⁴⁶ Simester and Brookbanks, above n1, assert that “[f]rom a moral and legal perspective, no other conclusion [than the former] is possible, even though this type of case presents difficulty for philosophical accounts of causation” (n94 at 60). However, as is shown in the text, the latter conclusion is also possible from a legal perspective (see Makoto Ida and Masao Maruyama, Case study keiho (Tokyo, Nihonhyoronsha, 1997)

85). Arguably, one could reach either conclusion by applying the “operating and substantial” test in accordance with one’s value judgment. That being the case, this test can be said to lack objectivity in the anomalous cases of alternative concurrence.

defendant for that consequence which could not have been prevented had the defendant refrained from his or her conduct. It is only where a defendant perpetrated an act and brought about a prohibited result when he or she could have prevented it by not acting as he or she did that the imposition of criminal liability for the result can be justified for that purpose. If the consequence was an unavoidable one (that is, if it would have occurred had the defendant refrained from acting as he or she did), it would not serve that purpose to ascribe it to the defendant.⁴⁷

Conclusion

The but for test is a useful means to determine whether there is factual causation between the defendant’s conduct and the consequence. To do so correctly, however, it needs to be applied to that act or omission which falls within the particular actus reus; and the consequence viewed in concrete terms rather than in the abstract.

The test for legal causation can be said to cover the but for test as properly applied. Enquiry for factual causation can theoretically be skipped as well. Further, since legal causation is not logically premised on factual causation, enquiry for factual causation can be skipped theoretically as well. Therefore, there is no need to stick to enquiry for factual causation. What counts “in the law is not whether there is a logical but for relationship between the defendant’s behaviour and the prohibited consequence, but whether there is a true causal relationship at law”.⁴⁸ The proof of but for causation need not be required as a separate requirement for the establishment of causation. This conclusion would enable the courts to escape theoretical, endless argument over the modification of, or substitution for, the but for test.

⁴⁷ Makoto Ida, “Ingakankeinoriron” (1999) Gendaikeijiho 4, 62-63. Although the author makes this argument at the stage of enquiry for factual causation, in New Zealand his reasoning is also applicable to the next stage of enquiry for legal causation where policy can be an overriding consideration.

⁴⁸ Simester and Brookbanks, above n1, 58 (emphases original).