I. INTRODUCTION
Accessory liability “raises some of the most difficult questions that can be asked … and can test almost to destruction our theories and concepts by which we hold an individual accountable in the relatively more simple case where he or she is acting in isolation”.Footnote 1 This comment was made in relation to criminal law, but the matter is no more straightforward in civil law. Its importance is obvious.Footnote 2
The common roots of both criminal and civil accessory liability can be found in the common law, but the two sets of rules have since diverged. Whereas in the criminal sphere liability can be imposed upon someone who “aids, abets, counsels or procures”, the civil law tends to limit liability to the last of these conduct elements: procurement.Footnote 3 This is obviously more restrictive. But should the civil law be so limited? Is it satisfactory that someone who knowingly assists a tort should escape liability?
This article aims to examine whether a divergence between the criminal law and law of tort is justifiable. In order to achieve this, the shape of the criminal law must first be outlined. The law of tort can then be considered. Torts which are also crimes merit particular attention: it currently seems that a person who is criminally liable as an accessory may be beyond the reach of the civil law. This state of affairs will be criticised; it will be suggested that tort law should impose civil liability upon those who knowingly assist the commission of a tort which is also a crime. For torts which are not also crimes, the case for imposing liability upon assisters is perhaps less strong. However, it will be suggested that even here assistance liability may be appropriate.Footnote 4
II. CRIMINAL LAW
Accessory liability in the criminal law is notoriously difficult. Only a brief outline of the law can be given here, but this should nevertheless be useful: assistance liability is well-established, in contrast to the situation in private law. Lessons learned by the criminal law may benefit the civil law; after all, although the conduct elements of “aid, abet, counsel or procure” are set out in section 8 of the Accessories and Abettors Act 1861, this statute simply embodies common law principles of even longer standing. The origins of accessory liability are the same across the criminal/civil law divide.Footnote 5
The four conduct elements in the 1861 Act have been said to be distinct and deserve to be given their natural meaning,Footnote 6 but the Law Commission has convincingly concluded that the conduct element for accessory liability is essentially assisting or encouraging the commission of the principal crime.Footnote 7 Although the Law Commission treated these two conduct elements as separate, the distinction between the two can be very fine: the emboldening of the principal offender assists the commission of the crime.
Assistance has been given a very wide meaning in the criminal law. It includes returning a jemmy to a known burglar who then uses it to commit a burglary,Footnote 8 and selling a baseball bat to a person who goes on to assault another with that bat.Footnote 9 The question of whether there has been assistance appears to be entirely a question of fact. It is difficult to establish a strong causal link between the principal crime and the assistance of an accessory; the notion that the assistance – or even, for that matter, the procurement – of an accessory has caused an autonomous being of full capacity, who can choose what he does and does not do, to act is obviously problematic.Footnote 10 As a result, it has been suggested that simply assisting a crime is insufficient to affix the accessory with responsibility for that crime.Footnote 11 Such contentions have been rightly rejected.
It is suggested that causation does play some role when establishing accessory liability. As Smith has put it:
It has always been implied in the concept of complicity that an accessory's involvement … did make some difference to the outcome, and, as a consequence of this, accessories have been implicitly linked to the harm element in the principal offence. No other plausible explanation exists for complicity's tenacious, fundamental requirement of the commission of the principal offence. … In sum then, under existing doctrine, the accessory's liability is derivative or parasitic of a principal offence and its harm content.Footnote 12
Such reasoning is persuasive. For an accessory to bear responsibility for the principal crime, it seems reasonable to demand that what the accessory did had some effect on the principal crime, even if the effect be simply on the precise way, or manner, or timing, of the commission of that crime.Footnote 13 However, if responsibility for the primary crime is to be attributed to the accessory then it would seem sensible to require that his acts have more than a minimal effect. Such a de minimis principle is well recognised in other branches of the law,Footnote 14 and Williams thought that as “a matter of common sense a person who gives very minor assistance ought not to be held liable as an accessory”.Footnote 15 Nevertheless, in Giannetto the Court of Appeal did not criticise a judge who thought that a person could still be found guilty as an accessory if all he did, upon being told by the principal of the principal's plan to commit a crime, was to pat the principal on the back and say “Oh, goody!”Footnote 16 It is submitted that such an approach is inappropriate and should not be followed: some causal connection – not necessarily great, but certainly beyond the de minimis threshold – should be found. This helps to mitigate the strength of the objection that accessory liability is an exception to the principle that the voluntary act of a person can break the chain of causation from a prior actor to the ultimate wrong.Footnote 17
The above quotation from SmithFootnote 18 raises another issue of fundamental importance: accessory liability in this context is not inchoate. Accessory liability requires a principal offence, and the liability of the accessory is generally for the same crime as the principal offender. The Law Commission explained that the main reasons for this were forensic considerations,Footnote 19 public acceptability,Footnote 20 culpability, condemnation and labelling,Footnote 21 and the connection between the accessory's conduct and the principal offence.Footnote 22 These reasons highlight the fact that an accessory who assists the principal crime is held to bear responsibility for that crime.Footnote 23
The conduct element for accessory liability is widely drawn, but the necessary mental element is narrow. Given the fact that the accessory is further removed from the principal crime than the primary party, the accessory must evince a high degree of culpability in order to be found guilty of the principal crime. What exactly that mental element is – or should be – is complicated and constitutes a matter of some debate.Footnote 24 The leading judgment on this point appears to be Johnson v. Youden, in which Lord Goddard C.J. found that:
Before a person can be convicted of aiding and abetting the commission of an offence he must at least know the essential matters which constitute that offence.Footnote 25
The “essential matters” to be envisaged include the conduct element and fault element of the primary wrong.Footnote 26 However, it need not be the purpose of the accessory that the primary crime be committed. As Devlin J. remarked in National Coal Board v. Gamble:
If one man deliberately sells to another a gun to be used for murdering a third, he may be indifferent whether the third man lives or dies, and interested only in the cash profit to be made out of the sale, but he can still be an aider and abettor.Footnote 27
It seems that an accessory will be found guilty if he knowingly encouraged or assisted the commission of the principal crime.Footnote 28
The final lesson of the criminal law that is relevant for the present discussion of the private law is the doctrine of “innocent agency”. This doctrine is important where the “principal” party is not convicted of an offence but the “assister” is still found guilty for the acts carried out by the “principal”. For example, an adult might give a child under ten years old a loaded gun, knowing that the child would like to shoot and cause serious harm to his teacher. If the child does shoot and kill his teacher, he cannot be found guilty of murder because he is not yet ten years old.Footnote 29 “Innocent agency”, which deems the child to “drop out of the picture”, might apply: the criminal act can be found to have been effectively carried out by the adult, who had the appropriate mental element for murder. This seems satisfactory.Footnote 30 However, it is important to note that in such an instance the adult is not guilty as an accessory, but rather as the principal.
III. TORT
The outline of the criminal law given above suggests a rational shape to accessory liability: a widely-defined conduct element,Footnote 31 coupled with a very narrow mental element. This may well be appropriate in the civil law also; a person who knowingly assists the commission of a civil wrong may be considered to be acting culpably. Such conduct should perhaps be deterred,Footnote 32 in order both to prevent undesirable conduct and to protect the rights of the victim of the primary wrong. The fact that there exists a causal link between the acts of the accessory and the primary wrong helps to explain why liability might be imposed: the accessory bears some responsibility for the violation of the victim's rights. The narrow mental element, which requires the accessory actually to know of the essential elements of the offence he is assisting, means that the scope of liability is not so broad that it catches innocent parties. A further safeguard could be provided by the defence of justification.Footnote 33 However, this does not describe the current state of the law of tort, which is best understood by examining the principles of joint tortfeasance.
A. Joint tortfeasance
1. Participation links
The leading exposition of the law on joint tortfeasance can be found in Carty's article, “Joint tortfeasance and assistance liability”.Footnote 34 She identified three relevant participation links: procurement, authorisation and combination. The final link of “combination” still seems to derive from The Koursk. This case actually held that the two tortfeasors were several concurrent tortfeasors rather than joint tortfeasors, in order to restrict the scope of joint and several liability, but the following quotation of Scrutton L.J. has had great effect:
I am of the opinion that the definition in Clerk and Lindsell on Torts, 7th ed., p.59, is much nearer the correct view: “Persons are said to be joint tortfeasors when their respective shares in the commission of the tort are done in furtherance of a common design.... but mere similarity of design on the part of independent actors, causing independent damage, is not enough; there must be concerted action to a common end.”Footnote 35
The need for there to be a common design has been prominent in many of the recent cases dealing with accessory liability. It has been used to support the idea that simply assisting the commission of a tort – without any aspect of combination – does not fall within the scope of joint tortfeasance as it is currently understood.Footnote 36 The second participation link of “authorisation” involves permission, or possibly an order, from a person having, or purporting to have, authority over the “immediate” wrongdoer.Footnote 37
It is the first participation link, however, which greatly restricts the scope of joint tortfeasance.Footnote 38 “Procurement”, according to Lord Templeman in CBS Songs v. Amstrad Consumer Electronics plc, means “inducement, incitement or persuasion”, such that advice and assistance are excluded. However, it is submitted that limiting this participation link to “procurement” rather than “assistance” is not beyond criticism.
It is important to understand the difference between the two potential conduct elements for accessory liability in this area. “Assistance” should be considered to be of a broad nature; it covers all acts which help, in any way, the primary wrongdoer to breach his duty. The scope of assistance should include facilitating the primary wrong.Footnote 39 “Procurement”, or “inducement”, by contrast, is much narrower. Hart and Honoré have suggested that an inducement can “make a given course of action more eligible or desirable in the eyes of the other than it would otherwise have been, or seem more eligible or desirable than it really is”.Footnote 40 It might be added that procurement tends to give another person an idea he would not otherwise have had.
Procuring has the advantages of ensuring a close link with the primary wrong and restricting the scope of liability. Assistance would widen the scope of liability. However, this would obviously be more consistent with the criminal law; the desirability of this is considered below.Footnote 41 Moreover, recognising assistance liability would afford better protection to private law rights.Footnote 42 Although it should not be assumed that all rights are protected to the same extent – tort law provides a remedy more readily for the infringement of property rights than for “pure” economic loss, for exampleFootnote 43 – it is nevertheless difficult to explain why, if a right is worthy of protection against an “immediate” wrongdoer, and a person who induces that wrong, it should not be similarly protected against a knowing assister as well. After all, the assister does bear some responsibility for the primary wrong: the need for a causal link should not be ignored.
Admittedly, the “connection” between assistance and a primary wrong appears to be less strong than that between inducement and a primary wrong, but this should not be overplayed: in both instances the primary wrongdoer is an autonomous party who is responsible for the acts he committed.Footnote 44 Whether the primary wrong was procured or assisted, ultimately the decision whether or not to commit the wrong lay with the primary wrongdoer himself.Footnote 45 This means that any “connection” between the ultimate wrong and the accessory is necessarily attenuated, but the fact that the acts of the accessory did have some causal effect upon the primary wrong means that he should bear some responsibility for that wrong. It is submitted that distinguishing between procurement and assistance is far from straightforward;Footnote 46 indeed, it may be possible to view inducement and assistance as simply points on a spectrum of conduct. A recent Scottish case has expanded the definition of “inducement” in the context of the tort of inducing a breach of contract, by stating that “A must induce B to break his contract with C by persuading, encouraging or assisting him to do so”.Footnote 47 The distinction between procurement and assistance may be very fine. In some instances, the relevant assistance may be crucial to the primary wrong being committed.Footnote 48 It is therefore suggested that assistance represents an entirely rational conduct element for accessory liability in tort. A wide conduct element should not be feared; there must still be some causal link between the assistance and the primary wrong,Footnote 49 and the scope of liability can be restricted through a narrow mental element.Footnote 50
2. Culpability
In order to justify liability the accessory must be culpable. It is not only the desire to protect rights which underpins accessory liability, but also a wish to ensure that culpable parties bear some responsibility for the consequences of their conduct. Thus the victim of a tort cannot just “shop around” in order to sue any third party who has somehow assisted the breach, but should only be successful against a person who knew that the assistance rendered would assist the primary tort. This narrow mental element is crucial in limiting the scope of accessory liability; negligent conduct cannot suffice.Footnote 51 By knowingly assisting the primary wrong the accessory has acted culpably; such culpability should prevent the assister from complaining if he is later made to compensate the victim of the wrong.Footnote 52 Although English law does not impose tortious liability for acts which are done with the intention to harm unless there has been an unlawful act,Footnote 53 this in no way prevents a wide conduct element for accessory liability. After all, in cases of accessory liability a victim's rights have been violated, and there is a connection between the acts of the accessory and the primary wrong. Tort law should not shy away from imposing liability upon undesirable behaviour where a person's rights have been infringed.Footnote 54 The combination of the violation of a right, responsibility through a causal link, and the culpability of the assister makes liability appropriate.
Moreover, imposing liability for assistance seems perfectly appropriate in order to discourage parties from knowingly assisting a primary wrong.Footnote 55 Deterrence is particularly relevant in the area here under consideration, as the conduct of the accessory is calculated ex ante.Footnote 56 Since a conscious decision is taken by the accessory, there is clearly the opportunity to deter the accessory from becoming “involved” in the primary breach. Cane has observed that “as a generalisation it is probably true to say that the deterrent theory of tort liability is most likely to be relevant to torts involving liability for calculated conduct”.Footnote 57 It is suggested that deterring parties from knowingly assisting the commission of primary wrongs would be a welcome development of the law. Deterrence is closely linked not only to preventing culpable conduct, but also to protecting rights. By deterring knowing assistance, the primary rights at stake are better protected.
If tortious liability were to be imposed upon a party who knowingly assists the commission of a wrong, the current restrictive requirements of joint tortfeasance would clearly be expanded. This would represent a departure from the current state of the law. In Sabaf SpA v. MFI Furniture Centres Ltd., Peter Gibson L.J. remarked that “[t]he underlying concept for joint tortfeasance must be that the joint tortfeasor has been so involved in the commission of the tort as to make himself liable for the tort. Unless he has made the infringing act his own, he has not himself committed the tort”.Footnote 58 It may be that by knowingly assisting the wrong a party does not make the tort his own in the same way as an inducer.Footnote 59 But it must be remembered that the mental element required for accessory liability is narrow, and often narrower than that required for the primary wrong.Footnote 60 As a result, the accessory may be more culpable than the primary wrongdoer, even if the conduct element for liability is recognised to be “only” assistance rather than inducement.Footnote 61 Liability is imposed upon an accessory as a result of the accessory's own, independent acts; liability should not be incurred without showing some responsibility for the primary wrong.Footnote 62 Accessory liability is parasitic, but it is imposed for reasons of protecting rights, culpability and deterrence.
B. Consistency with criminal responsibility
That the law of tort fails to impose liability upon an assister may sit uncomfortably with the criminal law on accessories in situations where the crime in question is also a civil wrong. Consider the case of a man who sells or lends his knife to a friend, knowing that it will be used to attack another, and that friend then goes on to stab another person to death with that knife. It is clear that on such facts the man may be found guilty as an accessory to the principal crime of murder. However, recent cases suggest that the man would not be liable to the victim in tort. Before considering such case law, it is helpful to consider the issue of principle: should a person who assists the commission of a crime, which is also a tort, ever be liable for that tort (where the accessory and principal are not party to a common design to commit a wrong)?
It is submitted that it should be possible to find an assister of a primary offence liable for that primary wrong. It seems remarkable that, although the criminal law may find that the assister bears responsibility for the murder, the civil law might not. This smacks of inconsistency, something which the law should clearly try to avoid.Footnote 63
One reason for such inconsistency may be the misconceived notion that the criminal law developed separately from the civil law.Footnote 64 Yet accessory liability finds its roots in the common law, and the same principles were originally applied to both criminal and civil cases. A close reading of Lumley v. Gye, for example, emphasises the link judges previously made between the civil and criminal law, and it is unfortunate that such a link appears to have been lost.Footnote 65
But it is not simply for historical reasons that assisting a crime should be recognised as potentially tortious. For example, imagine that, in a modified version of the example given above, a man lends a knife to his friend, and the friend goes on to stab someone, but the victim does not die. Instead, the victim is severely wounded and has to have his right arm amputated. The victim might give evidence at a lengthy and harrowing criminal trial, with the result that both the accessory and principal are deemed to bear responsibility for the offence committed and the resulting injuries of the victim. It seems entirely justifiable for the victim to expect to recover damages from both parties. Why should this not be the case?
Perhaps one reason can be found in the different functions of the civil and criminal law. The criminal law is primarily concerned with guilty defendants, whereas the law of tort is more victim-focussed and concerned with adequately compensating wrongs. But, if this is true, why is the civil law more restrictive than the criminal law? Surely one would expect it to be more difficult to prove criminal responsibility than civil responsibility, which is emphasised by the fact that criminal punishments are generally more severe. Civil liability can be seen as less intrusive since it is not (generally) concerned with punishment but simply compensating a wronged victim. Moreover, criminal liability engenders a stigma of much greater degree than that related to civil liability. It is submitted that once the hurdle of criminal liability has been cleared, it should be natural for civil liability to follow; after all, a crime has to be proved beyond reasonable doubt, whereas a tort only needs to be proved on the balance of probabilities.
The reticence of the courts to embrace assistance liability may be explained by the fact that the accessory could be jointly and severally liable with the primary wrongdoer.Footnote 66 In a situation where the primary wrongdoer has no money, but the accessory is extremely rich, the victim would be able to sue the accessory for the full extent of his injuries, rather than just the proportionate amount which could be attributed to the wrong of the accessory. This might seem somewhat unfair, particularly since the responsibility of the assister may be much less than that of the primary party. Indeed, to combat such perceived unfairness an increasing number of jurisdictions have replaced joint and several liability with a system of proportionate liability.Footnote 67 However, English law strongly favours joint and several liability. Given that an accessory who compensates a victim is able to claim contribution from the primary wrongdoer, the distinction between proportionate liability and joint and several liability is most pronounced when the primary party is insolvent and a claim in contribution would be worthless. It is suggested that there are good reasons why the law should put the risk of an insolvent primary party upon the accessory rather than the victim: after all, the victim has done nothing wrong, whereas the accessory has knowingly assisted the primary wrong. The Law Commission has decided not to recommend any change to a general principle of joint and several liability, and it may be inappropriate to gnaw away at this long-standing principle in a piecemeal manner.Footnote 68
Stapleton has suggested that the law is rightly concerned to ensure that causally important parties, rather than less (causally) important “peripheral parties”, compensate the victims of torts: peripheral parties should not owe a duty to the ultimate victim of the tort.Footnote 69 Such a contention does not affect the argument made here; it is not contended that accessories owe a duty of care to the victims of the primary tort, such that accessories can be primarily liable for their own breach of duty. Instead, liability is parasitic upon the primary wrong of another. Thus an assister might be liable as an accessory for knowingly assisting the tort of negligence committed by the primary party, but there is no need to show that the assister owed a duty of care to the victim of the primary tort of negligence. Indeed, to establish such a duty will generally be impossible.Footnote 70
However, Stapleton's point might lead to the conclusion that one would wish that the primary wrongdoer, rather than the accessory, compensate the victim. Stapleton justifies shielding peripheral parties from liability on the basis that both the expansion of liability for omissions and the dilution of the deterrent function of tort law should be avoided.Footnote 71 However, both rationales are of reduced importance in the area here under consideration. An assister will, in the vast majority of cases, have positively done something to aid the primary tort, so accepting assistance liability is unlikely greatly to expand the scope of liability for omissions. Moreover, since the assister must have knowingly assisted the primary wrong, it is strongly arguable that imposing accessorial liability might actually further the goal of deterrence. Only knowing assistance – not mere negligent assistance – should lead to liability. This stringent requirement should placate fears of an unreasonable burden being imposed upon a third party.Footnote 72
1. Stevens' objections
Professor Stevens, in his excellent book, Torts and Rights, has criticised the equiparation of the criminal and civil law. Admittedly, he does so in order to dismiss the idea of a wide-ranging principle of accessory liability in tort law, but his arguments also mean that assisting a crime should not, without more, be a tort. Stevens presents five main reasons why accessory liability in the criminal law is distinct,Footnote 73 all of which may be questioned.
First, Stevens points out that there are no inchoate torts whereas there are inchoate crimes. This is irrelevant to the present discussion: the issue is whether an accessory should bear responsibility for a primary wrong which has actually been committed. The common law does not impose either criminal or civil liability upon an accessory if the envisaged primary wrong fails to materialise.Footnote 74
Secondly, Stevens argues that only the act of the primary party, not the liability or wrong of the primary party, is attributed to the accessory in the civil law, in contrast to what occurs under the criminal law. However, even if this difference does represent the law,Footnote 75 it is unclear why there should be a difference between the criminal and civil law when considering the method by which an accessory is made liable. Stevens' assertion that the civil law is different because liability may be imposed despite the absence of any “primary” wrongdoer is unconvincing; the criminal law may also impose liability upon an accessory where the “primary” offender has a special defence.Footnote 76
Thirdly, Stevens highlights the fact that case law suggests that the modes of participation necessary to establish accessory liability differ across the criminal/civil divide. For example, Stevens argues that while authorisation and ratification suffice to make an accessory liable for the primary wrong, such conduct elements are insufficient for criminal liability. This may be doubted: acts of authorisation, at least, would probably be considered to assist or encourage the principal offence. In any event, this argument only serves to show that the requirements necessary to establish accessory liability under the criminal law are, if anything, more stringent than those in civil law. This strengthens the point that wherever there is criminal liability for an offence which is also a tort, there should also be tortious liability.Footnote 77
Fourthly, Stevens places weight on the fact that in tort law a joint tortfeasor is liable as a principal for the tort, and assumes that if an assister were to be liable for the primary tort he would be so as a joint tortfeasor, whereas in the criminal law liability is secondary. The nature of the liability of the accessory is a difficult issue,Footnote 78 but even if liable “as a principal”, it is submitted that this should not preclude liability in tort where tortious liability is appropriate. Arguments about the precise nature of tortious liability should not distract from the essential argument that liability for assisting a crime which is also a tort should give rise to a civil claim in damages. It is of lesser importance whether such liability be classified as primary or secondary.
Fifthly, Stevens observes that the terminology used differs across the criminal/civil law divide: the criminal law tends to use the term “principal” whereas the civil law prefers “agent”. Whilst true, this is clearly not a solid basis for a radical distinction between civil and criminal liability for the same wrong. Terminology is flexible and can evolve alongside the development of a coherent doctrine of accessory liability.
2. Judicial reluctance
The most detailed, recent judicial consideration of the criminal/civil law divide is to be found in the decision of the Court of Appeal in Credit Lyonnais Bank Nederland N.V. (Now Generale Bank Nederland N.V.) v. Export Credit Guarantee Department.Footnote 79 Although strictly obiter, Hobhouse L.J. in particular spent some time setting out the framework of the criminal law and contemplating whether the civil law should follow a similar path. He identified three legal categories into which the actions of a criminal accessory may be placed.Footnote 80 First, conduct which is not dependent on the commission of an offence; this would lead to liability for inchoate offences. The civil law equivalent might be liability for conspiracy. The second category concerns conduct which only becomes criminal because of the crime of another. The third category concerns agency: responsibility for the act of the primary actor. His Lordship rightly identified the crucial question to be:
whether in civil law there is a personal civil liability for conduct which only comes under the second head or its equivalent (aiding) without also coming under the first (e.g. conspiracy or incitement) or the third (agency).Footnote 81
Hobhouse L.J. used the “archetypical example”Footnote 82 of Devlin J. in Gamble, concerning a person who sells a weapon knowing that it will be used to commit a crime. Hobhouse L.J. emphasised that “[t]he shopkeeper has no interest in nor does he join in the commission of the crime”,Footnote 83 which appears to take the shopkeeper outside the perceived boundaries which limit the scope of joint tortfeasance.Footnote 84 The reasoning of Hobhouse L.J. – which relied heavily on the exclusion of “pure” facilitation liability from the scope of joint tortfeasance – means that in such a situation the shopkeeper would not be liable to compensate the victim of the principal offence.
Significantly, Hobhouse L.J. asserted at the end of a long exposition of the authorities that:
[t]he criminal law for obvious policy reasons goes further than the civil law. Acts which knowingly facilitate the commission of a crime amount to the crime of aiding and abetting but they do not amount to a tort or make the aider liable as a joint tortfeasor.Footnote 85
This significant statement should not be endorsed. It is unsatisfactory simply to assert that there are “obvious policy reasons” for a difference between the civil and criminal law without stating what they are. Such defective reasoning was replicated in the House of Lords by Lord Woolf.Footnote 86 It is readily accepted that the paradigm cases of criminal and civil law differ, as do some of their underlying rationales, but when considering whether someone should bear responsibility for a wrong which is both a criminal and civil wrong, what are the “obvious policy reasons” for the criminal law going further than the civil law? It would be easier to comprehend why the criminal law should be more restrictive than the civil law, given the severe nature of a criminal sanction, but Hobhouse L.J. suggests that the exact opposite is true. It is submitted that this view is flawed; once responsibility for the primary wrong has been established by the criminal law, there is no reason for the civil law not to follow suit.
Carty has rightly observed that the conclusion to be drawn is that “the vendor of a knife who knows of the murderous use to which the purchaser wishes to put it would not be liable to the victim in the civil law”.Footnote 87 That she does so untroubled is surprising. More intuitive, perhaps, is the following passage from Winfield & Jolowicz on Tort:
D1 is attacking C. D2, a malicious bystander, throws a knife to D1, with which D1 stabs C. It seems extraordinary to suggest that D2 is not civilly liable for C's injury. Yet it is difficult to say that there is any procurement, common design or conspiracy.Footnote 88
It is submitted that it is right to view such a suggestion as “extraordinary”. It is to be hoped that a court would impose civil liability upon an assister found guilty of an offence as a secondary party. As one judge has remarked in the comparable context of deciding whether there should be civil liability for inciting a crime which is also a tort, in the absence of any persuasive authority to the contrary, one would expect the civil and criminal law to be in harmony.Footnote 89
C. Consistency with other aspects of private law
The previous section was concerned with situations where the primary tort was also a crime. It is submitted that a person who knowingly assists the commission of such a tort which is also a crime should incur civil liability. However, assistance liability in tort should not be restricted to wrongs which are criminal offences. Instead, assistance liability should be more wide-ranging. This accords with the approach taken in other areas of law: liability for assisting a breach of fiduciary duty is well recognised, and some cases also appear to recognise liability for assisting a breach of contract.Footnote 90
Stevens has argued that the principles of attribution mean that any analogy between assistance of breaches of contract or fiduciary duties and the commission of a tort is inherently flawed. This is because, although both the Lumley tort and dishonest assistance are accessorial in nature, they are distinct from the primary wrong, which cannot be attributed to the accessory: neither the contractual nor fiduciary obligations can legitimately be attributed to the accessory.Footnote 91 A breach of contract can only be committed by a contracting party, and a breach of fiduciary duty can only be committed by a fiduciary. By contrast, however, a tort can be committed by anybody, and Stevens concludes that this makes tort different; the act of the primary tort can be attributed to the assister, and the assister can then be liable in tort for that tort. Someone who assists a breach of contract is not liable in breach of contract.
This reasoning rests on the distinction between primary and secondary liability. However, it is not clear that a party who knowingly assists the commission of a tort should be liable as a primary party for the commission of the tort. Rather, the accessorial nature of this liability should be emphasised, in a similar manner to the liability recognised for assisting breaches of contract and fiduciary duties. Assistance liability in tort should still be recognised, regardless of how that liability is classified. It is submitted that it would not be incoherent always to describe the liability of the assister as accessorial. After all, the assister would not incur any liability without the existence of a primary wrong. Given that a person can be liable as an accessory to a crime, a breach of contract, and a breach of fiduciary duty, it appears strange to contemplate the possibility that a person can never be liable as an accessory to a tort.
Moreover, support can be found for a general principle of assistance liability in tort. In Petrie v. Lamont Tindal C.J. found that “[a]ll persons in trespass who aid or counsel, direct or join, are joint trespassers”,Footnote 92 and in M'Laughlin v. Pryor, Erskine J. stated that “encouragement” would suffice to make a person a joint trespasser.Footnote 93 It is submitted that such comments should not be limited to the tort of trespass. Indeed, Atiyah was explicit in writing that “[j]ust as in the criminal law relating to misdemeanours any person who “aids or abets” the commission of the offence is guilty as a secondary party, so it is clear that in the law of torts any one who assists the commission of a tort is liable as a secondary party”.Footnote 94 Such an approach is logical and deserves full consideration. It is a pity that Hobhouse L.J. in Credit Lyonnais dismissed such writers in a rather peremptory matter.Footnote 95
D. Assistance Rejected? Two Decisions of the House of Lords
Carty's rejection of assistance liability is consistent with the tenor of the two leading House of Lords decisions in this area, Credit Lyonnais Bank Nederland N.V. (Now Generale Bank Nederland N.V.) v. Export Credit Guarantee Department Footnote 96 and CBS Songs v. Amstrad Consumer Electronics plc.Footnote 97 These two decisions must now be examined. It appears as if the decision in CBS poses the greatest obstacle to finding assistance liability in tort law. This decision therefore merits serious scrutiny. However, it will be argued that some of the reasoning in this case should not, perhaps, be followed.
1. Credit Lyonnais
Credit Lyonnais Bank Nederland N.V. (Now Generale Bank Nederland N.V.) v. Export Credit Guarantee Department Footnote 98 concerned a fraud perpetrated on the bank, Credit Lyonnais. Mr Chong was the principal figure in the fraud; he fabricated contracts of sale, and supported these contracts with fraudulent bills of exchange which were signed by imaginary buyers. Mr Chong effectively sold these bills of exchange to Credit Lyonnais. The fraud was helped by the fact that these bills of exchange were guaranteed by the Export Credits Guarantee Department (“ECGD”). A senior employee of the ECGD, Mr Pillai, was corrupted by Mr Chong; in return for payment, Mr Pillai ensured that there were no problems with the underwriting of ECGD guarantees. Mr Chong made over £10 million from this fraudulent scheme, but then disappeared. The bank therefore sought to recover its losses from ECGD. The bank's claim failed at every level up to and including the House of Lords. There were various heads of claim, including breach of contract, but the only aspect of the case that is relevant to the present discussion is whether ECGD could be liable for the ultimate fraud.
The actual ratio of the House of Lords decision is very narrow: ECGD could not be made vicariously liable for the wrongs committed by its employee, Mr Pillai. Lord Woolf made it clear that this was the only question that their Lordships had to decide:
As the action has proceeded through the courts, many of the issues between the parties have been resolved so that there is now only one issue which we have to determine. This issue was accurately and succinctly identified by Stuart-Smith L.J. in his judgment [1998] 1 Lloyd's Rep. 19, 36 in these words:
“Where A becomes liable to B as a joint tortfeasor with C in the tort of deceit practised by C on B on the basis that A and C have a common design to defraud B and A renders assistance to C pursuant to and in furtherance of the common design, does D, A's employer, become vicariously liable to B, simply because the act of assistance, which is not itself the deceit, is in the course of A's employment with D?”Footnote 99
Lord Woolf, giving the only reasoned speech, answered this question in the negative. As his Lordship stated, “before there can be vicarious liability, all the features of the wrong which are necessary to make the employee liable have to have occurred in the course of the employment”.Footnote 100 For ECGD to be vicariously liable for Mr Pillai's wrong, Credit Lyonnais would have had to have proved that his mental element, conduct element and the primary fraud itself were within the scope of his employment. This is because without the primary wrong, Mr Pillai could not possibly be liable for a tort. This is a salutary reminder that accessory liability is not free-standing and is not inchoate; the primary wrong itself must be shown before accessory liability can be found.
The result of the case is welcome. It prevents liability being imposed upon a person too far removed from the primary wrong. Such a danger was highlighted by Coleridge J. in Lumley,Footnote 101 but can be avoided by only allowing an “immediate” accessory to incur liability for the primary wrong; any person further removed from the primary tort than an “immediate” accessory is too remote from the primary wrong to be held responsible for it. A person who is vicariously liable for a person who assists a tort should not be liable for that primary tort.Footnote 102 Therefore, in Credit Lyonnais, ECGD could not incur liability for the act of another accessory, Mr Pillai; ECGD was too far removed from the primary fraud.
This reasoning remains the same whether or not the assistance is part of a common design between the accessory and the primary wrongdoer. The comments that Lord Woolf made concerning liability for assistance beyond the scope of joint tortfeasance were therefore obiter. His categorisation of assistance liability as primary liability must be treated with caution,Footnote 103 and the suggestion that the bank was seeking compensation for a “separate tort” of assistance must not obscure the parasitic nature of such liability. Lord Woolf's comments might be confined to the context of vicarious liability.Footnote 104
The Credit Lyonnais litigation was further hampered by the fact that it was not even clear that Mr Pillai had knowingly assisted the fraud at all relevant times. As Longmore J., at first instance, pithily remarked, “[i]t is very difficult to form any view of Mr Pillai's state of mind now that he is dead”.Footnote 105 Given the importance of the mental element to accessory liability, this was always a significant obstacle to the bank's quest to establish accessory liability.
In the Court of Appeal, Stuart-Smith L.J. had insisted that “[i]t is not enough that he merely facilitates the commission of the tort unless his assistance is given in pursuance and furtherance of the common design”.Footnote 106 Such an approach appears to have gained the support of Lord Woolf in the House of Lords. However, it is submitted that this approach is too restrictive and renders assistance both irrelevant and insufficient; irrelevant, since the assister would also be liable for being a part of a conspiracy, and insufficient, since assistance without any conspiracy would not lead to liability. This is unsatisfactory, but has been replicated in a number of cases which habitually cite The Koursk as authority for this proposition.Footnote 107 Yet assistance liability, and accessory liability generally, is of much longer standing than The Koursk, and it is lamentable that this case seems to have influenced some judges to believe in an unnecessarily limited scope of accessory liability.Footnote 108
The comments of Lord Woolf in Credit Lyonnais need to be treated with great caution. His assertion that accessory liability in the criminal law and civil law have different roots is inaccurate, and his reluctance to consider assistance liability beyond the confines of it being a part of a common design unwarranted. If assistance liability is to be rejected by the courts, a stronger case needs to be made.
2. CBS Songs
CBS Songs Ltd. v. Amstrad Consumer Electronics plc involved hi-fi systems, manufactured by Amstrad, which enabled high speed recording from pre-recorded cassettes on to blank tapes. CBS Songs (“CBS”), who brought the action on behalf of themselves and other copyright owners in the music trade, were concerned that many purchasers of the hi-fi systems would use the recording facility to make pirate or unlicensed copies of copyright works. CBS claimed that Amstrad should face liability as a joint infringer of copyrights. Amstrad sought to have the claim struck out; it failed at first instance, but succeeded both in the Court of Appeal (by a majority) and in the House of Lords.
The case proceeded on the assumption that no crime had been committed.Footnote 109 Lord Templeman, giving the only reasoned speech, emphasised that there had been no authorisation, no common design, and no procurement which could lead to Amstrad being liable as a joint tortfeasor. He thought that mere facilitation was insufficient for joint liability.
This clearly strikes a real blow against any wider theory of assistance liability. However, the decision is not beyond criticism. For example, Lord Templeman's discussion of causation is unconvincing. His Lordship thought that a person could become a joint infringer if the accessory did “procure an infringement by inducement, incitement or persuasion”,Footnote 110 but excluded facilitation liability on the apparent basis that there is an insufficient causal link.Footnote 111 It is worth setting out the following passage of the judgment in full:
But in the present case Amstrad do not procure infringement by offering for sale a machine which may be used for lawful or unlawful copying and they do not procure infringement by advertising the attractions of their machine to any purchaser who may decide to copy unlawfully. Amstrad are not concerned to procure and cannot procure unlawful copying. The purchaser will not make unlawful copies because he has been induced or incited or persuaded to do so by Amstrad. The purchaser will make unlawful copies for his own use because he chooses to do so. Amstrad's advertisements may persuade the purchaser to buy an Amstrad machine but will not influence the purchaser's later decision to infringe copyright.Footnote 112
The problem with this passage is that “traditional” principles of causation find it difficult to accommodate accessory liability at all, even for procurement: an autonomous person makes decisions as a result of a number of factors, but responsibility for that person's acts rests with that person alone.Footnote 113 Lord Templeman elsewhere observed that “the operator of an Amstrad tape recording facility, like all other operators, can alone decide whether to record or play and what material is to be recorded”.Footnote 114 If this principle of autonomy is rigidly adhered to, then there should be no scope for any form of accessory liability – including procurement. This approach should not be favoured: accessory liability protects rights, deters wrongs, and makes the culpable liable. All that needs to be shown to satisfy the causal requirement is that the acts of the accessory had a more than minimal effect upon the primary wrong.Footnote 115
Furthermore, it is unclear to what extent Lord Templeman was considering the general law of tort rather than the particular context of the case itself. His Lordship relied upon the common law when assessing the scope of joint tortfeasance, but in discussing facilitation liability he relied heavily on Belegging v. Witten, which was another case concerning intellectual property.Footnote 116 Admittedly, Hobhouse L.J. in Credit Lyonnais stated that “the principles applied are drawn from the general law of tort. Infringement of a patent or copyright is a tort”,Footnote 117 but other judges have not voiced the same opinion. For example, Oliver L.J. in Paterson Zochonis Ltd. v. Merfarken Packaging Ltd. thought that:
These casesFootnote 118 do, it is true, establish the proposition that to supply goods which to the knowledge of the supplier will be used by a third party for the production of an article which constitutes an infringement of patent or copyright does not make the supplier himself an infringer. Speaking for myself, I am not convinced that they are necessarily also authority for the further proposition that knowingly to supply goods for the purpose of enabling a tort to be committed can never involve any liability on the part of the supplier short of inducement or common design.Footnote 119
The context of the claim in CBS undoubtedly preyed heavily on Lord Templeman's mind. He considered it to be “lamentable”Footnote 120 that home copying was widespread, and yet rendered illegal by the Copyright Act of 1956. He stated that “[a] law which is treated with such contempt should be amended or repealed”,Footnote 121 and it seems fairly clear from the tenor of his judgment that Lord Templeman felt that the Copyright Act 1956 did not warrant the strong intervention of the law against home copiers. As a result, it was, perhaps, always unlikely that Lord Templeman would impose liability upon any accessory to such a wrong under the Act. Moreover, this statute explicitly imposes certain forms of accessory liability in defined circumstances. This could feasibly be taken to mean that other common law forms of accessory liability should be impliedly precluded.Footnote 122 After all, Lord Templeman stated that the rights of the claimants “are to be found in the Act of 1956 and nowhere else”.Footnote 123 However, it is not clear whether the existence of the statute should exclude general principles of the common law where this is not explicitly provided for.
Another possible difficulty with the case brought in CBS was that the claimants failed to identify any primary wrong. Such wrongs were presumed to be many in number, but without identifying the precise wrongs in question, imposing accessory liability may have smacked of inchoate liability, which is entirely unprincipled.Footnote 124 On the other hand, it was abundantly clear that breaches of copyright were at issue in the case. But did Amstrad knowingly assist such breaches? This raises squarely the difficult issue of what it is that an assister has to know. The criminal law suggests that an accessory needs to know the “essential elements” of an offence, and this should be adopted in the civil law also. To be liable as an accessory to a tort, an assister must know that the acts in question would assist a primary wrong. The difficulty lies in determining what type of knowledge is required. It is suggested that an objective test of knowledge would lead to an unduly wide scope of accessory liability. A subjective test of knowledge should be preferred: liability should only be imposed if an accessory actually knew that his acts would assist the commission of a primary wrong. Such conduct is culpable. It is suggested that this requirement of knowledge should also include deliberately turning a blind eye to obvious facts of which the defendant was actually aware.Footnote 125
There remains, however, a tricky issue regarding the degree of knowledge required. Simester, discussing a similar situation in the criminal law, has written:
a … distinction should be drawn between general and specific knowledge: the latter is needed. As a computer manufacturer, I may know, statistically speaking, that 30% of my customers will use the computer illegally to copy music. In each particular sale, I therefore know there is a substantial possibility that my conduct will facilitate a crime of music piracy. But I do not know whether this customer will use the product for such a purpose.Footnote 126
Given that the accessory is further removed from the actual commission of the primary wrong than the primary wrongdoer, it seems only right that the primary wrong be identified and that the claimant should have to prove (on the balance of probabilities) that the defendant knowingly assisted that particular primary wrong. Simply asserting vagaries will not do. As Lord Templeman himself remarked when discussing liability for procurement:
Generally speaking, inducement, incitement or persuasion to infringe must be by a defendant to an individual infringer and must identifiably procure a particular infringement in order to make the defendant liable as a joint infringer.Footnote 127
CBS handicapped its claim by failing to establish a primary wrong, let alone a series of primary wrongs. Indeed, this might not have been very easy to achieve; as Lord Templeman recognised, “[t]he infringements of a home copier are almost impossible to detect”.Footnote 128 CBS may have been unable to circumvent such problems of proof by simply asserting that the acts of Amstrad probably assisted a probable tort.
This may suggest that accessory liability should not be an “easy option”, and should not be easier to establish than the primary wrong itself. Indeed, if the relevant primary wrongs are not established, it is very difficult to determine the appropriate remedy against the assister: what loss has the victim suffered? But the need to prove that the assister knew of specific wrongs will often prove troublesome. A manufacturer such as Amstrad is unlikely to know whether or not a particular customer will use the product for unlawful purposes. Moreover, if Amstrad would not be liable as an accessory for an individual's breach of copyright, it seems strange to contemplate the possibility that the shop owner who sold the hi-fi system to that primary wrongdoer, knowing that the hi-fi would be used to commit breaches of copyright, could nevertheless be liable.Footnote 129 This distinction may, perhaps, be defended on the basis that it mirrors the situation in the criminal law; for example, the making of knives is not prohibited, even though the manufacturers know that some of their knives will be used to commit crimes, but the selling of the knives by shopkeepers who know that a particular knife will be used to commit an offence is illegal. In a case such as CBS, it may seem counter-intuitive for a shopkeeper to be circumspect about selling a product which Amstrad lawfully sells on the open market, but it is suggested that shopkeepers should not sell a product if they know it will be used to commit an unlawful act.Footnote 130 This may seem harsh, but it should be remembered that the knowledge requirement for accessory liability is restrictively defined.
But should a manufacturer such as Amstrad inevitably escape accessory liability? Take the scenario, more difficult than that given by Simester,Footnote 131 where the computer manufacturer knows that, statistically speaking, it is not 30% of its customers who will use the computer illegally, but 90%. Should it still matter that the manufacturer does not know that this particular customer will use the computer in such a manner? On the balance of probabilities, the manufacturer does know that the computer will be used unlawfully – say, in order to commit a breach of copyright – but he is still not sure that this particular customer will do so. He might be considered to be turning a blind eye to the obvious – 90% is a very high probability. But would the manufacturer still have the mental element necessary for accessory liability if there were a 60% chance of his product being used unlawfully? It may be that knowledge only has to be shown on the balance of probabilities, but that it is knowledge that a particular wrong be committed that is required, rather than simply a suspicion that unascertainable parties may commit wrongs.Footnote 132 Thus a general suspicion does not suffice for liability, even if the manufacturer suspects that most of its customers will use the product for unlawful purposes. Short of actual knowledge, a particular suspicion in a given situation is required, and that suspicion must be deliberately ignored.
Under such an approach, the mental element for accessory liability is obviously very narrow. This makes it very difficult to find manufacturers such as Amstrad liable as accessories. In a more current context, this means that providers of peer-to-peer software are unlikely to incur civil liability as it would need to be shown which particular primary wrong they assisted, with the necessary mental element. This is not obviously inappropriate. If the product provided is not unlawful in itself, then the fact that it is used by others in an unlawful manner should not necessarily lead to those supplying the product being held liable. On the other hand, if the provider actively encourages the public at large to use the product unlawfully, knowing that this will be done by the majority of its purchasers, then it is difficult to explain why that provider should not incur liability. Perhaps such conduct might be categorised as “procuring”.Footnote 133
However, apparently for pragmatic reasons, where it is reasonably obvious that a large number of torts have been committed, the courts do not inevitably require each and every tort to be proved.Footnote 134 Although this seems to expand accessory liability, it should be remembered that a manufacturer such as Amstrad, if sued as an accessory, may also have a defence to the claim; just because a party has knowingly assisted the commission of a tort does not lead to the inevitable conclusion that the assister is liable as an accessory. The assister may be able to avail himself of the defence of justification. This defence is well recognised in the context of accessory liability for breach of contract,Footnote 135 and is equally appropriate in the law of tort. A person can know that he is assisting a primary wrong and yet have a legitimate reason for this; in such a situation the defendant could not fairly be said to be culpable. Perhaps the most striking example of when the defence of justification should apply occurs when the primary wrong in question “deserves” to be committed. In Brimelow v. Casson, for example, the “accessories” who induced theatre proprietors to break contracts with a theatre manager were justified in so doing, since the theatre manager was paying his chorus girls so little that the girls had to turn to “immoral earnings”.Footnote 136
Justification is a flexible defence which is difficult precisely to define. However, acting in good faith will not, of itself, be a defence – the defendant needs a stronger reason to outweigh the claimant's “primary” rights.Footnote 137 The scope of this defence will probably only become clear through decided cases,Footnote 138 but Carty's observation should be borne in mind: “clarity – obviously important in the commercial world – demands that liability be imposed on grounds that are certain and that the defence of justification should be residual”.Footnote 139
The defence of justification may be of real relevance in intellectual property cases where there exists a tension between protecting copyright and furthering technological advancement. In CBS, for example, the hi-fi systems were clearly capable of lawful uses of great significance. If the decision in CBS had been in favour of the claimants, the protection afforded to the owners of copyright would have been strengthened, although this may well have stunted technological innovation. The best balance between these competing elements may be to decide that knowingly assisting a breach of copyright does, prima facie, lead to accessory liability, but that there should be a defence of justification if the product or technology at issue is capable of substantial lawful uses.Footnote 140 In this manner, knowledge that a product may be exploited by some parties for unlawful purposes does not mean that the product cannot be distributed if it may also be used for lawful purposes. However, the manufacturer should not encourage customers to commit unlawful acts; although good faith is not sufficient in itself to make out the defence of justification, the defence should not be manipulated by those acting in bad faith.Footnote 141
IV. CONCLUSION
The fact that knowingly assisting a tort does not currently seem to lead to liability may represent a gap in the law. This is particularly so if the tort in question is also a crime. There is no reason for the criminal and civil law to differ in this area; assistance liability should be recognised. Moreover, there is no reason for the scope of knowing assistance in tort law to be limited to torts which are also crimes; expanding the scope of accessory liability would help to protect rights, deter wrongs, and sanction the culpable. This may even provide a suitable framework for other areas of private law.Footnote 142
