I. The “A→B” Scenario
A notoriously difficult problem for criminal complicity law is what might be called the “A→B” scenario. Scenarios of this kind arise where a secondary party (S) intentionally aids or abets a principal (P) to do one crime (Crime A), but P introduces a variation by perpetrating a crime (Crime B) that is in some significant manner different from Crime A. For example, S might agree with P to inflict minor injury upon V (Crime A), but P stabs V (Crime B). Or, in more difficult cases, S might agree with P to cause serious injury to V with fists (Crime A), but P causes serious injury by stabbing V (Crime B). S will be liable straightforwardly for Crime A if that was (also) done by P. Yet what, if any, principles does the criminal law have to navigate S’s liability for Crime B?
Historically, one set of principles was supplied by the so-called doctrine of “joint enterprise” or, as it was sometimes named, “parasitic accessorial liability”.Footnote 1 As is well known, joint enterprise liability was concerned with offences that arose as a result of the commission of agreed crimes. It had an explicit two-crime, A→B structure. If S foresaw that P might commit Crime B in the course of their joint enterprise to commit Crime A, and P did so, S may become liable as an accessory for Crime B. S and P’s common purpose to commit Crime A could be combined with S’s foresight of Crime B to construct liability for Crime B, without the need to prove that S directly aided or abetted P’s Crime B, or that she intended to do so. In 2016, this controversial doctrine was abolished in England and Wales by the Supreme Court in R. v Jogee; Ruddock v The Queen (Jogee).Footnote 2
What replaced joint enterprise? According to the Supreme Court, nothing. This means that there is nowadays just one, standard path to ascribing secondary liability to S for P’s offending. S must:Footnote 3
(1) aid, abet, counsel or procure P’s offence;Footnote 4 and
(2) intend to aid, abet, counsel or procure P’s offence, having knowledgeFootnote 5 of the “essential matters” of that offence.
In what follows, we call this the “Standard Test” for aiding and abetting.
By abolishing joint enterprise, the Supreme Court effected a considerable simplification of English and Welsh complicity law. Even so, there remains much complexity,Footnote 6 especially in A→B scenarios. In Section II, we shall see that the Standard Test contains various resources to help determine liability in situations where S contends that P’s actual conduct differed significantly from what S had in mind at the time of her own act of intended assistance or encouragement. Historically, supplementing those resources, a rule had developed in the context of joint enterprise whereby, if P’s Crime B was committed in a manner “fundamentally different” from the conduct S had foreseen, S was not liable for Crime B. In this paper, we contend that, by abolishing joint enterprise liability, Jogee made the fundamental difference rule redundant. All that was left behind were the resources, set out in Section II, of the Standard Test.
The Supreme Court, alas, failed to acknowledge the full implications of its own analysis. It created space for a version of the fundamental difference rule to live on and to find a new home within the Standard Test. The rule’s continued existence was sustained by certain open-ended remarks in Jogee about the possibility that an “overwhelming supervening act”Footnote 7 by P might place a limit on S’s liability for aiding and abetting. In subsequent homicide cases, the Court of Appeal has capitalised upon those remarks to develop an apparent dichotomy between overwhelming supervening events (which will negate S’s liability for manslaughter) and mere foreseeable “escalations” of the violence that was intentionally encouraged or assisted by S (which will not). In escalation cases, it seems that S can once again be held liable for something going beyond what she intentionally assisted or encouraged. Our contention is that neither overwhelming supervening acts nor escalation has any place in a post-joint-enterprise world, whether in the context of homicide or elsewhere, and that the law is currently being misapplied, founding homicide liability that ought not to exist.
In order to defend these claims, we begin in Section II by setting out some of the basic principles of the Standard Test, and their limits, before turning in Sections III and IV to explain why those principles leave no room for doctrines of overwhelming supervening acts and escalation that the Supreme Court toyed with in Jogee and the Court of Appeal has subsequently developed. In Section V, we offer a tentative explanation of the Supreme Court’s lack of clarity on this point: having ruled that joint enterprise liability had died out in the nineteenth century before being resurrected in the 1980s, the court was compelled to interpret certain decisions from the 1960s on overwhelming supervening acts in terms of standard aiding and abetting liability. This reading suggests, in turn, that those 1960s authorities survived the demise of joint enterprise in Jogee – thereby creating space for the Court of Appeal’s recent innovations.
II. Back to Basics: Bridging between Variations in P’s Offending
Our concern in this section is to outline how the Standard Test accommodates cases where the crime that P ultimately perpetrates involves a variation from what S understood herself to be aiding or abetting. Joint enterprise doctrine made it possible to bridge even very significant variations between agreed Crime A and its offshoot Crime B, provided that S foresaw the possibility that Crime B might be committed by P. By contrast, the paths to liability under the Standard Test are more restrictive. There are a number of doctrines guiding the courts within the Standard Test. Broadly, they address two kinds of criteria:
(1) Whether Crime B is of the same type as Crime A; and
(2) Whether Crime B falls within the scope of S’s encouragement.Footnote 8
We take these criteria in turn.
A. Variations within a Type: Bainbridge and Maxwell
In general, S will remain liable for P’s offence notwithstanding that its precise mode of commission differs from what S envisaged. The locus classicus of this proposition is R. v Bainbridge.Footnote 9 In that case, the Court of Appeal confirmed that, on a charge of aiding a burglary, it is unnecessary that S should have knowledge of the “particular date and particular premises” of the burglary which was in fact committed.Footnote 10 Dismissing S’s appeal against conviction, the court endorsed the trial judge’s direction, which had required that S intended to assist the type of crime perpetrated by P. Lord Parker C.J. clarified that this meant “the felony of breaking and entering premises and the stealing of property from those premises”.Footnote 11
The basic idea here is that the location of the burglary is legally irrelevant, because it is not part of the offence definition.Footnote 12 Just as one who steals a painting wrongly thinking it belongs to Victor, when in fact it belongs to Victoria, remains guilty of theft, S remained guilty of burglary even if he had thought it was a different bank that would be burgled. Similarly, it does not matter that an accomplice thought that the serious injury he encouraged or assisted P to inflict would be inflicted on a Tuesday, when in fact it was caused on a Thursday.
Understood in this way, Crimes A and B are of the same type when they have the same essential elements of liability. In such cases, the legally relevant ingredients of the offence that P commits are the same as the one that S intended to aid or abet.
“Type”, on this view, is relatively narrow. It does not extend to the broad category of P’s offence, such as “offences against the person” or “property offences”.Footnote 13 A useful analogy can be drawn here with the law of conspiracy. Where D1 agrees to commit theft, but D2 and D3 agree to commit theft with violence, D1 is rightly not guilty of conspiracy to commit robbery.Footnote 14 It makes no difference that both crimes can be, and often are, categorised as property offences.Footnote 15 Similarly, if S aids P to commit theft by informing P of where V keeps her laptop, it is submitted that application of the principle in Bainbridge means that S does not become guilty of robbery if P uses force to relieve V of her laptop.Footnote 16 Indeed, if all offences within a category were deemed to be of the same type, it is unclear why the legal mechanism of joint enterprise would have been thought necessary: despite its potentially having a wider reach, joint enterprise liability was applied by the courts principally in cases where S had agreed with P to use a certain level of violence against V, and P used greater violence, killing V. Non-fatal offences against the person and homicide offences might be thought to be of the same category (and we take no view on that debate here), but they clearly have different essential elements and – on the analysis here – are not of the same type.
One can, of course, put pressure on this analysis of what it means for offences to be of the same type. Consider David Ormerod, Karl Laird and Matthew Gibson’s example where S provides P with a jemmy to break into V’s home in order to steal: P uses the jemmy to break in but does so in order to cause grievous bodily harm.Footnote 17 Both scenarios involve the same crime (burglary), yet does P’s differing ulterior intention render these offences of a different “type”? The case law offers no clear answer to this question. Arguably, however, the answer is “No”. The difference does not matter, legally: either intention suffices for burglary under section 9(1)(a) of the Theft Act 1968.Footnote 18
Despite its unclear boundaries, Bainbridge demonstrates that the Standard Test accommodates at least some unexpected variations in the commission of P’s offence. Moreover, it is complemented by two supplementary doctrines that extend the range of A→B variations captured within the Standard Test.
1. The “essential matters” doctrine
The first of these is the “essential matters” doctrine. S must be proved to have known the “essential matters” of P’s crime at the time of providing the relevant encouragement or assistance.Footnote 19 According to that doctrine, a strict-liability element is dealt with differently depending on whether: (1) it is essential to make P’s conduct criminal in nature; or (2) it merely operates to aggravate P’s already criminal conduct. Let us elaborate this distinction.
As to (1), where a strict-liability element is part of what makes P’s conduct criminal in the first place, it is not a matter of strict liability for S. For instance, in Callow v Tillstone,Footnote 20 S, a vet, negligently inspected meat that P proposed to sell. P sold the meat, which was in fact unfit to be sold. Although the quality of the meat was a matter of strict liability for P, S was acquitted on the basis that S had not known that it was unfit to be sold.Footnote 21
By contrast, where (2) the strict-liability element of P’s offence merely aggravates what is already a crime, the same element is also a matter of strict liability for S.Footnote 22 This form of strict-liability aggravation is often referred to by academics as “constructive liability”.Footnote 23 Such constructive-liability, aggravating elements lie outwith the essential matters pertaining to P’s conduct that accessories must intend or know about. Consider, for example, murder, which at common law requires the killing of a human being with intent to kill or to cause grievous bodily harm.Footnote 24 If S intentionally aids or abets the deliberate causing of grievous bodily harm to V and V dies, S is liable for murder under the Standard Test. The “constructive” nature of murder liability carries over to secondary liability.Footnote 25
The foregoing approach was confirmed in Jogee,Footnote 26 and restated by the Court of Appeal in R. v Grant and others:Footnote 27
On a charge of murder, if the accessory intentionally assisted or encouraged the perpetrator and intended that the perpetrator should cause grievous bodily harm with intent, he or she will have satisfied the elements of the offence of murder. The precise manner in which the victim happens to be killed and whether the perpetrator intended to kill as opposed to inflict really serious harm are by the way.
Similarly, in unlawful-act manslaughter, only the unlawful act counts as an essential matter for the purposes of secondary liability.Footnote 28
For the purposes of complicity law, Crime B is of a different type from Crime A only when it has different essential matters, and constructive-liability elements of P’s crime are simply not considered essential. Hence, S need not know about them in order to be an accessory to the aggravated offence.
2. Secondary liability and multiple potential offences
A second supplementary doctrine was articulated in Director of Public Prosecutions for Northern Ireland v Maxwell.Footnote 29 As was confirmed by the House of Lords in that case, S can be liable for P’s crime when S has in contemplation the type of offence that P in fact commits, albeit alongside other possible types of crime:Footnote 30
[S] may have in contemplation only one [type of] offence, or several: and the several which he contemplates he may see as alternatives. An accessory who leaves it to his principal to choose is liable, provided always the choice is made from the range of offences from which the accessory contemplates the choice will be made. Although the court’s formulation of the principle goes further than [Bainbridge], it is a sound development of the law and in no way inconsistent.
Maxwell extends the reach of complicity to a case where S knows that P will commit one of crimes X, Y or Z, but lacks certainty about which of them it will be. Like Bainbridge, it allows some (limited) flexibility for the law to accommodate offence-variations by P, provided that S was sufficiently aware of the essential matters regarding the crime that P ultimately commits. As such, the foregoing doctrines articulate the most important restrictions that the Standard Test contains for A→B scenarios. They operate to ensure that S’s conviction is for a crime of the same type that she intended to encourage or assist.
Our argument is, then, that Crime B will be of a different type from Crime A only where it contains distinct, legally salient, essential matters. Indeed, on one view, the type-based approach in Bainbridge and Maxwell is simply another way of explaining the essential matters doctrine, which has been part of the Standard Test since at least Johnson v Youden and Others.Footnote 31
B. Going beyond the Scope: The Calhaem Principle
There is an exception to the type-based analysis adopted in Bainbridge and Maxwell, and endorsed in Jogee. Sometimes, albeit rarely, S may be absolved of responsibility for Crime B even though that crime is of the same type as Crime A. As was explained in R. v Calhaem, P must also act “within the scope of the [S’s] authority or advice, and not, for example, accidentally”.Footnote 32 If P goes beyond that scope, S is not liable.
While nothing said in Jogee contradicts it, the limits of the Calhaem principle remain uncertain.Footnote 33 Something like the principle is found in ancient discussions of cases where S had procured P to commit an offence. Where P “wilfully and knowingly commiteth”Footnote 34 an offence different from the one S had procured P to perform, whether having different legal elements or directed against a distinct victim, S was not guilty of P’s offence. Such cases were contrasted with variations “in circumstances of time and place, or in the manner of execution”, where P “in substance complieth with the temptation”.Footnote 35 In the latter situations, secondary liability flowed. There is no bright line here and the courts have never resolved how “finnicky”Footnote 36 S can be when, effectively, insulating herself from secondary liability for variations by P.
In Calhaem itself, S was alleged to have hired P to murder V. Having taken a down payment, P apparently decided not to kill V but, rather, to stage a failed attempt – even going to the lengths of removing the shot from his shotgun cartridges. P then visited V, shotgun in hand. When V understandably screamed in response to P’s charade, P went berserk and attacked V with a hammer, before stabbing her with a knife. V died as a result of P’s attack. The prosecution put its case on the basis that S had counselled or procured P to commit murder and S’s conviction of murder was upheld by the Court of Appeal on the basis that even a “remote” instance of counselling could suffice for secondary liability.Footnote 37
The facts of Calhaem were differentiated by the Court of Appeal from two other scenarios. First, the court considered variations where the relevant result was brought about by P “accidentally when the mind of the final murderer did not go with his actions”.Footnote 38 The language of “accidents” here is potentially misleading. As the Supreme Court noted in R. v Gnango,Footnote 39 if S encourages P to shoot at V and P misses but hits T, an innocent bystander, the principle of transferred malice means that S can be liable for the injury suffered by T.Footnote 40 The shooting of T is, as the Supreme Court itself noted, “accidental” in the sense that neither P nor S meant for T to be shot. Primary liability for P, and secondary liability for S, nevertheless flows undisturbed, at least in terms of doctrine. This kind of accidental variation remains within the scope of S’s encouragement, because it does not change the type of offence committed by P.Footnote 41
The second comparison made in Calhaem involved an “unlikely”Footnote 42 example where it was asserted that S would not be a secondary party to P’s offending. Suppose that P kills V during a riot, not realising that V is the person whom P was previously counselled by S to murder. In such a case, P would not “have been acting within the scope of [S’s] authority; he would have been acting entirely outside it, albeit what he had done was what he had been counselled to do”.Footnote 43 In other words, it is not sufficient that P coincidentally commits an offence with identical legal elements to the one that S has counselled: there must be some rational connection between S’s contribution and P’s decision.Footnote 44
This is about as far as one can safely go on the authorities. Calhaem is a narrow restriction upon the Standard Test for secondary liability. The restriction is, in our view, a justified one. Secondary liability is an inherently problematic area because it goes against the general grain of the criminal law, by placing S’s liability in P’s hands.Footnote 45 The Calhaem principle means that S can sometimes set parameters on her encouragement or instruction, and thus “delimit the scope of [her] participation in an intended crime”, even if it remains, legally, the same type of crime envisaged by S.Footnote 46 For example, if S makes it sufficiently clear that she is encouraging or instructing the killing specifically of V, P’s decision to kill T instead negates S’s secondary liability, as in the famous case of R. v Saunders and Archer.Footnote 47 Similarly, if S is clear that her support is limited to a certain means of causing V grievous bodily harm (say, a “punishment” beating, or a knee-capping where the intention is that V survives) and P unexpectedly produces a knife and cuts V’s throat, S should not be liable for murder.Footnote 48 That conclusion has been doubted as a matter of doctrine, post-Jogee,Footnote 49 and its defence lies beyond the scope of this article; but it can potentially be supported on our reading of Calhaem.
Cases such as Bainbridge, Maxwell and Calhaem help to delimit secondary liability in A→B variation cases under the Standard Test. They operate as qualifications upon the actus reus and mens rea requirements of complicity. They are not things apart from those basic building blocks of complicity liability. In terms of substantive legal doctrine, they do not operate as “supervening” defences. It is an unnecessary complication to the law of secondary liability to set the doctrines analysed above up as distinct, supervening doctrines. Clarifying this point allows one to see, across the next two sections, why, despite their appearance in Jogee and subsequent cases, there is no distinctive place for overwhelming supervening acts, let alone escalations, in the Standard Test.
III. Overwhelming Supervening Acts: An Exculpatory Doctrine?
Joint enterprise liability was controversial in part because it extended the liability of participants to Crime B when they merely foresaw that it might be committed in pursuit of their common purpose to commit Crime A. There was no need for the prosecution to prove that S had actually encouraged or assisted Crime B, or that S had intended that Crime B would be encouraged or assisted by their conduct.
Because the extension of liability to Crime B was predicated merely on foresight of its possible commission by P, this form of liability was potentially very broad and was subject to certain limiting exceptions. One such exception was the doctrine of “fundamental difference”. If S formed a common purpose to perpetrate Crime A, with foresight that P might commit Crime B during that enterprise, S was nonetheless not guilty of Crime B should P perpetrate Crime B in a manner fundamentally different from what S had foreseen.Footnote 50 As explained in R. v Powell and Another; R. v English, “if the weapon used by [P] is different to, but as dangerous as, the weapon that the secondary party might use, [S] should not escape liability because of the difference in weapon”; otherwise, however, it was possible under certain circumstances for S to escape liability for murder or manslaughter.Footnote 51 For the most part, as this quotation suggests, the fundamental difference constraint was focused on matters such as whether S knew that P was armed and, sometimes, what kind of weapon P was armed with.Footnote 52 In general, where P committed Crime B in a manner that was far more likely to cause death than what S had foreseen, S was off the hook for Crime B altogether.Footnote 53 This point bears restatement: if Crime B was murder, then the fundamental difference rule meant not only that S was not guilty of murder; she was also not guilty of manslaughter.
After Jogee abolished joint enterprise, the fundamental difference rule became unnecessary. Secondary liability based simply on foresight of potential collateral offending was now unavailable and so no brake on such widespread liability was required.Footnote 54 Unfortunately, the Supreme Court did not disavow entirely the concept of fundamental difference. It recast it in terms of overwhelming supervening acts.Footnote 55 In the relevant extracts from Jogee, what appears to be at issue is whether P’s ultimate crime was a foreseeable upshot of S’s existing contribution, namely something that comes after S became prima facie an accessory to a crime. If the variation was unforeseeable, then S is not liable: “[I]t is possible for death to be caused by some overwhelming supervening act by the perpetrator which nobody in the defendant’s shoes could have contemplated might happen and is of such a character as to relegate his acts to history.”Footnote 56
This type of thinking was apt to joint enterprise cases: to scenarios in which P went further than S’s standard accessorial liability for aiding, abetting, counselling or procuring stretched. To be sure, if S has already become prima facie an accessory to one crime (A) and P’s further crime (B) is an overwhelming supervening act, the Standard Test would not render S guilty of Crime B. But, once more, this is not because of any special supervention doctrine. It is because S would not be liable for Crime B on basic principles: she did not intentionally encourage or assist the commission of Crime B. Since Jogee, that liability for Crime B must be established directly, in its own terms. It cannot be constructed upwards from any liability that S has for Crime A. Previously, a bridge to Crime B liability was provided by joint enterprise: that bridge is now fallen.
Taken together, the comments in Jogee invite the following misreading: “Where P does something that was not reasonably foreseeable following S’s intended contribution, then P’s act is fundamentally different from the one that S intended to encourage/assist; it is an overwhelming supervening act, and S is not liable. Otherwise, S is liable for P’s offence.”
A reading of this sort fits poorly with the passages in Jogee which suggest that the language of overwhelming supervening acts simply denotes situations where the actus reus and/or mens rea of complicity under the Standard Test cannot be proved. Consider, for instance, the following statement from Jogee:Footnote 57
[T]here may be cases where anything said or done by [S] has faded to the point of mere background, or has been spent of all possible force by some overwhelming intervening occurrence by the time [P’s] offence was committed. Ultimately it is a question of fact and degree whether [P]’s conduct was so distanced in time, place or circumstances from the conduct of [S] that it would not be realistic to regard [P]’s offence as encouraged or assisted by it.
Similarly, immediately after referring to overwhelming supervening acts, the court confirmed that “[w]hat matters is whether [S] encouraged or assisted the crime …. He need not encourage or assist a particular way of committing it, although he may sometimes do so”.Footnote 58
Both of these quotations suggest that there is no liability-limiting exception at work here. It is simply that, on the facts of the case at hand, it may be that the prosecution cannot prove beyond reasonable doubt that S in fact encouraged or assisted P’s offence; notably, where what S did is “so distanced in time, place or circumstances” that its possible connection to P’s offending is “spent”.Footnote 59 It is not news that such factual encouragement or assistance is required for the actus reus of secondary liability – that requirement has been clear for decades.Footnote 60
Consider a situation in which S joins P in confronting V and others. S knows that P has a knife, but P unexpectedly produces a revolver and shoots V at point-blank range.Footnote 61 Here, it might be argued that S had not in fact intentionally encouraged or assisted P’s action, because S never anticipated that P would perform an act of this type. S’s liability will, of course, depend on what the prosecution can prove. Perhaps they can prove an intention on S’s part to encourage the use of the knife to cause grievous injury with intent, in which case S should be found guilty of murder, notwithstanding P’s use of a more life-threatening means of killing V. Alternatively, perhaps it cannot be proved that S intended to do anything more than encourage the use of the knife to frighten V, in which case P’s crime of murder is different, in terms of its legal definition, from what S intended to encourage (an assault). But that possibility has nothing to do with overwhelming supervening acts as a distinct legal category. Rather, it concerns the basic actus reus and mens rea of secondary liability – namely the very prima facie case against S.
Instead of using the language of overwhelming supervening acts, it would be more profitable to speak only in terms of those actus reus and mens rea requirements, just as does the Standard Test. So, for example, if S encourages P to kidnap V and demand a ransom and P uses a knife to cut off V’s finger, talk of overwhelming supervening acts is an unnecessary distraction.Footnote 62 Without more, this is simply a case where S cannot be proved to have either encouraged or intended to encourage P to cause grievous bodily harm with intent.
The language of overwhelming supervening acts is, accordingly, unnecessary. But it is also dangerous. It risks misdirecting our focus, because we tend to think of “supervening” arguments as being things that come after the establishment of an actus reus and mens rea. For instance, the nomenclature of “supervening defence” is sometimes used to signify a defensive argument that “denies neither actus reus nor mens rea but, rather, seeks to avoid liability by reference to accompanying considerations not contemplated in the offence definition”.Footnote 63 P hit V (actus reus). P meant to (mens rea). But P was defending himself using proportionate force (supervening defence). The problem with certain aspects of the overwhelming supervening act discussion in Jogee is that they suggest some form of supervening defence exists in variation cases. In R. v Rowe, the Court of Appeal compounded matters by talking about overwhelming supervening acts as an “exclusion” of secondary liability, as opposed to being relevant to the establishment of the actus reus.Footnote 64 Language such as this leads Matthew Dyson to refer to overwhelming supervening act as a “specific defence”Footnote 65 to secondary liability and Beatrice Krebs to conceptualise it as a free-standing guarantee of just attribution of liability, something she believes to be particularly needed given the constructive nature of murder under English law.Footnote 66 But, as demonstrated above, there is no “supervenience” here; no work for bare foreseeability (as opposed to intention) to do. If one wishes to talk in terms of defence labels, there is merely a “failure of proof” argument:Footnote 67 the prosecution has failed to prove the elements of the Standard Test.
That this is the position under the current law is corroborated by the fact that the law’s erstwhile fixation with the type of weapon that P used – the hallmark of joint enterprise’s fundamental difference rule – is gone, post-Jogee. In murder, “the question is whether [S] intended to assist the intentional infliction of grievous bodily harm at least”.Footnote 68 Evidence regarding S’s knowledge of what P was armed with is relevant to that question, but not determinative of it.Footnote 69
If the Standard Test is kept firmly in mind, ostensibly tricky cases can become relatively straightforward. In Grant,Footnote 70 S and P had a shared purpose to cause grievous bodily harm to V. Post-Jogee, that can be interpreted as constituting mutual encouragement to cause such harm.Footnote 71 For that purpose, they had knives, a sledgehammer and a baseball bat in the boot of their car as P, with S beside him, drove around the neighbourhood looking for V. Suddenly, P spotted V and drove straight into him, killing him.
In post-Jogee currency, S’s liability must be analysed in terms of abetment of murder. On this approach, S’s liability is uncomplicated: S encouraged P to deliberately cause grievous injury to V and, within the scope of that encouragement, P intentionally killed V.Footnote 72 The killing was within the scope of the encouragement, at least so far as can be discerned from the facts recorded in the Court of Appeal’s judgment, because S had not insisted that V must be caused grievous bodily harm using only the weapons in the car. We think that, on the reported facts, it would have been odd, for instance, for S to shout “What on earth are you doing?” as P accelerated towards V.
The Court of Appeal nevertheless entertained the submission that overwhelming supervening events were relevant to accomplice liability post-Jogee and gave the following guidance on cases involving such phenomena:Footnote 73
[T]he principal focus … will be on whether there is a credible basis for suggesting that anything said or done by the accessory by way of encouragement or assistance “has faded to the point of mere background, or has been spent of all possible force by some overwhelming intervening occurrence by the time the offence was committed” … and which “nobody in the defendant’s shoes could have contemplated might happen and is of such a character as to relegate his acts to history” …. [U]ltimately the question will be whether the accessory’s conduct may have been “so distanced in time, place or circumstances from the conduct of [the principal] that it would not be realistic to regard [the principal’s] offence as encouraged or assisted by it”.
It would have been simpler, and more accurate, for the Court of Appeal to restrict itself to the point it had already made: that the precise means of killing V is not a legally relevant matter in a case like Grant. It is the actus reus of murder to kill V by whatever means, not specifically by stabbing, running over or the like. P’s actions constituted a mere variation within the type of activity that S was abetting and no Calhaem point arose. Additionally, due to the operation of constructive-liability (i.e. aggravating strict-liability) elements of offences like murder, it is abetment of murder for S intentionally to encourage P to deliberately cause grievous bodily harm to V, if V thereby dies. One might baulk at the fairness of this approach (and, perhaps, at the fairness of similarly expansive constructive-liability elements in the law of involuntary manslaughter). Appealing to a redundant hangover from a dead mode of liability is not, however, an appropriate remedy.
The question in Grant was, accordingly, whether the type of crime that P perpetrated, and which caused V’s death, was in fact encouraged by S. The same question could be put a slightly different way by asking whether the crime P perpetrated was rationally connected to the encouragement provided by S.Footnote 74 Plainly, the answer was affirmative: the group set out to cause serious injury to V and P used the car to effect that purpose (and perhaps even to kill, albeit that further intention did not matter, legally, when assessing S’s liability).Footnote 75
A. Should the Law Recognise Such an Exculpatory Doctrine?
Of course, one might contend that, even if there is no need for it following Jogee, there should be an overwhelming supervening act rule. Krebs, for instance, argues that the overwhelming supervening act doctrine can augment the law’s existing, scope-based limitations on complicity found in cases such as Bainbridge, Maxwell and Calhaem.Footnote 76 Suppose that S agrees with P to inflict serious injury in a fist fight with another gang. Rather than pulling out a knife (a foreseeable escalation, say), P suddenly dashes to his car and returns with an assault rifle, with which he kills their antagonists. Prima facie, this variation is irrelevant to S’s liability for murder: S intended to encourage the deliberate causing of grievous bodily harm and that is what happened; again, the constructive-liability elements of murder apply to accessories as well as principals. Yet it is, let us concede, unforeseeable that P would carry out the offence using that weapon. In R. v Lanning and another, the Court of Appeal dismissed the notion that use of a knife was an overwhelming supervening event by highlighting that “knives are produced in situations of this kind with a high degree of frequency”.Footnote 77 The same cannot be said of assault rifles.
There is something to be said for recognising a novel doctrine of this sort and perhaps it is these kinds of possibilities that encouraged the Supreme Court to cling to the idea of “overwhelming supervening acts”.Footnote 78 The essential matters doctrine allows for an expansive approach to constructive-liability crimes even under the Standard Test: something that a distinct overwhelming supervening act doctrine has the potential to trammel. In the light of Grant itself, however, it seems unlikely that a scope restriction of this sort is part of the current law. Neither is the case for it incontestable. Unlike the special case discussed in Calhaem, the outcome in our example lies within the scope of what S abetted; moreover, there is no variation in the type of crime. Only the means of causing that outcome is different, a difference that is plausibly morally as well as legally irrelevant to S’s culpability.
B. When Variations Will Matter
This is not to claim that all variations resulting in death are legally irrelevant under the current law. As the Court of Appeal noted in R. v Tas,Footnote 79 the facts of R. v Rafferty Footnote 80 offer an example where secondary liability is not made out on the Standard Test. S had participated in an attack on V, which took place on a beach. S then left to obtain money using V’s bank card. Whilst S was away, the principals carried V out into the water and drowned him. Here, the Court of Appeal thought there was a credible case, albeit “on the unusual facts”,Footnote 81 that S had not encouraged or assisted the principals to kill V. Neither had S intended to do so. S was a party to the initial attack, but not to the drowning; and the initial attack was not a cause of V’s death.
Explaining the case by reference to the Standard Test is preferable to invoking a supervening doctrine of exculpation, either by relying on the claim that the killing was “a new and intervening act” (which suggests that Rafferty was already in principle liable for a homicide offence, before something came along to alter that fact),Footnote 82 or that S had “withdrawn” from a joint enterprise (which is simply inapposite on the facts, given the reason why Rafferty was absent when the victim was drowned).Footnote 83
Notions of overwhelming supervening acts and fundamental differences are unnecessary distractions from the basic principles of secondary liability as reasserted in Jogee. Those notions should have been interred alongside joint enterprise, for they lack independent legal significance. As will be seen in the next section, however, the Supreme Court’s failure to acknowledge this has effectively allowed for a resurrection of joint enterprise-type thinking by the Court of Appeal. This leads to convictions for manslaughter in cases where, we argue, there should be no liability for homicide at all.
IV. “Escalation”: An Inculpatory Doctrine?
To some extent, our objection to talk of overwhelming supervening acts could be seen as no more than formal. We have critiqued some “enigmatic and unsatisfactory”Footnote 84 aspects of Jogee, but, if one looks at the actual decisions post-Jogee, one finds that the Court of Appeal has not been willing to exculpate defendants on that basis. Effectively, the court has treated claims of overwhelming supervening acts as red herrings. One might, therefore, be inclined to sanguinity about overwhelming supervening acts. If the category is, in practice, empty, what does it matter if in theory it continues to exist?
Unfortunately, there is a practical difficulty here. It lies in the fact that invoking the language of overwhelming supervening acts has served to obscure what should be the courts’ focus on the Standard Test for complicity’s actus reus and mens rea elements. If one takes those requirements seriously, then, unless S intentionally encourages or assists the type of crime that causes death, S should not be guilty of manslaughter. S may be guilty of whatever non-fatal offence she deliberately encouraged or assisted P to perform, but – crucially – unless that offence was (also) a cause of V’s death, she cannot be guilty even of manslaughter.Footnote 85
Contrast the situation that has arisen post-Jogee, as a result of judicial preoccupation with overwhelming supervening acts. Where there is no overwhelming supervening act following S’s encouragement or assistance of a crime of violence (Crime A), further violence (Crime B) is generally treated as a mere escalation of Crime A and V’s death in such circumstances will render S guilty of manslaughter. In other words, unless there is an overwhelming supervening act in such cases, there is, apparently, intentional assistance or encouragement of (the essential matters of) a manslaughter. The effect of this misperception is to create a novel inculpatory doctrine for A→B scenarios. Moreover, the new doctrine appears to be one of substantive law, rather than an evidential thesis or a bare description of the facts of a particular case. If the jury finds that death did not result from an overwhelming supervening act by P, it ought to find that death resulted from an escalation of the violence that S intentionally encouraged or assisted. As a result, it ought to convict S of manslaughter.
This liability-constructing approach finds support in Jogee itself: “If a person is a party to a violent attack on another, without an intent to assist in the causing of death or really serious harm, but the violence escalates and results in death, he will be not guilty of murder but guilty of manslaughter.”Footnote 86
Recall, too, that the Supreme Court and the Court of Appeal have generally explained overwhelming supervening acts by reference to what was unforeseeable in the circumstances of S’s contribution to P’s offending. When these two points are taken together, we find that the Supreme Court has facilitated the creation by the Court of Appeal of an apparently new form of manslaughter liability premised on the foreseeability of P’s using more serious violence than was intentionally encouraged or assisted – or even foreseen – by S.Footnote 87 So, for example, in Tas, the Court of Appeal thought that the question to be resolved was whether “the production of a knife is a wholly supervening event rather than a simple escalation”.Footnote 88 If the use of the knife was merely an escalation, it seems, S could not complain about a conviction for manslaughter.Footnote 89 But the Standard Test, which is the only form of complicity that Jogee leaves us with, makes it clear that what counts for S’s liability is whether S intentionally encouraged or assisted the type of act that caused V’s death. If she did not, then she is not guilty of a homicide offence. Escalation obscures this by looking beyond the type of act that S intentionally encouraged or assisted. It reintroduces joint enterprise-type thinking via the back door.
To illustrate further the impact of the wider test adopted by the Court of Appeal in the aftermath of Jogee, consider the facts of English.Footnote 90 S and P agreed to assault V with fence posts. During the attack, P produced a knife, which S did not know P had, and stabbed V to death. It was held that, even if S had foreseen the possibility of P’s deliberately causing serious injury to V, P’s act was fundamentally different from what S had foreseen. In consequence, S was not guilty of murder or manslaughter. Presumably it was foreseeable that the planned violence might escalate. In consequence, S would now become guilty of manslaughter even if S had not foreseen the possibility of P’s going further than planned.
Strikingly, the escalation doctrine has been invoked to date only in situations where V dies. Imagine that V had survived being stabbed in English. On joint enterprise principles, presumably the knife wound would still have been fundamentally different from what S had foreseen might happen and so S would not be liable for a section 18 offence. This same result would appear to follow subsequent to Jogee: however foreseeable it was that the violence would escalate, S is only responsible for that escalation if it was directly, and intentionally, aided and abetted by her on standard principles.Footnote 91 No justification is given in Jogee for why homicide should be anomalous.
It is also striking that the escalation doctrine seems only to have been invoked in situations where S is present and participating – that is, in the kind of scenario that was the paradigm of joint enterprise. Suppose that S lends P some brass knuckles in order to assist P’s plan to attack and injure V the next day (for the sake of argument, S intends to assist P to inflict actual bodily harm). On the day, however, V fights back. P then produces a knife and stabs V, intending to cause grievous bodily harm. V dies of the stab wound. Certainly, S is guilty of a section 47 offence. But if S was absent throughout the events of that fatal day, it seems unlikely that the courts would be willing to convict her of manslaughter on the basis of escalation.Footnote 92 Certainly, that option would not be available on the Standard Test, even if it was foreseeable that P might end up using a knife. Yet, in principle, the same law applies, post-Jogee, whether or not S is present. There is nothing in Jogee that draws a legal distinction between present and antecedent assistance or encouragement – and if that is right, any new escalation doctrine ought to apply when S is absent too. Once again, any such distinction between presence and absence appears to have been relevant to joint enterprise alone, since the technical distinctions between aiding, abetting, counselling and procuring (including whether S was present at the time of P’s offending) were excised by the Criminal Law Act 1967.
There are, to repeat, no longer any two-crime, A→B scenarios. S’s guilt of a lesser offence cannot be constructed into the more serious crime merely on the basis that such a variation in P’s offending was foreseeable – or even that it was (merely) foreseen. S must have directly and intentionally aided or abetted that more serious crime according to the basic principles of aiding and abetting that undergird the Standard Test. One can readily imagine why the Supreme Court and the Court of Appeal have been resistant to that conclusion. The retributive urge is strong when a life has been lost. But the point of law is to temper such urges.
All of this is to say that the question is not whether an escalation has become an overwhelming supervening act, such that accessorial liability is negated.Footnote 93 The question is whether S intentionally aided, abetted, counselled, or procured the type of offence that P perpetrated (or at least its essential matters). Notions of overwhelming supervening acts and escalations are simply distractions. And dangerous ones at that, in as much as they have the potential to generate manslaughter convictions that are unwarranted on the Standard Test.
V. A Tentative Diagnosis
One might wonder why the appellate courts have not restricted themselves to the traditional criteria in the Standard Test, and have instead deployed the concepts and language opposed here. Here, we offer some tentative thoughts.
First, there are comments in Jogee that suggest that the Supreme Court shared the view articulated here, concerning the limited relevance of overwhelming supervening acts. As was shown above in Section III, the Supreme Court appears to have seen clearly at points that the question is one of satisfying the basic actus reus and mens rea elements of complicity liability, not a distinct problem of supervening interventions.Footnote 94 It may be that “overwhelming supervening act” is simply being used as a synonym for variations of crime-type that fall outside the range of Bainbridge and Maxwell.Footnote 95 On that view, the phrase remains compatible with the Standard Test, but it does no work of its own and serves only to confuse matters. Unfortunately, the judgment in Jogee is not always perspicuous, inviting the manoeuvres criticised above.
Second, the history of complicity is notoriously murky and this opaqueness may help to explain the Supreme Court’s concern with overwhelming supervening acts and the language of escalation – preoccupations that have been seized upon in the Court of Appeal’s subsequent jurisprudence. The Supreme Court’s view in Jogee was that the Privy Council in Chan Wing-Siu and Others v The Queen Footnote 96 resurrected joint enterprise liability in 1984, following its demise at some point during the nineteenth century. On that version of history,Footnote 97 it becomes more pressing to find a role for overwhelming supervening acts in post-Jogee complicity. After all, the major references to this concept – in Smith,Footnote 98 R. v Anderson Morris,Footnote 99 R. v Betty Footnote 100 and R. v Reid Footnote 101 – are from the 1960s and 1970s. On the Supreme Court’s version of history, these decisions could not be about joint enterprise. Accordingly, these decisions had to be understood as meaning something in relation to the Standard Test for complicity, or else be accounted wrong or misunderstood. For whatever reason, the Supreme Court chose to endorse those Court of Appeal decisions, arguing that they involved application of the Standard Test and not joint enterprise. That has facilitated continued interest in, and reliance upon, redundant concepts.
The historical account of complicity preferred in Jogee has been doubted,Footnote 102 although it is unnecessary to reopen that debate here. Our suggestion here is merely that, given the sheer opacity of the history, it is plausible that twentieth-century courts prior to Jogee had not explicitly recognised, let alone acknowledged, the demise of joint enterprise liability. As such, statements by those courts about overwhelming supervening acts and escalation could be read as being (per Jogee, misguided) references to joint enterprise liability and thus restricted to that context. Indeed, this seems to have been the view of Lord Hutton in Powell and English: his Lordship saw the “fundamental difference” rule as being an example of those 1960s and 1970s decisions, not a thing apart from them.Footnote 103 That explanation remains viable even if one accepts that Chan Wing-Siu did indeed resurrect a dead head of liability.
Reasonable disagreement is possible on the matters addressed in this section. It is a matter of great regret that the courts have never delineated the grounds of complicity liability as clearly as they might have and this is, of course, part of the reason why joint enterprise was so controversial.
VI. Conclusion
Whatever the history, there is no doubt where we are today as a matter of the law’s logic. Jogee abolished joint enterprise liability. No longer is there need for a restriction upon such liability of the sort contemplated by rules about fundamental differences and overwhelming supervening acts. There is also no place for the view that the absence of such differences and acts establishes that there was simply an escalation in P’s offending, rendering S liable as an accessory on account of her participation in a distinct, less serious offence. There are only single-crime scenarios. All that is left, as the Supreme Court repeatedly insisted in Jogee, is the approach set out in the Standard Test. The core questions post-Jogee are whether S in fact encouraged or assisted the essential matters of the crime committed by P and whether S intended to do so. Variations in P’s offending are relevant only insofar as they alter the type of offence envisaged by S (as recognised in Bainbridge and Maxwell), or go outside of the explicit scope of S’s encouragement (the Calhaem principle).
Circumnavigating these core questions, the Court of Appeal has created a foreseeability-based mode of homicide liability that goes entirely against the intention-focused approach left in the wake of Jogee. One might note that this mode of homicide liability is less harsh than its predecessor, insofar as it will impose liability for manslaughter (with its discretionary life sentence) and not murder (with its mandatory life sentence). But that is little consolation when, on the Standard Test, there should be no homicide liability at all.
As the Supreme Court noted in Jogee itself, where an apex court creates a problem with the common law, it has a responsibility to fix it.Footnote 104 This is particularly so when a corrective measure is needed in order to reduce the potential for criminal liability. The Supreme Court should revisit this area and begin the process of resolving the problems left in the wake of Jogee.
Open access
