II. The Sixteenth Century

The sixteenth-century law on accessorial liability was undeveloped,Footnote ¹⁵ but an early version of PAL is detectable.Footnote ¹⁶ In R. v Salisbury,Footnote ¹⁷ a servant joined in a fight involving his master (Offence A). During the fight, the master killed another combatant (Offence B).Footnote ¹⁸ The servant was ignorant of his master's pre-existing plan to commit murder.Footnote ¹⁹ The servant was held liable for manslaughter, because he was ignorant of the murderous plan, and also lacked the “malice prepense”Footnote ²⁰ required for murder. This decision was generous.Footnote ²¹ In the mid-sixteenth century, the sole question appears to have been whether the collateral offending – even if unforeseeable – occurred in pursuit of the parties’ common unlawful purpose. For example, in R. v Herbert,Footnote ²² a group was assembled by Herbert to steal from and fight Mansell, but apparently did not intend to kill anyone. A disturbance ensued, and a stone thrown by a member of Herbert's group (aimed at someone else) killed Mansell's sister. Half of the judgesFootnote ²³ thought that Herbert's group was liable for her murder, given its connection to their initial unlawful purpose. In essence, “if a man takes the risk of doing an unlawful act, he must take all the consequences of that act even though he had neither intended nor foreseen them”.Footnote ²⁴ For the remaining judges, the group was liable for murder only if it was proved that Mansell's sister had come in defence of her brother, binding the murder to the group's initial unlawful purpose. If Mansell's sister had not sided with either party, then her death was a separate incident, and Herbert's group could be liable, at most, for manslaughter. The evidence could not resolve this factual issue, but the law was clear.Footnote ²⁵ Importantly, it is implicit in the approach of both groups of judges that, had the initial purpose of Herbert's group been lawful, there would not have been liability for murder or manslaughter, at least without evidence of actual encouragement or assistance being given to the stone thrower.Footnote ²⁶

The judges’ approach in Herbert is consistent with PAL. A party to a common unlawful purpose could be liable for a collateral offence despite not intentionally assisting or encouraging it – the distinguishing mark of PAL. The reaches of this early PAL doctrine are, nevertheless, unclear. In murder cases, the felony murder rule removed any need for proof of fault with regard to killing.Footnote ²⁷ Even if no felony were commanded, however, a collateral murder would be the responsibility of all parties to the common purpose.Footnote ²⁸ Furthermore, although those who commanded “beatings” were responsible for collateral murders, an intention to beat was sufficient mens rea for murder at the time.Footnote ²⁹ Examples not involving murder are thus more useful.

One rare example of a (hypothetical) case not involving murder is provided in Plowden's commentary on R. v Saunders and Archer.Footnote ³⁰ For Plowden, it was “reasonable” to hold secondary parties liable for what “follows from [the secondary party's procurement of the principal's offence], but not from any other distinct thing”.Footnote ³¹ Collateral crimes were not always “distinct things” – an approach that stretched beyond murder. For instance, if a person commanded robbery (Offence A), and more severe violence was used against the victim (Offence B), the commander was liable for both offences.Footnote ³² No explanation regarding the connection between the offences is given, so no intention to encourage Offence B is mentioned.

Plowden's “distinct things” lacked a “connection or affinity” with the secondary party's command.Footnote ³³ For instance, if the defendant specified that V1’s house should be burned and the principal deliberately burned down V2’s house, the defendant would not be a secondary party to the property damage. The principal's offence would be “another distinct thing, to which [the commander] gave no assent nor command”.Footnote ³⁴ Similarly, if the parties’ original unlawful purpose was accomplished, and the collateral offence occurred afterwards, distinct evidence of assistance or encouragement in relation to the collateral offence had to be proved.Footnote ³⁵

It is not clear how far beyond commanding (counselling and procuring, in modern terms)Footnote ³⁶ Plowden's comments went, or what the “assent” he referred to involved. These points do not harm the thesis that PAL existed in the sixteenth century. It did not appear necessary for there to be, even in homicide cases, a common purpose to additionally resist by force opposition to the plan, although brief references to such resolutions appear.Footnote ³⁷ As will be seen below, the common purpose to resist opposition was to reappear at various points in history, and was capitalised upon in Jogee.

III. The Seventeenth Century

The focus on collateral offence committed in pursuit of the parties’ common purpose was maintained, as far as the reported sources suggest, in the early-to-mid-seventeenth-century authorities – beyond, it appears, procurement cases “when several men joyn in an unlawful act they are all guilty of whatever happens upon it… . For a man must take heed how he joineth in any unlawful act as fighting is, for if he doth, he is guilty of all that follows”.Footnote ³⁸

The law was, however, beginning to narrow. A useful contrast can be drawn between two hypothetical cases addressed by the Court of King's Bench in 1692:

Here, there is no common unlawful purpose, and so even the nascent PAL doctrine recognised in the sixteenth century would be inapplicable: actual assistance or encouragement of the murder (presumably provided intentionally) would be required in such a case:

The answer to question 5 is reminiscent of the two-offence analysis employed in modern PAL cases. A is intentionally encouraging B only to beat C (Offence A), yet he is liable for C’s murder (Offence B).Footnote ⁴⁰ An intention to beat was, however, sufficient fault for murder at the time,Footnote ⁴¹ and other seventeenth-century authorities suggest that ignorance of the principal's plan to kill might (as in Salisbury) relieve the defendant of secondary liability for murder.Footnote ⁴² Less instructive still is the alternative example where A intentionally encouraged B to commit a felony, because of the felony murder rule.

It is worth noting, before proceeding, that the court viewed question 5 as being an example of aiding and abetting; the encouragement for Offence B came from supporting the principal's plan to commit Offence A. The long-running dispute about whether PAL was truly a form of aiding and abetting, or a distinct head of liability, does not appear to have got off the ground by this point in history. It will, however, be returned to at various points below.

Hale, writing in the 1670s (albeit his work was not published until the 1730s), also includes examples structurally reminiscent of PAL, but restricts secondary liability – at least in procurement and counselling (accessory before the fact) cases – to “objectively” probable results of the initial common purpose. First, “If A. command B. to beat C., and he beat him so that he die thereof, it is murder in B. and A”.Footnote ⁴³ No explicit connection was required between the beating and the death. By contrast:

The “compass of [the parties’] original intention”Footnote ⁴⁵ could stretch only so far, and its outer limit was, at least in relation to accessories before the fact (counsellors and procurers), probable collateral offending.Footnote ⁴⁶

Importantly, given later developments, Hale maintains that, for principals in the second degree (aiders and abettors), certain collateral offences were presumed to be within the parties’ common purpose to commit felonies: if a group embarks on a plan to steal deer:

If the group's joint purpose was lawful, however, secondary liability for murder required proof it was actually encouraged or assisted.Footnote ⁴⁸ Embarking on Offence A was itself, then, the basis for a presumption that opposition to Offence A would be met with force sufficient to ground a conviction for murder (Offence B) if death resulted. In other words, Hale's writing about the law in the 1670s is compatible with the existence of PAL, even if a distinction between accessories before the fact (probability test) and principals in the second degree (the scope of the common purpose, with – at least in felonies – additional presumed force, test) was visible.

The law's approach continued to narrow during the eighteenth century.

IV. The Eighteenth Century

In R. v Plummer, Holt C.J. said that, if a group of smugglers were to be liable for the murder of an officer committed by one of their number:

The implication is that no knowledge of the principal's further offence was necessary if it was committed in pursuance of a common unlawful purpose. The (in fact) collateral offence was only (in law) a collateral offence if it was not done in pursuance of the initial unlawful purpose. This inclusion of what were, in fact, collateral offences within the parties’ “common purpose” was an intellectual dishonesty that was to dog judgments until Chan Wing-Siu.

In R. v Ashton, Holt C.J. gave an additional example: “Two, three or more, are doing an unlawful act, as abusing the passers-by in a street or highway, if one of them kill a passer-by it is murder by all.”Footnote ⁵⁰ There is no mention of the need for a common plan to commit murder, or even contemplation that murder might happen. Furthermore, there is no felony. There is a simple connection between the unlawful abuse (Offence A) and the murder (Offence B) that flows from it. This is, structurally, reminiscent of PAL. A similar, two-step analysis can be applied to R. v Wallis, where it was suggested that “If a man begins a riot… and the same riot continue, and an officer is killed, he that began the riot… is a… murderer; though he did not do the fact”.Footnote ⁵¹ Again, liability for Offence B (murder) was imposed on the basis that it was an incident of the common purpose to commit Offence A (affray/rioting).

Assuming the initial common purpose of the parties was illegal,Footnote ⁵² it was only if murder was not connected to Offence A – as in R. v Hodgson,Footnote ⁵³ where a boy, unconnected with an affray, was killed – that liability for Offence B was ruled out in the absence of explicit assistance or encouragement. Holt C.J. and Pollexfen C.J. thought that the group in Hodgson was to be convicted of murder partially on the basis that they had armed themselves with offensive weapons.Footnote ⁵⁴ Holt C.J. delivered the opinions in Plummer and Ashton, so it is plausible that the weapons were relevant to establishing the probable dangers posed by the original unlawful purpose, suggesting that the killing was committed in furtherance of that purpose, and was not a distinct act.

The law's approach thus does not appear to have moved on a great deal by the 1730s. Contemporary secondary literature does, however, suggest that the probability of collateral offending remained important, at least sometimes. FosterFootnote ⁵⁵ corroborates Hale's comments about actions in pursuit of criminal ventures in cases of counselling and procuring: if “A. adviseth B. to rob C., he doth rob him, and in so doing either upon resistance made, or to conceal the fact, or upon any other motive operating at the time of the robbery, killeth him. A. is accessory to this murder”.Footnote ⁵⁶ He explains that “The advice, solicitation, or orders in substance were pursued, and were extremely flagitious on the part of A. The events, though possibly falling beyond his original intention, were in the ordinary course of things the probable consequences of what B. did under the influence, and at the instigation of A”.Footnote ⁵⁷ In cases involving counsellors and procurers, the focus was thus on what was probable, rather than intended. It must be accepted that, when Foster was writing, probability and intention would have been conceptually closer than they are nowadays, yet Foster distinguishes intention from probability. This unsettles the assumption in Jogee that a “subjective” version of Foster's probability test would be the modern concept of intention.Footnote ⁵⁸ The “subjective” equivalent of “objective” probability (foreseeability) is presumably “subjective” foresight of the relevant probability, which is (almost) what Chan Wing-Siu endorsed.

The probability test certainly softened the law's approach: a bare felony (robbery) was not enough to affix liability for murder – death had to be a probable result. This perhaps explains an awkward aspect of Foster's account.Footnote ⁵⁹ He cites a 1697 case, where three soldiers went to steal apples.Footnote ⁶⁰ One soldier was confronted by the orchard owner's son and murdered him. The other soldiers were acquitted of murder. They were engaged in “small inconsiderate trespass”, and it was not clear that they had a “general resolution against all opposers”.Footnote ⁶¹ This result might appear puzzling: Offence B (murder) seems to have been committed in pursuit of Offence A (stealing apples), and that was enough – it appears – to found secondary liability in the seventeenth century. But the answer to this quandary might be probability – if that test was to be applied beyond accessories before the fact. In the soldiers’ case, murder was not a probable consequence, given the “minor” nature of the common purpose to steal apples. This analysis (which is, it is submitted, less ahistorical than an alternative one based on “conditional intention”) could be used to explain away the orchard example, and indeed it later was.Footnote ⁶²

By the end of the eighteenth century, then, the question of whether Offence B was a probable consequence of the common intent to commit Offence A had begun to assume importance, most clearly in cases of accessories before the fact, but perhaps also in cases of principals in the second degree. Admittedly, what the parties knew (e.g. whether the principal (in the first degree) was armed or not) or had planned was relevant to the probability assessment.Footnote ⁶³ The Supreme Court/Privy Council agreed with this general picture in Jogee,Footnote ⁶⁴ but alleged that things changed markedly in the nineteenth century.

V. The Nineteenth Century

In Jogee, it was contended that “Cases in the nineteenth century showed … a significant change of approach. It was no longer sufficient … to prove that the principal's conduct was a probable consequence, in the ordinary course of things, of the criminal enterprise … The prosecution had to prove that it was part of their common purpose, should the occasion arise”.Footnote ⁶⁵

The Supreme Court/Privy Council is right about the intellectual muddle the courts created in the nineteenth century (finding the collateral offence to be part of the parties’ “common purpose” when it was, in fact, a departure from it). That conclusion does not, however, ground a requirement that the defendant had to intend to encourage or assist every one of the principal's offences if he was to be a secondary party to them. This becomes clear when a full view of the cases is taken.

East suggests that the presumption of a resolution to resist opposition when a felony was embarked upon remained in 1803,Footnote ⁶⁶ but also mentions the secondary party's “contemplation” of the principal's crime(s).Footnote ⁶⁷ It is not clear what relevance such “contemplation” had to determining the liability of parties – or indeed what “contemplation” was taken by East to entail – but it is significant that, even by 1803, the language of “contemplation” was used in relation to secondary liability. As demonstrated below, “contemplation” became very important in the early twentieth century, and was important to Chan Wing-Siu itself.

During the nineteenth century, the presumption regarding resistance gave way to the necessity of proof that the parties’ “common purpose” included such contingencies,Footnote ⁶⁸ but caution must be exercised when pondering whether this required proof that the parties actually shared a common intention. Some authorities might support this jump. In R. v White, it was held that “if the prisoners came with the same illegal purpose, and both determined to resist, the act of one would fix guilt upon both…”.Footnote ⁶⁹ Rather than presuming such a determination existed, the court found that White's running away as soon as the alarm was raised secured his acquittal for the principal's violent crime. The Supreme Court/Privy Council is thus right insofar as it does appear that the parties would need to be proved to have “determined” to resist opposition with force. But it remains something of a leap from the common “determination” mentioned in White to a discrete intention to assist or encourage all of the principal's offences. It is not certain how far White departed from the earlier probability model. Lawyers at the time still presumed natural and probable consequences of actions to be “intended”.Footnote ⁷⁰ It would thus be unwise to read statements about “determination” to be broadly synonymous with intention, and then use that to support a thesis involving the modern understanding of intention.

The concept of a common resolution to resist opposition was opaque. The first edition of Russell on Crime (1819) states that principals in the second degree (aiders and abettors) must have “a general resolution against all opposers … whether such resolution appears … to have been actually and explicitly entered into by the confederates, or may be reasonably collected from their number, arms, or behaviour, at or before the scene of the action”.Footnote ⁷¹ The “resolution” to resist opposition could thus be tacit, but it is unclear what such a “resolution” amounted to. If this “resolution” was a synonym for intention, it is noteworthy that the word “resolution” remained unchanged in the text by 1950,Footnote ⁷² when the law's approach to fault elements (and principally intention) had begun to solidify.Footnote ⁷³ Significantly, when assessing the Supreme Court/Privy Council's historical view, the common purpose scenario was juxtaposed in the 1819 edition with liability on the basis of having “actually aided and abetted him in the fact”,Footnote ⁷⁴ suggesting that PAL was becoming at least a special variety of aiding and abetting, if not a discrete doctrine, by the early nineteenth century.

The cases were similarly unclear about how a common “resolution” to resist opposition was proved. For instance, in R. v Hawkins,Footnote ⁷⁵ some poachers beat up a gamekeeper. The principal returned later and robbed the gamekeeper. The common purpose was assumed to be “to kill game, and perhaps to resist the keepers”.Footnote ⁷⁶ The use of the word perhaps suggests that, even when violence was actually used, the courts might not find that a common purpose to use it had existed, but this idea is not explored further.

Aside from concerns over what a common “resolution” to resist opposition involved, and how it was to be proved, it was not always insisted upon. In Redford v Birley and Others, Holroyd J. noted that, if a group's purpose was lawful, it would not be liable for the principal's crimes unless actual aiding and abetting were proved, but “If persons go together, go united in an unlawful design, to commit a felony, or a breach of the peace, and, in the course of effecting that purpose any one does an act in pursuit of the common purpose, they are all answerable because that which they set about, upon a common design, was originally unlawful”.Footnote ⁷⁷

The probability of Offence B's occurrence is not noted in Reford v Birley and Others (perhaps suggesting that approach was still restricted to accessories before the fact), never mind the resolution to resist opposition.Footnote ⁷⁸ Holroyd J.’s statement would not look out of place in the early seventeenth century.

The law's approach to most issues at the time was inconsistentFootnote ⁷⁹ and so discrepancies are not unexpected. The suggestion in Jogee is that the law began to solidify in R. v Collison (1831):

Collison is similar, factually, to Foster's orchard example, and suffers from the same difficulty concerning the probability of murder being committed (if that test had begun to filter into aiding and abetting). Garrow B.’s judgment is, however, supportive of the Supreme Court/Privy Council's historical thesis that the common purpose, in murder cases, had to include (if conditionally) “severe violence”.Footnote ⁸¹ It is not clear, though, what Garrow B. meant by “entertained the common guilty purpose”, and what evidence would have established the necessary “concert”. The Supreme Court/Privy Council's reading is that an intention to assist or encourage severe violence would be required. But “entertaining” the purpose of committing a collateral offence could equally have meant contemplating the risk of a member of the enterprise needing to use force in pursuit of the common purpose, and continuing regardless. That reading of Collison would be consistent with the modern cases on PAL, such as Chan Wing-Siu, insofar as mere contemplation of Offence B does not entail an intention to encourage or assist Offence B (as the Supreme Court/Privy Council noted correctly).

Other statements from that period are similarly opaque. In R. v Duffey,Footnote ⁸² decided a year before Collison, an important question was whether the secondary party was “cognizant” of the principal's collateral offence and “concurring” in it.Footnote ⁸³ It was not clear what the relevant “cognition” and “concurrence” required.

Collison’s resolution to use force appears in other nineteenth-century cases, but the matter of what could establish it remained unclear.Footnote ⁸⁴ In R. v Scotton,Footnote ⁸⁵ the defendants (poachers) were not liable as the principal was alone at the time of the shooting, and there was no evidence that they had intentionally helped or encouraged him to shoot. This decision might support a necessary requirement of help or encouragement, provided intentionally – supporting the historical argument in Jogee. The judgment in Scotton is, however, short and the relevance of the alleged secondary parties’ absence might have been that the original common purpose of the parties to poach was exhausted, rendering the principal's shooting a gamekeeper his responsibility only. That principle can exist alongside a form of PAL, which simply imposed liability for unintended collateral offences committed in pursuit of the original unlawful purpose.

R. v Macklin and Murphy Footnote ⁸⁶ suggests the Collison approach was, anyway, generous to defendants:

There is no requirement, here, to show that a collateral offence was more than a probable consequence (assuming that deliberate, as opposed to panicked, shooting is not a probable consequence of frightening) of the execution of the common purpose. All that is clear is that an action completely divorced from the initial “common intent” was not something to which secondary liability would attach. If one is not distracted by the word “intent” (still used loosely at the time), this view is consistent with the older authorities, and the existence of a (harsh) PAL doctrine.

Similarly, in R. v Howell, Littledale J. told the jury that “[A]ll those who assemble themselves together with a felonious intent, the execution thereof causes either the felony intended or any other to be committed, or with intent to commit a trespass, the execution whereof causes a felony to be committed, and continuing together abetting one another until they have actually put their design into execution [were liable].”Footnote ⁸⁸

There is, here, no requirement of a common design beyond the first felony, or “trespass” (misdemeanor). Littledale J. continued that “It appears … that there are cases in our law, where persons setting out engaged in a particular object, and in promotion of that object a felony was committed, though not originally intended, and where death ensued… all have been found guilty of murder or manslaughter”.Footnote ⁸⁹ If what marks PAL out is the fact that the secondary party is held liable for Offence B, without the need to have intended to encourage or assist Offence B, then Littledale J. appears to be explaining PAL in 1839. (Indeed, the felony murder rule cannot be used as an explanation here, as the common purpose seems to have led to a separate felony which led to death.) Other cases from that period can be analysed similarly.Footnote ⁹⁰ If Collison did in fact require an actual intention to assist or encourage the collateral offence, then this requirement was applied inconsistently.Footnote ⁹¹

Bramwell B.’s direction to the jury in R. v Jackson raises further doubts about the necessity of an intention to encourage or assist Offence B:

In short, the probability of violence being used was relevant not to an intention to encourage or assist it, but simply in working out whether it was committed in pursuit of the original common purpose.

Further support for this view comes from R. v Harrington,Footnote ⁹³ where Martin B. simply asked the jury if the death had occurred in pursuit of the common purpose to commit a breach of the peace and assault. Harrington was cited in the 1896 edition of Russell on Crime as an example of when a common purpose to resist oppressors was established,Footnote ⁹⁴ when there is no mention in the report of any purpose to do more than breach the peace and commit an assault. It is at least possible, then, that the common “resolution” was, in fact, determined by something other than (so to speak) actual intention, perhaps even the probability assessment that was already employed in relation to accessories before the fact. This was muddled by talk of probable consequences being within a “common purpose”, which meant some collateral (i.e. unintended) offences were artificially brought within a “common purpose” that did not exist factually. This confusion is not unexpected. Intention would have been presumed on the basis of probable consequences of the defendant's acts, and so references to “resolutions” or “intentions” must be read in that light. It was certainly not clear, by the mid-nineteenth century, that anything more demanding than contemplation of the principal's probable collateral crime was required.Footnote ⁹⁵ Even contemplation might have been unnecessary, if collateral offending was probable.

The second half of the nineteenth century, following the passing of the Accessories and Abettors Act 1861, saw little clarity added. In R. v Franz,Footnote ⁹⁶ the jury was directed that it had to be sure that, in light of all the evidence, a murder was committed to enable the planned burglary, or any other felony, before the parties were all liable for it as secondary parties. The reporters record that the doctrine that simply being involved in the burglary was enough to secure liability for additional offences was “long since exploded”.Footnote ⁹⁷ But that doctrine, influenced strongly by the harshest incarnation of the felony murder rule, had since been supplemented by ones asking if the collateral offence was in pursuit of the common purpose and, at least in the cases of counselling and procuring, one asking whether the collateral offending was a probable outcome of pursuing the common purpose. Even if the reporters’ posited “active or passive” participation in the collateral crime was meant as an additional ingredient of liability, it is not clear what it required. Could it be satisfied by contemplation of the need to cause such injury, as some previous cases can be read to suggest? Finally, it will be noted that the “resolution” in Franz (to inflict injury) is less demanding than the one in Collison (to use “extreme violence” or kill).Footnote ⁹⁸ A two-stage analysis of Franz in terms of Offence A (injury) and Offence B (murder) is thus possible, suggesting it was not necessarily the burglary doing the work in establishing liability for Offence B.

Other authorities are harder to explain. For example, the footnoted commentary to R. v Luck explains that the law's approach was more “humane” by 1862, because there had to be “a common design to commit a felony, and a felony homicidal in nature and likely to lead to homicide”.Footnote ⁹⁹ Similarly, in R. v Turner, Channell B. held that there must be a common purpose to use “murderous violence” before all parties could be liable for murder.Footnote ¹⁰⁰ There is still no indication, however, of how these cases built on the approach developed in the eighteenth century. The footnote in Luck speaks, for example, of felonies “likely to lead to homicide”, suggesting that a probability focus was beginning to seep into cases involving principals in the second degree. Turner is harder to square with a probability-based analysis, but the matter of how a common purpose to use “murderous violence” was to be proved was unaddressed. All that was clear was that, if the common purpose was to beat the victim, but the principal produced and used a knife, which the other parties did not know about, only the principal was liable for murder.Footnote ¹⁰¹ That analysis is consistent, however, with a probability-based account of PAL: murder was less likely to flow from a beating than from a knife attack.

Later cases are, admittedly, difficult to fit into the probability assessment model of PAL, suggesting it might have remained limited to counselling and procuring. In R. v Skeet, Pollock C.B. noted that there could be secondary liability “where all the parties were aware that deadly weapons are taken with a view to inflict death or commit felonious violence, if resistance is offered”.Footnote ¹⁰² “Awareness” is, however, a loose term.Footnote ¹⁰³ A person is not only aware of a possible outcome if she intends that possible outcome (I can be aware that someone might die if I x simply by believing that to be a possible, if unlikely, result of x-ing). This casts doubt on the claim in Jogee that a discrete intention to assist or encourage the principal's collateral offending was required. Mere awareness was apparently sufficient to establish a “felonious design to carry out the unlawful purpose at all hazards, and whatever may be the consequences”,Footnote ¹⁰⁴ suggesting that the “common design/intention” was (still) a construct, and its limits were set sometimes by probability and other times by contemplation.Footnote ¹⁰⁵

The way in which Skeet was interpreted in contemporary secondary sources is instructive. For example, the 1867 edition of Archbold states that “it is not sufficient that the common purpose is merely unlawful; it must either be felonious, or if it be to commit a misdemeanor, then there must be evidence to show that the parties engaged intended to carry it out at all hazards”.Footnote ¹⁰⁶ The use of “either” here suggests that, if Offence A was a felony, then liability for murder could flow in the absence of a common resolve to resist at all costs (perhaps a simple application of felony murder). There was thus no universal requirement of intention to assist or encourage collateral offending. If this was an erroneous statement of the law, it was not corrected, even by the time of the 26th edition, published in 1922.Footnote ¹⁰⁷

The Supreme Court/Privy Council based its account of the nineteenth-century law on only five of the above-mentioned cases (Collison, Macklin, Luck, Turner and Skeet). Enough has been done to suggest that these cases were “exceptional”,Footnote ¹⁰⁸ “progressive”Footnote ¹⁰⁹ statements of the law, unclear in their precise implications, and unrepresentative of the entirety of contemporary jurisprudence. Indeed, by the close of the nineteenth century, it appeared that probability/foreseeability was becoming a standard test, beyond cases of counselling and procuring. In 1877, for instance, Russell on Crime recorded that a party would be liable for collateral offences that he “ought to have known” would follow from the common purpose.Footnote ¹¹⁰ Although perhaps “editorial kite-flying”,Footnote ¹¹¹ this statement remained in subsequent editions and is found in the 1950 edition, edited by arch “subjectivist” J.W.C. Turner.Footnote ¹¹²

The English attempts at codification also suggest that “objective” probability, rather than intention, was assuming core importance.Footnote ¹¹³ The 1843 and 1846 reports of the Criminal Law Commissioners envisaged liability for collateral offences perpetrated “in pursuance of and in accordance with [the parties’] design”Footnote ¹¹⁴ (but with no indication of what this required, in terms of proof), with a probability analysis remaining limited to counselling and procuring.Footnote ¹¹⁵ By the 1879 Criminal Code (Indictable Offences) Bill, however, this distinction had collapsed: parties were liable for what they “ought to have been known to be a probable consequence of the execution of [their] common intention”.Footnote ¹¹⁶ This was apparently thought to be a statement of the contemporary legal position,Footnote ¹¹⁷ but it is far clearer than the cases at the time were in its confirmation that probability was a general test.

The nineteenth-century colonial codes also focus on probable collateral offendingFootnote ¹¹⁸ and what ought to have been foreseen by the parties as probable.Footnote ¹¹⁹ These codes were also thought to be rationalisations of existing English law,Footnote ¹²⁰ suggesting further that probability was considered to be the overarching test in establishing whether a collateral offence was (at least in legal fiction, if not in fact) included within the parties’ common purpose. Significantly, the codes deal with ordinary accessorial liability separately from common purpose scenarios, rendering the argument that ordinary aiding and abetting could suffice for all scenarios (which is, ultimately, the Supreme Court/Privy Council's thesis) questionable.

It is submitted that, once a proper view is taken of the nineteenth-century authorities, a “subjective” element (but not necessarily intention to encourage or assist all of the principal's offences, at least as the word intention is nowadays understood) was beginning to be insisted upon by some judges, but it was not consistently explained in terms of intention, rather than contemplation. Furthermore, intention was used loosely, and often in conjunction with probability assessments, in the nineteenth century. It is difficult to take much from them, and certainly very difficult, when everything is taken into consideration, to divine a consistent trend of authority requiring an intention to assist or encourage each of the principal's offences.

The next section will demonstrate that the twentieth-century cases suggest a move to greater “subjectivity” (in conjunction with the defendant's ability to give evidence at trial), but not necessarily towards intentional encouragement or assistance of each of the principal's offences. It will be argued that the move was more clearly towards “subjective” foresight of collateral offences, culminating in Chan Wing-Siu.

VI. The Twentieth Century (to Chan Wing-Siu)

As in the nineteenth century, most early twentieth-century cases were vague on the requirements of secondary liability.Footnote ¹²¹ The law's development in common purpose scenarios remained “erratic”.Footnote ¹²² Far from being directed precisely on the parties’ “subjective” intentions, juries were still told to establish the scope of the parties’ “unlawful design” with little (recorded) guidance on how to do so.Footnote ¹²³

Where guidance was given to juries about collateral offending, it does not necessarily support the Supreme Court/Privy Council's view of history. In R. v Pridmore, the trial judge directed the jury that the defendant was liable for collateral offences that “naturally follow”Footnote ¹²⁴ from pursuing an agreement to poach. Phillimore J. disagreed, and suggested that there needed to be a common purpose that included the intention to resist opposition, rather than merely poach.Footnote ¹²⁵ Yet it remained unclear what was required to establish such a common purpose, and the trial judge's comments about the secondary parties’ “realisation” of the principal's potential collateral offending were not criticised. Ultimately, the way in which the secondary party held his stick (aggressively, it appears) was taken to be evidence enough of the common purpose to use force.Footnote ¹²⁶

To add to the interpretational difficulties, it is unclear from Pridmore how much force the secondary party had to have (implicitly) appreciated was part of the common purpose.Footnote ¹²⁷ It is not clear that the “common purpose” (however established) had to stretch to serious violence or lethal force in order to affix liability for murder, as Collison had suggested. If it did not have to stretch that far, then, structurally, Pridmore can be read consistently with the existence of PAL: intentional encouragement of the use of (some) force (Offence A) could ground liability for a collateral murder that the secondary parties did not intend to encourage or assist (Offence B).

Pridmore can be compared with R. v Pearce,Footnote ¹²⁸ where the jury was told it would have to find an “arrangement” to use force if disturbed, and the existence of a common purpose “not merely to poach, but also to resist apprehension at all costs, even by violence if necessary”,Footnote ¹²⁹ if it was to convict. The defendant had run away when his group was disturbed, and there was a concern that the trial judge had “misled” the jury to think that “the mere fact that [the parties] were out poaching together by night justified them in convicting both of any assault committed by one”.Footnote ¹³⁰ But the earlier authorities are clear that secondary parties can withdraw from a criminal venture and thereby relieve themselves of further liability.Footnote ¹³¹ Pearce had withdrawn. Furthermore, it is simply not clear what exactly would establish Pearce's liability,Footnote ¹³² and the presumption of intention from natural and probable consequences still applied.Footnote ¹³³ Judges were thus not as careful to distinguish between intention/foresight and probability/foreseeability. Judicial comments must therefore be read cautiously.

Contemporary textbooks fail to resolve matters, even if they broadly support the need for a common purpose to resist opposition with force in (at least some) murder cases. For instance, the 1926 edition of Kenny's Outlines states that

No relevant cases are cited in support of this proposition, and there is no discussion of what this common resolution involved in terms of proof. Other secondary sources from the 1920s suggest that contemplation of the collateral offence was vital in such cases involving aiders and abettors, yet do not say that this was mere evidence that the secondary party had intentionally encouraged or assisted the collateral offenceFootnote ¹³⁵ (which is what the Supreme Court/Privy Council suggested that it was).Footnote ¹³⁶

Despite these moves in the direction of “subjectivity”, traces of the “objective”, probability-based approach occurred relatively far into the twentieth century in cases involving principals in the second degree (aiders and abettors). In R. v Betts and Ridley,Footnote ¹³⁷ for example, the parties agreed to rob the victim (Offence A). The victim was killed during the robbery (Offence B). Ridley claimed that he agreed to a push, not further violence. It was held that Betts had merely altered the level of violence required to reach the agreed criminal objective.Footnote ¹³⁸ Ridley was thus “actually a party and privy to an act which was calculated in the judgment of ordinary people to cause death”Footnote ¹³⁹ and rightly convicted of murder. It is not clear whether this was a simple invocation of felony murder, or an application of a probability-based model of PAL.Footnote ¹⁴⁰ Either way, intentional assistance or encouragement of each of the principal's offences was not insisted upon.Footnote ¹⁴¹ Other judicial statements from around the same time suggest an acceptance that acts beyond what the parties actually intended were at issue in common purpose cases.Footnote ¹⁴² Even though it was (still) said that such acts were part of the common purpose, it was becoming clearer that this was intellectually dishonest.

By the mid-twentieth century, then, not all cases involving homicide rested on a settled common purpose to resist opposition with force, and those that did remained opaque regarding what this involved in terms of proof, and when the felony murder rule would straightforwardly apply. In the 1950 edition of Russell on Crime, edited by evangelical subjectivist Turner,Footnote ¹⁴³ it was still suggested “that the true rule of law is, that where several persons engage in the pursuit of a common unlawful object, and one of them does an act which the others ought to have known was not improbable to happen in the course of pursuing such common unlawful object, all are guilty”.Footnote ¹⁴⁴ If the law had developed by the late nineteenth century to require an actual intention to encourage or assist Offence B (as the Supreme Court/Privy Council suggested), or even contemplation of Offence B, this was not noted over 50 years later by a leading criminal law commentator.Footnote ¹⁴⁵

Even when “subjectivity” appeared to be insisted upon, an intention to assist or encourage the collateral offence was not. For instance, in Davies v DPP, Lord Simonds L.C. noted, obiter, that:

It is not clear whether contemplation was intended by Lord Simonds to be a sufficient condition for liability,Footnote ¹⁴⁷ but it was viewed as an alternative to intention. This casts doubt on the proposition that contemplation was mere evidence of intention, and nothing more, as the Supreme Court/Privy Council suggested in Jogee.Footnote ¹⁴⁸

Although one should hesitate before making grand claims based on vague, pre-1950 authorities and an obiter comment, it was at least possible that, by the mid-1950s, mere contemplation of the collateral crime was beginning to be required to bring an act within the parties’ (constructed) common purpose. On the argument defended in this article, there had never been strong recognition of the need to prove that a secondary party intended to encourage or assist all of the principal's crimes, and so no climb-down to contemplation/foresight occurred.

Indeed, the view that the requisite “subjective” element was only just beginning to form by the 1950s might explain why references to probability persevered. In R. v Grant,Footnote ¹⁴⁹ the defendants plotted to tie up and use light force against the porter at the hotel they were burgling. The principal used extreme violence on the porter, killing him. The parties’ convictions for murder were upheld, but Grant did not agree to the level of violence used. It is not even clear if he contemplated it. The level of violence used was, however, viewed as a probable consequence of the plan to commit burglary and assault the porter.Footnote ¹⁵⁰ Of course, the felony murder rule still applied at that time, but it was abolished in 1957.Footnote ¹⁵¹

In the 1958 edition of Russell on Crime, the statement about what the secondary party ought to have known was replaced by Turner with a reference to what “the evidence shows was within [the secondary party's] contemplation”, though no cases are cited to explain this change in approach.Footnote ¹⁵² The abolition of the felony murder rule would only explain a change of approach in murder cases: a bare felony would certainly no longer be enough (though it had not been treated as such, at least consistently, for some time). Yet it is instructive that Turner did not require that the secondary party intended to encourage or assist Offence B. Williams, by contrast, did note cases such as Collison in 1961, and viewed contrary authorities as unreliable.Footnote ¹⁵³ He nevertheless suggested that a common intention to threaten violence “is equivalent to a common intent to use violence, for the one so easily leads to the other”.Footnote ¹⁵⁴ This savours of recognition that collateral Offence B (violence) was a probable consequence of pursuing the plan to commit Offence A (the threat). It is not clear from Williams's discussion that contemplation of the use of violence was even required.

The cases continued to be largely inconsistent into the 1960s. In R. v Spraggett,Footnote ¹⁵⁵ the Court of Appeal held that a “preconceived intention” to commit a crime of violence as well as the burglary was necessary before a conviction for the collateral offence of murder could follow. The court did not explain what this “preconceived intention” entailed in terms of proof. The presumption of intention from natural and probable consequences still applied at this stage, so it might have been sufficient that the collateral offence was a probable offshoot of the planned crime, or had at least been contemplated as such. Without this gloss, Spraggett is difficult to square with subsequent cases. In R. v Betty,Footnote ¹⁵⁶ for example, the parties agreed to attack the victim, but did not intend to kill him. The trial judge had explained that the defendant could not be convicted of murder as a secondary party unless he contemplated the use of fatal or serious violence and the conviction was upheld. This suggests that contemplation was acceptable as the lowest form of fault required for secondary liability flowing from a common unlawful purpose.Footnote ¹⁵⁷ The defendant was convicted of manslaughter on the basis that he had not envisaged death or serious bodily harm resulting from the common purpose.

There then followed two five-judge decisions from the Court of Appeal. In R. v Smith,Footnote ¹⁵⁸ the court focused on the secondary party's “foresight of consequences”, noting that:

All Smith did, if the argument presented above is accepted, is give support to the developing line of authority that mere probability of collateral offending was no longer enough for PAL, and that contemplation/foresight of the collateral crime was necessary.Footnote ¹⁶⁰ The trial judge's statements regarding intention are – on this view – to be treated with extreme care.Footnote ¹⁶¹ Just because the trial judge said “intention” (still, at that time, presumed from “objective” probability) was required, does not mean that that was the minimum level of fault consistent with secondary liability. The decision was read at the time to suggest that contemplation of the collateral offence was necessary and sufficient.Footnote ¹⁶² If a contemplation/foresight-based rot did set in during the twentieth century, then this happened long before 1984.Footnote ¹⁶³

In the second five-judge decision, R. v Anderson and Morris,Footnote ¹⁶⁴ the court approved the contention of Geoffrey Lane Q.C. (later Lord Lane C.J.) that the scope of the “joint enterprise” was crucial to establishing secondary liability,Footnote ¹⁶⁵ but said nothing about how it was to be established.Footnote ¹⁶⁶ The question is not answered in Anderson and Morris,Footnote ¹⁶⁷ which was mainly about unforeseeable changes to the common plan by the principal – which would relieve the putative secondary party of liability. But unforeseeable collateral offending would not even have fallen within the probability-based version of PAL that had existed (initially perhaps only in relation to counselling and procuring) from around the late seventeenth century, never mind the contemplation/foresight-based version that was developing (if falteringly) by the 1960s. It is noteworthy that emphasis was placed in Anderson and Morris on whether the principal “acted in a way which no party to [the] common design could suspect”.Footnote ¹⁶⁸ Again, it is not clear if suspicion would have been sufficient fault to bring the principal's offending within the responsibility of the secondary party, but it is not ruled out.

Following the demise of the presumption of intention and foresight from probable consequences (October 1967),Footnote ¹⁶⁹ and the abolition of the accessories before the fact/principals in the second degree distinction (January 1968),Footnote ¹⁷⁰ it might be expected that the law's approach would at last become consistent, but it did not. In R. v Lovesay,Footnote ¹⁷¹ for example, liability required a “common design … [to use] whatever force was necessary to achieve the robbers’ object”,Footnote ¹⁷² but it seemed to be enough that the defendant “envisaged” that resistance would need to be overcome.Footnote ¹⁷³ That said, neither the decided cases, nor contemporary secondary literature,Footnote ¹⁷⁴ are crystal clear.

Indeed, some cases do appear to insist on an intention to encourage or assist each of the principal's offences. In R. v Reid, Lawton L.J. held that:

Reid appears to be strong support for the Supreme Court/Privy Council's argument that, prior to 1984, secondary parties had to intend to encourage or assist, and actually encourage or assist, each of the principal's offences.Footnote ¹⁷⁶ The court's decision in Reid was, however, viewed by J.C. Smith at the time as merely being “in accordance with the established law” described in the third edition of Smith and Hogan.Footnote ¹⁷⁷ There, it is said that, if the secondary party expected the principal to use the knife with intent to cause grievous bodily harm, that was murder,Footnote ¹⁷⁸ and the importance of foresight of the principal's crime was emphasised, but not connected (at least explicitly) with a requirement of intention.Footnote ¹⁷⁹ Once again, it is not clear that intentional assistance or encouragement of Offence B was, in the end, required.

Other authorities from around the same period support the view that contemplation or foresight of the principal's crime could be sufficient to secure secondary liability. In R. v Penfold,Footnote ¹⁸⁰ Shaw L.J. noted that the defendants could “hardly fail to contemplate” the necessity of violence, and they had implicitly agreed “to put themselves under the dictates of any arising necessity”.Footnote ¹⁸¹ It remained unclear whether this contemplation was a necessary and sufficient condition of liability. Although an alternative reading of the case is that the secondary parties had accepted that violence was indeed what they were planning (and “conditionally intended” it), contemplation of violence could well have been necessary and sufficient to affix liability. This dubiety is sufficient to cast doubt on the Supreme Court/Privy Council's historical account. Chan Wing-Siu could, on this contemplation-as-sufficient view, simply have confirmed what, by the late 1970s, was the best gloss put on the cases regarding common unlawful purpose.Footnote ¹⁸² The law had narrowed after the sixteenth century's harsh approach, from concentrating on probable crimes committed in pursuit of the common purpose (through the seventeenth to nineteenth centuries – initially in counselling and procuring, and later generally) to contemplated crimes committed in pursuit of the common purpose. In another sense, the law's reach had broadened, in that possible (even if not probable) foreseen collateral crimes were seemingly included.

This question of possibility versus probability was addressed first in Australia. In R. v Johns,Footnote ¹⁸³ Barwick C.J. contended that the parties’ common purpose included “all those contingencies which can be held to have been in the contemplation of the participants, or which in the circumstances ought necessarily to have been in such contemplation”.Footnote ¹⁸⁴ The defence conceded that contemplation of Offence B was enough, but objected to contemplation of a possibility being sufficient.Footnote ¹⁸⁵ It was concluded that it was.Footnote ¹⁸⁶ The Supreme Court/Privy Council thought that Johns supported its approach, and was consistent with its reading of Collison.Footnote ¹⁸⁷ Admittedly, the judges’ statements in Johns are inconsistent,Footnote ¹⁸⁸ but that makes claims about it reflecting an “orthodox”Footnote ¹⁸⁹ and “long-standing” position,Footnote ¹⁹⁰ or having a clear ratio,Footnote ¹⁹¹ problematic.

It has been argued that the trend in the twentieth-century cases was towards identifying a “subjective” element beyond the intention to encourage or assist Offence A, but this was not necessarily an intention to encourage or assist Offence B. The Supreme Court/Privy Council bases its contrary thesis on eight twentieth-century cases (Spraggett, Betty, Smith, Anderson and Morris, Reid and Johns).Footnote ¹⁹² Many of these authorities are at best inconclusive, and at worst inconsistent with the intention-based account. Requiring simple contemplation of Offence B is, it is submitted, consistent with a far greater number of authorities, and represents a proper “subjective” equivalent of the “objective” probability focus that dominated earlier discussions of accessorial liability. That, it is submitted, is the best historical account of the law before 1984. Indeed, in the 1983 edition of his Textbook, Williams suggested that “It is now clear that if D1 and D2 set out to rob, and D2 knows that D1 may be carrying a gun and may use it, D2 will be an accomplice in an offence committed by D1 with the gun, if it is in the course of the robbery or of escaping afterwards”.Footnote ¹⁹³ The pivotal question can now, thus, be answered: did the Privy Council take a bold new step in Chan Wing-Siu?

VII. Was Chan Wing-Siu New?

In light of the discussion above, the thesis that the Privy Council's decision in Chan Wing-Siu bucked a clear trend in endorsing intention as the fault element in all instances of secondary liability, and introduced a “new principle”Footnote ¹⁹⁴ that changed “the common law in a way which made it more severe… widening the scope of secondary liability by the introduction of new doctrine”,Footnote ¹⁹⁵ is highly suspect. Although it was claimed in Jogee that Chan Wing-Siu was “based on an incomplete, and in some respects erroneous, reading of the previous case law”,Footnote ¹⁹⁶ the Supreme Court/Privy Council cannot escape a similar charge. The court cited selectively, giving the most sympathetic view of history that it could. On the alternative history argued for in this article, which it is submitted explains more of the decided cases, Chan Wing-Siu (despite the Privy Council citing few cases) simply confirmed what was already becoming clearer in the case law: there was a “wider principle” (i.e. PAL) beyond standard aiding and abetting, “whereby a secondary party is criminally liable for acts by the primary offender of a type which the former foresees but does not necessarily intend”.Footnote ¹⁹⁷ Indeed, the appellants in Chan Wing-Siu do not seem to have argued otherwise, basing their case on the requirement of foresight of a probability of collateral offending, rather than suggesting intention was the relevant standard (although they maintained the intellectual muddle of claiming there to be “tacit agreement” regarding Crime B where in fact there was none).Footnote ¹⁹⁸

The intellectual honesty of Chan Wing-Siu was that it was no longer pretended that the collateral offence was within the common purpose, even conditionally; at last, the law admitted that it was not, yet remained the secondary party's responsibility. Unfortunately, some intellectual dishonesty remained due to the mentions of “authorisation”, “consensus” and “tacit” agreement, where in fact there was bare foresight of the collateral offending.Footnote ¹⁹⁹ But these were unnecessary embellishments. As the Privy Council explained: “The criminal culpability lies in participating in the venture with [foresight of Crime B].”Footnote ²⁰⁰ The House of Lords later clarified this point in Powell and English,Footnote ²⁰¹ where it was confirmed (albeit with some reservations)Footnote ²⁰² that foresight of the collateral offending was necessary and sufficient in PAL cases. Short of being aberrations, then, it is submitted that Chan Wing-Siu and Powell and English were helpful clarifications of the law.

Indeed, had the decision in Chan Wing-Siu been as revolutionary and flawed as the Supreme Court/Privy Council suggested, one might have expected this to be noticed. Yet, in 1984, J.C. Smith described the Privy Council's decision as “a valuable restatement and clarification of … the law”.Footnote ²⁰³ This was not simply a case of the significance of the decision being missed in its immediate aftermath. In 1997, Smith stated that “It would be quite wrong to suppose that parasitic accessorial liability … is a recent development in the law, an innovation by the Privy Council in Chan Wing-Siu”.Footnote ²⁰⁴ Similarly, Spencer, writing in 1985, thought that the Privy Council's decision was consistent with the contemporary English approach,Footnote ²⁰⁵ not a new development.Footnote ²⁰⁶ The 1985 edition of Archbold mentions Chan Wing-Siu only once, without adverse comment.Footnote ²⁰⁷ Although the case was not initially applied universally,Footnote ²⁰⁸ the Court of Appeal endorsed it fairly promptly.Footnote ²⁰⁹

Once the dust had settled, some authors began to question the efficacy of the decision in Chan Wing-Siu.Footnote ²¹⁰ In 1987, Dennis presaged the Supreme Court/Privy Council's decision in Jogee by reading Chan Wing-Siu to be a case about so-called “conditional intention”; otherwise, it was a “dubious” decision, due to the tightening in the mens rea of murder (for principals) after 1984.Footnote ²¹¹ Yet, in his magisterial 1991 work on complicity, K.J.M. Smith thought that Chan Wing-Siu had merely gone “some way to (finally) settling the [fault element of secondary liability for collateral crimes], although distinct traces of confusion are still discernible in the use of consensus terminology”.Footnote ²¹² As noted above, that confusion was removed in Powell and English.

The claim that Chan Wing-Siu was a “wrong turn”, and departing from it was in large part a matter of precedent (and therefore constitutionally straightforward), is thus unconvincing. The more compelling reading of Jogee is that the Supreme Court/Privy Council engaged in substantive law reform. This could have been made explicit, and the decision sold as a continuation of the historical narrowing of PAL. Just as the law had moved from a focus on: (1) furthering the common purpose, to (2) probable collateral offending, to (3) contemplated/foreseen collateral offending, it was now moving to focus on (4) intentional encouragement/assistance of the collateral crime, thus rendering PAL conceptually identical to ordinary aiding and abetting (and redundant). Many would have viewed such a judicial change as desirable, regarding PAL – a judicial development – as the genesis of much injustice.Footnote ²¹³ Being explicit about the change would, however, have meant engaging more directly with the proper process of revisiting previous decisions,Footnote ²¹⁴ and raised more clearly constitutional concerns about judicial activism. Although the Supreme Court/Privy Council is no doubt right that corrections of clear common law “errors” are largely unproblematic constitutionally,Footnote ²¹⁵ there are clearer concerns raised by more substantive reform of even the common lawFootnote ²¹⁶ (as the prosecution maintained in Jogee).Footnote ²¹⁷ The reasons the change in Jogee is problematic, once the alleged precedential “error” has been dismissed as a smokescreen, are as follows: (1) the law as stated (defensibly, as shown above) in Chan Wing-Siu had been relatively settled for over 30 years; (2) requiring intentional encouragement or assistance for all secondary liability was not a reform the Law Commission had proposed when it had considered accessorial liabilityFootnote ²¹⁸ ; and (3) Parliament had not apparently contemplated reforming the law, despite recent encouragement to do so.Footnote ²¹⁹ Once that shield of precedent and history has been shattered, the question is whether the other reasons provided by the Supreme Court/Privy Council in Jogee Footnote ²²⁰ justified such dramatic law reform being undertaken by the courts, not the legislature. It is unfortunate that this question was so easily avoided in Jogee, but this paper opens up the possibility for it to be addressed more straightforwardly in the future. Jogee should be seen for what it is: significant judicial law reform, not common law housekeeping.