Miller v The Queen; Smith v The Queen; Presley v Director of Public Prosecutions (SA)

FRENCH CJ, KIEFEL, BELL, NETTLE AND GORDON JJ. In this matter, the Court is asked to review the doctrine of complicity in the criminal law known as "extended common purpose" or "extended joint criminal enterprise" enunciated in McAuliffe v The Queen and abandon or confine it. Although of general application, the doctrine is commonly invoked, as here, as a means of establishing the secondary offender's liability for murder. In this context, the doctrine holds that a person is guilty of murder where he or she is a party to an agreement to commit a crime and foresees that death or really serious bodily injury might be occasioned by a co-venturer acting with murderous intention and he or she, with that awareness, continues to participate in the agreed criminal enterprise.

The doctrine has been criticised for being inconsistent with the principles of accessorial liability and for being incongruous in light of the mental element of reckless murder. More generally, the criticism is of "over-criminalising": attaching criminal liability to the secondary offender in circumstances in which his or her moral culpability is suggested not to justify that liability. These criticisms were invoked in support of an application to re-open and overrule McAuliffe in Clayton v The Queen. By majority, the Court declined to do so. Among the majority's reasons for that refusal was the observation that principles consistent with McAuliffe form part of the common law in other countries. These principles are commonly traced to the decision of the Privy Council in Chan Wing-Siu v The Queen. Recently, the Supreme Court of the United Kingdom and the Privy Council in R v Jogee; Ruddock v The Queen ("Jogee") held that the common law took a "wrong turn" in Chan Wing-Siu and that there is no place for extended joint criminal enterprise liability in the law. Jogee makes it appropriate to reconsider McAuliffe; however, for the reasons to be given, the principle of extended joint criminal enterprise liability stated in McAuliffe should remain part of the common law of Australia. Liability as a secondary party to a joint criminal enterprise

"Common purpose", "common design", "concert" and "joint criminal enterprise" are expressions variously used in the Australian jurisdictions to describe one means of establishing the complicity of the secondary party in the commission of a crime. The expression "joint criminal enterprise" is used in South Australia and it is the expression that generally will be used in these reasons.

The law, as stated in McAuliffe, is that a joint criminal enterprise comes into being when two or more persons agree to commit a crime. The existence of the agreement need not be express and may be an inference from the parties' conduct. If the crime that is the object of the enterprise is committed while the agreement remains on foot, all the parties to the agreement are equally guilty, regardless of the part that each has played in the conduct that constitutes the actus reus. Each party is also guilty of any other crime ("the incidental crime") committed by a co-venturer that is within the scope of the agreement ("joint criminal enterprise" liability). An incidental crime is within the scope of the agreement if the parties contemplate its commission as a possible incident of the execution of their agreement. Moreover, a party to a joint criminal enterprise who foresees, but does not agree to, the commission of the incidental crime in the course of carrying out the agreement and who, with that awareness, continues to participate in the enterprise is liable for the incidental offence ("extended joint criminal enterprise" liability).

The expression used in Jogee to describe extended joint criminal enterprise liability is "parasitic accessory liability". The expression was coined by Professor JC Smith in an influential article on accessorial liability. Professor JC Smith observed that liability for a crime not intentionally assisted and encouraged by the secondary party, but merely foreseen by him, was not an innovation by the Privy Council in Chan Wing-Siu: the rule imposing liability for offences committed in the course of committing the offence which the secondary party assists or encourages is long-standing. History

In the mid-18th century, Foster described the liability of the accessory before the fact for the incidental crime committed by the principal in this way: "[I]f in the event the felony committed was a probable consequence of what was ordered or advised, the person giving such orders or advice will be an accessary to that felony."

Among Foster's illustrations of the proposition was the following: "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 accessary to this murder."

The statement appeared in the part of the Discourses dealing with accessories before the fact. Nonetheless, it would seem the rule applied equally to aiders and abettors. As Professor JC Smith has observed, it is difficult to suppose that Foster thought a party who would have been liable if he had stayed away would not be liable if he "turned up to bear a hand should he be needed".

In light of modern notions of criminal responsibility, the objective test has been replaced by one that focuses on the subjective state of mind of the secondary participant. The wrong turn in the law that Jogee held the Privy Council took in Chan Wing-Siu was the substitution of foresight as the subjective counterpart to Foster's objective probable consequences test. Their Lordships in Jogee held that the proper subjective element of liability is intention: the secondary party must intend by his or her participation in the joint criminal enterprise to assist the principal to commit the incidental offence. Where that offence requires specific intention, the secondary party must intend that the principal act with that specific intention.

The paradigm case of joint criminal enterprise liability is where the parties agree to commit a robbery and, in the course of carrying out their plan, one of them kills the intended victim with the requisite intention for murder. Applying the principles of joint criminal enterprise liability explained in Johns v The Queen, the secondary party is equally liable if the parties foresaw murder as a possible incident of carrying out the agreed plan. It can be seen that the rejection of foresight as a sufficient mental element would affect the foundation of joint criminal enterprise liability generally in Australian law. Jogee addresses the paradigm case of joint criminal enterprise liability by the adoption of the concept of "conditional intent": the parties may have hoped to carry out their planned robbery without violence but the prosecution must establish it was their intention, in the event the need arose, that a party would administer violence with the intent for murder.

The conclusion in Jogee reflects considerations of policy to which it will be necessary to return. It also reflects an analysis of a line of 19th century decisions which evidenced a shift from the objective probable consequences test, to the requirement that the incidental offence form part of the parties' common purpose should the occasion arise. Of critical importance to their Lordships' reasoning is that Chan Wing-Siu did not refer to two English decisions, R v Smith (Wesley) and Reid, in which it was held that a party to an unlawful attack, in which the principal acts with murderous intent not shared by the party, is guilty of manslaughter and not murder.

Professor KJM Smith traces the history of the development of the doctrine of common purpose in his Modern Treatise on the Law of Criminal Complicity. He observes of the shift from the objective test for liability to some form of subjective requirement that no tolerably clear authoritative principle emerges from the case law. He instances the division in judicial opinion as to the responsibility of an accessory for the violence of a confederate in carrying out a robbery: some decisions hold the accessory liable for any action performed in pursuance of the common purpose, while others require evidence of a common design to execute the common purpose with all necessary force. He suggests that the imposition of liability on the secondary party for the acts of the principal which, although outside the common design, were carried out in its pursuance and were objectively foreseeable was the subject of some controversy in the 19th century. He points to the absence of reference to the objective probable consequences test in the editions of Russell under the editorship of Charles Greaves. It remains that the 1877 edition, under the editorship of Samuel Prentice, concluded the section on common purpose with this summary: "It is submitted 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."

This statement of the law was reproduced in successive editions of Russell until 1958. In 1877, Stephen's Digest of the Criminal Law stated the law, consistently with Foster, in Article 41 under the heading "where crime committed is probable consequence of crime suggested": "If a person instigates another to commit a crime, and the person so instigated commits a crime different from the one which he was instigated to commit, but likely to be caused by such instigation, the instigator is an accessory before the fact."

Stephen's second illustration of the operation of the principle stated in Article 41 was: "A instigates B to rob C, B does so, C resists and B kills C. A is accessory before the fact to the murder of C."

Stephen's statement of the law has been influential in this country. The 1879 Criminal Code (Indictable Offences) Bill (UK), largely his work, proposed the imposition of liability for any offence committed in pursuance of the parties' common purpose which "ought to have been known to be a probable consequence of the prosecution of such common purpose". This statement was taken by Sir Samuel Griffith to reflect the common law at the close of the 19th century when he came to draft the Criminal Code (Q). It remains in s 8 of the Criminal Code (Q) and in the Criminal Codes of Western Australia and Tasmania.

The appellants submit that Stephen's statement of the principles was not a correct reflection of the law as it stood in the latter part of the 19th century. The submission was not developed but may be understood as based on the decisions collected in Jogee which form part of the second of the two streams of judicial opinion identified by Professor KJM Smith. The cases are not easy to reconcile. As late as 1930, there are decisions in England, and in this country, in which the conclusion of liability of a secondary party for murder or infliction of grievous bodily harm with intent committed by the principal in the course of carrying out a planned robbery reveals more than a trace of Foster's objective test. Johns

It was against this background that the Court of Criminal Appeal of the Supreme Court of New South Wales considered the liability of the accessory before the fact to a robbery for the murder of the intended victim of the robbery in R v Johns. Johns was a party to an agreement with two other men, Watson and Dodge, to rob a man named Morriss, who was believed to be a receiver of stolen jewellery. It was Johns' role to drive Watson and Dodge to a location near the planned scene of the robbery and to wait there to collect the proceeds from them and conceal those proceeds at an agreed location. Johns knew that Watson always carried a pistol and that he was quick tempered and would not let Morriss "get on top of him". In the event, there was a struggle and Watson shot and killed Morriss. Watson died before the trial. Dodge and Johns were both convicted of the murder of Morriss.

Johns appealed against his conviction, contending that the doctrine of joint criminal enterprise cast the net too widely in attaching liability to the accessory before the fact. Street CJ acknowledged that, in light of the fundamental alteration to proof of criminal liability effected by Woolmington v Director of Public Prosecutions, a subjective element had come to replace Foster's test. His Honour took the modern law to be correctly stated in the 12th edition of Russell: "Nowadays, it is submitted, the test should be subjective and the person charged as accessory should not be held liable for anything but what he either expressly commanded or realised might be involved in the performance of the project agreed upon. It would, on this principle, therefore be a question of evidence to satisfy the jury that the accused did contemplate the prospect of what the principal has in fact done."

Adopting this analysis, Street CJ stated that the secondary party bears criminal liability: "for an act which was within the contemplation of both himself and the principal in the first degree as an act which might be done in the course of carrying out the primary criminal intention – an act contemplated as a possible incident of the originally planned particular venture."

This statement of the principle was adopted in the joint reasons in Johns v The Queen. As their Honours explained, the act is within the scope of the agreed criminal enterprise because it is within the parties' contemplation and foreseen as a possible incident of its execution.

Jogee describes Johns as an "entirely orthodox" decision. Their Lordships observed there was ample evidence from which the jury could infer that Johns gave his assent to a criminal enterprise which involved the discharge of a firearm should the occasion arise. Nonetheless, there may be discerned a difference in principle between the parties' contemplation of the possible commission of the incidental offence and a requirement of proof of conditional intent that the incidental offence be committed.

In Chan Wing-Siu, the Privy Council took the law of joint criminal enterprise to be as stated by Lord Parker CJ in R v Anderson: where two persons embark on a joint enterprise each is liable for the acts done in pursuance of it, including for "unusual consequences" arising from the execution of the agreement, but if one party goes beyond the tacit agreement, the other parties are not liable for the consequences of that unauthorised act. Sir Robin Cooke, giving the judgment of the Privy Council, noted the absence of analysis of the test that the jury is to apply in deciding whether an act is within the scope of the agreement. Sir Robin adopted the statement of the principle in Johns, which he considered to be in accord with Lord Simonds LC's reference to the secondary party's contemplation of the infliction of mortal injury in Davies v Director of Public Prosecutions.

The prosecution case in Chan Wing-Siu was that three men armed with knives forced their way into an apartment with the intention of robbing the occupants. The male occupant died as the result of stab wounds inflicted by one or more of the men. The evidence of the role played by one of the three, Tse, was less precise than the evidence concerning the other two. All three were tried for murder. One way the prosecution case was put was that the accused must have contemplated the possible commission of the murder in carrying out the robbery. In light of this, Sir Robin said that liability depended on the "wider principle" whereby a secondary party is criminally liable for acts done by the primary offender of a type which he foresees but does not necessarily intend. Sir Robin continued: "That there is such a principle is not in doubt. It turns on contemplation or, putting the same idea in other words, authorisation, which may be express but is more usually implied. It meets the case of a crime foreseen as a possible incident of the common unlawful enterprise. The criminal culpability lies in participating in the venture with that foresight."

The principle stated in Chan Wing-Siu was affirmed by the Privy Council in Hui Chi-Ming v The Queen and by the House of Lords in R v Powell. As noted earlier, putting to one side considerations of policy, the reason for departing from it in Jogee was the neglect, in Chan Wing-Siu, of relevant English authority and, in particular, the decision of the five member Court of Criminal Appeal in Wesley Smith.

Smith, Atkinson and two other men were involved in a brawl in a hotel. It was the prosecution case that the four men were a party to an agreement to "tear up the joint". Smith and a confederate were outside the hotel throwing bricks at its glass door while Atkinson and the fourth man remained inside brawling, in the course of which Atkinson fatally stabbed the barman. All four men were indicted for murder. Two were acquitted and Atkinson and Smith were convicted of manslaughter. The decision is not without difficulty. Smith challenged his conviction on the ground that the fatal knife attack was outside the ambit of the agreement. In rejecting this challenge, Slade J, delivering the judgment of the Court of Criminal Appeal, said that: "It must have been clearly within the contemplation of a man like Smith ... [that Atkinson] might use it [the knife] ... as Atkinson did. By no stretch of imagination, in the opinion of this court, can that be said to be outside the scope of the concerted action in this case."

This passage was cited by Lord Hutton, giving the leading judgment in Powell, as supporting the principle enunciated by the Privy Council in Chan Wing-Siu. In Jogee, it was concluded that Wesley Smith was misinterpreted in Powell. This is because the Court of Criminal Appeal in Wesley Smith also approved the trial judge's directions to the jury, which included the direction, with reference to the secondary party: "Only he who intended that unlawful and grievous bodily harm should be done is guilty of murder. He who intended only that the victim should be unlawfully hit and hurt will be guilty of manslaughter if death results."

It was unnecessary for the Court of Criminal Appeal to resolve any tension between its view that the scope of the enterprise was to be determined by Smith's contemplation of what his co-venturer might do and the jury direction, in circumstances in which their Lordships were not considering the liability of the secondary party for murder. McAuliffe

The question raised in McAuliffe, which was not explicitly raised in Chan Wing-Siu, was whether it was necessary for there to be mutual contemplation of the commission of the incidental crime. It was common ground in McAuliffe that three youths had agreed to bash a person or persons in a park. All three were convicted of the murder of a man whom they had set upon. The jury was directed that the accused whose case it was considering would be guilty of murder if he shared the common intention, with the accused who did the act causing death, of inflicting grievous bodily harm or if he "contemplated that the intentional infliction of grievous bodily harm was a possible incident of the common criminal enterprise". The emphasised part of the direction was the subject of challenge on appeal. The McAuliffe brothers' case on appeal was that joint criminal enterprise liability required the prosecution to prove a shared contemplation that grievous bodily harm might intentionally be inflicted as a possible incident of the agreement to assault.

The Court observed that the challenged direction conveyed to the jury that, even if the criminal enterprise embarked upon by the three youths did not embrace the intentional infliction of grievous bodily harm, there was "a sufficient intent" for murder if the accused contemplated the intentional infliction of grievous bodily harm by one of them as a possible incident in carrying out their agreement and, with that awareness, the accused continued to participate in the enterprise. The accused was as much a party to the incidental crime as when its commission was within the common purpose. Participation in the joint criminal enterprise, with the requisite awareness, was not relevantly distinct from an understanding or arrangement that included foresight of the intentional infliction of grievous bodily harm.

The Court in McAuliffe adopted the rationale for the imposition of extended joint criminal enterprise liability given in Chan Wing-Siu: "As Sir Robin Cooke observed, the criminal culpability lies in the participation in the joint criminal enterprise with the necessary foresight and that is so whether the foresight is that of an individual party or is shared by all parties. That is in accordance with the general principle of the criminal law that a person who intentionally assists in the commission of a crime or encourages its commission may be convicted as a party to it."

As will appear, the second sentence proceeded upon acceptance of Professor JC Smith's analysis that the secondary party "lends" himself to the enterprise thereby giving encouragement to the commission of an offence which he knows may involve committing the incidental offence. It is also to be understood in the context of the following paragraph, in which their Honours, returning to what the challenged direction conveyed, reiterated that individual contemplation of the intentional infliction of grievous bodily harm as a possible incident of the enterprise amounted to "a sufficient intention on the part of either of them for the purpose of murder".

Jogee held that the law stated in Chan Wing-Siu, and the decisions following it, extended the reach of criminal liability too far. It is a conclusion that is in line with the views of a number of distinguished commentators. Their Lordships' conclusions that there is no occasion for extended joint criminal enterprise liability, and that the proper counterpart to Foster's objective probable consequences test is intention, are conclusions about the policy that the law should pursue. They are conclusions which reflect their Lordships' preference for the view of the editors of the 14th edition of Smith and Hogan's Criminal Law that extended joint criminal enterprise does not come within the principles of accessorial liability.

The relationship of joint criminal enterprise and extended joint criminal enterprise to general concepts of complicity is contested. Professor JC Smith saw no difficulty locating both within ordinary principles of accessorial liability. As noted, in the case of extended joint criminal enterprise, the secondary party "lends" himself to the enterprise and in so doing gives assistance and encouragement to the principal in carrying out an enterprise that he knows may involve the incidental offence. The alternative view, proposed by Professor Simester, is that joint criminal enterprise is a sui generis form of secondary participation in a crime and not merely a sub-species of accessorial liability.

In Clayton v The Queen, the joint reasons adopt Professor Simester's analysis distinguishing the liability of the aider and abettor from the liability of a party to a joint criminal enterprise. The wrong in the case of the aider and abettor is grounded in his or her contribution to the principal's crime. The wrong in the case of the party to the joint criminal enterprise lies in the mutual embarkation on a crime with the awareness that the incidental crime may be committed in executing their agreement. Acknowledgement of the sui generis nature of the secondary liability that arises from participation in a joint criminal enterprise may be thought to resolve at least some of the anomalies that are suggested to arise from allowing foresight of the possible commission of the incidental offence by a co-venturer as a sufficient mental element of liability.

The appellants contend that the doctrine of extended joint criminal enterprise diminishes the state of mind necessary for criminal liability. They argue that, if McAuliffe is reversed, the prosecution may have to make more of an effort to establish liability in joint enterprise cases but that to recognise so much is not to justify the imposition of liability without a sound doctrinal foundation. That submission is apt to overlook the difficulty, in the case of group criminal activity, of establishing the contributions made by individual members. Commonly enough the identity of the principal offender will not be known.

In Powell, Lord Steyn identified an important justification for the doctrine grounded in practical considerations: "In the real world proof of an intention sufficient for murder would be well nigh impossible in the vast majority of joint enterprise cases. Moreover, the proposed change in the law must be put in context. The criminal justice system exists to control crime. A prime function of that system must be to deal justly but effectively with those who join with others in criminal enterprises. Experience has shown that joint criminal enterprises only too readily escalate into the commission of greater offences. In order to deal with this important social problem the accessory principle is needed and cannot be abolished or relaxed."

McAuliffe builds on the principles enunciated in Johns. The appellants do not contest that the parties' foresight of the possible commission of the incidental offence suffices for liability for that offence on the principles of joint criminal enterprise liability enunciated in Johns. In cases in which the participants in a joint criminal enterprise acknowledge that an incidental crime is a possible consequence of carrying out their agreement, the commission of the offence is within the scope of the agreement and the parties must be taken to have authorised or assented to its commission even if it is their preference that it be avoided. It is the authorisation or assent which is said to justify the imputation of the acts of the principal to all the participants in the agreement. The wrong turning in the law enunciated in McAuliffe, in the appellants' submission, was the discarding of the concepts of mutuality, authorisation and assent.

The reason for McAuliffe's rejection of the mutuality of foresight of the commission of the incidental offence as the criterion of liability is well illustrated by the example given by Professor JC Smith in his commentary on R v Wakely: A knows that P is carrying a weapon which he will use to kill or cause grievous bodily harm if it is necessary in carrying out the agreed enterprise and A says to P "I do not agree to your using that weapon" but nevertheless A continues to participate in the enterprise. As Professor JC Smith observed, A's words deny tacit assent to the use of the weapon. Moreover, adopting the Jogee analysis, A can hardly be said to have conditionally intended the use of the weapon. It is not self-evident, however, that the policy of the law should be against the imposition of liability for murder in such a case. Certainly A's moral culpability is not less than that of the secondary party in a case such as Johns.

The principles applied to the re-opening of decisions of this Court need not be recited. McAuliffe was a unanimous decision. It has since been affirmed on a number of occasions. Many prosecutions have been conducted on the law stated in it in the Australian common law jurisdictions. Jogee held that the effect of "putting the law right" will not be to invalidate convictions arrived at over many years by faithfully applying the law laid down in Chan Wing-Siu, as leave to appeal out of time would only be granted where the applicant can demonstrate substantial injustice. The position in Australian law in this respect cannot be regarded as settled and it cannot be said that to depart from the law as it has been consistently stated and applied would not occasion inconvenience. Of course, were the law stated in McAuliffe to have led to injustice, any disruption occasioned by departing from it would not provide a good reason not to do so. However, here, as in Clayton, the submissions are in abstract form and do not identify decided cases in which it can be seen that extended joint criminal enterprise liability has occasioned injustice.

In Gillard v The Queen and again in Clayton, this Court rejected arguments that the doctrine of extended joint criminal enterprise should be abandoned or extensively modified. In addition to observing that the doctrine had not been shown to occasion injustice in the application of the law, the joint reasons in Clayton rejected that its application had made criminal trials unduly complex. Moreover, most of the arguments in favour of change had been thoroughly considered and rejected by the House of Lords in Powell. Importantly, in Clayton it was said that no change should be undertaken to the law of extended joint criminal enterprise without examining the whole of the law with respect to secondary liability for crime. As was observed, it would be undesirable to alter the doctrine as it applies to the law of homicide, which is its principal area of application, without consideration of whether the common law of murder should be amended to distinguish between killing with intent to kill and killing with intent to cause really serious injury.

The undesirability of altering the doctrine of extended joint criminal enterprise without examining the law with respect to secondary liability generally is underlined by the report of the Law Commission of England and Wales Inchoate Liability for Assisting and Encouraging Crime, in which a proposal to abolish secondary liability for a collateral offence committed in the course of a joint criminal enterprise was rejected. As emerged from that report, the projected and other possible ramifications of change were such that, if any change were to be made, it should be made by the Parliament.

In the decade since Clayton was decided, the Parliament of Victoria has amended the Crimes Act 1958 (Vic), abolishing the common law of complicity and in its place imposing liability on persons "involved in the commission of an offence". The New South Wales Law Reform Commission undertook a review of the law of complicity. The Commission proposed retention of extended joint criminal enterprise liability along the lines adopted in the Criminal Code (Cth) with a further modification in the case of liability for homicide. In such cases the Commission recommended that the secondary party's foresight be of the probability of the commission of the offence. The Parliament of New South Wales has to date not chosen to act on the Commission's recommendations. The Parliament of South Australia has also not chosen to reform the law as stated in McAuliffe.

In light of this history, it is not appropriate for this Court to now decide to abandon extended joint criminal enterprise liability and require, in the case of joint criminal enterprise liability, proof of intention in line with Jogee. For the same reasons, it is not appropriate to depart from McAuliffe by substituting a requirement of foresight of the probability of the commission of the incidental offence. As Johns explains, the difficulty with such a requirement is that it "stakes everything on the probability or improbability of an act, admittedly contemplated, occurring". This is not to accept the submission that since "anything is possible", the secondary party may bear liability for a crime contemplated by him or her as no more than a fanciful possibility.

In Chan Wing-Siu, the Privy Council considered that there may be a case in which the commission of the incidental offence occurs to the accused fleetingly and is genuinely dismissed by him or her as a negligible risk. It was held that, in such a case, the secondary party would not possess the necessary foresight to support liability. That conclusion should be accepted. Proof of the accused's foresight of the possibility of the commission of the incidental offence usually will be an inference from what the accused is proved to have done and to have known. Cases in which the evidence leaves open that the accused may have contemplated the incidental offence, but dismissed it as a fanciful possibility, are likely to be few. In such a case, it would be necessary to direct the jury that a possibility dismissed as negligible would not suffice for liability.

Before parting with Jogee, it is necessary to say something about the suggestion that the expression "joint enterprise liability" occasions public misunderstanding. The misunderstanding that their Lordships identified was that the expression allows a form of "guilt by association" or "guilt by simple presence without more". Nothing in McAuliffe supports either conclusion. It is to be appreciated that in the paradigm case of murder, the secondary party's foresight is not that in executing the agreed criminal enterprise a person may die or suffer grievous bodily harm – it is that in executing the agreed criminal enterprise a party to it may commit murder. And with that knowledge, the secondary party must continue to participate in the agreed criminal enterprise. The procedural history

Everard Miller, Wayne Smith and Johnas Presley were all convicted of the murder of Clifford Hall following a trial in the Supreme Court of South Australia (Stanley J). The deceased was fatally stabbed by Joshua Betts in the course of an assault to which Miller, Smith and Presley were said to be parties. Liability for the murder of the deceased was left in each case on the basis of either joint criminal enterprise or extended joint criminal enterprise.

Miller, Smith, Presley and Betts were also charged with causing harm to a man named Wayne King with the intention of causing harm. That offence was alleged to have been aggravated by being committed in company and by the use of offensive weapons particularised as a baseball bat and a pole. Presley pleaded guilty upon arraignment to this offence, and the jury returned verdicts of guilty against Miller, Smith and Betts. The jury found, in each case, that the offence was aggravated by having been committed in company. The jury did not find in any case that the offence was aggravated by the use of offensive weapons.

Miller, Smith, Presley and Betts had all been drinking alcohol in the hours leading up to the killing of the deceased and the assault on King. It was open to find that each was intoxicated at the material time. All four appealed unsuccessfully against their convictions to the Court of Criminal Appeal of the Supreme Court of South Australia (Gray, Sulan and Blue JJ). Miller, Smith and Presley each contended, among their grounds of appeal, that the verdicts were unreasonable and could not be supported by the evidence having regard to their states of intoxication.

On 13 November 2015, Keane and Nettle JJ granted Miller special leave to appeal on a ground which contends that the Court of Criminal Appeal erred in holding that his convictions are capable of being supported by the evidence. On 12 February 2016, French CJ and Kiefel J referred applications for special leave made by Smith and Presley to an enlarged Court with a view to those applications being heard with Miller's appeal. The question raised by these applications is also the capacity of the evidence to support the convictions when account is taken of the evidence of the applicants' intoxication.

The judgment in Jogee was delivered on 18 February 2016. Miller applied for leave to amend his grounds of appeal to contend that the trial miscarried as the result of liability for the murder of the deceased being left for the jury's consideration on the basis of extended joint criminal enterprise principles. Smith and Presley applied for leave to amend their proposed grounds of appeal to raise the same point. Leave was granted in each case.

For the reasons given, McAuliffe remains a correct statement of the common law of Australia and it follows that the amended grounds of appeal must be dismissed. Nonetheless, as will appear, the Court of Criminal Appeal did not review the sufficiency of the evidence to sustain the verdicts consistently with the task explained in M v The Queen. Special leave to appeal must be granted to Smith and Presley. The appeals in each case must be allowed and the proceedings in each case must be remitted to the Court of Criminal Appeal for determination of the ground that the verdict is unreasonable. The facts

The appellants and Betts are Aboriginals. They spent much of 12 December 2012 together drinking. In the evening, they were at Presley's home in Hayles Road, Elizabeth Park. Sometime before 11:00pm, Betts and Presley left those premises to obtain some marijuana. On their way back, Betts urinated against the fence of a residential property in Grant Street, Elizabeth Park. The deceased, a resident of Grant Street, remonstrated with them. He was joined by a neighbour, King. There was a heated exchange between the four men in the course of which the deceased made racial slurs. Betts or Presley was pushed by either the deceased or King and Betts was punched to the face.

On their return to the Hayles Road premises, Betts reported that he had been struck by three "white fellas", who had jumped him in the alleyway. His lip was bloodied. Presley was angry. He took hold of a baseball bat and said "let's go back and see what these people – go and see what the problem is". The appellants and Betts left the Hayles Road premises. Presley had the baseball bat. Betts was armed with a 332 mm long knife. One issue at the trial was whether the prosecution had established that the appellants must have been aware of the knife.

The appellants, Betts and, perhaps, others made their way to the scene of the earlier altercation in Grant Street. At least some of the group walked up a laneway that runs between Butterfield Road and Grant Street. The deceased, King and some of their neighbours were still gathered in Grant Street near the entry to the laneway when the group arrived. Estimates of the interval between the initial altercation and the second, fatal altercation varied from between three to 20 minutes. Estimates of the number of Aboriginals who formed part of the hostile group that made its way along the laneway to Grant Street varied from four to eight persons.

The group was described as hitting the fence with objects as they made their way down the lane. One of them, a big Aboriginal man, on the prosecution case Smith, was brandishing a shovel. The deceased, who was in the laneway, called out "run they've got weapons". King, who was also in the laneway, started to run back towards his house. He was struck on the left shoulder by something solid from behind. He turned and saw Betts and Presley. Each was holding an object and each struck him. He raised his arm and was struck by the taller of the two men with an object described as "like a baseball bat". King sustained a fracture to the upper arm, which required surgery. Witnesses saw King being kicked and hit by two men as he lay on the ground. One of the men had a tattoo of a crucifix on his face. This was Betts.

The deceased was struck from behind with the shovel. He fell to the ground, where one witness saw him being kicked to the jaw. The same witness described a "hand coming at him like at the back of him and hitting him" just below the shoulder blade. On the prosecution case, this was the fatal knife blow struck by Betts. The deceased was kicked and hit as he lay on the ground by a number of Aboriginal men. The estimates varied as to the number of his assailants. One was armed with a shovel or shovel-like implement. The deceased sustained a scalping wound to the head which was consistent with the use of a shovel. A witness saw the deceased being struck with a bottle.

The appellants and Betts returned to the Hayles Road premises after the second altercation. Gary Willis, who had been drinking with the group earlier in the evening, recalled one of them saying on their return "We smashed them. We had a fight" and Betts saying "I think I stabbed him, stabbed a bloke in the guts".

Presley was seen by a police officer at the front of the Hayles Road premises shortly after 11:30pm. He was pacing in an agitated state saying "you're fucked, dog" and "you dog". He was arrested at about 12:25am. Betts and Miller were arrested at premises in Northampton Crescent, Elizabeth East at about 1:52am. Smith was arrested asleep in a car outside the Northampton Crescent premises at about 5:15am.

None of the appellants gave evidence at the trial.

In an interview with the police, Betts admitted to stabbing the deceased and said that he had done so in self-defence. Following his arrest, Betts directed police to premises in Butterfield Road, where he showed them the knife, which had been placed in a drain. The blade of the knife was 202 mm long. It was stained with blood matching that of the deceased. A shovel was seized from the rear yard of the Butterfield Road premises. Hair and skin were observed on the leading edge of the blade. DNA profiling matched that of the deceased.

Presley declined to be interviewed on the morning of his arrest. On 22 December 2012, he was interviewed at his request. He said that he obtained a baseball bat after the initial altercation and that he ran, or jogged, back to the scene with Smith, Betts and a man whom he did not know. He said that Smith had a cricket bat, that he thought the man he did not know had a shovel and that Betts had a knife. He admitted to striking one man to the elbow but he said that he did no more than this. He did not see Betts stab the deceased.

The police located an empty Passion Pop bottle at the scene in Grant Street. Blood-like stains were observed on the neck of the bottle. The mouth of the bottle contained DNA consistent with Betts' DNA profile. Smith's fingerprints were found on the bottle. The location of the fingerprints was consistent with the bottle having been held by Smith both upside down and upright.

A baseball bat was seized from the lounge room at the Hayles Road premises. A swab obtained from the handle of the bat contained a mixed DNA profile from four contributors. Betts, Miller and Smith were excluded as contributors to the DNA deposit. Presley was not excluded, although the result did not establish that he was a contributor.

A triangular 1.9 kg piece of concrete was located in Grant Street in close proximity to the deceased. There were two blood-like spots on the edge of the block, which matched the DNA profile of the deceased. A witness described a concrete block as having been thrown from the direction of the laneway towards her during the course of the assault. Intoxication Evidence of Miller's intoxication

At the time of his arrest, Miller was observed by Constable Penn to be "extremely intoxicated by something". He was unsteady on his feet, his speech was slurred, he struggled to keep his eyes open and he smelt of alcohol. Constable Penn kept constant observations on him while Miller was in the holding cell at the Elizabeth Police Station. He noted that Miller was "extremely lethargic and fell asleep in the holding cell". A breath test administered by a police officer recorded an alcohol concentration of 0.167 grams of alcohol per 100 millilitres of blood at the time of Miller's arrest. At about 9:20am on 13 December 2012, a blood sample was taken from Miller. This revealed an alcohol concentration of 0.139 grams of alcohol per 100 millilitres of blood. The blood sample also revealed 0.5 grams per 100 millilitres of blood of diazepam and 0.04 grams per 100 millilitres of blood of nordiazepam. Nordiazepam is the metabolite of diazepam. The combined effect of these drugs was at the lower end of the therapeutic range. Diazepam is a drug of the benzodiazepine family. Among its effects is that it operates as a sedative and muscle relaxant. The sample also revealed three micrograms per litre of blood of THC, the active chemical in cannabis.

Dr Majumder, a pharmacologist, gave evidence in Miller's case. She estimated, based on Miller's blood sample, that at around the time of the second altercation Miller's blood alcohol reading would have been 0.292. The estimate assumed that alcohol was eliminated from Miller's system at 0.015 per cent per hour. The rate of elimination, which varies between individuals, is between 0.01 to 0.02. Taking into account that range, Dr Majumder considered that Miller's blood alcohol concentration at the time of the second altercation was between 0.241 and 0.342.

Dr Majumder was asked about the effect on her calculations if Miller had consumed two standard drinks after the second altercation. She said two standard drinks would not have raised Miller's blood alcohol levels by more than 0.04 per cent on average. She was asked what effects a blood alcohol concentration in the range of 0.272 to 0.322 would have on the mental state and behaviour of the subject. Dr Majumder said that, at these levels, there would be "significant effects on the behaviour and mental state" of the person. She went on to explain that the person would be "very drunk" and that it would be obvious to an observer because the person's speech would be slurred and he or she may have stumbling gait and glazed eyes. Dr Majumder said that levels of 0.272 to 0.322 would be "close to the level that is generally considered very high level". It is a level at which a person can lose consciousness. A person who has not lost consciousness and who has this level of alcohol in the blood is a person with a degree of tolerance to the effects of alcohol.

Dr Majumder said that at the assumed levels of alcohol concentration in Miller's blood, an individual's "thinking process, decisionmaking process will be significantly impaired". The person would have problems concentrating and would have a short attention span. An experienced drinker may be slightly less affected than a non-experienced drinker. At this high level of blood alcohol concentration the person would have "significantly impair[ed] decisionmaking and also planning ... so the person may not be able to foresee or predict the consequences of certain decisions". Dr Majumder explained that alcohol can release aggressive behaviour and cause disinhibition. Its effects may cause the person to act without thinking. At these high levels a person may be too intoxicated to be aggressive.

Dr Majumder considered it possible that the interaction of diazepam and nordiazepam could enhance the effects of alcohol. The impairment of concentration might be more pronounced taking into account the interaction with diazepam. It was possible that this enhanced effect would include the impairment of foresight.

Dr Majumder considered the level of THC in Miller's blood to be a moderate level. She was not able to say whether, at the time of the second altercation, Miller would have been under the influence of cannabis. It was one of the possibilities. The ingestion of cannabis could have potentiated the impairment from alcohol in combination with the diazepam and nordiazepam.

There was evidence upon which it may have been open to find that Miller was present with Smith and Betts at premises in Halsey Road, Elizabeth East after the second altercation, drinking a pre-mixed alcoholic drink described as a "cowboy". Betts, Smith and a third man, who may have been Miller, were at those premises for half an hour or an hour drinking. A witness present at the Halsey Road premises, who was herself drinking, did not consider that either Smith or Betts looked drunk. In the witness' estimate, they "most probably had a few drinks". The witness made no observation of the third man, whose face she did not really see and who had been sitting in the dark.

Dr Majumder was asked to assume that Miller had consumed five standard drinks between 11:30pm and 9:20am the following day when the blood sample was taken. On this assumption, Miller's range of blood alcohol concentration at the time of the second altercation was estimated at between 0.192 and 0.242. If the rate of elimination was calculated as between 0.01 per cent and 0.02 per cent per hour, the range, allowing for the consumption of five standard drinks after 11:00pm, came down to 0.141 to 0.242. At this lower range the effects on the individual would be as described for the higher range but to a lesser degree. Evidence of Presley and Smith's intoxication

The police officers who observed Presley outside the Hayles Road premises at about 11:30pm considered that he was moderately affected by alcohol. A blood sample taken from Presley at 8:28am on 13 December recorded a concentration of 0.054 grams of alcohol per 100 millilitres of blood. If Presley had not consumed alcohol after the second altercation, his blood alcohol level at the time was likely to have been about 0.2. Dr Majumder said that a person with a blood alcohol concentration of the order of 0.2 would have appreciable deficits in the person's perception of events occurring around them and their decision-making processes would be impaired to some degree.

King said that the two Aboriginal men involved in the first altercation were affected by alcohol. Ms Bateman, one of the neighbours in Grant Street, said the man accompanying the one who urinated on the fence was staggering a bit. Two other neighbours described the Aboriginal men involved in the first altercation as being drunk or intoxicated in some way.

In his interview, Presley gave an account of drinking Passion Pop and Jack Daniel's. Willis said that, while it was still light on 12 December 2012, he had driven Miller and Smith from the Elizabeth Tavern to the Hayles Road premises, where they met up with Betts and Presley, who were drinking. The group were drinking West End Draught. Willis recalled Presley going to the Elizabeth Tavern and returning with a bottle of Bundaberg Rum. Willis himself had been drinking for two days. He said all of the men were drunk that evening. A few, including Smith, had been smoking marijuana that night.

Smith was not alcohol breath tested. The only blood sample taken from him was performed about 24 hours after the fatal altercation. This showed a zero blood alcohol concentration. Traces of prescription drugs and cannabis were detected. There was uncontradicted evidence that Smith had been drinking throughout 12 December 2012. It was his case that the blood alcohol concentrations of his co-accused recorded in analysis of samples taken much closer to the event provided cogent circumstantial evidence of his likely state of intoxication at the material time. The zero blood alcohol concentration in his blood, 24 hours after the events, did not cast doubt on a conclusion that he was significantly intoxicated at the time. The Court of Criminal Appeal

Each appellant relied on a number of grounds of challenge to his convictions – in Presley's case, his conviction for murder – before the Court of Criminal Appeal. Each appellant's Notice of Appeal included a ground challenging the trial judge's directions on intoxication. Miller and Smith both abandoned this ground prior to the hearing before the Court of Criminal Appeal. Presley maintained the ground, contending, among other things, that the trial judge failed to instruct the jury as to how its findings respecting his state of intoxication might affect its findings as to the scope of any joint criminal enterprise to which he was a party and whether "he turned his mind to possible consequences of participation". The Court of Criminal Appeal extracted the trial judge's general directions on intoxication and specific directions as this issue applied to Presley. The Court of Criminal Appeal concluded that the summing-up on this topic could not be fairly criticised.

As noted earlier, in common to each appeal was a ground contending that the verdict was unreasonable because it was not supported by the evidence. The ground was largely advanced by reference to the unchallenged evidence of the appellant's intoxication. Determination of this ground in each case required the Court of Criminal Appeal to engage with the evidence and consider for itself the findings that were open as to what the appellant did in connection with the second altercation. The inferences to be drawn from these findings needed to be assessed in the context of the findings that were open as to the appellant's state of intoxication. It was necessary to consider whether the prosecution had excluded the reasonable possibility that, by reason of his intoxication, the appellant had not in fact come to an understanding or arrangement with the others to inflict grievous bodily harm or to assault a person or persons in Grant Street. In the event that the Court of Criminal Appeal was satisfied that it was open to find that the appellant was a party to an agreement at least to assault a person or persons in Grant Street, it remained to consider whether the prosecution had excluded the reasonable possibility that, by reason of his intoxication, the appellant did not in fact foresee that one of his co-venturers might kill or inflict grievous bodily harm on a person or persons in Grant Street intending so to do.

In dealing with each appellant's ground that the verdict was unreasonable, the Court of Criminal Appeal did no more than refer to its summary of the evidence and the way the prosecution had put its case at the trial. In no case did the Court review the evidence as it related to the appellant and address the asserted deficiencies in its capacity to establish the nature, if any, of his participation in the second altercation. Nor did the Court assess the significance of the evidence of the appellant's intoxication to its conclusion. In Presley's appeal, the whole of the Court's reasoning for rejecting the ground that the verdict was unreasonable is: "Earlier in these reasons we have set out the prosecution case against Presley of his presence and participation. We have identified the evidence led in the trial to support this case. In our view, the evidence allowed the jury to conclude that Presley was present and did participate in the attack on Mr Hall and that he did so with the necessary criminal intent."

In Miller's appeal, the whole of the Court's reasoning for rejecting the ground that the verdicts were unreasonable is: "In our view, given the concession that Miller was present at the time of the incident, the other facts were capable of establishing that Miller was acting with the others and was either party to a plan to inflict grievous bodily harm or foresaw that possibility. With respect to the other charge, the other facts were capable of establishing that Miller was party to a plan to assault. We do not consider that any basis has been made out to establish that the verdicts are unreasonable or cannot be supported having regard to the evidence as far as Miller is concerned."

In Smith's appeal, the whole of the Court's reasoning for rejecting the ground that the verdicts were unreasonable is: "We have earlier set out the evidence against Smith. In our view, there was sufficient evidence to leave it open to the jury to find that Smith was present at the scene and participated in the attack with the necessary intent."

Miller and Presley submit that the Court of Criminal Appeal erred in rejecting the ground that the verdicts are unreasonable. They invite this Court to allow their appeals and to substitute verdicts of acquittal. On the hearing in this Court, Smith made a less ambitious submission. He accepted that it is not usually appropriate for this Court to embark on an assessment of the sufficiency of evidence to support a verdict in circumstances in which the Court of Criminal Appeal has not undertaken that task. He submitted that the appropriate order is to remit the proceedings to the Court of Criminal Appeal. That submission should be accepted. The proceedings in each case must be remitted to the Court of Criminal Appeal for determination of the ground which contends that the verdicts in the appeals of Miller and Smith and the verdict in the appeal of Presley are not supported by the evidence. Orders

The following orders should be made: Matter No A28/2015 Appeal allowed. Set aside the order of the Court of Criminal Appeal of the Supreme Court of South Australia made on 28 April 2015. Remit the matter to the Court of Criminal Appeal for determination. Matters No A22/2015 and No A17/2015 Special leave to appeal granted. Appeal treated as instituted and heard instanter and allowed. Set aside the order of the Court of Criminal Appeal of the Supreme Court of South Australia made on 28 April 2015. Remit the matter to the Court of Criminal Appeal for determination.

GAGELER J. The common law imposes criminal liability on one person, a secondary party, for an offence committed by another person, a primary party, where the secondary party intentionally assists or encourages the commission of the offence by the primary party. The common law also imposes criminal liability on a secondary party where the primary party commits the offence as part of a criminal enterprise in which the secondary party participates. The criminal liability of the secondary party in the first of those circumstances is commonly referred to as "accessorial liability". The criminal liability of the secondary party in the second of those circumstances is commonly referred to as "joint criminal enterprise liability".

There is a real question as to whether accessorial liability and joint criminal enterprise liability are distinct in concept, and in particular as to whether joint criminal enterprise liability is anything more than a subcategory of accessorial liability. The question has been debated academically and conflicting answers have been suggested judicially. The question has not previously arisen for definitive resolution in this Court and does not arise for definitive resolution now.

Accessorial liability and joint criminal enterprise liability overlap in practice. Procedural and substantive differences attaching to different subcategories of accessorial liability have long been abrogated by statutory provisions in the form of s 267 of the Criminal Law Consolidation Act 1935 (SA). With the abrogation of those differences, where accessorial liability and joint criminal enterprise liability have overlapped in practice there has seldom been seen to be any practical need to distinguish between them.

Common to accessorial liability and joint criminal enterprise liability is the mental element of intention: in order to be liable, a secondary party must intend the commission of the offence by the primary party. The common law for a long time treated intention as a matter for objective determination: a party was taken to intend a probable consequence of an act which that party did or to which that party agreed. Early commentaries on criminal liability at common law, particularly those of Sir Michael Foster in the middle of the eighteenth century and Sir James Stephen in the second half of the nineteenth century, need to be read cautiously in that light. Not until towards the middle of the twentieth century did the common law firmly settle on understanding intention as wholly subjective – as an actual state of mind.

Take, then, a simple case. Three men set out to rob a bank. They adopt a simple plan. One of them, the driver, is to wait in the car. The other two are to enter the bank. One is to wave a gun. The other is to put the money in a bag. The two who enter the bank encounter a security guard. The gunman shoots him and he dies. Who of the three is liable for murder?

The traditional answer of the common law is that the criminal liability of each depends on the intention of each. The gunman is liable for murder if he shot the security guard intending to cause death or grievous harm. If the gunman is liable for murder, the bagman (who might in earlier times have been described as an accessory at the fact) and driver (who might in earlier times have been described as an accessory before the fact) are also liable for murder if they intended that the gunman would shoot with intention to cause death or grievous harm. Their intention need not have been absolute; it need only have been contingent. They may have hoped to get away with robbing the bank without anyone getting hurt. They need only have intended that the gunman would shoot to kill or cause grievous harm as a possible means of carrying out the plan – if worst came to worst.

But what if shooting to kill or cause grievous harm was never part of the plan? The gunman went too far. The gun was not meant to be loaded. The gun was meant only to frighten. The common law's traditional answer has been that the bagman and driver cannot be liable for a criminal act of the gunman that they never intended to occur.

The common law has of late given a different answer. The bagman and driver need not have intended that the gunman would shoot to kill or cause grievous harm as a possible means of carrying out the plan to rob the bank. It is enough for them to be liable for murder that they foresaw the possibility that the gunman would take it upon himself to shoot to kill or cause grievous harm and that they participated in the plan to rob the bank with that foresight.

The distinction between intention and foresight as a basis for imposing criminal liability, in this instance for murder, might seem a fine one where the group is three men, the weapon is a gun, and the plan is to take coordinated action to rob a bank. The distinction comes into sharp relief where the group is an indeterminate number of youths, the weapon is a knife or a baseball bat, and the plan is an evolving tacit agreement to assault or to engage in an affray. One of the group is more prone to violence. He goes further than the rest. He stabs or hits with intent to kill or cause grievous harm and someone dies. Other members of the group may never have intended things to turn out that way. Each member of the group is nevertheless liable for murder if he foresaw the possibility that the one more prone to violence would go beyond the plan and would stab or hit with intent to kill or cause grievous harm.

Just when the common law came to admit of foresight as a sufficient basis for criminal liability can be traced to the advice of the Privy Council on appeal from Hong Kong in Chan Wing-Siu v The Queen. The author of the advice was Sir Robin Cooke. Equating "contemplation" with "authorisation", the advice propounded a "wider principle" of secondary liability according to which a secondary party would be criminally liable for the commission by a primary party of an offence which the secondary party did "not necessarily intend" but which the secondary party did foresee as a "possible incident" of their "common unlawful enterprise". The "criminal culpability" of the secondary party according to that wider principle was said to lie in "participating in the venture with that foresight".

The timing of the common law development can be traced more precisely to a comment on Chan Wing-Siu by Professor JC Smith in the Criminal Law Review in 1990. Professor Smith there pointed out in relation to the critical passage in Chan Wing-Siu that "contemplation" is not the same thing as "authorisation" and that "the general effect of the passage is that contemplation or foresight is enough". Until then, Chan Wing-Siu had not been understood that way by courts in England or in Australia.

Professor Smith went on to comment in 1997 that "[i]t may be that the law is too harsh and, if so, it could be modified so as to require intention (or even purpose) on the part of the accessory that, in the event which has occurred, the principal should act as he did". But by 1997 the common law had turned in the harsh direction to which Professor Smith had pointed in 1990.

The turn occurred in England in 1990 in the decision of the Court of Appeal in R v Hyde. Referring to its own emphatic rejection as late as 1989 of the suggestion that "a mere foresight of the real or definite possibility of violence being used is sufficient to constitute the mental element of murder", the Court of Appeal then said: "On reconsideration, that passage is not in accordance with the principles set out by Sir Robin Cooke which we were endeavouring to follow and was wrong, or at least misleading. If B realises (without agreeing to such conduct being used) that A may kill or intentionally inflict serious injury, but nevertheless continues to participate with A in the venture, that will amount to a sufficient mental element for B to be guilty of murder if A, with the requisite intent, kills in the course of the venture. As Professor Smith points out, B has in those circumstances lent himself to the enterprise and by so doing he has given assistance and encouragement to A in carrying out an enterprise which B realises may involve murder."

Within a year, that new view of secondary criminal liability was adopted in a further decision of the Privy Council on appeal from Hong Kong. Within 10 years, it was endorsed by the House of Lords in R v Powell. The House of Lords there held that "it is sufficient to found a conviction for murder for a secondary party to have realised that in the course of the joint enterprise the primary party might kill with intent to do so or with intent to cause grievous bodily harm".

The equivalent turn occurred in Australia in 1995 in the decision of this Court in McAuliffe v The Queen. The Court there upheld a direction of a trial judge to the effect that all would be liable for murder where: three youths go to a park with a common intention of bashing whoever might be there; an act of one of them causing death is done with the intention of inflicting grievous harm; and each of the other two either share the intention of inflicting grievous harm or "contemplate the intentional infliction of grievous bodily harm as a possible incident of the common criminal enterprise" of bashing whoever might be in the park.

McAuliffe was a unanimous decision of five members of the Court. Acknowledging that occasion had not arisen in Johns v The Queen for the Court to "turn its attention to the situation where one party foresees, but does not agree to, a crime other than that which is planned, and continues to participate in the venture", the Court in McAuliffe followed Chan Wing-Siu and R v Hyde to hold that the liability of the secondary party for the crime committed by the primary party was relevantly the same in each situation. The Court said: "As Sir Robin Cooke observed, the criminal culpability lies in the participation in the joint criminal enterprise with the necessary foresight and that is so whether the foresight is that of an individual party or is shared by all parties. That is in accordance with the general principle of the criminal law that a person who intentionally assists in the commission of a crime or encourages its commission may be convicted as a party to it."

The awkwardness of the resultant common law doctrine, by which a member of a group setting out to commit one offence would become liable for a different offence committed by another member of the group if he or she foresees the possibility of that other member committing that different offence, was reflected in the labels the new doctrine came to be given. Professor Smith called it "parasitic accessory liability" and noted that it had "a savour of 'constructive crime'". In Australia, it became known as "extended common purpose" or "extended joint criminal enterprise liability".

Very recently, the common law doctrine has been revisited by the Supreme Court of the United Kingdom in R v Jogee and by the Privy Council on appeal from Jamaica in Ruddock v The Queen. "[T]he benefit of a much fuller analysis than on previous occasions when the topic [had] been considered" allowed the Supreme Court and the Privy Council to conclude that their earlier acceptance of the doctrine had been "based on an incomplete, and in some respects erroneous, reading of the previous case law, coupled with generalised and questionable policy arguments". The doctrine was said to have remained "highly controversial and a continuing source of difficulty for trial judges" and to have "led to large numbers of appeals". The Supreme Court and the Privy Council concluded that acceptance of the doctrine had been a "wrong turn". They held that the doctrine should be reversed and that the common law should be restated in accordance with previously established principles.

Unsurprisingly, the appellant and applicants for special leave to appeal moved promptly to invite this Court to follow the course taken by the Supreme Court and the Privy Council in restating the common law of Australia. In responding to that invitation, the procedural and pragmatic considerations identified by the Supreme Court and the Privy Council for reversing the doctrine ushered in by Chan Wing-Siu must be acknowledged from the outset to have limited purchase in Australia.

Prosecution reliance on extended joint criminal enterprise liability in Australia has been noted to have been a source of difficulty for judges, to have added to the complexity of jury directions, and to have contributed to the number of appeals. But the problem has not been ignored by legislatures and law reform bodies in Australia. The common law of secondary liability has not for some time applied to offences under Commonwealth or Territory law and the entirety of the common law of secondary liability has recently been abolished by legislation in Victoria in the implementation of recommendations of Weinberg JA, the Judicial College of Victoria and the Victorian Department of Justice. Extensive legislative reform of the common law of secondary liability has been recommended by the New South Wales Law Reform Commission.

Apart from the effect on past convictions (an important topic to which I will need to return), for this Court now to reverse the doctrine of extended joint criminal enterprise would in practical terms amount to declaring that doctrine no longer to form part of the common law as it continues to govern secondary liability only in South Australia and New South Wales.

This Court, moreover, cannot be said to have failed carefully to consider the doctrine until now. Of particular significance is that in 2006 the Court entertained a fully argued attempt to reopen McAuliffe in the course of refusing applications for special leave to appeal in Clayton v The Queen. For reasons then elaborately given, the attempt to reopen McAuliffe was rejected by a majority of six to one. Reasons for refusing applications for special leave to appeal are not binding as precedents. The indication of views of current members of the Court contained in reasons for refusing applications for special leave to appeal can nevertheless have a significant effect on legal practice. The Clayton refusal had just that effect.

In 2007, Kirby J, who had been the sole dissentient in Clayton, said that "[w]hatever doubts or hesitations existed earlier concerning the common law of Australia in this respect, the decision in Clayton has to be taken as settling the matter, at least for the present". Since then, the Court has reapplied McAuliffe, declining an express invitation in 2012 "to establish a more principled and unified approach to when a person should be criminally responsible for the acts of another".

If the common law of Australia is now to be returned to the path it was on before McAuliffe, the only justification could be that the return is compelled by principle. Consideration of principle must examine the reason for following Chan Wing-Siu and R v Hyde stated by all five members of the Court more than 20 years ago in McAuliffe. Consideration of principle must also grapple with the reasons for not reopening McAuliffe given by six members of the Court nearly 10 years ago in Clayton. One of the reasons given in Clayton for not reopening McAuliffe was that other countries continued to apply a similar doctrine. That reason has been overtaken by Jogee and Ruddock. Other reasons have not.

Just one reason was stated in McAuliffe for following Chan Wing-Siu and R v Hyde. That reason, as has already been noted, was that to hold a secondary party liable for a crime committed by a primary party on the basis of the secondary party's participation in a joint criminal enterprise with foresight of that crime accorded with the general principle of the criminal law that a person who intentionally assists in or encourages the commission of a crime may be convicted of that crime.

McAuliffe's identification of the applicable general principle of the criminal law is undoubtedly correct: a person who intentionally assists in or encourages the commission of a crime may be convicted as a party to that crime. The principle explains accessorial liability and (if there is a difference) joint criminal enterprise liability.

The problem is that the general principle does not explain why a secondary party should be liable for a crime committed by a primary party which the secondary party neither intentionally assisted nor encouraged. In short, the principle does not explain McAuliffe's extension of criminal liability beyond accessorial liability or joint criminal enterprise liability.

Of the numerous criticisms of the extension of criminal liability ushered in by Chan Wing-Siu and R v Hyde which are to be found in Jogee and Ruddock and in the formidable dissent of Kirby J in Clayton, two predominate. The first is that making a party liable for a crime which that party foresaw but did not intend disconnects criminal liability from moral culpability. The second is that making the criminal liability of the secondary party turn on foresight when the criminal liability of a principal party turns on intention creates an anomaly.

To my mind, those two criticisms are unanswerable. The first is fundamental, and the second is related to the first. The anomaly demonstrates incoherence in the imposition of criminal liability. The incoherence in turn highlights the disconnection between criminal liability and moral culpability.

The common law has developed ordinarily to insist that justice requires that a primary party become criminally liable only by acting with intention, albeit that in the case of murder the requisite intention is not confined to an intention that the victim be killed but can be an intention that the victim suffer very serious injury, and albeit that in a case of manslaughter special considerations apply. Exceptions to the principle that intention is an element of an offence at common law have been few, and the overall trend of the case law has been for the exceptions to become fewer.

The imposition of liability in the category of case sometimes described as murder by recklessness is not an exception, at least in any presently meaningful sense. To the contrary, contrasting liability of a secondary party for extended joint criminal enterprise murder with liability of a primary party for reckless murder illustrates both the incoherence and the disconnection.

According to the narrow view of murder by recklessness, which has prevailed in Australia, the concept of recklessness is confined to engaging in an act expecting its probable result to be death or grievous harm. Acting with that expectation has been seen to be acting with a state of mind "comparable" to acting with an intention to kill or to do grievous harm in that acting with that expectation is "just as blameworthy". But even on the wide view of murder by recklessness, now rejected, the concept of recklessness was understood to involve more than mere foresight of a possible result: it required foresight to be coupled with willingness to run the risk of the result occurring so as to amount to indifference to a foreseen result.

Underlying the Australian common law's preference for the narrow view over the wide view of murder by recklessness has been acknowledgement of a basic distinction in terms of moral culpability between acting with an intention or an equivalent expectation and acting with mere foresight. Acknowledgement of that basic distinction in terms of moral culpability has in turn been seen to be reflected in the common law distinction between murder and manslaughter. Gibbs J explained in La Fontaine v The Queen: "There is a great difference between the state of mind of an accused who is prepared to risk the consequences of death or grievous bodily harm that he foresees as probable and that of an accused who does no more than take the chance that death or serious injury may ensue although it seems an unlikely consequence. The act of the former is much more worthy of blame than that of the latter. To treat knowledge of a possibility as having the same consequences as knowledge of a probability would be to adopt a stringent test which would seem to obliterate almost totally the distinction between murder and manslaughter."

Consistently with accepting higher moral culpability to attach to acting with intention and lower moral culpability to attach to acting with mere foresight, Gibbs ACJ spelt out the gradations of criminal responsibility of participants in a joint criminal enterprise resulting in death in Markby v The Queen: "When two persons embark on a common unlawful design, the liability of one for acts done by the other depends on whether what was done was within the scope of the common design. Thus if two men go out to rob another, with the common design of using whatever force is necessary to achieve their object, even if that involves the killing of, or the infliction of grievous bodily harm on, the victim, both will be guilty of murder if the victim is killed ... If, however, two men attack another without any intention to cause death or grievous bodily harm, and during the course of the attack one man forms an intention to kill the victim, and strikes the fatal blow with that intention, he may be convicted of murder while the other participant in the plan may be convicted of manslaughter ... The reason why the principal assailant is guilty of murder and the other participant only of manslaughter in such a case is that the former had an actual intention to kill whereas the latter never intended that death or grievous bodily harm be caused to the victim, and if there had not been a departure from the common purpose the death of the victim would have rendered the two participants guilty of manslaughter only. In some cases the inactive participant in the common design may escape liability either for murder or manslaughter. If the principal assailant has gone completely beyond the scope of the common design, and for example 'has used a weapon and acted in a way which no party to that common design could suspect', the inactive participant is not guilty of either murder or manslaughter".

Those very clear gradations of criminal responsibility of participants in a joint criminal enterprise resulting in death have been blurred by the choice made in McAuliffe. The gradations should in my view have been maintained.

To hold a secondary party liable for a crime committed by a primary party which the secondary party foresaw but did not intend does not measure up against the informing principle of the common law "that there should be a close correlation between moral culpability and legal responsibility". In the language of King CJ, who stood against the introduction of the doctrine of extended joint criminal enterprise into the common law of Australia during the period after Chan Wing-Siu and before McAuliffe, the doctrine results in "the unjust conviction of persons of crimes of which they could not be said, in any true sense, to be guilty".

The fundamental problem that the doctrine fails to align criminal liability with moral culpability was not, to my mind, answered by the majority in Clayton in the suggestion that "criminal culpability lies in the continued participation in the joint enterprise with the necessary foresight" or in the observation that a primary party as well as a secondary party can be liable for murder without intending that a victim be killed. Neither the suggestion nor the observation explains how it is consistent with justice and principle that a secondary party is criminally liable for acting merely with foresight of the possibility of the primary party acting with intent.

The prosecution seeks to provide the missing explanation. The prosecution asserts that the imposition of criminal liability on a participant in a joint criminal enterprise for acting with foresight of the commission of a more serious crime is necessary to prevent a "gap" in the law. To support the existence of and need to fill that gap, the prosecution invokes the policy justification that the doctrine is necessary to address the important social problem of escalating gang violence. The prosecution points in support of that policy justification to social science research said to show that individuals behave differently when they are in groups – they take more risks, feel pressure to conform to the majority, and feel less personal responsibility.

What the prosecution seeks to characterise as a gap in the law is nothing more or less than the difference between the limit of secondary criminal liability as traditionally understood and the limit of secondary criminal liability as extended following Chan Wing-Siu. There is in truth no gap to be filled. Absent the extension of secondary criminal liability, there would be no hole in the legal fabric which would need to be mended. There would be an absence of secondary criminal liability in circumstances now covered solely by the extension. There would be an alignment of criminal liability with moral culpability.

What the prosecution advances as the policy justification for the extension is a highly contestable normative judgment about the appropriate legal response to a particular social problem. The policy justification was once proffered in the House of Lords, but is now rejected by the Supreme Court of the United Kingdom and the Privy Council.

Courts must of course make normative judgments in the course of adapting the common law to meet contemporary social conditions. But courts must be extremely cautious about refashioning common law principles to expand criminal liability. Escalating gang violence is hardly a new social phenomenon. Whether some, and if so what, modification of common law principles of secondary criminal liability is needed to address that particular social problem in a contemporary setting is appropriately a question for legislative consideration. Significantly, no law reform body considering the problem has seen fit to recommend that the appropriate response is to impose secondary criminal liability by reference only to foresight.

Whether the social science literature to which the prosecution points provides an empirical basis for drawing any general conclusion about gang behaviour has been questioned academically and was not scrutinised in argument. The literature does nothing to dispel the concern expressed by Kirby J in Clayton that the extension of secondary criminal liability to individuals unable to extricate themselves from a group as violence gets out of hand operates to catch potentially weak and vulnerable secondary offenders, fixing them with "very serious criminal liability because they were in the wrong place at the wrong time in the wrong company".

The majority in Clayton gave as another reason for refusing to reopen McAuliffe that it would be inappropriate to reconsider the doctrine of extended joint criminal enterprise without reconsidering other aspects of common law criminal responsibility including the whole of the law with respect to secondary liability for crime. I cannot agree. Adoption of the doctrine was a discrete judicial development. The doctrine is capable of discrete judicial reversal. Whatever room there may be for debate as to their jurisprudential foundation, and however much they might yet be improved by reconsideration and re-expression, the common law principles of secondary liability apart from the doctrine of extended joint criminal enterprise would remain unaffected by its excision. The distinction between murder and manslaughter in a case of joint criminal enterprise would re-emerge with clarity.

One further consideration, not mentioned in Clayton and not now raised by the prosecution, must be addressed. It is the systemic consideration of stability. To declare the common law in a case such as this is to declare the common law for the past as well as the future. To reopen and overrule McAuliffe would be to hold that the doctrine McAuliffe introduced has never been part of the common law of Australia. The overruling of McAuliffe would not of itself alter the legal rights of persons whose criminal liability has already merged in conviction. The overruling would nevertheless create a legitimate sense of injustice in persons who have been convicted on the assumption that the doctrine of extended joint criminal enterprise formed part of the common law of Australia and raise the real prospect of many of them seeking to have their convictions overturned by invoking such avenues of legal redress as may remain available to them. The overruling would also raise the prospect of criticism of a court system which could proceed on an erroneous view of the common law for more than 20 years.

Troubling as that consideration is, it cannot be decisive. The doctrine of extended joint criminal enterprise is neither deeply entrenched nor widely enmeshed within our legal system. The problem the doctrine has created is one of over-criminalisation. To excise it would do more to strengthen the common law than to weaken it. Where personal liberty is at stake, no less than where constitutional issues are in play, I have no doubt that it is better that this Court be "ultimately right" than that it be "persistently wrong".

The doctrine of extended joint criminal enterprise is anomalous and unjust. The occasion for its reconsideration having been squarely presented, I cannot countenance its perpetuation. Dissenting from the view of the majority, I would reopen and overrule McAuliffe.

To overrule McAuliffe would mean that one of the pathways to criminal liability for murder left to the jury in respect of the appellant and each applicant for special leave to appeal was not open in law. On that basis, I would accede to each application for special leave to appeal and, in each appeal, allow the appeal and set aside the order of the Full Court of the Supreme Court of South Australia. In place of that order, I would quash the convictions and order a retrial on counts other than those which rely on extended joint criminal enterprise.

KEANE J. I agree that the applications for special leave should be granted and the appeals should be allowed for the reasons given by French CJ, Kiefel, Bell, Nettle and Gordon JJ. I also agree with their Honours' reasons for concluding that the common law in Australia should not be altered by the rejection of the principle of criminal responsibility associated with the doctrine known as extended joint criminal enterprise. I wish to add some brief observations upon the issues of principle and policy exposed by the divergence of approach which has emerged between this Court and the Supreme Court of the United Kingdom following its decision in R v Jogee. Jogee

In Jogee, the Supreme Court held that the Privy Council's decision in Chan Wing-Siu v The Queen should be overruled in so far as it supports the proposition that if two people set out to commit an offence (crime A), and in the course of it one of them commits another offence (crime B), the second person is guilty as an accessory to crime B if he or she foresaw it as a possibility, but did not necessarily intend it.

The Supreme Court (Lord Hughes and Lord Toulson JJSC, with whom Lord Neuberger of Abbotsbury PSC, Lord Thomas of Cwmgiedd CJ and Baroness Hale of Richmond DPSC agreed) said: "The error [in Chan Wing-Siu] was to equate foresight with intent to assist, as a matter of law; the correct approach is to treat it as evidence of intent. The long-standing pre Chan Wing-Siu practice of inferring intent to assist from a common criminal purpose which includes the further crime, if the occasion for it were to arise, was always a legitimate one; what was illegitimate was to treat foresight as an inevitable yardstick of common purpose."

Their Lordships considered that Chan Wing-Siu "brings the striking anomaly of requiring a lower mental threshold for guilt in the case of the accessory than in the case of the principal." In this regard, it was said to be anomalous that "foreseeability of death or really serious harm was not sufficient mens rea for the principal to be guilty of murder, but was sufficient in a secondary party." The Australian position

In Gillard v The Queen, Hayne J, with whom Gummow J agreed, explained that the decision in McAuliffe v The Queen is founded on the notion that the criminal culpability of a participant in a criminal joint venture for an "incidental crime, when its commission is foreseen but not agreed ... lies in the participation in the joint criminal enterprise with the necessary foresight." Similarly, Gleeson CJ and Callinan J, with whom Kirby J agreed, identified the basis of responsibility for the incidental crime in continued participation in the venture with foresight of that crime as a possibility. Continued participation with that foresight "is regarded as intentionally assisting in the commission" of that crime.

In Clayton v The Queen, six members of this Court declined to reconsider the position established in McAuliffe and Gillard, and emphasised that, while accessorial responsibility is grounded in the secondary party's intentional contribution to a particular crime, the criminal responsibility of a participant in a joint criminal enterprise is grounded in the authorisation of a crime which is incidental to the enterprise. Considerations of principle

The position established in Australian law by these decisions does not deny or diminish the importance of the overarching concern that criminal responsibility should reflect the moral culpability of the individual offender. Rather, the Australian position recognises that deliberate participation in a joint criminal enterprise which carries a foreseen risk of an incidental crime itself has an important bearing upon the individual moral culpability of each participant for the incidental crime. The implications of deliberate participation in a criminal enterprise for the moral culpability of each individual participant are ignored if one adopts an analysis of criminal responsibility which starts from an assumption that the person who commits the actus reus of the incidental offence is the principal offender and all others complicit in that offence are to be regarded as having accessorial responsibility only. The moral culpability of a participant in a crime will not always be revealed by an analysis which assumes that the participant has merely aided or abetted the commission of the actus reus by the principal offender.

In particular, where two or more persons agree to commit a crime together knowing that its execution includes the risk of the commission of another crime in the course of its execution, there is no obvious reason, in terms of individual moral culpability, why the person who commits the actus reus should bear primary criminal responsibility, as between himself or herself and the other participants to the joint criminal enterprise, for the incidental crime. Because of the fact of the agreement to carry out jointly the criminal enterprise, the person who commits the actus reus of the incidental crime is necessarily acting as the instrument of the other participants to deal with the foreseen exigencies of carrying their enterprise into effect.

The decision in Jogee proceeds squarely on the basis that cases of complicity in a crime must be analysed as a subset of accessorial liability. To insist that the liability of participants in a joint criminal enterprise be analysed exclusively in terms of accessorial liability is to fail to recognise that each participant in a joint criminal enterprise is not merely an accessory to a crime committed by someone else. Where parties commit to a joint criminal enterprise, each participant becomes, by reason of that commitment, both the principal and the agent of the other participants: for the purposes of that enterprise they are partners in crime. Each participant also necessarily authorises those acts which he or she foresees as possible incidents of carrying out the enterprise in which he or she has agreed, and continues, to participate. It is to be understood, of course, that the agreement to participate in a joint criminal enterprise, while it may be inferred from the circumstances, must be proved as a fact beyond reasonable doubt.

As to the anomaly identified in Jogee, namely that under the approach in Chan Wing-Siu foreseeability of death or really serious injury is not a sufficient mens rea for the principal to be guilty of murder, but is sufficient in a secondary party, two points may be made. First, the anomaly is said to lie in a comparison of the position of the person who performs the actus reus as the principal, with that of the participant who is described as the secondary party. But to say this is to identify the person who commits the actus reus as the principal. For reasons already stated, in cases of an agreed pursuit of a criminal purpose, one is not concerned with the criminal responsibility of a party whose involvement is merely "secondary". One is concerned with the criminal responsibility of each of two principal parties to a criminal enterprise for incidents which occur in carrying it out.

Secondly, there is little reason to conclude that the person who commits the actus reus of the incidental crime should bear a greater degree of moral culpability for that crime than those of his or her consorts whose instrument he or she became for the purpose of dealing with the exigencies of carrying the joint enterprise into execution where those exigencies have subjectively been foreseen by them. There is little reason why one who organises a crime should be regarded as less morally culpable for the risks of carrying it out, which he or she foresees, than those deployed to deal with those risks. To say that those who join together to organise the commission of a crime, in circumstances which involve the acceptance of the risk of the commission of an incidental crime in the course of carrying out their enterprise, are less morally culpable for the incidental crime than their consort who actually does the dirty work, is to appeal to a sense of morality which could commend itself only to the criminal elite.

In Jogee, Sir Robin Cooke's reasons in Chan Wing-Siu were also criticised because, as was said, they "elided foresight with authorisation". But the reasoning of Sir Robin Cooke does not equate contemplation with authorisation, or otherwise merge these concepts. Rather, his Lordship held that participation in the commission of a crime, with foresight of the risk of the incidental crime, establishes authorisation of the incidental crime where the foreseen risk eventuates. The concept of authorisation of an incidental crime is no less apt to capture the degree of individual moral culpability for that crime than the concept of "conditional intent" propounded in Jogee.

In Jogee, it was said that authorisation of crime B cannot "automatically be inferred from continued participation in crime A with foresight of crime B" because it "makes guilty those who foresee crime B but never intended it or wanted it to happen." But no one disputes that a person may be held criminally responsible for an occurrence even though he or she does not want it to happen. The circumstance that a party to a robbery might fervently hope that no one will be killed in the course of carrying it out is hardly a reason to hold that that party did not authorise, and thereby intend, an intentional killing where that occurrence was actually foreseen as a possible incident of carrying out the robbery.

The final criticism in Jogee of the reasoning of Sir Robin Cooke in Chan Wing-Siu which may be noted here was that it is "illegitimate ... to treat foresight as an inevitable yardstick of common purpose." But, as is apparent from McAuliffe, Gillard and Clayton, foresight is not a yardstick of common purpose; rather it is a basis for concluding that what occurred in the pursuit of the common purpose was subjectively authorised by each participant. Considerations of policy

As a matter of policy, it is well-recognised that the pursuit of a joint criminal enterprise necessarily involves a substantial element of unpredictability, which exposes the participants, their victims and the general public to the unacceptable risk that a crime additional to that which motivated the enterprise might be committed. It is perfectly intelligible, as a matter of policy, that the law should expose each participant in a joint criminal enterprise to punishment for an incidental crime if he or she actually foresees the risk of the commission of the incidental crime and authorises the eventuation of that risk as part of his or her continued participation in the enterprise.

The appellants' contention that a participant in a joint criminal enterprise does not, by the "mere" fact of agreement to participate, make any causal contribution to the commission of the actus reus of the incidental crime is not at all compelling. It is trite that a group is a more potent force, even in purely physical terms, than an individual acting alone: there is strength in numbers. In addition, members of groups acting together tend to exhibit higher levels of moral disinhibition in their interactions with other persons than occurs when individuals act alone. McAuliffe is a powerful illustration of this phenomenon. The strength of the policy of protecting the public from the dangerous exigencies of the pursuit of joint criminal enterprises was acknowledged in England in R v Powell; nothing has occurred since to cause a reappraisal of the strength of that consideration.

In addition, where a joint criminal enterprise is in the nature of a business activity on the part of the participants, as was the case in Gillard, the reasons of policy which support the Australian position are no less strong. Where crime is a business the conduct of which puts the lives of innocent citizens at risk, it is not sound policy to minimise the criminal responsibility of those who organise crime, and in so doing create the foreseen risk of an incidental crime, merely because they are able to engage others as their agents for that purpose.

For this Court now to accept the invitation to depart from the position established in McAuliffe would create a serious "want of continuity in the interpretation of the law." Only the most compelling grounds would justify such a course, given that since McAuliffe this Court has twice affirmed that it correctly stated the common law in Australia. For the reasons stated above, the grounds for taking that course are far from compelling. On the contrary, considerations of principle and policy provide strong support for maintaining the Australian position.