The Supreme Court has just given judgment in an appeal concerning the principles of so-called accessory liability in the criminal law. The typical case is of the person who acts as look-out during an armed robbery, but who does not himself fire the shots that kill or injure the victims. The court unanimously cut back the criminal culpability of accessory defendants. They were clearly right to do so. The principles that they set down are appropriate in cases of murder, and the result reached although radical on the basis of the law as previously understood, is right in principle. However, I will argue that in laying down the law for subsequent decisions, the court crucially failed to adequately distinguish the questions of whether the actus reus of an offence can be attributed to a person other than the one physically carrying it out, and the mens rea required to convict the accessory.
Actus Reus and Mens Rea
To be guilty of a crime it must be my doing. So for murder, the action is the killing of another. Theft the appropriation of property belonging to another. Battery the unlawful use of force upon another. And so on. The actus reus of the offence does not concern the state of the defendant’s mind. [In some extreme cases, although the defendant’s body may have caused the outcome that constitutes the offence it may not be the defendant’s doing, as where I grab your hand and strike another with it, or where I am drugged and in an unconscious state cause the victim to be shot.]
Most offences also require some mental element, which is roughly synonymous with blameworthiness. So, murder requires an intention to kill or cause serious injury. Theft dishonesty and an intention to permanently deprive. Battery the intention to apply immediate force, or the foresight that such force may be applied by the action being done.
Accessories: the Actus Reus
Qui facit per alium facit per se: he who acts through another does the act himself. Although this is a legal principle, it is also a principle of personal responsibility that we recognise in every day life. Corporations, governments, football teams all act through real world human agents, and their agents’ actions are attributed to them. One difficult question (which we do not need to take further) is the kind of participation that is sufficient to attribute one person’s actions to another. The language of the law is that D2 must “aid, abet, counsel or procure” D1’s actions. [The rules on conspiracy are, today in England, seen as separate. This is in my opinion an historical aberration. Conspiracy is simply another method by which one person’s actions may be attributed to another.]
If therefore D1’s actions are attributed to D2, and D2 has the requisite blameworthiness for the offence, D2 is guilty of the offence. Notice that D2 is guilty of the principal offence. If the action procured is the killing of another, and D2 has the mental blameworthiness sufficient for the offence, D2 is a murderer (see s 8 of the Accessories and Abettors Act 1861). D2 is not guilty of any separate offence of “procuring murder”. Further, D1 may not be guilty of the principal offence as lacking the necessary mens rea. This should not prevent D2’s conviction if it can be proven that he had the requisite blameworthiness. As in morality generally, D2’s guilt or innocence should not turn upon the state of D1’s mind.
Accessories: the Mens Rea
Should a lower level of blameworthiness suffice in order to convict the accessory than is required for the principal? If, for example, the look-out intends the shop to be robbed and foresees that it is possible that the principal may lose his head and kill someone, should that suffice for murder if the principal does in fact kill a bystander?
In principle the answer is clearly no. Any legal system that says that the level of blameworthiness for crime X is Y, but that we will convict you of crime X when your degree of blameworthiness is less than Y when you have assisted another, has got itself into a mess. And a mess is what English law has been in this field, for at least 32 years.
In Chan Wing-Siu the Privy Council had held that where in the course of a joint enterprise to commit an offence (such as robbery) D1 commits another offence (such as murder), D2 is also guilty of that other offence as an accessory if he could foresee that D1 might act as he did. This had been confirmed by the House of Lords in R v Powell and R v English. This made no sense at all. It meant that you could be convicted of murder despite lacking the degree of blameworthiness required for murder. Rightly, and predictably, the Supreme Court overturned this line of authority.
The opportunity was there to align the mens rea required for the accessory with that required for the principal offence. This would mean that the mens rea requirement for the accessory would vary according to the offence charged, just as it does for the principal. The degree of blameworthiness required should be exactly congruent with that required for the offence. Unfortunately this step was not taken, and the Supreme Court has left the law in an unsatisfactory state. Why?
For crimes of intent the court state
If the crime requires a particular intent, D2 must intend to assist or encourage D1 to act with such intent
This is wrong. It is too narrow. For purposes of D2’s conviction, D1’s blameworthiness is neither here nor there. For D2 to be convicted of a crime of intent it should not matter what D1’s intentions were, nor what D2 intends about those intentions. If D1’s actions are attributed to D2 and D2 himself has the requisite intent of the principal offence, D2 should be capable of conviction. For purposes of convicting D2, if he procures the killing of V by D1, intending that V should be killed, why should it matter that he doesn’t intend D1 to act with the intent to kill or cause serious injury?
The court continue
[T]here can be cases where D2 gives intentional assistance or encouragement to D1 to commit an offence and to act with the mental element required of him, but without D2 having a positive intent that the particular offence will be committed. That may be so, for example, where at the time that encouragement is given it remains uncertain what D1 might do; an arms supplier might be such a case.
This is also wrong. It is too broad. If the crime of murder requires an intent to kill or cause serious injury, a defendant who lacks the intent that another should be killed or seriously injured should not be guilty of it. If they can foresee death or serious injury they may be guilty of another lesser offence (here manslaughter). An intent to assist someone who may or may not kill should not be enough. Their blameworthiness is very high, equivalent to the bomber who plants a bomb knowing it may or may not go off, but indifferent as to whether it does. But it is not murder.
Further the court adopts the position that it is always necessary to intend to assist the perpetrator to do the prohibited act. So we are told 
Where the offence charged does not require mens rea, the only mens rea required of the secondary party is that he intended to encourage or assist the perpetrator to do the prohibited act, with knowledge of any facts and circumstances necessary for it to be a prohibited act:
This is wrong, and is again too narrow. If, for example, the look-out intends a theft, but foresees that it may be necessary to use physical force, he should be convicted as an accomplice to robbery. It should not be necessary to intend to assist the use of force (which is part of the actus reus for robbery) as an intention to use force is not a constitutive element of the offence. Similarly, if you procure an offence of strict liability, if the actions of the principal are attributed to you, the intention with which you acted should be irrelevant (the ratio of NCB v Gamble does not contradict this).
If the proposed principle represented the law in relation to offences of strict liability then, a fortiori, the same should apply in relation to offences requiring recklessness or gross negligence. Is it really the case that to find someone liable as an accessory to manslaughter that it be proven that they intended to encourage or assist the killing? Are defendants such as these not clearly guilty of at least manslaughter, given their degree of blameworthiness? Clearly the Supreme Court do not think that is the law, and concluded that the submission that the defendant Jogee was not at least guilty of manslaughter was ‘hopeless.’ However, what they consider to be the appropriate mens rea test in cases of manslaughter is, to me, hopelessly obscure (see ).
The Supreme Court are correct that there would be a ‘striking’ anomaly if there were a lower mental threshold for guilt in the case of the accessory than in the case of the principal for conviction of the same offence .
The correct solution to this ‘striking’ anomaly ought to have been to align the mental threshold for guilt in the case of the accessory with that of the principal. Such a move would have been a bold one, but no less bold than the step taken of overturning over thirty years of authority. The court failed adequately to distinguish between rules for the attribution of action, and the necessary mental element constitutive of the offence.
The problem of accessories in cases of murder has now been solved in a generally satisfactory way: because it is a crime of intent and the mental elements for principal and accessory have been nearly aligned. The problem has not been solved for other offences, and we will in my view be back before the Supreme Court in quick order in relation to other offences to require the court to review what it has done.