Abstract
Complicity provides a perfect place from which to take steps towards a doctrinally clear and coherent criminal law. In particular, by acknowledging in the complex mass of cases a requirement that the accomplice contribute to the principal's crime. That takes effect differently in assistance and encouragement compared to procuring: (a) an accessory's assistance or encouragement must make a significant contribution to the principal's crime, but does not need to be a but-for cause, and (b) to procure an offence, an accessory must cause it in a but-for sense. This requirement flows from how complicity can be justified and determines the linguistic form of complicity. It extends to the end point of complicity: overwhelming supervening events and withdrawal on the one hand, and sentencing on the other. Fidelity to how we express, and label, the wrongs within participation are important parts of the work we expect it to do. That includes what the wrong in complicity is (and how participants are labelled), the limits in what one can do through another or intend another can do, the causal claims we make in complicity and the differences between forms of complicity. Without even a significant contribution, an accomplice is not meaningfully involved in the principal's crime
Keywords
In complicity, like many other areas of English law, legal actors routinely and mistakenly sacrifice precision about doctrine for the sake of a purportedly practical application of the law. It is too easily accepted that vague rules are the best answer for complex problems, even when that vagueness subverts the law's normative and linguistic foundations. Even the most extreme form of complicity, parasitic accessorial liability, was, in 1999 recognised as difficult: ‘Intellectually, there are problems with the concept… but they do not detract from its general practical worth’, 1 highlighting the choice being made between intellectual rigour and practical value. That choice could be made differently. English lawyers can take small but valuable steps to address this problem, and one sensible place to start is complicity. In particular, we could clarify the contribution a secondary party has to make to the principal's crime. The best way to develop the current law would be that: (a) an accessory's assistance or encouragement must make a significant contribution to the principal's crime, but does not need to be a but-for cause, and (b) to procure an offence, an accessory must cause it in a but-for sense. This would improve the criminal doctrine at the point of establishing liability, in ending liability, and in determining sentence. An accompanying advantage is that this proposal does not require that we forge a new consensus on complicity, or indeed a first consensus, on the theories underpinning complicity. Furthermore, by using the contribution in complicity as a starting point to push for greater doctrinal clarity, a wider improvement to criminal law more generally might be begun. Clarifying the need for contribution in complicity has been something the criminal law should have done already. 2 In the face of seemingly unresolvable tensions of principle, we can at least start to make progress towards a better criminal law by accepting that small steps towards coherent doctrine have value.
This is not a paper about the merits of Jogee. 3 What is now the leading decision on complicity, from 2016, has its proponents and some detractors. The decision removed the category of Parasitic Accessorial Liability. Now, for each crime, an accessory is only liable if she assists, encourages or procures it, intending to assist, encourage or procure it; where the crime requires fault, the accessory must intend the principal to have that fault. There is certainly space to debate the value and effect of the decision. 4 However, those debates can too easily treat complicity as a special box of its own, seeing or promoting as exceptional problems which predate Jogee and/or apply beyond it. For example, conditional intention, and the substantial injustice test for appealing out of time, predate Jogee. So too did the removal of separate doctrinal rules for participants in a common purpose; what remains is the predominant prosecution narrative that the defendants were ‘all in it together’. On the substantive merits of Jogee itself, there is no good reason why, as a general rule, it should be easier to convict a person of a second crime merely because a first crime has been committed. Even if it were, there is certainly no good reason why such an extreme form of liability as parasitic accessorial liability should be used to do so. Under parasitic accessorial liability, even the absence of contribution to the principal's commission of the second crime, and the accomplice's mere foresight of the possibility that the principal would commit it, make the accomplice liable in full for it. 5 In addition, an accomplice should have to intend to contribute to each crime of the principal. And, also as now, the accomplice should also know enough of the principal's conduct and plans, to know of the criminality on the normal basis that the law is deemed to be known. We have already felt the consequences of failing to articulate and test for doctrinal components in complicity and instead relying on lay-led applications of vague doctrine. This is highly problematic. It makes inconsistent results more likely. It makes the law less coherent and less predictable. It reduces the ability of the subjects of the criminal law to have fair warning of what conduct is prohibited and live lives avoiding the criminal law. Courts relying on vague doctrine is not uncommon, and many legal systems criminalise complicity without having a robust and precise explanation for it driving precise legal rules. However, English law seems to have adopted, or fallen into, a far more opaque and conviction-maximising position. That has permeated criminal legal culture, so that any steps away from the permissive and often punitive vagueness are criticised (and some of the reactions to Jogee should be seen in that light 6 ).
Instead of being about Jogee itself, this piece picks out something Jogee passed by without exploring or justifying: what contribution the accomplice needs to make to the principal's crime. Often there is little doubt that an accomplice has either made a crime more likely and/or easier to commit, by supplying assistance, encouragement or bringing the crime about in another way. However, English law is too willing to assume that such a contribution has occurred and has little detailed law to test for it. The best position is a two-part approach: to be liable for assisting or encouraging a crime, the accomplice must make a substantial contribution to the principal’s commission of it; to be liable for procuring the principal's crime, the accomplice must bring the crime about.
The article considers complicity and its reform in four stages:
The normative gap in complicity; The importance of language in complicity; The threshold of complicitous conduct; and The end of complicity.
The key point is that even if we cannot define complicity perfectly, we need to do better at containing the risks caused by the flexibility and comprehensibility we crave. That is particularly so where, at the moment, those risks are being borne by defendants, and often, the most marginalised defendants in society.
7
Small improvements in our law can easily be implemented by judges recognising threads from the existing law, while still centering the role of juries in finding facts and applying normative tests to those facts. The law on complicity has continued for decades with an approach of blurring justifications and foundational distinctions, such as between principal and accessory, and it could potentially continue doing so. However, this proposal offers the opportunity to make the law clearer, more intellectually rigorous, more internally coherent and align better with what lay people, and lawyers, think the law is doing. Such a law is more likely to be better understood and has a claim to be more likely to be followed. It also avoids treating complicity as an inchoate offence, only worse, since the accomplice would not need to contribute at all and yet would still be labelled as having done so.
8
Normative Gaps
A first question is how complicity fits within the criminal law. If complicity liability is justified under mainstream criminal law, there would normally be some form of responsibility for conduct or outcome which the law has prohibited, and culpability in how that conduct or outcome came about. In fact, English law has a gap in how it justifies participation through complicity, yet relies on language which asserts that gap has been filled. Indeed, the language asserts to the world that the secondary party contributed to the crime.
Derivate Trial and Full Labelling
In complicity, the prohibited conduct/outcome is defined by the crime committed by the principal, and so too is the culpability required of the accomplice; accomplices are not made liable ‘some self-subsisting crime’. 9 ‘Derivative liability’ is therefore key to understanding English law.
The accomplice's liability is ‘derived from’ the principal's liability, but it is a special form of derivation. Derivation here does not have any sense of ‘lesser than’, like the work of a lesser artist copying a master. Rather, it means ‘equal to’. The accomplice is liable to be indicted, tried and sentenced as if a principal. This is set out in law from the 19th century, its current form being the Accessories and Abettors Act 1861, s. 8. Indeed, it does not matter if some of the jury think the defendant was a principal, and some think the defendant was an accessory. 10 Some special cases aside, discussed below, the principal must commit the offence before the accessory can be liable. All participants are then labelled in the same way as having committed the substantive offence.
This ‘derivative but treated as equal’ rule is well-established law, despite its inherent difficulties. The strongest practical argument in favour is in the ease for prosecutor narratives that the parties were ‘in it together’ and no distinctions need to be drawn between the participants. In some cases, that takes the form of a claim that distinctions cannot be drawn: that the evidence permits the conclusion that each defendant was either a principal or an accessory but we cannot prove which. In that situation, convicting them all for the same offence, rather than no offence, is an attractive proposition. However, there is an alternative: convict all as secondary parties. That would require a principal offence be proven, but no principal needing to be found. Arguably it does less unfairness to accept that, than to label and potentially punish all but one as a principal when they might be accessories.
English law's approach of equal labelling for principals and accessories arguably mischaracterises and misrepresents what the accessory did. 11 It groups all participants under the same label, such as “person convicted of murder”, wiping out all nuance in discussing a potentially huge range of contribution and culpability. 12 This can be problematic both for the label the accessory receives, but also for the asserted equality in what the accessory and principal did. It is far too blunt a labelling process. It does not properly engage with social understandings of law by labelling offences and offenders, and it is difficult to justify the automatic parity of labelling for all parties. 13 Depending on one's particular views about the role of criminal law, this might take the form of a critique of the morality of the wrongdoing, 14 or a less morally-based position. The key levels of labelling appear to be about its audience: the criminal justice system, the offender and the wider public, including any victim(s). On each level, offences generally, and participation in offences in particular, should be defined in a way that differentiates enough to obviously capture the wrongdoing they purport to prohibit. 15
The debate about whether the accessory should be convicted of the same offence as the principal is worth having, and is briefly referred to again, below (section ‘The Formulation of the Wrong in Complicity’). Fuller discussion would require more space, but the issue does tie in with the theme of this paper, of places where English law prioritises practicality over intellectual rigour.
Derivate Sentencing
While all participants are also liable to receive the same sentence, subject to narrow statutory exceptions,
16
an accessory might well receive a lower sentence than the principal. Taking the example of Jogee itself, the principal was sentenced to nearly 23 years in prison, while Jogee was initially given nearly 21 years, reduced on appeal against sentence to 18
17
(and on retrial for manslaughter, to 12
18
). The precise grounds for deciding any differentiation are part of the normal sentencing matrix, turning on aggravating and mitigating factors. The Sentencing Council guidelines do not give details of how to sentence accomplices specifically. Indeed, there seems to be little by way of formal rules to differentiate the accomplice as a class of participant, further reinforcing the idea that the underlying conduct/outcome and associated culpability are derivative of crime, the crime committed by the principal.
19
The structure of the guidelines suggests that being an accessory could be a factor at Step 1 assessing the seriousness of the offence on harm and culpability. At the very least, the role of the accomplice in the relevant harms in the principal's offence, and the culpability for it, might be mitigating factors compared to the principal. Judicial discussion of this question has been imprecise, the most common context has been murder, and the application of what was originally section 269 and Schedule 21 of the Criminal Justice Act 2003, now Schedule 21 of the Sentencing Act 2020. Courts have had to push back against regular appeals by denying a formal distinction between accessory and principal in applying those guidelines determining the minimum term for the life sentence being imposed.
20
The question is the relative roles of responsibility and culpability. Significant discretion is being left to the sentencing judge, but the Court of Appeal also seems to approve of, not just prefer not to intervene to change, similarly onerous sentences for accomplices. In some recent cases, the Court of Appeal has confirmed sentences which arrive at a similar period for the principal and accessory. What is strange is that the rare and minimal discussion that has taken place seems to focus on culpability, and does not also discuss responsibility, particularly not responsibility through causation. In R v Sanchez,
21
in 2008, the Court of Appeal held that: Although the culpability of the secondary party may in many cases be less than the principal, the sentences must be viewed proportionately in the light of the policy of the law, that he who encourages the commission of a murder or assists with the commission is to be dealt with as a murderer.
22
The line from Sanchez was picked up in Semusu in 2021, 24 where the principal's sentence of 20 years was confirmed by the Court of Appeal, and the accomplice's sentence was increased from 16 years to 19 years. While the trial judge held that the accomplice had played a secondary role, 25 the Court rejected any significant discount: ‘There is no reason to find that his culpability was markedly less than that of [the principal] in what was truly a joint attack with a shared murderous intention’. 26 (That was technically true, although it had only been an intention to cause serious harm. 27 ) The court itself found 20 years for the principal to be at the lowest end of the permissible range, 28 and that might have been part of why they did not want the accessory's minimum term to be much below that. The court maintained the distinction between the principal and accessory that the trial judge had seen, but only by a one-year gap. 29 That difference is apparent in the list of aggravating features, where only the principal was described as attacking or injuring the deceased, but it was also described as a joint attack with both offenders playing a part. 30 This mention only of culpability in the formal test from the Court of Appeal was contrasted in both cases with the approach of the trial judges, who also noted differences in the level of contribution made by the accessories as factors relevant to their lower sentences. In Semusu, the court expressly noted that Jogee reduced the number of cases where the accessory will be convicted of murder with the principal, despite a ‘wide gap’ in culpability, because such accessories are more likely to be convicted of manslaughter. That would be on the basis that the accessory only foresaw, and did not intend, that the principal would cause death or serious bodily harm. 31
The culpability line from Sanchez, and that implication from Jogee drawn in Semusu, were picked up in Fearon at the end of 2021. 32 The accomplice had only known of the presence of a knife moments before it was used by the principal. That knowledge moments before the use of the knife seemed to be more operative on sentencing than the fact that the accomplice was not the person using it. However, this particular discussion about culpability in relation to a knife used to kill is slightly unusual. In Kelly in 2011, the Court of Appeal heard a number of appeals on Schedule 21, particular for multi-handed offences. One of the most significant aspects was that where the accessory does not take, nor is a party to the taking of a knife, the starting point set by s. 5A, of 25 years, does not apply. Instead, participation in the crime knowing of the knife would be an aggravating feature of the offence. In that sense, the accomplice's responsibility for the principal's offence is lower when not linked to the knife, and that appears to be linked to a causative argument. Kelly was cited in Fearon, though in Fearon the court did not discuss responsibility or causation expressly.
It might be noted that far more structure in sentencing nuance between accessories and principals has been achieved in other legal systems. In some, defined discounts in sentencing for secondary parties exist in the criminal code, 33 and others have guidelines for a structured reduction based for some accomplices, for example, those who were ‘minimal participants’ in a criminal offence. 34 It is not completely clear whether this flows only from culpability differences, or also from responsibility, particularly causation.
Normative Gaps
Here is the conundrum: what contribution and culpability is enough to make a person an accomplice (and thus liable to be treated, labelled and sentenced as a principal) but also one who might be sentenced to less punishment for reasons at least related to that contribution and culpability?
In part, the conundrum might come down to how legal actors construct liability. Liability is first determined by thresholds, binary tests which are sequentially passed. This is then expressed in a scalar output, a sentence. However, the threshold appears to be very low, bringing the trial, label and sentencing outcome into play for many more cases.
More than that, the threshold appears to ignore the very thing that the labelling of participation requires: that there was aiding, abetting, counselling or procuring. The meaning of those words will be discussed in the next section. We turn first to consider the normative gap in justification for complicity liability: how does the criminal law justify and delimit liability in complicity now.
The short answer, well-known amongst those working in the field, is that there is no clear justification and limitation. Courts rarely do better than the Supreme Court in Jogee, which merely said that the accomplice ‘shares’ in the physical act and culpability of the principal: In the language of the criminal law a person who assists or encourages another to commit a crime is known as an accessory or secondary party. The actual perpetrator is known as a principal, even if his role may be subordinate to that of others. It is a fundamental principle of the criminal law that the accessory is guilty of the same offence as the principal. The reason is not difficult to see. He shares the physical act because even if it was not his hand which struck the blow, ransacked the house, smuggled the drugs or forged the cheque, he has encouraged or assisted those physical acts. Similarly he shares the culpability precisely because he encouraged or assisted the offence.
35
Principals
Quantifying the portion of the main offence shared by the accessory is difficult. For a start, how English law decides who commits a crime as a principal is itself unclear. The simple understanding is that it is the person who fulfils the physical elements/actus reus of the offence. However, as soon as there is more than one person involved in the crime, more than one might be involved in bringing about those physical elements. There does not appear to be any clear test for how much is enough. In Gnango,
36
a difficult case on the edge of complicity and principalship, Lord Kerr said: In Smith & Hogan's Criminal Law, 13th ed (2011), p 215, para 8.5.1.1, the following definition of joint principals is given: ‘D1 and D2 are joint principal offenders where each does an act which is a cause of the actus reus’. Unlike the position in a joint enterprise, no common purpose is required in order to render those who cause or contribute to a cause of the actus reus guilty as joint principals. What is required is that each must contribute by his own act to the commission of the offence with the necessary mens rea.
Under Lord Kerr's analysis, ‘a contribution’ from each individual was enough to make a person a joint principal. The very fact that the definition of principal remaining unclear is made less practically problematic by accomplices being able to be treated as principals for trial and sentencing. 37 There is still the difficult decision of how to distinguish the principal from the accessory, and whether contribution to the commission of the crime is a useful mode of analysis. This article argues that contribution is a useful marker, but there is a great difference in scale and role between the principal and the accomplice.
Causation in Complicity?
The case authorities on causation within complicity do not offer a clear answer to our conundrum of what contribution and culpability is enough to make a person an accomplice. Most judges seem to appreciate some causative role, but without being able to label it as a but-for cause, they have not agreed an alternative test. To take one of the most recent examples, prior to moving to the Supreme Court, Toulson LJ attempted a formulation of a causal requirement in the Court of Appeal in R v Mendez and Thompson
38
in 2010. He described the minimum requirement as ‘some causative link’.
39
The case was about parasitic accessorial liability, but the Court was speaking of secondary liability more broadly. The difficulty came in defining that link, as the only practical example given related to counselling, not to aiding. This was in the context of R v Calhaem,
40
discussed below. Toulson LJ, giving the judgment of the court,
41
held: The prosecution do not have to satisfy a ‘but for’ test, i.e. that P's act would not have happened but for D's assistance or encouragement: Attorney General v Able [1984] QB 795, 812 and R v Calhaem [1985] QB 808. To require the prosecution to satisfy a ‘but for’ test would be to place an impossible burden on them in many cases and would be liable to produce perverse and unprincipled results. Where a victim (V) is attacked by a group, it may well be the case that if any one of the group had not taken part in the attack the outcome would have been the same. If the prosecution had to satisfy a ‘but for’ test in relation to each defendant, the result would be that no defendant had committed the offence, whereas it is proper to regard each as having contributed to it. So it is no defence for D to say that without his assistance or encouragement the offence would still have occurred. However, in both Attorney General v Able and R v Calhaem the court recognised that there must be a connecting link between D's assistance or encouragement and P's act, without attempting a precise definition of the connection, other than to say (in R v Calhaem) that P's act must be done within the scope of D's authority or advice. As an example, the court postulated a case where D encouraged P to kill V, and soon afterwards P became involved in a football riot in which he used a weapon and killed V, without realising that the person he attacked was the same person as D had encouraged him to kill. Although P had done what D encouraged him to do, there would have been no link between D's encouragement and P's act. For D to be found guilty jointly with P, D's conduct must (objectively) have constituted assistance or encouragement at the time of P's act, even if P (subjectively) did not need assistance or encouragement.
In Calhaem, the secondary party wanted to have another woman, V, killed, and thus she hired P to kill V. P decided not to carry out the murder, intending to fake the murder attempt instead. This fake attempt became a real homicide when, at, V's property, Panicked in the face of V's alarm. The accomplice was liable for counselling murder. In Calhaem the Court of Appeal decided that there did not have to be any causal connection between the counselling and the commission of the offence.
44
There, the Court of Appeal acknowledged that some earlier cases had interpreted “counselling” to require causation but nonetheless, liability would be found by focusing on the scope of the authority or advice once contact had been made, rather than its causative impact. This represented a significant expansion of liability, or rather, a solidification of the most expansive view that had been argued, most likely as the minority view. Parker LJ, giving the judgment of the court, held: We must therefore approach the question raised on the basis that we should give to the word ‘counsel’ its ordinary meaning, which is, as the judge said, ‘advise’, ‘solicit’, or something of that sort. There is no implication in the word itself that there should be any causal connection between the counselling and the offence. It is true that, unlike the offence of incitement at common law, the actual offence must have been committed, and committed by the person counselled. To this extent there must clearly be, first, contact between the parties, and, secondly, a connection between the counselling and the murder. Equally, the act done must, we think, be done within the scope of the authority or advice, and not, for example, accidentally when the mind of the final murderer did not go with his actions… We see, however, no need to import anything further into the meaning of the word, unless authority drives us to do so. It is of course possible, and both counsel took this course, to take examples of cases which appear to be anomalous, whichever construction is put upon the wording. Such anomalies do not, in our view, assist. There will always be cases in which difficulties are raised and they are disposed of in the ordinary way by sensible action on the part of the prosecuting authority. So far as the authorities are concerned, there are in them, we accept, phrases which may appear, at any rate at first sight, to assist the view put forward by Mr. Carman.
45
proffered advice or encouragement which has no effect on the mind of the principal offender is not counselling… This is not to say that the counselling must be a cause of the commission of the offence. So to require would be to insist that the counselling must be a sine qua non of the offence—i.e., that Dl would not have committed the offence if he had not been counselled.
47
In a case of encouragement, as contrasted with assistance, the Law Commission suggested that the encouragement ‘must have the capacity to act on P's mind’. Similarly, KJM Smith argued that rather than proving an effect in fact, it should be shown the conduct ‘at least carried the potential to cause or help the principal to act in the way he acted’. 48 Smith's work is persuasive. It highlights how frustrating the authority is on the relevance of causation was in the mid-1980s, and it is not clear it has significantly improved since then. Smith’s view was that the fact that principal offence has to be proven to have been committed shows that even in the case of encouragement, there is an underlying causal basis. For the assisting form of complicity, there an inescapable assumption of some causal contribution or influence. 49
Causation Calling for a Formulation
As already noted, the same underlying thread of causation has not been tied up into a practical test. The way that the court put it in R v Mendez was that D's conduct must (objectively) have constituted assistance or encouragement, even if P (as a matter of fact about him/herself) did not need assistance or encouragement. Whereas the provision of assistance need not involve communication between D and P, encouragement by its nature involves some form of transmission of the encouragement by words or conduct, whether directly or via an intermediary. An un-posted letter of criminal inspiration would not be encouragement unless P chanced to discover it and read it. Similarly, it would be unreal to regard P as acting with the assistance or encouragement of D if the only encouragement took the form of words spoken by D out of P's earshot. 50
Perhaps the most important difficulty with Calhaem is that an event can be causally relevant to an outcome without being a but-for cause. This is hardly a novel proposition, but it is one recently approved at the level of the Supreme Court in respect of the Covid pandemic and insurance cover. 51 Within complicity, the Law Commission has taken the view that while in most cases, there will be no issue as to whether P was aware of S's encouragement, a rule requiring at least awareness of the encouragement was necessary for where there was doubt. They gave the example of words shouted by D, a football spectator, in a large crowd some distance away from an incident on the pitch involving P: in this situation, the prosecution would have to prove that P had been aware of the encouragement. 52
A case also sometimes cited to support the argument that causation is needed in complicity is Luffman in 2008.
53
The accessory was alleged to have counselled the principal to commit murder. The Court of Appeal held that, following Calhaem, there must be some connection or causal link between the counselling, aiding and abetting and the commission of the offence… whether that causal link can be described as substantial or simply a connection, in our opinion there was sufficient evidence before the jury upon which it could properly conclude that such a causal link or connection was sufficiently substantial to satisfy this requirement.
54
The court raised a formulation of ‘material contribution’ to an offence: the accessory's conduct should be ‘treat[ed]…as materially contributing to the commission of the offence, whether or not P would have acted in the same way without such encouragement or assistance’. 56 The idea of material contribution is known to tort law, covering situations where a defendant's conduct was not necessarily a but-for cause but influenced in a measurable way the outcome. 57 It has a niche use in tort law in narrow situations where requiring the claimant to prove that but for the defendant's wrong, the claimant would not have suffered the loss, would lead to unacceptable results.
Toulson LJ went on: D's conduct must have some relevance to the commission of the principal offence; there must, as it has been said, be some connecting link. The moral justification for holding D responsible for the crime is that he has involved himself in the commission of the crime by assistance or encouragement, and that presupposes some form of connection between his conduct and the crime.
58
Toulson LJ thus went further than just countenancing a ‘material contribution’, he also considered what might happen after the accomplice acted. He raised an idea that would, in Jogee, find expression as the defence of an overwhelming supervening event: There may be cases where any assistance or encouragement provided by D is so distanced in time, place or circumstances from the conduct of P that it would be unjust to regard P's act as done with D's encouragement or assistance.
59
We are left, in the words of Toulson LJ (and Glanville Williams), trusting to the good sense of juries as triers of questions of fact as well as the fairness of prosecutors in selecting charges. This is quite a surprising amount of trust since the logical consequence is that they are given no directions on the meaning of ‘aid’ nor how to apply it to the case in hand. In effect, the case law does not offer clarity and we are forced to leave the matter to ‘common sense’ in laypeople and prosecutors; again, prioritising what some see as a practical approach over doctrinal or intellectual rigour: Whether D's conduct amounts to assistance or encouragement is a question of fact. Professor Glanville Williams commented in Criminal Law: The General Part, 2nd ed (1961), p 356, that it is sometimes difficult to know what degree of assistance is to be regarded as aiding. Several centuries of case law have not produced any definitive legal formula for resolving that question. This is unsurprising because the facts of different cases are infinitely variable. It is for the jury, applying their common sense and sense of fairness, to decide whether the prosecution have proved to their satisfaction on the particular facts that P's act was done with D's assistance or encouragement (subject to the qualification that if no fair-minded jury could properly reach that conclusion, the judge should withdraw the case).
62
There is a further, instrumental, reason for complicity, in that it might provide ways to convict despite problems of evidence. Evidence can be difficult to obtain for many offences, but it might be particularly difficult in complicity cases. Multiple parties involved in the crime might make the evidence opaque. Hence a defendant can be convicted without it being proven whether she or he was an accessory or a principal. This is particularly valuable for a group of offenders, and especially for unplanned crimes where there might be no earlier trail of evidence. However, this kind of reasoning is not a justification, just an intuitive appeal to convict because it suits a particular policy. Lord Hutton came close to saying as much in R v Powell and Daniels; English in 1999, 64 when he referred to ‘practical concerns’ and the ‘need to give effective protection to the public against criminals operating in gangs’, both of which overrode ‘considerations of strict logic’. The difficulty is how it protects the public to be able to easily convict of the most serious crime that took place those who are not meaningfully connected to it.
Academic Engagement with Causation in Complicity
Academics have presented a range of options for the core justification of complicity. This paper does not claim to be building a new justification, and indeed, aligns with a significant body of existing work. It attempts to show how a concrete proposal for recognising a significant contribution to the principal's crime as a requirement of liability for assisting or encouraging would improve specific parts of the law. This might be useful across the wide range of roles that an accomplice might perform in connection with a principal. 65
Many commentators have argued that some level of causative contribution is needed. Perhaps most fully articulated was KJM Smith' argument that there has been some residual sense of consequence and linkage between the accessory and the principal: It has always been implied in the concept of complicity that [D's] involvement … did make some difference to the outcome and as a consequence of this, accessories have been implicitly linked to the harm element in the principal offence.
66
By contrast, Kadish frames the accessory's contribution not in terms of causation about the crime, but rather in terms of influencing the will of the principal. 71 He rejects the possibility of describing the accomplice as causing the action of the principal, instead the principal ‘freely and voluntarily chose to act’. 72 Other arguments in this direction have been made, at times including a preference for inchoate liability. 73
Finally, Rebecca Williams has recently argued for the importance of causation in complicity, perhaps as a form of ‘semi-causation’. 74
There is another group who point out some difficulties with how to apply causation within complicity. A leading critic was Glanville Williams, and that criticism was recently amplified by Catarina Sjölin.
75
Williams wrote in 1989 that: Principals cause, accomplices encourage (or otherwise influence) or help. If the instigator were regarded as causing the result he would be a principal, and the conceptual division between principals…and accessories would vanish. Indeed, it was because the instigator was not regarded as causing the crime that the notion of accessories had to be developed.
76
Williams’ expression of the issues was quoted with approval by Lord Bingham in one of the leading criminal law cases on causation, R v Kennedy (No 2). 78 Williams seemed to prefer to talk in terms of an accessory influencing, persuading or inspiring, rather than causing the principal's act. 79
The claim, made by Williams, Bingham and Sjölin, that to require causation by the accessory would be to make the accessory into the principal, and therefore must be resisted, seems appealing at first. It provides something that looks like a distinction between principal and accessory. However, it might be doubted on both practical and theoretical grounds.
First, it must not be forgotten that that does not seem to be a distinction the rest of the law of complicity values. For over 170 years English law already treats accessories as if they were principals for trial, labelling and sentencing, so the only difference remaining is in the test for being a principal or an accomplice. That is, as discussed above, one that seems to turn first on causation, and to a lesser extent, as discussed below (under ‘The Importance of Language in Complicity’) on the fault elements. And that difference does not therefore have any effect in practice. More generally, not requiring causation at all simply makes it easier to convict, and that person will then be treated as a principal. This is hardly protecting the value of a difference between principals and accessories.
Second, on theoretical grounds, if the definition of a principal's role in causation is imprecise, it is not obvious that it formed the grounds for differentiating complicity and principalship. Instead, there is an opportunity now for different levels of causation to ground those two forms of participation. Picking up what Lord Kerr said in Gnango, causal contribution might be shared between co-principals. There seems to be no reason in principle that accessories might not also causally contribute. Indeed, if they do not it seems very difficult to meaningfully say they have assisted/aided the offence, and impossible to say they have procured it. The only possible place where causation could be said to be irrelevant seems to be encouraging/abetting/counselling, but even there a role for significant contribution is needed for the law to make sense. Put another way, if the defendant can show that his words did not contribute to the principal's crime at all, or even made it less likely, how can they meaningfully be said to be encouraging, abetting or counselling it? The key step seems to be to accept that a person can make a significant contribution to the crime of another, but at the same time not be liable at all without further elements being satisfied. This takes us onto the relevance of language to claims about complicity.
The Importance of Language in Complicity
As noted already, the language we use to describe wrongdoing is an integral element of the existence of that wrongdoing, for the public, the defendant, the victim(s), and the wider justice system. The labelling of the offence in criminal law is important. Less commonly remarked upon is the need for the right language for the constituent parts of liability. That applies for components of liability like fault terms including intention and recklessness no less than elements like causation and consent. For present purposes, it is the language of complicity that is important. Four examples of the importance of phrasing complicity correctly will be considered here: (a) the wrong in complicity, (b) the limits of complicitous verbs, (c) the forms of complicity and (d) distinguishing between forms of complicity.
The Formulation of the Wrong in Complicity
The first example of the importance of language in complicity is what the accomplice is connected to: the principal, or the crime. ‘It is the assistance, not the ultimate crime, that must be intended by the secondary party’, or so the English Court of Appeal wrote in Bryce in 2004.
80
There a person was liable as an accessory for transporting a principal to the location of an attack, even when trying to dissuade that attack. Nonetheless, the judgment of the Supreme Court in Jogee was a recent example of a court noting that in practice, fault components are often described in terms of P's crime, even if strictly they apply to P herself: If the crime requires a particular intent, D2 must intend to assist or encourage D1 to act with such intent. D2's intention to assist D1 to commit the offence, and to act with whatever mental element is required of D1, will often be co-extensive on the facts with an intention by D2 that that offence be committed. Where that is so, it will be seen that many of the cases discuss D2's mental element simply in terms of intention to commit the offence. But there 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.
81
In cases of concerted physical attack there may often be no practical distinction to draw between an intention by D2 to assist D1 to act with the intention of causing grievous bodily harm at least and D2 having the intention himself that such harm be caused. In such cases it may be simpler, and will generally be perfectly safe, to direct the jury (as suggested in Wesley Smith and Reid) that the Crown must prove that D2 intended that the victim should suffer grievous bodily harm at least. However, as a matter of law, it is enough that D2 intended to assist D1 to act with the requisite intent. That may well be the situation if the assistance or encouragement is rendered some time before the crime is committed and at a time when it is not clear what D1 may or may not decide to do. Another example might be where D2 supplies a weapon to D1, who has no lawful purpose in having it, intending to help D1 by giving him the means to commit a crime (or one of a range of crimes), but having no further interest in what he does, or indeed whether he uses it at all.
82
Stating that the defendant was an ‘accessory to murder’ is a simple and common shorthand. The full form, something like ‘accessory to the principal committing murder’ or ‘assisted/encouraged the principal to commit murder’, is clearly more cumbersome. English law has adopted the derivative liability principle, thus in general the accessory's liability is tethered to the principal's. As the Supreme Court note, in group violence cases there might be little harm in the slip of language that the accessory is connected to the crime not to the principal. However, that slip can risk further imprecision which does matter. The error typically starts by saying that the accessory has to have a fault element about the crime itself. The Supreme Court even noted that risk in the text above, but their judgment was not as clear as it could have been on the issue. As the example of supplying weapons shows, the accessory need not intend that the crime take place. The accessory might have no interest in it at all and might even desire the crime not to take place. 83 This kind, what might be called a ‘disinterested assistor’, was central to the case of National Coal Board v Gamble, 84 cited with approval in Jogee. 85 However, if the accessory assists or encourages the principal, intending to assist or encourage the principal, knowing ‘any facts necessary to give the principal's conduct or intended conduct its criminal character’, 86 that is enough liability linked to the principal. For example, if the principal commits a homicide offence, this is enough or manslaughter. To be liable for more, attention turns to a further element of the accomplice's fault. If the principal's crime is one requiring fault, the accessory must intend for the principal to act with that fault, as noted in the first quote above.
In addition, this language issue becomes more pressing if the fair labelling difficulties with labelling an accessory as having committed the offence are taken seriously. One option would be to replace the Accessories and Abettors Act 1861 with a secondary party is convicted of the offence of ‘being a secondary party to X offence’. In form it would mirror the inchoate liability under, say, s. 1 of the Criminal Attempts Act 1981. For example, D was a secondary party to murder in the same way that D attempted to commit criminal damage. This has the advantages that you better label the wrong and allow flexibility of sentencing. This approach is tried and tested in other jurisdictions, and has long been known to English lawyers. 87 The same principle was inherent in plan of the Law Commission, in proposing what became the Serious Crime Act 2007. There, under the heading of ‘parity of culpability’, 88 at least in the sense that without S intending P to commit the offence, only liability for ‘encouraging or assisting’ the offence should be available. This might require a fact-finder, including a jury to determine whether D was a secondary party or a principal, which is no bad thing.
The Limits of Complicitous Verbs
We turn now to purported problems of language legal actors which affect participation generally, and complicity in particular. This brings us to a discussion of when you can act through another, and what you can intend another to do. Both are important in setting limits on what causative role an accomplice can perform, as well as the underlying argument about the importance of language in complicity.
What Can You Do Through Another?
The first example is about what one can do through another. The doctrine of innocent agency applies where the human agent of the physical elements of the offence was ‘innocent’, that is, lacking in capacity and/or lacking fault. When it applies, it turns the human being who used the innocent agent into a principal. The most common example is of delivery: the postperson who delivers a blackmailing or explosive letter or the child administering a poison according to the plan of the mother. 89 The innocent human agent falls out of the legal picture, much as the baseball bat used to kill is irrelevant to the liability of the hitter.
Some have argued 90 that the English language does not always permit a person to be described as performing a verb through another. The reasoning appears to be about the limits of the concepts involved in the relevant verbs. The fact patterns raised by English courts have been having sexual intercourse and driving. Those relevant verbs, at least in their modern forms, are ‘penetrate’ and ‘drive’. The claim is that it cannot be said that ‘A, using B's penis, penetrates C’; nor can it be said that ‘A drives using B's body at the wheel of a car’. It is not clear whether the claim is that one can never perform conduct through another. That would be the broadest claim, since the other large class of offences, result offences, can clearly have their ends achieved through the use of an innocent agent. A can, by packaging a bomb delivered by a postwoman, B, cause the death of C.
It is difficult to be sure of the logic underlying this claim about the limits of some verbs. It appears to be a difficulty with an unknown number and range of verbs being used indirectly, mediated through another human agent. Agency like that in everyday and non-criminal use might not have featured these kinds of fact patterns, leaving a sense of unease in the use of the verbs. Agency, in its wider and lay sense, of one acting on behalf of another, might not normally in everyday usage have included sexual activity. So too, driving seems to involve more individual voluntariness than might in common parlance be described as permitting the control to be another's. The same might be true in other languages, particularly cognate ones. This is not an instance where the claim is articulated that legal English must track common usage; it is most commonly stated or assumed simply that the verbs do not permit this use at all. No evidence for this restrictive reading is typically cited. And even if verbs were so restricted in the past, it does not seem to be necessary now. No confusion would be caused using verbs indirectly, or mediated through another agent, if used carefully and appropriately, as all words should be in a legal context.
If the claim, that there are some actions which one cannot commit through another, is treated as correct, innocent agency will be ruled out and such situations will have to be handled through complicity. In particular, English law developed an exceptional form of complicity, procuring the physical components of those offences. 91 Since the principal was not liable because of lacking the fault required, the principle of derivative liability would prevent the accomplice being liable. To get around this problem, English law accepted the possibility of the accomplice procuring something less than the full offence. The classic example is Cogan and Leak, 92 where D1 tricked his friend, D2, into thinking D1's wife consented to have sex with D2. D1's wife did not consent; D2's conviction for rape was quashed as he, having believed his friend, was not reckless about the wife not consenting while D1's conviction was affirmed. However, it was expressly decided on ‘justice and common sense’, and because D1 admitted liability as a procurer, not on technicalities of pleading. In this particular case, the relevant offences have changed since Cogan and Leak was decided. As of 1 May 2004, to avoid liability for rape, D2 would have had to have a reasonable belief in consent: s. 1(1)(c), Sexual Offences Act 2003.
However, after Jogee there is the difficulty that the fault element in complicity has changed and it is unlikely that the procurement route is still available. While Jogee itself did not discuss procuring, it treated the discussion of the fault element of assisting and encouraging as if it were universal. If the same rules apply to procuring, then the accomplice will now have to intend that the principal have the relevant fault element for the offence (as shown in the quote from Jogee on the preceding page). In a fact pattern like Cogan and Leak that will mean the husband would have to intend that his friend not have a reasonable belief in the wife's consent. That was clearly not what he intended on the facts as understood. 93 The fact pattern leads us to infer that the friend would not have had sex with the wife without the lie making him believe the wife consented. Alternative scenarios, like the lie simply made the sexual encounter more exciting for the friend, do not seem plausible on the facts. In the particular case of sexual offences, an alternative offence might be found, such as s. 4 of the Sexual Offences Act 2003, causing a person to engage in sexual activity without consent. The logic might be that the husband causes his wife to engage in sexual activity without consent and without reasonably believing in her consent. It might also be possible to argue that the husband commits the s. 4 offence against the friend. It seems doubtful the husband reasonably believes in the friend's consent, when the husband has lied about the wife's consent in order to trick the friend into having sex. While this might provide some coverage for the facts of Cogan and Leak it would clearly be labelling less transparently the wrong, and invoking less stigma. The s. 4 offence has a much wider range of conduct, from a woman forcing a man to have sex, to making another person watch pornography. Of course, more generally, other offences with a ‘causing another’ form like this are not widely available. The result would be that if innocent agency does not accommodate the having sex through another, and procuring does not any longer, there might well be no offence committed. The best solution is to confront, and reject, the argument that the English language does not permit the legal sense of ‘a person has sex through another’.
Similarly, what of where the wrong only exists because of some quality of the person who actually carries out the physical components? Thus, the offence of bigamy is constituted when D marries during the course of the lifetime of his wife. If P, unmarried, tricks D into marrying again while D's wife still lives, it is difficult to say that P has ‘married during the course of the lifetime of P's wife’ since P has no wife. The principal is a person who brings about the commission of the physical components of the offence, it does not matter whether those components are results, conduct, or even states of affairs. Nor need P have a wife, only that he causes D to marry during the course of D's wife's life. Thus, the doctrine is not deeming S a principal to avoid the limits of language, it is recognising the mechanics of principalship.
What Can You Intend About Another?
The second example of a problem of linguistics is what one can intend. There is a view within philosophy of action, that ‘intention’ can only meaningfully be used in the description of a person's own body or which that person can sufficiently control.
94
This view holds that it would not normally be possible to say that ‘D intends P to kill’ nor indeed, that ‘D intends P both to kill and to intend to kill’. The argument starts by saying that P can only intend to control his own body, and D is a separate human body. It concludes by offering control in some other sense. However, outside of situations where D does in fact control P, which would be addressed through procuring, that control ability would be absent between D and P. In commenting on Jogee, Simester has argued in exactly the same terms: Save in special cases of causing or procuring, S cannot intend that P will do X. S can of course act in order to help or encourage P to do X: S can intend his or her own act of helping or encouraging P's action. But S can’t intend P's action. Neither can S intend P's mens rea.
95
This is deep philosophical water. It might help to separate intention into bare intention, without any act, intentional action, of doing something intentionally or intending to do so, and further intentions, which expands beyond intentional action, that is, it extends to the ends towards which present actions serve. 96
The law currently has no other terms but intention to capture purpose with regard to an object within the canon of mens rea. If we wish to describe the accessory's mental state as a purpose about the principal's conduct and mental state, intention is the only word we have. If ‘intention’ is inapt to describe an accomplice's mental state about the principal's acts and mental state then that would leave a gap. Note, the idea of ‘intending to help’ another, which the law still requires in any case, is not directly about the mental state of the other, it is about the accessory's mental state about the effect of the accessory's conduct.
One option is to describe A's conduct in a different formulation. The closest we have is offences where an additional fault element is added on to conduct, but not in the common method of an ‘ulterior intent’, like ‘intending thereby to injure, aggrieve or annoy’ in s. 24 OAPA 1861. The main example of this form of non-‘intention’ purpose in mens rea is a formulation about purpose, in particular ‘for the purpose of’. This formulation of the fault of a person is relatively rare. One example is in the offence of voyeurism, ‘for the purpose of obtaining sexual gratification’ in s.67(1) Sexual Offences Act 2003). This phrase does not perform the same semantic work as ‘intention/intend’, and its core use is in offences where there is already some conduct and other mental state already engaged. However, it does lead to a similar outcome. One problem in using it is how rare the formulation is in existing law. The other problem is that if we were to start using it here, it would be unclear why we do not use it in the rest of the law. The argument might go that, if intention is meant to capture purpose, why not just refer to ‘purpose’ throughout. This might be worth considering. Nonetheless, while using a mens rea term of ‘intention’ clearly has some problems, but it is unlikely that anytime soon we could replace it.
Similarly, while a new term might be found, though inserting it into the criminal law might be difficult, since English law has not added a new fault concept for nearly a century, not since recklessness was inserted (albeit that it was not given its present two-limbed form until the late 1970s). 97 It is also not clear what this term would be.
At least for the moment, the better option is to consider whether the linguistic limits of ‘intention’ are, ultimately, a choice. Just like Glanville Williams having a preference for influencing, persuading or inspiring, rather than causing as noted above, we are delimiting between shades of meaning. The legal sense of ‘intend’ might be wider than the philosophical sense. It might, for instance, match lay use, where ‘intend’ and ‘purpose’ might merge more. Lay definition is indeed a standard touchstone in law generally. A football fan might have as his purpose that his team score. He might grasp that purpose in his mind and even through it at the television, where he sees the image of his team's player approaching the opposition goal. The fan might be described as wishing, hoping, desiring, wanting and perhaps even praying that his team score. In the loosest sense, football teams are thought to play better when their supporters support them; in the most direct sense, this was the will of the individual supporter. A simple comparison of prayer might be revealing. In the lay sense, it might be thought of as a request, expressing a desire. It is done with an acknowledgement that the entreaty might not be granted but typically with the belief that it might be. That seems to track many of the component elements of an intention.
The same logic looks unremarkable elsewhere in the criminal law, such as within impossible attempts. Under the Criminal Attempts Act 1981, s. 1(2), a person is liable for attempting to carry out a crime that is in fact impossible to commit. This includes something being impossible in the way the defendant attempted. Section 1(3) makes clear that even if the intention to carry out the course of conduct which would amount to an offence is in some way defective, it will still be sufficient if the facts were as the defendant believed them to be. The logic seems to be that the defendant must believe that his means can influence the outcome of the attempt. In the same way in complicity, when assisting, encouraging or procuring the principal, the accessory must intend to assist, encourage or procure. What makes the accessory liable for, say, murder, is his intention to be an accessory to murder. That involves the accessory intending that the principal intend to kill or intend to cause grievous bodily harm, an intention that the accessory must have while encouraging, assisting or procuring.
Causal Claims in Complicity Language
Having addressed the limits of the language we use for verbs of complicity, we return to the underlying problem: that those verbs and the very language of complicity relies on causal claims. The opposite claim would seem ridiculous but for the fact that it is routinely made. How can A ‘assist’ B if B does not benefit in any way? How does A ‘encourage’ B if B is no more likely to commit, or interested in, the crime? To create a separate and technical legal meaning to ‘assist’ or ‘encourage’ might be permissible in certain circumstances, but to strip them of a core of their lay meaning is illogical and damaging to the criminal justice system generally.
That the terms assist and encourage, no less than procure, require a form of causation is also supported by other rules in the same context. One option would be to prosecute under the Serious Crime Act 2007, Part II. Encouraging or assisting a crime, intending or believing it would take place under ss. 44–46 of the Act covers conduct which is capable of encouraging or assisting a crime. There are some ill-defined limits to what that covers, along with some difficult issues of impossibility under the Act, but the very fact that the Act had to use language of ‘capable of encouraging or assisting’ in those sections, makes it clear that it sought a wider scope than normal accessorial liability. An act which in fact could not assist or encourage should not be included within the scope of liability there, just like it should not be sufficient for complicity.
Indeed, Parliament removed the possibility of liability for attempting to assist or encourage in the Criminal Attempts Act 1981, s. 1(4)(b). However, if complicity does not require any proof of causation, the criminal law seems at the least to be subverting much of the effect of that Parliamentary rejection. That is, if the assistance or encouragement need not have any causal effect within complicity, there is much less pressure to consider liability for doing acts which were more than merely preparatory to assistance and encouragement.
The place that these causal claims are strongest is in procuring, but there are further nuances to the meaning of these terms. We therefore turn to consider the different forms of complicity.
Distinguishing Between Forms of Complicity
The modern use is to truncate the ‘aid, abet, counsel and procure’ found in the Accessories and Abettors Act 1861, s. 8. These four physical element words are often, in modern language, summarised as liability for one who ‘assists or encourages’. There ‘aids’ equates to ‘assists’ without difficulty. However, ‘abets, counsels or procures’ is mostly, but not completely, equivalent to ‘encourages’. In particular, while presence at the scene of the crime is the only thing that distinguishes abetting (present) from counselling (absent), procuring could cover more than encouragement, in that it need not involve communication with the principal. For that reason, while three modes are practically important (excluding abetting as a separate mode), all four will be discussed below.
More than simply a matter of language, the content of each term has had a significant impact on the development of secondary liability more broadly, especially its fault element. That is, different modes of being a secondary party tend to provide different justifications for liability as well as different standards of, and evidence for, fault. The nuance that we need to engage in here is in respect of what has been called ‘consensus’, a ‘meeting of the minds’.
Perhaps the clearest exposition of the four modes of accessory liability came in Attorney General's Reference (No.1 of 1975) 98 the Court of Appeal held that each of the four words, aid, abet, counsel and procure, had a separate meaning. S had ‘laced’ his friend's drink, and when the friend drove home, he was stopped by the police and ultimately convicted of driving with an excess quantity of alcohol in his body under section 6(1) of the Road Traffic Act 1972. The Attorney-General was particularly motivated by the fear that a person might claim to have laced the drink in order to avoid a friend being disqualified from driving under section 93(1) of the 1972 Act: if that claim had legal consequences, particularly liability as an accessory, such claims and the fear of them distorting the justice system, would subside. Liability as an accessory could only flow from the Accessories and Abettors Act 1861. The Act was being interpreted in an unusual context, since unlike the majority of accessory liability situations, in a ‘lacing’ case, S would not have to have contact with P before the offence took place, to the extent that P might be completely unaware of S's actions.
Lord Widgery C.J., giving judgment for the Court of Appeal, agreed in large part with the trial judge who: took the view that in the absence of some sort of meeting of minds, some sort of mental link between the secondary party and the principal, there could be no aiding, abetting or counselling of the offence within the meaning of the section. So far as aiding, abetting and counselling is concerned we would go a long way with that conclusion. It may very well be, as I said a moment ago, difficult to think of a case of aiding, abetting or counselling when the parties have not met and have not discussed in some respects the terms of the offence which they have in mind. But we do not see why a similar principle should apply to procuring. We approach section 8 of the Act of 1861 on the basis that the words should be given their ordinary meaning, if possible. We approach the section on the basis also that if four words are employed here, ‘aid, abet, counsel or procure’, the probability is that there is a difference between each of those four words and the other three, because, if there were no such difference, then Parliament would be wasting time in using four words where two or three would do. Thus, in deciding whether that which is assumed to be done under our reference was a criminal offence we approach the section on the footing that each word must be given its ordinary meaning.
99
procuring requires causation but not consensus; encouraging requires consensus but not causation; assisting requires actual help but neither consensus nor causation
100
Turning now to each of the four historic terms: aiding, abetting, counselling and procuring, it is evident that their common meanings lend themselves more easily to certain fault element configurations and justifications for liability and raise questions for causation too.
There was not accidental presence in this case. The appellant paid to go to the concert and he went there because he wanted to report it. He must, therefore, be held to have been present, taking part, concurring, or encouraging, whichever word you like to use for expressing this conception. It was an illegal act on the part of Hawkins to play the saxophone or any other instrument at this concert. The appellant clearly knew that it was an unlawful act for him to play. He had gone there to hear him, and his presence and his payment to go there was an encouragement. He went there to make use of the performance, because he went there, as the magistrate finds and was justified in finding, to get ‘copy’ for his newspaper.
105
To procure means to produce by endeavour. You procure a thing by setting out to see that it happens and taking the appropriate steps to produce that happening. We think that there are plenty of instances in which a person may be said to procure the commission of a crime by another even though there is no sort of conspiracy between the two, even though there is no attempt at agreement or discussion as to the form which the offence should take. In our judgment the offence described in this reference is such a case.
109
Indeed, the line between procuring and counselling/encouraging can be difficult to draw. In practice, it is not normally necessary to do so, and hence how the Supreme Court in Jogee sought to avoid even discussing procuring. However, we continue to strain the normal meaning of these terms, and their doctrinal clarity, by eliding them when we could do better. We can do better if we consider the notion of consensus referred to already or perhaps even better, ‘furthering the purpose of the principal’. To counsel or encourage, the normal meaning of encourage requires a positive effect on the principal. More than that, the positive effect must be in line with the aims of the principal. It must further the purpose that is then carried out by the principal. If the encouragement involves a change of the principal's plan, that change must be accepted by the principal. By contrast, procurement needs neither communication from the accessory to the principal, nor any sense of the accessory furthering the purpose of the principal. An example might help. A, a scout, sees B, an elderly person, who was standing by a pedestrian crossing, looking from one lane of traffic to the other. A, determined as a scout to support others, assertively takes B's arm, and walks B across the crossing. Does A ‘help’ B? Arguably, only if it was B's purpose to cross the street. If B is in fact watching the road for a car to arrive to collect B, crossing to the other side might not further that purpose at all. It could, however, be procuring: A might be said to procure B crossing the road. The difference is that procuring does not require furthering B's purpose. Similarly, if A merely shouted ‘Go on, cross the road’, does A ‘encourage’ B? First, B might not hear A, so there would be no encouragement even on the current interpretation of the rules. Second, B might not care what A thinks, and the encouragement might have no effect. Indeed, it is even possible that A's comment makes B less likely to proceed; that B might still proceed cannot make A's comment into encouragement. And third and finally, A only encourages B if A shares the aim of B crossing the road.
To summarise the current way complicity is conventionally said to apply: would look like this.
The better position would be to articulate the requirement that, outside of procuring, of ‘significant contribution’ is needed. It would make the law clearer, reducing the risk of misunderstanding and the transaction costs in applying it. It would better justify complicity liability, as well as align with the language the nature of the offending relies on to be understood and to convey censure. It also helps to avoid unjust results. In particular, it meaningfully links the liability a person faces for their responsibility and culpability for the actions of another. That would be expressed in the following way:
In these diagrams, ‘furthering’ is used to indicate that the accomplice is supporting the plans of the principal. Common Purpose (‘CP’) is not needed for any form of complicity. It is sufficient, but it is not necessary. The accessory need not intend the crime to take place. The next section turns to consider what ‘significant contribution’ is.
111
Threshold for Complicitous Conduct
The paper so far has shown that even though there are weak normative foundations for the law of complicity as it is now, it has adopted a doctrinal position which does not make use of the most normatively defensible set of requirements for liability. Some minimum level of contribution by the accessory to the principal's crime must be required. Without it, we would be in conduct crime or inchoate crime and that is clearly not how the law is treating accessories, for trial, sentencing or labelling. The criminal law is certainly capable of employing a non-but-for form of causation and complicity is the only place where that would be justified so far. The weak normative justifications can support that, at least.
Now is the ideal time to bring some rigour to the contribution that an accessory must make to the principal's crime. This is a step judges can take immediately. This would be a much smaller change than those in Jogee, and that change has by now clearly been shown not to have unbalanced the criminal justice system. It is also on a part of the law which Jogee did not have to rely on, so it does not undermine the UKSC's decision. It would also mark a small step towards improving one part of the criminal law which is practically important as well as doctrinally difficult. It would offer hope that other lax formulations of the criminal law might be improved over time. More generally, we have already seen that if errors like this are not corrected, they tend to propagate. The model seems to be this: (a) create form of criminal liability labelled as if there were a causative contribution to a crime; (b) deny that that causal contribution is needed; and (c) where some effect of that need for contribution is felt, in effect, irrebuttably presume that a causative role was played. This pattern even played out in casting the net of criminal liability even wider pre-Jogee too. Parasitic Accessorial Liability was, in part, based on the idea of escalation, that one crime leads to another. 112 That escalation was then the very thing that the formulation of PAL assumed, without testing for it. PAL has been removed from the law, and such sloppy underlying reasoning should be removed as well.
The best position is a two-part approach: to be liable for assisting or encouraging a crime, the accomplice must make a substantial contribution to the principals’ commission of it; to be liable for procuring the principal's crime, the accomplice must bring the crime about. Put in reverse, why should we be criminalising a purported accessory for an insubstantial contribution? The only reason seems to be that criminalising more easily is a good thing. To defend that on deontological grounds would be very difficult; to defend it on instrumental grounds would need a lot of evidence, that has so far not been shown. If there was no contribution, then we are dealing not with derivative liability, but inchoate liability; that is, complicity is not the appropriate area of criminal law but there might be an offence elsewhere, particularly under the Serious Crime Act 2007.
The change to recognising ‘substantial contribution’ would be open to courts to make as an incremental development in the law from Calhaem, Stringer and Mendez. We can pick up by way of example a case mentioned tangentially already: Attorney-General v Able. 113 The defendant had published a suicide manual to distribute to members of a euthanasia organisation and was subsequently charged with having assisted and encouraged suicide. On appeal, the Court of Appeal referring to the need for but-for causation when the form of participation alleged was procuring, but held that the same causal strength was not required for assistance and encouragement. 114 It held that ‘it does not make any difference that the person would have tried to commit suicide anyway’, so long as relevant person was assisted or encouraged to commit suicide. 115 The only additional element is that the causal contribution has to be significant. The jury would have had to be sure that there was assistance or encouragement of a level that was sufficient to merit conviction. Articulating that test a little more fully would benefit the jury in their work.
Alternatively this recognition could be given effect by legislation, but that does not seem necessary in fact. Indeed, legislation already uses this causative language elsewhere.
The phrasing of ‘significant contribution’ is used elsewhere in the criminal law, to highlight non-but for causes which are none the less relevant to our decisions about the relevance of an issue. The most obvious example is in respect of the partial defence of diminished responsibility, which leads to a conviction for manslaughter, not murder. 116 The Supreme Court has felt able to rely on the jury's good sense of what a sufficient enough contribution was. 117 This would be, like many things within complicity, is still a jury question. Jogee emphasises that in general terms. While there are reasons to re-consider how much we use juries to decide substantive questions of law, the level of contribution can meaningfully be given to a jury. Another phrasing, ‘material contribution’ is used in tort law, as already noted. However, that approach would normally lead to liability for the proportion of the harm caused, not the whole loss. That proportional causation approach does not fit the binary approach to liability for a given offence that the criminal law displays, and as such a different form of language would better mark out the role the test is playing within criminal law. “Significant contribution” uses language better suited to the role of the jury and chiming with existing language elsewhere in criminal law. It remains a jury question, but one where it might be possible over time to give examples if they were truly needed.
Note the new test is not formally a reversal of the burden of proof. Its application might nonetheless might in practice ask whether the evidence raises an inference of contribution. Doing acts which could and in the normal course of life would assist or encourage, might give the jury the sense that they did, unless other evidence is available. That would not be a formal reversal of the burden of proof, but it would highlight the decision-making the jury should engage in carefully, before convicting.
The End of Complicity: Overwhelming Supervening Events and Withdrawal
A final benefit of tightening the expression of the requirement of an accessory's contribution is that it might assist with working out the ending of a person's liability through complicity. The core issue is how causation, if formally irrelevant, can be relevant to some concepts designed to remove or reduce the accessory's criminal liability.
It has already been shown that the current approach to sentencing does not categorically differentiate the principal from the accessory as a matter of law, nor formally, as a matter of sentencing guidelines. The two factors that are raised in judicial discussion about sentencing accessories seem to be the responsibility, and the culpability, of each defendant. The responsibility seems to include any causative role for both principals and accessories, but responsibility's role in sentencing is less remarked upon in complicity cases. Tightening the law on the requirement of a significant contribution by the accomplice who assists or encourages is itself a measurable and valuable achievement. As noted above, it would bring even greater benefits if it fed through to raising the significance of contribution as expressing responsibility for the crime for the purposes of sentencing as well.
Aside from a conviction and sentencing, the two ways we are interested in are the newly formulated ‘overwhelming supervening event’ and the established defence of withdrawal. In a sense, they are the flip side of the same coin. In an overwhelming supervening event, defendants argue that their contribution has been eclipsed by something the principal did; in withdrawal, defendants argue that they themselves have removed their participation from the scope of liability. It would be absurd if only accessories who had causally contributed to the principal's crime could avail themselves of overwhelming supervening event or withdrawal. That does not appear to be what the limited case law says either. If it did those who had done less towards the crime would have fewer ways to reduce or remove liability, which is hard to justify. It would also raise doubts about the very function of the defences: if no responsibility for the principal's crime is required, such as when there was an overwhelming supervening event or withdrawal, why have defences which at least in part deny causation?
Overwhelming Supervening Events
The precise character of this concept is unclear, but in practice it operates as something the defendant must first raise. The defendant argues that the prosecution has not proven its case by raising evidence that there was in fact not assistance or encouragement (or surely, procuring) at the point the crime was committed. There might have been in the past, but overwhelming supervening events intervened. In Jogee's terms, there was, some overwhelming supervening act by the perpetrator which nobody in the defendant's shoes could have contemplated might happen and is of such a character as to relegate his acts to history.
118
Conversely, there may be cases where anything said or done by D2 has faded to the point of mere background, or has been spent of all possible force by some overwhelming intervening occurrence by the time the offence was committed. Ultimately it is a question of fact and degree whether D2's conduct was so distanced in time, place or circumstances from the conduct of D1 that it would not be realistic to regard D1's offence as encouraged or assisted by it. An early example is R v Hyde (1672), described in Hale's Pleas of the Crown (1682), vol 1, p 537, and in Foster's Crown Law (1762), p 354. This was Foster's description and explanation: A, B and C ride out together with intention to rob on the highway. C taketh an opportunity to quit the company, turneth into another road, and never joineth A and B afterwards. They upon the same day commit a robbery. C will not be considered an accomplice in this fact. Possibly he repented of the engagement, at least he did not pursue it. Nor was there at the time the fact was committed any engagement or reasonable expectation of mutual defence and support, so far as to affect him. In other words, on the particular facts A and B were not regarded as having committed the robbery with C's encouragement or assistance. Any original encouragement was regarded as having been spent and there was no other assistance. (It appears from Hale's account that C parted from A and B at Hounslow and that the later robbery took place three miles away.)
119
He did not assent that the daughter should be poisoned, but only that the wife should be poisoned, which assent cannot be drawn further than he gave it, for the poisoning of the daughter is a distinct thing from that to which he was privy, and therefore he shall not be adjudged accessory to it.
123
Plowden's commentary on the case in his report attempts to draw a distinction using other examples, set out below. His commentary is a good summary of the views of earlier commentators. Before turning to the specific examples he draws on, however, it is important to understand the difficulty he faced in drawing such a distinction. This is highlighted by how Plowden himself bookends his examples: Note, it seems to me reasonable that he who advises or commands an unlawful thing to be done shall be adjudged accessary to all that follows from that same thing, but not from any other distinct thing… But I greatly approve of the said opinions of the justices concerning the accessary in the case before reported, because the poisoning of the daughter was a distinct fact, to which Archer gave no advice nor counsel, and whose death he did not procure.
124
This example might highlight the continued relevance of the rule that an intentional departure from an agreed plan can end any liability the secondary party had, a rule that predates the modern formulation of overwhelming supervening event. 125
There is now appellate case law on overwhelming supervening events which addresses this issue,
126
in Grant, Khan et al in 2021.
127
The question was whether first running down a victim with a car was an overwhelming supervening event as a change of plan from the attack with weapons, which then happened and led to the victim's death. These are already rather unpromising facts for such a defence. Starting at [31], the Court of Appeal cited [12] of Jogee: Once encouragement or assistance is proved to have been given, the prosecution does not have to go so far as to prove that it had a positive effect on D1's conduct or on the outcome: R v Calhaem [1985] QB 808. In many cases that would be impossible to prove. There might, for example, have been many supporters encouraging D1 so that the encouragement of a single one of them could not be shown to have made a difference. The encouragement might have been given but ignored, yet the counselled offence committed. In our view, that passage is an insuperable obstacle to the suggestion that the concept of OSA should be viewed through the lens of causation. To the contrary, as the Supreme Court in the next two sentences in paragraph 12 explain, it is encouragement and assistance principally that count. As it seems to us, the Supreme Court in these paragraphs in Jogee significantly limited the circumstances in which a jury will need to consider the possibility that there had been a departure from the agreed plan. Although paragraph 98 is not expressed in absolute terms (viz. ‘(t)his type of case apart, there will
Instead, the UKSC's earlier reluctance to see as relevant a level of causation short of but-for, misled the Court of Appeal into rejecting the obvious mode of operation of the overwhelming supervening event concept.
Accepting that some causative link must be required, in the form of a substantial contribution for assistance or encouragement, neatly brings overwhelming supervening event within the mainstream and guides juries in their decisions. The question is first, what level of contribution through assistance or encouragement did the accessory provide. Then, second, did that persist to the point that the principal committed the offence? The alternative that Jogee perceives is that any contribution had been relegated to history, background, or otherwise been distanced in time, place, or circumstances from the conduct of the principal that it was no longer realistically a form of assistance or encouragement?
Withdrawal
Withdrawal is a complex area of law, with a deeply uncertain set of tests. While inchoate liability for assisting or encouraging under the Serious Crime Act 2007, conspiracy and attempt will be unaffected by a withdrawal, liability through complicity can be ended by the defendant in certain circumstances. It is generally accepted that a person can withdraw up to the point that the principal goes far enough to have committed an attempt. 130 The test is unclear largely because two elements are in tension: a reduction in the contribution to P's crime, and a sufficient removal of culpability.
First, the erstwhile accessory must ‘serve unequivocal notice upon the other party to the common unlawful cause that if he proceeds upon it he does so without the further aid and assistance of those who withdraw’, according to Beccera. 131 The appellant, B, broke into the house with two other men C and G. Their intention was to steal from the householder. While in the house, the tenant of a flat on the first floor surprised them, and B calling ‘let's go’ climbed out of a window followed by G and ran away. C, meanwhile, who had earlier been handed a knife by B, stabbed and killed the tenant. However, the conviction of B was upheld on appeal (G did not challenge his conviction). The actions of B and G were not enough to withdraw.
Second, in O’Flaherty,
132
the Court of Appeal clearly approved language of causation in deciding when an accessory had withdrawn. for there to be withdrawal, mere repentance does not suffice. To disengage from an incident a person must do enough to demonstrate that he or she is withdrawing from the joint enterprise. This is ultimately a question of fact and degree for the jury. Account will be taken inter alia of the nature of the assistance and encouragement already given and how imminent the infliction of the fatal injury or injuries is, as well as the nature of the action said to constitute withdrawal. In cases of assistance it has sometimes been suggested that, for there to be an effective withdrawal, reasonable steps must have been taken to prevent the crime. It is clear, however, this is not necessary. In R. v Whitehouse [1941] 1 W.W.R. 112, a decision of the Court of Appeal of British Columbia approved by this Court on several occasions (see R. v Becerra and Cooper (1975) 62 Cr.App.R 212 …) Soan J.A. stated (at p.115) that after a crime has been committed:
[I]n the absence of exceptional circumstances, something more than a mere mental change of intention and physical change of location by those associated who wish to disassociate themselves from the consequences attendant upon their willing assistance up to the moment of the actual commission of that crime. I would not attempt to define too closely what must be done in criminal matters involving participation in a common unlawful purpose to break the chain of causation and responsibility.
133
It is entirely unsurprising that breaking the causative link should be relevant to withdrawal since, as this article argues, substantial contribution should be needed to generate liability to begin with. A test for withdrawal is hard to set up in general terms, but the courts have made clear that the accessory normally need not prevent the crime. This is the flipside of the accessory's contribution not needing to be a but-for cause of the principal's offence at the start.
O’Flaherty is also useful as a reminder of the scope of ‘joint enterprises’, or what is a more specific name, common purposes. Where there was no pre-arranged plan but spontaneous violence by a group who jointly attacked common victims and the ‘enterprise’ arose on the spur of the moment, in many cases the scope of the joint enterprise would be ascertained by considering the knowledge and the actions of those participating, and juries would usually have to make inferences. 134 Where death had resulted, the jury should be directed that they must be satisfied (a) that the fatal injuries were sustained when the joint enterprise was continuing and that the defendant was still acting within that joint enterprise and (b) that the acts which caused the death were within the scope of the joint enterprise. A person who unequivocally withdrew from the joint enterprise before the moment of the actual commission of the crime by the principal should not be liable for that crime, although his acts before withdrawing might render him liable for other offences. The key test starts with removing his contribution, and continues to require a clear withdrawal to express his changed position of culpability.
In short, if we more fully acknowledge the requirement that encouragement be causal, then this helps clarify what contribution must be cancelled in order to successfully withdraw.
Concluding Contribution
This paper has suggested some small but significant changes to the operation of the law of complicity to bring it greater doctrinal and linguistic rigour. Building a criminal justice system on claims of practicality, but in fact calling on lay people to apply vague doctrine is too uncertain, too risky and too likely to produce unjust results across the jurisdiction. In particular, the apparent practicality of the current interpretation, that assisting and encouraging are non-causal, harms the value of the claim that the law of participation has. It exacerbates further the failures of fair labelling already in this area of law. The article proposes recognising that for an accessory to assist or encourage a principal, the accessory must make a ‘significant contribution’ to the principal's crime. These are terms and concepts well-known to the criminal law, and can be left to juries to apply. If the accessory has not even made a significant contribution, the liability inchoate, rather than being a form of participation, and should not be shoehorned into complicity. By putting slightly more attention into the doctrinal rigour of the contribution an accomplice has to make to the principal's crime, we improve the law of complicity generally. That includes both substantive legal issues, such as the test for being accomplice. It also includes matters of sentencing and ending complicity by clarifying overwhelming supervening event and withdrawal. These changes in complicity would also highlight the potential for small sequential changes to make a tangible contribution to the wider improvement of the criminal law.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship and/or publication of this article.