CASES VONNISSE

 

Clarifying the law of complicity S v Mbuyisa [2023] ZAKZPHC 132; 2023 JDR 4950 (KZP)

 

 

1 Introduction

The single perpetrator is the paradigmatic offender. Apart from very limited exceptions, all the elements of criminal liability, including both actus reus and mens rea, apply to the perpetrator, and a conviction cannot ensue unless the presence of each element is established, and the blameworthiness of the accused is duly considered to be proved beyond reasonable doubt. The problems posed by complicity (more commonly described as participation in South African law) emerge where more than one person is involved in criminal conduct, and yet, the rules governing criminal liability in respect of groups of two or more offenders do not admit of collective responsibility, but depend on the individual contribution of the participant to the criminal enterprise, and the application of the elements of liability to such contribution.

The question of the operation of the rules governing co-perpetrator liability and common purpose has frequently arisen in the courts and in academic writing on criminal law. These issues once again arose for consideration in the recent case of S v Mbuyisa ([2023] ZAKZPHC 132; 2023 JDR 4950 (KZP)), which is discussed below.

 

2 Judgment

2 1 Findings

The Mbuyisa case (supra) concerned an appeal against convictions of robbery with aggravated circumstances in the Pongola Regional Court. The first part of the judgment dealt with the problem of the trial record not being complete, and consequently whether the existing incomplete record was "adequate for a just consideration of the issues ... raised in this appeal" (par 8). After due consideration, the court concluded that the record was indeed adequate for this purpose (par 22) and, having set out the appropriate approach of an appeal court to a trial court's findings - essentially that a court of appeal will not overturn a trial court's factual findings unless they are shown to be wrong (par 23-27) - the court proceeded to examine the findings of the trial court. Upon assessing the evidence before it, the court on appeal upheld the correctness of the convictions of all four appellants. In respect of the first appellant, his positive identification by the complainant and his employee, and his arrest very shortly after the robbery, was determinative (par 42, 49). The fourth appellant contested his identification, which the State erroneously conceded to be invalid (par 80), but, based on its evaluation of the evidence, the appeal court confirmed the correctness of the finding of guilt by the trial court (par 77-85). The second and third appellants were similarly unsuccessful in their contention that they were wrongly convicted in the face of their decision not to testify, despite the extensive direct and circumstantial evidence against them (par 50-60). The argument on behalf of the second and third appellants was that there was insufficient evidence to link them to the crime. The court on appeal pointed out that there was no finding made by the trial court convicting the second and third appellants on the basis of common purpose (par 43), and that in fact their conviction was on the basis of co-perpetrator liability (par 56). The appeals of the four appellants were therefore all dismissed (par 90).

Notably, the State conceded the appeal of the second and third appellants on the basis that they were not aware of the reliance on common purpose (par 61). The appeal court was at pains to point out that the basis of the conviction of the appellants was merely their respective roles as perpetrators, and that "the doctrine of common purpose played no role in the decision of the trial court" (par 61). The analysis that follows assesses this issue.

2 2 Court's reasoning regarding the common purpose doctrine

As noted above, the court in Mbuyisa had no difficulty in dismissing the claims of the appellants that there was insufficient evidence to justify their convictions beyond reasonable doubt. The findings of the trial court were affirmed. However, the court was required to deal with the concession by the State on the merits of the appeal of the second and third appellants, on the basis that they had not been advised of the State's reliance on the doctrine, as evidenced by a failure to mention the doctrine in the charge sheet (par 63). Both the State and counsel for the appellants took the view that since the second and third appellants were not the "main perpetrators" and were not involved "with the wielding of weapons and the removal of items from the complainant's possession this necessarily means that they could only be convicted in our law on the basis of the doctrine of common purpose" (par 62).

As previously mentioned, the court did not agree with the approach of counsel in this case, and stated that common purpose was not the basis of any argument in the trial court, and was not even mentioned in the judgment of the trial court, where the court convicted the appellants as perpetrators (par 64). The court referred (par 65) to the case of S v Hlongwane (2014 (2) SACR 397 (GP)) for the following dictum:

"The starting point is that a person can commit an offence directly or vicariously through another and that where two or more persons agree to commit a specific crime, such as robbery, it is irrelevant what task each was assigned for its execution. Each is a co-perpetrator because he or she had agreed to commit the crime and either intended that force would be applied in order to rob or foresaw that possibility. Furthermore their agreement can be established through circumstantial evidence alone." (Hlongwane supra par 41)

The court pointed out (citing Snyman (Hoctor Snyman's Criminal Law 7ed (2020) 222-223 par 66-67)) that the legal principles governing co-perpetrator liability are well established, and that where a number of persons commit a crime, and all comply with the requirements for perpetrator liability, they are simply to be regarded as co-perpetrators. The court further cited R v Parry (1924 AD 401) in support of the fact that each accused acting together with another may be convicted on the basis of his own acts and mental state (par 68), before referring (Mbuyisa supra par 69-70) to the cases of S v Williams (1980 (1) SA 60 (A) 63A-B), S v Khoza (1982 (3) SA 1019 (A) 1031B-F) and S v Kimberley (2004 (2) SACR 38 (E) par 10) in order to illustrate the distinction between perpetrator, co-perpetrator and accomplice.

Finally, the court (Mbuyisa supra par 71-73) referred to Kruger (Hiemstra's Criminal Procedure (Service Issue 16 February 2023) 22-29, 22-30) to explain why it regards the State's concession as flawed. Quoting from this source, the court approved the discussion in this source as "logical" and "correct" when it stated that the common purpose doctrine is frequently applied where it is "unnecessary and inappropriate" (echoing the earlier comment in this source that the doctrine is "sometimes unnecessarily invoked"), which leads to confusion in both the legal principles and in the evaluation of the evidence by the person who is required to establish the facts (Mbuyisa supra par 71). The following words from Hiemstra's Criminal Procedure are highlighted by the court in this regard (Mbuyisa supra par 71): "The doctrine postulates as point of departure the absence of an agreement to commit the offence alleged", prior to citing the following passage:

"Common purpose thinking is irrelevant where an agreement to commit the offence has been proved by means of direct or circumstantial evidence or both. Botha JA's discussion in S v Mgedezi and Others 1989 (1) SA 687 (A) at 705I of the prerequisites for liability based on the doctrine is expressly based on the premise: 'In the absence of a prior agreement Holmes JA in S v Ngobozi takes as point of departure the absence of an agreement to murder. To invoke, as is sometimes done, common purpose in the case of a hired assassin is wrong in principle and calculated to confuse the judex facti. In such cases the parties are simply co-perpetrators, with the person hired as direct actor, and the person who hires as vicarious actor." (Mbuyisa supra par 72)

The court proceeded to point out the presence in the facts of this case of both circumstantial and direct evidence of a prior agreement to rob the complainant (Mbuyisa supra par 72), before citing a further example from Hiemstra's Criminal Procedure (22-30), which in the view of the court "makes the point even clearer":

"[F]ive robbers, all members of a gang, commit a bank robbery in the central business district in broad daylight. A sits waiting in the getaway car around the corner; B is the sentry across the road; C enters the bank with a suitcase in which to load the spoils; and D and E, both armed with AK47s, walk into the bank and open fire as they enter and fatally wound several bystanders. All five are guilty of murder, not as a result of a forced application of the doctrine but simply as co-perpetrators. Against each one the inference would be irresistible that he agreed that shots would be fired (by himself or one of the others), with the intent to kill bystanders or, at best for him, that he foresaw the real risk of such death and was indifferent thereto. Each of the members of the gang had the direct intent to apply deadly force in order to rob as to the murders there was thus, at the very least, intention by foresight of possibility (legal intention). Each fulfilled his agreed role in the execution of such intent. Each is thus a co-perpetrator in the commission of the murder, albeit vicariously in the case of those who did not directly participate in the shootings but nevertheless participated fully in the crime. In such case invocation of the doctrine of common purpose is superfluous. The correct result would be reached by a simple application of the principles of the law of participation on the given facts." (Mbuyisa supra par 73, emphasis added by the court)

 

3 Discussion

3 1 Development of terminology

While the notion of co-perpetrator liability has always constituted a part of the rules of criminal liability in South African law, it has assumed different forms as the rules of participation have developed. In describing the law of participation in Gardiner and Lansdown's work on criminal law, the discussion was, in all editions from 1917 to 1957, consistently placed under the heading "Principals and accessories before the fact", and the opening words remained the same:

"In English law a person who actually commits, or takes part in the actual commission of, a felony is a principal in the first degree, and a person who aids and abets the commission is a principal in the second degree ... South African law knows no such distinction: all persons who aid or abet in the commission of a crime are socii criminis, companions or partners in guilt, and are indictable and punishable as principals." (Gardiner and Lansdown South African Criminal and Procedure Volume I: General Principles and Procedure (1917) 81-82)

Although the term "principal in the first degree" was only mentioned in a few early judgments (see, e.g., R v Abrams (1880-1882) 1 SC 393 397-398), it seems clear that the co-perpetrator, who would be regarded as a joint principal in the first degree in English law (Kenny Outlines of Criminal Law (1902) 85), would fall within the broad notion of a socius criminis in South African law. This is exemplified by the Appellate Division decision in R v Parry (supra 401), where the court held (404, per Innes CJ) that "[b]y our law one who knowingly assists in the commission of a crime is a socius criminis and may be charged as if he were the actual perpetrator of the deed". The court in Parry elaborated that "[i]t is the existence of criminal intent in each of those who jointly commit a crime which entails on each a criminal responsibility" (406).

The imprecision associated with the application of the term socius criminis in the courts was lamented by De Wet, who notes that this "muddled terminology" ("verwarde terminologie") incorporated both the co-perpetrator and the accomplice (De Wet De Wet & Swanepoel Strafreg 4ed (1985) 198). Burchell and Hunt seek to distinguish between an "actual perpetrator" and the socius criminis in the first edition of South African Criminal Law and Procedure (Vol I: General Principles of Criminal Law (1970) 350), but this distinction is not aided by the attempt to link the categories to the English law equivalents, which are foreign to South African legal practice. The necessary clarification of terminology ultimately occurred in the Appellate Division, when the court authoritatively distinguished between perpetrator, co-perpetrator and accomplice liability:

"'n Medepligtige se aanspreeklikheid is aksessories van aard sodat daar geen sprake van 'n medepligtige kan wees sonder 'n dader of mededaders wat diemisdaad pleeg nie. 'n Dader voldoen aan al die vereistes van die betrokke misdaadomskrywing. Waar mededaders saam die misdaad pleeg, voldoen elke mededader aan al die vereistes van die betrokke misdaadomskrywing." (S v Williams supra 63A-B)

(Translation (Burchell Cases and Materials on Criminal Law 4ed (2016) 644): "An accomplice's liability is accessory in nature, so that there can be no question of an accomplice without a perpetrator or co-perpetrators who commit the crime. A perpetrator complies with all the requirements of the definition of the relevant offence. Where co-perpetrators commit the crime in concert, each co-perpetrator complies with all the requirements of the definition of the relevant offence.") (When this dictum is cited in the Mbuyisa case at par 69, the court provides its own translation.)

Ever since the Williams case, the nature of what a "co-perpetrator" entails has been settled in the law, and the concept has been applied in a number of cases (see, e.g., S v Frederiksen 2018 (1) SACR 29 (FB) and S v Tilayi 2021 (2) SACR 350 (ECM)). Indeed, the Mbuyisa judgment further endorses the application of this form of liability, and there is nothing to suggest that the court is incorrect in doing so, given the court's careful analysis of the facts of the case.

3 2 Conceptual concerns

3 2 1 The nature of common purpose liability

It is however submitted that the court's reliance on Hiemstra's Criminal Procedure in making its analysis of the common purpose doctrine requires closer scrutiny. (This analysis was introduced into this work under the authorship of Kriegler in Hiemstra Suid-Afrikaanse Strafproses 5ed (1993), and has been followed in this work ever since.) First, the statement that the point of departure of the doctrine is the "absence of agreement to commit the offence alleged" is difficult to reconcile with the otherwise uniformly accepted nature of common purpose liability in South African law (as expressed by the Constitutional Court in S v Tshabalala (2020 (2) SACR 38 (CC)):

"The liability requirements of a joint criminal enterprise fall into two categories. The first arises where there is a prior agreement, express or implied, to commit a common offence. In the second category, no such prior agreement exists or is proved. In the latter instance the liability arises from an active association and participation in a common criminal design with the requisite blameworthy state of mind." (par 48)

In adopting the approach propounded by Hiemstra's Criminal Procedure, the Mbuyisa court is dismissing the possibility that the common purpose doctrine can be based on prior agreement, clearly undermining the historical development of the common purpose doctrine, which initially only took the form of prior agreement (for further discussion, see Hoctor "The Genesis of the Common Purpose Doctrine in South Africa" 2023 26 Potchefstroom Electronic Law Journal DOI http://dx.doi.org/10.17159/1727-3781/2023/v26i0a163851-29).

Moreover, the court is required to deny the existence of the body of authoritative judicial precedent that has formed the basis of the increasing resort to the common purpose doctrine since the Constitutional Court declared the doctrine to be constitutional in the landmark case of S v Thebus (2003 (2) SACR 319 (CC)). In the Thebus case, Moseneke J, writing on behalf of the court, summed up the basis of common purpose liability as follows:

"The liability requirements of a joint criminal enterprise fall into two categories. The first arises where there is a prior agreement, express or implied, to commit a common offence. In the second category, no such prior agreement exists or is proved. The liability arises from an active association and participation in a common criminal design with the requisite blameworthy state of mind." (par 19)

This dictum was cited with approval in the judgment of Zondo DCJ (as he then was) in the Constitutional Court case of S v Jacobs (2019 (1) SACR 623 (CC) par 128 and 150). In addition, it has been followed, inter alia, in S v Gedezi (2010 (2) SACR 363 (WCC) par 49) and S v Tilayi (supra par 19). (For further discussion of the forms of common purpose, see Hoctor "Distinguishing the Forms of Common Purpose Liability - S v Govender 2023 (2) SACR 137 (SCA)" 2023 Obiter 913). More significantly perhaps, this understanding of the nature of common purpose liability has been fully accepted into the application of the doctrine in the Constitutional Court (see the recent judgment of Mathopo AJ (with which all the other judges concurred) in S v Tshabalala (supra par 48) (cited above), which repeated the words of Moseneke J in Thebus (supra par 19), without the need for attribution) and in the Supreme Court of Appeal (see, for example, the recent case of S v Pooe 2021 (2) SACR 115 (SCA) par 57).

In the passage cited by the court in Mbuyisa (supra par 65) from Hlongwane (supra par 41), which addressed the question whether the doctrine of common purpose applied in Hlongwane, that court also agreed that common purpose can come about through a prior agreement: "[W]here two or more persons agree to commit a specific crime ... [liability follows] because he or she had agreed to commit the crime ..." (my emphasis). Furthermore, in seeking to explain that the common purpose doctrine is unnecessarily invoked, Hiemstra's Criminal Procedure uses the example proffered in S v Ngobozi (1972 (3) SA 476 (A) 478D-E) to illustrate the operation of the doctrine, in terms of which each participant acting with a common purpose to assault can be held liable for murder on the basis of dolus eventualis:

"Suppose A and B, each carrying a knife, form an unlawful common purpose, in the execution whereof each is to play a contributory part, to assault C by stabbing him. In the ensuing scuffle, first A gets in the first and only stabbing-blow; and as a result C falls dead. Each is guilty of murder if he subjectively foresaw the possibility of the execution of their unlawful common purpose causing the death of C, but nevertheless persisted therein, reckless whether the possibility became fact." (Hiemstra's Criminal Procedure 22-29)

It is, ironically, clear that the example contemplates common purpose liability being based on a prior agreement.

The Mbuyisa court (supra par 72) clarifies and supports its position, distinguishing prior agreement from common purpose liability by further reference to Hiemstra's Criminal Procedure (22-29 and 22-30; full passage cited above under heading 2 2), which dismisses common purpose as "irrelevant" where an agreement to commit the offence has been proved. In this regard, reference is made to the cases of Mgedezi (supra) and Ngobozi (supra). While the court in Mgedezi does use the words "[i]n the absence of proof of a prior agreement" (705I) prior to setting out the authoritative requirements for active association common purpose (705I-706C), this merely serves to distinguish the application of this form of common purpose liability from the other form of common purpose liability, namely prior-agreement common purpose. There is no question of this phrase excluding prior agreement as a basis for common purpose liability; instead, it merely indicates that prior agreement was not relevant to the particular factual complex in Mgedezi. As for the reference to Ngobozi, the court accepted that at the time of the initial blow there was no (prior agreement) common purpose between the appellant and his companion (to assault the deceased) (478G), and further that no common purpose was formed between them thereafter, at the time of the fatal attack on the deceased (478G-479B). In this case, the appellant and his companion were not even held to be co-perpetrators in respect of the murder of the deceased, with the appellant being convicted of assault with intent to do grievous bodily harm, and his companion being convicted of murder. The analysis in Hiemstra's Criminal Procedure proceeds to state that the invocation of common purpose in the case of a hired assassin is "wrong in principle", and is calculated to confuse the judex facti (trier of facts), as in this case there is simply co-perpetrator liability. However, there is no reason in principle not to charge and convict a person hiring an assassin to kill another on the basis of (prior agreement) common purpose. This is precisely what occurred in the recent cases of S v Panayiotou (2017 JDR 1739 (ECP)) and S v Soni (2021 (2) SACR 241 (SCA)).

The Mbuyisa court (supra par 73) concludes its reliance on Hiemstra's Criminal Procedure (22-30) to point out the superfluity or irrelevance of the common purpose doctrine in circumstances where each participant in a criminal enterprise had fulfilled his agreed role by referring to the example of the bank robbery (see full passage cited above under heading 2 2). In the example, it is reasoned that in respect of a fatal shooting during the course of a bank robbery, all the members of the gang of robbers "are guilty of murder, not as a result of a forced application of the doctrine [of common purpose] but simply as co-perpetrators". After explaining further that each of the robbers would have direct intent to apply deadly force in order to rob, it is concluded that the correct result "would be reached by a simple application of the principles of the law of participation" on the facts of the case. Further support for this approach may be gleaned from the dictum cited (Mbuyisa supra par 65) from the Hlongwane judgment (supra par 41; see full passage above under heading 2 2).

What is one to make of the approach of the court in Mbuyisa and in Hlongwane, which coincides with the approach adopted in Hiemstra's Criminal Procedure? As discussed earlier, it is clear that the doctrine of common purpose can indeed be founded on prior agreement, as well as on active association. That this is the correct legal position has found wide and authoritative support in the courts, as well as being the academic consensus. Is it then not true that co-perpetrator liability can be employed to establish criminal liability where the participants in a criminal enterprise, formed by prior agreement, intentionally commit unlawful conduct? In this regard, Snyman's words are clear, and unassailably correct: "If a number of people commit a crime and they all comply with the requirements for perpetrators ... they are all simply co-perpetrators" (Hoctor Snyman's Criminal Law 222). Does the characterisation of the common purpose doctrine as unnecessary in the context of prior agreement (as strongly contended for by Hiemstra's Criminal Procedure, upon which the courts in Hlongwane and Mbuyisa rely) not then undermine the usefulness and apparent ubiquitous use of this doctrine in cases of more than one person acting together to commit a crime?

By no means: while the court in Mbuyisa correctly determined that the appellants were soundly convicted, on the basis of each of them fulfilling all the requirements for robbery, this does not detract from the fact that the appellants could have been convicted on the basis of the common purpose doctrine. The common purpose doctrine provides invaluable assistance to the State in cases where a group of two or more accused are involved in the commission of a crime, and where it is therefore often difficult to determine that the individual conduct of each of the group satisfied the requirement of causation. The operation of the doctrine provides that where two or more people share a common purpose to commit a crime, and act together in order to achieve such purpose, the conduct of each of them in the execution of that common purpose is imputed to each of the others in the common purpose (Hoctor Snyman's Criminal Law 225; Burchell Principles of Criminal Law 5ed (2016) 477). The difficulties of proof of a causal link between the individual accused's conduct and the harmful result are thus circumvented. These difficulties are cogently summarised in S v Mzwempi (2011 (2) SACR 237 (ECM)):

"In many cases involving a consequence crime and committed by a group of people - such as, for instance, murder - it is often very difficult, if not impossible, to determine which offender caused the death. If a victim is beaten to death by four offenders, all hitting him with knobkieries, it is often impossible to determine which of the offenders delivered the fatal blow -causing the death. In cases of this nature the element of causation is not proved beyond reasonable doubt, and all four offenders must be acquitted. This was the injustice and mischief sought to be overcome by the introduction of the common purpose doctrine." (par 45)

The Constitutional Court in S v Tshabalala (supra) adopted this reasoning, echoing the balance of the justification for the common purpose doctrine set out in Mzwempi (supra par 46), as follows:

"The object and purpose of the doctrine are to overcome an otherwise unjust result which offends the legal convictions of the community, by removing the element of causation from criminal liability and replacing it, in appropriate circumstances, with imputing the deed (actus reus) which caused the death (or other crime) to all the co-perpetrators." (Tshabalala supra par 56)

3 2 2 Common purpose or co-perpetrator

While the explanation of the rationale of the common purpose doctrine is very lucidly expressed, unfortunately the Constitutional Court in Tshabalala follows the Mzwempi judgment into a conceptual difficulty - a difficulty that was perpetuated in Hlongwane and the Mbuyazijudgment under discussion.

The difficulty in question simply relates to the fact that by definition co-perpetrator liability, being a form of perpetrator liability, requires that all the necessary elements of liability be established in order for the accused to be convicted of the crime in question (Burchell Principles of Criminal Law 475; S v Williams supra 63A-B). Where all such elements can be proved, there is no need to consider common purpose liability, as the following passage indicates (echoing the statement from Hiemstra's Criminal Procedure cited in par 72 of Mbuyisa):

"The Court a quo, applying the guidelines itemised in S v Mgedezi and Others ... convicted appellant No 4 on the basis of the doctrine of common purpose. But of course if appellant No 4 was the man in brown, as he must be found to have been, the doctrine of common purpose is irrelevant. If appellant No 4 was the man in brown he was a co-perpetrator who passed the gun to appellant No 5 when he was being held by the deceased to enable appellant No 5 to shoot the deceased. Appellant No 4's actions contributed causally to the death of the deceased." (S v Majosi 1991 (2) SACR (A) 540A-B)

The fact that common purpose is not required for the purposes of liability where co-perpetrator liability can be established does not however exclude it from being applied. The common purpose doctrine, by definition, relieves the State of the proof of the element of causation for each individual accused acting in a group to commit a crime. It follows then that a participant in a common purpose cannot be a co-perpetrator, as in order to be a co-perpetrator, the accused would be required to fulfil all the necessary elements of criminal liability. Unfortunately, both Mzwempi (supra par 46) and Tshabalala (supra par 56) refer to participants in a common purpose as "co-perpetrators"; so does the Hlongwane case (supra par 43), one of the authorities on which the Mbuyisa decision relies (par 64) as providing "a good illustration of the law [relating to participation] on the facts"; and so too do a handful of other High Court judgments, and at least one Supreme Court of Appeal judgment (S v Leshilo 2020 JDR 1882 (SCA) par 15).

Is this inaccurate use of terminology worthy of concern? It is submitted that there is indeed cause for concern, because while co-perpetrator liability closely resembles common purpose liability, there are important distinctions that need to be drawn between these concepts. Co-perpetrator liability arises where a number of persons have co-operated in the commission of a crime, and each, in so doing, has satisfied the definitional elements of liability for the crime. In contrast, the common purpose doctrine does not require proof of causation. As pointed out by Nienaber JA in S v Majosi (supra 540B), the liability of the co-perpetrator is direct in nature, whereas the liability of the participant in the common purpose is imputed (my emphasis; Mzwempi supra par 51 -53). (Is it any wonder that wherever possible in cases where more than one offender commits unlawful conduct, the State employs the common purpose doctrine to achieve a conviction?) Where it is not possible to prove that each participant personally performed every act required for liability for the offence, in appropriate circumstances the proved conduct of another member of the group may be imputed or attributed to the accused so as to render him liable in his own right - in other words, as a perpetrator. However, it bears emphasising that the common purpose doctrine forms a special category of perpetration, in terms of which a person may be deemed to be a perpetrator by reason of their mere involvement with a person or persons who actually perpetrate a crime. In Thebus (supra), Moseneke J explained the common purpose doctrine as "a set of rules of the common law that regulates the attribution of criminal liability to a person who undertakes jointly with another person or persons the commission of a crime" (par 18). Evidently, the "attribution" or deeming of common purpose liability is a different enterprise from the proof of co-perpetrator liability in terms of the accepted categories of perpetrator liability. Co-perpetrator liability and common purpose liability are recognised as qualitatively different and distinct concepts - both in the courts (see, e.g., S v Khoza supra 1038F-G; S v Kimberley 2005 (2) SACR 663 (SCA) par 12; S v Thebus supra par 40), and among the writers (Burchell Principles of Criminal Law 475ff; Hoctor Snyman's Criminal Law 219ff; Kemp, Walker, Palmer, Baqwa, Gevers, Leslie and Steynberg Criminal Law of South Africa 4ed (2022) 280ff; Visser and Maré Visser & Vorster's General Principles of Criminal Law Through the Cases 3ed (1990) 675ff), each of which subject common purpose liability to a separate discussion from perpetrator liability simpliciter.

The drawing of this distinction reflects the fundamental theoretical and practical dichotomy to be noted between co-perpetrator and common purpose liability. Strictly speaking, the rules relating to the co-perpetrator have nothing to do with those relating to complicity or participation. The liability of the co-perpetrator is assessed in terms of their own conduct and state of mind, as is that of the single perpetrator. What others do around the co-perpetrator does not affect the co-perpetrator's culpability. In sharp contrast, the liability of a participant in a common purpose is by definition dependent on someone else. In order for common purpose liability to ensue, the unlawful act, with a causal link to the harmful result, must be committed by someone in the group, who need not be the particular accused. Moreover, the common purpose doctrine requires someone to agree with another to commit the crime in question, or someone to actively associate with someone else's acts, demonstrating the intention to commit the crime in question. It is axiomatic that the application of the common purpose doctrine requires the presence of a number of persons acting together, and an interdependent and mutual reliance between such persons on each other in committing the crime. Responsibility for the conduct of each of the participants in the common purpose is ascribed to all others, and this mutual act-attribution serves to establish liability in terms of the doctrine. By contrast, the co-perpetrator's liability is not dependent on the liability of their co-perpetrator - as, for example, the case of Parry (supra) confirms.

Moreover, although there may be no difference in the result (conviction) achieved by establishing either co-perpetrator liability or common purpose liability, it may well be adjudicated by the court that there is a difference in the respective levels of blameworthiness, particularly where the common purpose liability is based on active association. Thus, in this respect, drawing the distinction in this context is of significant probative and practical importance.

It is therefore contended that to refer to a participant in a common purpose as a "co-perpetrator" conflates significantly distinct concepts of criminal liability; this practice in the sources listed, including the Mbuyisa judgment, is unfortunate and does not assist the vital process of establishing legal clarity, particularly in the area of participation, where there has been much uncertainty over the years. Corbett JA in Khoza (1031B-C) reflects on the law of participation: "]I]mprecise and undefined use of legal terms can lead to misunderstanding and confusion of thought, especially in [this] juristic field." Similarly, De Wet bemoans instances where the courts have not consistently upheld the correct principles, and where cases of co-perpetratorship are sometimes presented as if they are cases where one actor is actually a sort of participant in the other's crime ("gevalle van mededaderskap soms voorgestel asof hulle gevalle is waar die een dader eintlik 'n sort deelnemer is aan die ander se misdaad" - De Wet & Swanepoel Strafreg 4ed (1985) 192).

3 2 3 Vicarious liability

There is a further concern in the passage cited by the Mbuyisa court (supra par 72); in the terminology employed in relation to the hired-assassin example (Hiemstra's Criminal Procedure (22-29 and 22-30; full passage cited above under heading 2 2)), as between the co-perpetrators, the person hired is described as a "direct actor" and the person who hires him as a "vicarious actor". There is no difficulty with the term "direct actor". Snyman distinguishes between a direct and indirect perpetrator (which he describes as "merely convenient terms") by explaining that an indirect perpetrator is "somebody who commits a crime through the instrumentality of another", while the other party, who actually carries out the unlawful conduct is the direct perpetrator (Hoctor Snyman's Criminal Law 223). However, the use of the term "vicarious actor" is problematic. Mbuyisa (supra par 65) is not the first place in which this term comes up. The passage cited earlier from the Hlongwane case (supra par 41) explains that a person can commit an offence "directly or vicariously through another" (my emphasis). Later in the Hlongwane judgment (supra), the court states:

"It is evident on the facts that the appellant readily meets the requirements of a co-perpetrator to the crime of robbery since an agreement that they together would rob the two women can be inferred. Moreover he at all times continued to associate with his co-perpetrator when the knife was drawn and after, rendering him at the very least vicariously liable." (par 44, my emphasis)

The term is also employed (Mbuyisa supra par 73) in the example of the bank robbery from Hiemstra's Criminal Procedure (22-30; see full passage above under heading 2 2), where reference is made to co-perpetrator liability for murder for those in the criminal enterprise who "did not directly participate in the shootings but nevertheless participated fully in the crime" as being vicarious in nature.

The difficulty with the use of this term is simply that no general principle of vicarious liability is recognised in criminal law (Burchell Principles of Criminal Law 449). Vicarious liability is possible only in relation to statutory offences, and then only where the legislature specifically creates such liability in particular legislation (Hoctor Snyman's Criminal Law 212). Employing the term "vicarious" in the context of perpetrator liability for murder (as in the above-mentioned examples in Hiemstra's Criminal Procedure) or robbery (as in Hlongwane) is therefore inaccurate and confusing.

 

4 Concluding remarks

After these critical observations, it bears iteration that the result in Mbuyisa providing for co-perpetrator liability for the four appellants is sound. It is submitted that justice was served in this case. However, as indicated above, the reasoning that the court adopted from Hiemstra's Criminal Procedure regarding the concepts relating to criminal complicity is subject to challenge on two fronts.

As indicated, the argument that the common purpose doctrine is essentially limited to where the accused actively associates themselves with the conduct of another is not at all consistent with the current legal position in South African law, as the jurisprudence of both the courts and other writers agree. The difficulty with the approach adopted in Hiemstra's Criminal Procedure is demonstrated by the concluding part of the argument that appears in the work after the passage cited in the case (see above). Here it is contended as follows:

"On close examination it appears that the only real place for the doctrine of common purpose is in cases in which a specific agreement is not proved, and the causal link between the acts of a particular accused and the result of a consequence crime cannot be established. With respect, the example of Holmes JA in Ngobozi can, with a slight adaptation, make the point clearer. Say each of A and B (without an agreement to murder) inflicts one stab wound on C. At the post-mortem examination it is found that there was one fatal and one superficial stab wound to the body. Nobody knows which wound was inflicted by which assailant. In such a case the doctrine can be usefully applied. its application makes it irrelevant which attacker is causally connected to the bringing about of C's death." (Hiemstra's Criminal Procedure 22-30)

Unfortunately for the purposes of this example, the argument proceeds to the Tshabalala case, of which it is said that it was held that "[s]imilarly ... in cases of group rape . [t]he accused can be convicted on the basis of common purpose, and it is not necessary to prove that each accused raped the victim" (Hiemstra's Criminal Procedure 22-30). However, the form of common purpose that was applied in the Tshabalala case was prioragreement common purpose, as is evident from the following extract from the court's reasoning:

"It is trite that a prior agreement may not necessarily be express, but may be inferred from surrounding circumstances ... After a careful analysis of the facts, the High Court found that the applicants were part of the group that moved from one plot to another as per their arranged sequence. The High Court further found that the group members must have been aware or associated themselves with the criminal enterprise. They must have hatched a plan before then, that they would invade different households. Included in that plan or understanding was the rapes of the complainants." (Tshabalala supra par 49-50)

It may be further noted, in respect of the example based on the adjusted Ngobozi facts that, rather than common purpose being "usefully applied" to obtain a murder conviction, the posited absence of intent to murder would simply exclude the possibility of either A or B being convicted of the crime of murder. Neither is it correct to say that the application of common purpose "to certain sets of fact is often problematic (especially when large amorphous groups are involved)" (Hiemstra's Criminal Procedure 22-29). This is the very situation that the common purpose doctrine seeks to remedy, as the rationale for the doctrine (discussed above in Mzwempi and Tshabalala) indicates. But it bears emphasising that each participant in a common purpose must meet all the requirements for liability contained in the doctrine, whether the applicable form of common purpose liability is prior agreement or active association.

The second issue arising out of the argument (and sources relied on) in the Mbuyisa judgment is the conflation of co-perpetrator and common purpose liability. As explained above, these are distinct concepts. While both arise in the context of two or more actors being involved in the commission of a crime, co-perpetrator liability deals with direct-perpetrator liability, as opposed to common purpose liability, which deals with liability by means of imputation. While co-perpetrator liability (which as argued should never be expressed in terms of vicarious liability) is founded upon the actor meeting all the requirements for liability, common purpose liability is based on the imputation of the conduct of the participants in the common purpose, each to every other, along with the requisite elements of the particular form of common purpose liability in issue. It follows that to use the term "co-perpetrator" in the context of the application of the common purpose doctrine is misleading, and unhelpful.

Shannon Hoctor

Stellenbosch University