CASES / VONNISSE

 

Distinguishing the forms of common purpose liability - S v Govender 2023 (2) SACR 137 (SCA)

 

 

1 Introduction

In the case of S v Govender (2023 (2) SACR 137 (SCA)), the Supreme Court of Appeal (SCA) was required once again to examine the common purpose doctrine, which although it has been in use in South African law for the past century, has in recent years seen significant development, in its expansion from being applied solely to a prior agreement, to also being applied in the case of an active association between two or more persons. The importance of distinguishing between these different forms of the common purpose doctrine has concomitantly also become increasingly important.

As the court points out in S v Mzwempi (2011 (2) SACR 237 (ECM) par 56), prior-agreement common purpose encompasses "any conduct which falls within the wide and general common design", whereas active-association common purpose is "restricted to particular conduct". Thus, in active-association common purpose, the association is with a "specific act" by which the crime was committed by another participant in the common purpose (S v Tilayi 2021 (2) SACR 350 (ECM) par 27, citing Snyman Criminal Law 6ed (2014) 260). It follows that, given the "marked differences" (S v Tilayi supra par 27) between the two forms of common purpose:

"in a case where the state seeks to place reliance on the doctrine of common purpose, the trier of fact will be required to determine the nature of the common purpose relied upon, what the scope of that common purpose happened to be, and whether the accused was a participant, and remained a participant, in the common purpose." (S v Tilayi supra par 28)

The discussion below examines the significance of the distinction between the different forms of common purpose doctrine, in light of the SCA judgment in the case of Govender.

 

2 Facts

The appellant was convicted on two counts of murder, along with contraventions of the Firearms Control Act (60 of 2000), in the Gauteng Division of the High Court, following events that took place at a club in Kyalami in Johannesburg. After leave to appeal to a full bench of the High Court was initially refused, the Supreme Court of Appeal granted the appellant special leave to appeal (par 1).

The context for the fateful events (set out in par 2-4) was that the appellant attended a function at the club in question, along with his wife and a group of friends, including one Latchman (first accused), who was the appellant's co-accused in the trial court. The appellant and his wife were waiting outside the venue when two of the appellant's friends emerged from the club, one with a bloody nose. The appellant took a firearm from his holster, and brandished it in the presence of witnesses, one of whom tried to restrain the appellant from going back into the club. These urgings fell on deaf ears.

At this point, the first accused appeared, whereupon the appellant and first accused walked up the stairs in the direction of the club. When halfway up the stairs, the first accused took the firearm from the appellant. The exchange of the firearm occurred without any force on the part of the first accused, and there was no evidence of anything militating against a consensual handing over of the weapon on the part of the appellant (par 5). Very shortly thereafter, shots were fired, as a result of which the two victims were shot dead. Both victims had earlier been involved in an argument with the appellant and his friends, which had turned violent (par 7-8).

The first accused's defence involved a simple denial that he was involved in any of the events that took place (par 9). The appellant chose not to give evidence in his defence, although through his counsel he put forward a different version of events to that of one of the witnesses, which indicated that he returned to the club to retrieve his gun, which had been taken from him (par 10).

 

3 Judgment

The SCA (per Siwendu AJA, Schippers and Carelse JJA, and Nhlangulela and Unterhalter AJJA concurring) dismissed the appeal against conviction (par 23, 26) and reaffirmed the correctness of the mandatory life sentence handed down by the court a quo (par 24-25). It found that the State had proved all the requirements for liability beyond reasonable doubt (par 14), and that the submissions by appellant's counsel seeking to introduce a different version of events were implausible and "unsustainable" (par 17). The failure of the appellant to testify himself was also a factor that counted against him (par 21 -22), and appellant's failure to report the loss of his firearm to the police further undermined his allegations of a scenario absolving him from liability (par 20). By contrast, the witnesses testifying for the prosecution were unshaken in their testimony, which clearly evidenced the appellant's guilt (par 17-19).

The basis upon which the court found the appellant to be liable for the two killings was that he had acted (with dolus eventualis (see par 15-16)) in common purpose with the first accused (see par 11-13). It is submitted that the court's path to the verdict, and the verdict itself, are beyond reproach, being logically ordered and clear. However, some of the reasoning employed in reaching such verdict is worthy of closer consideration, and such reasoning is consequently examined in the discussion that follows.

 

4 Discussion

4 1 The doctrine of common purpose in Govender

The appellant's challenge to the finding that he and the first accused had acted in common purpose was unsuccessful; he had argued that such a conclusion amounted to "conjecture and speculation" and was not consistent with the evidence (par 11). The court noted that although there was no evidence of any prior agreement (between the first accused and the appellant) to murder the deceased, common purpose can be established without any prior conspiracy, and may be "inferred from the conduct of the participants" (par 12). This form of common purpose liability, founded on active association, may be established if the authoritative requirements laid down in the Appellate Division case of S v Mgedezi (1989 (1) SA 687 (A) 705I-706C) are met (as the court points out (par 13)):

"The accused must have: (a) been present at the scene where the violence was committed; (b) been aware of the assault on the victim by someone else; (c) intended to make common purpose with the person perpetrating the assault; (d) manifested his sharing of a common purpose by himself performing an act of association with the conduct of the perpetrator; and (e) have the requisite mens rea."

The court proceeds to note that dolus eventualis suffices for liability - that is, where the accused has "foreseen the possibility that the acts of the perpetrator may result in the death of the victim, and reconciled himself with that eventuality" (par 13).

In finding that the appellant could be held liable for murder based on active-association common purpose, the court made the following statement (par 12):

"The concept of active association is wider than that of agreement, since it is seldom possible to prove a prior agreement. Consequently it is easier to draw an inference that a participant associated himself with the perpetrator."

4 2 The nature of common purpose liability

It is well established that common purpose liability comprises two categories: liability arising from "prior agreement, express or implied, to commit a common offence"; and, where no such prior agreement can be established, liability that "arises from an active association in a common criminal design with the requisite blameworthy state of mind" (S v Thebus 2003 (2) SACR 319 (CC) par 19). These categories operate distinctly, although Burchell argues that the two forms of common purpose may overlap in practice. Burchell contends that if there is a prior agreement among members of a group to commit crime A (which is then committed) and crime B is also foreseen by members of the group as a possible consequence of committing crime A, then if the conduct requirement of crime B is fulfilled, the group members can be liable for both crime A and crime B (Burchell Principles of Criminal Law 5ed (2016) 477), provided that there is "a specific agreement to commit crime A and ... some similarity in nature between crime A and crime B" (Burchell Principles of Criminal Law 478).

There are two basic difficulties with Burchell's formulation. (For further critique, see the views of Hoctor "A New Category of Common Purpose Liability" 2016 Obiter 666.) These difficulties are: (i) it does not demonstrate any "overlap" between the two "extremes" (as Burchell Criminal Law 477 terms them), as the commission of both crime A and crime B are incorporated in the prior-agreement form of common purpose liability, albeit that dolus eventualis applies to crime B; and (ii) the requirement that there be some similarity between crime A and crime B is not consistent with any legal precedent (not even in the case of S v Mzwempi (supra par 42), which supports Burchell's critique of the common purpose doctrine, and which Burchell in turn supports (Criminal Law 479)). In the recent case of S v Tilayi (supra), the court states that the two forms of common purpose, though distinct, are not necessarily mutually exclusive. The court proceeds to point out (par 24):

"[A] finding that the unlawful act falls outside an existing prior agreed to common design, does not mean that it cannot also be found to have been done in the furtherance of a common purpose that arose spontaneously, or by active association before, during or after the execution of the earlier agreed to common design. Furthermore, the execution of the agreed common purpose may also satisfy the requirements for active association."

It is clear that what the court has in mind is not a commingling of, or relationship between, the forms of common purpose (as Burchell suggests), but rather that even if an act was not contemplated in the prior agreement, it may still form part of a common purpose to act - albeit that the common purpose is based on spontaneous agreement or active association, rather than a plan to commit a crime. This approach, which is correct, is elaborated upon below, in the context of the discussion of the development of the different forms of common purpose.

4 3 The development of the different forms of common purpose liability

The notion of common purpose as a prior agreement (or "prior conspiracy") has been part of South African law from the first cases in which the doctrine was employed. So, for example, in R v Taylor (1920 EDL 318 327), the court held that there was "a common purpose or agreement to proceed to violence", and in R v Garnsworthy (1923 WLD 17 19), common purpose was characterised as when "two or more persons combine in an undertaking for an illegal purpose". Such a conception of common purpose is closely analogous to the inchoate offence of conspiracy (where two or more persons agree to (i) commit, or (ii) assist in, or (iii) procure the commission of a crime (Kemp, Walker, Palmer, Baqwa, Gevers, Leslie and Steynberg Criminal Law in South Africa 4ed (2022) 313), where liability is similarly founded upon the parties coming to an agreement.

By the middle of the twentieth century, it was established that a common purpose could arise spontaneously. In R v Mkize (1946 AD 197 206), Greenberg JA, writing on behalf of the court, and following the judgment of Tindall JA in R v Duma (1945 AD 410 415), stated that a "previous conspiracy between the persons concerned" was not required, but

"it is sufficient if they act in concert with the intention of doing an illegal act, even though this co-operation has commenced on an impulse without prior consultation or arrangement."

This approach was further confirmed by Murray AJA in R v Mtembu (1950 (1) SA 670 (A) 684), where it was stated that "even in the absence of a plan determined in advance", liability could follow in terms of common purpose if "upon the impulse of the moment" the appellant had joined in an unlawful attack (see also, inter alia, R v Tsosane 1951 (3) SA 405 (O) 408A; S v Maree 1964 (4) SA 545 (O) 553D-F).

It seems that the first use of the term "active association" in the context of the common purpose doctrine was also in the Mtembu case (supra), where Murray AJA describes a situation where the accused "sees the actual perpetrator about to take action which may have certain consequences and by active association with the perpetrator in such action he incurs liability for what the perpetrator thereafter causes" (685). The term was sparingly used in this context thereafter, appearing in R v Mgxwiti (1954 (1) SA 370 (A) 374, 375) and S v Mavhungu (1981 (1) SA 56 (A) 66C-D) before it was used by Botha AJA in S v Khoza (1982 (3) SA 1019 (A) 1053H), prior to the same judge employing this term in the seminal cases of S v Safatsa (1988 (1) SA 868 (A)) and S v Mgedezi (supra). Since then, the notion of active-association common purpose has been very frequently applied in practice, receiving its constitutional imprimatur in S v Thebus (supra); it has consequently appeared regularly in criminal case law.

However, even if active-association common purpose by definition arises on the spur of the moment (S v Safatsa supra 898A-B), this does not mean that prior-agreement common purpose cannot also originate spontaneously. As Van Zyl DJP has stated in Tilayi (supra par 21, footnotes omitted):

"When the common purpose is founded on an agreement, the agreement need not be express. It may be implied, in that it is inferred from all the circumstances. An agreement to commit an offence is generally a matter of inference deduced from certain acts of the parties accused, done in pursuance of a criminal purpose in common between them'. A common purpose may consequently be found to have arisen extemporaneously. Its existence is inferred from the fact that a number of persons act together in circumstances which are indicative of an intention to achieve a single common objective."

That an agreement may be established by implication, and inferred from the circumstances of the case, has been confirmed by both the Constitutional Court (in S v Tshabalala 2020 (2) SACR 38 (CC) par 49) and the SCA (see, e.g., S v Carter 2007 (2) SACR 415 (SCA) par 26; and S v Sibuyi 1993 (1) SACR 235 (A) 249h). That a common purpose based on a prior agreement may arise spontaneously may be illustrated by the example proffered by Botha AJA in S v Khoza (supra 1053D-E):

"[I]f, immediately before he commenced his assault on the deceased, accused No 2 had said to the appellant: 'Let us kill this man', and the appellant had replied: 'I agree', there can be no doubt that the appellant would have been guilty of murder, despite the fact that his own assault on the deceased in no way contributed to the deceased's death."

An example of such extemporaneous agreement establishing common purpose immediately preceding unlawful conduct may be found in S v Mambo (2006 (2) SACR 563 (SCA)), where it was held (par 17):

"The evidence against appellant 1, that he uttered the word 'skiet' as appellant 3 cocked the firearm ... constitutes sufficient proof that he shared a common purpose with appellant 3 - which might have been formed on the spur of the moment - to cause the death of the orderly. He, too, was therefore correctly convicted of murder."

(For an earlier example of a similar spontaneous common purpose, where the appellants collaborated in a deadly assault with the intent of avoiding arrest, see R v Du Randt 1954 (1) SA 313 (A).)

Given that both the prior-agreement and active-association forms of common purpose can arise in circumstances that can be classified as spontaneous, and given that "[t]he two forms apply to different sets of circumstances, have different conditions for their application, and must not be invoked when those circumstances and conditions are not present" (S v Tilayi supra par 28), it is crucially important properly to identify the form of common purpose that finds application in a particular case.

The functioning of the prior-agreement form of common purpose was authoritatively laid down in the Appellate Division case of S v Madlala (1969 (2) SA 637 (A) 640F-H):

"It is sometimes difficult to decide, when two accused are tried jointly on a charge of murder, whether the crime was committed by one or the other or both of them, or by neither. Generally, and leaving aside the position of an accessory after the fact, an accused may be convicted of murder if the killing was unlawful and there is proof-

(a) that he individually killed the deceased, with the required dolus, e.g., by shooting him; or

(b) that he was a party to a common purpose to murder, and one or both of them did the deed; or

(c) that he was a party to a common purpose to commit some other crime, and he foresaw the possibility of one or both of them causing death to someone in the execution of the plan, yet he persisted, reckless of such fatal consequence, and it occurred; see S v Malinga and Others, 1963 (1) SA 692 (AD) at p. 694F-H and p. 695; or

(d) that the accused must fall within (a) or (b) or (c) - it does not matter which, for in each event he would be guilty of murder."

This dictum has since been approved in numerous decisions in the SCA (formerly the Appellate Division) ( see e.g., S v Dhlamini 1971 (1) SA 807 (A) 817G-H; S v Maxaba 1981 (1) SA 1148 (A) 1155G-1156A; S v Daniëls 1983 (3) SA 275 (A) 323E-G; S v Safatsa supra 896F-897A; S v Nzo 1990 (3) SA 1 (A) 7B-D; S v Majosi 1991 (2) SACR 532 (A) 536I-537C; S v Sithebe 1992 (1) SACR 347 (A) 354G-H; S v Lebopa 1997 JDR 0297 (SCA) 11-12; S v Masango 1997 JDR 0379 (SCA) 6) as well as the High Court (see e.g., S v Tilayi supra par 25). Typically, these decisions have underscored the rule laid down in (c) of the Madlala dictum - that where X, a party to a common purpose to commit a certain crime, foresees the possible commission of another crime, then liability for such further crime may follow for X, even if there is no direct causal link between X's conduct and the coming about of the unlawful harm. Liability is thus founded on the common purpose, the prior agreement, which establishes the "conspiracy" between the parties. On this basis, the "conspirators" can be held liable not only for the crime that forms the basis of their collaboration, but also for any other crimes that are foreseen as possible consequences of their mutual endeavour. Liability can even extend to negligence-based crimes such as culpable homicide, where the members of the common purpose ought to have foreseen the possibility of death occurring as a result of their conduct (Burchell Principles of Criminal Law 498-499).

The prior agreement that establishes this form of common purpose liability approximates liability for the inchoate offence of conspiracy. Like the offence of conspiracy, there must be an agreement (a "meeting of the minds"), which may be either express or implied (Hoctor Snyman's Criminal Law 7ed (2020) 253). Unlike the conspiracy offence, the agreement itself in common purpose does not suffice for liability, and a party to the common purpose can withdraw from it after the agreement has been concluded, and in so doing avoid liability for the ensuing crime (Hoctor Snyman's Criminal Law 231).

The functioning of the prior-agreement form of common purpose is clearly demonstrated in S v Nzo (supra). The appellants were members of an undercover ANC group that had entered Port Elizabeth in order to engage in sabotage there. The deceased, Mrs Tshiwula, was killed by Joe (a member of the group), after she had threatened to expose what the group were doing to the authorities. While the appellants were not involved in the killing, their ongoing association with the group based on the common purpose (despite foresight of the possibility of the murder being committed) was held to be determinative of liability. The majority judgment, which approved and followed the dictum set out above in S v Madlala (supra 640F-H), was required to deal with the defence counsel's contention that common purpose liability was inapplicable (7D-G):

"Appellants' counsel argued, however, that this principle does not apply in a case like the present one. His argument went as follows: The ANC is an organisation with thousands of members in this country and several others. Some of its members are known to have committed a multitude of crimes in the execution and furtherance of its objectives. It is foreseeable that they might also do so in future. But, since liability cannot conceivably be imputed to every member for every foreseen crime so committed by all other members, the imputed liability of a member is limited to crimes with which he specifically associates himself. This is so because liability on the basis of the doctrine of common purpose arises from the accused's association with a particular crime and is not imputed to him where he associates himself, not with a particular crime, but with a criminal campaign involving the commission of a series of crimes. In such a case he can be convicted, apart from crimes in which he personally participated, only of those with which he specifically associated himself. And in the present case, although the appellants were actively involved in the campaign, there is no evidence that they associated themselves with Mrs Tshiwula's murder."

The majority of the court rejected this argument as "shrouded in a veil of irrelevant matter", in that neither "the general question of the liability of members of the ANC for crimes committed by other members" nor "the appellants' liability merely as members of the organisation" was in question; rather, the court was concerned with "the actions of three individuals" (Joe and the two appellants), who "formed the active core of the ANC cell in Port Elizabeth" and "functioned as a cohesive unit in which each performed his own allotted task" (7G-J). In fact, the majority held that the appellants' liability fell to be determined within a narrow ambit - that it was:

"[t]heir design ... to wage a localised campaign of terror and destruction; and it was in the furtherance of this design and for the preservation of the unit and the protection of each of its members that the murder was committed." (7I-J)

Hefer JA for the majority stresses that to argue that the appellants' participation in the execution of the common design is insufficient, and that evidence of their association with the murder as such is required to render them liable, entails a disavowal of the principles stated in S v Madlala (supra 640G-H (8E-F). Hefer JA continues, stating that to argue that the reference in par (c) to "some other crime" was intended as a reference to a particular crime and not a series of crimes is plainly not so, and in a case like the present one there is no logical distinction between a common design relating to a particular offence and one relating to a series of offences (8F-H).

MT Steyn JA wrote the minority judgment in Nzo, and held that in fact, contrary to the majority's view, the common purpose doctrine relates to specific crimes committed by a number of persons (15H-I). However, in making this statement, MT Steyn JA cites the Madlala case along with the cases of Safatsa (supra) and Mgedezi (supra 15G-H). Unlike the prioragreement form of common purpose, the active-association form, of which the latter two cases are textbook examples, does focus on specific criminality, and specific crimes (as MT Steyn JA points out 15G-I). MT Steyn JA (with whom the court in Mzwempi supra par 111 and Burchell Criminal Law 478-479 agree) unfortunately treats the circumstances in Nzo as an example of active-association common purpose, as opposed to prioragreement common purpose.

The broad approach envisaged by the prior-agreement form of common purpose may further be illustrated by the judgment in S v Mitchell (1992 (1) SACR 17 (A) 21). In this case, the appellants formed a common purpose to throw stones at persons they would pass on the road, from the back of the vehicle on which they were travelling, and duly loaded stones on the vehicle for this purpose (21D-G). The appellants therefore formed a common design to commit a series of assaults, in that whenever they passed anyone on the road, they would throw stones at them. The court further elaborates (21G-H):

"True to their design, when shortly after they left the café and a cyclist was encountered, second appellant threw one of the stones at him. Had the issue arisen, the other three would, on the basis of common purpose, also have been responsible for second appellant's actions. Similarly, had first appellant thrown one of the stones at deceased, second appellant and the other two would have been parties to his crime. In both cases this would be so even though they acted individually rather than in concert. The throwing of a stone by any one of them would have been imputed to each member of the group and not (as was argued) be regarded as the independent act of the individual perpetrator only."

It should be noted that while the prior-agreement common purpose provides for a broad ambit of liability, certain "stringent conditions", principally related to the proof of such agreement, nevertheless apply, as pointed out in S v Banda (1990 (3) SA 466 (BG) 501D-F); an accused "cannot be found guilty of sharing a common purpose with other accused by a process of osmosis".

However, in contrast, if reliance is placed on the active-association form of common purpose, there must be proof that the accused person associated himself, not with a wide and general common design, but with a specific criminal act that the other participant(s) committed (Hoctor Snyman's Criminal Law 229). This distinction in application between the different forms of common purpose is evident from the differing judgments in S v Nzo (supra). The majority, which held that there was a prior agreement between the parties to carry out criminal conduct, had no difficulty in finding that the appellants were correctly convicted on the basis of common purpose, given their foresight of the possibility that the victim could be killed, and their reconciliation with this possibility. The minority, which did not find liability could be attributed in this way while essentially applying active-association common purpose, held that the appellants could not be convicted of the murder of the victim.

A further illustration flows from the cases of S v Phetoe (2018 (1) SACR 593 (SCA)) and S v Tshabalala (supra). Without going into the broader discussion of whether common purpose could apply to rape (which was in any event definitively held to be so in the Tshabalala case) and how the differing perspectives of the courts affected the differing verdicts, these are the important considerations for the purposes of the current discussion. A group of young men, including Phetoe and Tshabalala, embarked on a rampage, including acts of housebreaking, rape, assault and robbery in the Tembisa township in Gauteng. Ultimately, the SCA held that Phetoe could not be found guilty in respect of the commission of the rapes. Phetoe was convicted in the court a quo as an accomplice to rape on the basis of common purpose. However, the SCA found that there was no proof of a prior agreement between the parties to commit rape, and given that he could not be identified at the dwellings where the rapes took place, he could not be convicted on this basis (S v Phetoe supra). By way of contrast, the Constitutional Court in S v Tshabalala (supra par 10) held that the High Court had correctly established that the attackers had acted in terms of a prior-agreement common purpose to rape (acting as a "cohesive whole", using the same adjective as did the majority judgment in Nzo), and as a result there was no difficulty in upholding the rape convictions.

One of the primary factors that differentiates the two forms of common purpose is that of presence at the scene of the crime. Whereas in the case of a prior agreement to commit a crime, it is not required for the purposes of liability that the actor be present when the harm occurs (see e.g., S v Yelani 1989 (2) SA 43 (A) 46D-H; S v Nzo supra; S v Khundulu 1991 (1) SACR 470 (A) 479E-F; S v Lungile 1999 (2) SACR 597 (SCA) par 14), for active-association common purpose, the courts have, following the authoritative list of requirements set out in S v Mgedezi (supra 705I-706C), consistently required that the actor be present if they are to be held liable for the crime in question (S v Motaung 1990 (4) SA 485 (A) 510I-J; S v Jacobs 2019 (1) SACR 623 (CC) par 106). It is, however, further required for active-association common purpose liability that the actor do more than merely be present at the scene of the crime: there must be an active association with the common purpose by means of some kind of overt or objectively ascertainable conduct (S v Khumalo 1991 (4) SA 310 (A) 357A-E; S v Mzwempi supra par 123). Once again, this would not be a requirement in respect of prior-agreement common purpose liability.

As pointed out in S v Banda (supra 501E-G), the active-association form of common purpose

"may not be used as a method or technique to subsume the guilt of all the accused without anything more. It cannot operate as a dragnet operation systematically to draw in all the accused. Association by way of participation, and the mens rea of each accused involved, are necessary and essential prerequisites."

Having examined the nature of common purpose liability, and its development through the South African case law, we can return to an assessment of some aspects of the application of the doctrine in Govender.

4 4 Difficulties with the application of the common purpose doctrine in Govender

There are a couple of issues that bear clarifying in the Govender judgment. First, in explaining the basis on which it found (correctly, it is submitted) that the appellant had the necessary intention in order to be found guilty of murder by way of the common purpose doctrine, the court indicates that (at a minimum) the appellant had dolus eventualis in respect of the use of the gun in the club by the first accused (par 15). The court indicates that this conclusion may be established on the basis of inferential reasoning, taking into account the appellant's reaction to the armed entrance into the club (citing S v Kramer 1972 (3) SA 331 (A) 334F), and consequently holding that the appellant "must have foreseen" that the first accused would use the firearm, which is indeed what occurred (par 16). In the course of this reasoning, the court states that "[t]his was not a case where the common purpose arose spontaneously or on the spur of the moment" (par 16). This is a curious remark. As discussed above, there would be no obstacle to a prior agreement developing spontaneously, where this is borne out by the evidence. However, the court specifically indicated that its finding was that there was no prior agreement between the appellant and the first accused, and that the finding was that the common purpose was in the form of active association (par 12), which, by definition, develops on the spur of the moment.

A further query relates to the statement (par 12) that the active-association form of common purpose is "wider" than that of prior agreement, as it is "seldom possible to prove a prior agreement", and it is thus "easier to draw an inference that a participant associated himself with the perpetrator". The source of this statement is Snyman (Criminal Law 5ed (2014) 267). Snyman continues his discussion to explain that agreement "whether express or implied, is merely one form of active association" (Snyman Criminal Law 5ed (2014) 267, citing Matzukis and Whiting). The statement thus refers to the broad classification of common purpose as active association, in all its forms. Averring that all forms of common purpose amount to active association (see Matzukis "The Nature and Scope of Common Purpose" 1988 SACJ 226 231-2, followed by Snyman in his earlier editions of Criminal Law, including 5ed (2008) 267n33, where Whiting "Joining in" 1986 SALJ 38 39-40 is also (incorrectly) cited in support of this view) can however only give rise to confusion and difficulty in keeping the distinction between the two forms of common purpose settled and clear, as is required by their different natures, requirements and modes of application (as discussed above). In any event, Snyman's statement, adopted in Govender, seems to be clearly incorrect in light of the development of the doctrine, as well as recent authoritative case law.

That it may be difficult to establish an agreement, and that it would be easier to draw an inference from the accused's conduct as to whether he intentionally associated with the perpetrator, are axiomatic from an evidentiary perspective, and these statements require no further comment. However, it is now well established that the active-association form of common purpose is in fact narrower than the prior-agreement form. In Dewnath v S ([2014] ZASCA 57 par 15), the SCA held:

"[T]he most critical requirement of active association is to curb too wide a liability. Current jurisprudence, premised on a proper application of S v Mgedezi & others, makes it clear that (i) there must be a close proximity in fact between the conduct considered to be active association and the result; and (ii) such active association must be significant and not a limited participation removed from the actual execution of the crime."

This statement was cited with approval by the SCA in S v Machi (2021 JDR 1741 (SCA) par 36) and the Constitutional Court in Makhubela v S; Matjeke v S ([2017] ZACC 36 par 38).

As Snyman indeed points out, the distinction between prior-agreement and active-association common purpose is that the former merely requires that the accused agreed with the "wide and general" common design of the conspirators, whereas the latter requires that the accused associate himself with "the specific act whereby the other participant(s) committed the crime" (Hoctor Snyman's Criminal Law 229). Taking into account the fact that the accused need not be present at the crime in respect of prior-agreement common purpose, and that where there is a prior-agreement common purpose, actual active association with the crime at the time of its commission is not required (Hoctor Snyman's Criminal Law 229; see further discussion above), it is evident that "liability arising from active association is more restrictive in nature than liability arising from a prior agreement" (Hoctor Snyman's Criminal Law 229; S v Mzwempi supra par 77).

The statement in Govender is therefore unfortunately misleading (as is an identical statement in the case of S v Ntshaba 2022 JDR 0279 (ECG) par 12; see the criticism of Hoctor "Recent Cases: General Principles of Criminal Law" 2022 SACJ 222 224-225). As the dictum in Dewnath authoritatively states, the role of active-association common purpose is to limit the liability of the accused in the context of group crime. Where there is prior-agreement common purpose, the basis for liability is much broader. Theron J in S v Jacobs (supra par 72) employs the terminology used in Kemp, Walker, Palmer, Baqwa, Gevers, Leslie and Steynberg Criminal Law in South Africa 2ed (2012) 235, now 4ed 284), referring to prior conspiracy (or prioragreement common purpose) as a bilateral or multilateral act of association and common purpose by conduct (spontaneous or active association) as a unilateral act of association. Drawing the distinction between the different forms of common purpose in this way is useful in underlining the disparity in their respective breadth of application. While the prior-agreement form of common purpose is, like the inchoate offence of conspiracy, founded upon mutual agreement between the participants to commit a crime or crimes, in respect of active-association common purpose, the accused is held individually liable based on his or her own action in joining in the commission of a crime that is already underway (but not yet completed; where the harm has already been caused at the point that the accused acted, this is the so-called "joining-in" situation, which by definition excludes common purpose liability (Hoctor Snyman's Criminal Law 232-233)). This approach, stated authoritatively in Mgedezi (supra), has been applied in the SCA cases of S v Buthelezi (1999 JDR 0587 (SCA) 45-46) and S v Botha (2006 (1) SACR 105 (SCA) par 14), where the basis for liability was, in each case, the appellant intentionally joining with or associating in the ongoing assault of the victim, despite the absence of proof of prior agreement. The Mgedezi judgment was also cited in S v Le Roux (2010 (2) SACR 11 (SCA) par 17), where the SCA held that unlike a "general and all-embracing approach", which as has been demonstrated above applies in the case of prior-agreement common purpose,

"the conduct of the individual accused should be individually considered, with a view to determining whether there is a sufficient basis for holding that a particular accused person is liable, on the ground of active participation in the achievement of a common purpose that developed at the scene."

 

5 Concluding remarks

Despite the ongoing development in the common purpose doctrine (see most recently its extended application to the crime of rape in S v Tshabalala supra), and its regular appearance in criminal cases, the crucial distinction between the forms of common purpose is not always drawn, to the detriment of legal clarity. Thus, when Kemp et al (Criminal Law in South Africa 4ed 285n35) state that the requirements for active-association common purpose laid down in S v Mgedezi "in fact apply equally to all cases involving common purpose" and that [i]t is just that they are more or less self-evident in cases where there is a prior conspiracy", this fails to take account of the important difference in the elements and ambit that the authors otherwise clearly delineate. The requirements for active-association common purpose are necessarily more narrowly constrained, since they deal with the particular factual scenario where an accused spontaneously joins in the commission of a crime. Since there is no question of the establishment of the existence of a prior agreement, and the conduct of the accused is required to be evaluated, often in the case of the criminal activity of a large group, there need to be stricter safeguards in place than would apply in the case of a prior agreement, where proof of the conspiracy (or even spontaneous agreement) is the essential requirement.

An example of the kind of difficulty that lack of clarity on the distinction between prior-agreement common purpose and active-association common purpose can bring may be found in the case of S v Bantom (2019 JDR 1784 (ECP)). It should immediately be noted that justice was done in this case, and that the court's treatment of the evidence is exemplary. However, it is noteworthy that the attack on the victims, with fatal consequences for one of the victims, was categorised as "not pre-planned, but was decided upon shortly before the attack started" (par 12). The mutual decision to attack is clearly a prior agreement, which is affirmed by the court's assessment of the state witness's testimony that the accused "spontaneously decided to carry out a robbery" (par 33). However, the court later contrarily asserts that the attack, was not pre-planned, nor subject to a prior agreement, but occurred when the accused "identified the opportunity to rob", upon which the accused "did not hesitate to act as a predatory group intent on achieving their objective" (par 92). The presence of an agreement is indicated by the court noting that resistance from the victim fatally wounded was "not only foreseen but anticipated", which was why the victim was subject to a rushed attack from behind (par 94). The court then, however, proceeds to apply the active-association common purpose doctrine (citing the cases of Safatsa supra, Thebus supra, and Mgedezi supra in this regard) and finds the accused guilty of murder on this basis (par 98-110). The point may simply be made that if the accused were indeed acting on the basis of a prior agreement, even if such prior agreement was formed very shortly before the attack ensued - as the court indicates occurred - it would not be necessary to prove all the elements for active-association common purpose to establish liability (cf the view of Kemp et al Criminal Law in South Africa above), and the process of proof would have unfolded very differently.

In sum, courts should strive for clarity in differentiating between the two forms of common purpose, and should not merely elide this crucial distinction by a default resort to bolstering their reasoning by employing the elements of active-association common purpose, no matter how tempting this may be. While both forms of common purpose play an important role in founding liability where two or more persons together engage in criminal conduct, prior agreement (aligned to "conspiracy" or "common design") is essentially and fundamentally different in its innate quality from spontaneous active association in criminal conduct that has already begun through the acts of another or others.

Shannon Hoctor

Stellenbosch University