IN THE HIGH COURT OF SOUTH AFRICA
(EASTERN CAPE HIGH COURT : MTHATHA)
CASE NO. 284/04
Date heard: 17/09/2010
Date delivered: 28 /04/20/11
REPORTABLE
In the matter between:
ZINGELEZA MZWEMPI Appellant
and
THE STATE Respondent
JUDGMENT
ALKEMA J
[1] One of the archaic remnants of tribal life in the deep rural areas of South Africa is faction fighting. Reminiscent of the massacre of the McDonald clan by the Campbells in Glencoe, Scotland, during the night of 13 February 1692, the Manduzini clan attacked the Makhwaleni clan at Lusikisiki before sunrise early the morning of 3 October 2000.
[2] On the version of the McDonalds, the Campbell clan arrived at their village, Glencoe, earlier that evening. Masquerading as peaceful travelers and capitalizing on the misguided hospitality of the McDonalds, they occupied the latter’s homes for the night. At the given hour and whilst their hosts were asleep, the Campbells massacred almost the entire McDonald clan. (On the version of the Campbells, they were simply giving effect to the order from King William to “… fall upon the rebels, the McDonalds of Glencoe, and put all to the sword, under 70.) (Prologue: The Glencoe song))
[3] The Manduzini, sometimes referred to by the Makhwaleni as “Nombola’s” (meaning the “illiterate people”), used less stealth. They announced their attack before sunrise on 3 October 2000 by blowing on horns, firing shots in the air and shouting war cries. The Campbells used knives and swords; the Manduzini used rifles, shotguns and spears. The McDonalds were massacred in their sleep; the Makhwaleni fled their huts and ran into the bush, followed by their assailants.
[4] It is not known whether the Makhwaleni fled in one group, or whether they broke up in groups and scattered in different directions. The impression I get from a reading of the record is that they scattered in different directions, some being followed and others not.
[5] In the course of the next few hours, five of the Makhwaleni were killed, seven were severely wounded, and twenty eight of their huts were burnt down. Four accused from the Manduzini attacking force were duly charged with five counts of murder, seven counts of attempted murder and twenty eight counts of arson. The trial eventually started on 8 April 2006 before the High Court, sitting as a circuit court in Bizana, and presided over by Tshiki AJ (as he then was).
[6] Based on the doctrine of common purpose, all the accused were found guilty as charged. Each accused was sentenced to life imprisonment in respect of each count of murder; ten years imprisonment in respect of each count of attempted murder; and five years imprisonment in respect of each count of arson. The court ordered all the counts to run concurrently with the sentence of life imprisonment imposed in respect of the murder counts.
[7] The accused applied for and were granted leave to appeal to the Full Bench of this division. Only accused No. 4 prosecuted the appeal. The others abandoned their appeal. I will refer to accused No.4 as the appellant in this judgment. His appeal is against conviction only. This is the judgment on appeal.
[8] The facts of the case are relatively straight-forward and do not present any difficulties. As always, the difficulty lies with the application of the facts to the legal principles. The broad issue in this appeal concerns the proper approach to the application of the common purpose doctrine on the proven facts. The narrow issue is more daunting and worthy of early identification. It is entirely a legal issue. The background thereto is the following:
[9] The leading cases on the subject of common purpose remain the judgments of Botha JA in S v Safatsa and Others 1988 (1) SA 868 (AD) and S v Mgedezi and Others 1989 (1) SA 687 (AD). To avoid repetition and for the sake of convenience I will in the course of this judgment refer to the approach by the Appellate Division (as it was then known) in the aforesaid two judgments simply as the Safatsa/Mgedezi rule.
[10] As I will hopefully demonstrate in the course of this judgment, the carefully constructed jurisprudence in Safatsa/Mgedezi was disturbed in material respects by a subsequent judgment of the same Court slightly one year later in S v Nzo and Another 1990 (3) SA 1 (AD). Again, to avoid repetition, I will refer to this judgment simply as Nzo. For the reasons I will attempt to explain, Nzo extended the scope of liability under the Safatsa/Mgedezi rule beyond recognition of the very rules laid down in Safatsa (supra) and Mgedezi(supra).
[11] The rule in Safatsa/Mgedezi was constitutionally challenged before the Supreme Court of Appeal in S v Thebus and Another 2002 (2) SACR 566 (SCA). It was held to pass constitutional muster and one year later the judgment was confirmed by the Constitutional Court in S v Thebus and Another 2003 (2) SACR 319 (CC). Paradoxically, no reference whatsoever is made to Nzo in either the Supreme Court of Appeal decision of Thebus (SCA) or in the Constitutional Court judgment of Thebus (CC). (Hereinafter referred to as Thebus (SCA) and Thebus (CC) respectively). By contrast, it is clear from both judgments that what was considered in these cases was the constitutionality of the Safatsa/Mgedezi rule. In this context Thebus (SCA) refers at 578 para 28 to S v Mgedezi (supra) and Thebus (CC) refers throughout its judgment to both S v Safatsa (supra) and S v Mgedezi (supra).
[12] The constitutional approval of the rule in Safatsa/Mgedezi (supra) raises the question of how, if at all, the stare decisis rule remains applicable to Nzo. Whereas it is clear that the well defined scope of liability under the common purpose doctrine as expressed in the Safatsa/Mgedezi rule has now been upheld in both the Supreme Court of Appeal and in the Constitutional Court, has the extension of the rule as expressed in Nzo now, at least by implication, been disapproved? Or does it still apply? Are the lower Courts free to follow either the Safatsa/Mgedezi rule or the extended application in Nzo? Or are the lower Courts by virtue of stare decisis nevertheless obliged to follow Nzo where the facts fit the case? These questions invite further reflection and constitute the core issue in this appeal. But, first the facts.
[13] The State case against the appellant rests on the evidence of three witnesses. There is, of course, other evidence implicating the other accused, but such evidence is of little relevance to the appellant, save that it clearly establishes a pre-meditated and revenge attack against the Makhwaleni on the given day and time. The evidence implicating the appellant can be summarized as follows.
[14] The first State witness, Magilase, testified that his two huts were burnt down on the day of the attack. He said he had gone to the village the previous day, where he overheard a conversation between the appellant and his three companions. The appellant and one of his companions (identified as accused No. 3 during the trial) were known to him, the other two not. He heard appellant saying to his companions words to the effect that they (the appellant and his companions) “should not miss 4 o’clock …,” and that “by 5 o’clock we should be finished.”
[15] Magilase conceded that the words are meaningless, but he explained that he understood the appellant to say that they should attack the Makhwaleni clan at 4am the following morning “… to kill us …” and that they should be finished by 5 am. He did not say, and nor was he asked by the prosecutor, why he attributed such meaning to the words. He did not explain the context of the conversation. He did not elaborate any further or say that he had any prior knowledge or no knowledge at all concerning the pending attack, or on what basis he believed the words referred to the pending attack. On the argument of Mr Siyo, who appeared for the State, these issues are unimportant because, as a matter of common cause fact, the attack did occur the following morning at 4 am or shortly thereafter.
[16] The argument, with respect, misses the point. As I will point out later in this judgment when dealing with the legal principles, the issue is not so much whether or not the attack occurred – this is common cause - but whether or not the appellant was party to a prior agreement to carry out the attack and kill members of the Makhwaleni clan. The learned Judge a quo found that the appellant was indeed a party to such prior agreement, and it seems the appellant was convicted on this ground. The factual basis for such a finding is the conversation between the appellant and his companions the day prior to the attack referred to above. It is convenient to dispose of this finding now.
[17] Because Magilase, in consequence of the above conversation which he overheard, feared for his life, he decided not to go back to his home but rather to spend the night with his in-laws in another locality. He returned the following morning after sunrise. On his return the attack was over and he did not see anyone causing violence or committing a crime. His two huts were burnt down. Other huts were also burnt down. He assisted in the recovery of bodies. He did not see the appellant.
[18] The words uttered by the appellant to his companions and mentioned above are equivocal in context. They could mean, and/or may have been intended to mean, innumerable things. They may or may not even have referred to the pending attack. To even begin ascribing any meaningful purpose or intent to those words will amount to baseless speculation. In short, an agreement (to which the appellant was a party) to attack the Makhwaleni clan cannot be inferred from this conversation. Mr Siyo argued, however, that there may be another basis upon which such an agreement may be inferred.
[19] He submitted that the general incontrovertible tenor of the evidence is that the attack was expected and did not arise spontaneously. The warring factions were in continuous feud with one another, and the attack on the day in question was a revenge attack, anticipated by both factions. On this basis, Mr Siyo argued, it is an inescapable inference that the attack and its strategy was pre-planned and agreed.
[20] Whilst I have no problem with the logic of the above argument, the missing link is the absence of any evidence that the appellant was a party to such planning or agreement. As I will show later in this judgment, a prior agreement to commit a crime may invoke the imputation of conduct committed by one of the parties to the agreement which falls within their common design, to all the other contracting parties. Subject to proof of the other definitional elements of the crime such as unlawfulness and culpa, criminal liability may in these circumstances be established. The test, and requirements, however, of criminal liability under the common purpose doctrine in the absence of proof of a prior agreement, as I hope to demonstrate, is very different and more restrictive. It is therefore essential that before convicting under the common purpose doctrine on the strength of a prior agreement, the Court must be satisfied beyond reasonable doubt that such a prior agreement was proved, and that the accused was a party thereto.
[21] Our Courts are regularly faced with evidence of, say, a pre-planned robbery or burglary. It is trite that a prior agreement may not necessarily be express but may be inferred from surrounding circumstances. The facts constituting the surrounding circumstances from which the inferences are sought to be drawn must nevertheless be proved beyond reasonable doubt. And in this case there are no facts proved from which an implied prior agreement involving the appellant may be inferred. It is common cause that his shotgun was licensed and that he regularly used it for hunting purposes. The inference that he spontaneously joined the attacking force whilst on a hunting venture, or that he simply joined the attacking force without having been a party to any prior agreement to do so cannot be reasonably excluded.
[22] The appellant’s liability under the common purpose doctrine must therefore, in my view, be tested on grounds other than being a party to a prior agreement to commit the crimes. I am therefore of the respectful view that no evidential basis exists for a factual finding that appellant was a party to such a prior agreement. In making such a finding, I am of the respectful view that the Court a quo had misdirected itself.
[23] The next question is whether, on the evidence before the Court a quo, it was entitled to convict the appellant on the other facts found proven. It is to this question I now turn.
[24] The second State witness relied on by the prosecution is one Sijeni, who also knows the appellant. He testified that the pending attack was well known to the Makhwaleni. Because they expected the attack, most of them, if not all, did not sleep in their huts during the night of 2 and 3 October. He, Sijeni, slept outside (“… in the grass …”). He returned home shortly before sunrise to release his chickens. Whilst busy, he heard a commotion and saw people from neighbouring huts shouting and running away, chased by a group armed with firearms. Shots were fired. He joined those fleeing and ran.
[25] While fleeing they came across an old man by the name of Mzikinya who was armed with a bush knife and dressed in white overalls. They were still being pursued and continued running. The body of Mzikinya was subsequently found. He had been shot and his body placed in a hut which was burnt down.
[26] During the flight Sijeni had crossed the border of the Makhwaleni into another location. He was with Zandisile Gagadu. Whilst crossing a rivulet they came across some elderly women. Sijeni hid in the banks of the rivulet near the women.
[27] While hiding, he heard a voice saying “… you should not hide because that would not be clever.” He recognized the voice as that of the appellant who was known to him. The appellant had a peculiar hoarse voice. He looked and saw the appellant standing on the other side of the slope, approximately half a kilometer away. The appellant was carrying “… a big firearm.” He heard the appellant calling the other pursuers, whereafter appellant turned and walked away. He did not see him again.
[28] Sijeni testified that he did not see the appellant firing a shot. He said that on his return he found his hut burnt down, together with a number of other huts. It is impossible to say from the evidence how far and how long he fled before his encounter with the appellant. On his own evidence he had already crossed into another location. The probabilities are overwhelming that he only encountered the appellant long after sunrise, and also long after the other crimes of murder, attempted murder and acts of arson were committed. I will shortly revert to this issue.
[29] The third State witness, Gagadu, corroborated the evidence of Sijeni in broad general terms. His evidence, as recorded, is disjointed and difficult to follow.
[30] It appears that he spent the night somewhere else, and the morning of the attack he returned to his homestead. On arrival he found his home to be burnt down. He heard a whistle and a horn being blown, and shots were being fired. He ran away. He recognized the appellant as one of the pursuers. He, the appellant, was carrying “a big gun.” He fled and shots were fired at them. He does not know who fired the shots.
[31] He crossed a rivulet and came across Sijeni. He does not refer to the appellant when he crossed the rivulet with Sijeni or say if he saw the appellant at that stage. When he saw the appellant for the first time, he (the appellant) was approximately 100 to 200 meters away from him. It is unclear precisely where and when he saw the appellant for the first time. There is no evidence that he saw the appellant engaging in any activity. He simply places the appellant on the scene.
[32] The appellant testified in his own defence. In short, his defence is simply a denial of the evidence of the three State witnesses mentioned above. He stated that he was at work as a Security Guard near Lusikisiki on the day in question. He did not participate in the attack, and nor was he present when it occurred. He admitted to owning a licensed pump-action shotgun which he used for hunting.
[33] The appellant called two witnesses in his defence. The first did not contribute anything meaningful to the case, and the second confirmed that he (the appellant) was at work on 3 October 2000 where he, the witness, saw him.
[34] The learned trial Judge evaluated the evidence of both State witnesses and the defence witnesses in his judgment. It serves no purpose to repeat or discuss the trial court’s evaluation of the evidence. It suffices to say that the court a quo recognized the shortcomings in both the State case and the defence case. The learned trial Judge properly looked at the evidence as a whole and had proper regard to the contradictions, the probabilities, the demeanor of the witnesses, their credibility and powers of observation, and the extent and nature of corroboration of their evidence. It is trite that a court of appeal will not interfere with credibility and factual findings of the court below unless it has misdirected itself in a material respect either on the facts or on the law.
[35] In the present case the learned trial Judge accepted the evidence of the three State witnesses who placed the appellant on the scene, and rejected as false and unreliable the evidence of the appellant and his two witnesses who said he did not participate in the attack. Again, it is unnecessary to refer to and discuss the reasoning of the trial court in reaching the conclusions which it did. I am unable to fault the reasoning process, and in my respectful view, the court a quo did not misdirect itself in any material respect in arriving at these conclusions. The learned trial Judge therefore found that the appellant formed part of the attacking force, and this appeal must be decided on such factual basis.
[36] The question is whether those findings justify a conviction of the appellant on the five counts of murder, seven counts of attempted murder and twenty eight counts of arson. The facts upon which the appellant’s criminal liability must be judged may therefore be summarized as follows:
An attack by the Manduzini clan on the Makhwaleni clan occurred during the early hours of 3 October 2000.
In the course of the attack five members of the Makhwaleni were murdered, seven were severely wounded, and twenty eight of their huts or homesteads were burnt down.
The appellant was a member of the group of the Manduzini clan who attacked the Makhwaleni clan.
During the attack, the appellant was armed with his own, licensed shotgun.
The post-mortem reports show that the five deceased died from both gunshot and bullet wounds. The inference that the assailants were armed with rifles, shotguns and handguns is supported by the post-mortem reports and the general tenor of the evidence.
The post-mortem report relating to Mzikinya, the old man armed with the bush knife and dressed in white overalls who Sijeni came across during his flight, shows that he died only of multiple pellet wounds.
Ballistic tests conducted on cartridges found at the scene could not link appellant’s shotgun to any of the murders and there is no evidence whatsoever that he was the only attacker that was armed with a shotgun.
There is no direct or even circumstantial evidence specifically linking the appellant individually to any of the particular crimes except, of course, that he was an armed member of the attacking force and that he made common purpose with its general aims.
It is unclear from the evidence whether the Manduzini at all times acted and moved in one group, or whether they broke up in smaller parties and scattered throughout the battlefield in pursuance of the Makhwaleni. The overwhelming probabilities, as I remarked earlier, are that both groups broke up in smaller parties and the circumstances under which each murder and other crimes were committed differ in time, place and factual circumstances.
Notwithstanding Mr Siyo’s submissions to the contrary, there is no evidence which shows beyond a reasonable doubt that at the time the appellant was seen by Sijeni and Gagadu (outside the area of the Makhwaleni), all the crimes had not already been committed. On the contrary, the probabilities are that all the acts of murder, attempted murder and arson had already been completed when the appellant was observed as testified by Sijeni and Gagadu at the rivulet.
There is no credible evidence that the appellant was a party to a prior agreement to commit any of the crimes.
There is no evidence that the smaller attacking force of which the appellant was a member, and which pursued the State witnesses mentioned earlier, killed or attempted to kill any of the fugitives, or that they committed any acts of arson, or that the accused individually actively assisted or associated himself with any such other attacking force or forces (except, of course, that he was part of the general main group of the Manduzini).
13. It will be naïve not to hold that the appellant reasonably foresaw the possibility of death and or destruction of property in the course of the general attack, and I make such finding.
[37] I now turn to the legal principles applicable, and more particularly to the principles involving the common purpose doctrine. To place the doctrine in proper perspective and to read the rule in Safatsa/Mgedezi and the judgments in Nzo and Thebus (supra) in their proper context, it is unavoidable, I believe, to briefly refer to the history and nature of the common purpose doctrine.
[38] The doctrine was unknown to Roman law, Roman-Dutch law and the early common law of South Africa. It was introduced from English law to South African common law after the occupation by the English of the Cape and the other South African territories during the 19th century. One of the first leading cases on the subject is McKenzie v Van der Merwe 1917 (AD) 41. Although the case dealt with delictual liability, Innes CJ declared at 45:
“The delicts sued upon, if brought home to the respondent, would involve criminal as well as civil liability: Now a man may be guilty of a crime in which he took no physical part and rendered no assistance, if the perpetrator was in law his agent for the purpose … But then a mandatum sceleris must be established ... .”
[39] As I will indicate later in this judgment, the mandatum sceleris today, cast in the form of a prior agreement, plays a pivotal role in the modern common purpose doctrine in South African criminal law.
[40] The doctrine soon found its way into criminal law. See R v Garnsworthy and Others 1923 (WLD) 17 at 19; R v Duma and Another 1945 (AD) 410; R v Mkize 1946 (AD) 197 at 205-206; R v Shezi and Others 1948 (2) SA 119 (A).
[41] The application of the doctrine almost immediately evoked severe criticism from leading academics; most notably from Prof. J.C. de Wet in De Wet en Swanepoel, Strafreg (1st Ed.) (1949). Conflicting judgments from both Provincial Divisions and the Appellate Division (as it was then known) increased the uncertainty. See, for instance, R v Tsosane and Others 1951 (3) SA 405 (O); S v Ngobozi 1972 (3) SA 476 (A); S v Williams en ’n ander 1980 (1) SA 60 (A) at 65E-F; and S v Maxaba en Andere 1981 (1) SA 1148 (A) at 1155E-G.
[42] The debate continued and produced a mass of legal literature and divergent opinions and conflicting judgments. See, for instance, the later edition of J.C. de Wet (supra) (4th Ed.) (1985) pp.192 to 197. It serves no purpose to enter this debate and indulge in jurisprudential philosophizing. Suffice it to say that during 1988 and 1989 the (then) Appellate Division produced the two judgments in Safatsa (supra) and Mgedezi (supra) which settled most of the issues and brought some calm to troubled legal waters. Then came Nzo and the constitutional approval of the rule in Safatsa/Mgedezi in Thebus (SCA) and Thebus (CC) in the manner explained at the outset of this judgment. However, as I will shortly indicate, the debate has not yet been settled entirely. The correctness of Thebus (CC) is convincingly questioned by Jonathan Burchell, Principles of Criminal Law, 3rd Edition, 2008 pp. 580-588.
[43] Before dealing with the rule in Safatsa/Mgedezi and Nzo, it is important to understand the reason for the assimilation of the common purpose doctrine in South African law and the mischief it sought to remedy. The three judgments referred to above can only be understood and interpreted, in my respectful view, against such background, which is briefly the following: The traditional definitional elements of a crime under South African common law did not always, and in all circumstances, adequately satisfy the quest for crime control and a just criminal system in an increasingly lawless society. The problem lay with the requirement of causation.
[44]Absent the doctrine of common purpose, the South African common law of criminal liability recognizes four separate and distinct elements or requirements, namely; (i) an act (actus reus); (ii) which is unlawful (unlawfulness); (iii) causing the crime (causation); and (iv) committed with the necessary intent or culpa (mens rea). The doctrine of common purpose concerns only the element of causation.
[45] 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 overcome by the introduction of the common purpose doctrine. See, for instance, S v Madlala 1969 (2) AD 637 at 640F-649A.
[46] The object and purpose of the doctrine was therefore to overcome an otherwise unjust result which offended the legal convictions of the community. It did so 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.
[47] That the doctrine is aimed solely at removing the obstacle of proving causation and not any of the other requirement, is clear from the leading case of Safatsa (supra). Having referred to a number of authorities, Botha JA said at 898A-B:
“In my opinion these remarks constitute once again a clear recognition of the principle that in cases of common purpose the act of one participant in causing the death of the deceased is imputed, as a matter of law, to the other participants.”
[48] In regard to some judgments from the (then) Appellate Division which still advocated the retention of the element of causation in criminal law, (i.e. S v Williams and Another 1980 (1) SA 60 (A) per Joubert JA) Botha JA said in Safatsa (supra) at 898D:
“This remark [proving causation] has given rise to the question whether, in relation to cases of common purpose, some kind of causal connection is required to be proved between the conduct of a particular participant in the common purpose and the death of the deceased before a conviction of murder can be justified in respect of such a participant. In my view the clear answer is: No.”
[49] The Constitutional Court in Thebus (CC) held that the doctrine passed constitutional muster notwithstanding the abolition of causation as a requirement for criminal liability. The Court, per Moseneke J, said at 341, para [34] d-g:
“In our law, ordinarily, in a consequence crime, a causal nexus between the conduct of an accused and the criminal consequence is a prerequisite for criminal liability. The doctrine of common purpose dispenses with the causation requirement. Provided the accused actively associated with the conduct of the perpetrators in the group that caused the death and had the required intention in respect of the unlawful consequence, the accused would be guilty of the offence. The principal object of the doctrine of common purpose is to criminalise collective criminal conduct and thus to satisfy the social ‘need to control crime committed in the course of joint enterprises’. The phenomenon of serious crimes committed by collective individuals, acting in concert, remains a significant societal scourge. In consequence crimes such as murder, robbery, malicious damage to property and arson, it is often difficult to prove that the act of each person or of a particular person in the group contributed causally to the criminal result. Such a causal prerequisite for liability would render nugatory and ineffectual the object of the criminal norm of common purpose and make prosecution of collaborative criminal enterprise intractable and ineffectual.”
[50] To return to liability under the common purpose rule in circumstances where there is no prior agreement, I believe the starting point is the definition of common purpose. The best known and often quoted definition comes from Jonathan Burchell, Principles of Criminal Law (supra) at 574 which reads as follows:
“Where two or more people agree to commit a crime or actively associate in a joint unlawful enterprise, each will be responsible for the specific criminal conduct committed by one of their number which falls within their common design.”
[51] The definition embodies two elements or stages. The first stage refers to the conditions which must be fulfilled before the principle of imputation of conduct can operate; and the second stage refers to the scope and extent of imputing the conduct of one party to the others. The second stage, to repeat, only comes into operation when the conditions of the first stage are fulfilled.
[52] The conditions in the first stage which trigger the principle of imputation, are either a prior agreement or an active association in the joint venture. Any one of these conditions must exist. On the facts of this case, a prior agreement was not proved. The question is therefore whether or not the appellant actively associated himself with the aims of the attack on the Makhwaleni. The answer to this question depends on the meaning and content which our Courts, and in particular the Supreme Court of Appeal, have given to the concept of ‘active association’. Both Safatsa (supra) and Mgedezi (supra) deal with this issue. In neither of these two cases was any reliance placed on a prior agreement. It is, with respect, important to bear this distinction in mind.
[53] The second stage of the definition imputes conduct to an accused which falls within the common design or purpose (my emphasis). Conduct which falls within the common design seems to be any or all conduct in the execution of the common design or purpose. In the case of a prior agreement therefore, all the parties thereto will be held liable for the act of any one of their members which either falls within the common design or is executed in the course of the implementation of the agreement (provided, however, the other definitional requirements such as dolus are also present.
[54] It follows that where a prior agreement is proved, the accused is not required to be present at the scene where the crime is committed, and neither is he required to have actively participated in the conduct which caused the crime. Provided that the conduct imputed to him falls within the common design or the execution of the agreement, and that he had the necessary mens rea (either direct or dolus eventualis), he may be held liable under the common purpose rule.
[55] The above principle usually operates in cases, for instance, where a gang agrees to rob a bank or commit some other crime. A party to the agreement whose function it is to wait in a second getaway car some five or ten kilometers away, will be equally guilty of the offence of robbery under the common purpose doctrine notwithstanding that he or she may not have been at the scene where the robbery was committed, or that he or she may not have physically participated in the act of robbery. A classical example of this type of case is S v Majosi and Others 1991 (2) SACR 532 (A).
[56] The imputation of conduct – any conduct which falls within their common design – under the second stage must not be confused or conflated with the conduct which constitutes active association as a condition precedent under the first stage. As I will endeavour to demonstrate, the conduct in the former case seems to be any conduct which falls within the wide and general common design. The conduct in the latter case (active association) seems to be restricted to particular conduct and not to any conduct. That was the conduct considered in Safatsa (supra), Mgedezi (supra) and Nzo (supra), and to which I now, at long last, turn.
[57] In Safatsa (supra) (also known as the Sharpville 6 case) the salient facts were these:
[58] Eight accused were charged with murder and subversion under the (then) Internal Security Act. The charges followed the murder of one Dlamini, the Deputy Mayor of the town council of Lekoa, on 3 September 1984. A mob of people of about 100 had attacked his house, first by pelting it with stones and then by hurling petrol bombs through the broken windows, thus setting the house alight. Dlamini’s car was removed from the garage, parked in the street, turned on its side, and set on fire. As his house was burning, Dlamini fled from it to a neighbouring home. He was caught by some members of the mob and was assaulted. Stones were thrown at him and his head was battered with stones. He was dragged into the street, petrol was poured over him and he was set alight. He died.
[59] The eight accused were part of the attacking mob of approximately 100. Their participation may be summarized as follows:
(1) Accused 1 was one of the persons who caught the deceased when he fled his house. He wrestled with the deceased, and was the first who struck the deceased with a stone.
(2) Accused 2 was one of the mob who stoned the deceased’s house. When the deceased fled his burning house, the accused threw stones at him which struck his back, presumably causing him to fall and be caught by the other pursuers.
(3) Accused 3 was one of the small group of men who caught the deceased as he fled his house. He wrestled him to the ground and disarmed him.
(4) Accused 4 was part of the crowd, carrying a placard. She shouted repeatedly: He is shooting at us, let us kill him” (the deceased). When petrol was poured over the deceased a member of the crowd objected to him being set alight. The accused slapped this woman in the face, ostensibly to stop her from complaining.
(5) Accused No. 5 and 6 were part of the vanguard of the crowd, but they were not seen to throw stones. Save for being present and part of the leaders of the vanguard, there was no evidence against them of any active participation in any acts which contributed to the death of the deceased.
(6) Accused 7 was part of the stone-throwing mob. He made petrol bombs, poured petrol over the kitchen door of the deceased’s house and set it alight. He assisted in pushing the deceased’s car into the street.
(7) Accused 8 made petrol bombs which he handed to other members of the mob with instructions to surround the house and set it alight
[60] The court a quo found, which was confirmed on appeal, that with the exception of accused no. 5 and 6 all other 6 accused “…had actively associated themselves with the conduct of the mob, which was directed at the killing of the deceased.” ( at 893G) (my emphasis).
[61] It is clear, in my respectful view, that the general tenor of the judgment of Botha JA is that the test for imputing to the accused the actions of the group is the active association with the conduct which caused the deceased’s death. At 901 H-I the learned Judge of appeal states:
“… there can be no doubt, in my judgment, that the individual acts of each of the six accused convicted of murder manifested an active association with the acts of the mob which caused the death of the deceased.” (my emphasis).
[62] It was also due to the absence of any evidence that accused no. 5 and 6 had actively associated themselves with the acts causing the death, that they were acquitted.
[63] As is quite evident from the facts in Safatsa (supra), the participation of each accused in the death of the deceased must be separately analyzed. If it cannot be said that a particular accused actively participated or associated him or herself with the conduct which caused the death or other crime, then the actions of those who caused the death cannot be imputed to the particular accused and he must be acquitted.
[64] I now turn to consider the facts in Mgedezi (supra).
[65] In Mgedezi (supra) the trial court found that members of a large mob had attached and killed the deceased. The accused was part of the mob. The court a quo found that the accused foresaw the death of the deceased and associated themselves with such consequences. However, there was no evidence that any of the accused committed any act which was directly and physically linked to the causing of the death of any of the deceased. (at 698 F-G).
[66] The Court of Appeal (again per Botha JA) held that a view of the totality of the evidence cannot legitimately be used as a brush with which to tar each accused individually, nor as a means of rejecting the defence versions en masse (at 703B). The learned Judge of Appeal made the following remark at 703B-C:
“A view of the totality of the defence cases [of all the accused] cannot legitimately be used as a brush with which to tar each accused individually, nor as a means of rejecting the defence versions en masse. The global view taken by the trial Court of the defence cases led it to draw two inferences: (a) that each accused was present at the scene (at room 12) and participated in the execution of the threat against the mpimpi’s; and (b) that the defences of all of them were false beyond reasonable doubt. With respect, as a matter of simple logic I consider both inferences to be wholly insupportable.”
[67] At 703E-F he said:
“The trial Court erred by precluding itself from performing its duty to consider the evidence of each accused separately and individually, to weigh up that evidence against the particular evidence of the individual State witness or witnesses who implicated that accused, and upon that basis then to assess the question whether that accused’s evidence could reasonably possibly be true.”
[68] At 703H-704A he concluded:
“The reference, in purely general terms, to liability on the basis of common purpose, in para (3) of the above quotation from the judgment, cannot warrant an inference of liability in respect of all the accused en bloc. The trial Court was obliged to consider, in relation to each individual accused whose evidence could properly be rejected as false, the facts found proved by the State evidence against that accused, in order to assess whether there was a sufficient basis for holding that accused liable on the ground of active participation in the achievement of a common purpose. The trial Court’s failure to undertake this task again constituted a serious misdirection.”
[69] It is quite evident from the above that the conduct and activity of each individual accused in the participation of the crime must be considered. The general finding on the totality of the evidence that an accused was merely present and formed part of the crowd, without any evidence of his or her active participation in events which resulted in the death or other crimes with which he or she is charged, is insufficient to warrant a finding of “active participation.”
[70] The requirement that in the absence of a prior agreement the State must prove an active association with the events which are causally connected to the death, and that the accused must have been present at the scene where these events occurred, appear from the following extract of the judgment in Mgedezi (supra)at 705E-I:
“It would appear from the judgment of the trial Judge (para (2) of the quotation given earlier) that the trial Court might have based its decision on a finding that there had been a prior agreement between the accused to kill the mpimpi’s, i.e. the team leaders. There was, however, no evidence to substantiate such a finding. The attack which resulted in the killing or wounding of the team leaders was confined to room 12 and its occupants. Consequently any enquiry into common purpose must be directed at the events that occurred there. As far as accused No 6 is concerned, there is nothing in the evidence to show that he had agreed that the inmates of room 12 were to be assaulted. There is no suggestion of an express agreement and there is no proof of an implied agreement. As to the latter, the acts that accused No 6 was proved to have committed in the vicinity of blocks 4 and 5 do not give rise to an inference beyond reasonable doubt that he had agreed with any other person that the occupants of room 12 were to be killed. At the time when, and at the place where, accused No 6 participated in the activities of the group who were calling for mpimpi’s to be killed, those activities constituted no more than threats and intimidation, which had not reached any stage of actual execution, as we know from what happened in room 108, and it would be too much of a leap in time and place to infer from those events that accused No 6 had agreed to the events that occurred at room 12.”
[71] The restrictive meaning of “active association” is evidenced by the four requirements for liability under common purpose as formulated in Mgedezi (supra) at 705I-706C as follows:
“In the absence of proof of a prior agreement, accused No 6, who was not shown to have contributed causally to the killing or wounding of the occupants of room 12, can be held liable for those events, on the basis of the decision in S v Sefatsa and Others 1988 (1) SA 868 (A), only if certain prerequisites are satisfied. In the first place, he must have been present at the scene where the violence was being committed. Secondly, he must have been aware of the assault on the inmates of room 12. Thirdly, he must have intended to make common cause with those who were actually perpetrating the assault. Fourthly, he must have manifested his sharing of a common purpose with the perpetrators of the assault by himself performing some act of association with the conduct of the others. Fifthly, he must have had the requisite mens rea; so, in respect of the killing of the deceased, he must have intended them to be killed, or he must have foreseen the
possibility of their being killed and performed his own act of association with recklessness as to whether or not death was to ensue.”
[72] I pause to remark that the fifth requirement of mens rea is a definitional element of any crime which must in any event be proved, and is not a requirement of “active association.” Only the first four requirements referred to above are necessary to prove an “active association” in the crime under the doctrine of common purpose. The enquiry into dolus is a separate and further enquiry.
[73] In S v Thebus (supra) the Constitutional Court approved the approach in Mgedezi (supra). The issue is put beyond any doubt by Moseneke J at 341e (para 34) as follows:
“Provided the accused actively associated with the conduct of the perpetrators in the group that caused the death and had the required intention in respect of the unlawful consequence, the accused would be guilty of the offence” (my emphasis).
[74] The approach to adjudicate the actions of the accused individually and not to paint his conduct with a collective brush is described by the Constitutional Court in Thebus (supra) at 345, para [45] as follows:
“[45] A collective approach to determining the actual conduct or active association of an individual accused has many evidentiary pitfalls. The trial court must seek to determine, in respect of each accused person, the location, timing, sequence, duration, frequency and nature of the conduct alleged to constitute sufficient participation or active association and its relationship, if any, to the criminal result and to all other prerequisites of guilt. Whether or not active association has been appropriately established will depend upon the factual context of each case.”
[75] This state of the law in respect of criminal liability under the common purpose doctrine seems to have become settled by the rule and approach adopted in Safatsa/Mgedezi. The salient features of the rule may perhaps be summarized as follows:
[76] First, a distinction needs to be drawn between liability based on a prior agreement, and liability based on active association. On either basis, the conduct imputed to the accused is the conduct by the participants in the execution of their joint venture.
[77] Second, in the absence of a prior agreement, only the active association of the accused in the particular events which contributed to or caused the crime, triggers the principle of imputation in the manner described above. In this sense, liability arising from active association is much more restrictive. Such association will depend on the factual context of each case and must be decided with regard to the individual actions of each accused. In the assessment of the individual’s actions of each accused, the first four requirements for active association as set out in Mgedezi (supra) at 705I-706B and referred to above must be satisfied.
[78] Third, the other definitional elements of the crime, such as unlawfulness and culpa, must be present.
[79] The jurisprudential objections to liability under the common purpose doctrine was to a great extent met by the approach and rule in Safatsa/Mgedezi, in that the definitional element of causation was replaced with active association with the conduct which caused the death or other crime. The causal element thus remained between the conduct and the death. The actus reus constitute either the conclusion of the prior agreement, or the active association. Either of these events trigger the imputation principle. In this sense the invasion of common purpose liability into the common law requirement of causation is limited and serves the need for criminal expediency. This seems to have been the state of our law for one year.
[80] But then came the judgment in Nzo. The facts in Nzo were the following:
[81] During the relevant time (early eighties of the 20th century) the armed struggle waged by the then banned African National Congress (the ANC) was at its height. The ANC infiltrated the country from, inter alia, Lesotho. A small group (termed as “terrorists” in the law reports) had been sent to operate in Port Elizabeth. Their targets were public buildings, such as the Magistrate’s Court building, Administration Board buildings, shopping centres and railway tracks which were all damaged by explosions.
[82] A fuel depot was selected as the next target, but before it could be attacked certain incidents occurred which caused a dramatic turn of events. Mr. and Mrs. Tshiwula operated a so-called safe house which harboured part of the group. Mrs Tshiwula objected to the continued harbouring of the “terrorists” and threatened to reveal their activities. One of the group members, Joe, threatened to shoot her in order to protect the safety of the group. The threat was overheard by Nzo, the appellant.
[83] During the morning of 8 May 1983 the appellant, Nzo, passed through Aliwal North en route to Lesotho. He was accosted by members of the local security branch and found to be in possession of a false identity document. He was arrested, detained and questioned. It seems he made full disclosure and gave the security police all the information they required. During the evening of 8 May 1983, whilst Nzo was still in custody, Joe murdered Mrs Tshiwula. There was no evidence that Nzo was aware of the murder or that he knew it was going to be committed.
[84] Nzo was nevertheless charged, inter alia, with the murder of Mrs Tshiwula. He and his co-accused were both convicted by the trial court. It was found by the trial court that Nzo and the Second Appellant could not be convicted for the murder as co-perpetrators since they had no part in it (Nzo was under detention). They were, however, both convicted on the basis of common purpose. The conviction was based on the finding that Nzo and the Second Appellant foresaw the possibility of the murder and continued to associate in the common purpose of the group.
[85] The court a quo found a common purpose “… on the part of the terrorists… to commit acts of sabotage, in the execution of which design the possibility of certain categories of fatality must have been foreseen and, by inference, were foreseen by the participants to that common purpose” (Nzo (supra) at 4G-H).
[86] The court a quo held, as reported at 5G-H:
“To sum up, we find beyond reasonable doubt that accused No 1 must, on all the evidence and at all material times from 10 April onwards, have foreseen the killing of the deceased possibly occurring in the prosecution of the common purpose. In other words, he must have foreseen the possibility that it might become necessary for Joe to kill her in order to preserve the security and success of the mission on which they were engaged. With that foresight, and reckless as to whether such death occurred, he continued to associate in the common purpose right up to the time of his arrest eight hours before the murder….”
[87] The above reasoning and findings of the court a quo were confirmed on appeal to the Appellate Division. In argument, and undoubtedly following the reasoning process of the rule in Safatsa/Mgedezi, appellant’s counsel argued as follows: (I can do no better than to repeat the summary of the learned Judge of Appeal at 7D-G).
“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.”
[88] I must confess that on my reading and understanding of the rule in Safatsa/Mgedezi (supra), the above argument neatly and precisely encapsulates the rule.
[89] However, Hefer JA, writing also on behalf of Nestadt JA, was unable to agree. He regarded the argument as “… shrouded in a veil of irrelevant matter.” (at 7G). The learned Judge of Appeal found that the group, of which the appellants were members, functioned as a cohesive unit in which each performed his own allotted task. Their design was to wage a localized 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. In view of their continuing participation in the execution of the common design, despite their foresight of the possibility of murder, they are liable for every foreseen offence committed by any of them in the execution of the design (Nzo (supra) at 7 H-J).
[90] As authority for the basis of liability, the Court of Appeal relied on S v Madlala 1969 (2) SA 637 (A) at 640H to the effect that the parties to a common purpose are liable for every foreseen offence committed by any of them in the execution of the design if they persist, reckless as to its possible occurrence (at 7 C-D). Interestingly, the judgment in Nzo does not refer to or mention either Safatsa (supra) or Mgedezi (supra).
[91] The minority judgment in Nzo by Steyn JA embraced the appellant’s argument referred to above “as relevant and to the point.” (at 16G-H). Steyn JA referred to the rule in Safatsa/Mgedezi and pointed out that the extension of the rule as contained in the majority judgment was not considered in Safatsa (supra) or Mgedezi (supra) (at 15G-I).
[92] For the sake of completion I should mention that there were two appellants in Nzo. The first appellant, Nzo, was successful on appeal and was acquitted on the ground of disassociation only. It was found that by making a full disclosure of his involvement and that of his co-perpetrators in the joint venture after his arrest, he disassociated himself from any further involvement. The murder was committed, as I said earlier, only after his arrest and after his disassociation from the common design.
[93] The second appellant, however, was convicted on the basis of his foresight of the murder and his continued association with the common design. The point, however, is that but for his disassociation, Nzo would also have been convicted on such basis if he did not confess his involvement in the common design, even in circumstances where he was in detention when the murder was committed. I shall shortly return to the issue of disassociation.
[94] What is the effect of the judgment in Nzo on the present day status of the law of criminal liability under the common purpose rule? As I pointed out at the outset of this judgment, the rule in Safatsa/Mgedezi has received constitutional blessing in Thebus (CC). In approving the approach in Safatsa/Mgedezi, the Constitutional Court does not refer to Nzo, and neither does the Supreme Court of Appeal in Thebus (SCA). Is this Court nevertheless obliged by the stare decisis rule to apply Nzo? In considering these questions it must be recognized that this Court is not empowered, and nor is it sufficiently presumptuous, to express any views on the correctness or otherwise of Nzo. Although Mr Siyo did not refer to Nzo, it nevertheless remains binding authority on this Court. The issue to be decided is whether it remains binding authority in view of the constitutional approval of the Safatsa/Mgedezi rule in Thebus (CC).
[95] The starting point may be to consider the extent to which the three judgments in Safatsa, Mgedezi and Nzo respectively have been followed in South Africa before Safatsa and Mgedezi received constitutional consideration. The result is interesting.
[96] As far as I could ascertain, the judgment in Nzo has only been followed in South African case law on the issue of disassociation, and not on the issue of its extended scope of liability. The cases following Nzo on the issue of disassociation (correctly so), are inter alia, S v Singo 1993 (2) SA 765 at 771F; Molimi v S 2008 (5) BCLR 451 at 466C; S v Maelangwe 1999 (1) SACR 133 at 152C-D (NC); S v Musingadi and Others 2005 (1) SACR 395 at 407B-D (SCA); and S v Nduli 1993 (2) SACR 501 (A) at 504D-E.
[97] The second appellant in Nzo was convicted on the extended ground of foresight of death, notwithstanding the absence of any evidence that he was present or involved in the murder. There are no reported cases, as I said, which follow Nzo on the issue of the extended scope of liability. On the other hand, there is nearly an endless list of cases which follow Safatsa (supra) and Mgedezi (supra) on the issue of the restricted scope of liability. I mention only a few:
[98] The cases following Safatsa (supra) are:
S v Jama and Others 1989 (3) SA 427 (A) at 436D; S v Motaung and Others 1990 (4) SA 485 (A) at 486F-G; Magmoed v Janse Van Rensburg and Others 1993 (1) SA 777 (A) at 788E-F; S v Singo 1993 (2) SA 765 (A) at 772C-E.
[99] The cases following Mgedezi (supra) are:
S v Thebus 2003 (2) SACR 319; S v Singo 1993 (2) SA 765 (A) at 772C-E; S v Khumalo 1991 (4) SA 310 (A) at 351E-F; S v Mbanyaru & Another 2009 (1) SACR 631 (C) at 637H-J.
[100] It was submitted on behalf of the appellant in the present case that, in any event, he disassociated himself from the common design by turning his back on Sijeni after he found him and walked away. If this submission is correct, then Nzo should be followed on the issue of disassociation.
[101] The submission is tempting. However, in my respectful view, an honest appraisal of the facts do not show a disassociation on the part of the appellant from the common purpose and general design to attack the Makhwaleni.
[102] The appellant was part of the attacking force of the Manduzini. He was armed with a shotgun and actively participated in the pursuit of the Makhwaleni. Until the time he turned his back on Sijeni and walked away, he made common purpose with the aims of the attack. He foresaw the possibility that huts may be burned down and members of the Makhwaleni clan may be shot and killed. Notwithstanding, he recklessly and actively persisted with his association with the attacking force, right up to the time of his disassociation with the common design in the manner described above.
[103] The evidence discloses that he only disassociated himself from the attack when he saw Sijeni and Gagadu hiding in the banks of the rivulet near the women. He called out to them to the effect that, “…you should not hide because that would not be clever …” He thereupon called his other co-pursuers, turned and walked away.
[104] The disassociation evidenced by the above facts, however, on all probabilities took place after sunrise and after the acts of arson, murder and attempted murder by his fellow members of the attacking force had already been committed. I am therefore unable to find, on the facts, that his act of disassociation took place before the crimes were committed and that he therefore was no longer party to the common purpose of the grand design.
[105] I have no doubt that if Nzo remains binding authority in respect of its extended scope of liability, and is applied to the facts of this case, then the appellant was correctly convicted of murder and the other crimes. The question remains whether Nzo continues to be binding authority in this respect in view of Thebus (CC).
[106] The extension of the application of the common purpose rule in Nzo has, in my respectful view, far reaching and profound implications, both constitutionally and dogmatically. It is not merely an extension of the scope of liability under the Safatsa/Mgedezi rule; in many respects it is contrary to and even destructive of such rule.
[107] First, it does not substitute causation with an active association with the conduct which caused the death as required by Safatsa/Mgedezi. The element of causation between conduct and death is removed in its entirety. It is replaced by foresight, which is an element of culpa and not of causation.
[108] Second, the extended scope of liability under Nzo does not serve the purpose for the introduction of the common purpose doctrine into South African criminal law. It does not only remove the definitional element of causation, but it also removes a foundational requirement for criminal liability; namely the element of an actus reus. In this regard the foresight of harm and, notwithstanding, the continued association with the common design or purpose, is an element of culpa in the form of dolus eventualis, and must not be confused and/or conflated with the element of actus reus. In addition to culpa, the definitional requirement of an actus reus remains to be proved. Also, and for the reasons more fully discussed in the next paragraph, the association with the general design in the absence of a prior agreement is insufficient; to qualify as an actus reus it must be an active association with the particular conduct which caused the death or other consequence crime. Finally, the approach in Nzo elevates mere intent to commit murder (namely the foresight of death in the execution of the general design), to a conviction of murder. This is a concept not only foreign to South African criminal common law, but to most legal systems in the world.
[109] Third, the appellant’s contention in Nzo, namely that there must be an active association with a particular crime and that a collective brush may not be used to tar all the accused collectively, was rejected by the majority judgment in Nzo in the following words at 8G:
“This is plainly not so. 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, and I can conceive of no reason for drawing such a distinction.”
This finding seems to be in direct contrast with the approach in both Safatsa (supra) and Mgedezi (supra), and also to the judgment in Thebus (CC).
[110] Fourth, and except for the requirement of dolus (foresight), the facts in Nzo do not comply with any one of the other four requirements for liability under the common purpose rule as set out by Botha JA in Mgedezi (supra) and referred to earlier in this judgment. This approach is contrary to the approval of the rule in Safatsa/Mgedezi and therefore to the judgments in the Thebus (SCA) and Thebus (CC).
[111] Fifth, and with great respect, far from blurring the issues, the practical implication of the extended rule in Nzo when taken to its logical conclusion, is indeed that, (as held by Steyn JA in his minority judgment) by reason of an active association with the aims of the armed struggle and foresight that lives may be lost in the execution of such aims, every member of the ANC at the time would have been guilty of murder, a consequence not worthy of serious thought and, as far as I know, never applied in South Africa. If applied to the facts of this case, every single member of the attacking force of the Manduzini is guilty of all the crimes, whether they were present or even knew about the crimes or not.
[112] Sixth, and finally, the constitutional requirement that all accused persons must be treated equally, prohibits any rule which applies different criteria or requirements for criminal liability. Effectively, and depending on whether the Court applies the rule in Safatsa/Mgedezi; the basis of liability in Nzo; or the definitional elements of a crime under the traditional common law, the results will differ depending on the rule applied. The constitutional justification for crime control and an effective criminal justice system as applied in Safatsa/Mgedezi finds, in my respectful view, no application to the Nzo model, and does not carry the constitutional approval from the Constitutional Court.
[113] Of course, it must be remembered that Nzo was decided nearly four years prior to the interim constitution being promulgated on 27 April 1994. What may today be regarded as constitutionally objectionable was not considered as any legal impediment to the rule as formulated in Nzo at the time the judgment was delivered. Nevertheless, the judgment in Nzo has come in for severe criticism. See, for instance, Jonathan Burchell, S v Nzo 1990 (3) SA (A) – Common Purpose Liability SACJ (1990) (3) pp. 345-354; and by the same author, Joint enterprise and common purpose : perspectives in English and South African criminal law, SACJ (1997) 10 pp. 125-140 and 133-139.
[114] There is a final aspect of the judgment in Nzo which calls for comment. It relies, as mentioned earlier, on Madlala (supra) for extending the rule in Safatsa/Mgedezi: The passage in Madlala (from the unanimous judgment of Holmes JA) reads as follows at 640F-H:
“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-
that the individual killed the deceased, with the required dolus e.g. by shooting him; or
that he was a party to a common purpose to murder, and one or both of them did the deed; or
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 (A.D.) at 694 F-H and 695; or
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.”
[115] Hefer JA in Nzo relied, in particular, on (c) above as authority for the proposition that foresight, coupled with an active association in the grand design of the common purpose, constitute sufficient grounds for liability under the common purpose rule. This approach, with respect, is at this point in time and at this point of the development of the law on criminal liability under the common purpose rule, problematic. Madlala (supra) was decided some twenty years before Safatsa/Mgedezi. The rule in Safatsa/Mgedezi narrowed the scope of liability on the ground of active association considerably. And this was the approach approved by both Thebus (SCA) and Thebus (CC). It follows that the meaning and content of the concept “common purpose” in (c) above must today bear the meaning given to it by Safatsa/Mgedezi as approved by Thebus (CC). If this approach is correct, then Madlala (supra) is no longer authority for the extended scope of liability under the common purpose rule.
[116] In any event, what I think Holmes JA had in mind with (c), is a situation our trial courts are regularly faced with, and that is this. A gang of four robbers either agree or embark on a common purpose venture to rob a bank, but not to commit murder. Two of the robbers are armed. The duty of one of the unarmed robbers is to keep watch at the door of the bank. They do not have any direct intention to kill. During the robbery a shoot-out occurs in the course of which a bank employee (or robber) is shot and killed. All four robbers, including the two unarmed robbers and the one keeping watch, foresaw the possibility of death but notwithstanding such foresight they recklessly pursued their aims of robbery and associated themselves with the deeds of the others. In these circumstances the two unarmed robbers, including the one who stood watch at the door, are also guilty of murder on the common purpose rule. The result will be the same, even when applying the principles in Safatsa/Mgedezi and the requirements for liability under the common purpose rule in Mgedezi (supra).
[117] It is therefore my respectful view that Madlala (supra) can no longer be interpreted as authority for the approach in Nzo. For the above reasons, and following the judgments in Thebus (SCA) and Thebus (CC), I believe this Court must follow the rule in Safatsa/Mgedezi in preference to the extended application in Nzo.
[118] The final chapter on the subject is still to be written by either the Supreme Court of Appeal or the Constitutional Court (or both), but for the purposes of this judgment I do not believe that in the light of the judgments in Thebus (SCA) and Thebus (CC) the judgment in Nzo remains binding on this Court. The development of the common purpose doctrine in South African criminal law since the judgment in Nzo, and particularly the constitutional development as formulated by Thebus (CC), has, in my respectful view, overtaken the judgment in Nzo. Notwithstanding, it bears repetition that the Supreme Court of Appeal or Constitutional Court will undoubtedly have the last say on the subject.
[119] Applying that rule to the facts of this case, the following emerge.
[120] First, the appellant was not present at the scene where the violence which resulted in the criminal charges against him was being committed. The first requirement in Mgedezi (supra) is therefore not met.
[121] Second, there is no evidence that he was even aware of the crimes committed elsewhere by the other members of the attacking force. The fact that he foresaw that those crimes may be committed and nevertheless pursued the common purpose venture, is relevant only to the issue of culpa; it does not constitute a physical, active association with the conduct which caused the death (or other crimes). Mere intent to commit a crime in the absence of an actus reus (active association in the case of common purpose) is insufficient for criminal liability. The second requirement in Mgedezi (supra) is also not met.
[122] The third requirement for liability in terms of Mgedezi (supra) is that an accused person must have intended to make common cause with those who were actually perpetrating the crime. This means with those whose conduct caused the crimes. For the reasons more fully discussed in Safatsa/Mgedezi and mentioned earlier, the association must be with those particular individuals and not with the general, broad common design of the attacking mob. It accordingly follows from the absence of the first two requirements that also the third requirement in Mgedezi (supra) is not met.
[123] The fourth requirement in Mgedezi (supra) is that the appellant must have manifested his sharing of a common purpose with the perpetrators of the assault by himself performing some act of association with the conduct of the others. The “conduct of the others” is a reference to the conduct of those who actually perpetrated the crimes as contemplated by the requirement. Again, by virtue of the absence of the first two requirements, the fourth requirement is not met. Save for having been part of the attacking force, there is no evidence of any act committed by the appellant to manifest his association with the actual individual perpetrators of the crimes.
[124] I accept that the appellant had the necessary mens rea in the form of dolus eventualis, in that he foresaw death and arson and recklessly continued with his association of the attacking force in general. The fifth requirement is thus met, but in the absence of the first four requirements mere intent to commit a crime is insufficient for criminal liability.
[125] If I am correct that Nzo can no longer be interpreted as an extension of the principles in Safatsa/Mgedezi, then on my interpretation of Safatsa/Mgedezi the requirements for criminal liability under the common purpose rule are not met on the facts of this case. I therefore believe, with respect, that to the extent that the Court a quo either followed the judgment in Nzo, or failed to apply the principles in Safatsa/Mgedezi (supra) as set out in this judgment, it had misdirected itself.
[126] It follows that, in my respectful view, the appeal should succeed and the appellant should be found not guilty.
[127] Before I propose the order I intend to make in this appeal, I should mention that if this proposed order is to withstand scrutiny by a Higher Court, then those accused who did not prosecute their appeal, may languish innocently in custody. I therefore request the Director of Public Prosecutions to take this aspect of the matter up with the local Legal Aid Board with a view to prosecuting those appeals.
[128] I propose the following order:
The appeal of the appellant against his conviction succeeds, and his conviction and sentence is set aside and is substituted by the following order:
“The accused is found not guilty and he is discharged on all counts.”
I agree :
_____________________
EBRAHIM J
I agree
_____________________
MAQUBELA AJ
It is so ordered :
_____________________
ALKEMA J
Counsel for the Appellant : Adv. Qitsi
Instructed by : Legal Aid Office, Mthatha
Counsel for the Respondent : Adv. Siyo
Instructed by the : DPP office, Mthatha