S v Mzwempi

• Jurisdiction: South Africa
• Judge: Alkema J, Ebrahim J and Maqubela AJ
• Judgment Date: 28 April 2011
• Citation: 2011 (2) SACR 237 (ECM)
• Docket Number: 284/2004
• Hearing Date: 17 September 2010
• Counsel: KD Qitsi (instructed by the Legal Aid Board) for the appellant. WM Siyo (DPP, Mthatha) for the State.
• Court: Eastern Cape Division

Alkema J: C

[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 in the morning of 3 October 2000.

[2] On the version of the McDonalds, the Campbell clan arrived at their D village, Glencoe, earlier that evening. Masquerading as peaceful travellers and capitalising 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 E 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 F '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. G

[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. H

[5] In the course of the next few hours, five of the Makhwaleni were killed, seven were severely wounded, and 28 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 28 counts of arson. The trial eventually started on 8 April 2006 before the High Court, sitting as a circuit court in Bizana, and I 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; 10 years' imprisonment in respect of each count of attempted murder; and five years' imprisonment J

Alkema J

A 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 B 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 straightforward and do not present any difficulties. As always, the difficulty lies with the application of the C 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:

D [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 (A); and S v Mgedezi and Others 1989 (1) SA 687 (A). 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 E 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 (A). Again, to avoid F 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 G 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) (2003 (6) SA 505; 2003 (10) BCLR 1100). Paradoxically, no reference whatsoever is made to Nzo in H 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) I 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 J the common purpose doctrine as expressed in the Safatsa/Mgedezi rule

Alkema J

has now been upheld in both the Supreme Court of Appeal and in the A 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? B 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 C it clearly establishes a premeditated and revenge attack against the Makhwaleni on the given day and time. The evidence implicating the appellant can be summarised 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 D 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'. E

[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 4 am the following morning 'to kill us', and that they should be finished by 5 am. He did F 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 G 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 H 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 I 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 J

Alkema J

A 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 B 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 C 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...