A bi-annual update complementing the

Commentary on the Criminal Procedure Act


NO 1 OF 2014

Andrew Paizes, Author (Editor)

Steph van der Merwe, Author



In this edition we have, because of the importance of the cases and the interest they are likely to generate, four feature articles. The first considers the Constitutional Court’s assessment of the constitutional validity of s 2(2) of the Prevention of Organised Crime Act 121 of 1998, which allows the courts, in respect of certain offences set out in that Act, to hear evidence which might otherwise be inadmissible, provided that such evidence would not render the trial unfair. The second asks whether corroboration, when it is required, may be found in a person’s own statements (so-called ‘self-corroboration’). The third assesses a decision of the Supreme Court of Appeal in which it was held that an admission by a co-accused is not admissible against an accused. And the fourth returns to a topic considered in a feature article in 2013 (1) CJRdolus eventualis – and evaluates recent decisions in the light of what was held in cases considered in that article.

Other important issues raised in the cases include: (i) whether and in what circumstances criminal proceedings may be broadcast to the public through audio, audio-visual and photographic means (see the trial of Oscar Pistorius below); (ii) the constitutional validity of s 50(2)(a) of the Criminal Law (Sexual Offences and Related Matters) Amendment Act 32 of 2007, which provides that a court must make an order that the particulars of a person convicted of a sexual offence against a child or mentally disabled person must be included in the National Register for Sex Offenders (see S v IJ and J v NDPP below); (iii) whether a period spent by an accused awaiting trial can, on its own, constitute ‘substantial and compelling circumstances’ justifying a departure from the minimum sentence prescribed by the Criminal Law Amendment Act 107 of 1997 (see S v Radebe and DPP v Gcwala); and (iv) whether and in what circumstances an attorney’s feenotes may attract the legal professional privilege (see A Company & Two Others v Commissioner for SARS).

In addition, as indicated in the Editorial Note in 2013 (2) CJR, there was an appeal against the decision in Freedom Under Law v NDPP & others. The decision of the Supreme Court of Appeal is discussed and its implications considered. This is an important decision. It was, as Brand JA said, time to put to rest the issue whether prosecutorial decisions to discontinue or withdraw a prosecution were reviewable under the Promotion of Administrative Justice Act 3 of 2000. The role of the principle of legality as an alternative pathway to judicial review is also considered.

There was, too, legislation of some consequence. The Criminal Law (Forensic Procedures) Amendment Act 37 of 2013 was assented to on 23 January 2014, but will only come into operation on a date determined by the President in the Gazette. The scope and purpose of the Act in setting out specific procedures for DNA and related matters, as well as the relationship between it and the Criminal Law (Forensic Procedures) Amendment Act 6 of 2010, are discussed and analysed.

Andrew Paizes






(a) Criminal Law

(b) Criminal Procedure and Evidence: Pre-sentence

(c) Sentencing


The constitutional validity of s 2(2) of the Prevention of Organised Crime Act 121 of 1998

In Savoi & others v National Director of Public Prosecutions & another 2014 (1) SACR 545 (CC) the Constitutional Court was called upon to pronounce on the constitutional validity of a number of provisions of the Prevention of Organised Crime Act (POCA). In this article, I will consider only one of these, section 2(2); the rest will be dealt with elsewhere in this Review (see below).

Section 2(2) provides as follows:

The court may hear evidence, including evidence with regard to hearsay, similar facts or previous convictions, relating to offences contemplated in subsection (1), notwithstanding that such evidence might otherwise be inadmissible, provided that such evidence would not render a trial unfair.

The applicants in Savoi contended that s 2(2) infringed the right of an accused to a fair trial set out in s 35(3) of the Constitution. This contention, said the court, was based on an assumption that the admission of evidence that is subject to an exclusionary rule (such as that excluding hearsay or similar fact evidence) would, of necessity, render a trial unfair. If, then, it was true that the admission of such evidence would in all instances always lead to an unfair trial, the applicants’ case was on solid ground; if not, said the court, it would fail (at [36]).

Madlanga J (with whom the other justices agreed) then proceeded to examine each of the three specified classes of evidence, but only after making the point that s 2(2) was not limited to these but applied also to other exclusionary rules such as those relating to opinion, character evidence and confessions. What follows is, then, a careful examination as to whether, and, if so, in what circumstances, evidence excluded in terms of each of the three specified rules might, if it is nevertheless received, threaten the fairness of the trial.

One wonders, with respect, whether this lengthy exegesis was strictly necessary. There are, after all, only two possibilities. One is that some evidence would, even if it would be inadmissible under one or other of the exclusionary rules, nevertheless not render the trial unfair if received. If so, as Madlanga J correctly observed, the objection to s 2(2) must fail. The other is that no such cases arise, and that all evidence rendered inadmissible by each of the exclusionary rules would, if received, lead to an unfair trial. If this is the case, there would still be no danger of this actually occurring because of the proviso to s 2(2). This would mean that s 2(2) would be redundant and have no purchase or effect. But it could not be impugned for violating an accused’s rights under s 35.

Madlanga J’s analysis is, however, interesting and instructive. In respect of hearsay, first, he pointed to the anomaly that arises out of these propositions: first, that hearsay is excluded at common law because it is generally unreliable; second, that some hearsay that does not fall within the borders of a recognised exception to the rule is reliable; and, third, that the courts had closed the door to the creation of new exceptions to the rule, not ‘because the fountain whence the exceptions existing up to that point had been drawn had gone dry, with the result that no exceptions worthy of recognition could possibly ever come out of it again’ (at [41]), but because of the uncertainty that would arise from the piecemeal introduction of exceptions by the courts (see Myers v Director of Public Prosecutions [1965] AC 1001 at 1021–2). That being so, said Madlanga J, ‘there must still be categories of hearsay evidence not falling within the recognised exceptions, which are nevertheless reliable and thus deserving of admission’. The fact that Canadian courts refused to follow Myers and ‘opted to continue finding new exceptions to the hearsay rule on a principled basis . . . buttress[ed] the point that, outside of the recognised exceptions to the hearsay rule, there are other categories of hearsay, the admission of which will not necessarily lead to unfair trials’ (at [45]). If, in a given case, the reception of otherwise inadmissible hearsay would be objectionable in that it would undermine the fairness of the trial, the ‘filter’ in the form of the proviso to s 2(2) would, he added, be engaged.

As a result, the court concluded, any unconstitutionality that there might be would be a result of ‘a failure to use the filter in a constitutionally compliant manner’ (at [48]).

What is most curious about this analysis is that no reference is made in this context to the effect of s 3 of the Law of Evidence Amendment Act 45 of 1988. That section, in effect, replaces the common-law approach to hearsay with a far more flexible one that removes the tyranny of rule-and-exception and allows a court enormous discretionary power to receive hearsay evidence whenever it ‘is of the opinion that such evidence should be admitted in the interests of justice’ (in s 3(1)(c)). And it is very difficult to imagine how evidence excluded under this provision (which means that the court was of the view that its admission would not be in the interests of justice) could ever constitute evidence that would not undermine the fairness of an accused’s trial.

Even more curious is the fact that the court did refer to the provisions of s 3(1) in another context: it pointed to how s 2(2) of POCA, in doing away with the hearsay rule on charges under s 2(1), effectively did away with the hearsay rule (and other exclusionary rules), whereas s 3(1) of the Law of Evidence Amendment Act retained the exclusionary rule and allowed for the admission of hearsay only under the circumstances set out in paragraphs (a) to (c) of the section. Furthermore, after remarking that ‘it would be ill-advised to attempt to anticipate instances where the admission of hearsay would be so unfair as to infringe an accused’s fair trial right’ under the proviso to s 2(2), and that this would be ‘something best left to a trial court’ (at [49]), he pointed to some of the issues that ‘would have to be considered’. These included ‘the nature of the evidence, its reliability or lack of it, its probative value and prejudice to the accused’ – all factors that would, in an inquiry under s 3(1)(c), already have been considered in any finding by a court that it would not have been in the interests of justice to admit that item of hearsay evidence if, first, that section were to be invoked and, second the evidence were held to be inadmissible under its provisions.

Madlanga J then turned to similar fact evidence, the admission of which he observed to be ‘surrounded by some degree of confusion; but perhaps less so in recent times’ (at [50]). At the centre of this confusion was the decision in Makin v Attorney-General for New South Wales [1894] AC 57. It was held there that ‘if all that evidence of similar facts shows is proclivity of a particular kind, it is not admissible regardless of the probative value of that disposition’ (at [52] of Savoi). The problem, as Madlanga J correctly pointed out, is that ‘[i]t was not readily apparent why . . . propensity in and of itself might not, in a given set of circumstances, be sufficiently relevant to an issue before a trial court’. Unfortunately, and '[p]resumably because of its reference to design, accident and rebuttal of a defence, the Makin formulation came to be understood as having laid down rigid categories in which similar fact evidence would be relevant and admissible; the converse being the inadmissibility of similar facts not falling within those categories’ (at [52]) (see Commentary, s 210, sv Applications of the concept of relevance under the sub-heading Similar fact evidence where the Makin dictum is set out, analysed and criticised at length).

The decision of the House of Lords in DPP v Boardman [1974] 3 WLR 673 is, as Madlanga J said, ‘credited with having relaxed the stereotypical approach to the admission of similar fact evidence’ (at [53]), even though the exact ratio decidendi in that case was, as Madlanga J pointed out, ‘not without ambiguity’ and ‘difficult to discern’ (at notes 83 and 84, respectively). A major and, in the court’s view, welcome relaxation of the rule came, more clearly expressed, with DPP v P [1991] 3 All ER 337 (HL), [1991] 2 AC 447, where the House of Lords set out (at 460–1 (AC)) what Madlanga J called ‘a salutary proposition’ (at [54]):

Once the principle is recognised, that what has to be assessed is the probative force of the evidence in question, the infinite variety of circumstances in which the question arises demonstrates that there is no single manner in which this can be achieved. Whether the evidence has sufficient probative value to outweigh its prejudicial effect must in each case be a question of degree.

Significantly, too, the court recognised and endorsed the emphasis placed by the House of Lords on applying this principle in a flexible manner: in particular that ‘restricting the circumstances in which there is sufficient probative force to overcome prejudice of evidence relating to another crime to cases in which there is some striking similarity between them is to restrict the operation of the principle in a way which gives too much effect to a particular manner of stating it, and is not justified in principle.’ (Emphasis added.)

In identifying what the position was in South Africa, Madlanga J restricted himself to a consideration of two decisions of the Supreme Court of Appeal (or Appellate Division, as it was in the first case), S v D 1991 (2) SACR 543 (A) and S v Nduna 2011 (1) SACR 115 (SCA). In S v D the Appellate Division’s insistence on ‘striking similarity’ might, said Madlanga J, ‘lead to sophistry and technicality and raise more questions than provide answers’; the ‘real question should be whether, when looked at in its totality, evidence of similar facts “has sufficient probative value to outweigh its prejudicial effects” and that is a matter of degree in each case’ (at [55]; emphasis added).

In S v Nduna the court ‘followed the Makin formulation and the category based admission that seems to have come after Makin’ (at [56]) (see Commentary where Nduna is discussed and criticised). While making ‘no attempt at suggesting what the ideal development of the law on similar fact evidence should ultimately be’, Madlanga J expressed the view that the debate around these cases ‘adequately demonstrate[d] that in South Africa there is still ample room for a less restrictive approach to the admission of similar facts’ (at [58]), and that it emerged from that debate that not all similar fact evidence that is inadmissible according to South African Law would automatically render a trial unfair if admitted.

Subject to what was said above about whether it was necessary even to address these issues in the light of the proviso to s 2(2), this assessment of the flaws of our courts’ approach to similar fact evidence is to be welcomed. For too long the category and label approach fostered by the Makin formulation has flourished, forcing some courts to avoid undesirable results by distorting the formulation to allow highly probative evidence to be received. The error in such an approach is demonstrated by the decision in R v D 1958 (4) SA 364 (A) (not S v D to which Madlanga J referred), where the Appellate Division excluded highly probative evidence of previous acts performed by the accused on the ground that it showed merely that he had a propensity to perform such acts and did not use the (rather spurious) category of ‘actus reus’ so effectively employed in cases such as R v Ball and R v Katz (discussed in Commentary). Thus, by correctly applying the spirit of Makin, R v D stands as one of the few wrongly decided cases by that court in this area.

The Constitutional Court, by setting its face against Makin, the category and label approach, as well as an insistence on ‘striking similarity’ as a prerequisite to admissibility in certain cases, has cleared the way for an unequivocal acceptance of the ‘salutary’ propositions set out in DPP v P. The way should be open, too, to understanding that similar fact evidence is necessarily a species of circumstantial evidence, with the result that any assessment of the standard of proof required in respect of any item of similar fact evidence has to take account of a broader debate: whether that item is an indispensable part of the process of reasoning towards a conclusion that the accused is guilty (in which case it may be likened to a link in the chain of reasoning, which must be established beyond a reasonable doubt to allow, at all, for proof of guilt beyond a reasonable doubt), or whether it is merely one of a number of other items – each relevant but not necessary – indicating guilt (in which case it may be likened to a strand in a rope or cable, which merely adds strength to that rope if sufficiently probative, and which may do so even if not established beyond a reasonable doubt).

Flexibility of thought is thus essential – in assessing the role of the evidence in the broader scheme and the kind of inferential reasoning it generates and in assessing the extent to which its probative value as an item of circumstantial evidence in this context may or may not be sufficient to overcome any prejudice it might cause in the process.

And it is this flexibility which the Makin formulation resisted and which the judgment of Madlanga J endorses.

The court turned, finally, to evidence of previous convictions. Such evidence is, ordinarily, admissible only under the circumstances set out in s 211 of the Criminal Procedure Act – where, that is, it is ‘otherwise expressly provided by [the] Act or the Child Justice Act, 2008, or . . . where the fact of a previous conviction is an element of any offence with which an accused is charged’.

Since, however, ‘[e]vidence of previous convictions might be used where it would serve as relevant similar facts’, and, although the court did not say so expressly, since the principles governing the admissibility of relevant evidence are provided for in s 210 and, possibly, the residual provision set out in s 252, what was said by the court on similar fact evidence was applied, too, to evidence of previous convictions.

In conclusion (at [64]): It was ‘misconceived to suggest that under all circumstances the admission of otherwise inadmissible evidence would automatically result in a trial being unfair in violation of section 35(3) of the Constitution’.

Andrew Paizes


Sexual offences: Corroboration, self-corroboration and the probative value of the victim's report

In S v AM 2014 (1) SACR 48 (FB) the appellant’s conviction of the rape of his 14-year-old stepdaughter was confirmed by two judges. In a joint judgment, they pointed out that the victim’s evidence was corroborated by DNA tests linking the appellant to the semen found on the victim’s underwear (at [6]). This was an incriminating fact which the appellant, having advanced the defence that no intercourse with the complainant had taken place, could neither explain (at [7]) nor refute (at [9]). The physical and emotional condition of the victim some six hours after the incident and as described by the aunt to whom the victim had reported the rape, was also treated as corroboration of the victim’s evidence that her stepfather had raped her (at [6]).

Indeed, there is ample Supreme Court of Appeal authority to the effect that evidence of the victim’s distressed condition can, where appropriate, serve as corroboration. See S v S 1990 (1) SACR 5 (A) at 11a-c where the victim’s condition of severe shock was accepted as strong corroboration (‘sterk stawing’) that she was raped. See also S v Hammond 2004 (2) SACR 303 (SCA) at [21] where Cloete JA referred with approval to the following statement by Lord Parker CJ in R v Redpath (1962) 46 CAR 319 at 321: ‘[T]he distressed condition of a complainant is quite clearly capable of amounting to corroboration’. Evidence of this nature can assist ‘to show that sexual contact took place, where this is denied’ (S v Hammond (supra) at [22]).

Ashworth ‘Corroboration and Self-corroboration’ 1978 Justice of the Peace 266 explains that in terms of the English common law there is a rule that a witness cannot corroborate himself except for ‘one carefully circumscribed set of circumstances where self-corroboration is possible – by means of the victim’s distressed condition after the alleged incident’ (at 267). Of course, the court must be satisfied that the emotional condition was not simulated and, if genuine, that it was indeed the result of the fact that the witness was the victim (S v Hammond (supra) at [23]; S v Balhuber 1987 1 PH H22 (A); R v Redpath (supra) 322; R v Chauhan (1981) 73 Cr App R 232). The same reasoning applies to bodily injuries.

However, in S v AM (supra) the court of appeal – in assessing the evidence and dismissing the appeal against conviction – relied not only on the corroboration provided by the DNA evidence and the evidence of the bodily and distressed condition of the victim. The court of appeal went further and concluded that the report to her aunt also corroborated the victim’s evidence (at [6]). Furthermore, at [9] of the joint judgment, it was concluded that the trial court, in addressing the discrepancies in the testimony of the young victim, had correctly found that ‘such discrepancies were not material, taking into account that the evidence was corroborated by the first report she made to her aunt and, critically, the DNA results’.

It is submitted, with great respect, that in S v AM (supra) the court of appeal erred in treating the complaint as corroboration of the victim’s evidence. The complaint can, at most, only serve to prove consistency. See R v M 1959 (1) SA 352 (A) at 355H; S v Mohlakane 2003 (2) SACR 569 (O) at [6].

In S v Gentle 2005 (1) SACR 420 (SCA) Cloete JA (Farlam and Ponnan JJA concurring) refused to treat as corroboration the alleged victim’s complaint made to her sister-in-law, who had arrived on the scene when the victim was without her panties and in what appeared to be an injured state. At [19] it was said that the complaint ‘is not corroboration’ and only proves consistency: ‘It is relevant solely to her credibility. The complaint cannot be used as creating a probability in favour of the State case . . .’ See also S v Hammond (supra) at [17]. The rule that a victim’s complaint cannot corroborate the victim can be traced to an English case decided exactly a century ago. In R v Christie [1914] AC 545 the trial judge had directed the jury that the evidence of a young boy, the alleged victim of an indecent assault, was corroborated by the complaint made by him to his mother and a policeman shortly after the incident. This direction, held the House of Lords, was wrong. This approach must be supported. In S v Hanekom 2011 (1) SACR 430 (WCC) at [27] Saner AJ said that there is a ‘rule against self-corroboration by self-consistent statements’. Corroboration must come from an independent source (S v Gentle(supra) at [18]); and the victim’s report, being nothing other than a prior consistent statement of the victim, lacks the required independence to qualify as corroboration. See generally S v Scott-Crossley 2008 (1) SACR 223 SCA at [17]; S v Mkohle 1990 (1) SACR 95 (A) at 99d. After all, a lie can be repeated as often as the truth (R v Rose 1937 AD 467); and if a witness could corroborate himself, said the court in R v Whitehead 1929 1 KB 99 at 102, it would only be necessary for him to repeat his report some twenty-five times in order to secure twenty-five corroborations of it.

In Director of Public Prosecutions v Kilbourne [1973] 1 All ER 440 the court, having noted that ‘[t]here is nothing technical in the idea of corroboration’, stated as follows (at 456, emphasis added):

And the law says that a witness cannot corroborate himself. In ordinary affairs we are often influenced by the fact that the maker of the doubted statement has consistently said the same thing ever since the event described happened. But the justification for the legal view must, I think, be that generally it would be too dangerous to take this into account and therefore it is best to have a universal rule.

It is therefore respectfully submitted that the treatment of a victim’s report as corroboration (as was done in S v AM (supra)), is totally incompatible with our case law and English common-law principles as set out above. Finally, it should be noted that the fact that the admissibility of the victim’s report has, since 16 December 2007, been regulated by the provisions of s 58 (as read with s 59) of the Criminal Law (Sexual Offences and Related Matters) Amendment Act 32 of 2007, has had no impact on the principles governing corroboration and self-corroboration. See also Commentary on the Criminal Procedure Act, s 190, sv Complaints in sexual cases, under the main heading Previous consistent statements. See further Schwikkard in Smythe & Pithey (eds) Sexual Offences Commentary (2011) at 23-5.

The dismissal of the appeal in S v AM (supra) was, it would seem, justified on the facts referred to in the appeal judgment. However, treating the victim’s report as corroboration was impermissible.

Steph van der Merwe

Is an admission by a co-accused admissible against an accused?

It is clearly stated in s 219 that ‘[n]o confession made by any person shall be admissible in evidence against another person’ (emphasis added). So a confession by a co-accused is not admissible against an accused. But what of an admission? This question is considered at great length in Commentary in the notes to s 219. The position has, for some time and certainly since the decision in S v Ndhlovu & others 2002 (2) SACR 325 (SCA), been the subject of much debate and controversy. It was held in Ndhlovu that an admission, which constitutes a class of hearsay evidence, may be used against a co-accused if it satisfies the requirements of s 3(1)(c) of the Law of Evidence Amendment Act 45 of 1988 (see, too, the cases discussed in Commentary, especially S v Ralukukwe 2006 (2) SACR 394 (SCA) and S v Molimi & another 2006 (2) SACR 8 (SCA)).

But the decision in Ndhlovu did not meet with universal approval. The Constitutional Court in S v Molimi 2008 (2) SACR 76 (CC) was asked to invoke the right to equality in such a way as to require admissions to be treated in the same way as confessions. It declined to decide the issue but conceded that ‘the argument may be sound’. But deep unhappiness with the Ndhlovu approach was voiced in decisions of the Supreme Court of Appeal in Balkwell & another v S [2007] 3 All SA 465 (SCA), S v Libazi & another 2010 (2) SACR 233 (SCA) and S v Shilakwe 2012 (1) SACR 16 (SCA) (all discussed in Commentary).

This rising tide of dissatisfaction has culminated, finally, in an emphatic and unequivocal rejection of the principle enunciated in Ndhlovu by the Supreme Court of Appeal in S v Litako & others [2014] ZASCA 54 (unreported, SCA case no 584/2013, 16 April 2014). Navsa and Ponnan JJA (with whom Leach and Petse JJA and Swain AJA agreed) examined the English common-law position, which they found, ‘usefully summarised’ in Halsbury’s Laws of England 4 ed (1990) vol 11 (2) para 1131, to be that ‘[w]here several persons are accused of an offence, and one of them makes a confession or an admission, that confession or admission is evidence only against the party making it’; and that ‘statements which are not made in pursuance of the common design are evidence only against the makers’.

The South African law, said the court, had developed along the same lines. Until Ndhlovu it had not been suggested that s 3 was applicable to an admission of a co-accused, and cases such as S v Matsitwane 1942 AD 213, R v Baartman & others 1960 (3) SA 535 (A) and R v Nkosi & another 1950 (1) PH 91 (A) followed the approach of the English common law until the ‘seismic shift’ in Ndhlovu. And in that case, said the court at [50], ‘no attention at all was paid to the common law rule’.

The court turned to what it considered the weaknesses of the reasoning in Ndhlovu. The court there treated the problem as one relating only to hearsay, and considered that the provisions of s 3(1)(c) of the Law of Evidence Amendment Act, which allowed for the reception of hearsay whenever it was in the interests of justice to admit it, contained sufficient safeguards to ensure that there was no infraction of constitutional principles (see [44]). The problem, said Navsa and Ponnan JJA, was that the common-law rule was ‘not only an aversion to the admission of hearsay evidence, but it developed because of the inherent dangers of permitting the use of extra-curial statements by one accused against another’ (at [51]). It ‘recognised the potential conflicts between the interests of co-accused persons’ and, because a co-accused could not be forced to testify, it appreciated ‘that fair trial rights, including the right to fully challenge the State’s case, may be hampered’.

The common-law rule was thus not based solely on the hearsay objection although, said the court, ‘that in itself would have constituted sound reason for excluding such evidence’ (at [65]). It had, too, always been said that such evidence was ‘normally irrelevant’ and that it was evidence which, ‘in itself, ought to bring with it a caution’. Moreover, since a co-accused often disavowed his or her statement and often chose not to testify, the reception of such evidence effectively nullified the accused’s right to challenge the truthfulness of the co-accused in terms of s 35(3)(i) of the Constitution.

Other arguments against the Ndhlovu approach followed: first, that s 219A had indirectly, for admissions, secured the same protection for an accused as s 219 had for confessions by providing that an admission ‘by any person . . . shall . . . be admissible in evidence against him’ (the court’s emphasis), thus implying that it should not be admitted against anyone else. Second, that there was no sound jurisprudential basis for the distinction between confessions and admissions in this context: ‘[t]hat the characterisation of a statement as a confession or an admission could determine, without more, whether it falls to [be] admitted as evidence against a co-accused in and of itself provokes anxiety’ (at [54]). Incorrect characterisation, furthermore, may leave an appeal court with difficult choices, lead to a host of potential irregularities, and raise the danger of a conviction being vitiated. And, third, it would lead to an unhappy result if the rights of an accused to challenge evidence were more circumscribed under the constitutional order than under the old regime.

The argument of commentators that s 3(1) of the Law of Evidence Amendment Act contained sufficient safeguards to ensure the preservation of fair trial rights in that the evidence could be received only if the interests of justice allowed it, met with this rebuttal (at [67]): ‘Considering the rationale at common law for excluding the use of extra-curial admissions by one accused against another, it appears to us that the interests of justice [are] best served by not invoking the Act for that purpose’. The court found itself ‘compelled to conclude that our system of criminal justice underpinned by constitutional values and principles which have, as their objective, a fair trial for accused persons, demands that we hold, s 3 of the Act notwithstanding, that the extra-curial admission of one accused does not constitute evidence against a co-accused and is therefore not admissible against such co-accused’.

We have argued, in Commentary in the notes to s 219, that s 3(1)(c) is, despite the strong reservations expressed pre-Litako, sufficiently sensitive to meet all the concerns expressed by the courts and sufficiently flexible to ensure an accused a fair trial. Litako now makes it clear that the Supreme Court of Appeal disagrees. Litako creates a blanket exclusion of all admissions made by a co-accused. There can be no exceptions, and as a co-accused’s admission ‘does not constitute evidence against an accused’, the Supreme Court of Appeal has set its face against any inquiry into whether, in a particular case, such evidence could ever be admitted. Our argument in Commentary that such evidence – even though it will, in the vast majority of cases fall to be excluded – might, in a particular case call for admissibility in the interests of justice, was not considered.

There are important questions that one cannot help asking in the wake of Litako. First, how will courts deal with situations such as that which arose in S v Robiyana & others 2009 (1) SACR 104 (Ck)? In that case the appellants had cross-examined the accomplice on the contents of a statement he had made, which also implicated them, in order to show that he had contradicted what he had said earlier and that he should not be believed. This, as Greenland AJ observed, they were entitled to do in order to prove the witness a liar and to secure their acquittal. In fact, however, they failed to destroy his credibility. A blanket ban against the reception of this evidence against the appellants would have led to what the court in Robiyana described as an ‘incongruity’: to claim that the contents of the statement should be ignored and not used to confirm the accomplice’s credibility after a failed attempt to use the evidence to destroy his credibility would, said Greenland J, be ‘plainly absurd’. He concluded, quite correctly in our view, that the statement should be received under s 3(1)(c).

Robiyana, then, is one case where the interests of justice – if one were allowed to take them into account – would favour admissibility. A blanket exclusionary rule would, however, make such an inquiry impossible. And might there not be other cases where the interests of justice might be frustrated by a blanket ban? What if, for instance, the co-accused were shown to be a person of the highest credibility; his statement was made under oath in circumstances where the accused had ample opportunity to question him and to challenge his evidence; where there existed strong indicia of the reliability of his statement; and where there was, in fact (as opposed to theory) no cause to question the fairness of the trial if the evidence were adduced? Such circumstances would not arise often. But what if they did? Must we simply ignore the interests of justice and defer to an inflexible rule?

Our experience with issues such as relevance and hearsay is that fixed rules do not work. Rules that allow for flexible results – such as s 3(1)(c) – do much better if judicial officers are guided by being required to consider an unbounded set of factors to the end of reaching a result that best serves the interests of justice. There is no reason to believe that s 3 could not do a proper job in this context. The factors mentioned by the court in Litako in its attack on Ndhlovu are not, we believe, outside the proper compass of s 3(1)(c). The objections relating to the potential for an unfair trial flow from the hearsay nature of the evidence. The fact that the accused’s right to challenge the truthfulness of the co-accused in terms of s 35(3)(i) is compromised follows from the fact that the probative value of the evidence depends on the credibility of the absent declarant (the co-accused), the fact, that is, that it is hearsay. The objection relating to ‘irrelevance’ is, in our view, unfounded. There is nothing irrelevant about material in A’s statement that incriminates B. The relevance argument arises out of the fact that an admission by A will be relevant as an admission against B only if it is made in execution of a common purpose between A and B. Such a principle says nothing about relevance outside of this very narrowly articulated rule (see Commentary), and certainly does not stand in the way of receiving evidence, not as an admission made by A, but in terms of s 3(1)(c), which needs no such labels to render hearsay admissible and which lists, as one of the factors to be considered, the probative value of the evidence.

The argument arising out of the wording of s 219A, too, does not convince. That section sets out circumstances in which an admission made by A will be admissible against A. It thus creates an exception to the hearsay rule limited within this narrow ambit. That it says nothing about the use of a statement made by A against B cannot be interpreted as an indication that the legislature impliedly intended that such evidence would necessarily be inadmissible. If that statement by A is not an admission at all (as may, in fact, have been the case in Litako since the statement there was exculpatory in respect of its maker), then clearly s 219A cannot in any way be in point, and the matter falls squarely within the province of s 3(1) – which creates a separate, independent exception to the hearsay rule. Why should the position be different if the statement is an admission if tendered against A but not an admission if tendered against B (which it would be if it were made, say, in execution of a common purpose)? In such a case, s 219A is, again, simply not in point, and to bar the use of s 3(1) on the strength of the court’s reading of s 219A would be unjustifiable.

The fact that the cautionary rule, which would have been applied had the co-accused testified, has been frustrated flows, too, from the hearsay quality of the evidence. Caution could, in any event, be applied to the evidence of the statement itself.

How, then, if s 3(1)(c) is properly applied, can it not conduce to the right result if that result (to admit or to exclude) correctly reflects what is dictated by the interests of justice? If the blanket rule set out in Litako is a sign that the Supreme Court of Appeal does not have faith in the ability of the lower courts to determine where the interests of justice lie – and would, as a result, rather have an inflexible rule that produces the correct answer in the vast majority of cases, and wrong ones in a few cases, rather than one that is flexible enough to do get it right in all cases – that would be a sad indictment on the state of our system of criminal justice.

It would be interesting to see what the position of the Constitutional Court would be on this question.

Andrew Paizes

Dolus eventualis again

In S v Ndlanzi [2014] ZASCA 31 (unreported, SCA case no 318/2013, 28 March 2014) the Supreme Court of Appeal restated and followed the test for dolus eventualis set out by it in S v Humphreys 2013 (2) SACR 1 (SCA) and S v Tonkin 2014 (1) SACR 583 (SCA). The appellant in Ndlanzi was a taxi driver who, in an apparent attempt to evade an angry mob after striking one pedestrian with his vehicle, drove his taxi onto the pavement in order to flee from the mob. In doing so, he struck a newspaper stall, a dustbin and another pedestrian, who was walking on the pavement next to the entrance to the taxi rank. His vehicle continued to move on until it hit a stop sign, after which the appellant reversed and drove over the pedestrian, who had fallen to the ground. That pedestrian died as a result of the injuries. The appellant’s evidence, undisputed on this point, was that he had not seen the deceased; had seen only a newspaper stall and a concrete pole; and was not even aware that he had collided with the deceased.

The court held that the appellant must have foreseen the possibility of causing the death of a pedestrian by his actions. However, it accepted that he ‘believed he would be able to avoid colliding with the pedestrians on the pavement by turning to the right back onto the road’ (at [39]), and that, as a result, ‘it [could] not be inferred that it was immaterial to the appellant whether he collided with a pedestrian on the pavement’. It ‘[could] also reasonably be inferred that he may have thought that a collision with a pedestrian, which he subjectively foresaw, would not actually occur’. ‘In other words’, said Bosielo JA, ‘the appellant “took a risk which he thought would not materialise”’ (quotation from S v Humphreys at 10d).

The second element of dolus eventualis, said the court, was not satisfied, and his conviction for murder was set aside and replaced with one for culpable homicide.

The approach of our courts to the question of dolus eventualis was criticised in a feature article in (2013) 1 CJR. The meaning, utility and scope of the second element of dolus eventualis, as it was employed in Humphreys, was subjected to particular attack.

I find it necessary to return to these in the light of the court’s reiteration of the test set out in Humphreys in both Tonkin and Ndlanzi. It is difficult, in particular, to understand the view (as set out in Tonkin at [11]) that, ‘[i]f the perpetrator genuinely believed – despite the unreasonableness of that belief – that the foreseen consequences would not materialise, the element of reconciliation cannot be said to be present’. What does this mean? There can, it seems to me, be only two possibilities:

(i)         It means that A started off foreseeing that his act might cause B’s death but, when he actually acted, he believed it would not, so that he no longer foresaw the possibility of that result taking place. If it means this, then the initial foresight is irrelevant. Foresight is tested at the time of acting; if he no longer foresaw death at that point, it is the first element of dolus eventualis that is not satisfied, and the second need not be considered.

(i)         It means that, had you asked A at the time he acted, what the odds were of his act causing death, he would have said something like this: I accept that there is a real chance that B could be killed if I commit the act, but I accept, at the same time, that there is an even stronger chance that he will not be. There is nothing unusual about this equivocal state of mind. After all, the first element of dolus eventualis requires foresight of the possibility, not the probability, of the result occurring. So, in the ordinary course, an accused who satisfies this element will not uncommonly foresee the non-happening of the result as more likely than its happening. So, if this is the meaning to be given to the state of mind that would negate the second element, the result is that satisfying the first element will necessarily lead to a failure to satisfy the second unless what is foreseen is the probability of causing the result. But this is not the law and neither should it be. Dolus eventualis is built on the conscious acceptance of risk. Our courts have accepted, quite properly, that conscious risk-taking is based on what is foreseen as a real possibility, not a probability (see, for instance, R v Thibani 1949 (4) SA 720 (A) at 729–30). Interpreting the ‘second element’ in a way that requires the latter instead of the former sets the bar too high for the prosecution and leads, if taken to its logical conclusion, to inappropriate decisions.

The Supreme Court of Appeal seems to consider that ‘foresight’ and ‘belief’ exist on entirely different planes, so that an assessment of one cannot have a bearing on the measure of the other. This cannot be true. Once it is accepted, as it must in my view, that the planes intersect, the core of the problem is revealed: you cannot have two ostensibly separate elements that use essentially the same currency – in this case the extent of the appreciation of the risk. Any attempt to give meaning to the second element by identifying how much ‘foresight’ will make up a ‘belief’ necessarily makes inroads into the meaning of the first element by eroding what constitutes foresight of the real possibility of causing the prohibited result.

I submit, once more, that one can find no useful place in the analysis for the second element of dolus eventualis set out by our courts, and that a more nuanced approach (as suggested in 2013 (1) CJR) is necessary. That it is entirely inappropriate to be talking, in the second element, about whether X ‘believed’ the event would not happen or whether it was ‘immaterial’ to him whether or not it did, can be illustrated by these examples:

(i)         X deliberately sets out to torture Y in order to extract information from him, and subjects Y to a kind of Russian roulette situation, but one in which (hypothetically) the risk is, say, one in a hundred or, even, a thousand. If death ensues, it cannot be doubted that dolus exists, even if X ‘believed’ it would not or if it was not immaterial to him that that result might ensue.

(ii)          X places bombs in public places in the course of terrorist activities. He is a caring person, and places the bombs in ways that will minimise the loss of life, which he foresees as a real possibility but hopes fervently will not occur. It is not ‘immaterial to him whether this result comes about, but, if it does, he will certainly have dolus eventualis’ in that regard.

Another situation to test the validity of the approach in Humphreys, Tonkin and Ndlanzi is that which faced the courts in two recent cases. In S v Nyalungu 2013 (2) SACR 99 (T) the court held that a person who raped another in the knowledge that he was infected with the HIV virus and who did not take preventative measures could be convicted of attempted murder if the element of intention was satisfied. This decision has now been followed in S v Phiri 2014 (1) SACR 211 (GNP). The difference between the two cases is that the accused in Nyalungu had raped the complainant; in Phiri the act of sexual intercourse was consensual and took place within a love relationship.

This difference was not mentioned by the court. It was enough in its view that the appellant, who knew he was HIV positive, engaged in sexual intercourse with the complainant, whom he knew to be HIV negative, without any preventative measures, since he had mens rea in the form of dolus eventualis. The court took judicial notice of the fact that, at present, the disease has no cure and is likely to lead to a reduced life span, and seems to have inferred that the appellant must have been aware of that fact as well.

The court did not, however, consider the second leg of the test for dolus eventualis currently favoured by the courts. Did the appellant ‘reconcile himself’ to the result foreseen by him? If the appellants in Humphreys and Ndlanzi could, in cases involving motor accidents, escape findings of dolus eventualis on the ground that it was ‘not immaterial’ to them whether their conduct caused the death of the victims in question, should it not have been asked whether the appellant in Phiri, who was in a love relationship with the complainant, and who had had intercourse with her on no more than two occasions, should have been treated in the same way?

It is unlikely that the appellant in Phiri ‘reconciled himself’ with the possibility of causing his lover’s death in the sense in which that term has been understood by the Supreme Court of Appeal. It is not suggested, however, that that approach should be employed in the first place. The situations in Nyalungu and Phiri demonstrate the need for the more nuanced approach to dolus eventualis advocated in 2013 (1) CJR. Important questions need to be asked in the light of these two cases: is it appropriate to be speaking of dolus eventualis and murder when one is dealing with acts of sexual intercourse within a loving relationship? Is it different when the intercourse is an act of rape? Does it matter that the acts of intercourse were infrequent or regular features of a long-term relationship?

And these are questions best considered as part of the fine-grained approach to dolus set out in 2013 (1) CJR, and not within the inquiry into the second leg as that has come to be understood by the courts: to conclude, for instance, that the appellant in Phiri should escape liability for murder because it was ‘not immaterial’ to him whether his lover died or not would allow the terrorist in the above example to escape for the same reason. So, too, a rapist who hoped not to infect his victim even though he accepted the real risk of doing so by raping her in the first place. It is far better to ask whether it is appropriate to speak about dolus (and murder) when dealing with acts of consensual intercourse within marriage or other love relationships, or whether any unlawful causing of death in such circumstances, is, as in the case of motor accidents, more properly the province of culpa.

Andrew Paizes


The Criminal Law (Forensic Procedures) Amendment Act 37 of 2013

The President assented to the above Act (hereafter the ‘DNA Act’) on 23 January 2014. See GG 37268 of 27 January 2014. However, the DNA Act will only come into operation on a date determined by the President by proclamation in the Gazette. See s 9 of the DNA Act.

The DNA Act should not be confused with the Criminal Law (Forensic Procedures) Amendment Act 6 of 2010, which supplemented and amended Chapter 3 of the Criminal Procedure Act 51 of 1977 extensively, with effect from 18 January 2013. Act 6 of 2010 amended the existing s 37 by further regulating the powers of the police and the courts as regards the taking of body-prints or the ascertainment of bodily appearances of accused and convicted persons. Act 6 of 2010 also inserted three new sections into Chapter 3: ss 36A, 36B and 36C. These new sections deal with a wide variety of matters pertaining to the ascertainment of bodily features of persons, ranging from definitions to be used in the interpretation of Chapter 3 to the taking, retaking and destruction of fingerprints in specified circumstances.

Act 6 of 2010 did not introduce any specific DNA procedures. The idea was that DNA and related matters would be dealt with in a separate Act which would, once again, amend certain sections in the Criminal Procedure Act. And this will now be done by the DNA Act referred to above.

One of the main purposes of the DNA Act is to amend the Criminal Procedure Act, so as to provide for the compulsory taking of specified samples from certain categories of persons for the purpose of forensic DNA analysis. Several new definitions will be inserted into the existing s 36A once the DNA Act comes into operation. One example is the definition of ‘DNA’ as ‘deoxyribonucleic acid which is a bio-chemical molecule found in the cells and that makes each species unique’. This definition, it would seem, should be extended to include a reference to the uniqueness of the DNA of individuals within a species (excluding identical twins). This reference is important for forensic investigations.

Further examples of definitions are:

In terms of a new s 36A(3) buccal samples would have to be taken, with strict regard to decency and order, by an authorised person who is of the same gender as the person from whom such sample is required.

Section 2 of the DNA Act inserts two new sections in Chapter 3 of the Criminal Procedure Act, namely ss 36D and 36E. Section 36D will regulate powers in respect of buccal samples, bodily samples and crime scene samples. This section covers aspects such as compulsory taking of buccal samples (s 36D(1)); supervision where a person requests to take his own buccal sample (s 36D(3)); retaking of samples on certain conditions (s 36D(5)); use of a forensic DNA profile to conduct a comparative search (s 36D(6)) and criminalisation where a bodily sample, crime scene sample or forensic DNA profile is used for any purpose other than the purposes permitted in the Act (see s 36D(7)(c) as read with s 36D(7)(a)).

Section 36E regulates the taking of samples for comparative investigation purposes, for example where an authorised person not only suspects that a person, or one or more persons in a group, have committed an offence identified in the newly inserted Schedule 8, but also believes that the buccal sample or the result of the forensic DNA analysis thereof ‘will be of value in the investigation by the excluding or including of one or more of those persons as possible perpetrators of the offence’. Where a person refuses to provide a buccal sample for purposes of s 36E, a warrant may be issued by a judge or magistrate in the circumstances specified in s 36E(2).

Section 6 of the DNA Act inserts a completely new chapter in the South African Police Service Act 68 of 1995. Chapter 5B will deal with the establishment, administration and maintenance of the National Forensic DNA Database of South Africa (NFDD). Section 15F of Act 68 of 1995 states that the objective of Chapter 5B would be to establish and maintain a national forensic DNA database in order to perform comparative searches for the following purposes:

(a)     to serve as a criminal investigative tool in the fight against crime;

(b)     to identify persons who might have been involved in the commission of offences, including those committed before the coming into operation of this Chapter;

(c)     to prove the innocence or guilt of an accused person in the defence or prosecution of that person;

(d)     to exonerate a person convicted of an offence; or

(e)     to assist with the identification of missing persons or unidentified human remains.

The other sections in Chapter 5B of Act 68 of 1995 will deal with the following: a national forensic DNA database (s 15G); a crime scene index (s 15H); an arrestee index (s 15I); a convicted offender index (s 15J); an investigative index (s 15K); an elimination index (s 15L); a missing persons and unidentified human remains index (s 15M); comparative forensic DNA search and communication of information (s 15N); foreign and international law enforcement agencies (s 15O); analysis retention, storage, destruction and disposal of samples (s 15Q); infrastructure (s 15R); offences and penalties (s 15S); awareness and training programmes (s 15T); access to and security of the NFDD (s 15U); the establishment and composition of a National Forensic Oversight and Ethics Board, and various matters pertaining to this Board, such as meetings, functions and disciplinary recommendations (ss 15V–15AB); parliamentary oversight (s 15AC) and regulations (s 15AD).

It is clear that the success or otherwise of the DNA Act, once it is put into operation, will depend on the training that police officials and all other role players have received; the quality of equipment used; the integrity of the system and, most importantly, the verification of the authenticity of all samples taken, analysed and compared for forensic purposes as identified in the DNA Act.

The setting up of the NFDD as provided for in Chapter 5B of the South African Police Service Act 68 of 1995, is – in terms of financial and human resources – a massive enterprise. However, it is absolutely necessary that the South African criminal justice system should enjoy the benefits and services that the NFDD would be able to offer. As recently as 27 March 2014 the Supreme Court of Appeal expressed its concern about the lack of DNA testing and evidence, which contributed to the acquittal of two convicted rapists on appeal. See S v Mugwedi [2014] ZASCA 23 (unreported SCA case 694/13, 27 March 2014). The decision in S v SB 2014 (1) SACR 66 (SCA) contains valuable insights into the technical and scientific basis upon which DNA evidence rests (at [8]–[16]). In dealing with this evidence, Van der Merwe JA expressed his debt to Meintjes-Van der Walt’s DNA in the Courtroom: Principles and Practices (Juta 2010).

It is difficult to predict when the DNA Act will be put into operation. According to press reports, the forensic training of police officers – which was supposed to have commenced in April – was moved to September, and is now expected to start even later as various logistical matters and tenders must still be finalised (Cape Argus, 17 May 2014).



(a) Criminal Law

The principle of legality: Statutory offences in Prevention of Organised Crime Act 121 of 1998 (POCA) challenged: Validity of provisions of POCA considered

In Savoi & others v National Director of Public Prosecutions & another 2014 (1) SACR 545 (CC), one particular aspect of which is the subject of a feature article above (see above), the Constitutional Court considered a number of different attacks on various provisions of POCA (offences relating to racketeering and other related crimes) in the context of the dictates of the principle of legality.

One attack was on the definitions of a ‘pattern of racketeering activity’ and ‘enterprise’ in ss 1 and 2(1)(a) to (g). It was argued that the entire concept of a ‘pattern of racketeering’ was vague because the factors that affected the requirements for establishing it were so numerous and varied; and that the definition of ‘enterprise’ was overbroad and thus unconstitutional. This attack failed in both respects. The first because: (1) the participation had to be in an offence specified in Schedule 1; (2) the word ‘planning’ indicated that the offences had to be interconnected to form a sequence, a part of an elaborate plan; (3) the fact that some of the offences were somewhat obscure did not signify since what the rule of law required was reasonable certainty, not absolute or perfect lucidity; and (4) the objection that some of the offences were not, if viewed in isolation, those ordinarily associated with organised crime and appeared to be ‘ordinary’ or ‘garden variety’ commercial offences ignored the complex web-like manner in which organised crime operated by infiltrating different areas of economic activity – some relatively minor at face value – over time in a complex combination (at [25]).

The second, the attack on ‘enterprise’ for being overbroad, was, said the court, ‘misconceived’ (at [31]), since overbreadth was not, in our constitutional jurisprudence, a self-standing ground of statutory constitutional invalidity. It comes into the equation in the justification analysis in terms of s 36(1) once a law of general application had been found to limit a right.

The next attack aimed at the alleged retrospectivity of some of the offences, arising out of the fact that the definition of the ‘pattern of racketeering activity’ in s 1 of POCA required at least two offences listed in Schedule 1, one of which was allowed to pre-date POCA, the other not. It was held that the rule against retrospectivity was not violated since ‘[a] person that completes the circle [by committing the second offence] does so with their eyes wide open’ and has to ‘live with that choice’ (at [79]). As a result, ‘[t]he inherent injustice, unfairness and cruelty, which [made] a criminal offence created retrospectively constitutionally impermissible, [were] simply not there’.

The final challenge centred around the words ‘ought reasonably to have known’ in s 2(1)(a)(ii), (b)(ii), (c)(iii) and (f) of POCA. These words, it was argued, exposed an accused to ‘the possibility of punishing an unintended, insensible or unconscious conduct’ and were, moreover, vague and unintelligible. This attack, too, was unsuccessful. The language used, said the court, classically captured the idea of negligence (at [88]), and it was the legislature’s choice to make intention or negligence the basis of the required fault element of a statutory offence. There was nothing to suggest, in these circumstances, that the standard of negligence was in any way unsuitable.

Common purpose: Active association – requirements

The requirements for holding an accused liable for an offence on the basis of common purpose formed as a result of an active association – as opposed to a mandate or agreement – were restated by the Supreme Court of Appeal in S v Dewnath [2014] ZASCA 57 (unreported, SCA case no 269/13, 17 April 2014). The appellant was, at a certain point, with his parents while they were negotiating with X to assassinate Y, the appellant’s uncle. At that stage the price was being negotiated, and the appellant said this to X: ‘But why are you asking for so much money? The person that we are asking you to kill is absolutely worthless. I would understand if he was a member of the taxi business. If I wasn’t involved in the police, with the police, I would kill him myself’. He then left the room, leaving X with his parents as they continued the negotiations. There was no evidence that the appellant was privy to any further negotiations or that he approved of what was decided in his absence. The appellant’s parents, however, paid X for killing Y.

The appellant was convicted of murder by the trial court. In the appeal, it was argued that the court erred in convicting him on the basis of common purpose. It was contended that the state had ‘failed to prove that the words uttered by him were sufficient to form an active association with the common purpose, between his parents and their co-accused, to kill the deceased’ (at [7]).

Mocumie AJA (with whom the rest of the court agreed) set out the requirements of common purpose by active association as they were laid down in S v Mgedezi & others 1989 (1) SA 687 (A) at 705–6, but warned that it was ‘critical’ to ‘curb too wide a liability’ (at [15]). ‘Current jurisprudence,’ he added, ‘premised on a proper application of S v Mgedezi & others, [made] 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’.

It was held that there was no evidence that the appellant actively participated in the murder apart from the words uttered by him. Mere approval of the murder, said the court, was not enough. What he said did not amount to active association; his ‘participation’ was insignificant; it was limited and removed from the actual action. There was, moreover, no basis to conclude that he intended to kill Y.

The court was clearly correct in finding that there was no active association. Indeed, it is difficult to understand why active association was even considered by the trial court, since that form of common purpose is ordinarily and more properly invoked in situations of mob violence, where parties join others who are already involved in violent acts and where no form of agreement, whether express or implied, can be said to exist. It might, I submit, have been more profitable for the state to have argued for common purpose in its more common manifestation – that based on agreement or ‘mandate’. The trial court seems to have ‘accepted the appellant’s version that he had no prior agreement with [X] and his parents to kill the deceased’ (at [13]). But it found, nonetheless, that ‘the only reasonable inference to be drawn was that when the appellant uttered those words, he did so with the intent to persuade [X] to carry out the plan and force him to abandon or forgo his demand for a down payment or deposit’ (at [13]). Was this, then, not enough to suggest the giving of a ‘mandate’ to X to murder Y?

Dissociation from a common purpose

The court in S v Wana & others (unreported, ECP case no CC 16/2013, 19 & 20 March 2014) examined what circumstances might warrant the conclusion that a party to a common purpose by mandate or agreement (as opposed to active association) had done enough to dissociate himself from the common purpose and escape liability.

It had been stressed in S v Nduli & others 1993 (2) SACR 501 (A) at 504D that the more advanced his participation in the commission of the crime was, the more pertinent and pronounced his conduct would have to be to establish dissociation. It was, said the court, a matter of fact and degree. The court in S v Beahan 1992 (1) SACR 307 (ZS) at 324bc said it depended on the role played by the person:

Where a person has merely conspired with others to commit a crime but has not commenced an overt act toward the successful completion of that crime, a withdrawal is effective upon timely and unequivocal notification to the co-conspirators of the decision to abandon the common unlawful purpose. Where, however, there has been participation in a more substantial manner something further than a communication to the co-conspirators of the intention to dissociate is necessary. A reasonable effort to nullify or frustrate the effect of his contribution is required

In Nduli the Appellate Division found it unnecessary to decide whether the dictum in Beahan constituted a rule of law in South Africa or was merely at best a ‘rule of thumb’ (see 506–7). So, too, in S v Lungile & another 1999 (2) SACR 597 (SCA) at 603, where it was held that whatever view one took of the matter, there was no effective dissociation, since the reason for the accused fleeing from the scene was, in all likelihood, because of fear of arrest or injury or to make good his escape with the stolen property.

In Wana, however, Goosen J said that it was necessary to consider the status and effect of the Beahan dictum. That case involved a conspiratorial criminal enterprise which had already advanced to a substantial degree by the time of the alleged dissociation. He considered that the dictum proceeded from the logical premise that the greater the involvement and the more advanced the execution of the criminal enterprise, the more clearly an accused had to establish a basis for a finding of dissociation. He hastened to add that this would not place an onus on an accused. It pointed to the need, rather, to establish the performance of some positive act of withdrawal or dissociation as well as a clear and unambiguous intention to withdraw from the criminal enterprise.

As to what conduct was necessary, Goosen J, relying on the views of Snyman Criminal Law 5 ed (2008) at 267, said that this depended on a number of considerations, and might, depending on the circumstances, include an attempt to dissuade co-conspirators from proceeding with the plan or taking steps to thwart or prevent the performance of the enterprise, such as reporting timeously to the police or providing the police with the means to prevent the commission of the offence.

The dictum in Beahan commended itself to Goosen J as ‘sound in principle and . . . consistent with our law’ (at [204]). In his view a co-conspirator who had, by his conduct and actions, played a central role in the initiation of the enterprise and also proceeded to facilitate its execution had, if he wished to establish effective dissociation from the enterprise, actively to set out to undo the conspiracy, or, if that was not possible, to thwart its execution. What would suffice would depend, of course, on the facts of each case.

In Wana the accused had played a significant role in devising the plan to commit the planned robbery. He functioned as the ‘key link’ in securing buyers for the platinum that was to be stolen; he had done so previously; he brought other persons into the planning process; he was, for a period, in custody but returned immediately afterwards to the central role he had previously played.

His own account of what he did to withdraw included: telling a co-conspirator of his wish to withdraw; going ahead with meetings with another ostensible co-conspirator (who, it turned out, was a police trap) only so as not to alert the others of his decision to withdraw; and informing a policeman in vague terms – but only on the evening before the robbery, and only by giving him the name of a general area. The policeman was not given any details as to what was to take place or where, specifically, it would take place. It was clear from this, said Goosen J, that ‘no effective steps could then be taken to frustrate the carrying out of the prior agreement to commit an offence nor to dissuade the co-conspirators from continuing with that criminal enterprise’ (at [209]). There was, moreover, evidence to suggest that the accused continued to play his role in the robbery on the day of the incident, which was to monitor the movement of the security vehicle carrying the platinum from the airport to the spot where the robbery was to take place, and to furnish his co-conspirators with information to that effect.

In conclusion (at [211]), the accused’s conduct did ‘not, in the circumstances of this case, establish that he manifested an unequivocal intention to dissociate himself from the commission of the offences or that his positive act was sufficient to establish dissociation from the commission of the crimes which he had conspired to commit with his co-conspirators’.

Conspiracy – agreement as to means of carrying out enterprise not necessary – timing of violence for robbery

It was held in S v Wana & others (unreported, ECP case no CC 16/2013, 19 & 20 March 2014) that it is sufficient for establishing a conspiracy if the evidence discloses an agreement to commit an unlawful act which is the object or purpose of the agreement, and it is not necessary, further, to establish the means by which the enterprise is to be carried out (see R v Adams & others 1959 (1) SA 646 (SCC)).

In Wana the evidence had established that agreement had been reached, since matters had clearly gone beyond the stage of negotiating about whether the act should be done or not. The only reasonable inference from the facts was that the accused foresaw the possibility that, in the execution of their criminal enterprise, violence or the threat of violence might be necessary in order to achieve the object of the enterprise – the theft of platinum from an armed vehicle in a public street in broad daylight where at least one person not a party to the enterprise was armed.

The court turned to the essential elements of the crime of robbery and affirmed what was said in S v Yolelo 1981 (1) SA 1002 (A): that the use of force or violence did not have to precede the taking of the stolen property. It is enough if the violence is sufficiently closely connected to the process of the taking that the violence and the taking constitute a single course of conduct (see [142]).


(b) Criminal Procedure and Evidence: Pre-sentence

Chapter 1: Review of prosecuting authority’s decision to withdraw charges, and the validity of a mandatory interdict to prosecute

National Director of Public Prosecutions & others v Freedom Under Law [2014] ZASCA 58 (unreported, SCA case 67/2014, 17 April 2014)

In Freedom Under Law v National Director of Public Prosecutions & others 2014 (1) SA 254 (GNP) a public interest organisation known as ‘Freedom Under Law’ succeeded in obtaining a mandatory interdict directing the National Prosecuting Authority to reinstitute, and prosecute to finalisation, several withdrawn criminal charges against the fifth respondent in that matter. Murphy J held that a review of the decision not to prosecute was constitutionally permissible and required and, furthermore, not excluded by the Promotion of Administrative Justice Act 3 of 2000 (at [124]–[126]). In terms of s 1(ff) of PAJA only decisions to institute or continue a prosecution are excluded from the definition of administrative action. And a decision ‘to withdraw criminal charges or to discontinue a prosecution’ held Murphy J at [131], ‘. . . meets each of the definitional requirements of administrative action’. See further the discussion of this case in Commentary on the Criminal Procedure Act, Chapter 1, sv The decision in Freedom Under Law v National Director of Public Prosecutions & others.

The above case resulted in an appeal to the Supreme Court of Appeal in National Director of Public Prosecutions & others v Freedom Under Law [2014] ZASCA 58 (unreported, SCA case no 67/2014, 17 April 2014). Brand JA, writing for a unanimous full bench, took the view that ‘the time’ had come ‘to put . . . to rest’ the issue whether prosecutorial decisions to discontinue or withdraw a prosecution were reviewable under PAJA, even though the answer to this issue was ‘by no means decisive of the matter’ (at [19]). In rejecting the court a quo’s view that a decision of non-prosecution is of a different genus to the decision to institute a prosecution, Brand JA observed and reasoned as follows (at [24]):

To say that the validity of a decision to prosecute will be tested at the criminal trial which is to follow, is, in my view, fallacious. What is considered at the criminal trial is a determination on all of the evidence presented in the case of the guilt or lack thereof of the accused person, not whether the preceding decision to prosecute was valid or otherwise. The fact that an accused is acquitted self-evidently does not suggest that the decision to prosecute was unjustified. The reason advanced by the court a quo itself, namely, that a decision not to prosecute is final while a decision to prosecute is not, is in my view equally inaccurate. Speaking generally, both these decisions can be revisited through subsequent decisions by the same decision-maker, by in the one case re-instituting the prosecution, and by withdrawing the prosecution in the other.

At [25] it was noted that an important policy consideration for courts limiting their own power to interfere in prosecutorial decisions was the safeguarding of the independence of the prosecuting authority. It was accordingly held that although on a strictly textual interpretation the exclusion in s 1(ff) in PAJA is limited to decisions to prosecute, it must be understood to incorporate decisions not to prosecute (at [27(d)]). In doing so, the court aligned itself with an earlier dictum in Democratic Alliance & others v Acting National Director of Public Prosecutions & others 2012 (3) SA 486 (SCA) at [27], where Navsa JA noted that there appeared to be ‘some justification for the contention that the decision not to prosecute [was] of the same genus as a decision to institute or continue a prosecution . . .’

Having come to the above conclusion, Brand JA was quick to point out that ‘[t]he legality principle has by now become well-established in our law as an alternative pathway to judicial review where PAJA finds no application’ (at [28]). Indeed, it was conceded by the first appellant that the impugned decisions not to prosecute were subject to a principle of legality or a rule of law review by the court (at [19]). This form or ground of review is a safety net, providing the courts with a measure of control over action which does not qualify as ‘administrative action’ under PAJA, but nonetheless concerns the exercise of public power (at [29]–[35]).

It was argued on behalf of the appellants that the withdrawal of the charges was provisional because it was by s 6(a) and not s 6(b) of the Criminal Procedure Act, and that such provisional withdrawal could not be subjected to review. Section 6(a) provides for the withdrawal of a charge before an accused has pleaded, and a plea of prior acquittal is not available should the withdrawn charges be reinstituted at any time. However, stopping of the prosecution – as provided for in s 6(b) – takes place after an accused has pleaded. In such an instance the court must acquit the accused; and upon reinstitution of the same charge(s), a plea of prior acquittal would be successful. The mere fact that the impugned decisions were in fact decisions under s 6(a) made no difference to Brand JA for purposes of the competency of the review (at [34]):

Although I am in agreement with the premise of the argument, that both decisions to withdraw were taken in terms of s 6(a), my difficulty with its further progression is twofold. First, I can see no reason why, at common law, a decision would in principle be immune from judicial review just because it can be labelled ‘provisional’ however illegal, irrational and prejudicial it may be. My second difficulty is more fundamental. I do not believe a decision to withdraw a criminal charge in terms of s 6(a) can be described as ‘provisional’ just because it can be reinstituted. It would be the same as saying that because a charge can be withdrawn, the institution of criminal proceedings is only provisional. As I see it, the withdrawal of a charge in terms of s 6(a) is final. The prosecution can only be recommenced by a different, original decision to reinstitute the proceedings. Unless and until it is revived in this way, the charge remains withdrawn. 

Brand JA held, in relation to the withdrawal of the fraud and corruption charges, that the appellant concerned had not taken the decision in accordance with the dictates of the empowering statute (at [38]–[42]). However, as far as the withdrawal of the murder and seventeen other related charges was concerned, Brand JA accepted that the purpose was to avoid ‘a fragmentation of trials’ – a purpose which was not irrational on account of facts as set out in [44] of the judgment. In this respect the appeal succeeded, except for the fact that the following undertaking furnished on behalf of the National Director of Public Prosecutions was recorded as paragraph 3 of the court’s order:

 (a) To decide which of the criminal charges of murder and related crimes that were withdrawn on 2 February 2012, are to be reinstituted and to make his decision known to the respondent within 2 months of this order.

(b) To provide reasons to the respondent within the same period as to why he decided not to reinstitute some – if any – of those charges.

Brand JA was satisfied that the mandatory interdict amounted to an inappropriate transgression of the separation of powers doctrine: ‘In terms of the Constitution the NDPP is the authority mandated to prosecute crime . . . the court will only be allowed to interfere with this constitutional scheme on rare occasions and for compelling reasons . . . The setting aside of the withdrawal of the criminal charges . . . [has] the effect that the charges and the proceedings are automatically reinstated and it is for the executive authorities to deal with them. The court below went too far’ (at [51], emphasis added). The ball is now, for all practical purposes, in the court of the NDPP to ensure that all the circumstances which formed the subject matter of the litigation as described do not ‘develop’ into a rare occasion where there might be compelling reasons for a court to issue a mandatory interdict.

The decision of the Supreme Court of Appeal should be read with the decision of Gorven J in Booysen v Acting National Director of Public Prosecutions & others [2014] 2 All SA 391 (KZD). In this case the applicant – who did not rely on PAJA – succeeded in having certain prosecutorial authorisations and the decision to prosecute him, set aside on review. He relied on the principle of legality.

s 73: Legal representation and the presumption of regularity

Ndlanzi v S [2014] ZASCA 31 (unreported, SCA case no 318/2013, 28 March 2014)

In the above matter the appellant alleged that, at the trial, his former counsel and attorney had adopted a trial strategy in conflict with his instructions and, furthermore, had failed to discuss this trial strategy with him (at [20]). Bosielo JA, writing for a unanimous full bench, noted that the allegations, if true, might justify a conclusion that the accused had had an unfair trial (at [25]). However, at [25] he also noted that the appellant’s allegations were not supported by evidence ‘other than the mere say-so of the appellant’ through his subsequent counsel.

Several factual details contained in the trial record rendered the appellant’s allegations unlikely. At [25] Bosielo JA observed that at his trial the appellant had more than enough time to raise an objection, if any, to the manner in which the trial was conducted but that he – ‘[q]uite inexplicably’ – had failed to do so.

One is left with the impression that the appellant sought a new strategy when it became clear that the initial trial strategy was not going to produce the outcome he desired. ‘The appellant’ said Bosielo JA at [27] ‘had met his Waterloo and the only escape route was to put the blame on his counsel. Hence the belated volte face’.

In rejecting the appellant’s contention, it was said that an ‘important fact which weighs heavily against the belated volte face is that, given the strict ethics governing the lawyer’s profession, a presumption of regularity operates in favour of accepting that the lawyers acted in terms of their mandate from the appellant’ (at [28]). It is, with great respect, strange that the court should in the context of this case have resorted to reliance upon a presumption of regularity which, somehow, stems from the ‘strict ethics’ of the profession. It is submitted that courts should resist the temptation to solve lawyer–client conflicts on the basis that the conduct of lawyers must be presumed to have been regular. In Ndlanzi Bosielo JA indicated that ‘cogent evidence’ is required ‘to displace this presumption’ (at [28]). This observation seems to indicate that the presumption of regularity, as identified by the court, is one of fact and, fortunately, not of law which would place a burden of proof on a client.

s 112: Irregularity where prosecutor explained rights to an undefended accused

S v Mbathsha [2013] ZAECGHC 114 (unreported, ECG case no CA&R 328/13, 7 October 2013)

A prosecutor must display the highest degree of fairness to an accused, especially if the latter has no legal representation. See S v Mofokeng 1992 (2) SACR 261 (O) at 264c.

However, even in the course of a prosecutor’s bona fide efforts to achieve this admirable goal, matters can go wrong – as is evident from S v Mbathsha (supra). In this case the prosecutor informed the court at the commencement of proceedings and before the accused was asked to plead, that he had earlier and out of court explained the accused’s rights to him, including the right to silence, and that the accused understood his rights and wished to represent himself. The magistrate accepted these statements and directed the prosecutor to proceed. The accused then pleaded guilty to the charge of housebreaking with intent to steal and theft. After judicial questioning in terms of s 112(1)(b) of the Criminal Procedure Act, the accused was sentenced to two years’ imprisonment.

On review Roberson J was concerned that the prosecutor, and not the magistrate, had explained the accused’s rights to him. In response to the review court’s query, the magistrate stated that he had no reason to doubt what had been conveyed to him by the prosecutor (at [5]). However, Roberson J found it problematic that the record did not reflect that the accused’s rights were properly explained (at [7]). In this context reference was also made to S v Malatji & another 1998 (2) SACR 622 (W) where it was found unacceptable that a magistrate should delegate his judicial duty to explain the rights of the accused to the interpreter in circumstances that left no record of precisely what was conveyed to the accused.

In Mbathsha (supra) at [8] Roberson J concluded that the impropriety of delegating to the prosecutor the judicial duty of explaining rights to an undefended accused ‘is rendered even more serious when, within the adversarial criminal justice system, the duty is delegated to the very person who is prosecuting the accused’. This was a fatal irregularity in that it negated the constitutional right to a fair trial, vitiating the whole trial (at [9] and [10]). It should be noted that the bona fides of the prosecutor concerned was at no stage questioned. The presence or absence of bona fides could have no effect on the fact that fair trial considerations require that a prosecutor should not perform judicial duties. The fairness of the trial was compromised.

Mbathsha (supra) should also be compared with S v Bothma 1971 (1) SA 332 (C), where it was held that a prosecutor should not administer the oath to a witness.

s 113(2): Consequences of a court’s correction of a plea of guilty to a competent verdict accepted by the prosecutor

S v Swartz 2014 (1) SACR 461 (NCK)

Section 113(2) of the Criminal Procedure Act provides that if a court records a plea of not guilty under s 113(1) of the Act ‘before any evidence has been led, the prosecution shall proceed on the original charge laid against the accused, unless the prosecutor explicitly indicates otherwise’. The effect of this provision is that where a prosecutor has accepted an accused’s plea of guilty to a charge which is a competent verdict on the main charge and a plea of not guilty is recorded by the court after the section-112(1)(b) questioning of the accused in respect of the competent verdict, the prosecutor can once again consider whether the main charge against the accused should be pursued. See S v Swartz 2014 (1) SACR 461 (NCK) at [15]–[28]. However, in the absence of any indication by the prosecution that further proceedings should be confined to the competent verdict charge, the trial shall proceed on the main charge. At [41] Olivier J held that there is nothing in the wording of s 113(2) requiring ‘an election by the prosecutor, as to the charge in respect of which the prosecution is to proceed, before the trial can proceed’ (emphasis in the original). To put the matter differently: the default position is that the trial would proceed on the original charge(s).

s 153: Televised criminal proceedings and the trial of Oscar Pistorius

The electronic, broadcast and print media approached the High Court in Multichoice (Pty) Ltd & others v National Prosecuting Authority & another: In re S v Pistorius 2014 (1) SACR 589 (GP) for permission to broadcast the entire criminal proceedings in the murder trial of Oscar Pistorius. They sought to do so through audio, audio-visual and photographic means. It was a matter which, as Mlambo JP observed, brought into sharp focus the relation between the functioning of the criminal justice system, on the one hand, and the quest by the media and press to participate in that system, on the other.

As a result, Mlambo JP pointed out, a number of important constitutional rights were set to collide: the rights of an accused person, on one hand, and the freedom of expression rights of the media and the principle of open justice, on the other.

It was argued, for Pistorius, that the live broadcasting of his trial would infringe his rights to a fair trial in that: the mere presence of cameras would inhibit him as well as his witnesses while testifying; his counsel might be inhibited in the questioning of witnesses and the presentation of his case; and witnesses still to testify would be enabled to fabricate and adapt their evidence based on their knowledge of what prior witnesses might have testified.

The applicants, on the other hand, relied on the principle of open justice and s 16 of the Constitution which, in s 16(1)(a), guarantees everyone the freedom of expression, including the freedom of the press and other media as well as the freedom to receive and/or disseminate information and ideas (see South African National Defence Union v Minister of Defence & another 1999 (4) SA 469 (CC) at [7] and Khumalo & others v Holomisa 2002 (5) SA 401 (CC) at [22]).

After considering the decisions in Dotcom Trading 121 (Pty) Ltd t/a Live Africa Network News v King NO & others 2000 (4) SA 973 (C), SA Broadcasting Corporation Ltd v Thatcher & others [2005] 4 All SA 353 (C), South African Broadcasting Corporation Ltd v Downer SC NO & others [2007] 1 All SA 384 (SCA) and South African Broadcasting Corporation Ltd v National Director of Public Prosecutions 2007 (1) SA 523 (CC), Mlambo JP considered how best to resolve the contestation of rights in this matter. He turned to s 173 of the Constitution, which gives the courts the ‘inherent power to protect and regulate their own process, and to develop the common law, taking into account the interests of justice’, and found it inevitable that a balancing exercise would have to be engaged to resolve the problem, the overarching objective of which was serving the interests of justice.

The test most apposite to the attainment of this end was, he considered, that laid down by the Supreme Court of Appeal in Midi Television Pty Ltd t/a E-TV v Director of Public Prosecutions (Western Cape) 2007 (5) SA 540 (SCA) at [9], where it was stressed that opposing rights ‘cannot be reconciled by purporting to weigh the value of one right against the value of the other and then preferring the right that is considered to be more valued, and jettisoning the other, because all protected rights have equal value. They are rather to be reconciled by recognising a limitation upon the exercise of one right to the extent that it is necessary to do so in order to accommodate the exercise of the other (or in some cases, by recognising an appropriate limitation upon the exercise of both rights) according to what is required by the particular circumstances and within the constraints that are imposed by s 36 [of the Constitution]’.

His approach was, then, ‘to ensure that each of the rights asserted find proper expression and enjoyment without being unduly limited’ (at [19]). He was not persuaded by the objection by Pistorius that the coverage of the trial should be jettisoned to the degree he suggested, and found merit in the applicants’ argument that to do so would limit the information on the trial to only a small minority having access to the electronic media and leave the majority reliant on second-hand reporting which was liable to be inaccurate and truncated.

To accede to the objection would be to jettison entirely ‘the noble objectives of the principle of open justice, when one takes cognisance of our development along the democratic path’ (at [22]). Mlambo JP concluded that he could not, in this day and age, ‘countenance a stance that seeks to entrench the workings of the justice system away from the public domain’. The Constitution was underpinned by values of openness and accountability, and ss 34 and 35(3)(c) made it clear that criminal proceedings were to be public, a fact that encouraged public understanding as well as accountability.

There was, however, merit in the objection relating to the inhibitory effect of televising proceedings on defence witnesses. Audio coverage, in his view, was less inhibitory and less intrusive, and should be allowed in respect of the testimony of defence witnesses in the light of the enormous public interest and the importance of dispelling negative and unfounded perceptions about unequal treatment of rich and poor in South Africa.

Subject to the trial court’s discretion to vary aspects of the relief granted, Mlambo JP then made an order (see [30]) giving detailed technical specifications, giving effect to these findings. It was held, too, subject to this constraint, that audio-visual broadcasting or recording could take place in respect of: first, the evidence of all experts called by the state, excluding the evidence of the accused and his witnesses; second, the evidence of any police officer or former police officer in relation to the crime scene; and, third, the evidence of all other witnesses for the state unless any such witness or witnesses did not consent to the broadcasting or recording and the presiding judge ruled that it should not take place.

The question of the televising of criminal trials will be explored in greater depth in Revision Service 53 to Commentary.

s 166: Duties of a prosecutor when irregularities arise that undermine the fairness of a trial

S v Macrae & another [2014] ZASCA 37 (unreported, SCA case no 93/2013, 28 March 2014) is an important reminder of the duties owed to the court by the prosecution. The magistrate in that case had committed a number of irregularities that had rendered the trial unfair. These included an inordinately lengthy questioning of the accused (10 pages in the record) after he had testified; an instruction to the accused that he could not put questions to his own witness at a certain stage of the trial; the cutting short of legitimate cross-examination by the accused; and a ‘bizarre’ instruction to the accused that he could not cross-examine on a statement without first proving its authenticity, even though that statement had been furnished to the accused by the prosecution.   

The prosecution’s fault was that it did not concede at the outset that the trial was unfair when it clearly was. Instead, the office of the Director of Public Prosecutions not only pursued the prosecution, but defended the conviction in the full court and even resisted leave to appeal being granted to the Supreme Court of Appeal.

Wallis JA stressed ‘once again that the duty of prosecutors is not to secure a conviction at all costs or to defend convictions once obtained. Their duty is to see that so far as possible justice is done.’ He added (at [28]):

Where an appeal is being argued one expects the prosecutor to do so in an objective and fair manner and, if satisfied that the conviction is flawed, to draw that to the attention of the court, particularly where the flaw goes to the heart of the fairness of the trial at which the accused person was convicted.

 s 201: Legal professional privilege – does it apply to the invoices of attorneys?

In A Company & Two Others v Commissioner for the South African Revenue Services [2014] ZAWCHC 33 (unreported, WCC case no 16360/2013, 17 March 2014), the court had to decide whether the feenotes (invoices) of attorneys were subject to legal professional privilege. The applicants had supplied the invoices to SARS in terms of the provisions of s 46 of the Tax Administration Act 28 of 2011 but had redacted the portions that they claimed were privileged.

Binns-Ward J, finding no South African cases directly in point, turned to the English authorities, which he found to be divided. In his view, the modern English law would be likely to take the view that feenotes do not, as a rule, attract privilege, but may do so in a particular case if the conditions for that privilege were satisfied.

He found the general approach to privilege in the English law to be best stated by Taylor LJ in Balabel & another v Air India [1988] 2 All ER 246 (CA), where it was said that it had to be recognised that there was a ‘continuum’ of communications and meetings between solicitor and client, and privilege arose whenever information was passed from one to the other as part of that continuum, aimed at keeping both informed so that advice might be sought and given as required. It was too narrow, said Taylor LJ, to say that privilege attached only if the communication in question specifically sought or conveyed advice.

The feenotes of an attorney, Binns-Ward J found, were not amenable to a blanket rule of privilege: they were not created for the purpose of giving advice and were not ordinarily related to the performance of the attorney’s professional duties as a legal adviser. Their purpose was to obtain payment for duties already performed and did not form part of the ‘continuum’ of communications described by Taylor LJ. In a particular case privilege could arise if the invoice set out the substance of the advice or contained sufficient particularity of its substance to constitute secondary evidence of the advice; mere reference to the advice would not be enough.

In the present case the applicants had provided no context for such a determination to be made. SARS had consented to the court taking a ‘judicial peek’ at the covered up parts to make an assessment. Such a practice, although allowed by the courts on occasion, has elicited strong reservations from both the Supreme Court of Appeal and the Constitutional Court (see President of the Republic of South Africa v M & G Media Ltd 2011 (2) SA 1 (SCA) and 2012 (2) SA 50 (CC)), since it is important that courts earn the trust of the public by conducting their business openly and with reasons for their decisions. It was, said Binns-Ward J, to be taken only as a last resort or where absolutely necessary (at [49]).

Even after taking a judicial peek, the court held that most (but not all) of the redacted portions of the feenotes were not privileged.

s 225: Prosecution’s failure to obtain DNA evidence and to procure evidence of examining doctor in rape case

The Supreme Court of Appeal in S v Mugwedi [2014] ZASCA 23 (unreported, SCA case no 694/13, 27 March 2014) was highly critical of two omissions by the prosecution in its presentation of a rape case. First, there was no attempt to obtain DNA sampling for analysis. In spite of the same court’s emphatic statement in S v Carolus 2008 (2) SACR 207 (SCA) at [32] that it was imperative in sexual assault cases, especially those involving children, that DNA tests be conducted, the necessary procedures had not been followed. Saldulker JA stressed that the relevant kits had to be made available for this to happen and he had ‘difficulty in understanding why repeated judicial pronouncements [were] not acted upon by the relevant authorities’ (at [2]) (see, too, S v Nedzamba 2013 (2) SACR 333 (SCA) at [35]).

Second, the doctor who had examined the two young girls after the incident was not called to testify, and no effort was made by the prosecution to secure his attendance. There was not even an attempt to present his findings on affidavit in terms of s 212(4)(a) of the Criminal Procedure Act. Instead, the state called a second doctor who said he knew the first doctor but did not know where he was that day. He proceeded to read out the contents of both J88 forms in court and opined that there was vaginal penetration of both girls. This evidence was, as the court held, clearly inadmissible, based as it was on the hearsay evidence of the first doctor.

It is hoped that the National Prosecuting Authority takes note of the growing impatience of the courts in such cases and takes steps to introduce procedures and practices that would prevent what might otherwise lead to such serious injustice as to do irreparable harm to the repute of the system of criminal justice as a whole.

Another matter that fell to be considered by the court in Mugwedi was the status of identification evidence by one of the complainants (aged 7 at the time). It was held that the evidence could not be relied upon since it was not a spontaneous identification but had clearly been prompted by an adult. There were, moreover, contradictory versions as to how the identification came about.

s 252A: Traps

In S v Wana & others (unreported, ECP case no CC 16/2013, 19 & 20 March 2014) it was held that the information supplied by the police trap and his contribution to the planning of the robbery went no further than creating the opportunity for the commission of the offence in terms of s 252A. The essential terms of the conspiracy to commit the offence had been agreed at a very early stage of the meetings between the trap and the perpetrators of the robbery. What followed that agreement was a process of planning to facilitate the carrying out of the enterprise (see [110]).


(c) Sentencing


s 50(2)(a) of the Criminal Law (Sexual Offences and Related Matters) Amendment Act 32 of 2007: Children and the constitutional invalidity of the National Register for Sex Offenders

S v IJ 2013 (2) SACR 599 (WCC) and J v National Director of Public Prosecutions & others [2014] ZACC 13 (unreported, CC case no CCT 114/13, 6 May 2014)

Section 50(2)(a) of Act 32 of 2007 provides that a court must make an order that the particulars of a person convicted of a sexual offence against a child or mentally disabled person must be included in the register referred to above. For a detailed discussion of all statutory measures relating to this register, see Le Roux and Williams in Smythe & Pithey (eds) Sexual Offences Commentary: Act 32 of 2007 at 17-1 to 17-44.

In S v IJ 2013 (2) SACR 599 (WCC) Henney J, writing for a full bench, declared that s 50(2) of Act 32 of 2007 was invalid and inconsistent with the Constitution in that it failed to allow the court to inquire and decide, after having afforded the accused an opportunity to make representations, whether or not the accused’s particulars should be included in the National Register for Sex Offenders (at 600h–i). It was ordered that the declaration would not have retrospective effect. Its effect was also suspended for 18 months to provide sufficient time for Parliament to amend s 50(2) so that it could be ‘constitutionally compliant’ (at 600j). The order was referred to the Constitutional Court (at 601e). At 601a–e the full bench also made an order as regards the procedure and principles that had to be followed during the period of suspension.

The Constitutional Court dealt with the matter on 6 May 2014. See J v National Director of Public Prosecutions & others [2014] ZACC 13 (unreported, CC case no CCT 114/13, 6 May 2014). It should be noted that the convicted person concerned in this matter was himself a child at the time of the commission of the offences: he was 14 years old when he raped a seven-year-old boy and two six-year-old boys in contravention of s 3 of Act 32 of 2007. The Constitutional Court accordingly held that the constitutionality of s 50(2) in relation to adult offenders was not before the court.

Skweyiya ADCJ, writing for a unanimous court, identified s 28(2) of the Constitution as the essential point of departure: ‘A child’s best interests are of paramount importance in every matter concerning the child’. At [36] it was stated that on a previous occasion in S v M (Centre for Child Law as Amicus Curiae) 2008 (3) SA 232 (CC) at [19], the Constitutional Court had emphasised the developmental impetus of the best-interests principle in securing the right of children to ‘learn as they grow how they should conduct themselves and make choices in the wide and moral world of adulthood’. At least three key principles, held Skweyiya ADCJ, stem from this approach to the best interest of the child offender:

This infringes the best interests of the child. The opportunity for an individuated response to the particular child offender, taking into account the child’s representations and views, is excluded both at the point of registration and in the absence of an opportunity for review. The limited circumstances in which an offender can apply for his or her removal from the Register are insufficiently flexible to consider the particular child’s development or reform.

The serious consequences for an offender whose particulars are recorded in the register were also identified and considered by the court (at [20] and [43]–[44]). In the case of a child offender, ‘the consequences of registration will, for the most part, only be felt as an adult’ (at [43]).

At [44]–[45] Skweyiya ADCJ concluded: ‘Given that a child’s moral landscape is still capable of being shaped, the compulsory registration of the child sex offender is an infringement of the best-interests principle . . . [T]he provision [s 50(2)(a)] limits the child offender’s right in terms of section 28(2) of the Constitution’.

The court was also satisfied that the offending provision was not a constitutionally permissible limitation as provided for in s 36 of the Constitution: the limitation will not always achieve its purpose (at [49]) and there are ‘less restrictive means’ to achieve the aims of the register (at [50]).

The following observations give a clear indication to Parliament as to what is required to remedy the defects (at [50]):

Affording courts a discretion and the concomitant opportunity to the child offender to lead evidence and make argument on the question of registration would permit the possibility of greater congruence between the limitation and its purpose. Where a court decides on matters affecting children, discretion plays an important role in allowing for an individuated response to meet the child’s best interests. Modifications to registration parameters (such as when registration is triggered and how it is terminated) may also permit for more individualised concerns to be taken into account in a consistent fashion.

Having set aside the order of the Western Cape High Court, the court suspended its own declaration of invalidity of s 50(2)(a) for 15 months in order to afford Parliament an opportunity to correct the defect in the light of this judgment. An order that the respondents should also furnish certain information to the court was also made, so that the court, in turn, could make such information available to persons or organisations seeking to assist those child offenders on the existing register (at [57]).

Sentencing and consideration of the period in detention prior to sentencing

S v Radebe 2013 (2) SACR 165 (SCA) and DPP v Gcwala [2014] ZASCA 44 (unreported, SCA case no 295/13, 31 March 2014)

In both cases referred to above it was held that the period spent in prison awaiting trial cannot on its own constitute ‘substantial and compelling circumstances’ justifying a departure from the minimum sentence prescribed by the Criminal Law Amendment Act 107 of 1997.

At [14] in Radebe Lewis JA said that ‘the period in detention pre-sentencing is but one of the factors that should be taken into account in determining whether the effective period of imprisonment to be imposed is justified: whether it is proportionate to the crime committed’. The judgment in Gcwala was also written by Lewis JA.

In both decisions the so-called ‘mechanical rule of thumb’ was rejected. In terms of this rule – which can be traced to cases such as S v Brophy & another 2007 (2) SACR 56 (W) and S v Stephen & another 1994 (2) SACR 163 (W) – a court could, in determining an appropriate sentence, take into account the period spent as awaiting trial prisoner, double it and then deduct this double period from the period of imprisonment that would have been imposed but for the awaiting trial period. The demise of this rule of thumb must be noted but not mourned. It stood on flimsy grounds and was often criticised in other high court decisions. See S v Vilikazi & others 2000 (1) SACR 140 (W); S v Seboko 2009 (2) SACR 573 (NCK) at [21]–[22]. In S v Mahlangu & others 2012 (2) SACR 373 (GSJ) Satchwell J called it ‘an arithmetical calculation’ (at 376c). The ‘mechanical formula’ is ‘unhelpful’ (Radebe (supra) at [13]).

It is submitted that as far as an awaiting trial period in detention is concerned, a court should exercise its sentencing discretion on the basis of the following broad but immutable principles:

(a)     The relevant period must be considered, with all other factors, on the basis that being a pre-trial incarcerated person is a great hardship (S v Mahlungu & others (supra) at 376ab).

(b)     There is no room for a mechanical rule of thumb (Radebe (supra); Gcwala (supra)).

(c)     It is wrong to ignore the relevant period on the basis that the accused could, by a plea of guilty, have avoided it (S v Bhadu 2011 (1) SACR 487 (ECG) at [5]).

(d)     Where there was an ‘inordinate time spent awaiting trial . . . it would be appropriate to factor in that period in mitigation of the cumulative effect of the sentences’ (per Shongwe JA in S v Kruger 2012 (1) SACR 369 (SCA) at [12]).

See further the discussion of the antedating of prison sentences in Commentary on the Criminal Procedure Act, s 282, sv General.

Table of Cases

Click here to see the full Table of Cases of the Criminal Justice Review No 1 of 2014  



Click here for a printable .pdf version of this Review



Principles of Criminal Law

Sexual Offences - Criminal Law (Sexual Offences and Related Matters) Amendment Act 32 of 2007 & Regulations; Sexual Offences Act 23 of 1957
Criminal Law Casebook / Strafregsakebundel
Criminal Procedure Act 51 of 1977 & Regulations
Domestic Violence Act 116 of 1998 & Regulations
Probation Services Act 116 of 1991 & Regulations
Bail: A Practitioner’s Guide
Guide to Bail Applications, A
Criminal Procedure Workbook
Commentary on the Criminal Procedure Act (Du Toit)
Sexual Offences Commentary






> Click here to opt-in to continue to receive these newsletters.

> Click here to unsubscribe from this communication


> Jutastat e-publications user helpdesk

> Juta Customer Services

> Juta Law Marketing


Jutastat electronic subscribers and print subscribers may receive important service-related information relating to their Juta publications from time to time.

Visit the Juta Law Website

2019 © Juta and Company (Pty) Ltd