A bi-annual update complementing the

Commentary on the Criminal Procedure Act


NO 1 OF 2017


Andrew Paizes, Author (Editor)

Steph van der Merwe (Author)


For a printable .pdf version of this Review click here.



There are three feature articles in this edition. The first examines critically the so-called ‘cardinal rules of logic’ set out in the locus classicus in the area of circumstantial evidence, R v Blom 1939 AD 188, which were, once again, relied upon by the Supreme Court of Appeal in a recent decision. The question is asked whether these rules, in spite of their firm grasp on the judicial imagination, are now more a hindrance than a help, and whether the time has now come to abandon them in favour of a more sophisticated and nuanced approach.

The second considers two recent cases which examine the legal duty of the executive to address matters relating to parole in a manner that is consistent with legal principles and requirements: the first case concerns the duty of the executive to ensure that its decision to release a sentenced offender on parole is based on adequate information; the second the duty to follow fair administrative action in any decision by the executive to refuse parole.

The third feature article concerns the meaning of ‘hearsay evidence’ in the Law of Evidence Amendment Act 45 of 1988 and challenges the view, expressed in a recent case, that this definition is much the same as that which applied at common law.

The section on legislation describes various statutory provisions but the focus is mostly on the Criminal Procedure Amendment Bill, which seeks to make important amendments to Chapter 13 of the Criminal Procedure Act. This chapter deals with matters relating to an accused’s mental illness, criminal responsibility and the capacity of an accused to understand proceedings. The discussion includes a consideration of the extent to which the legislation successfully addresses constitutional objections raised by the courts in respect of the existing provisions. The section covers legislation up to 12 June 2017.

The case notes, as usual, cover a wide range of issues. Two very interesting questions relating to the substantive criminal law are considered. First, is a person who assists another to commit suicide guilty of murder, and is there a difference in this context between ‘physician administered euthanasia’ and ‘physician assisted suicide’? Second, are the statutory offences prohibiting the use of cannabis in small quantities for purely personal use valid in the light of the constitutionally protected right to privacy?

Among the procedural issues canvassed, the following may be mentioned:

  • When is residential property an ‘instrumentality of an offence’ for the purpose of a forfeiture order in terms of the Prevention of Organised Crime Act (POCA)?
  • When can a statutory juristic body institute a private prosecution?
  • In what circumstances will bail application proceedings be rendered unfair by remarks made by a judicial officer that compromise his or her impartiality?
  • Can an accused waive his or her rights to participate in the process of reconstructing a lost record of the trial proceedings?
  • What practical guidelines should a court follow in avoiding the pitfalls of (a) identification evidence; and (b) evidence of a confession?
  • Does the failure to warn an accused prior to conviction that he or she faces a possible sentence of life imprisonment render the trial unfair?
  • What considerations should a court take into account in ensuring trial fairness when it considers imposing a non-parole period as part of the sentence?
  • When should a court exercise its discretion to order a concurrent running of sentences in order to avoid inhumane or unfair sentences?
  • What happens if a court of appeal fails in its duty to give an appellant notice of a possible increase of sentence?
  • When may a court of appeal dispose of an appeal without the hearing of oral argument?

Andrew Paizes





Circumstantial evidence and inferential reasoning: Is it time to jettison the rules in R v Blom?

Perspectives on parole: Two recent cases

What is the meaning of ‘hearsay evidence’ in s 3(4) of the Law of Evidence Amendment Act 45 of 1988?


The Justice Administered Fund Act 2 of 2017

Implementation of the Rome Statute of the International Criminal Court Act Repeal Bill [B23—2016]: Withdrawal by Minister

Children’s Amendment Act 17 of 2016

Judicial Matters Amendment Bill [B14—2016]

Criminal Procedure Amendment Bill [B2—2017]

Proposed amendments to s 77(6)(a)(i)    
Proposed amendments to s 77(6)(a)(ii)    
Proposed amendment to s 78(6)    
Proposed amendments to s 79    
Sections 77, 78 and 79: ‘mental defect’ to be replaced by ‘intellectual disability'  


(a) Criminal Law

Is a person who assists another to commit suicide guilty of murder? Physician administered euthanasia and physician assisted suicide considered

Constitutional validity of statutory offences prohibiting use of cannabis for personal consumption and related matters

(b) Criminal Procedure and Evidence

  (i) Pre-sentence

ss 7 and 8: Private prosecution and s 6(2)(e) of the Societies for the Prevention of Cruelty to Animals Act 169 of 1993   

60: Bail: Right to a fair bail hearing

ss 151 and 166: Onus of proof—implications of. Judicial officer descending into the arena—how to deal with inexperienced prosecution   

s 152: Reconstruction of the record: Can the accused waive his or her rights to participate in the process?   

s 208: Identification evidence guidelines and pitfalls

ss 217, 219, 166: Admissibility of confessions under the Criminal Procedure Act and the Constitution   

       (ii) Sentencing

Sentencing and trial fairness: No reference in charge-sheet to minimum sentence legislation

s 276B: Procedural fairness where court considers a non-parole period   

s 280(2): Sentencing court’s discretion to order concurrent running of sentences   

       (iii) Forfeiture and confiscation

s 20: Sections 48(1) and 50(1) of the Prevention of Organised Crime Act 121 of 1998: When is residential property an ‘instrumentality of an offence’ for the purpose of a forfeiture order? 

(iv) Appeal and Review

Appeal: Requirement that court of appeal gives notice of possible increase of sentence   

Appeal: Disposal of an appeal in terms of s 19(a) of the Superior Courts Act 10 of 2013 


Circumstantial evidence and inferential reasoning: Is it time to jettison the rules in R v Blom?

The recent decision of the Supreme Court of Appeal in S v Mahlalela [2016] ZASCA 181 (unreported, SCA case no 396/16, 28 November 2016) raises some important and interesting issues concerning the appropriate mode of reasoning when inferences are sought to be drawn from primary or intermediate facts. The locus classicus in this regard is the much-cited decision in R v Blom 1939 AD 188, in which Watermeyer JA set out what he described (at 202–203) as ‘two cardinal rules of logic’ which could not be ignored when it came to reasoning by inference:

  1. The inference sought to be drawn must be consistent with all the proved facts. If it is not, then the inference cannot be drawn.
  2. The proved facts should be such that they exclude every reasonable inference from them save the one sought to be drawn. If they do not exclude other reasonable inferences, then there must be a doubt whether the inference sought to be drawn is correct.’

These cardinal rules have become so entrenched in our law that there is hardly a case on circumstantial evidence that does not cite them or purport to rely on them for guidance. They have achieved a status not far short of canonical imprimatur, to the extent that one hesitates to criticise them for fear of being accused of heresy.

But it cannot be denied that the rules in Blom are far from perfect or, even, satisfactory. They beg more questions than they answer; they are unclear in important respects; they cannot be rules of ‘logic’ in any meaningful sense; they seem, in so far as they are rules of ‘law’, to find application (at least the second rule) in the criminal law alone; and yet they seem to be imperfect servants of the rules governing the onus of proof. They are, in my view, misleading and dangerous and they lead to wrong answers to complex questions.

Such sharp criticism will surely be viewed by some as heretical or, at least, misconceived. But I remain unrepentant. I have subjected the rules in Blom to careful scrutiny in The South African Law of Evidence (Zeffertt and Paizes, 2 ed (2009) at 100–13), and have levelled a number of accusations at them.

At the very least, it must be accepted that the rules raise a number of questions that are very difficult to answer with any precision or certainty and which give the lie to the notion that they at least provide a simple or helpful direction for untying an intricate knot.

These are some of those questions: (1) What constitutes a ‘proved fact’? (2) What standard of proof is to be deployed in determining whether a ‘fact’ has indeed been ‘proved’? (3) Is there only one such standard, to be used as a universal solvent, or are there different standards to be used in different situations? (4) Does the standard—if, in fact, it does vary from one situation to another—depend on whether the proceedings are criminal or civil? (5) Does it depend, further, on whether the ‘fact’ relied upon as a ‘proved fact’ is a ‘primary’ or an ‘intermediate’ fact? (6) Does the consideration that the ‘fact’ rests on inferential reasoning drawn from a large number of prior (primary or intermediate) facts, as opposed to merely one or two, signify at all? (7) Is the second rule in Blom a restatement or an expression of the onus of proof in criminal proceedings (ie proof beyond a reasonable doubt)? (8) If so, how can the rule be one of logic and not of law? (9) And, further, how can it be made to apply to civil proceedings? (10) And, further still, how does its application in criminal cases vary, if at all, in respect of issues directly related to guilt (such as the essential elements of the offence and other closely related issues) and those much less directly related?

A consideration of all these questions leaves the rules very vulnerable to attack. In The South African Law of Evidence I have not hesitated to shine a light on their frailties and shortcomings in an attempt to weaken their grip on the imagination of our judges and commentators. But their ubiquity in judicial pronouncements, age, succinct formulation, apparent (albeit deceptive) simplicity and kinship with the criminal onus of proof, have given them an aura that will be very difficult to dispel. Who will be bold enough to shake the firmament or, even, give it a gentle prod? What messy, complex, tangled web of inter-related ideas, notions and questions will be unearthed if we do so? Can we bear to engage with the complexity that will emerge if we expose the rules for what they really are: an anodyne that provides some comfort but fails to cure or treat the real disorder?

I will not repeat all the criticisms I have raised elsewhere but will limit myself to one crucial aspect of those criticisms, which has been given fresh impetus by what has been said in a recent decision of the Supreme Court of Appeal.

In S v Mahlalela [2016] ZASCA 181 (unreported, SCA case no 396/16, 28 November 2016), the appellant had been convicted of murder and robbery with aggravating circumstances. The case against him (and his two co-accused) had rested almost entirely on circumstantial evidence. A vital item of circumstantial evidence, received by the trial court and relied upon by it in convicting the appellant, was that the appellant and one of the co-accused had, shortly after the robbery, each made deposits of R2000 into their respective bank accounts on the same day, at the same bank, and with the same teller. An amount of approximately R8000 had been taken in the robbery. The appellant did not testify at the close of the State’s case.

Dlodlo AJA (with whom Shongwe, Van der Merwe and Mocumie JJA and Potterill AJA agreed) was ‘persuaded that the State did not prove beyond reasonable doubt that there was a factual basis entitling the trial court to draw inferences either intermediate or otherwise’ (at [19]). The discovery of the bank slips evidencing the two deposits did ‘not go far enough to assist the State’ since the evidence ‘hardly satisfie[d] the first rule of logic in Blom’: there was ‘a host of reasonable inferences to be drawn from the discovery of these deposit slips’, and it did not follow, said Dlodlo AJA, that the deposited money belonged to the deceased who had been robbed of R8000 in cash.

If it were to be ‘accepted that these deposit slips constitute[d] a fact from which an inference of guilt [could] be drawn, that would’, said Dlodlo AJA, ‘stretch inferential reasoning too far’ (at [20]; emphasis added). The ‘proved facts’ envisaged in Blom were, according to Dlodlo AJA, ‘facts proved beyond reasonable doubt’, and ‘intermediate inferences’ too, had to be based on such ‘proved facts’ (at [15]).

Moreover, said the judge (at [20]), for an inference to be permissible, it ‘not only had to be based on proved facts, but also had to be the only reasonable inference from those facts to the exclusion of all other reasonable inferences’. The ‘ultimate question [was] whether, in the light of all the evidence adduced at trial, the guilt of the appellant was established beyond reasonable doubt’.

This judgment touches on several of the concerns I have expressed in The South African Law of Evidence and raises fresh anxieties as well. One of the concerns expressed above related to the uncertain meaning of a ‘proved fact’ and the standard of proof required to render a fact a ‘proved fact’ for the purpose of the rules in Blom.

The court in Mahlalela, however, seemed to have no uncertainty at all. It spoke unequivocally about such facts having to be proved beyond a reasonable doubt, and made it quite clear that intermediate inferences had to be based on facts so proved. It is submitted, however, that such certainty was entirely misplaced and that an insistence on ‘facts’ being proved beyond a reasonable doubt in all cases before being able to sustain legitimate inferences, is wrong in principle and in law. In R v Mtembu 1950 (1) SA 670 (A) at 679–80, Schreiner JA emphatically opposed an approach that would require all intermediate facts in a criminal trial to be proved beyond a reasonable doubt. He set out a broad test for determining when proof on this standard would be necessary, but made it plain that it would not always be required. He said:

'I am not satisfied that a trier of fact is obliged to isolate each piece of evidence in a criminal case and test it by the test of reasonable doubt. If the conclusion of guilt can only be reached if certain evidence is accepted or if certain evidence is rejected then a verdict of guilty means that such evidence must have been accepted or rejected, as the case may be, beyond reasonable doubt. Otherwise the verdict could not properly be arrived at. But that does not necessarily mean that every factor bearing on the question of guilt must be treated as if it were a separate issue to which the test of reasonable doubt must be distinctly applied. I am not satisfied that the possibilities as to the existence of facts from which inferences may be drawn are not fit material for consideration in a criminal case on the general issue whether guilt has been established beyond reasonable doubt, even though, if the existence of each such fact were to be treated by the test of reasonable doubt, mere probabilities in the Crown’s favour would have to be excluded from consideration and mere probabilities in favour of the accused would have to be assumed to be certainties. Circumstantial evidence, of course, rests ultimately on direct evidence and there must be a foundation of proved or probable fact from which to work. But the border-line between proof and probability is largely a matter of degree, as is the line between proof by a balance of probabilities and proof beyond reasonable doubt. Just as a number of lines of inference, none of them in itself decisive, may in their total effect lead to a moral certainty (Rex v de Villiers (1944, A.D. 493 at p. 508)) so, it may fairly be reasoned, a number of probabilities as to the existence of the facts from which inferences are to be drawn may suffice, provided in the result there is no reasonable doubt as to the accused’s guilt.’

This dictum has been referred to and applied in a number of cases (see, for instance, R v Manda 1951 (3) SA 158 (A) at 166B–D, R v Sibanda & others 1965 (4) SA 241 (SRA) at 246–7, S v Shepard & others 1967 (4) SA 170 (W) at 172–3 and S v Smith en andere 1978 (3) SA 749 (A) at 754–5; see, too, S v Ntsele 1998 (2) SACR 178 (SCA) at 189c–d and S v Morgan & others 1993 (2) SACR 134 (A) at 172–3). However, it is, perhaps, not as well known, understood or even appreciated as it should be. That it is not is probably due to the unfortunate ‘lines of interference’ emanating from the rules in Blom, which regrettably entangle the rules for reasoning by inference with those governing the criminal onus. That there are certainly instances where inferences may properly be drawn from ‘facts’ that are probable—albeit not proved at the level excluding reasonable doubt—is demonstrated by the decision in R v Manda (supra), where the appellant had been convicted of murder and evidence had been adduced at his trial that (a) the appellant had received scratches on his face at approximately the time that the deceased, a nine-year-old boy, had been killed; and (b) material had been scraped from under the fingernails of the deceased which was probably human skin. The latter probability, said Schreiner JA, was something that could be used in arriving at a conclusion.

This perspective is one that the High Court of Australia reached only after a false start. In Chamberlain v The Queen [No 2] (1984) 153 CLR 521, that court, in the famous ‘dingo trial’, insisted (as Dlodlo AJA seems now to have) that all primary facts in a criminal case had to be proved beyond a reasonable doubt, since ‘[n]o greater cogency can be attributed to an inference based upon particular facts than the cogency that can be attributed to each of those facts’ (at 599; emphasis added). In a strong dissent, however, Deane J took a very different view, one that squares with the pronouncement of Schreiner JA in Mtembu. In considering what standard should be imposed for proving primary facts, he said (at 626–7):

'In the view I take, it is impossible to give a general theoretical answer to that question. There is certainly no requirement of the law that the members of a jury must examine separately each item of evidence adduced by the prosecution and reject it unless they are satisfied beyond reasonable doubt that it is correct. Nor is it the law that a jury is in all circumstances precluded from drawing an inference from a primary fact unless that fact is proved beyond reasonable doubt. If a primary fact constitutes an essential element of the crime charged, a juror must be persuaded that that fact has been proved beyond reasonable doubt before he or she can properly join in a verdict of guilty. Whether or not a juror must be satisfied that a particular fact has been proved beyond reasonable doubt will, however, otherwise depend not only on the nature of the fact but on the process by which an individual juryman sees fit to reach his conclusion on the ultimate question of guilt or innocence. If, for example, the case against an accused is contingent upon each of four matters being proved against him, it is obvious that each of those matters must be proved beyond reasonable doubt. Indeed, it would be appropriate for the presiding judge to emphasise to the jury in such a case that even a minimal doubt about the existence of each of these matters would be greatly magnified in the combination of all. On the other hand, if the guilt of an accused would be established by, or a particular inference against an accused could be drawn from, the existence of any one of two hundred different matters, each of which had been proved on a balance of probabilities, it would be absurd to require that a jury should disregard each of them unless satisfied, either in isolation or in the context of all of the facts, that any particular one of those matters had been proved beyond reasonable doubt.’

Deane J’s exposure of the flaws in the majority’s reasoning prompted a quick retraction and about-face. In Shepherd v The Queen (1990) 170 CLR 573 the High Court did what appellate courts often do when they know they have been wrong: maintain that what they were understood to have said was not what they actually said. The majority in Chamberlain, said the court in Shepherd, intended their remarks to apply only to those intermediate facts which were ‘a necessary basis for the ultimate inference’ (at 581). The court distinguished between situations where intermediate facts constituted ‘indispensable links in a chain of reasoning toward an inference of guilt’, where proof of those facts would have to be on the criminal standard and would have to exclude reasonable doubt, and situations where the evidence could better be described as consisting of ‘strands in a cable rather than links in a chain’ (at 579), where that standard would not apply.

The crucial question, then, in cases of inferential reasoning, is this: Which form of reasoning better describes the process one is dealing with—‘chain’ reasoning or ‘cable’ reasoning? In answering this question, these incidental questions may assist: How important is the inferred ‘fact’ in the broader scheme of things and, in particular, in establishing the ultimate issue of the accused’s guilt? How many facts are relied on to establish the inferred ‘fact’? The more facts that are relied upon, the less likely it is that each or any primary fact is important enough to be decisive of the issue by itself, and the more likely it is that guilt itself (or an element of it or an intermediate fact itself crucial to establishing guilt or such element) may be proved beyond a reasonable doubt even though a particular primary fact or, even, each of them, is shown to be no more than probable.

Consider a murder trial where the court has taken the view that it has been shown beyond a reasonable doubt that the accused had a motive for the murder and that the murder weapon belonged to him, but that these facts still left the matter of his guilt short of being proved beyond a reasonable doubt. Assume, further, that the court accepted it as no more than ‘probable’ that the accused had been seen near the scene of the crime shortly before the time of the murder and had been inexplicably absent from his place of work during the crucial time period. Would not the inference that the accused had, impelled by his motive, used his weapon to kill the deceased, not be significantly strengthened by this evidence—perhaps to the point of proving his guilt beyond a reasonable doubt? McHugh J in Shepherd believed, correctly in my view, that it would.

The lesson to be learned is that it is surely naïve to believe that the facile rules expressed in Blom’s case can effectively and comprehensively engage with all the complexities and nuances presented by cases of this kind. In fact, far from helping, they get in the way of clear thinking and may obstruct in answering the most important question of all—whether the accused’s guilt has indeed been proved beyond a reasonable doubt.

To ensure that the correct answer is achieved in respect of what Dlodlo AJA correctly described as the ‘ultimate question’, there is, in my view, only one method that will work in all cases: to begin with the ultimate question, work backwards, and ensure that the mode of reasoning by inference is, in each instance and at every level of the inferential process, sufficiently sound to secure proof of the ultimate issue (guilt) beyond a reasonable doubt.

Consider this example: X is charged with the murder of his wife, Y, and a man, Z. The State’s case is that X killed both of them out of anger and jealousy because Y and Z were having an affair. No other motive is put forward and the State’s case rests heavily upon the intermediate fact that Y and Z were, in fact, having an affair. X’s guilt, in such a case, cannot be established beyond a reasonable doubt if this intermediate fact is not also proved beyond a reasonable doubt. That fact, then, may be likened to a link in the probative chain.

Assume, further, that the State, in order to prove the existence of this affair, calls as a witness the receptionist at a hotel who testifies that, a year before the trial, she checked in at the hotel for a night a couple who, to the best of her recollection, were Y and Z, even though they checked in under false names. Cross-examination reveals weaknesses in the identification, and the court is prepared to accept no more than that it is probably accurate. If, then, it is reasonably possible that it is not, can the inference that Y and Z were having an affair be drawn? If this is the only evidence to support that inference, it cannot be drawn, as the existence of the affair cannot be established beyond a reasonable doubt if there is reasonable doubt attaching to the only evidence tendered to prove it. However, the position would be quite different if the State could call several different receptionists, working at various hotels, who could testify to the same facts. Even if the accuracy of each of these accounts was accepted as being no more than probable, it could legitimately sustain an inference that Y and Z were having an affair. If so, the intermediate fact (the affair) would be proved beyond a reasonable doubt and the probative edifice would stand. The evidence of each of these witnesses could appropriately be likened to a strand in the cable, which, given the support of a sufficient number of other similar strands, would be strong enough to secure the necessary evidentiary connection to allow, at the end of the process, for proof of guilt beyond a reasonable doubt.

It is unprofitable, in such cases, to work ‘forward’ from primary to intermediate to ultimate ‘facts’. That is one of the biggest flaws in the rules in Blom. And that is why the rules seek—in an unscientific and muddled fashion—to tie the rules of inferential thinking with the standard that encapsulates the criminal onus. If we, instead, work ‘backwards’ from that standard, it has these benefits: First, it guarantees that we can never fail to meet that standard in criminal cases. Secondly, we are able to develop a far more nuanced approach to inferential reasoning that depends on the amount of probative work each intermediate fact is required to do as well as the mode of reasoning (‘chain’ or ‘cable’) we are employing to connect it with the fact (ultimate or other) we are seeking to establish.

The rules in Blom provide little guidance for charting the murky waters of inferential reasoning. Worse, still, they have led to or allowed for many misconceptions. Some of these are: (i) All primary or intermediate facts have to be proved beyond a reasonable doubt (Chamberlain and Mahlalela). (ii) Inferences may not be drawn from other inferences (Mahlalela at [15]). There is no reason why a whole series of cascading inferences cannot be upheld provided the ultimate issue (guilt) is proved beyond a reasonable doubt. The model described ensures that it cannot fail to be, and care will be given to ensure that the mode of reasoning (cable or chain) does not allow for a violation of what is a constitutional entitlement—the right to be presumed innocent until proven guilty. (iii) A court is ‘entitled to draw inferences but not to speculate’ (see, for instance, S v Essack & another 1974 (1) SA 1 (A) at 16D, S v Cooper 1976 (2) SA 875 (T) and S v Nkuna 2013 (2) SACR 541 (GNP) at [11]). Why should a court not be allowed to come up with an innocent explanation, one that was, for some reason, not furnished by the accused (he or she may wish to protect another, or may wrongly believe the true explanation to be unwise because it is incriminatory in respect of a lesser offence) if that ‘speculative’ explanation has a reasonably possible chance of being true? To disallow such ‘speculation’ would not only violate the accused’s constitutional right (the presumption of innocence) but would disregard the authority of the majority judgment in R v Mlambo 1957 (4) SA 727 (A). It is a strange irony that that case is usually cited more for the attack on ‘speculation’ in the minority judgment of Malan JA than for the proposition, correctly accepted by the majority, that the appellant’s conviction for murder could not stand because of the reasonable possibility—not raised by the appellant—that he had unintentionally killed the deceased in the course of a sexual assault (see The South African Law of Evidence at 121–6). Labels, in my view, are at best unhelpful and at worst dangerous in this context. If there is a reasonable doubt that the accused is innocent, the reason for that doubt—whether it be speculation, conjecture or ‘solid evidence’—should not signify. The mere fact that it exists, does. This view is strengthened by the recent decision of the Supreme Court of Canada in R v Villaroman 2016 SCC 33 at [35].

I do not believe that the court in Mahlalela reached the wrong result. The evidence in that case was clearly insufficient to sustain the convictions. But the process of reaching that result was not facilitated by the rules in Blom. Instead, those rules muddied the waters even further by leading to unnecessary complications regarding the meaning of ‘proved facts’, the relation between the rules and the criminal onus, and a consideration of whether inference could follow inference.

It is time to see the rules in Blom for what they were probably intended to be: an early, tentative step on a challenging journey. We have been on this journey for long enough now to see more of the terrain and to understand more of its dangers. It is time to let them go.

Andrew Paizes

Perspectives on parole: Two recent cases

Moses Parole in South Africa (2012) 10 describes parole as ‘a form of conditional release of prisoners into the community, before the expiry of their sentences, subject to, and on the basis of, an undertaking by the parolee that he will...comply with such conditions as...imposed and prescribed...’.

Our courts have over many years consistently maintained that, in line with the principle of separation of powers, the decision to grant parole to a sentenced offender is ‘the exclusive field’ of the executive. See S v Jimmale 2016 (2) SACR 691 (CC) at [11]; S v Mhlakaza & another 1997 (1) SACR 515 (SCA) at 521f–i. In S v Botha 2006 (2) SACR 110 (SCA) Ponnan JA said that courts do not have the power to prescribe to the executive branch of government how long a sentenced offender should be in jail, ‘thereby usurping the function of the executive’ (at [25]). A court’s limited power to determine a non-parole period as provided for in s 276B of the Criminal Procedure Act 51 of 1977 is an exception. This section came into operation only on 1 October 2004. See further the discussion of s 276B in Commentary, sv The judiciary and the executive, as well as the brief discussion of S v Ndlovu [2017] ZASCA 26 (unreported, SCA case no 925/2016, 27 March 2017) in Section C of this issue of CJR, sv ‘s 276B: Procedural fairness where court considers a non-parole period’. In all cases falling outside the ambit of s 276B, the executive enjoys the prerogative of granting parole. See generally S v Van Loggenberg 2012 (1) SACR 462 (GSJ) at [13].

It is sometimes argued that parole and the system of parole lead to unacceptable interference with a judicial decision, that is, the sentence imposed by the court. See generally Van Heerden 2011 24(2) Acta Criminologica: Southern African Journal of Criminology 17 at 26–7. However, the fact of the matter is that parole serves as an incentive to prisoners to rehabilitate themselves so that they can be released without having served their full prison sentence. A court is required to sentence an accused on the basis of all the relevant facts and without regard to the possibility of parole. See S v Matlala 2003 (1) SACR 80 (SCA) at [7]; S v Motloung 2016 (2) SACR 243 (SCA) at [18]. The executive is better placed than the courts to determine parole because release on parole ‘requires the assessment of facts relevant to the...[sentenced offenders’]...conduct after the imposition of sentence’ (S v Madolwana (unreported, ECG case no CA&R 436/12, 19 June 2013) at [7], emphasis in the original).

The two cases discussed below provide insightful perspectives on parole and the executive’s legal duty to address parole matters in a manner consistent with legal principles and requirements. The first case deals with the duty of the executive to ensure that its decision to release a sentenced offender on parole is based on adequate information. The second case deals with a situation where the executive’s refusal of parole was set aside because of unfair administrative action.

Naidu v Minister of Correctional Services [2017] 2 All SA 651 (WCC)

In the above matter the plaintiff claimed damages arising from an attack on her by one MM who was on parole at the time of the attack. The attack, she claimed, was a direct result of the negligent release of MM on parole—a matter for which the defendant, the Minister of Correctional Services, was liable. At the trial, counsel for the defendant made some concessions: first, the plaintiff was attacked by MM as alleged in the particulars of claim; second, if it were to be found that the defendant was negligent in releasing MM on parole, such negligence was causally linked to the harm suffered by the plaintiff; and, third, there was a legal duty on the defendant to ensure the safety of members of the public, such as the plaintiff, and this duty was sufficient to found liability. The defendant’s legal duty, noted Meer J at [4], was akin to that relied upon in Minister of Safety and Security & another v Carmichele 2004 (3) SA 305 (SCA), a case which concerned liability for the negligent failure to oppose bail for a dangerous offender. At [47] Meer J also made brief reference to s 131 of the Correctional Services Act 111 of 1998 (hereafter the ‘CS Act’). This section refers to patrimonial loss and provides as follows: ‘In the event of a person serving community corrections being liable in delict for an act or omission in the course of such service, the damages sustained may be recovered from the State.’

Given the concessions as outlined above and the fact that the parties had agreed that the merits and quantum be separated, the only question was whether the defendant—acting through the Brandvlei Prison Correctional Supervision and Parole Board (hereafter ‘the Board’)—was negligent in releasing MM on parole. According to the plaintiff the Board was negligent in that it had failed to act with reasonable care and diligence in determining whether MM could be released on parole: inadequate attention was given to the previous convictions of MM and the fact that he had previously violated his parole conditions (at [3]). The plaintiff also asserted that in its assessment of MM as a candidate for parole, the Board had relied on inadequate information and had decided the matter in the absence of certain reports which the Case Management Committee (hereafter ‘the CMC’) should in terms of the CS Act have submitted to the Board. One of the core functions of a CMC is to report to the relevant Board on the possibility of placing a sentenced offender on parole.

But in Naidu it turned out that the CMC had failed to perform its statutory duty to place certain crucial reports and information before the Board hearing MM’s parole application. There was no report on the mental state of MM (s 42(2)(d)(iii)). Vital information as required by s 42(2)(d)(iv) was also not furnished. This section requires a report on ‘the likelihood of a relapse into crime, the risk posed to the community and the manner in which this risk can be reduced...’. There was also no report or information regarding ‘the assessment results and progress report’ in respect of MM’s correctional sentence plan as contemplated in s 38 and required by s 42(2)(d)(v).

The Board, in turn, also failed to comply with its legal duty because it proceeded to assess MM’s parole application in the absence of the above prescribed information (at [48]). Clearly, in these circumstances no proper assessment could be made of all the issues and risks involved in releasing a prisoner on parole.

Meer J had good reason to find that there was ‘ clear evidence that...[MM]...had been rehabilitated and could be granted parole’ (at [51]). In this regard the court also relied on the evidence of one P, an expert witness called by the plaintiff. P qualified as an expert on account of his experience over two decades as a social worker specialising in the rehabilitation of offenders (at [18]). The defendant did not call any expert.

In P’s opinion ‘the decision to release or not to release on parole...had become a logistical consideration rather than an enquiry into rehabilitation and readiness to be released into society’ (at [22]). P expressed the view that in his experience MM should, on the information that was available to the Board, not have been released on parole and that further intervention by the Board was required for it to have determined MM’s parole readiness (at [23]). On the evidence before the Board, MM had in prison attended two programmes—of only three days’ duration each—on ‘Aggression and Life Skills’. According to P, however, it was impossible to rehabilitate a person with MM’s background after a few days of group sessions. The defendant, testified P, used attendance of a programme as the ‘yardstick’, whereas the true criterion was ‘how much impact the programme had on an inmate’ (at [23]). Meer J concluded that P’s ‘extreme scepticism about the prospects of a person with...[MM’s]...criminal profile being rehabilitated by the mere attendance of group sessions over a few days,...[was]...more than warranted’ (at [51]).

MM’s ‘criminal profile’ consisted of ‘a long list of previous convictions dating back some 26 years to 1980’ (at [7]). For present purposes only some of these convictions need to be identified as the pertinent highlights in an impressive criminal career. In November 1987 MM was sentenced to 12 years’ imprisonment for murder. This sentence ran concurrently with an earlier sentence of nine months’ direct imprisonment imposed for assault with intent to do grievous bodily harm. In December 1996 MM was released on parole in respect of the murder sentence imposed in 1987. Eleven months later and whilst still on parole, he committed theft for which he was sentenced to four and a half years’ direct imprisonment (at [7.12]). In 2004 MM was again convicted of, amongst others, two thefts, assault with intent to do grievous bodily harm and assault (at [8.1]–[8.2]). The result was that by the time he was placed at Brandvlei Correctional Centre he had a sentence of nine and a half years’ imprisonment to serve, with March 2013 as the date on which his sentence was to expire (at [9]–[10]). In 2007 MM was considered but not recommended for parole (at [11]). On three different occasions during the course of 2008, MM’s conduct as a prisoner was such that he was found guilty of internal ‘disciplinary infringements’ as provided for in s 23 of the CS Act (at [11]–[12]). In March 2010 MM was granted parole—some three years prior to his normal release date. And it was within four months of having been released on parole that MM assaulted the plaintiff in her home.

At [50] Meer J stated that the reasonable person in the position of the Board and appraised of the following three facts would, in the absence of clear evidence of rehabilitation, have foreseen the reasonable possibility that MM, if released on parole, ‘would cause harm of the kind he ultimately caused to the plaintiff’ (at [50]): first, MM’s record indicated that he was ‘a habitual violent criminal who was historically not rehabilitated by time spent in prison, nor by early release on parole’ (at [50.1]); second, MM had previously violated parole conditions whilst serving a sentence for murder (at [50.2]); third, after 2004 MM had committed offences in prison and ‘his aggression was flagged on more than one occasion’ by prison officials (at [50.3]). At [53] Meer J concluded:

‘Given the absence of evidence that...[MM]...had been rehabilitated, the Board ought in the circumstances to have taken reasonable steps to guard against the foreseeable harm of...[MM’s]...release on parole, by refusing his parole application. Failure to do so was an act of negligence. A further act of negligence was the failure to comply with the mandatory statutory requirements of...[s 42(2) of the CS Act].’

In Naidu the defendant’s negligence ‘was causally connected to the harm suffered by the plaintiff’ (at [53]). The plaintiff, it can be argued, was the primary victim of the defendant’s negligence. However, seen from a different and more subtle angle, there was also an indirect or secondary ‘victim’, namely MM, the hapless and irresponsible parolee whose parole readiness was never properly assessed before being released on parole. MM was indeed convicted of the crimes he had committed against the plaintiff while he was on parole but should have been in prison. This, in turn, resulted in a further but well-deserved hefty and effective sentence of 15 years’ direct imprisonment. It would appear that in Naidu the parole system failed not only society and the plaintiff, but also MM, who was negligently released on parole in circumstances where his parole was, in the absence of rehabilitation, simply an opportunity to commit further crimes. To put the matter in proper perspective: it is also in the best interests of a convicted offender that parole readiness be assessed properly and, where required, parole be refused.

Kelly v Minister of Correctional Services & others 2016 (2) SACR 351 (GJ)

In the above matter Satchwell J set aside the Minister’s refusal of parole and the recommendations made by the National Council for Correctional Services (the ‘NCCS’). The NCCS is a council established in terms of s 83 of the CS Act. Its functions and duties are set out in s 84. One of its functions is to advise the Minister.

The applicant in Kelly was serving a sentence of life imprisonment. He sought a review of the Minister’s refusal of his parole, which was based on recommendations of the NCCS and required him as a sentenced offender to comply with certain conditions before being released.

One condition of the NCCS was that the applicant was ‘to participate further in the Gang Management Strategy’ programme (the ‘GMS programme’). However, it turned out that no GMS programme ‘was ever availed to the applicant prior to the NCCS recommendation’ and that there ‘had been no initial participation to be taken further’ (at [21]). The applicant was also not offered the opportunity to attend such a GMS programme at some other prison at any time before and after the recommendation and decision. There was no fair administrative action (at [25]). Satchwell J explained as follows (at [27]–[28]):

‘If the prisoner had been told that his parole depended on the condition that he fly to the moon by waving his arms, there would be no doubt in finding that such condition is an impossibility and therefore the recommendation and decision predicated upon such flight to the moon were unreasonable and must be set aside[.]

In the present circumstances, attendance of a course on ‘‘Gang Management Strategy’’ has proven equally impossible for respondents to procure and applicant to attend prior to making the recommendation and decision (as well as thereafter). It cannot be considered to have been anything other than an inadmissible or incompetent consideration.’

The second condition set out in the NCCS recommendation, and adopted by the Minister in refusing parole, was that the applicant be encouraged and assisted to develop academic and/or practical skills ‘to compete in the labour market once released on parole’ (at [7]). This condition was, according to Satchwell J, too vague and lacked details such as how or when or where, or in what manner, the applicant should develop skills which will assist him in the ‘labour market’ (at [37]). It appeared, furthermore, that the NCCS or the Minister had paid insufficient attention to the many courses successfully completed by the applicant in prison, such as ‘basic computer skills’ and ‘information technology’ (at [30], [33] and [41]). Satchwell J also referred to a matter which she preferred to call a ‘general concern’, namely that sentenced offenders with an advantaged background and academic qualifications might somehow be released on parole earlier than those of disadvantaged background and less education.

Satchwell J also found it unacceptable that the Minister should have set a condition that the offender ‘should participate in Restorative Justice Processes involving the family of the victims as well as the community’ (at [42]). She observed as follows (at [46]–[47]):

‘If any restorative justice is desired by victims or their families, or the community, it is hardly possible for the applicant to go and visit the victims or hold a town-hall meeting, or devise a programme. Respondents are absolutely silent on what could or should be done, and who should do it, and when or how.... It may be, of course, that none of the victims or their families wish to participate in any restorative-justice programme. That is their right. They are not obliged to interact with, meet with, communicate or engage with—and certainly not forgive—the offender. And no burden should ever be placed on such a victim, that he or she is responsible for the continued incarceration of the offender.’

In setting aside the Minister’s refusal of parole as well as the NCCS recommendations, Satchwell J found it necessary to emphasise that whilst the applicant was not as of right entitled to parole, he had the right to a fair hearing which involves a fair process (at [56]). The process is unfair where the authorities set up ‘hurdles which he cannot overcome’ (at [57]).

Having set aside the refusal of parole and the recommendations, the court ordered that the prescribed processes be carried out ‘in a manner which requires expedition’ (at [59]). For this purpose, the court ordered the respondents to reconsider the applicant’s parole in terms of a strict and tight time frame which was duly set out by the court in [60](b) to [60](g) of its judgment.

Remarks in conclusion

In Democratic Alliance v President of the Republic of South Africa & others 2013 (1) SA 248 (CC) at [13] the Constitutional Court observed that ‘an effective criminal justice vital to our democracy’. It is also true that parole and the system of parole form an important component of the criminal justice system (Joubert (ed) Criminal Procedure Handbook 12 ed (2016) 8).

In Naidu and Kelly—the two cases discussed above—the High Court was required to deal with situations where it was the poor and imperfect implementation of the rules governing parole—and not parole itself—that yielded results like delictual liability (in Naidu) and judicial intervention in the parole process (in Kelly).

The risk of a parolee violating his parole conditions by reoffending is a risk inherently present in the system of parole. What is required, though, is that parole readiness be properly assessed on the basis of reliable information concerning rehabilitation. Clearly, in Naidu the Board’s decision to release MM on parole was unacceptable—not because it turned out that MM had committed crimes whilst on parole, but because MM’s release on parole was approved in the absence of evidence of rehabilitation and without proper consideration of his criminal profile and history of violent crime. Indeed, in Naidu the Board’s negligence defeated one of the main objectives of parole, namely the successful reintegration of the sentenced offender into society.

Moses Parole in South Africa (2012) 14 states: ‘The tension in the relationship between the State and its imprisoned citizens is acutely manifested in the assessment and consideration of prisoners for parole and the ultimate decision whether or not to release such prisoners on parole.’ It is against this particular background that Kelly (supra) must be understood. The ‘tension’ that exists is a healthy and necessary one. However, it gets disturbed when parole processes and decisions lack reasonableness and fairness. This much happened in Kelly where a robust judgment by Satchwell J was necessary to enforce fair administrative action (at [56]).

To be sure, the question whether parole should be granted or not is at the best of times not an easy one. However, those who have the responsibility to assess parole readiness and decide on parole would benefit greatly from the principles and guidelines identified and applied in Naidu and Kelly.

Steph van der Merwe


 What is the meaning of ‘hearsay evidence’ in s 3(4) of the Law of Evidence Amendment Act 45 of 1988?

This was one of the questions that Eksteen J had to consider in Pentree Limited v Nelson Mandela Bay Municipality [2017] 2 All SA 260 (ECP). He was requested, in that case, to ‘deal generally with the issue of whether a party in expropriation proceedings may, through a valuer called as an expert witness, adduce evidence of statements made to the valuer by other persons in respect of matters which have influenced her valuation of the land, in circumstances where such other persons are not called as witnesses’ (at [9]).

Eksteen J looked at the position before the enactment of the Law of Evidence Amendment Act 45 of 1988 and considered what effect the Act had to change that position. In Estate De Wet v De Wet 1924 CPD 341 Watermeyer J described hearsay evidence as ‘evidence of statements made by persons not called as witnesses which are tendered for the purpose of proving the truth of what is contained in the statement’. In R v Miller 1939 AD 106, the same judge, now elevated to the Appellate Division, went further (at 119):

‘Statements made by non-witnesses are not always hearsay. Whether or not they are hearsay depends upon the purpose for which they are tendered as evidence. If they are tendered for their testimonial value (ie as evidence of the truth of what they assert), they are hearsay and are excluded because their truth depends upon the credit of the asserter which can be tested only by his appearance in the witness box. If on the other hand they are tendered for their circumstantial value, to prove something other than the truth of what is asserted, then they are admissible if what they are tendered to prove is relevant to the enquiry.’

These pronouncements, said Eksteen J, led Botha J in Lornadawn Investments (Pty) Ltd v Minister van Landbou 1977 (3) SA 618 (T), where the court ‘gave consideration to a similar issue’ to that before Eksteen J, to ‘conclude that evidence which is apparently of a hearsay nature tendered for a purpose other than establishing the truth of the content thereof is not struck by the hearsay rule and that such evidence is admissible provided that such other purpose for which it is tendered is relevant to the issues in dispute’ (at [12]).

It was argued in Pentree that the validity of these pronouncements ‘had been impugned by the promulgation of the Evidence Act and that the common-law rules relating to hearsay [found] no application after 1988’ (at [12]). Counsel relied, for this submission, on this passage in The South African Law of Evidence 2 ed (2009) by Zeffertt and Paizes at 389:

‘No longer is hearsay defined along assertion-oriented lines, with the result that the hearsay status of the evidence no longer depends upon whether or not it is tendered to prove the truth of what it (expressly or impliedly) asserts. In its place is a declarant-oriented definition that, in effect, identifies hearsay according to whether or not the traditional ‘‘hearsay dangers’’ are present.’

The definition in s 3(4) of the 1988 Act provides that ‘‘‘hearsay evidence’’ means evidence, whether oral or in writing, the probative value of which depends upon the credibility of any person other than the person giving such evidence’. Eksteen J, while he conceded that ‘the common law rules in respect of hearsay no longer find application’, was, however, not persuaded that the statement of Zeffertt and Paizes could be accepted ‘without qualification’ (at [13]). His reasons were these:

‘At common law, if the evidence of a statement by a non-witness is tendered for a purpose other than its testimonial value then the truth or otherwise is irrelevant and it is not dependent for its reception on the credibility of the asserter. It is not hearsay....The position under the Evidence Act appears to me to be much the same. If it is not tendered for its testimonial value its probative value is not dependent on the credibility of any person other than the person giving such evidence. It is therefore not hearsay as defined in the Evidence Act.’

He drew further on what was said in Mdani v Allianz Insurance Ltd 1991 (1) SA 184 (A), where Van Heerden JA (at 189–90) made these remarks in considering the effect of s 3:

‘If A testified that B made such an admission, A’s evidence in itself is clearly not hearsay. Whether B in fact made the admission depends upon A’s credibility and can be tested by cross-examination. What is hearsay is the content of the admission if it is to be used to establish the truth of what was said. Whether the content is true or not, depends entirely on B’s credibility.’ (Emphasis added by Eksteen J).

In the light of these observations and pronouncements, reinforced by the approach of Botha J in Lornadawn (at 622F–H), Eksteen J was convinced ‘that the purpose for which the evidence is tendered may still determine whether it is hearsay or not within the definition in the Evidence Act’ (at [15]).

So was the evidence in question hearsay? In Lornadawn Botha J described the function of the court in expropriation matters, since there was no lis between the parties and no onus on the plaintiff, as being that of a ‘super valuer’: it had to place itself in the shoes of the notional informed seller and buyer and have regard to everything which a seller and buyer would have experienced in the open market as well as all the information that would have been at their disposal in order to determine market value. The court, of course, could not go into the market to gather information, but it was part of the function of an expert valuer to do so and to found his opinion on that information. In determining compensation, a court was, according to Botha J, involved in ‘an enquiry into the probable conduct of two fictitious persons to whom we should attribute full knowledge of all circumstances which would play a role in the manner in which they would have acted in their negotiations with one another in determining a purchase price for the subject property’ (per Eksteen J at [21]). And those fictitious persons would be influenced in their actions and conduct by relevant communications received by other people. These communications are, said Botha J (at 627A–D), ‘direk ter sake in die ondersoek warmee die Hof besig is, en dit is so afgesien van die vraag of die inhoud van sulke medelings objektief gesproke die waarheid is of nie’ (cited by Eksteen J at [21]).

In essence, Eksteen J concluded (at [22]), ‘evidence of such communications are not dependent for their reception on the credibility of the asserter and the probative value thereof in expropriation proceedings depends on the weight which the notional willing buyer and willing seller would have attached thereto in the light of all the known facts’. It was therefore ‘not hearsay at all in the context of the expropriation proceedings for the probative value thereof is not dependent on the credibility of any other person, but rather on the weight which a notional willing buyer or seller would have attached to the statement in his negotiation’ (at [27]).

It is submitted that Eksteen J was correct in labelling the evidence non-hearsay. It was similar, in nature, to the situation that arose in International Tobacco Co (SA) Ltd v United Tobacco Cos (South) Ltd 1953 (3) SA 343 (W): if W were to testify that he is a commercial traveller and that other people had told him that rumours were circulating that a particular brand of cigarette caused a particular disease, and this evidence were tendered to prove that such rumours were, in fact, circulating, would it be hearsay? It would, in my submission, not be hearsay because what the other people told W would be no less proof of the circulation of those rumours than what others told them, so that the probative value of W’s evidence that rumours were circulating would depend entirely on his own credibility.

I have, therefore, no quarrel with Eksteen J’s conclusion that the evidence was not hearsay. What I do, with respect, take issue with is his reasoning concerning the impact of s 3(4) on the meaning and definition of hearsay, and, in particular, his assertion (at [15]) that ‘the purpose for which the evidence is tendered may still determine whether it is hearsay or not within the definition in the Evidence Act’ (emphasis added).

It is important, in my view, to understand why the old approach to hearsay, which hinged around whether the evidence was tendered to prove the truth of its contents, was dysfunctional and unsatisfactory. Consider these situations:

Situation 1: The issue in a particular matter is whether X, following a traumatic event, was able to speak. W, who testifies on the issue, tells the court that he came upon the traumatised X and asked him if he was too shocked to speak, whereupon X told him: ‘I can speak’. This is clearly adduced to prove the truth of what X said to W, that he could, indeed, speak. But it is not hearsay; nor should it be so labelled. The probative value of what was said by X to W clearly depended entirely on the credibility of W, who heard the words used by X, irrespective of what those words actually were. The same would apply to similar expressions, such as ‘I am alive’ or ‘I am here’, if used to prove those very facts. The ‘assertion-oriented definition’, then, fails and the ‘declarant-oriented’ definition (as embodied in s 3(4)) provides the correct solution. It can, of course, still be said that the evidence is not being tendered to prove the truth of what is asserted, and that we are concerned here only with the circumstantial use of language by X and not with the content of the words uttered by him. But my point is that we should not even be engaging in such subtle distinctions since the plain wording of s 3(4) simply renders this unnecessary.

Situation 2: The issue is the sanity of the testator, T, whose will is challenged by those who would benefit from its invalidity. In order to show that T was sane at the time of making the will, evidence is adduced of letters addressed to and received by him. They are tendered to prove, not anything expressly said by the writers of those letters, but that those writers, given that the contents were on matters and in language appropriate to communications with a person of reasonable understanding and intelligence, believed T to have possessed an adequate comprehension of the affairs described. This type of situation—the facts of which are those of an old English case, Wright v Doe d Tatham (1837) 7 Ad & El 313—has given rise to the notorious problem of so-called ‘implied hearsay assertions’: is evidence hearsay where it is tendered to prove, not what it expressly asserts, but something that it asserts by implication? The assertion-oriented definition of old did not provide an answer to this question, and it tended to lead into a semantic swamp where fine and often unhelpful distinctions would be made between the circumstantial use of the evidence and its tendency to assert something indirectly by implication. The declarant-oriented definition in s 3(4) avoids such dilemmas by going straight to the heart of the matter: if the probative value of the evidence depends on the credibility of the absent declarant, the evidence is hearsay. In the example taken from Wright’s case, the evidence would be hearsay, since the assessment of T’s sanity by the writers of the letters depends strongly on their credibility: were they sincere in writing these letters or were they part of an elaborate charade; did they know T well enough to make any such assessment; how much time had elapsed since last they saw him? These are questions for cross-examination. The prejudice caused to the opposing party by the denial of the opportunity to cross-examine the writers of the letters on these matters is the strongest—but not the only—ground for classifying the evidence as hearsay.

It is significant that it took the Canadian courts as long as it did to identify that no distinction should be drawn between express and implied assertions for this purpose. For, in the absence of a declarant-oriented definition of the kind employed in s 3(4), there was less to alert the courts in Canada to what Fish J, in the landmark decision of the Supreme Court in R v Baldree 2013 SCC 35, [2013] 2 SCR 520 (discussed in Commentary in s 216 sv Definition of hearsay) described as the absurdity ‘that anything should turn on the grammatical form of the declarant’s assertions’ (at [42]). The key to encouraging this breakthrough is that in Canada, as in South Africa since the enactment of s 3, there is a flexible and nuanced regime for dealing with the admissibility of hearsay. Many implied assertions will be highly reliable, especially since the danger of insincerity tends to be reduced given the indirect, implicit use of what was said. Where the hearsay dangers (discussed in Commentary in s 216) are satisfactorily accounted for, these will be admitted in South Africa under s 3(1)(c). Without this liberal approach to admissibility, courts would be much less inclined to bring implied assertions within the hearsay fold and to risk, by doing so, that they might be lost as evidence despite their reliability in a given case.

Situation 3: A ship has sunk and the question raised by the insurer is whether it was seaworthy when it sailed. Evidence is tendered that the captain of the ship, after conducting an inspection, embarked on the voyage, taking with him his entire family. Is this evidence hearsay? The assertion-orientated definition does not even get to grips with the problem, as it is meaningless to talk about the truth of the contents of the communication—express or implied—where the communication is non-verbal and not even intended to be communicative in the first place.

The declarant-oriented definition does. It asks if the evidence—tendered in oral form by a witness who saw all these events (and so satisfying the requirement in s 3(4) that the evidence be ‘oral or in writing’)—depends for its probative value on the captain’s credibility. That it clearly does, is illuminated by the kinds of question the opposing party, denied the opportunity to cross-examine him, might have liked to ask: how rigorous was the examination; how experienced was the captain in assessing such matters; was he sober at the time; was he a risk-taker; how good was the lighting when he inspected the ship?

It would take some boldness for courts employing the common-law definition to identify such evidence as hearsay. Interestingly, even the Canadian Supreme Court in Baldree (at [63]) thought the issue ‘best left for another day’, adding merely this comment by PR Rice (1992) 65 Temple Law Review 529 at 536: ‘[O]ne can engage in conduct without ever intending to communicate anything to anyone [but] the same is not true of speech or a combination of speech and conduct (for example, placing a bet) because the sole purpose of speech is communication.’ It is submitted that, whereas the danger of insincerity will usually be significantly reduced in such cases, those of erroneous memory and perception will frequently remain. Whether they are strong enough to signify should be something a court will take into account in deciding admissibility under s 3(1)(c), not categorisation. The mere potential for such danger—and the prejudice it could cause—is reason enough to label the evidence hearsay. Admissibility is another matter.

In short, the triumph of s 3 is that it liberates us from a number of straitjackets: the tyranny of rule and fixed exceptions; the forcing of evidence into ill-fitting and synthetic categories; the service of formalism at the expense of substantive justice and effective function; the unquestioning deference to stale precedent; the dead hand of linguistic and grammatical artifice; and the muddle of hair-fine distinctions.

It is in this spirit that I made the comments in The South African Law of Evidence that are referred to by Eksteen J. To say, as Eksteen J did in Pentree, that the position under the Evidence Act ‘appears to be much the same’ as that at common law, and that the ‘purpose for which evidence is tendered may still determine whether it is hearsay or not within the definition in the Evidence Act’ (at [14] and [15]; emphasis added), is to deny or negate some of these gains. It runs, moreover, counter to the very wording of s 3(4), which makes no mention of whether or not the evidence is tendered to prove the truth of its contents.

It must be acknowledged that the old definition will produce the correct results in the majority of cases. But it is a flawed theory that fails entirely to account for some cases. To borrow an analogy from the field of theoretical physics: Newton’s laws served us well for centuries and were good enough to get us to the moon. But Einstein’s theory showed Newton’s physics to be flawed and incomplete: it could not account for the motion of Mercury, the bending of light from the stars by the sun, and many other phenomena. We will probably never reach the stars. But Einstein’s revelations at least cleared away a fatal impediment.

Andrew Paizes


During the period under review (1 February 2017 to 12 June 2017) there were no amendments to the Criminal Procedure Act 51 of 1977. Bills which will affect provisions in this Act are referred to below. Other legislative activity pertinent to our criminal justice system is also identified below.

(i)  The Justice Administered Fund Act 2 of 2017

The above Act was signed by the President on 4 April 2017 and published for general information on 6 April 2017. See GN 344 in GG 40773 of 6 April 2017. The Act will come into operation on a date to be fixed by the President by proclamation in the Gazette.

The Act establishes a Justice Administered Fund (the ‘Fund’) and provides for the management, control, investment and utilisation of money in this Fund.

Section 3(b) of the Act provides that money received as bail, payable in terms of the Criminal Procedure Act 51 of 1977 or any other Act, must be administered through the Fund. It should be noted that this section regulates the administration of bail money through the Fund and has no effect on payment and repayment of bail money as provided for in Chapter 9 ‘Bail’ (ss 58–71) of Act 51 of 1977.

(ii) Implementation of the Rome Statute of the International Criminal Court Act Repeal Bill [B23—2016]:    Withdrawal by Minister

The Minister of Justice and Correctional Services has withdrawn the above Bill in accordance with Rule 334 of the Rules of the National Assembly. See Announcements, Tablings and Committee Reports of the Parliament of South Africa (13 March 2017) at 2.

(iii)  Children’s Amendment Act 17 of 2016

The above Act was signed by the President on 18 January 2017 and was published for general information on 19 January 2017. See GN 42 in GG 40773 of 19 January 2017. The Act (the ‘Amendment Act’) will come into operation on a date to be fixed by the President by proclamation in the Gazette.

In terms of s 120(1)(b) of the Children’s Act 38 of 2005, a finding that a person is unsuitable to work with children may be made by ‘any...court in any criminal or civil proceedings in which that person is involved...’. Section 2(b) of the Amendment Act inserts a new s 120(4) which sets out the circumstances and offences when a person, in criminal proceedings, ‘must be deemed unsuitable to work with children...’. However, in terms of s 120(4A)—as inserted by s 2(c) of the Amendment Act—a court may not make an order of unsuitability where the person concerned was a child at the time of the commission of the offence unless:

‘(a)  the prosecutor has made an application to the court for such an order;
(b) the court has considered a report by the probation officer referred to in section 71 for the Child Justice Act, 2008,  which deals with the probability of committing an offence contemplated in subsection (4), against a child;
(c) the person concerned has been given the opportunity to address the court as to why his or her particulars should not be included in the Register; and

(d) the court is satisfied that substantial compelling circumstances exist based upon such report and any other evidence, which justify the making of such an order.’

The ‘Register’ referred to in s 120(4A)(c) above is the National Child Protection Register (the ‘NCPR’) as created and governed by ss 111–128 of the Children’s Act 38 of 2005.

The new s 120(4A) of Act 38 of 2005 is similar to s 50(2)(c) of the Criminal Law (Sexual Offences and Related Matters) Amendment Act 32 of 2007. Section 50(2)(c) was inserted by s 7(b) of Act 5 of 2015 which came into operation on 7 July 2015. See GG 38977 of 7 July 2015. Section 50(2)(c) was introduced as a result of the Constitutional Court decision in J v National Director of Public Prosecutions & others 2014 (2) SACR 1 (CC), a decision that followed S v IJ 2013 (2) SACR 599 (WCC). See further the notes on these two cases in the discussion of s 276 in Commentary, sv The National Register for Sex Offenders (‘NRSO’): The convicted sex offender who was a child at the time of the commission of the offence. The new s 120(4A) of Act 38 of 2005 will ensure that placement of a child’s name on the NCPR would be consistent with the decision of the Constitutional Court in J (supra).

(iv)  Judicial Matters Amendment Bill [B14—2016]

In para 1 of the memorandum on the objects of the above Bill published in GG 40274 of 14 September 2016, it is stated that the Bill seeks to amend provisions of several Acts in order to address practical and technical issues of a non-contentious nature. The important clauses seeking to amend the Criminal Procedure Act 51 of 1977 were discussed in CJR 2016 (2) at 9–10.

The invitation by Parliament’s Portfolio Committee on Justice and Correctional Services to make written submissions on the Bill, closed on 15 March 2017. It would therefore appear that the Bill might become law before the end of this year.

(v)  Criminal Procedure Amendment Bill [B2—2017]

The above Bill was published in GG 40487 of 9 December 2016. It seeks to make various amendments to Chapter 13 ‘Accused: Capacity to Understand Proceedings: Mental Illness and Criminal Responsibility’. This chapter consists of three sections: s 77 (capacity of accused to understand proceedings); s 78 (mental illness or mental defect and an accused’s criminal responsibility) and s 79 (panel for purposes of an inquiry and report under ss 77 and 78).

Proposed amendments to s 77(6)(a)(i)

The above section provides that where a court finds that an accused is incapable of understanding the proceedings so as to make a proper defence, the court must—in the case of certain serious crimes and circumstances as identified in s 77(6)(a)(i)—direct the detention of the accused in a psychiatric hospital or a prison pending the decision of a judge in chambers in terms of s 47 of the Mental Health Care Act 17 of 2002. However, in a judgment delivered on 26 June 2015 the Constitutional Court declared s 77(6)(a)(i) inconsistent with the Constitution to the extent that it provides for (a) the compulsory imprisonment of an adult and (b) compulsory hospitalisation or imprisonment of children. The declaration of invalidity was suspended for two years from the date of the judgment in order to allow Parliament to address the matter having regard to the defects raised in the judgment. See De Vos NO & others v Minister of Justice and Constitutional Development & others 2015 (2) SACR 217 (CC) at [69]. This case is discussed in the notes on s 77 in Commentary, sv Section 77(6): The unconstitutionality of subparagraphs (i) and (ii).

Clause 1(b) of the Bill aims to amend s 77(6)(a)(i) in accordance with constitutional requirements identified by the Constitutional Court in De Vos (supra). In terms of this clause a court will have a discretion—in dealing with a s 77(6)(a)(i) situation—to order that the accused be detained in a psychiatric hospital (or temporarily in a correctional health facility of a prison where no bed is available in a psychiatric hospital) ‘if the court is of the opinion that it is necessary to do so on the grounds that the accused poses a serious danger or threat to...[himself]...or to members of the public or to any property belonging to him...or any other person’. In terms of clause 1(b) the court will also, depending on all the circumstances, be able to order the detention of the accused in a designated health establishment as if he were an involuntary mental health care user contemplated in s 37 of the Mental Health Care Act 17 of 2002. A further alternative for the court would be to order the release of the accused subject to such conditions as the court considers appropriate.

It is submitted that clause 1(b) is consistent with constitutional requirements and meets all the problems identified in De Vos (supra) at [46], [48], [51] and [52]. Given the time limit set by the court, it is to be expected that clause 1(b) will become law on or before 25 June 2017.

Proposed amendments to s 77(6)(a)(ii)

Section 77(6)(a)(ii) deals with the situation where a court finds that the accused has committed an offence other than the serious offences stipulated in s 77(6)(a)(i) or that the accused has not committed any offence. Before the decision of De Vos (supra) on 26 June 2015, the court was simply required to make an order that an accused in the s 77(6)(a)(ii) situation be admitted to and detained in an institution as if he or she were an involuntary mental health care user contemplated in s 37 of the Mental Health Care Act 17 of 2002. In De Vos it was found that s 77(6)(a)(ii) breached the right to freedom and security of the person as guaranteed in s 12 of the Constitution (at [57]) and amounted ‘to an arbitrary deprivation of freedom under s 12’ (at [58]). ‘The mere fact’, it was said at [66], ‘that an accused person brushes shoulders with the criminal justice system is not a just cause for institutionalisation and renders the provision constitutionally invalid in respect of such a person.’ However, at [67] the court concluded that a ‘reading-in’ was appropriate and that s 77(6)(a)(ii) could be remedied by extending the options available to the court. The Constitutional Court accordingly ordered that as from 26 June 2015 two further options had to be read into s 77(6)(a)(ii), namely an order by the court that the accused be ‘released subject to such conditions as the court considers appropriate’ or an order by the court that the accused ‘be released unconditionally’ (at [69]).

In proposing a new s 77(6)(a)(ii), clause 1 of the Bill incorporates the Constitutional Court’s ‘reading-in’ as set out above. Indeed, in De Vos (supra) at [67] the Constitutional Court noted that the legislature could amend s 77(6)(a)(ii) ‘as it deems fit...provided such amendment complies with this judgment and the Constitution’.

Proposed amendment to s 78(6)

An accused must be found not guilty if the court finds that the accused was at the time of the commission of the offence ‘by reason of mental illness or intellectual disability not criminally responsible’ for his act. See s 78(6)(a) as read with s 78(1). Once an accused has been acquitted on this basis, the court is required to make one of several orders as set out in s 78(6). Clause 2 inserts a further possible order that the court can make, namely that the acquitted person be

‘temporarily detained in a correctional health facility of a prison where a bed is not immediately available in a psychiatric hospital and be transferred where a bed becomes available, if the court is of the opinion that it is necessary to do so on the grounds that the accused poses a serious danger or threat to himself or herself or to members of the public or to any property belonging to him or her or any other person...’.

The above proposed amendment is an eminently sensible one. Clause 2 also seeks to amend s 78(6) to the extent that detention in a prison could no longer be ordered unless it is in the circumstances as described in the proposed amendment quoted above.

Proposed amendments to s 79

Section 79(1) governs the constitution of the panel of experts who should report to the court for purposes of ss 77 and 78. The size of the panel can vary according to the nature or classification of the offence(s) involved. In S v Pedro 2015 (1) SACR 42 (WCC) attention was drawn to inconsistencies in the interpretation and application of s 79(1)(b) as far as the composition of the panel was concerned. See in this regard the discussion of Pedro in the notes on s 79 in Commentary, sv Section 79(13): Directives issued by the NDPP.

Clause 3 of the Bill aims—in line with the findings and views in Pedro (supra)—to clarify the composition of the panels and secure a consistent application of s 79(1) in all courts across the country. The proposed new s 79(1) will read as follows:

‘(1) Where a court issues a direction under section 77(1) or 78(2), the relevant enquiry shall be conducted and be reported on—

(a)  where the accused is charged with an offence other than one referred to in paragraph (b), by the head of the designated health establishment designated by the court, or by another psychiatrist delegated by the head concerned; or

(b)  where the accused is charged with murder or culpable homicide or rape or compelled rape as provided for in section 3 or 4 of the Criminal Law (Sexual Offences and Related Matters) Amendment Act, 2007, respectively, or another charge involving serious violence, or if the court considers it to be necessary in the public interest, or where the court in any particular case so directs—

(i)   by the head of the designated health establishment, or by another psychiatrist delegated by the head concerned;

(ii)   by a psychiatrist appointed by the court;

(iii)  by a psychiatrist appointed by the court, upon application and on good cause shown by the accused for such appointment; and

(iv)  by a clinical psychologist where the court so directs.’

Clause 3 of the Bill also proposes the repeal of s 79(13). This section empowers the National Director of Public Prosecutions to issue directives concerning the prosecutor’s application for the appointment of a psychiatrist in certain specified circumstances. See also Pedro (supra) at [25] and [48]–[69]. Repeal of s 79(13) is a consequence of the fact that the proposed new s 79(1)(b) does not include a prosecutorial right to apply for an extra psychiatrist to be appointed in certain circumstances.

Sections 77, 78 and 79: ‘mental defect’ to be replaced by ‘intellectual disability’

The term ‘mental defect’ appears in ss 77, 78 and 79. The Bill proposes the replacement of ‘mental defect’ with ‘intellectual disability’. In para 2.1.1 of the Memorandum which accompanied the Bill, it is explained that ‘intellectual disability’ is the ‘more acceptable term’ and that ‘mental defect’ is an ‘outdated term’. On the use of the term ‘intellectual disability’, see the discussion of s 78 in Commentary, sv Section 78(6): ‘mental illness or mental disability’. The term ‘intellectual disability’ is already used in s 78(6). It is submitted that the introduction of this term in all other relevant provisions of ss 77, 78 and 79 would also promote consistency without disturbing any substantive criminal law principles governing capacity to understand proceedings and criminal responsibility.

 (a) Criminal Law

Is a person who assists another to commit suicide guilty of murder? Physician administered euthanasia and physician assisted suicide considered

Minister of Justice & others v Estate Stransham-Ford 2017 (3) SA 152 (SCA)

This case arose out of the wishes of a terminally ill person, SF, who had approached the High Court claiming an order that a medical practitioner could either end his life by administering a lethal substance, or provide him with the lethal substance to enable him to administer it himself. He sought an order that the common law in relation to the crimes of murder and culpable homicide should, in terms of s 39(2) of the Constitution, be developed to allow for such an order to be made. Two hours after SF died, the High Court, unaware of that fact, granted an order that the medical doctor who acceded to SF’s request would not be acting unlawfully, and hence would not be subject to prosecution or to disciplinary proceedings for such conduct.

The appeal to the Supreme Court of Appeal was against this order. The appeal was successful, and the order was set aside for ‘three inter-related reasons’ (at [101]), any one of which would have been decisive. First, the death of SF two hours before the order was made meant that the cause of action had ceased to exist. Once the judge who had made it was told of SF’s death the following week, before delivering his reasons, he refused to recall his order on the ground that his judgment had broader societal implications. It was held that he was wrong not to rescind his order since it had been granted on the erroneous basis that SF was still alive.

Secondly, there was no full and proper examination of the present state of our law on the subject. And, thirdly, the order had been made on an incorrect and restricted factual basis, without complying with the Uniform Rules of Court and without affording all interested parties a proper opportunity to be heard.

Since, however, there was the ‘dilemma’ that the decision of Fabricius J in the High Court (see 2015 (4) SA 50 (GP)) stood as a ‘reasoned and reported judgment by the High Court’ which might, unless the Supreme Court of Appeal, at least to some extent, addressed the merits, ‘be taken as having some precedential effect’ (at [27]), the court felt compelled to deal with the merits.

Wallis JA, who gave the judgment of the court, considered it necessary to undertake a ‘brief exposition of the current state of our law in this area’ (at [29]). He found the succinct assertion of the High Court that ‘[t]he current legal position is that assisted suicide or active voluntary euthanasia is unlawful’ (at [10]), to be, ‘not only not supported by the authorities relied upon’, but also ‘a wholly inadequate analysis of the relevant law in this area’ (at [29]).

The judgment of Wallis JA covered a wide expanse and included a close examination of the decided South African cases, the provisions of the Constitution and comparative foreign law. The following are some of the propositions expounded and issues considered:

(1)   Neither suicide nor attempted suicide is a crime in South Africa: see R v Peverett 1940 AD 213.

(2)  A person may refuse medical treatment that would otherwise prolong life. This is an aspect of personal autonomy that is constitutionally protected and it would not ordinarily be regarded as suicide: see Re Conroy 486A 2d 1209 (NJSC 1985) at 1224. The constitutional provisions involved are ss 10 (right to dignity) and 12(2)(b) (right to bodily integrity).

(3)  The refusal of treatment can be made by a patient only if he or she has the mental and legal capacity to make such a decision. In such cases doctors do not commit a criminal offence ‘by ceasing treatment or other forms of medical intervention that serve neither a therapeutic nor a palliative purpose’ (at [33]). Where patients lack capacity, decisions are usually made by the doctors in conjunction with family members and others having responsibility for the patient. In cases of uncertainty, it might be desirable to seek a declaratory court order (see Clarke v Hurst NO & others 1992 (4) SA 630 (D)).

(4)  A medical practitioner ‘commits no offence by prescribing drugs by way of palliative treatment for pain that the doctor knows will have the effect of hastening the patient’s death’. This is referred to as ‘the ‘‘double effect’’, where the drugs serve the purpose for which they were prescribed, but have potentially detrimental side effects’ (see Clarke v Hurst supra). There were, said Wallis JA (at [35]), many steps available to avoid interminable and purposeless treatment or the preservation of life ‘as a purely mechanical process artificially maintained’, and considerable advances had been made in palliative care to alleviate suffering.

(5) On the other hand, a ‘mercy killing’ ‘undoubtedly constitutes the crime of murder’ at [36]): see S v Hartmann 1975 (3) SA 532 (C); S v De Bellocq 1975 (3) SA 538 (T) and S v Marengo 1991 (2) SACR 43 (W). These cases, however, had nothing to do with either physician assisted suicide (PAS) or active physician administered euthanasia (PAE), since in none of them did the deceased ask for his life to be ended. They were relevant in confirming, however, that consent is not a defence to murder.

(6) In S v Robinson & others 1968 (1) SA 666 (A), the deceased did arrange for others, including family members, to kill him for financial (insurance-related) reasons. The case confirmed that PAE did constitute murder, so that a ‘medical practitioner who administers a lethal agent to a patient at the latter’s request commits the crime of murder’ (at [40]). Wallis JA could see no reason for distinguishing between the actions of a medical practitioner from those of a family member or friend in such circumstances.

(7) The question in this case was, therefore, whether the law should be changed. The case before the court was, in effect, challenging both Peverett and Robinson. However, neither the principles nor the cases in question were addressed by the High Court. If the law were to be developed, this ‘needed to be confronted squarely and the scope and ambit of the requisite exception to, or departure from, existing principle had to be defined’. An ‘order making such a profound change to our law of murder, without any consideration of applicable principles, should not have been made’ (at [41]). Even if the deceased had not died when he did, since no doctors had come forward to say they were willing to do as requested by SF, the matter was purely ‘academic’ and the court should not have entered into such an academic inquiry.

(8) Wallis JA considered next the issue of assisted suicide (PAS), which was sought as an alternative order. In this context the decision in Ex parte Die Minister van Justisie: In re S v Grotjohn 1970 (2) SA 355 (A) was relevant. In that case the deceased was a bipolar, partially paralysed woman, whose husband was having an affair with another woman. In the course of a heated argument, she said that she would shoot herself. The accused (her husband) obtained a bullet from elsewhere in their apartment, loaded the gun in her presence, and said: ‘Skiet jouself dan as jy wil want jy is a las.’ She did so. The Appellate Division was asked to consider whether encouraging, providing the means for or helping another to commit suicide was a crime, and, if so, what crime. There was, said the court, no simple ‘yes’ or ‘no’ answer to this question. The ordinary principles of the criminal law—including those relating to causation—had to be invoked in the context of the particular facts of each case. The court, however, recognised the possibility that murder may be committed in such circumstances if the requisite intent were present and there was no break in the chain of causation between the accused’s actions and the death of the deceased (as to which, compare the ‘suicide pact’ cases of Peverett (supra) and S v Gordon 1962 (4) SA 727 (N) discussed by Wallis JA). If the accused lacked intent but had mens rea in the form of negligence, the crime of culpable homicide would be committed if all the requirements (including causation) were present. But it could not be said ‘that in the current state of our law PAS is in all circumstances unlawful’ (at [54]); emphasis added); the High Court’s statement to that effect ‘went too far’.

(9)  A court confronted with a case of PAS would thus have to consider how the principles articulated in Grotjohn ‘should be applied and adapted to the present day’ (at [55]), with particular attention to (i) the facts of the particular case; (ii) the changes in medical science in the 50 years since that decision; and (iii) the requirement in s 39(2) of the Constitution that, in the development of the common law, the courts must strive to give effect to the nature, purport and objects of the Bill of Rights, in respect of which assistance could profitably be sought by looking at other jurisdictions. In particular, a court, if it was of the view that the common law should be developed, would have to consider the area of the criminal law that should be so developed: whether causation, intention or unlawfulness. Such considerations were absent in the High Court decision, and this rendered it impossible to assess whether any limitation of a constitutional right was reasonable or justifiable in terms of s 36 of the Constitution.

(10)  Wallis JA turned next to the foreign law, which was examined and compared. One aspect of this treatment stands out: the fact that the High Court’s ‘too ready adoption’ of the reasoning of the Supreme Court of Canada in Carter v Canada (Attorney General) 2015 SCC 5; [2015] 1 SCR 331 ‘ignored the very different context in which that case was decided’ (at [72]). It was, said Wallis JA, important to note that it was ‘only on the question of ‘overbreadth’ that the Supreme Court of Canada held in Carter that the criminalisation of aiding and abetting suicide unjustifiably infringed a protected right’. A South African court faced with the same issue would need to examine the ‘very different circumstances in this country; the availability of medical care and especially palliative care; the wide diversity of our society in its cultures and belief systems; our sense of the need to protect the poor, the weak and the vulnerable and the value attached to providing such protection’. ‘The notion of a dignified death’, said Wallis JA (at [100]), had to be ‘informed by a rounded view of society, not confined to a restricted section of it’. This was not done in this case and could not have been done ‘because of the inadequacies of the evidence and the haste with which it was decided’. When an appropriate case comes before our courts, the common law will, he concluded, ‘no doubt evolve in the light of the considerations outlined... and the developments in other countries’.

Constitutional validity of statutory offences prohibiting use of cannabis for personal consumption and related matters

Prince v Minister of Justice and Constitutional Development & others; Acton & others v National Director of Public Prosecutions & others (unreported, WCC case no 8760/2013, 31 March 2017)

How, in a constitutional democracy such as South Africa, do the courts ‘demarcate the border between the terrain in which disputes lie in the province of the courts and those controversies which are better located on a terrain belonging to the legislature and/or the executive’? This broad question faced the court in Prince, in which Davis J (in whose judgment Saldanha and Boqwana JJ concurred) had to consider this particular question: to what extent are the statutory laws that prohibit the use of cannabis and the possession, purchase or cultivation thereof exclusively for personal consumption, valid? Another aspect of the inquiry was ‘whether the invalidity of the relevant legislature framework in terms of which these prohibitions [were] couched should be determined by courts or left for further legislative and/or executive consideration’ (at [2]).

Among the legislative provisions that were impugned were: ss 4(b) and 5(b) of the Drugs and Drug Trafficking Act 140 of 1992 (the ‘Drugs Act’), read with part 3 of Schedule 2 to the Drugs Act in so far as it related to ‘simple possession, cultivation, transportation and distribution of cannabis for personal communal consumption’; s 22A(10) of the Medicines and Related Substances Control Act 101 of 1965 (the ‘Medicines Act’), read with Schedule 7 of that Act; and s 40(1)(h) of the Criminal Procedure Act, which empowers a peace officer without a warrant to arrest any person who is reasonably suspected of having committed an offence under any law governing the making, supply, possession or conveyance of, inter alia, dependence-producing drugs.

Section 4(b) of the Drugs Act prohibits, except in specific cases, the use or possession of cannabis (among other substances), and s 5(b) prohibits, again except in specific cases, any dealing in such substances. Section 22A(9)(a)(i) of the Medicines Act prohibits the use, possession, manufacture or supply of cannabis. The applicants invoked what Davis J (at [11]) called a ‘veritable constitutional laundry list’ to argue that the criminal prohibition of the use and possession of cannabis in their own homes and in ‘properly designated places’ was unconstitutional. The list included the rights to equality, dignity and freedom of religion, but the thrust of the argument hinged around the right to privacy.

The court considered that the doctrine of res judicata was no impediment in this case since the decision in Prince v The President, Cape Law Society & others 2002 (2) SA 794 (CC), relied on by the respondents, turned on the narrow question involving limited exemptions, for religious reasons, from the criminal provisions. The majority in that case found the limitation of the right to religious freedom to be justifiable on the ground that it would be impractical to administer a religious exemption without fundamentally undermining the general prohibition against possession of cannabis.

The core of the matter, said Davis J (at [20]), was whether the infringement of the right to privacy caused by the impugned legislation could be justified in terms of s 36 of the Constitution.

The court examined the nature and extent of the right to privacy, and made the following remarks: First, the right extends only to those aspects in regard to which a legitimate expectation of privacy can be harboured (see Bernstein & others v Bester & others NNO 1996 (2) SA 751 (CC) at [75]). Secondly, there is a narrowly construed, invisible core in respect of which a very high level of protection is given to an individual’s intimate personal sphere of life, which is ‘beyond interference from any public authority’, and where no justifiable limitation can take place (Bernstein at [77]). This core is left behind once an individual enters into relationships with persons outside this sphere, so that the right then becomes subject to limitation as the activities acquire a social dimension. Thirdly, it is established law that the right to privacy ‘becomes more powerful and deserving of greater protection the more intimate the personal sphere of the life of a human being which comes into legal play’ (at [22] of Prince). Fourthly, there is a connection between the right of privacy and that to dignity (see s 10 of the Constitution): privacy fosters human dignity and a right to make intimate decisions, and to have one’s personal autonomy protected is central to individual identity. Privacy is linked, too, to the right to freedom; it ‘fosters and encourages the moral autonomy of citizens which is a central requirement of governance in a democracy’ (at [24]).

On the general question of the limitation of rights under s 36(c), Davis J made these observations: First, the burden of justification rested upon the State (see Moise v Greater Germiston Transitional Local Council: Minister of Justice and Constitutional Development Intervening 2001 (4) SA 491 (CC) at [19]). Secondly, the provisions of Chapter 2 of the Constitution (including s 36) had to be interpreted in the light of values underlying an open and democratic society based on human dignity, equality and freedom. Privacy, in this context, was clearly deserving of constitutional protection, absent a clear justification to the contrary.

Was there, then, a substantial State interest to warrant such a limitation? Davis J traced the history of the prohibition against cannabis (which began in 1928) and found it to be ‘replete with the racism that has saturated South African society for more than three centuries’ (at [33]). He insisted that it was necessary for the State to provide ‘clear and plausible evidence tailored to the contemporary context to justify an infringement of an important right ... viewed in terms of autonomous acts in one’s home’ (at [34]). The ‘private moral views of a section of the community’ would not suffice.

The State’s evidence pivoted around an affidavit by a pharmacist and registrar of medicines who attested to the acute effects of cannabis, including neurological impairment of driving skills and its harmful effects on the brain and the body, evident in both learning and pregnancy. There was also an affidavit by a general medical practitioner, whose qualifications as an expert in this field were ‘hardly beyond contest’.

Against these was a detailed report by an expert in criminology, who noted that countries with more punitive policies did not tend to have lower drug use prevalence levels than those with more liberal policies. He contested claims regarding aspects of the effects of cannabis on cognitive function, and asserted that many of those who wished to end prohibition did so ‘out of a keen appreciation’ of the harms at play—and out of a belief that the most successful management of those harms requires that the task be brought into the ‘sphere of legal, transparent and constitutionally guided public institutions’ (at [50]). There was, the expert maintained, no evidence that the decriminalisation of cannabis in certain states in the United States (for instance, Colorado) had resulted in any spike in criminal conduct.

Davis J, on a balance of the medical evidence, accepted (at [61]) that it was clear that ‘uncontrolled consumption of cannabis, especially where it is consumed in large doses poses a risk of harm to the user’. But the evidence was equally clear that there was ‘a level of consumption that is safe in that it is unlikely to pose any risk of harm’, although the medical evidence on record was silent on what that level of consumption was.

The court then turned to the position in other comparable democratic societies for guidance. The respondents relied on the decision of the Supreme Court of Canada in R v Malmo-Levine [2003] SCC 74, in which a constitutional challenge to the criminal provisions under s 7 of the Charter (providing for the right to life, liberty and security of the person) was dismissed by the majority on the basis that the criminalisation of possession of cannabis was ‘not arbitrary but [was] rationally connected to a reasonable apprehension of harm’ (at 642–3). The ‘compelling’ minority judgment of Arbour J, on the other hand, examined the ‘harm principle’, which John Stuart Mill described thus: ‘The only purpose for which power can be rightfully exercised over any member of a civilized community, against his will, is to prevent harm to others.’ The record, said Arbour J, showed that ‘the prohibition of simple possession of marihuana attempts to prevent a low quantum of harm to society at a very high cost’.

In view of the fact that s 7 of the Charter was not the equivalent of s 14 of our Constitution, which provided for a clearly demarcated right to privacy, Davis J found the approach of Arbour J to be more applicable to the issue before him than that of the majority. After examining decisions in Argentina, Mexico and Alaska, and considering legislation legalising possession of cannabis in small quantities for personal use in eight states in the United States and soon, it seems, in Canada, as well as several other jurisdictions around the world, he found (at [90]) that there was ‘no longer a consensus that can regard such limitations as justifiable’. The evidence, in his view, ‘supports the argument that the legislative response to personal consumption and use is disproportionate to the social problems caused as a result thereof’ (at [102]).

Both ss 4 and 5 of the Drugs Act had, accordingly, to be amended to ensure that those provisions did not apply to those who used ‘small quantities of cannabis for personal consumption in the privacy of a home’, since the present position unjustifiably limited the right to privacy. It would be for the Legislature to determine the extent of what would constitute ‘small quantities in private dwellings’. Even if the respondents were correct in identifying the objections of the legislation, they would still have to show why a less restrictive means to achieve them did not exist.

The court expressed itself to be acutely aware of the perils of drug abuse, and so was at pains to stress the narrow confines of its judgment. It maintained, on the other hand, that there were several options to fight this problem other than the ‘blunt use of the criminal law’. The impugned legislation was, as a result, disproportionate to the harms which the Legislature sought to curb. The quantity of cannabis in a person’s possession constitutes, said Davis J at [109], ‘an objective, established and readily enforceable basis upon which to distinguish possession for personal consumption from dealing or other, more serious conduct’, so it was both practical and objectively possible for future legislation to distinguish the two forms of use or possession.

The contested statutory provisions (ss 4(a) and 5(b) of the Drugs Act and s 22A(9)(a)(i) of the Medicines Act, read with their respective relevant schedules) were accordingly declared invalid, but ‘only to the extent that they prohibit the use of cannabis by an adult in a private dwelling where the possession, purchase or cultivation of cannabis is for personal consumption by an adult’ (at [132]).


 (b) Criminal Procedure and Evidence
 (i) Pre-sentence

ss 7 and 8: Private prosecution and s 6(2)(e) of the Societies for the Prevention of Cruelty to Animals Act 169 of 1993

National Society for the Prevention of Cruelty to Animals v Minister of Justice and Constitutional Development & others 2017 (1) SACR 284 (CC)

In the above Constitutional Court case (hereafter the ‘CC case’) it was declared that the National Society for the Prevention of Cruelty to Animals (hereafter the ‘NSPCA’) has the statutory power of private prosecution conferred upon it by s 6(2)(e) of the Societies for the Prevention of Cruelty to Animals Act 169 of 1993 (hereafter the ‘SPCA Act’) as read with s 8 of the Criminal Procedure Act 51 of 1977.

The above declaration of the Constitutional Court was preceded by decisions of the High Court and Supreme Court of Appeal. See National Society for the Prevention of Cruelty to Animals v Minister of Constitutional Development & another (unreported, GP case no 29677/2013, 8 October 2014) and National Society for the Prevention of Cruelty to Animals v Minister of Justice and Constitutional Development & another 2016 (1) SACR 308 (SCA). These two cases are discussed in the notes on s 7 in Commentary, sv Section 7(1)(a): Constitutionality of excluding juristic persons from the right to institute a private prosecution. In both cases the NSPCA’s constitutional challenge to s 7(1)(a) of Act 51 of 1977, on the basis that it permitted only a private person (and not a juristic person) to institute a private prosecution, was unsuccessful. In terms of s 7(1)(a) only private persons (natural persons) are allowed to institute a private prosecution—and then only in limited circumstances and after having obtained a certificate nolle prosequi from the Director of Public Prosecutions who has jurisdiction over the alleged offence.

In the CC case it was, however, ultimately not necessary to make a finding concerning the constitutionality or otherwise of s 7(1)(a). Writing for a unanimous Constitutional Court, Khampepe J pointed out that the NSPCA’s claim that it had a right privately to prosecute animal cruelty offences, had to be determined with reference to s 8 of Act 51 of 1977 and s 6(2)(e) of the SPCA Act, as read with the Animals Protection Act 71 of 1962 (hereafter the ‘AP Act’) which provides ‘the groundwork for the animal protection-regime’ (at [43]). Section 8 of Act 51 of 1977 provides for private prosecutions under statutory right: ‘Any body upon which or person upon whom the right to prosecute in respect of any offence is expressly conferred by law, may institute and conduct a prosecution in respect of such offence in any court competent to try that offence’ (emphasis added).

The crucial issue in the CC case was whether s 6(2)(e) of the NSPCA Act gave the NSPCA the statutory power to institute private prosecutions as envisaged in s 8 of Act 51 of 1977. Section 6(2)(e) permits the NSPCA ‘to institute legal proceedings connected with its functions...’. At [36] Khampepe J noted that this phrase in s 6(2)(e) can include the power to prosecute privately because ‘the language is broad and permissive...[and]...does not distinguish between civil and criminal proceedings’. The SPCA Act, it was said at [38], had to be read with the AP Act. Sections 2 and 2A of the latter Act contain an extensive list of offences that constitute animal cruelty. It was therefore clear that legal proceedings stemming from the AP Act ‘are most likely to be criminal’ (at [46]). Khampepe J concluded as follows (at [47]–[48], emphasis added):

‘Functionally, the NSPCA is best placed to conduct a private prosecution and give effect to preventing and enforcing the offences set out in the animal-protection regime. To understand the SPCA Act as conferring the power of private prosecution is to give effect to the objects and purposes of the regime. This effect to the NSPCA’s primary purpose: to protect animal welfare.... To read s 6(2)(e) as excluding the right of private prosecution would render the regime a toothless tiger.... The term ‘‘institute legal proceedings’’ takes on a specific and nuanced meaning in this context ...including the power to institute private prosecutions.’

However, the fact that the NSPCA is now considered a ‘body’ with the right to prosecute privately as provided for in s 8(1) of Act 51 of 1977, does not mean that the NSPCA’s exercise of this power is beyond the control or oversight of the Director of Public Prosecutions (DPP) concerned. The NSPCA, for example, may only exercise its right after consultation with the DPP and after the latter has withdrawn his or her right to prosecute. See s 8(2) of Act 51 of 1977. In withdrawing the right to prosecute, the DPP may set conditions he or she deems fit, including a condition that the NSPCA’s appointment of a prosecutor shall be subject to the DPP’s approval. See s 8(3) of Act 51 of 1977. Another possible condition mentioned in s 8(3) is that the DPP ‘may at any time exercise with reference to any such prosecution any power which he might have exercised if he had not withdrawn his right of prosecution’.

It must, finally, be pointed out that the CC specifically refrained from determining the constitutionality of s 7(1)(a). The matter was no longer ‘a live dispute’ (at [63]) since the NSPCA’s right to prosecute privately was decided without relying on s 7(1)(a). The result is that the Supreme Court of Appeal’s decision as referred to earlier must stand, namely that the exclusion of juristic persons in s 7(1)(a) is not unconstitutional. However, Khampepe J made it clear that nothing in the CC judgment ‘should be construed as barring a future challenge to that provision, if the appropriate factual scenario arises’ (at [64]).

s 60: Bail: Right to a fair bail hearing

S v Kobese (unreported, ECG case no CA&R 34/2017, 23 February 2017)

In the above matter the appellant appealed unsuccessfully against the refusal of a magistrate in the Grahamstown district to release him on bail pending the outcome of his trial on charges of having raped a 14-year-old girl on three separate occasions and having assaulted another child with intent to do grievous bodily harm. At the bail hearing and on appeal it was common cause that on account of the rape charges, the burden of proof was on the accused as provided for in s 60(11) of Act 51 of 1977. On the nature of the bail applicant’s burden in this regard, see the discussion of s 60 in Commentary, sv Section 60(11) and the burden of proof.

In Kobese Malusi J noted, with reference to s 65(4), that a court hearing an appeal against a refusal of bail will not set aside the bail court’s refusal ‘unless...satisfied that the decision was wrong’ (at [10]). See also the discussion of s 65 in Commentary, sv Approach on appeal: s 65(4). Indeed, in Kobese it was clear that the appellant had failed to discharge his burden of proving on a balance of probabilities that there were exceptional circumstances which in the interests of justice permitted his release (at [14], [17] and [19]).

However, in dismissing the appeal (at [23.1]), Malusi J also made an order requiring the Registrar to ensure that a copy of the High Court’s judgment in the bail appeal be delivered to the Chief Magistrate, Grahamstown, so that the latter could—within a reasonable time from delivery of the High Court judgment—facilitate ‘sensitivity training’ for the magistrate who had presided at the bail hearing and had refused bail (at [23.2]).

The above order concerning ‘sensitivity training’ was a consequence of certain unacceptable remarks or observations made by the bail magistrate. In her ex tempore judgment the bail magistrate, having stated that in the Grahamstown community ‘children are abandoned or raped by their parents’, asked rhetorically whether it would be safe if the appellant were released on bail in this ‘situation’. The magistrate claimed that her remarks were based on her four years residence in Grahamstown as well as her experience in the Children’s Court. However, Malusi J pointed out that her remarks were ‘unwarranted’ and that there was ‘no evidence of such a situation nor...[was] such a notorious fact for her to take judicial notice of it’ (at [20]).

In her ex tempore judgment the bail magistrate also stated that ‘we are living in a country...[where]...we who are animals...’. According to Malusi J this observation was ‘[t]he most egregious comment by the magistrate’ and amounted to gender stereotyping that could ‘well be in breach of the equality clause in section 9(3) of the Constitution’ (at [21]).

The order regarding sensitivity training was absolutely necessary. However, it would appear that there is ample room for an argument that the bail magistrate’s unacceptable and objectionable remarks in the course of her ex tempore judgment not only fatally tainted her impartiality but also constituted a gross irregularity that was of such a fundamental nature that there was a failure of justice which rendered the bail proceedings a nullity. See, for example, S v Sibeko en ’n ander 1990 (1) SACR 206 (T). In Sibeko a bail magistrate had earlier during the police investigation taken down a confession by one of the bail applicants. It was held that the magistrate could not thereafter have presided at the bail hearing. The irregularity committed was of such a nature that the bail proceedings were declared a nullity. It was fundamental to the administration of justice, said Kriegler J in Sibeko at 207i–j, that a presiding judicial officer should be an impartial and open-minded (‘onbevange’) adjudicator who decides matters on the basis of information placed before him or her in the usual way.

It is furthermore submitted that the bail hearing presided over by the magistrate in Kobese fell well short of what was said in S v Majali (unreported, GSJ case no 41210/2010, 19 July 2011) at [33]: ‘A bail inquiry is a judicial process that has to be conducted impartially and judicially and in accordance with relevant statutory and constitutional prescripts.’ It must also be appreciated that the constitutional right to a fair trial includes proceedings such as bail proceedings. Fairness and the constitutional right to a fair bail hearing are manifestly under threat where a bail magistrate—like the one in Kobese—compromises her impartiality by making the remarks and observations which she did in the course of an ex tempore judgment dealing with the question whether a male bail applicant charged with the rape of a young girl should be released on bail in the district of Grahamstown. What our law requires is that a bail magistrate, like any judicial officer presiding over a trial, should conduct proceedings open-mindedly, impartially and fairly, and that such conduct must indeed be manifest to all concerned, especially the bail applicant. See the instructive pronouncements made by Ponnan JA in S v Le Grange & others 2009 (1) SACR 125 (SCA) at [14] as regards trial proceedings. At [18] Ponnan JA also made the following telling observation: ‘A judicial officer can only perform his demanding and socially important duty properly if he also stands guard over himself, mindful of his own...personal views.’

Regrettably, it would seem that in Kobese the bail magistrate’s personal views on rapes of children in the Grahamstown district as well as her views on the conduct of some men in South Africa, had tainted her impartiality to the extent that the bail applicant’s right to a fair bail hearing was fatally compromised. The bail proceedings should have been declared a nullity, accompanied by an order—like the one in Sibeko (supra) at 208c—that a fresh bail application be heard by another magistrate on a specific day shortly after the date of the order of nullity.

The fact that there were sufficient grounds in Kobese for the refusal of bail, need not necessarily be an obstacle to a declaration of nullity. An objective test comes into play. It is well known that a vital element of a fair trial is that justice must not only be done, but must also be seen to be done. In the context of Kobese this would mean that a reasonable and informed person’s perception of bias on the part of the bail magistrate, would have the same impact as actual bias in that the proceedings are vitiated whether the bias is perceived or real. On the nature of the objective test in assessing the impact of alleged bias on fair trial principles, see generally the discussion of s 145 in Commentary, sv Guidelines for the recusal of judicial officers (including assessors).

Kobese (supra) calls for one more comment. Our courts, and especially the Supreme Court of Appeal, have on numerous occasions for purposes of sentencing noted that rape has become prevalent not only in the specific jurisdiction of the court concerned but countrywide as well. See S v Mgibelo 2013 (2) SACR 559 (GSJ) at [4]; S v SMM 2013 (2) SACR 292 (SCA) at [14]; S v Nkunkuma & others 2014 (2) SACR 168 (SCA); S v Mhlongo 2016 (2) SACR 611 (SCA) at [25]. Indeed, in S v Hewitt 2017 (1) SACR 309 (SCA) it was noted for purposes of sentencing that Interpol has identified South Africa as ‘the rape capital of the world’ (at [9]). However, there is a vast and fundamental difference between the situation where a sentencing court notes the prevalence of rape for purposes of determining a suitable sentence for a convicted rapist and the situation where a bail court which has to decide on the pre-trial liberty of an alleged rapist, makes observations like the ones made by the bail magistrate in Kobese (supra).

It is, in the final analysis, submitted that Malusi J’s order regarding sensitivity training for the bail magistrate in Kobese was appropriate for purposes of assisting her to realise that judicial duties must be performed with sensitivity and circumspection. But this order only addressed the future conduct of the magistrate. It is respectfully submitted that the court of appeal in Kobese should also have addressed a broader but very relevant issue, namely the impact of the bail magistrate’s comments on the appellant’s constitutional right to a fair bail hearing. In fact, at [20] it was noted that the comments of the bail magistrate could have caused accused persons appearing before her to experience ‘a sense of injustice’. Must one assume that the bail applicant himself had not experienced such a sense of injustice? Or that the reasonable person in the position of the bail applicant would not have experienced such a sense of injustice? It is respectfully submitted that the appeal court in Kobese should also have examined and answered the following question: did the comments of the magistrate constitute a fatal irregularity that nullified the bail proceedings?
ss 151 and 166: Onus of proof—implications of. Judicial officer descending into the arena—how to deal with inexperienced prosecution

S v Quayiso 2017 (1) SACR 470 (ECB)

The accused had been charged with two counts of assault with the intention to cause grievous bodily harm arising out of an incident in which he was alleged to have assaulted two men by stabbing them with a knife. The men had allegedly robbed him of his cell phone the previous day. The magistrate found the evidence of the second complainant to be unreliable and acquitted the accused on the second count. On the first count, the magistrate convicted the accused, even though she accepted that his version that he had acted in self-defence was reasonably possibly true, on the basis that he had exceeded the bounds of self-defence by causing excessive harm to the first complainant.

This, said Mbenenge J (with whom Van Zyl J agreed), the magistrate was not entitled to do. Having accepted the version of the accused—that he had acted in lawful self-defence—she was not entitled to reject it on the basis that the accused had exceeded the bounds of that defence. That was never the case of the prosecution in the first place and it flew in the face of the accused’s version (which she accepted) that he had been threatened by the complainants and their companions and even received a stab wound in his thigh from one of them.

Mbenenge J cited this helpful and important test laid down in S v Sithole & others 1999 (1) SACR 585 (W) at 590g–i:

‘There is only one test in a criminal case, and that is whether the evidence establishes the guilt of the accused beyond reasonable doubt. The corollary is that an accused is entitled to be acquitted if there is a reasonable possibility that an innocent explanation which he has proffered might be true. These are not two independent tests, but rather the statement of one test, viewed from two perspectives. In order to convict, there must be no reasonable doubt that the evidence implicating the accused is true, which can only be so if there is at the same time no reasonable possibility that the evidence exculpating him is not true. The two conclusions go hand in hand, each one being the corollary of the other. Thus in order for there to be a reasonable possibility that an innocent explanation which has been proffered by the accused might be true, there must at the same time be a reasonable possibility that the evidence which implicates him might be false or mistaken.’

The propositions stated in Sithole may be axiomatic, but it is remarkable how often they are overlooked or wrongly applied. It may be helpful, then, to set out this further explanation by Nugent and Schwartzman JJ in Sithole (at 590–1), which was not repeated in Quayiso:

‘Whichever way one phrases the test, it is to be applied upon an assessment of all the evidence, and not by a process of piecemeal reasoning. In other words, it cannot be applied by looking only at the evidence of the State, or the accused, in isolation. It may be that the evidence of the State is such that the conflicting evidence of the accused must, by a process of logical reasoning, be untrue, or it may be that the evidence of the accused is such that the possibility that it may be true cannot be said to be excluded by the State’s evidence; but in either event a court must bear in mind all the evidence when reaching the appropriate conclusion. Thus, for example, where there is evidence that the accused committed a particular act, and his evidence is that he was not present at the time the act was committed, quite obviously the two assertions cannot both be true. There cannot even be a reasonable possibility that they are both true. There can only be a reasonable possibility that the accused’s alibi is true if there is at the same time a reasonable possibility that the State’s evidence is false or mistaken.’

This pronouncement encapsulates the correct position with admirable clarity.

Apart from the erroneous application of the law relating to the onus, the court in Quayiso had other anxieties with the approach of the magistrate.

First, a number of questions were put to the accused about his version after he had been cross-examined by the prosecutor. The questions were ‘hardly questions to get clarity on certain unclear issues, but constituted cross-examination, with the result that the magistrate descended into the arena, transgressing the guidelines stipulated in S v Rall [1982 (1) SA 828 (A) at 831–2]’ (at [25]). Rall’s case and several others dealing with this issue are discussed in Commentary in the notes to s 166 sv Questioning by the court.

The impression that the magistrate had descended into the arena was further strengthened by her unjustified interference in the cross-examination conducted by the defence. The stabbing incident had been triggered by the alleged robbery involving the cell phone. The magistrate disallowed cross-examination of the first complainant on the issue of the robbery on the basis that the case was not about a cell phone, but an assault. This was clearly unwarranted, and by so acting she had, said the court (at [23]), ‘descended into the arena and associated herself with the prosecution, thus contaminating the proceedings by not acting in an open-minded and impartial fashion’.

The ‘magistrate’s lopsided approach’, said Mbenenge J (at [26]), ‘once again reared its ugly head’ when she thwarted his attempt to show the court the stab wound he had sustained to his thigh. Since the accused claimed to have been acting in self-defence, this was an important part of his case. The magistrate’s refusal was a sign that she ‘favoured the State in a way that could ‘‘only lead to the administration of justice falling into disrepute and a perception of bias’’ on her part’ (following Claasen J in S v Du Plessis 2012 (2) SACR 247 (GSJ) at [25]).

The final irregularity concerned the re-opening of the State’s case, which had been done improperly after the closure of the case for the State, with no formal application having been made to re-open the case, and without an opportunity being afforded to the defence to cross-examine the recalled witness, the second complainant. The State’s case was re-opened because there was no proof that the witness had been stabbed, as the medical form (J88) was not in the docket. The witness was recalled to show the court the stab wound to his shoulder, a ‘benevolence ... not extended to the accused’ (at [30]), who was not permitted to show the court the wound to his thigh.

To make matters worse, the magistrate seemed to shift the blame for this irregularity to the ‘inexperience of aspirant prosecutor’ (at [28]), as she put it in a response to a query by the regional magistrate. This response, said Mbenenge J, left ‘much to be desired’. It remained the duty of magistrates to ensure that proper procedures were followed at all stages in a trial. There was a ‘duty cast on the magistrate to guide the inexperienced prosecutor where there were indications of him floundering, as she is not merely a figurehead’. Her duty was ‘not only to direct and control the proceedings according to recognised rules of procedure, but to see that justice [was] done’ (at [29], the court relying on R v Hepworth 1928 AD 265 at 277, followed in S v Rall supra and S v Musiker 2013 (1) SACR 517 (SCA) at [18]).

In the trial justice was conspicuously not done; the proceedings were vitiated and the conviction was set aside.

s 152: Reconstruction of the record: Can the accused waive his or her rights to participate in the process?

S v Schoombee & another 2017 (5) BCLR 572 (CC)

Whether an accused can waive his or her constitutional rights in respect of the proper procedures to be followed in reconstructing a lost record of the trial was considered, but not answered, by the Constitutional Court in this case. The applicants in Schoombee, when they had sought to appeal against their convictions for murder, discovered after much searching that the record of their trial proceedings had been lost. A ‘reconstructed record’ had been prepared by the trial judge on the basis of the notes he had taken during the proceedings. After some hesitation, both applicants proceeded to appeal on this record—one against conviction and sentence, the other only against sentence. Both appeals were dismissed by the full court, before which neither applicant raised the issue of the inadequate record.

It was contended before the Constitutional Court that the criminal process was flawed, the inadequate reconstruction of their trial leading to a violation of their constitutional right to a fair trial. Without a proper trial record, it was argued, there could be no appeal; with no appeal, there could be no fair trial. The decision to proceed with the appeal using the reconstructed record, they contended, was taken out of desperation after years of delay, and after being advised by counsel that they had reasonable prospects of success on the record as it stood.

The Constitutional Court set out various aspects of the law relating to the relation between the reconstruction procedure and a fair trial, the methods that may be employed in reconstructing a record and the responsibilities of the State, the accused and the court in engaging in such a process (as to which see the notes on s 152 of Commentary, as well as the notes on ss 158 and 309). The appellant (or legal representative) ‘carries the final responsibility to ensure that the appeal record is in order’ (at [21], quoting from S v Sibelelwana [2012] ZAWCHC 150 (unreported, WCC case no A401/2011, 3 August 2012) at 10). At the same time, said the court, it was for the reviewing court ‘to ensure that an accused is guaranteed the right to a fair trial, including an adequate record on appeal, particularly where an irregularity is apparent’ (see, too, S v Zuma 1995 (2) SA 642 (CC) at [16]).

In this case, by choosing not to pursue a reconstruction process, could it be said that the applicants had waived their right to a fair trial regarding their participation in such a process? The State argued that their decision to launch their appeal using the ‘judge’s imperfectly reconstructed’ record was such a waiver. Was it? This was the Constitutional Court’s answer (at [25]):

‘Perhaps. This Court has emphasised that waiver of a constitutional right is difficult. The bar is high. To waive a right, a party must intentionally and knowingly abandon it. The onus to prove waiver is strictly on the party asserting it—here, the State. Even so, this Court has questioned whether waiver is applicable in relation to constitutional rights. And it has noted the distinction between waiver in the contractual sense and a mere choice not to exercise a constitutional right.’

Although the facts here did ‘suggest a possible waiver’ (at [26]), the court found it unnecessary, on the facts, to determine either whether the applicants did waive their rights or whether it was even possible for them to have done so. This was because the applicants had a fair trial, including a fair appeal. Although the record of their trial had been improperly and imperfectly reconstructed, it was ‘more than adequate to ensure the applicants exercised their constitutional right to appeal’ (at [27]): the notes taken by the trial judge were ‘unusually full and detailed’ and were not ‘scrappy, telegram-style annotations’; they appeared to be a complete narrative account of the evidence led in the trial; they recorded the witnesses’ evidence-in-chief as well as their cross-examination; and they gave a full picture of both the events on the night of the murder as well as the trial proceedings themselves.

As the Supreme Court of Appeal held in S v Chabedi 2005 (1) SACR 415 (SCA), a reconstructed record need not be perfect; it need only be adequate for a proper consideration of the appeal. Whether or not it is so is a question that ‘cannot be answered in the abstract’, but depends, inter alia, on ‘the nature of the defects in the particular record and on the nature of the issues to be decided on appeal’ (Chabedi at [5]–[6]). Thus, said the Constitutional Court (at [29]), it was ‘practical and sensible and just’, where the adjudication of an appeal on an imperfect record would not prejudice the appellants, not to set aside their convictions ‘solely on the basis of an error or omission in the record or an improper reconstruction process’ (see, too, S v Machaba 2016 (1) SACR 1 (SCA) at [4]–[5]; see, too, Commentary in the notes to s 152 sv Reconstruction of the record).

These principles, said the court, applied in Schoombee’s case for these reasons: the reconstruction was detailed and specific; the applicants reviewed this record; they took the advice of counsel; they chose to proceed with their appeal on the imperfect record in accordance with counsel’s advice; and, even if they did not waive their right to participate in reconstruction, they ‘certainly signified their assent to the substantive recital contained in the reconstructed record’ (at [30]). The court warned, however, that ‘[n]one of this detracts from the magnitude of the lapses that took place in constructing the record’ (at [38]). It stressed that the High Court had failed to ensure, as it was obliged to do, that the reconstruction process involved both parties. The duty to ensure that the process complied with the right to a fair trial was one that had to be ‘undertaken scrupulously and meticulously in the interests of criminal accused as well as their victims’.

It noted, finally, that the loss of trial court records was a ‘widespread problem’ which ‘raised serious concerns about endemic violations of the right to appeal’. Reconstruction, it lamented, ‘should not be the norm in providing appellants with their trial records’.

s 208: Identification evidence guidelines and pitfalls

S v Kotze [2017] ZASCA 27 (unreported, SCA case no 776/16, 27 March 2017)

S v Madondo (unreported, KZP case no AR 350/2016, 8 December 2016)

‘Authorities dealing with the dangers of incorrect identification’, said Mocumie JA in Kotze, ‘are legion’. This is true, and it is instructive to consider how the court in that case navigated through those authorities in determining whether the complainant had made an accurate and reliable identification of the appellant.

The appellant had been convicted of housebreaking with intent to commit indecent assault, as well as indecent assault itself. Much turned on the complainant, a 13-year-old girl. The trial court and the full court found the complainant to be an honest and impressive witness who had no reason falsely to implicate the appellant. In the appeal to the Supreme Court of Appeal, the accuracy of the identification was attacked.

Mocumie JA, who delivered the judgment of the court, referred to the ‘locus classicus’ in this regard, S v Mthetwa 1972 (3) SA 766 (A) at 768A, where the court warned that ‘because of the fallibility of human observation, evidence of identification is approached by the Courts with some caution’. Reference was also made to this passage in the court’s judgment in R v Dladla & others 1962 (1) SA 307 (A) at 310C–E (see, too, S v Arendse [2015] ZASCA 131 (unreported, SCA case no 089/15, 28 September 2015)):

‘One of the factors which in our view is of the greatest importance in a case of identification is the witness’ previous knowledge of the person sought to be identified. If the witness knows the person well or has seen him frequently before, the probability that his identification will be accurate is substantially increased. Even in the case when a witness has some difficulty in the witness-box in giving an accurate description of the facial characteristics and clothes of the person whom he has identified, the very fact that he knows him provides him with a picture of the person in the round which is a summary of all his observations of the person’s physiognomy, physique and gait, and this fact will greatly heighten the probability of an accurate identification.... In a case where the witness has known the person previously, questions of identification marks, of facial characteristics, and of clothing are in our view of much less importance than in cases where there was no previous acquaintance with the person sought to be identified. What is important is to test the degree of previous knowledge and the opportunity for a correct identification, having regard to the circumstances in which it was made.’

In Kotze’s case there were circumstances that enhanced the reliability of the identification. First, there was unrefuted evidence that the complainant knew the appellant prior to the incident. There was also unchallenged evidence that of all the people who visited at the house in question on the evening of the assault, the complainant knew only the appellant.

Secondly, the appellant was able to identify the appellant on three separate occasions on the night in question. The first was when she saw him inside her bedroom the first time, after he had switched on the light. On that occasion there was instant recognition. The second was when he returned to that room to assault her. On that occasion she called him by his name and, on the strength of his silhouette, correctly described his physique as ‘tall and chubby’. He had, she said, smelled of liquor, and it was common cause that the appellant was intoxicated that night. It was clear, further, that she could not have known this unless her evidence was true. The third occasion was when she was hiding in the garden after she had fled from her attacker. She saw him leave the main house through the kitchen door and described him as being unsteady on his feet, adding that he then walked to a nearby flatlet and sat on a bench next to the sliding door. The light at the back door was on and so was the light at the door of the flatlet. The appellant was later found asleep on this bench and it was clear from the evidence that the complainant could never have known that the appellant had gone to sit on the bench unless she had seen him.

On the strength of this evidence and this rigorous analysis of that evidence, the court accepted that the complainant’s identification of the appellant as the person who assaulted her was ‘undoubtedly both honest and reliable’ (at [20]).

A different conclusion was reached by the court in S v Madondo (unreported, KZP case no AR350/2016, 8 December 2016). The appellant’s conviction for robbery with aggravating circumstances in that case depended upon the identification of him as the robber by the complainant. There were several problems relating to this identification: the complainant did not know the appellant or his name; he had not seen the faces of his two assailants clearly; the only illumination was provided by street lights some eight to nine metres away; it was already dark when the attack occurred; the event was a very emotional and traumatising one, which ‘would no doubt have affected his powers of observation and recall’ (at [5]); and both the estimation by the appellant of the time of the event and, as a result, the opportunity to observe, were open to doubt since the ‘harrowing nature thereof probably made it feel much longer than it actually was’.

But added to these concerns were two extremely troubling features of the complainant’s account. The first was that when he was taken to the home of the appellant by a co-accused, he claimed to have recognised the appellant as the one who had robbed him ‘by his dreadlocks as well as his complexion’. The problem was that he had not provided this description to any law enforcement agency or even to any of the many members of the community who had been present at his home before this identification. The second was that the co-accused had taken him to the appellant’s home in the first place. He had not taken him there freely and voluntarily, and it seems that this was the result of the unfortunate conduct of members of the community—numbering between 20 and 30—who had decided to take the investigation into their own hands and had assaulted the co-accused before he agreed to lead them to the house ‘of the one who had been with him when they committed the robbery’ (at [7]).

These circumstances raised the ‘real danger that the seemingly positive identification of the appellant might have been caused or influenced (possibly exclusively) by the conduct of [the co-accused] leading the complainant and the others to the appellant as a co-assailant’ (at [9]). The position, said Koen J (with whom Chetty J agreed), would obviously have been different if the complainant had, right from the outset, mentioned the identifying features to the members of his community.

As for the statement made by the co-accused to the crowd following the assault, it was obviously inadmissible against the appellant. This was so, apart from the fact that it was not made freely and voluntarily, because it constituted an extra-curial admission which could, in line with what was held in S v Litako & others 2014 (2) SACR 431 (SCA), not be admitted against a co-accused (see the discussion in Commentary in the notes to s 219 sv Is an admission (as opposed to a confession) by an accused admissible against a co-accused?)

There was, as a result, a reasonable possibility that the complainant had mistakenly identified the appellant. The conviction was set aside.

ss 217, 219, 166: Admissibility of confessions under the Criminal Procedure Act and the Constitution

S v Mudau [2017] ZASCA 34 (unreported, SCA case no 1148/2016, 29 March 2017)

The trial of the two appellants, in which they were convicted of murder, assault with intent to commit grievous bodily harm and robbery with aggravating circumstances, contained so many serious misdirections, gross procedural irregularities, infringements of the statutory provisions governing confessions and violations of their constitutional right to a fair trial, that the Supreme Court of Appeal had no option but to exclude the confessions in question and to set aside the convictions.

The recording of the confessions in writing by the magistrate was ‘replete with omissions, incoherent and contradictory recording of answers by the appellants to questions, and serious non-adherence to some of the fundamental principles governing confessions’ (at [15]). Some of these irregularities were:

(1)   In the first appellant’s confession form, the paragraph setting out his constitutional right to remain silent and the right to legal representation, was left blank and not completed. So, too, was the paragraph which questioned whether the deponent was in his sound and sober senses and the paragraph which inquired whether the accused wished to make a statement notwithstanding what had been explained to him.

(2)  When questioned about a further blank paragraph, the magistrate, in his viva voce evidence, conceded that the first appellant had undoubtedly answered in the negative to the question whether or not he wished to make a statement. In the form itself, he said expressly that he did not wish or prefer to make a statement. The fact that the magistrate still proceeded to record the confession and the fact that the trial court ruled the confession to be admissible ‘resulted in a serious miscarriage of justice and rendered unfair the first appellant’s trial’ (at [19]), per Mbha JA, who delivered the judgment of the court).

(3)   The trial court also ‘inexplicably’ ignored or overlooked the testimony of a police inspector who conceded that he ‘duly advised the first appellant of the advantages of making a confession, namely, that such confession was going to allow the matter to proceed quickly’ (at [19]). This, said Mbha JA, amounted to undue influence.

(4)    The statement of the first appellant was, moreover, wrongly classified as a ‘confession’. He claimed that he had acted under the coercion of the second appellant, with the result that the statement did not constitute an unequivocal admission of every element of the offence.

(5)   The form relating to the second appellant’s confession also showed material omissions. The magistrate omitted, for instance, to record the answer given to the question whether he wished to make a statement and whether or not he had been assaulted or coerced into making a statement. The magistrate also failed to record his ‘observation whether the second appellant was in his sound and sober senses with specific reference to anxiety, nervousness, joviality and demeanour’. These qualities were relevant since the appellant claimed in his testimony that, when he was brought to the magistrate to make a confession, he was ‘dizzy because he felt he was being forced to admit an offence which he did not commit’ (at [22]). Even more disturbing, in the view of Mbha JA, was the answer given by the second appellant to the question of how it came about that the he was brought to make the confession. The answer—that he had been advised by the police inspector to report to the magistrate —was simply recorded by the magistrate, who should have seen this answer as a ‘red flag’ and followed up by asking questions to ensure that the appellant was not influenced in any way into making the confession.

Mbha JA then turned to the irregularities committed by the trial judge:

(1)  The trial judge had made an erroneous ruling on the onus of proof by maintaining, at the commencement of the trial within a trial, that the appellants had first to adduce evidence to prove that the confessions were not freely and voluntarily made or were not made without undue influence. The onus, as Mbha JA pointed out, rests always on the State, and the ‘erroneous shifting of the onus of the appellants rendered their trial unfair’ (at [23]).

(2)   The trial judge ignored the fact that the same magistrate recorded both appellants’ confessions, which were made a week apart. There was nothing in the record to indicate that he made any attempt to check whether another magistrate was available to take down the second confession, and it was ‘doubtful’ whether he had sufficiently cautioned himself not to be influenced by his prior knowledge of the first confession.

(3)   The trial judge applied the confessions and admissions made by the (then) three accused against one another in clear violation of s 219 of the Criminal Procedure Act (see Commentary on s 219). Instead of treating each confession separately against its specific maker, he treated all of them ‘in blanket fashion against all the accused’ (at [28]).

(4)   In addition to these irregularities in relation to the confessions, there was one further irregularity that sustained the conclusion that the judge ‘was not open-minded, impartial and fair during the second appellant’s trial within a trial’ (at [27]). After the second appellant had given his evidence-in-chief at the trial within a trial, and before the prosecutor could cross-examine him, the trial judge said to the prosecutor:

‘I do not know if you would like to cross-examine this witness. He made a very bad impression. I do not know, as a witness really. I do not know whether it would be worth for you to cross-examine him, but you can have that time. But he made a bad impression, really, I do not know. Do you have any questions for the witness?’

This ‘unfortunate remark’ showed that the trial judge had already decided that he was not going to accept the second appellant’s evidence, and, said Mbha JA, ‘a perception of bias on the part of the trial judge by any reasonable person was ... inevitable’ (at [25]). See, further, S v Basson 2007 (3) SA 582 (CC) at [27], Van Rooyen & others v The State & others (General Council of the Bar of South Africa intervening) 2002 (5) SA 246 (CC) at [32]—which followed what was said by the Canadian Supreme Court in Valente v The Queen [1985] 2 SCR 673—and S v Le Grange & others 2009 (1) SACR 125 (SCA) at [16]. See also, on the need for the judicial officer to avoid descending into the arena, Commentary on s 166 sv Questioning by the court.

 (ii) Sentencing

Sentencing and trial fairness: No reference in charge-sheet to minimum sentence legislation

S v Tshoga 2017 (1) SACR 420 (SCA)

Tshoga (supra) was an appeal against life imprisonment imposed by the High Court on the appellant for the rape of a 10-year-old girl. This sentence was—in the absence of substantial and compelling circumstances—the prescribed sentence in terms of s 52(1) as read with s 51(1) and Part I of Schedule 2 to the Criminal Law Amendment Act 105 of 1997 (at [1] and [5]).

One of the main grounds of appeal was that the appellant was at no stage prior to conviction warned that he faced a possible sentence of life imprisonment in terms of s 51(1) of Act 105 of 1997 (at [8]). This, it was argued, constituted an unfair trial.

The following facts were common cause: the charge-sheet did not refer to Act 105 of 1997; at no stage prior to conviction did the trial magistrate or prosecutor refer to Act 105 of 1997; the first time the magistrate mentioned s 51 of the Act was after conviction when he informed the appellant that having been convicted of the rape of a girl under the age of 16 years, it was in terms of s 51 of Act 105 of 1997 required that the appellant’s case be referred to the High Court for purposes of sentence (at [9]). It should be noted that at the time of the appellant’s trial this was indeed the procedure that was prescribed by Act 105 of 1997. See generally the discussion under the repealed s 277 of Act 51 of 1977 in Commentary, sv Minimum sentences for certain serious offences: A brief overview of Act 105 of 1997 as amended by Act 38 of 2007.

In Tshoga (supra) the majority (per Schoeman AJA, Dambuza JA and Nicholls AJA concurring) found that the appellant ‘did have a fair trial’ (at [26]) and had ‘suffered no prejudice, in the circumstances of the case’ (at [27]). The majority referred to an earlier decision of the court in S v Kolea 2013 (1) SACR 409 (SCA) where it was held that if a charge-sheet had incorrectly referred to s 51(2) instead of s 51(1) of Act 105 of 1997, it could not—in the absence of any prejudice to the accused—preclude a sentencing court from imposing the prescribed minimum sentence as if there was no incorrect reference in the charge-sheet. In Kolea a full bench held that the appellant had on account of at least three reasons suffered no prejudice: he had legal representation at his trial and ‘well knew of the charge he had to meet’ and that the prosecution ‘intended to rely on the minimum-sentencing regime created’ by Act 105 of 1997 (at [11]); it was never ‘demonstrated’ that the appellant would have conducted his defence differently ‘had the mistake not been made in the charge-sheet’ (at [14]); trial unfairness and prejudice were for the first time claimed before the Supreme Court of Appeal (at [15]).

The majority in Tshoga took care to point out that the facts in Tshoga were not comparable with those in Kolea: in Tshoga the charge-sheet made no reference to Act 105 of 1997 whereas the charge-sheet in Kolea did refer to that Act but contained an incorrect reference to s 51(2) instead of s 51(1) (see Tshoga at [20]). However, the majority in Tshoga at [22] was satisfied that the true principle applied in Kolea was that ‘substance was of paramount importance and that form was secondary’. At [22] in Tshoga Schoeman AJA also explained as follows:

‘I am of the view that a pronouncement that...[Act 105 of 1997]...had to be mentioned in the charge-sheet or at the outset of the trial would be elevating form above substance. Every case must be approached on its own facts and it is only after a diligent examination of all the facts that it can be decided whether an accused had a fair trial or not.’

The majority in Tshoga, relying on its own interpretation of the approach adopted in Kolea, concluded that the cumulative effect of all the circumstances was such that there was indeed no prejudice (at [23]). The appellant had pleaded not guilty to the rape charge, had ‘participated fully’ in the trial and had ‘effective legal representation throughout the trial until his conviction and the transfer to the High Court’. After his conviction there was no objection when the regional court ordered that the case, in terms of then-existing procedures, had to be transferred to the High Court for the latter to consider the applicability of prescribed minimum sentences, which included consideration of life imprisonment. After transfer to the High Court, no complaint of possible prejudice was raised despite four further proceedings (two sentencing procedures in the High Court and two appeals to the full bench of the High Court) during which the appellant was legally represented (at [23]). The appellant therefore had legal representation at all relevant stages, from the trial to the various High Court proceedings, and had never alleged prejudice.

The complaint of prejudice, said the majority at [23], was ‘only belatedly raised’ when the matter reached the Supreme Court of Appeal. At [24] it was pointed out, too, that appellant’s counsel could in argument before the court also not identify any prejudice the appellant had suffered as a result of ‘the failure to mention ...[Act 105 of 1997] the charge-sheet or at the outset of the trial’, except for the possibility of prejudice in that the appellant, if properly informed, might have tendered a plea of guilty. This possibility, observed the majority, was remote since the appellant—some eight years after the incident—had ‘still professed his innocence’ to a probation officer who had compiled a pre-sentence report for purposes of the second sentencing procedure in the High Court, where the accused had legal representation but chose not to testify in mitigation (at [6] and [24]).

The minority (per Bosielo JA, Tshiqi JA concurring) took the view that it was grossly unfair to inform an accused only after conviction of the applicability of such a ‘patently serious’ matter as ‘the applicability of the minimum-sentence legislation’ (at [47]). The trial in Tshoga, said the minority at [47], was ‘a trial by ambush, which is neither desirable nor permissible in a constitutional democracy underpinned by a Bill of Rights’ (at [47]). The minority considered that the prosecution’s failure to refer to Act 105 of 1997 in the charge-sheet, as well as its further failure to apply for the charge-sheet to be corrected as provided for in s 86(1) of the Criminal Procedure Act, inevitably caused prejudice to the accused and had rendered his trial unfair (at [46] and [52]).

At least two matters can be raised in an attempt to explain the differences between the minority and majority judgments.

First, the majority—unlike the minority—accepted for purposes of the merits of the case (that is, whether the appellant should have been convicted or acquitted on the charge of raping a 10-year-old girl) that the absence of any reference to Act 105 of 1997 could not have caused prejudice, especially since the accused—assisted by a competent legal representative—had participated fully in his trial. After conviction an awareness that the minimum-sentence legislation would for sentencing purposes be considered, was essential. But it was precisely at this stage that the magistrate had indeed informed the appellant of the applicability of such legislation, and the legal representative at the trial had raised no objection to the referral of the matter to the High Court for purposes of the latter’s consideration of the minimum-sentence legislation. The minority, however, took a firm stance that the appellant’s fair trial right was breached right from the start when he was required to plead to a charge that was incomplete and not later corrected. See Tshoga at [50] and [52].

Second, the majority as well as the minority accepted that the facts and issues in Tshoga were not covered by the decision of the full bench of the Supreme Court of Appeal in Kolea (supra). See Tshoga at [22] and [51]. However, the majority—unlike the minority—accepted that Kolea indicated that in assessing the presence or absence of prejudice, form should not be elevated to substance. It was this principle which ultimately formed the basis of the majority’s decision. And it is, with respect, on the basis of this principle that the majority’s decision should be supported.

s 276B: Procedural fairness where court considers a non-parole period

S v Ndlovu [2017] ZASCA 26 (unreported, SCA case no 925/2016, 27 March 2017)

Section 276B(1)(a) of Act 51 of 1977 provides that if a court sentences a convicted offender to imprisonment for two years or longer, the court may as part of this sentence fix a period during which the sentenced offender may not be released on parole. In terms of s 276B(1)(b) such period ‘shall be referred to as the non-parole period, and may not exceed two thirds of the term of imprisonment imposed or 25 years, whichever is the shorter’. On the origin and purpose of s 276B and its context in the parole system, where the executive as opposed to the judiciary should have the say, see Terblanche A Guide to Sentencing in South Africa 3 ed (2016) 259–60, Moses Parole in South Africa (2012) 40–3 and the discussion of s 276B in Commentary, sv General and sv The judiciary and the executive. See also the introductory part in the article ‘Perspectives on parole: Two recent decisions’ which appears in Section A of the present edition of CJR.

In S v Ndlovu [2017] ZASCA 26 (unreported, SCA case no 925/2016, 27 March 2017) the sentencing court had fixed a non-parole period in terms of s 276B (at [5]). On appeal it was common cause that the court never informed the appellant and the State of its intention to invoke the provisions of s 276B (at [12]). The prosecution and the State were also not invited by the sentencing court to make submissions as to whether a non-parole period should be fixed.

The basis for the appeal was that the sentencing court had misdirected itself when it fixed a non-parole period without first having invited the parties to make submissions on the matter (at [7]). Writing for a unanimous full bench, Mbatha AJA confirmed that a convicted offender is entitled to address the sentencing court when it considers the imposition of a non-parole period as part of a sentence (at [11]). See also S v Jimmale 2016 (2) SACR 691 (CC) at [16] and [21]. Indeed, a sentencing court should in terms of the decision in S v Stander 2012 (1) SACR 537 (SCA) at [22] give the parties an opportunity to address it on at least the following: should a non-parole period be ordered? And, if so, how long should the non-parole period be?

In Ndlovu (supra) it was concluded, in line with several other decisions, that a failure to afford the parties an opportunity to address the court was a misdirection by the sentencing court. See further the discussion of s 276B in Commentary, sv Procedural matters and the requirement that the court should give reasons’.

In Ndlovu (supra) the appellant also relied on another ground of appeal, namely that there were no exceptional circumstances which could have justified the court’s decision to fix a non-parole period (at [7]). In this regard it must be pointed out that the requirement that there should be ‘exceptional circumstances’ is not stipulated in s 276B, but stems from case law in which s 276B has been interpreted and applied. The Supreme Court of Appeal, for example, stated in Stander (supra) at [16] that an order in terms of s 276B should ‘only be made in exceptional circumstances, when there are facts before the sentencing court that would continue, after sentence, to result in a negative outcome for any future decision about parole’. See also S v Mhlongo (supra) at [5]–[7]; S v Motloung 2016 (2) SACR 243 (SCA) at [18] and further cases referred to in the discussion of s 276B in Commentary, sv Case law: Aggravating factors and the requirement that there must be exceptional circumstances. The reason why our courts insist that there should be exceptional circumstances before a court can fix a non-parole period must also be understood in the context of the following remarks by the Constitutional Court in S v Jimmale (supra) at [13]:

‘The order should be made only in exceptional circumstances, which can be established by investigation of salient facts, legal argument and sometimes further evidence upon which a decision for non-parole rests. In determining a non-parole period following punishment, a court in effect makes a prediction on what may well be inadequate information as regards the probable behaviour of the accused. Therefore, a need for caution arises because a proper evidential basis is required.’

In Ndlovu Mbatha AJA, having duly noted the approach as set out in Jimmale (supra), proceeded to point out that it was not clear from the judgment of the sentencing court what factors were considered for purposes of imposing the non-parole period (at [12]). This, in turn, also meant that exceptional circumstances were not identified. In this respect the sentencing court had also misdirected itself. Indeed in S v Mhlongo (supra) Mocumie JA—having noted that the right to a fair trial extends to the sentencing process (at [9])—also emphasised the importance of the principle that proper reasons must in the interests of justice be furnished whenever a non-parole period is determined (at [12], emphasis added):

‘Giving reasons for decisions is a longstanding and salutary practice that serves the interests of justice. Furthermore, it helps to show the rationale for the decision. Without reasons for a judgment on sentence, as is the case in this matter, in respect of the invocation of s 276B, such lack of reasons is highly prejudicial to the accused person. Thus the court a quo’s failure to state the rationale for its judgment is a vitiating factor.’

In Ndlovu (supra) Mbatha AJA accordingly had no problem in setting aside ‘the imposition of the non-parole period’ (at [13]).

In conclusion Mbatha AJA also identified another irregularity, namely that the non-parole period fixed by the sentencing court was in conflict with the provisions of s 276B(2). This section provides that where an accused who has been convicted of two or more sentences is sentenced to imprisonment and the court directs that prison sentences shall run concurrently, the court shall—subject to s 276B(1)(b)—‘fix the non-parole period in respect of the effective period of imprisonment’. The sentencing court in Ndlovu had indeed made a concurrency order: the sentence of 20 years’ imprisonment for murder was ordered to run concurrently with the sentence of 12 years’ imprisonment for robbery. The effective period of imprisonment was therefore 20 years. The sentencing court, however, had confined its non-parole period of 13 years to the sentence imposed for murder. This, said Mbatha AJA at [12], was another ‘misdirection’. See further S v Mthimkulu 2013 (2) SACR 89 (SCA) which is also discussed in the notes on s 276B in Commentary, sv The interpretation of s 276B(2).

s 280(2): Sentencing court’s discretion to order concurrent running of sentences

S v Zulu [2017] ZASCA 207 (unreported, SCA case no 226/2016, 21 December 2016) and S v Langa [2017] ZASCA 2, unreported SCA case no 640/2016, 23 February 2017)

Section 280(2) of Act 51 of 1977 provides a sentencing court with a judicial discretion to order that sentences consisting of imprisonment (or of correctional supervision) should run concurrently. See the discussion of s 280 in Commentary, sv Concurrent sentences. A concurrency order is a means of ensuring that the cumulative effect of two or more sentences is kept within acceptable levels. See Terblanche A Guide to Sentencing in South Africa 3 ed (2016) 200. In S v Chake 2016 (2) SACR 309 (FB) Murray J pointed out that a disregard for s 280 ‘may lead to inhumane and unfair sentences’ (at [8.6]).

Aspects concerning concurrency orders arose in two recent Supreme Court of Appeal decisions.

In S v Zulu [2017] ZASCA 207 (unreported, SCA case no 226/2016, 21 December 2016) the applicant was granted leave to appeal against sentence to a full bench of the High Court which had dismissed his petition for leave to appeal. At [6] Wallis JA pointed out that where various charges against an accused arose out of a ‘single criminal enterprise’, a sentencing court is required to consider concurrency.

In Zulu the regional court had convicted the appellant of two counts of robbery. On each count of robbery he was—in the absence of substantial and compelling circumstances justifying a lesser sentence—sentenced to 15 years’ imprisonment in accordance with the minimum sentence provisions in the Criminal Law Amendment Act 105 of 1997. Sentences imposed for firearm charges were ordered to run concurrently with the two sentences for robbery. However, the regional court made no concurrency order in respect of the two sentences for robbery, which meant that these two sentences had to be served consecutively—amounting to an effective 30 years’ imprisonment (at [2]).

In Zulu there was indeed, for purposes of sentencing, a single criminal enterprise. The appellant and his accomplice had at gunpoint robbed a man of his motor car and when this man and the woman with him had stepped out of the car, she was robbed of her handbag and its contents. In these circumstances the two charges of robbery, which resulted in two convictions, did not offend the rule against duplication of convictions. For an analysis of this rule and its guiding principles, see the discussion of s 83 in Commentary, sv The rule against duplication of convictions and sv Guiding principles to determine whether a duplication of convictions has occurred.

Wallis JA pointed out that whilst the appellant had been properly convicted of two counts of robbery arising ‘out of a single criminal enterprise’, the ‘problem’ to which this gave rise was that the two convictions were treated separately for sentencing purposes (at [5]). The reasons for requiring an order of concurrency were set out as follows by Wallis JA (at [6], emphasis added):

‘The result was that the chance fact that...[the woman]...was present and had her handbag stolen, was the only reason for an increase in sentence from 15 years to 30 years imprisonment. It should have been plain to the magistrate that this was irrational and resulted in a manifestly excessive sentence being imposed for a single criminal enterprise. The theft of the handbag added nothing to the moral culpability of...[the appellant]...and did not justify any significant increase in the sentence to be imposed upon him.’

Zulu (supra) should be compared with S v Langa [2017] ZASCA 2 (unreported, SCA case no 640/2016, 23 February 2017). In Langa the regional court had sentenced the appellant to ten years’ imprisonment for murder and four years’ imprisonment for attempted murder. It was specifically noted by the sentencing court that the sentences were not to run concurrently (at [1]). This meant that the appellant had to serve an effective 14 years.

One ground of the appeal against sentence was that the regional court should have exercised its s 280 discretion by making a concurrency order, since ‘the two offences were closely connected in space and time’ (at [13]).

Writing for a unanimous full bench, Potterill AJA said that while it was true that the murder and attempted murder were indeed ‘closely linked in time and space’, the evidence showed that the appellant ‘randomly’ shot ‘at patrons in a shebeen in an attempt to kill the deceased execution style’ (at [14]). ‘The appellant’, continued Potterill JA at [14], ‘thus intended not only to shoot the deceased, but foresaw that by randomly shooting in the shebeen, he could kill other patrons.’

The full bench concluded that on account of the above facts, the appellant was ‘not punished twice for the same actions’ and that the regional court had not misdirected itself ‘in not ordering that the sentences run concurrently’ (at [14]). In the court’s view the cumulative effect of 14 years’ imprisonment for murder and attempted murder was neither disturbingly inappropriate nor disproportionate to the offences. The regional court, it was concluded, had exercised its discretion judicially. There was no misdirection and there were no other grounds for interfering with the sentences on appeal.

It is submitted that there is no conflict between Langa and Zulu (supra). The essential principle is that an otherwise acceptable effective term of imprisonment is not rendered unacceptable simply because the offences concerned were ‘closely linked in space and time’ or ‘arose from ‘a single criminal enterprise’.


 (iii) Forfeiture and confiscation

s 20: Sections 48(1) and 50(1) of the Prevention of Organised Crime Act 121 of 1998: When is residential property an ‘instrumentality of an offence’ for the purpose of a forfeiture order?

Brooks & another v National Director of Public Prosecutions 2017 (1) SACR 701 (SCA)

The appellants, who were husband and wife, were the registered owners of residential immovable property which had been declared forfeit to the State in terms of s 50(1) of POCA on the basis that it was an ‘instrumentality of an offence’, illegal diamond dealing, as contemplated in POCA. Of the 19 transactions in question, ten took place at the property in the space of less than one year. The second appellant, the wife, had raised the so-called ‘innocent owner’ defence to the grant of the forfeiture order relating to the property, which was owned jointly by the spouses, who were married in community of property. It was accepted that she had no knowledge of the illegal diamond dealing on the residential property, which took place while she was at work. The court a quo therefore excluded her interest in the property from the forfeiture order, and ordered the curator bonis to pay her one half of the net proceeds from the sale of the property.

An appeal against this judgment drew a divided response in the Supreme Court of Appeal. For the minority, Schippers JA (Mocumie JA concurring) held that the property had correctly been held to have been an instrumentality of the offence. In the view of the minority judges the relationship between the use of the property and the commission of the offences was ‘neither tenuous nor remote’ (at [36]). The involvement of the property was ‘not merely incidental to the commission of the offences: it was put to use in a positive sense and was a means through which the crimes were committed’. The property was the base for a significant, organised and well-funded illicit diamond dealing business, and, but for their arrest, the first appellant (the husband) and the dealers would clearly have continued to use the property as their base. POCA defines ‘instrumentality of an offence’ as meaning, inter alia, ‘any property which is concerned in the commission or suspected commission of an offence’. The cases have made it clear (see Commentary on s 20 sv Seizure and forfeitures under the Prevention of Organised Crime Act 121 of 1998), however, that the definition must be restrictively construed and that the property has to ‘facilitate or make possible the commission of the offence in a real and substantial way’ (at [19]).

Schippers JA found that the evidence in support of instrumentality was compelling and that the NDPP had discharged the onus of establishing that fact on a balance of probabilities. He then turned to the issue of proportionality, which he described as an inquiry ‘aimed at balancing the constitutional imperative of law enforcement and combating crime and the seriousness of the offence, against the right not to be arbitrarily deprived of the property’ (at [39]). The balance, in his view, fell in favour of forfeiture, which in this case would serve the ‘broader societal purpose of: (a) deterring persons from using property to commit crime ... (b) eliminating or incapacitating some of the means by which crime may be committed; and (c) advancing the ends of justice by depriving those involved in crimes of the property concerned’ (at [43]): see NDPP v RO Cook Properties (Pty) Ltd; NDPP v 37 Gillespie Street Durban (Pty) Ltd & another; NDPP v Seevnarayan 2004 (2) SACR 208 (SCA), discussed in Commentary.

The majority, per Ponnan JA (with Willis and Zondi JJA concurring), disagreed. Ponnan JA assumed, without deciding, that the use to which the property was put did render it an instrumentality. But he did so only because this was not crucial in view of the conclusion he reached on proportionality, and even added that, because the commission of the crimes on the property was ‘purely incidental to their commission’ (at [61]), it was ‘difficult ... to see how property bearing no reasonably direct connection to the offences other than serving as the location for the transactions could constitute an instrumentality of those offences’.

The Bill of Rights, said Ponnan JA (at [62]), provides in s 25(1) that ‘no law may permit arbitrary deprivation of property’. The Constitutional Court has stressed that there must be sufficient reason for the deprivation of assets by forfeiture and that the nature of the relationship between the means and ends to satisfy the rationality requirement of s 25(1) depends both on the nature of the affected property and the extent of the deprivation (see Mkontwana v Nelson Mandela Metropolitan Municipality & others 2005 (1) SA 530 (CC) at [35]). Thus, although s 50(1) of POCA is couched in peremptory terms by providing that a court ‘shall’ make a forfeiture order if it finds on the civil standard of proof that the property in question is an instrumentality of an offence, our courts have consistently interpreted ‘shall’ to mean ‘may’ (see Mohunram & another v NDPP & another (Law Review Project as Amicus Curiae) 2007 (4) SA 222 (CC) at [121]). Accordingly, said Ponnan JA (at [63]), the courts have developed the notion of proportionality, not as a statutory requirement, but as a ‘constitutional imperative’ to curb the excesses of civil forfeiture. Its proper application ‘weighs the forfeiture and, in particular, its effect on the owner concerned, on the one hand, against the purposes the forfeiture serves, on the other’ (see, too, Mohunram at [58]).

The majority examined all the circumstances of the case and concluded (at [82]) that the forfeiture order made under s 50(1) of POCA was disproportionate. What factors are weighed in the inquiry into proportionality? According to the Constitutional Court in Prophet v NDPP 2006 (2) SACR 525 (CC) at [58], [63] and [68], the nature and gravity of the offence in question, the extent to which ordinary criminal law measures are effective in dealing with it, its public impact and potential for widespread social harm and disruption, are all to be considered. Ponnan JA observed that, according to Mohunram’s case, it would be wrong for POCA to be used in a manner which blurred the distinction between the purposes and the methods of criminal enforcement, on the one hand, and those of civil law, on the other. There was thus ‘no justification for resorting to the remedy of civil forfeiture under POCA as a substitute for the effective and resolute enforcement of ordinary criminal remedies’ (at [64]). The forfeiture had to advance the purposes of POCA itself; if not, the forfeiture, being the means, would be misaligned with the predominant ends pursued by POCA.

Was, then, the instrumentality of the offence sufficiently connected to the main purpose of POCA? The following considerations led the majority to answer in the negative: First, the appellant and others were, at the time, being prosecuted for contravening the Diamonds Act 56 of 1986, which set out fairly harsh penalties for its contravention and made specific provision for forfeiture, so that forfeiture in terms of POCA might be ‘doubly punitive’ (at [65]). Secondly, the forfeiture had to be ‘rationally related to its purposes’. These purposes, set out in the Cook Properties case at [(18)], included: (a) removing incentives for crime; (b) deterring persons from using or allowing their property to be used in crime; (c) eliminating or incapacitating some of the means by which crime may be committed; and (d) advancing the ends of justice by depriving those involved in crime of the property concerned. Although these motives were laudable, said Ponnan JA, there was ‘no escaping the draconian nature of the forfeiture’, particularly in the case of the second appellant (the wife) who had ‘committed no wrong of any sort, whether intentional or negligent, active or acquiescent’ (at [66], quoting from the Cook Properties case at [26]).

The third consideration involved the interests of the two minor children of the two appellants. In terms of s 28(2) of the Constitution, the best interests of the children are of ‘paramount importance’ in all matters concerning them. Their interests had to be given ‘specific and separate consideration, in addition to the attention they might get in the proportionality analysis’ (at [72]), so that a court might need to appoint a curator to conduct an independent assessment of their interests. No investigation had been conducted in this case and the interests of the children were not even mentioned in the proportionality analysis of the High Court. The onus of establishing proportionality rested on the NDPP, who made no attempt to adduce any evidence that, in respect of the wife and the two minor children in particular, the forfeiture would not be constitutionally excessive (at [78]).

The closing remarks of Ponnan JA at [81] encapsulate the sentiments of the majority as to why, in essence, the forfeiture order was disproportionate:

‘There can be no serious claim that the confiscation order sought by the NDPP is not punitive. Fundamental fairness prohibits the punishment of innocent people. I regard it as axiomatic that persons should not be punished when they have done no wrong. The unique facts of this case demonstrate that Ms Brooks and her children are entitled to the protection of that rule. ‘‘Improperly used, forfeiture could become more like a roulette wheel employed to raise revenue from innocent but hapless owners whose property is unforeseeably misused, or a tool to punish those who associate with criminals, than a component of a system of justice’’’ [quote from Bennis v Michigan 516 US 442 (1996) at 456].


 (iv) Appeal and Review

Appeal: Requirement that court of appeal gives notice of possible increase of sentence

S v Joubert 2017 (1) SACR 497 (SCA)

The appellant in the above matter was convicted in the regional court on 20 counts of fraud and sentenced to seven years’ imprisonment, wholly suspended on certain conditions for a period of five years. Aggrieved by his convictions but not his sentence, the appellant petitioned the High Court for leave to appeal against his convictions only. However, the High Court granted leave—‘erroneously it would appear’ (at [1])—to the appellant to appeal against sentence as well as conviction.

Riding on the back of the High Court’s error, the State ‘grasped the opportunity’ to give notice to the appellant and his attorneys that the State intended to seek an increase of sentence on appeal (at [2]). In his grounds of opposition, the appellant referred to the State’s failure ‘to follow the legal prescripts and requirements’ specified in the Criminal Procedure Act 51 of 1977. The State never formally sought leave to appeal against sentence, as required by s 310A of Act 51 of 1977. The High Court hearing the appeal was aware of this ‘procedural shortcoming’ (at [9]), but never notified the appellant that the court itself was indeed considering an increase on appeal.

Having heard the appellant’s appeal and having dismissed the appeal against conviction, the High Court set aside the sentence imposed by the regional court, replacing it with a sentence of seven years’ imprisonment, with four years suspended on certain conditions for a period of five years (at [1]).

The appellant appealed—with leave of the High Court—to the Supreme Court of Appeal against his increased sentence, claiming that his right to a fair trial had been infringed by the High Court’s failure to give him proper prior notice of its intention to consider increasing his sentence. Indeed, in the absence of a prior notice from the court, counsel for the appellant never explained to the appellant that the High Court ‘was minded to mero motu increase the sentence’ (at [6]). It was only in the course of argument before the High Court that she became aware of the High Court’s concern over the appropriateness of the sentence imposed by the regional court.

Writing for a unanimous full bench, Majiedt JA had ‘little doubt’ that the Constitutional Court decision in S v Bogaards 2013 (1) SACR 1 (CC) applied ‘squarely to this case’ (at [5]), despite the fact that the High Court appeal was concluded before Bogaards was decided (at [4]). For a summary of the principles established in Bogaards, see the discussion of s 309 in Commentary, sv Increase in sentence. Relevant to the appeal in Joubert were the following rules and principles as formulated in Bogaards (at [79]):

‘If the court forms a prima facie view before the hearing, that it is considering an increase in sentence, for instance, while reading the record, it should put the accused person on notice prior to the hearing. If the court forms this opinion during the hearing, then it must formally inform the accused person that it is considering an increase and give the accused person sufficient time, subsequent to the hearing, to make written submissions on this issue. Finally, even if the court is contemplating an increase after the hearing it must formally request the parties to make submissions on this point before making its final decision.’

It is clear that the High Court in Joubert had acted without regard to the rules in Bogaards. However, counsel for the State submitted that whilst its application to the High Court for an increase in sentence was irregular, the High Court’s failure to comply with the rules in Bogaards was cured by the fact that the State’s notice and application, even though irregular, had indeed alerted the appellant to the possible increase in sentence. Majiedt JA rejected this argument on the basis that it was ‘inconceivable that one fatal irregularity can be called in to cure another irregularity’ (at [8]).

Counsel for the State also advanced another argument in an attempt to place Joubert outside the ambit of Bogaards. It was argued that there had been no prejudice to the appellant since he was in fact afforded an opportunity to make submissions on sentence—in his opposing affidavit and through his counsel’s oral submissions in the course of argument when the High Court raised the matter. Having noted that not every irregularity is necessarily a fatal irregularity, Majiedt JA proceeded to point out that an accused person is required to demonstrate that the irregularity had caused material prejudice to him, ‘such that it...led to a failure of justice and an infringement of the right to a fair trial’ (at [8]). And fairness, continued Majiedt JA, had to be determined ‘upon the particular facts of the case context-specific’.

Turning to the facts in Joubert, it was concluded that the appellant had indeed been materially prejudiced. The prejudice present went beyond a mere inadequate opportunity to prepare. In preparing the response to the State’s notice, the appellant and his legal representatives obviously focused on the procedural defects of the notice and not the merits or demerits of an increase of the regional court’s sentence; and the possibility of an increase in sentence had not been foreshadowed in a prior notice from the High Court, as was required (at [9]).

At [10] Majiedt JA stated that there was ‘a further and even more compelling reason’ in support of the conclusion that the appellant had been materially prejudiced:

‘An accused person who has been given notice by an appellate court that it intends to increase the sentence imposed by the trial court has the option of withdrawing the appeal, with the leave of the appellate court. This practice, together with the requirement of prior notice to an accused person by the appellate court balances the appellant’s right to a fair trial and the court’s duty to ensure that the sentence is appropriate and where necessary, to increase an inappropriate sentence. In the present instance, the appellant had not been afforded the opportunity to consider such a course of action. As counsel for the appellant correctly contended, a notice of intention to increase sentence is a very weighty consideration, emanating as it does from the judges empowered to increase the sentence. The prejudice and its materiality are self-evident.’

The sentence imposed by the High Court on appeal was accordingly set aside, and the matter was remitted to the High Court for consideration of the appeal against sentence only, in accordance with the guidelines outlined by the Constitutional Court in Bogaards (supra) at [79].

Appeal: Disposal of an appeal in terms of s 19(a) of the Superior Courts Act 10 of 2013

S v Zulu [2017] ZASCA 207 (unreported, SCA case no 226/2016, 21 December 2016)

In the above appeal the Supreme Court of Appeal, having granted the appellant special leave to appeal to it, was required to determine whether the appellant had reasonable prospects of succeeding in his appeal against sentence if he were permitted to pursue such an appeal in the High Court (at [3]). In addressing this matter, the Supreme Court of Appeal had the relatively rare opportunity to exercise the power it now has in terms of s 19(a) of the Superior Courts Act 10 of 2013. This section determines that the Supreme Court of Appeal—or any Division of the High Court exercising appeal jurisdiction—‘may, in addition to any power as may specifically be provided for in any other law...dispose of an appeal without the hearing of oral argument...’.

In Zulu at [4] Wallis JA explained that ‘[a] reading of the record and the heads of argument made it clear to all the members of the court allocated to sit in the appeal that the appeal had to succeed’. It was accordingly decided that the appeal—which was set down for hearing on 16 February 2017—could be disposed of without the need to hear oral argument. Judgment was delivered on 21 December 2016. It was, said Wallis JA, ‘appropriate’ for the court to exercise its s 19(a) power ‘in the interests of the expeditious disposal of the appeal’. This also meant that there was ‘an appropriate use of judicial resources’ and the saving of costs ‘that would otherwise have been incurred by the public purse’ (at [4]). A further advantage was that the process of setting down the appellant’s appeal for hearing in the High Court was expedited. The decision in Zulu makes a lot of sense. But as a precedent it has and must have a very limited reach.

It is submitted that the Supreme Court of Appeal’s decision in Zulu (supra) to dispose of the appeal without the hearing of oral argument should be understood not only in the context of the fact that the judges concerned were all in agreement ‘that the appeal had to succeed’ (at [4]). It must be assumed that the judges in Zulu had, by implication, also decided that s 19(a) of Act 10 of 2013 could be invoked because they were convinced that nothing advanced in oral argument would or could have persuaded them to change their position. This type of decision, it is suggested, is not difficult to take where the issue happens to be elementary and where the solution or answer is of such a nature that the appellant’s rights are in no way curtailed or infringed, and the respondent is not deprived of arguing the merits of the matter elsewhere. It should be emphasised that in Zulu the issue was whether the appellant had reasonable prospects of success if he were allowed to have his appeal against sentence heard in the High Court. Matters of greater complexity and finality require a different approach. In these instances no court of appeal—however firm its viewpoint might be after having read the record and heads of argument—should deprive itself of the benefit of the salutary practice that a court, if it holds a view adverse to a particular party, should put such view to the party concerned in the course of oral argument.

Clearly, the decision in Zulu should not be interpreted as a case which seeks to reduce oral argument to an unnecessary formality.

 Table of Cases

Click here to see the full Table of Cases of the Criminal Justice Review No 1 of 2017


Click here for a printable .pdf version of this Review




Click here to register a profile and opt-in to receive these newsletters.

Click here to opt-out from this communication (please note that users are still required to have a profile in order to opt-out from receiving marketing communications from this website).


> 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

© Juta and Company (PTY) LTD  2019