NO 1 OF 2013


Andrew Paizes, Author (Editor)

Steph van der Merwe, Author



As editor, and on behalf of my co-contributor, Professor Steph van der Merwe, and our publisher, Juta & Co Ltd, it gives me great pleasure to introduce the Criminal Justice Review to subscribers to the Commentary on the Criminal Procedure Act. Commentary has recently passed a significant milestone, its 25th birthday, and Revision Service 50 marks a quarter of a century since the initial updating service was launched. The time was right, we believed, to extend our service to our subscribers by publishing, in electronic form, a Review which would keep our subscribers even more comprehensively informed of new developments in all areas of criminal justice.

Our aim is to time the publication of the Review so that, when subscribers receive the biannual Revision Services to Commentary, they will, almost contemporaneously, receive the Review, which will bring them up to date in respect of the three months that will have elapsed between the cut-off date for the material covered in the Revision Service and its receipt by subscribers. We will aim, in other words, to send out the Review at the end of June and November of each year.

The benefit to subscribers will, however, go beyond being completely up to date at those times. The Review will cover, selectively, developments in areas of criminal justice outside the scope of Commentary. It will, in particular, pay close attention to developments in the substantive criminal law as well as those areas of the law of evidence not dealt with in Commentary. Subscribers will also have the benefit of self-contained, critical feature articles which would not necessarily appear in Commentary at all.

The format of the Review will be such that cases or statutes of greater significance or interest will be discussed in the feature articles, while the rest will be selectively considered in brief summaries and short comments. In this first edition of the Criminal Justice Review, the feature articles concern important areas of the criminal law which have been the focus of much public interest and concern. The first addresses the incident concerning the killings at the Marikana mine, and considers whether it is good law to hold people in the position of those originally charged with murder liable for that crime where it was the police who shot and killed the victims, who were their fellow miners. This event triggered a massive public outcry, and much criticism was aimed – quite unfairly, it will be argued – at the much-maligned doctrine of common purpose.

The second concerns a decision of the Supreme Court of Appeal – S v Humphreys – on the meaning of a fundamental cornerstone of the criminal law – legal intention (or dolus eventualis) – and it considers the important question of whether, and, if so, in what circumstances, it is appropriate to convict the clearly reckless driver of a motor vehicle, who has caused the death of another in a collision, of murder once it is established (as it frequently will be in such cases) that he or she foresaw the real possibility that his or her conduct might cause the death of another person.

Andrew Paizes

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


Register at to receive future editions of the Criminal Justice Review

To receive future editions of this email newsletter, print and Jutastat CD-ROM and Online subscribers to Commentary on the Criminal Procedure Act and other interested Juta customers are required create a user profile on the Juta Law website (click here) and register to receive this FREE service under ‘Newsletters’. Juta Law customers with an existing website profile (which include Jutastat Online subscribers) can simply add this newsletter to their existing profile.



  • The conundrum of the Marikana miners: can there be liability for murder in such cases?
  • Dolus eventualis revisited: S v Humphreys 2013 (2) SACR 1 (SCA)


  • The Criminal Law (Forensic Procedures) Amendment Act 6 of 2010 as read with the subsequent Criminal Law (Forensic Procedures) Amendment Bill [B9–2013]


Criminal Law

  • Factual causation
  • Causation – death of infant arising out of injury to foetus
  • Dolus eventualis – foresight of real, not remote, possibility required
  • Common purpose – active association – mere presence not, by itself, sufficient
  • Fraud – prejudice – need not be proprietary prejudice
  • Robbery with aggravating circumstances – grievous bodily harm
  • Rape – consent – ostensible consent of a child merits great scrutiny

Criminal Procedure and Evidence: Pre-sentence

  • s 78(2): Basis on which accused must be referred for mental observation in terms of s 79
  • s 86 and s 88: Amendment to, or correction of, the indictment
  • s 94: Need for charge to allege commission of crime on diverse occasions
  • s 112(2): Prosecutor’s acceptance of plea of guilty to alternative charge
  • s 151: Effect of false evidence by accused
  • s 152 and s 158: Reconstruction of the record
  • s 164(1): No inquiry held to determine if minor child understood nature of oath
  • s 166: Duty of court to assist unrepresented accused and to avoid showing bias
  • s 166: Questioning by the court and its proper boundaries
  • s 203: Privilege against self-incrimination – does a litigant in civil proceedings have a right to have proceedings stayed pending outcome of related criminal proceedings?
  • s 208: Identification evidence by single witness
  • s 208: Cautionary rule – evidence of identification – identity parades
  • s 208: Cautionary rule – evidence of co-accused – corroboration not the only means of invoking caution
  • s 211: Evidence of previous convictions disclosed to presiding officer – irregularity – effect of
  • s 211: Evidence tending to disclose previous convictions of accused
  • s 216: Hearsay – when in interests of justice to receive it
  • s 216 and s 222: Hearsay – admissible in terms of s 222 – tendered for circumstantial value
  • s 217 and s 218: Pointing-out made without proper warning of right to legal representation
  • s 225: Illegally obtained evidence – accused not informed of right to a search warrant


  • s 276B(2): Sentencing court’s discretion to determine a non-parole period and the provisions of s 276B(2)
  • s 324(c): Lost record, acquittal and subsequent prosecution
  • Appeal to Constitutional Court on matters of sentencing: Principles to be applied
  • Minimum sentence in terms of s 51(1) of the Criminal Law Amendment Act 105 of 1997 and determining whether there was one or more rapes
  • Appeals: Sentencing and delays in hearing an appeal
  • Sentencing: White collar crime




The conundrum of the Marikana miners: can there be liability for murder in such cases?

The decision to charge over two hundred miners with the murder of over thirty of their fellow miners, shot and killed by the police at the Marikana mine in August 2012, created an outcry. Much was said and written in the media, and much of that bordered on the hysterical. Some blamed the doctrine of common purpose for what they saw as an outrageous decision. Charges were, eventually, dropped, but the question remains: If we leave aside the specific facts of the Marikana incident – which are the subject of commission of enquiry – is it good law to hold members of a mob, some of whom are armed with firearms, liable for the murder of other members of that mob killed by the police, if the accused foresaw that, in the circumstances: (1) one or more of their number might fire at the police; (2) the police might fire back at the mob; (3) at least one of their number might be hit by a bullet; and (4) he or they might be killed as a result?

It is not difficult to state the case for the prosecution in such cases. It would, first, not be difficult to establish that a common purpose existed between all of these people – either by reason of an implied (if not express) mandate, or by reason of active association – to commit the act of firing in the general direction of the policemen. The act of the miner or miners of shooting at the police would, then, be attributed to each of the other parties to the common purpose, so that that act would, in law, be deemed to have been performed by each of them just as if each had actually pulled the trigger.

It would, from that point, be easy to close the circle of liability. The act of shooting (which becomes, in law, the act of each of the group) is clearly a factual cause of the deaths of the deceased miners, in that they would not have died when they did were it not for the conduct. It is, further, a legal cause of that result, since it could hardly be maintained that the conduct of the police, in returning fire, was unforeseeable. It could, then, not amount to a novus actus interveniens. And, since it has been established that each of the accused must have foreseen, and, therefore, by necessary inference, did foresee the real possibility that one or more of their number might, as a result, be killed, all the elements of the offence are satisfied.

One’s sense of justice is, however, offended by this proposition. Authority, too, what little there is of it, is against it. In the (to my knowledge) only case to have considered directly this type of situation, S v Mkhwanazi & others 1988 (4) SA 30 (W), which involved a robbery by a gang of armed men, the argument that they were liable for murder in such circumstances was unsuccessful. Van Schalkwyk J found, rather generously, that dolus eventualis had not been proved (so, too, legal causation, erroneously in my view), but went on to add two significant remarks. First, that if he was wrong in reaching his conclusion about the absence of subjective foresight, he considered the proposition advanced by the state to be ‘inherently untenable’, one for which there existed no support in the authorities. And, second, that a finding of guilty was one ‘surely to be avoided’, whether on the grounds stated by him ‘or the moralistic idea that it is against public policy to convict on a charge of murder in circumstances such as this’.

I agree. But it is not easy to state clearly and succinctly why it would offend one’s sense of justice to convict an accused of murder in such circumstances. In order to isolate the reasons, and in an attempt to bring clarity to the issues involved, I will consider a number of situations, each more complex than its predecessor.

Situation 1: Consider the case of a single wrongdoer, A, who sets out to rob a bank. He knows that the bank is protected by two armed guards, so he takes a firearm with him, which he resolves to use, if necessary, to achieve his objective. He fires at the guards to frighten them off, but they return fire, and a bullet strikes and injures A. He foresees, of course, that this might happen and that he might, even, be killed as a result. Can A be convicted of assault or attempted murder? Obviously not. One cannot assault oneself, and murder is the unlawful and intentional killing of another human being, one, that is, other than oneself. Where, in other words, there is an identity of actor and victim, there can be no question of liability.

Situation 2: Assume that A, in the first situation, did not act alone but had a partner, B, who set out with him to rob the bank. What if A’s shooting at the guards caused them to fire back, resulting in a fatal injury to B? Can A be convicted of the murder of B? At first glance, the case against A is now much stronger than it was in situation 1. The act of A has caused the death of another human being, B, so that, with dolus eventualis being present, the requirements for murder would seem to be satisfied. If so, it must be noted, this is a conclusion reached without reliance having to be placed in any way on the doctrine of common purpose: there is clearly a causal nexus between the act of A and the death of B, since the return of fire by the guards, which was foreseen by A and B, cannot be regarded as a novus actus interveniens.

But convicting A of the murder of B strikes me – as it did the court in Mkhwanazi – as an affront to one’s sense of right and wrong. A and B set out to rob the bank together. There was a common purpose to do so. As a result, any act performed by either of them in the course of the common design is, in law, to be regarded as the act of each or both of them.

I have deliberately stated the principle in this way, even if it appears a little cryptic and, even, inelegant, since I believe that the words ‘each’ and ‘both’ are – in different contexts – both necessary to allow for a proper analysis of what we are doing when we invoke the doctrine of common purpose. The latter captures the sense of what one understands to be happening when a mob is involved in committing a violent act. When a victim, for instance, is stoned by a mob, one has no difficulty in accepting the notion that it was the mob who killed the victim. The throwing of the fatal stone thus becomes the act of not one but, say, twenty people. It squares with one’s visceral understanding of how the world works to imagine the throwing of that stone to be the product of the joint energies of all twenty who were involved in the murderous attack.

But sometimes the mechanics of the doctrine’s application to the facts demands a less descriptive and a more clinical appraisal. It is sometimes necessary – to lend functional utility to the doctrine – to consider the act in question to be the act of each of the wrongdoers in turn. The requirement, in a case of murder, for instance, that it must be shown that the accused foresaw that his or her act might cause the death of the victim, demands that one views the act in question, in the case of each accused, to be his or her own.

That this distinction is not merely a matter of semantics, is brought out more clearly by considering, again, the facts of situation 2. The firing of the shot at the guards by A fell clearly within the bounds of the mandate that existed between A and B. If we apply the doctrine of common purpose – as, indeed, we should – we could describe what happens in a number of ways. We could say that the act of A is imputed to B; that it becomes B’s act as much as it is A’s act; that it is the act of each of them; or that it is the act of both of them. To say that it is the act of each inclines one more in the direction of a conviction for murder, since we have A’s act of shooting and B’s act of shooting. B’s death cannot support a murder conviction as a result of B’s shooting, of cause, for the reasons set out in situation 1. But there would seem to be no obstacle to a conviction arising out of A’s shooting, since the requirement of ‘another human being’ would seem to be satisfied.

But if we look at the situation through a slightly different lens, one that sees the act of firing at the guards as the act of both A and B, as a kind of combined effort, the picture changes. B is killed as a result of an act that is as much his as it is A’s. And it is no longer at all clear that the rule that ‘another human being’ must be killed by the ‘act of the accused’ is satisfied. After all, if the act of firing at the guards is as much B’s act as it is A’s act, and if (assuming B was not killed but only injured by the return fire) B himself could not be held liable for injury to himself brought about by that act, then it is less clear that we can justify holding A liable for injury to B caused by that same act.

So which of these versions is preferable? Within this particular context it is my view that the second picture squares more with our sense of justice and public policy. The point is that, once an act is committed by one of the parties to a common purpose, and that act falls within the mandate given to that party by the other or others, the law ceases, quite properly, to have its normal concern for determining, precisely, who did what. Non-actors, in the strict sense, become actors, and identity, for the purpose of determining precisely who did what, is, for this purpose, lost to the collective. And when one of the parties is himself or herself a resultant victim of such an act, as in situation 2, it becomes inappropriate to ‘unscramble the eggs’, as it were, and to attempt to identify individuals within the collective for the purpose of attributing specific criminal responsibility to each individual other than the actual victim.

If this picture is preferred, a curious irony arises. I mentioned before that, in situation 2, the case for holding A liable for B’s murder did not, in any way, rely on the doctrine of common purpose. The irony is that this doctrine, so much maligned by so many in the Marikana incident, would, on the contrary, be the reason for not holding A liable if my argument found favour.

And so to situation 3: add one or more (even two hundred) parties to the common purpose to rob the bank. Let A again fire at the guards, who fire back, killing B. What is the liability of C, or D, or any other member of the mob? If it is right to hold A liable for B’s murder, then we will need the doctrine of common purpose to attribute A’s act to each of the others. But, I have argued, it is not right to do so, and so the doctrine (which was the very reason why it was wrong to do so), is not to be invoked to spread responsibility to the others.

For these reasons, than, it would be wrong, in my opinion, to convict accused persons in the position of the Marikana 200, of murder.

Andrew Paizes 

Dolus eventualis revisited: S v Humphreys 2013 (2) SACR 1 (SCA)

In this important case, the Supreme Court of Appeal considered, once again, the definition of dolus eventualis in the criminal law. It accepted as trite that, in a case of murder, the test for dolus eventualis was twofold: (a) did the accused subjectively foresee the possibility of the death of the victim ensuing from his conduct; and (b) did he reconcile himself with that possibility.

After finding that the accused in Humphreys satisfied both of these conditions, the trial court concluded that he was guilty of murder. After finding that the first condition was satisfied but that the second was not, the Supreme Court of Appeal reached the conclusion that he was not guilty of murder. It is my view that the trial court cannot be faulted in finding that both conditions were satisfied, but that its conclusion that dolus eventualis was present was wrong. And it is my further view that the Supreme Court of Appeal reached the correct result, but only by a process of combining two cancelling errors: one that the second leg of the test can negate dolus eventualis once the first leg has already been satisfied; and the second that, on the facts, the appellant had not reconciled himself with the foreseen possibility.

The facts of Humphreys were quite simple. The appellant, an experienced minibus driver who operated a shuttle service for school children, was driving his minibus on the same route that he had used for ten years. The minibus was carrying fourteen children. On reaching a railway crossing, he joined a queue of vehicles which had stopped because the large red warning lights had started flashing. There were, at this intersection, on both sides of the railway line, also large stop signs as well as other traffic signs indicating a railway crossing. The crossing was, further, controlled by two booms, positioned on different sides of the railway line, so that it was possible to avoid them, even when they were down, by going onto the lane intended for oncoming traffic and then returning to the correct lane to pass the boom on the other side. And this is what the appellant attempted to do. He overtook the line of vehicles after the red lights had started flashing, and dodged the boom that had come down mere seconds after the lights had come on. Upon entering the intersection, the minibus was hit on its left side by the train, the driver of which had no chance to avoid the collision. Ten of the children were fatally injured, and the accused was charged with murder in respect of each of them.

It was clear that the first element of dolus eventualis – subjective foresight of the possibility of death as a result of his conduct – was present. Everything, then, turned on the second element. This element, said Brand JA, had been explained by Jansen JA in S v Ngubane 1985 (3) SA 677 (A) at 685 A–H. He called it the ‘volitional element’, which is present when the accused ‘‘‘consents’’ to the consequences foreseen as a possibility, he ‘‘reconciles himself’’ to it, he ‘‘takes it into the bargain’’’.

Did the appellant in Humphreys ‘reconcile himself’ to the foreseen consequences of his conduct? The trial court held that he did; Brand JA (with whom Cachalia and Leach JJA and Erasmus and Van der Merwe AJJA agreed) reached the opposite conclusion. The ‘true enquiry under this rubric’, said the court, was ‘whether the appellant took the consequences that he foresaw into the bargain; whether it [could] be inferred that it was immaterial to him whether those consequences would flow from his actions’. Thus, ‘if it [could] reasonably be inferred that the appellant may have thought that the possible collision he subjectively foresaw would not actually occur, the second element of dolus eventualis would not have been established’ (at [17]).

On the facts, Brand JA accepted that the latter inference was not only a reasonable one, but, indeed, the ‘most probable one’. He gave two reasons for this conclusion. First, to reconcile himself to the death of the victims, he would have had to reconcile himself to the possibility of his own death, and there was no evidence to suggest this. He ‘foresaw the possibility of the collision, but he thought it would not happen; he took a risk which he thought would not materialise’ (at [18]). Second, the evidence showed that the appellant had, previously, successfully performed the same manoeuvre in virtually the same circumstances. The inference was ‘that, as a matter of probability, he thought he could do so again’ (at [19]). Dolus eventualis had, therefore, the court concluded, not been established. Culpa was however, clearly present, and the convictions of murder were replaced with culpable homicide.

The following submissions are respectfully put forward:

  1. Both elements of dolus eventualis, as articulated and understood by the courts, were present. The first was clearly present. And the second, which adds nothing of any substance to the first, is merely descriptive of what happens when an accused, having foreseen the possibility of causing death by his conduct, goes ahead with that conduct. How can X, who foresees the possibility of causing Y’s death if he commits act Z, not reconcile himself with the possibility (not the actual fact of) causing Y’s death if he goes ahead and commits act Z? (See, further, Whiting ‘Thoughts on dolus eventualis’ (1988) 3 SACJ 440 and Paizes ‘Dolus eventualis reconsidered’ (1988) 105 SALJ 636.)
  2. To say that the appellant ‘foresaw the possibility of the collision but he thought it would not happen; he took a risk which he thought would not materialise’, is to confuse the two elements of dolus eventualis. If he, at the crucial time of the manoeuvre, no longer believed it would happen, and this belief was so strong that it displaced entirely the foresight of the real possibility of a fatal collision (a requirement recently set out, once again, by the Supreme Court of Appeal in S v Makgatho (unreported, (732/12) [2013] ZASCA 34, 28 March 2013) (see the discussion under Case Law below), then it would be wrong to find that the first element was present. If this was not the case, and it seems clear that it was not, then a more accurate articulation of the appellant’s state of mind would be that he still foresaw the real possibility of such a collision, but believed it to be improbable (or, perhaps, even, highly improbable). Such a finding would leave both elements in place. The first because he still foresaw the real possibility of a fatal collision. The second because he must, by going ahead with the manoeuvre, have reconciled himself with this possibility.
  3. The trial court was, thus, correct in finding both elements to have been satisfied. But it was wrong in finding dolus eventualis to be present. How can this be? Simply because it has become clear over the years that the test set out by our courts is insufficiently subtle, nuanced or elastic to cover the entire range of situations that fall to be considered in such cases. Whiting, in his insightful article (supra) points the way to a better understanding of some of the factors which may affect the question. One is the type of activity involved. Where an accused is engaged in an activity which, although it involves some risk of harm, is not only socially acceptable but also legally permissible, different issues arise. Such activities include mining and the driving of a car. Whiting argues that, even if a person drove at a grossly excessive speed and was also aware that the brakes were defective, such that he must have realised that he was endangering the lives of others to a degree which was substantially beyond what was permissible, it would, in the case of a fatal collision, still offend one’s sense of what is right to convict him of murder. Dolus eventualis, in his view, would not have been established. Other factors, in his view, include the type of risk involved (was it of a generalised statistical nature or was it a specific, concrete risk?); whether the conduct involved was a positive act or a mere omission; and whether the accused’s purpose was to expose the victim to the risk of death or whether he merely accepted that the risk was present. Some of these factors will be relevant to the degree to which the accused must have foreseen the possibility of death; others to the question of whether it is appropriate to be speaking of dolus eventualis at all.
  4. The crucial point is that while the conventional test will produce the correct results in the vast majority of cases, there will be cases (and Humphreys is one) where it will not. In such cases, adhering correctly to the test (as the trial court did) will lead to the wrong result. And the right result will be reached only by reading into the second element something that should not be there.
  5. There are other factors that may affect the test for dolus eventualis: Whiting makes it plain that his list is not closed. One such factor may well be the attitude of the accused to the taking of the risk: was it immaterial to him that his victim was exposed to the risk, or did he hope strongly that it would not eventuate?

What emerges is that the law, like life, is sometimes messier and more nuanced than we would like it to be. Dolus eventualis, it seems, is not amenable to containment within a simple formula. It operates, rather, on a scale where black and white are separated by infinitely many shades of grey. All that can be said with confidence, is that the strongest case for dolus eventualis is likely to be found where there is foresight of a substantial possibility of causing the result in question; where the activity is part of an overtly dangerous and unlawful enterprise (such as robbing a bank); and where the accused is uncaring about whether the victim lives or dies as a result of his conduct. On the other hand, the weakest case will tend to be where there is foresight of only a slight possibility of death; where the activity is socially valuable and permissible (such as mining); and where the accused strongly hopes that life will not be lost in the enterprise, and has taken considerable care to ensure that the risk is minimised.

It is, obviously, not easy to state with certainty precisely where the dividing point on this line will be. But, in Humphreys, at least, it is clear that the requirements for dolus eventualis were not met: the accused was engaged in an activity which – although he was clearly carrying it out in a manner which was grossly negligent – is both useful and socially valuable. This is not to say that there can never be situations where fatal collisions involve dolus eventualis. But this was not one. The reason for taking the risk was not callous indifference to human life but, rather, impatience and foolish impetuosity. He was not indifferent to the fate of his passengers but, rather, unrealistic in his assessment of the degree of the risk.

It would, in short, be an affront to one’s sense of justice to convict him of murder. Not because of failure to satisfy the second element as defined by the courts; but because it is inappropriate to do so given the more fine-grained picture of dolus eventualis sketched above.

Andrew Paizes


The Criminal Law (Forensic Procedures) Amendment Act 6 of 2010 as read with the subsequent Criminal Law (Forensic Procedures) Amendment Bill [B9–2013]

The Criminal Law (Forensic Procedures) Amendment Act 6 of 2010 supplemented and amended Chapter 3 of the Criminal Procedure Act 51 of 1977 extensively, with effect from 18 January 2013 (GG 36080 of 18 January 2013). Act 6 of 2010 amended the existing s 37 by further regulating police and court powers as regards the taking of body-prints or the ascertainment of bodily appearances of accused and convicted persons. Act 6 of 2010 also inserted three new sections into Chapter 3: ss 36A, 36B and 36C. These new sections deal with a wide variety of matters pertaining to the ascertainment of bodily features of persons, ranging from definitions to be used in the interpretation of Chapter 3 to the taking, retaking and destruction of fingerprints in specified circumstances. The following two sections in Chapter 24 of the Criminal Procedure Act were also amended so as to align them with the refined and revamped Chapter 3: s 212 (proof of certain facts by affidavit or certificate) and s 225 (evidence of prints or bodily appearance of accused).

Commentary on the Criminal Procedure Act (as revised by Revision Service 50 of 2013) reflects all the above changes, with comment where necessary.

It should, however, be noted that Act 6 of 2010 did not introduce any procedures specifically related to DNA. It was decided that DNA procedures would be dealt with in separate legislation which would, once again, amend the Criminal Procedure Act. This process has been set in motion by the Criminal Law (Forensic Procedures) Amendment Bill [B 9–2013], as published in GG 36415 of 26 April 2013. This Bill, should it become law, will bring our criminal justice system a great step closer to the ideal of having detailed rules regulating all aspects of DNA evidence. The Bill covers DNA matters ranging from the taking of bodily samples from certain persons in specified circumstances, to the establishment of a National Forensic DNA Database (‘‘NFDD’’) in terms of a proposed s 15G to be inserted in the South African Police Service Act 68 of 1995.

The Bill proposes several insertions into the existing s 36A of the Criminal Procedure Act. An important feature of the Bill is that it identifies and describes at least four different ‘samples’ relevant to DNA procedures and Chapter 3 generally:

  • A ‘bodily sample’ is an over-arching term and means any type of sample taken from a person and includes ‘intimate’ and ‘non-intimate samples’
  • An ‘intimate sample’ means ‘a sample of blood or pubic hair or a sample taken from the genitals or anal orifice area from the body of a person’, but excludes a ‘buccal sample’ which means ‘a sample taken from the inside of a person’s mouth’.
  • A ‘non-intimate sample’ would be a ‘buccal sample’ or ‘a sample taken from a nail or from under the nail of a person’.

The precise category of the sample is important in determining who may take such a sample. The Bill seeks, for example, to amend the existing s 37 to make it clear that a police official would not be allowed to take ‘an intimate sample of any person’.

The taking of a buccal sample was described as ‘quick and painless’ by Kennedy J in Maryland v King 569 US 1 at 5 (2013).

The Bill determines that any forensic DNA profile derived from samples, shall only be used for the following purposes: detection or investigation of crime; conducting a prosecution; or identification of unidentified human remains or missing persons. Abuse of DNA samples or forensic profiles shall be a crime, punishable with a prison term not exceeding 15 years.

It is estimated that the provision of DNA sample collection kits, the analysis thereof, destruction of samples and essential infrastructure will cost R160 million per annum. Training of police officials and forensic awareness programmes would be in the region of R22 million. See paragraph 5 of the Memorandum accompanying the Bill. The importance of proper training for police officials must be emphasised. Detailed legislation and adequate laboratories would serve no purpose if crime scene samples are botched and ‘the chain of custody’ suspect. In S v Adams (unreported, ECG case no 73/2011, 25 June 2012) the court referred to the importance of the ‘chain of custody’ of a DNA sample as well as the need for verification of the authenticity and legal integrity of such a sample. It has rightly been said that DNA testing has the ‘unparalleled ability’ both to exonerate the innocent and to identify the guilty; and it can ‘significantly improve both the criminal justice system and police investigative practices’ (District Attorney’s Office for Third Judicial District v Osborne 557 US 52 at 55 (2009). See further S v Rululu 2013 (1) SACR 117 (ECG) where the chain of custody was admitted and the s 212 affidavit encapsulating the DNA test result proved the prosecution’s case as regards the identity of the rapist.

In the Commentary on the Criminal Procedure Act (see the comment on s 225 sv DNA profile) the dangers of relying on DNA evidence, which is a form of circumstantial evidence, are discussed. The need to avoid the so-called ‘prosecutor’s fallacy’ is stressed, as are the dangers of using mathematical techniques where the underlying statistics do not rest on firm foundations.



Criminal Law

Factual causation

Lee v Minister for Correctional Services 2013 (2) SA 144 (CC), 2013 (1) SACR 213 (CC)

Although this is not a criminal case, its treatment of factual causation is likely to be influential in the criminal context. The question was whether the applicant’s detention and the systemic failure to take preventative and precautionary measures by the Correctional Services authorities was a factual cause of the applicant’s infection with tuberculosis while in detention. The applicant had been incarcerated in a maximum security prison at Pollsmoor from 1999 to 2004. He did not have TB when he arrived there. He was diagnosed with the disease after three years in the prison, which was described (at [8]) by Nkabinde J (who delivered the majority judgment) as ‘notoriously congested’ and as providing ideal conditions for the transmission of what is ‘an airborne communicable disease which spreads easily, especially in confined, poorly ventilated and overcrowded environments’.

The applicant’s claim for delictual damages was upheld by the High Court (see 2011 (2) SACR 603 (WCC), 2011 (6) SA 564), but this decision was overturned by the Supreme Court of Appeal (see 2012 (1) SACR 492 (SCA), 2012 (3) SA 617). That court held that factual causation had not been proved by the applicant, since it was ‘just as likely as not that Mr Lee was infected by [an inmate] who the [responsible] authorities could not reasonably have known was contagious’ (at [63] SCA). Since he could not track the precise trail of acts by the prison authorities that caused his infection, he was compelled to rely on a ‘systemic omission’; that, in other words, the lack of overall management measures caused that result. And, to prove that a failure to implement a reasonable system of TB controls had probably caused him to contract that disease, he had to show that such a system ‘would have altogether eliminated the risk of contagion’, which he was unable to do.

The question then found itself before the Constitutional Court, the judges of which were unable to agree. Five, led by Nkabinde J, found in favour of Mr Lee, and four, led by Cameron J, held that the test for factual causation had not been satisfied, but that the Supreme Court of Appeal should, nevertheless, to prevent an infraction of the applicant’s constitutional rights to security of the person and dignity, have considered developing the common law of causation.

In the opinion of the majority, the Supreme Court of Appeal erred in two respects. First, it ignored the fact that ‘[o]ur existing law does not require, as an inflexible rule, the use of the substitution of notional, hypothetical lawful conduct for unlawful conduct in the application of the ‘‘but-for’’ test for factual causation’ (at [50]). There was, said Nkabinde J (at [55]), ‘nothing in our law that prevented the high court from approaching the question of causation simply by asking whether the factual conditions of Mr Lee’s incarceration were a more probable cause of his tuberculosis, than that which would have been the case had he not been incarcerated in those conditions’ (emphasis added). Second, ‘even if one accepts that the substitution approach is better suited to factual causation, there was ‘no requirement that a plaintiff must adduce further evidence to prove, on a balance of probabilities, what the lawful, non-negligent conduct of the defendant should have been’: ‘what is required is postulating hypothetical lawful, non-negligent conduct, not actual proof of that conduct’ (at [56]).

Cameron J disagreed with the majority’s approach to factual causation for three reasons: first, that it was not possible to infer probable factual causation from an increase in exposure to risk by itself. Moreover, where the actual origin of the infection could not be traced, it was ‘impossible to say that infection was probably caused by a negligent exposure to risk, as opposed to an exposure that no amount of care on the prison authorities’ part could have avoided’ (at [106]). Second, the very nature of negligent conduct was that it increased risk, so that the approach of the majority impelled one to accept that probable factual causation followed from every finding of negligence. And, third, that approach left no room for an assessment of the amount of risk exposure, so that factual causation could be inferred from any increase in risk.

Causation – death of infant arising out of injury to foetus

S v Mentoor (unreported, WCC case no A 300/2012, 27 February 2013)

This interesting case asks whether an assault on a pregnant woman by the accused which results in damage to the foetus, the premature birth of the child, and the child’s death a few hours later, may properly sustain a conviction of murder. The trial court held not, since the accused’s conduct took place before the live birth of the child, and since a foetus was not a ‘person’ for the purpose of the definitional requirements for murder. Louw J (Nyman AJ concurring), however, took the view that this was incorrect, since murder is a ‘consequence crime’, and the accused’s assault on the woman caused the child to be born prematurely and caused the child’s death shortly thereafter. However, since there had been no appeal by the state against the acquittal on this charge, it was not necessary for the court to decide this issue.

Dolus eventualis – foresight of real, not remote, possibility required

S v Makgatho (unreported, (732/12) [2013] ZASCA 34, 28 March 2013)

In a significant acceptance of what we submit to be the correct position, the supreme Court of Appeal in S v Makgatho emphasised that, for dolus eventualis to be established in respect of a consequence, ‘it must be shown that a real – as opposed to a remote – possibility of that consequence resulting was foreseen’ (at [9]) (italics added). Shongwe JA referred to S v Van Wyk 1992 (1) SACR 147 (Nms) at 161b, where Ackerman AJA said there must be ‘the subjective appreciation that there is a reasonable possibility that the prescribed consequence will ensue’. Any doubt that foresight of a remote possibility might suffice has now been eliminated.

Common purpose – active association – mere presence not, by itself, sufficient

S v Toya-Lee van Wyk (unreported, (575/11) [2012] ZASCA 47, 28 March 2013)

The Supreme Court of Appeal gave a salutary reminder that, while the inference of an active association such as to warrant the conclusion that there was a common purpose can sometimes be drawn from what occurred or was said during or after the event, care needs to be taken to avoid lightly inferring such an association with a group activity from the mere presence of the person who is sought to be held liable for the actions of some of the others in the group (at [16]).

In this case, said Pillay JA (at [18]), ‘even if the appellant had realised that the deceased was about to be killed when he returned into the thicket with the rest of the group, that does not justify an inference that he was in agreement with, or approved of, the crime which was about to be perpetrated, nor that he thereby manifested his association with the group’s criminal purpose’. The fact that he did not participate in the murderous assault on the deceased, meant, further, that there was no evidence of any act of association by the appellant with the actions of those who assaulted and murdered the deceased. The fact that he did take the knife from one of the group when ordered to do so, and went as far as asking where the heart was located, was not, in the court’s view, sufficient to establish either the intention to make common cause with the others or the performance by him of some act of association with them.

Fraud – prejudice – need not be proprietary prejudice

S v Tshopo & others 2013 (1) SACR 127 (FB)

The court confirmed that, for the crime of fraud, the prejudice caused by the accused need not be of the proprietary kind. This case involved a tender fraud; the appellants had failed to disclose a family connection in a tender application which specifically required an affinity declaration. This, said the court, was prejudicial to both other tenderers and the community at large. The state had an interest in keeping strict control over state tenders, which were being unscrupulously used for self-enrichment by public servants. Similarly, the general public, whose funds were being used to finance such projects, also had an interest in keeping strict control over these tenders.

Robbery with aggravating circumstances – grievous bodily harm

S v Maselani (unreported, (511/2012) [2013] ZASCA 21, 22 March 2013)

Where an accused is charged with robbery with aggravating circumstances, and a court is required to consider whether, in terms of s 1(b)(ii) of the Criminal Procedure Act, grievous bodily harm has been inflicted, is it appropriate to consider the consequences to the victim of the assault as a relevant factor? It was argued that it was not, since the inquiry was concerned solely with the ‘nature, position and extent of the actual wounds and injuries’, and the consequences stemming from these wounds and injuries was to be ignored. But this argument was rejected. The Supreme Court of Appeal agreed with what had been said in R v Jacobs 1961 (1) SA 475 (A), where the majority stipulated that whereas the intention of the accused in this regard was irrelevant, the infliction of grievous bodily harm had to be considered in relation to ‘the whole complex of objective factors involved in the accused’s assault,’ including ‘the results which flowed from’ the infliction of the wounds or injuries.

Common sense, too, impelled the court to this conclusion. To exclude the consequences could lead to absurd results, since aggravating circumstances would exist if there were a mere threat to kill, but not if the accused actually caused death if he did so by only the moderate use of force, as in Maselani.

Rape – consent – ostensible consent of a child merits great scrutiny

S v Mugridge (unreported, (657/12) [2013] ZASCA 43, 28 March 2013)

Any appearance of consent to sexual relations given by a child is, said Erasmus AJA at [42], ‘deserving of elevated scrutiny’. The inequalities in the relationship between an adult and a child in this regard are of great importance in understanding the construction, nature and scope of the child’s apparent consent, and they may manifest in the form of ‘forced choices, precluded options, constrained alternatives, as well as adaptive preferences’ (see S v M 2007 (2) SACR 60 (W) at [37] and Marx v S [2005] 4 All SA 267 (SCA)). The court pointed to the vulnerability of children and their openness to manipulation as dangers when assessing any consent apparently given by children. Erasmus AJA warned, in particular, of a ‘system of gifts and privileges being accorded as a reward for compliant behaviour’ – acquiescence to sexual relations in response to what has sometimes been described as ‘sexual grooming’. He described this notion as the utilising and manipulating by the perpetrator of a position of authority over the child victim and the child’s environment in a manner which opens the victim up to the intended abuse (at [45]).

In Mugridge the appellant, a man of the church, ‘had manipulated the complainant’s fragile state and his stature in the community to his advantage, slowly inviting her to acquiesce to his advances’ (at [52]). Her compliance was a ‘direct result of his calculated distortion of his position of authority over her’, which included his supplying her with drugs and alcohol to weaken further her resistance and cloud her judgment. There was, in short, no real consent.


Criminal Procedure and Evidence: Pre-sentence

s 78(2): Basis on which accused must be referred for mental observation in terms of s 79

In Malatji v S (unreported, NGP case no A259/10, 18 April 2013) the appellant’s legal representative, in addressing the trial court on sentencing after a conviction for murder, had pointed out that the appellant had at some stage been detained in a mental hospital but had recovered fully. The relevant SAP 69 form also showed that the appellant had been convicted of murder in 1984, and had been declared a State President’s patient. The trial court, without any further investigation, summarily sentenced the appellant to 20 years’ imprisonment.

On appeal Makgoka J and Ratshibvumo AJ held that the trial court had erred in that the court was in terms of s 78(2) obliged to send the accused for mental observation as provided for in s 79. It was held that there was no onus on the appellant to have persuaded the court to do so, as ‘there need only be a reasonable possibility’ that the accused might lack criminal responsibility (at [8]). The jurisdictional fact required by s 78(2) is reached when there is a reasonable possibility. At [7] it was held:

From the plain reading of s 78(2) it is clear that in the case of an allegation or appearance of mental illness or mental defect, the court is obliged to direct that an enquiry be made under s 79. However, if the allegation is made or appearance relates to ‘‘any other reason’’ the court has a discretion whether or not to direct that an enquiry be made under s 79. In my view the present case falls under the former category. The very fact of the appellant having in the past been detained as the State President’s patient, was sufficient for the court to direct that an enquiry be made under s 79. There was more than an allegation.

It should be noted that an order in terms of s 78(2) can be made at any stage of the proceedings, that is, even after conviction. See S v Mogorosi 1979 (2) SA 938 (A). 

s 86 and s 88: Amendment to, or correction of, the indictment

In Nedzamba v S (unreported, (911/2012) [2013] ZASCA 69, 27 May 2013) the High Court had convicted the appellant of two counts of ‘rape’ committed on 17 March 2008. The indictment made no reference to the provisions of the Criminal Law (Sexual Offences and Related Matters) Amendment Act 32 of 2007, which came into operation on 16 December 2007. On appeal it was argued that the accused had been charged with the common-law crime of rape at a time when it had already been abolished by Act 32 of 2007 and had been replaced by a statutory provision, namely s 3 of Act 32 of 2007. The state agreed with this argument.

A full bench of the Supreme Court of Appeal rejected this approach. Navsa JA pointed out that s 3 of Act 32 of 2007 actually preserved ‘rape’ as an offence whilst making it gender neutral and broadening it to cover other acts in order to provide ‘greater protection to victims of sexual misconduct’ (at [16]). The absence of a reference to s 3 of Act 32 of 2007 in the indictment was not fatal in that the indictment ‘undoubtedly asserted that the appellant was guilty of ... rape and the summary of substantial facts set out the details’ (at [17]). It was also emphasised that in terms of the provisions of s 86 of the Criminal Procedure Act, the amendment of a charge would be possible where it would not prejudice the accused (at [18]).

Navsa JA also pointed out that s 88 of the Criminal Procedure Act allowed ‘latitude’ in that a defect in a charge could be cured by evidence (at [19]).

Referring to S v Motha 2012 (1) SACR 451 (KZP), Navsa JA took the view that an amendment which amounted to the inclusion of a reference to s 3 of Act 32 of 2007, was possible on appeal and if it could be done without prejudice to the appellant (at [21]–[22]).

In Nedzamba (supra) the amendment was indeed possible because the appellant had all along been fully aware of the fact that he ‘faced a charge of rape’ (at [23]) and ‘his defence would have remained the same’ (at [24]). Prejudice was absent.

s 94: Need for charge to allege commission of crime on diverse occasions

 In Senwedi v S (unreported, NWM case no 27/12, 19 April 2013) the trial magistrate had sentenced an accused on a multiplicity of rapes, whereas the charge sheet identified only a single occasion of rape. On appeal it was held that this did not necessarily render the trial unfair and that, on the evidence, the appellant was ‘still guilty of the rape on a single count of a minor complainant’ (at [23]).

The prosecution, it was pointed out at [21], ‘was aware of the multiplicity of rapes that [had] occurred’ but alleged only ‘one count of rape’. Kgoele J referred to S v Mponda2007 (2) SACR 245 (C) where it was held that the right to a fair trial is placed in jeopardy where the state proceeds on a single count of rape while the evidence supports a multiplicity of rapes. The prosecution should, ideally speaking, rely on s 94 from the outset. This section states that where it is alleged that an accused on diverse occasions during any period of time had committed an offence in respect of a particular person, the accused may be charged in one charge with the commission of that offence on diverse occasions during a stated period. It follows that where it is impracticable (or even impossible) to specify each and every occasion on which the alleged crime was committed, the prosecution should draft the charge as provided for in s 94. This approach will not only prevent sentencing problems but will also avoid serious problems as regards the admissibility (relevance or irrelevance) of evidence pertaining to offences not identified in the charge. See S v Jones2004 (1) SACR 420 (C) at 426ij where it was held that the trial court had erred in receiving certain evidence relating to an alleged rape by the accused of the complainant but not specifically alleged by the prosecution as one of the charges.

s 112(2): Prosecutor’s acceptance of plea of guilty to alternative charge

In Tshilidzi v S (unreported, (650/12) [2013] ZASCA 78, 30 May 2013) the appellant had pleaded not guilty to the main charge but guilty to the alternative. The prosecutor informed the trial judge that she had read the appellant’s s 112(2) statement and that she accepted his plea of guilty to the alternative charge. The s 112(2) statement – prepared and signed by the appellant, who had legal representation – was then read into the record, confirmed by the appellant and handed to the court. The statement clearly supported the plea of guilty to the alternative charge. However, the trial judge refused to proceed on the basis of the plea of guilty. He based his refusal on the inconsistency between the plea and the s 144(3)(a) summary of substantial facts that accompanied the indictment. As a direct result of this refusal, the appellant withdrew his plea of guilty and pleadednot guilty to the main and alternative charge. The trial proceeded on this basis and the appellant was ultimately convicted of the main charge (rape) and sentenced to life imprisonment.

On appeal Van der Merwe AJA, writing for a unanimous full bench, concluded that the trial judge had committed a ‘gross irregularity’ (at [4]). The trial judge had interfered with the powers of the prosecutor in that the latter had the discretion to accept or refuse the plea of guilty on the alternative charge (at [5] and [6]). At [7] reference was also made to S v Ngubane 1985 (3) 677 (A) 683E–F where Jansen JA said that the prosecutor’s acceptance of a plea at the commencement of the trial, is ‘a sui generic act by the prosecutor by which he limits the ambit of the lis between the State and the accused in accordance with the accused’s plea’. This approach, it can be argued, stems from the fact that the prosecution as dominus litis determines the charges against the accused. See Commentary on the Criminal Procedure Act where s 112 is discussed, sv Acceptance of plea by prosecutor: Effect of. The prosecutor’s decision whether to accept or reject the plea exists up to the time the plea is tendered in court and before any evidence is led. See also S v Cordozo 1975 (1) SA 635 (T) referred to by Van der Merwe AJA at [6].

Van der Merwe AJA went further than merely confirming the approaches adopted in Ngubane (supra) and Cordozo (supra). He pointed out that the irregularity committed by the trial judge should also be seen in the context of the fact that ‘[t]he independence of the prosecuting authority concerning prosecutions is entrenched in s 179 of the Constitution’ and is given effect by ss 20(1) and 32 of the National Prosecuting Authority Act 32 of 1998.

It was concluded that, for purposes of the appeal, the conviction on the main count could not stand because the prosecutor’s acceptance of the plea on the alternative charge had removed the main charge from the indictment; and on appeal a conviction on the alternative charge was also not possible because the accused, in response to the trial court’s insistence that the trial should proceed on the main charge, had withdrawn his plea of guilty on the alternative charge (at [9]).

The conviction and sentence were set aside. 

s 151: Effect of false evidence by accused

S v Deppe and Douman (unreported, (512/2012) [2013] ZASCA 4, 7 March 2013)

The Supreme Court of appeal accepted the distinction set out in the Australian case of Zoneff v R [2000] 200 CLR 234 at 238 and made by Kirby J between ‘credibility lies’ and ‘probative lies’:

The former are said to be those which, according to their content, affect the credibility of the accused’s evidence and thus the weight which the jury may give to other testimony of the accused. In this sense, a conclusion that the accused has lied upon one matter, even peripheral to the offence charged, may make the jury scrutinise with more care (perhaps scepticism) other testimony given by the accused. It might, in this way, contribute indirectly to the rejection of the accused’s version of critical events and the acceptance of that pronounced by the prosecution. Probative lies, on the other hand, are those ‘‘which naturally indicate guilt ... a hard test to satisfy’’ ... This is a ‘‘hard test’’ precisely because it is rare that a lie about a particular matter will be so crucial as, of itself, if proved, to establish directly guilt beyond reasonable doubt of a criminal offence.

Schoeman AJA applied the test set out in S v Mtsweni 1985 (1) SA 590 (A) at 593–4, which corrected the narrow approach attributed to the remarks of Malan JA in R v Mlambo 1957 (4) SA 727 (A) at 738B–D, and concluded, correctly in our view, that the furnishing of a false account by the accused which indicated an awareness that the vehicle in question had been stolen, did not, by itself, justify the further inference that they had taken part in the housebreaking and theft. The cumulative effect of all the circumstantial evidence, however, was, in the court’s view, sufficient to establish their guilt.

Mtyhida v S [2013] 2 All SA 335 (ECG)

This case, too, states that the lies of an accused do not always warrant the most extreme conclusion. A portion of the accused’s testimony was found to be false, but its effect was considered to be ‘neutral’, as it did not improve the strength of the state’s case (at [28]), since no evidence had been led to prove the commission of the offence by the accused.

s 152 and s 158: Reconstruction of the record

Davids v S (unreported, WCC case no A571/12, 18 March 2013)

The requirements for reconstructing the record, set out in S v Gora & another 2010 (1) SACR 159 (WCC) (see Commentary on the Criminal Procedure Act on the above sections), were not observed by the trial court in Davids. As Bozalek J observed, Gora underlined that the reconstruction process is part and parcel of the fair trial process. Here neither the appellant nor the trial attorney nor the trial prosecutor was asked to make a contribution to the reconstruction. The court did add, however, that the very long delay (seven and a half years) between the trial and the reconstruction rendered it improbable that any of them could have made a meaningful contribution. The loss of all the recorded evidence, however, made it clear that the record was inadequate for a proper consideration of the appeal by the accused.

s 164(1): No inquiry held to determine if minor child understood nature of oath

Senwedi v S (unreported, NWM case no 27/12, 19 April 2013)

The court followed two decisions of the Supreme Court of Appeal (S v B 2003 (1) SACR 52 (SCA) and DPP, KwaZulu-Natal v Mekka 2003 (2) SACR 1 (SCA); see the discussion in Commentary on s 164) in holding that the failure by the trial court to hold a formal inquiry to establish whether the child understood the nature and import of the oath was not an irregularity: ‘the presiding officer is only expected to form an opinion that the child will/does not understand the nature and import of the oath, and can simply admonish him/her’ (at [14]).

s 166: Duty of court to assist unrepresented accused and to avoid showing bias

S v Sithole (unreported, (604/12) [2013] ZASCA 55, 4 April 2013)

The Supreme Court of Appeal took pains to remind judicial officers that, since an unrepresented accused has a limited appreciation of the legal process and is greatly disadvantaged in legal proceedings, they had to ‘ensure impartiality, objectivity and procedural fairness in respect of the unrepresented accused who lacks familiarity with courtroom technique and legal knowledge in order to ensure a fair trial’ (at [9]). They had an obligation to assist such an accused in all facets of the trial, ensuring that only admissible evidence is placed before the court.

S v Mofokeng 2013 (1) SACR 143 (FB)

In this case the court set out the rules of practice for judicial officers when dealing with unrepresented accused. These have evolved to assist illiterate and indigent accused to ensure a fair trial, and cover the full unfolding of the criminal trial, from an examination of the charge sheet before the accused is called upon to plead until the end of the trial. At all stages, the presiding judicial officer acts as the guide of the undefended accused; he or she should assist the accused whenever assistance is needed in the presentation of the case, should protect the accused from unfair cross-examination, and must ensure that he or she fully understands his or her rights. See the full exposition of the judicial officer’s duties in this regard at [17].

s 166: Questioning by the court and its proper boundaries

S v Maroeli & another (unreported, FB review no 338/12, 17 January 2013)

The trial magistrate was held to have exceeded the permissible boundaries of judicial questioning under s 112(1). He descended into the arena, put too many questions to the accused, caused prejudice by leading the accused to incriminate himself on other charges, and created a situation where no rational connection existed between the proven or admitted facts and the verdict returned. The procedure under s 112, said the court, was not designed to ensnare an unrepresented accused. The excessive questioning unduly exposed the unwary accused to the peril of a conviction on a more serious charge. It led, further, to the inference that the magistrate was not open-minded, impartial or fair. The mark set down in S v Rall 1982 (1) SA 828 (A) was overstepped, leading to a failure of justice.

s 203: Privilege against self-incrimination – does a litigant in civil proceedings have a right to have proceedings stayed pending outcome of related criminal proceedings?

Law Society of the Cape of Good Hope v Randell 2013 (3) SA 437 (SCA)

This question, which has been discussed at some length in the Commentary on the Criminal Procedure Act (s 203 sv The nature and scope of the privilege), and which has divided the lower courts, has finally been answered by the Supreme Court of Appeal. The case for granting such a right revolves around the argument that damning evidence against such a party in the civil case offers him or her no choice but to testify, that this constitutes compulsion to disclose his or her hand, and that there is a contravention of s 35(3)(j) of the Constitution, which entrenches the right not to be compelled to give self-incriminating evidence. Our courts have, in the main, however, not embraced this position, and have taken the view that such an application can only succeed where the party is under a legal compulsion or obligation to show his or her hand before the state has produced evidence in the criminal trial.

In the present case, the court below rejected the latter approach as being ‘a retrogressive step which cannot be justified in the context of our constitutional dispensation’ (see 2012 (3) SA 207 (ECG) at [23] and see Commentary supra). But the Supreme Court of Appeal disagreed with the view of Smith J. Mthiyane DP considered that ‘the golden thread that runs through the previous cases ... is that they all involved sequestration proceedings, in which the examinee respondent was required to subject himself or herself to interrogation or to answer questions put to him or her by the provisional trustee’ (at [12]). In such cases, he added, there was clearly an element of compulsion because s 65 of the Insolvency Act, before its amendment, provided that the person concerned was not entitled to refuse to answer questions, a position ameliorated only by a discretion, exercised by the court, not to interrogate the examinee.

The court concluded that the approach in Davis v Tip NO & others 1996 (1) SA 1152 (W) was sound; that a stay will be granted only where there is an element of state compulsion which has an impact on the accused’s right to silence; and that any extension of the court’s intervention to cases where an applicant merely has a ‘hard choice’ to make ‘would bring the right to remain silent into disrepute’ (at [31]). One has, however, some sympathy for the view expressed by Zeffertt (Annual Survey of South African Law 1996 at 835–7) that, where a person’s ‘hard choice’ is to forego his right to remain silent or to face the likelihood of losing his professional status and, even, his livelihood, a ‘layman might be excused ... for thinking that that consequence was something of a penalty’. 

s 208: Identification evidence by single witness

S v Sithole (unreported, (604/12) [2013] ZASCA 55, 4 April 2013)

The Supreme Court of Appeal stressed the need for caution where the identification evidence rested on a single witness. The evidence in this case was unreliable, being based on what other people, none of whom had testified, had told the witness. There was, further, nothing in the objective facts to corroborate the evidence.

s 208: Cautionary rule – evidence of identification – identity parades 

S v Masomo (unreported, ECG case no CA&R 275/2011, 27 March 2013)

The court restated some important principles relating to evidence of identification in general, and identification parades in particular:

  1. It is a general principle of the law of evidence that evidence of identification must be approached with caution.
  2. An identification parade is based on principles of fairness, and it is the duty of the court to determine whether the parade was proper before relying on the evidence.
  3. The parade should consist of at least eight people, all of whom are similar to the accused in general appearance: see S v Mohlathe 2000 (2) SACR 530 (SCA) at 541ad.
  4. The parade does not have to be perfect in every respect, but it must be easy for the court to conclude that it was conducted in a fair and reliable manner. It had not been in this case, as there had been only one person similar to the accused in the parade; the witness was unable to identify the other two suspects at all; he had, in fact, identified the wrong people in the parade; and his original exposure to the perpetrators had taken place after the intake of alcohol.

s 208: Cautionary rule – evidence of co-accused – corroboration not the only means of invoking caution

S v Maselani (unreported, (511/2012) [2013] ZASCA 21, 22 March 2013)

The Supreme Court of Appeal pointed out, once again, that although corroboration is one of the recognised safeguards to reduce the risk of a wrong conviction, it is not the only one. Absence of gainsaying evidence by a co-accused, or his mendacity as a witness, may be sufficient. If it can be said that the accomplice is beyond all question a satisfactory and convincing witness while the accused is the opposite, corroboration is not required (see S v Mpompotshe & another1958 4) SA 471 (A) at 476F–G).

s 211: Evidence of previous convictions disclosed to presiding officer – irregularity – effect of 

Ndleve v Director of Public Prosecutions, North Gauteng, Pretoria & another (unreported, GNP case no A877/11, 22 April 2013)

The presiding officer had been prematurely informed by the prosecutor of the accused’s previous convictions. It was held that this was clearly an irregularity. The only question was whether it was fatal or whether it could be cured by the fact that the accused, who was an advocate of the High Court, had conceded that the judicial officer could, nevertheless, continue with the trial.

The court pointed out that ss 89, 197, 211 and 271 of the Criminal Procedure Act, but especially s 211, were ‘all against the disclosure of previous convictions at any stage before the conviction of an accused person’ (at [10]). It was, said the court, a long-standing practice that to be so informed of previous convictions was an irregularity that nullified the proceedings as a whole (see S v Mavuso 1987 (3) SA 499 (A)). It did not change things even if the accused pleaded guilty (see S v Mdletye [2005] JOL 13933 Tk) or if the State’s case was overwhelmingly strong (see S v Mofokeng & others(unreported, GSJ case no A421/11, 12 October 2011)).

The concession by the appellant in this case did not cure the irregularity as: first, there was no evidence to show what his practice in the criminal courts was, if any; second, he had been encouraged by the magistrate not to object; and, third, he had been in custody after arrest at the time and, the evidence showed, he would have objected had he been released at the time.

s 211: Evidence tending to disclose previous convictions of accused

S v Khubeka 2013 (1) SACR 256 (GNP)

The practice of allowing prisoners to appear at their criminal trials in prison garb or with leg-irons has been criticised at length by our courts. It appears from this case that they may have run out of patience. Bertelsmann J deprecated, in the strongest terms, a standing instruction that seems to remind officials of the Department of Correctional Services that no person who is a sentenced prisoner or is an awaiting trial detainee may come to court without being shackled. The courts have accepted that such improper practices do not, by themselves, constitute a fatal irregularity which taints the fairness of the trial, but the great potential for prejudice in subverting notions of basic fairness and justice, the presumption of innocence, and the rights to dignity, freedom and equality, make this result far from unlikely. In Khubeka there had been no prejudice, since the same court had, itself, sentenced the accused to prison. But Bertelsmann J warned that the courts would, in future, not hesitate to impose fines if such practices were found to be contemptuous of the court.

s 216: Hearsay – when in interests of justice to receive it

S v Jenkins; S v Moosagie & another (unreported, ECP case no CC29/2010, 4 February 2013)

It was argued that hearsay evidence should be admitted in terms of s 3(1)(c) of the Law of Evidence Amendment Act 45 of 1988 to establish bias sufficient to warrant the recusal of a judicial officer. It was argued, further, that it was in the interests of justice to do so. The court, however, declined to do so, since the inference could properly be drawn that the application had been brought for an ulterior reason, ‘to frustrate the finalisation of the criminal trial’ (at [17]). The application for recusal was thus ‘the product of a collusive effort by the accused and clearly contrived’.

s 216 and s 222: Hearsay – admissible in terms of s 222 – tendered for circumstantial value

Le Roux v Pieterse NO & others 2013 (1) SACR 277 (ECG)

This case deals with the admissibility of a medical report (form J88), made by a doctor who had, since, emigrated to Australia, to the effect that the complainant in a rape case had a tear in her vagina and that her panties, too, had been torn. The court received the evidence under Part VI of the Civil Proceedings Evidence Act 1965, incorporated into criminal proceedings by s 222, on the ground that all the conditions set out in s 34(1) of the former Act had been satisfied. But it was at pains to point out that the report ‘was admitted, not for its testimonial value, but as a statement of the objective facts found by [the doctor]’. Why? If the conditions of Part VI (or any other hearsay exception, including s 3 of the Law of Evidence Amendment Act 45 of 1988) had been satisfied, then the evidence would be admissible for its testimonial or circumstantial value. It is only if the hearsay evidence does not fall within the borders of a recognised exception that this distinction may be material.

s 217 and s 218: Pointing-out made without proper warning of right to legal representation

S v Thyse (unreported, ECG case no CC 75/2012, 20 February 2013)

The accused in this case had unequivocally expressed a desire to secure the services of a legal representative. The police officer in question did not inquire whether he was prepared to proceed to a pointing-out without such assistance, and, further, did not investigate his apparent desire to proceed with the pointing-out, notwithstanding his prior express wish to have legal representation. The pre-printed questionnaire used by her did not specifically highlight a suspect’s right to engage legal representation for the purpose of deciding whether or not to make a pointing-out or to be assisted at such a pointing-out. The court held that the pointing-out was inadmissible, since it was not satisfied that the accused’s rights had properly been explained to him and that he had proceeded with the pointing-out with full knowledge of his rights.

The right to legal representation, said Goosen J, is a fundamental right which seeks to ensure that accused persons are adequately and properly protected at all stages of the pre-trial and trial procedures. At the pre-trial stage, there was ‘a clear duty upon the investigating authorities to ensure that the accused is properly informed of his rights’ (at [17]). There was ‘also a duty to facilitate the exercise of the right in order to ensure that an accused person’s right to silence is protected and that the fairness of the trial is assured’.

It was thus important to find, not only that the accused’s rights were explained to him, but also that he properly understood those rights (see S v Melani & others 1996 (1) SACR 335 (E) at 349g). The court accepted that the right could be waived, but, as was held in Melani, only if the accused knew and understood what he was abandoning.

In this case, where the accused was only 18 years old, and unable even to write or sign his name, it was ‘wholly insufficient to rely on the alleged say-so of the suspect that he/she will in due course obtain legal assistance when he/she appears in court’ (at [20]). If there was any doubt as to his understanding of his rights, the police officer was obliged to take all necessary steps, including the postponement of the pointing-out, to ensure that the suspect’s rights were adequately and properly protected.

s 225: Illegally obtained evidence – accused not informed of right to a search warrant

Mtyhida v S [2013] 2 All SA 335 (ECG)

Money had been found at the appellant’s home, but the police had not informed him of his right to a search warrant before his home was searched. The search of his premises without explaining his rights was unlawful, said Tshiki J, and could not be condoned. The ipse dixit of the police that he ‘simply allowed them to search his home [could] not be countenanced’ (at [34]), and, for the police to claim that he should be found guilty because he did not demand his rights to privacy under s 14 of the Constitution was ‘nonsensical to say the very least’ (at [35]), since he ‘could not have been aware of the existence of these rights because they were not explained to him’.



s 276B(2): Sentencing court’s discretion to determine a non-parole period and the provisions of s 276B(2)

Where an accused is sentenced to imprisonment for two years or longer, the sentencing court may – as part of the sentence – fix a non-parole period which ‘may not exceed two thirds of the term of imprisonment or 25 years, whichever is the shorter’ (s 276B(1)(a) and (b)). In Commentary on the Criminal Procedure Act in the notes to this section the provisions of s 276B(1) and the sentencing court’s discretion to act in terms of this section are discussed with reference to Supreme Court of Appeal decisions like S v Pakane & others 2008 (1) SACR 518 (SCA) and S v Stander 2012 (1) SACR 537 (SCA).

In Mthimkhulu v S (unreported, (547/12) [2012] ZASCA 53, 4 April 2013) a full bench of the Supreme Court of Appeal had an opportunity to interpret the rest of s 276B, that is, s 276B(2), which provides that where an accused who has been ‘convicted of two or more offences is sentenced to imprisonment and the court directs that the sentences of imprisonment shall run concurrently, the court shall, subject to subsection (1)(b), fix the non-parole period in respect of the effective period of imprisonment’.

In Mthimkhulu the trial judge, like Moses in his Parole in South Africa (2012) at 41, took the view that s 276B(2) impels or directs a sentencing court to fix – within the limits set by s 276B(1)(b) – a non-parole period in respect of the effective period of imprisonment whenever it orders two or more sentences to run concurrently. The trial court’s view was that the use of the word ‘shall’ in s 276B(2) made the fixing of a non-parole period peremptory, as opposed to the discretion it has where s 276B(1) applies.

Petse JA, writing for a unanimous full bench, rejected the trial court’s approach. At [16] he held: ‘What s 276B(2) ... does is to enjoin a sentencing court, once it has exercised its discretion under s 276B(1)(a) against the convicted person, to then fix the non-parole period of imprisonment taking cognisance of the provisions of s 276B(1)(b)’. Petse JA pointed out that the definite article ‘the’ in the phrase ‘fix the non-parole period’ – which appears in s 276B(2) – is inevitably a reference to the non-parole period determined in terms of s 276B(1)(a) of the Act. There is, accordingly, no room for an interpretation that the discretion which so clearly exists as regards s 276B(1)(a), is somehow replaced by an obligation once the sentencing court is faced with circumstances as identified in s 276B(2).

Reference was also made to the preamble to the Parole and Correctional Supervision Amendment Act 87 of 1997, which inserted s 276B. According to this preamble, one stated objective was to provide that ‘a court sentencing an offender ... may fix a non-parole period’. The word ‘may’, said Petse JA, strongly suggests [an] intention to give courts overall latitude in deciding whether or not to fix a non-parole period’ (at [14]).

Leach JA, apart from agreeing with the judgment of Petse JA, also wrote a separate judgment advancing additional reasons in support of the interpretation that s 276B(2) placed no obligation on a sentencing court. At [29] Leach JA (Shongwe JA concurring) pointed out that – in contradistinction to s 276B(1)(a) – there is in s 276B(2) no provision as regards a time period; and s 276B(1)(a)which limits a court’s discretion to sentences of imprisonment of at least two years before a non-parole period may be determined, must also govern s 276B(2) in that the ‘sentences of imprisonment’ referred to in s 276B(2) ‘must include at least one sentence which the court in exercising its discretion under s 276B(1)(a) has determined should be subject to a non-parole period’.

In Mthimkhulu the Supreme Court of Appeal also confirmed its earlier decision in Stander (supra) to the effect that the parties are entitled to address the court on, first, whether a non-parole period as provided for in s276B should be imposed and, second, what the duration of the non-parole period should be (at [20]). It was, furthermore, also reaffirmed that a court should exercise its discretion to impose a non-parole period only in exceptional circumstances (at [23]). See further the discussion of s276B in Commentary, sv Case law: Aggravating factors and the requirement that there must be exceptional circumstances

s 324(c): Lost record, acquittal and subsequent prosecution

The constitutional right to a fair trial, it was held in Davids v S (unreported, WCC A571/12, 18 March 2013), is infringed where the merits of the appeal cannot be assessed because of a missing record, or almost complete absence of a record, of the trial proceedings. In such an instance the conviction and sentence must on appeal be set aside.

But in Davids (supra) the respondent’s counsel submitted that the court of appeal should make an order in terms of s 324(c) to the effect that a prosecution de novo of the appellant by the state was possible, taking into account factors such as the original date of incarceration and the date of the appellant’s possible release on parole. Section 324(c) provides that a plea of prior acquittal shall not be valid where an appellant’s earlier conviction is on appeal set aside on account of ‘technical irregularity or defect in the proceedings’. See the discussion of this section in Commentary on the Criminal Procedure Act.

Bozalek J (Boqwana J concurring) declined to make such an order. Section 324, he said, does not envisage a prior order by the court of appeal to the effect that the conviction and sentence were on appeal set aside on account of a technical irregularity or defect. Furthermore, such a court order was also not a necessary prerequisite to the State reinstituting prosecution: ‘It is for the Director of Public Prosecutions or his/her delegee to form a view on the matter and take a decision on whether to re-institute proceedings or not’ (at [16]). Bozalek J’s refusal to provide the order or declaration as requested, must be supported. It maintains the distinction between judicial and prosecutorial powers. Making an order or declaration like the one requested could easily be interpreted as judicial encouragement that re-institution of a prosecution should take place, whereas that is solely within the discretion of the DPP. The question whether a plea of prior acquittal should succeed, is also something which the presiding judicial officer at the retrial should decide without reference to a specific order or declaration like the one sought in Davids (supra). See further Director of Public Prosecutions, Transvaal v Mtshweni 2007 (2) SACR 217 (SCA) as discussed in Commentary on the Criminal Procedure Act.

Appeal to Constitutional Court on matters of sentencing: Principles to be applied

In Houston v S 2013 (5) BCLR 527 (CC) the Constitutional Court considered an application by a prisoner for leave to appeal to the Constitutional Court against sentence on the ground that non-concurrent sentences, imposed on two different occasions by the Kwazulu Natal High Court, unfairly limited his right to be considered for parole. In September 1997 he was sentenced to an effective 30 years’ imprisonment and in February 1998 he was again sentenced to an effective 40 years’ imprisonment, which was not ordered to run concurrently with the first sentence. This left the applicant with an overall period of 70 years’ imprisonment. In determining the second sentence, the high court concluded that life imprisonment was not appropriate, implying that there were prospects of rehabilitation.

‘The quirk in this tale’, said the Constitutional Court at [5], ‘is that had [the applicant] been sentenced to life imprisonment in respect of his offences he would by now probably be eligible to be considered for parole’. The applicant claimed unfair discrimination and, in the words of the Constitutional Court at [6], argued that ‘[s]omething is wrong with a system which makes the granting of parole easier for persons sentenced to life imprisonment – and thus assumed to be unlikely to be rehabilitated or reformed – than for those, like [the applicant], who at the time of sentencing were considered to have potential for rehabilitation or reform’.

The Constitutional Court granted condonation but refused leave to appeal. Without expressing any opinion on the merits of the applicant’s argument, the Constitutional Court pointed out that the applicant’s ‘possible remedy’ was to seek a High Court review of the parole policies of the Department of Correctional Services (DCS). The Constitutional Court itself ‘should also not ordinarily deal with a review application of [this] kind as a court of first instance’ (at [7]). Furthermore, the applicant’s complaint was not that his trials were unfair, but that the parole policies of the DCS discriminated unfairly against him. The Constitutional Court would ordinarily allow an appeal against sentence only where the appeal ‘raises fair trial issues that may result in a failure of justice’ (at [3]). The question of sentence will generally not be a constitutional matter. See S v Bogaards 2013 (1) SACR 1 (CC) at [42].

In Houston v S (supra) the Constitutional Court, having dismissed the application for leave to appeal, requested Legal Aid South Africa to assist the applicant in deciding what further steps had to be taken in view of the court’s decision.

Minimum sentence in terms of s 51(1) of the Criminal Law Amendment Act 105 of 1997 and determining whether there was one or more rapes

In Tladi v S (unreported, (895/12) [2012] ZASCA 85, 31 May 2013) the accused, having been charged with and convicted of two counts of rape of the same victim, was in the absence of substantial and compelling circumstances sentenced to life imprisonment as prescribed in s 51(1) and Part 1 of Schedule 2 of the Criminal Law Amendment Act 105 of 1997. On appeal to the Supreme Court of Appeal one of the issues was whether the state had indeed proved two separate incidents of rape, proof of which was necessary to have impelled the trial court to impose life imprisonment as provided for in Act 105 of 1997. The Supreme Court of Appeal concluded that there was insufficient evidence to establish two separate acts of rape: the two rapes as alleged stemmed from one sexual encounter which, even on the complainant’s evidence, did not suggest ‘an interruption in the sexual intercourse to constitute two separate acts of sexual intercourse and, therefore, two separate acts of rape’ (at [13]). Indeed, the complainant’s evidence suggested that the sexual acts were closely linked, amounting to a single continuing course of conduct without any ‘appreciable length of time between the acts of rape to constitute two separate offences’ (at [13]). The conviction on the second count of rape was set aside. The compulsory life imprisonment was accordingly also set aside and replaced with twenty years’ imprisonment, the prescribed minimum sentence for a rape offender with two previous convictions. See Part III of Schedule 2 as read with s 51(1) of Act 105 of 1997.

Tladi v S (supra) can now be added to the body of case law dealing with the issue whether, for purposes of the prescribed minimum sentence legislation, one or more acts of rape took place. In S v Mavundla 2012 (1) SACR 548 (GNP) it was held, for example, that a prolonged act of rape without rest or interruption and involving three ejaculations was a single act of rape. Other cases dealing with this issue are discussed in Commentary on the Criminal Procedure Act in the notes to s 277, sv ‘Rape’ and substantial and compelling circumstances. All these cases merely illustrate the absurdity of the legislation concerned: statutory provisions are trying to achieve what should best be left to judicial common sense in the process of determining an appropriate sentence.

Appeals: Sentencing and delays in hearing an appeal

In Malgas v S (unreported, (703/12) [2013] ZASCA 90, 31 May 2013) at [20] Willis JA said: ‘The phenomenon whereby inertia descends upon an appeal ... once bail has been granted to an accused after conviction and sentence, has been recurring with increasing frequency, especially in certain parts of the land’. The problem with this type of delay is that appellants can somehow attempt to ensure that they derive some benefit from a situation engineered by them: There is a principle that – in exceptional circumstances – the long delay between the passing of a prison sentence and the hearing of an appeal, may justify interference with this prison sentence. This is an exception to the rule that a court of appeal should consider only facts known to the court a quo at the time of sentencing. See Malgas (supra) at [17]; S v Michele & another 2010 (1) SACR 131 (SCA): and Commentary on the Criminal Procedure Act in the notes to s 276A and s 277. This type of situation was also recently dealt with in S v Van Deventer 2012 (2) SACR 263 (WCC). In this case there was a delay of eight years between the sentencing of the appellants and the hearing of their appeals. At [78] Blignault J (Saldanha J concurring) held that ‘the delay in this case is ... a significant factor that must be taken into account in favour of both appellants’, given the prejudice it caused the appellants and the anguish they and their families experienced over eight years pending the outcome of the appeal.

However, it should be noted that in a case like Van Deventer (supra) the court of appeal specifically pointed out that it was ‘not ... in a position to blame anyone’ for the ‘inordinate delay’. This finding explains why the delay enured to the benefit of the appellants. Malgas (supra) stands on a different footing. Willis AJA (Navsa and Majiedt JJA concurring) noted that only one matter had to be considered, ‘namely whether the eight year delay ... in and of itself, justifie[d] a lighter sentence’. No such justification was found, because the appellants were responsible for the delay in finalising their appeals: ‘The predicament in which the appellants find themselves is largely of their own making’ (at [22]). The appellants had indeed adopted a supine attitude to the prosecution of their appeals, did as little as possible and hoped that the ‘problem’ would disappear (at [21]). At [22] Willis AJA said:

It will be hard on the appellants and their families that, ten years after their sentencing by the magistrate, they should now have to report to jail to commence serving their sentences. We have anxiously reflected upon the needs of justice in this case, including the requirement that this court should show mercy to and compassion for our fellow human beings. Having done so, the conclusion remains inescapable that, if this court were to regard this case as yet another ‘‘exception’’, it would undermine the administration of justice.

Sentencing: White collar crime

More than half a century ago authors like Taft & England Criminology 4 ed (1964) could with good cause point out that white collar crime ‘is attractive because it brings material rewards with little or no loss of status’ (at 201). The Supreme Court of Appeal has over the last two decades confirmed decisions and sentences which make white collar crime less attractive – by imposing prison sentences of substance. See generally S v Sadler 2000 (1) SACR 331 (SCA) and S v Shaik 2007 (1) SA 240 (SCA).

In Director of Public Prosecutions, Cape of Good Hope v Fielies (unreported, WCC case no A338/12, 21 May 2013) the DPP appealed against the non-custodial sentence and fine imposed by a regional court on an accused who had pleaded guilty to and had been convicted of 39 counts of corruption in terms of s 4(1) of the Prevention and Combating of Corrupt Activities Act 12 of 2004. The respondent, a municipal official, had awarded irregular tenders which resulted in ‘kickbacks amounting to R350 000’ to him (at [3]).

Boqwana J (Griesel J concurring) pointed out that the respondent had not only ‘abused his position of trust and authority’, but had also admitted that he was the initiator of the ‘corrupt relationship’ (at [13]). It was found that the sentence imposed by the trial court ‘was too lenient and not in keeping with the general sentencing approach followed by the courts in white collar crime’ (at [14]). Having found the trial court’s sentence disturbingly inappropriate, the court of appeal set it aside and imposed a sentence of five years’ imprisonment of which two were suspended on appropriate conditions (at [15]). It was also found that the fine paid by the respondent could be repaid to him by the state (at [14]). This repayment by the state to the respondent was probably justified given the fact that he had paid back the amounts received as kickbacks (at [9]) and the fact that he had a young family who would have needed support while he was in prison.


Table of Cases

Click here to see the full Table of Cases in the .pdf version of the Criminal Justice Review No 1 of 2013.






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

> Click here to unsubscribe from this communication


> Jutastat e-publications user helpdesk

> Juta Customer Services

> Juta Law Marketing


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

Visit the Juta Law Website

2019 © Juta and Company Ltd