Reformulating dolus eventualis Guidance from USA and Germany South African CRIME

Dolus eventualis has correctly been described as an ‘enigma’. Not only has it been variously described by the courts, but the courts have applied the two-stage test without providing an in-depth analysis of what it means. Both dolus eventualis required for murder and conscious negligence required for culpable homicide, contain an element of subjective foresight of the remote possibility of death occurring. As a result, the distinction between murder and culpable homicide has become confused over the years, and is evident in the courts vacillating between findings of murder and culpable homicide. Considering the lack of clarity, this article examines the test for dolus eventualis in the case of murder and determines whether it can be more clearly distinguished from culpa , in the case of culpable homicide. German and American law and academic opinion are consulted in order to establish how the respective countries have dealt with the conflation of murder and negligent killings.


Introduction
MacKinnon 2 correctly points out that the meaning of murder is not self-evident, and that 'both its definition and status relative to other forms of homicide present serious difficulties in criminal law theory'. Murder is defined as the intentional unlawful killing of a person, whilst culpable homicide is defined as the negligent unlawful killing of a person. 3 The sole difference and distinguishing feature between these two crimes lies in the fault element of the crime which determines whether the unlawful conduct was carried out intentionally or negligently. 4 Intention does not mean that the accused must have aimed, wanted or meant to commit the crime in question, 5 and therefore an accused's intention includes his 'conscious acceptance of such risks of unlawful conduct as he foresaw occurring whilst he was pursuing some other aim or object, whether lawful or unlawful'. 6 However, MacKinnon 7 states that it is this 'extension of the concept of intention to include foreseen consequences which is at the root of the mens rea problem'. Once we include reference to foresight of consequences, the blurring of the distinction between intention and negligence begins. 8

Dolus eventualis in South Africa
Dolus eventualis has been recognised by the Constitutional Court in S v Coetzee 9 and Thebus v S, 10 and forms an integral part of criminal liability in South Africa. The courts conduct a two-stage test to determine whether the accused possessed dolus eventualis. The first stage, the cognitive component, asks whether the accused subjectively foresaw the possibility of causing death. The controversy associated with the cognitive component has been framed in a question by Hoctor 11 who asks, 'should the cognitive component be limited to foresight of a real or reasonable possibility of harm, or does foresight of a remote possibility suffice for intention?' The vast majority of cases to date 12 have established that the degree of foresight needed to establish the cognitive component of dolus eventualis is merely 'the possibility of harm occurring'. 13 In terms of Black's Law Dictionary, 'possibility' has been defined as 'an uncertain thing which may happen' 14 and therefore the harm which results need not have been a certain result of the accused's conductbut there exists a chance that the harm may or may not ensue.
The second stage of the test, the conative component, entails that the accused subjectively reconciled himself to the possibility of death ensuing, 15 which means that the accused decided to proceed with his action despite possessing such foresight. In Director of Public Prosecutions, Gauteng v Pistorius, 16 the Supreme Court of Appeal held that 'the possibility of death is foreseen which, coupled with a disregard of that consequence, is sufficient to constitute the necessary criminal intent.' Therefore, to him it is immaterial whether death results from his actions, 17 and he does not allow the possibility of killing another human being to deter him from proceeding. In other words, he consciously accepted the risk. 18 There is rarely direct evidence of the existence of the conative component and, therefore, it is inferred from the accused's deliberation and preparation, together with a failure to render assistance. 19 An array of terminology exists for describing the conative component. Some judgments 20 refer to the conative component as 'insensitive recklessness' or 'callous indifference'. However, the accused's feelings toward the risk is irrelevant when determining the conative component, and it is immaterial whether the accused hoped that the risk would not materialise. 21 What matters is that the accused consciously proceeded to take the risk. Not only has the conative component been labelled redundant but a lack of clarity exists as to its exact meaning in that it has been variously defined. 22

Conflation of dolus eventualis with conscious negligence
The relationship between dolus and culpa has become a grey area of law since S v Ngubane 23 in which the Appellate Division held that 'a man may foresee the possibility of harm and yet be negligent in respect of that harm ensuing'. The court held that 'the concept of conscious negligence clearly establishes that foresight per se does not exclude negligence'. 24 Conscious negligence occurs when the accused foresaw only a remote possibility of harm ensuing, but unreasonably trusts or is confident that the harm will not occur, and therefore failed to take the steps that a reasonable person would have taken to prevent harm. 25 Kemp et al 26 state that the greater the risk of the possibility of death ensuing, the greater the duty placed on a person to prevent the possibility from materialising. Conscious negligence therefore differs from traditional objective negligence which consists of a failure to measure up to the foresight required of the reasonable person. 27 However, the courts rarely refer to conscious negligence and the most commonly quoted passages on conscious negligence come from S v Van Zyl 28 in which the Appellate Division found that an accused who foresaw the harm may be guilty of negligence only. Therefore, the main difference between dolus eventualis and conscious negligence rests on whether the accused reconciled himself to the foreseen possibility -irrespective of the degree of foresight.
It is therefore not the degree of foresight which is the determinant of whether dolus or culpa exists but how the accused reacts to foreseeing the possibility of death. In terms of this line of reasoning, dolus eventualis will be present where the accused, accepting that death may result from his conduct, willingly decides to take a chance that it will not follow. However, where the accused unreasonably decides that death will not ensue, he will be guilty of culpable homicide based on conscious negligence. Jansen JA in S v Ngubane 29 held that 'the distinguishing feature' of dolus eventualis is the 'volitional component' and that provided this component is present, it makes no difference whether the accused foresaw the possibility 'as strong or faint, probable or improbable'.
Kemp et al 30 believe that the Appellate Division 'took a wrong turn with S v Ngubane' 31 in that it has introduced 'unnecessary confusion into the requirements for dolus eventualis'. Academic opinion, however, mostly favours a conative component requiring that the accused should have 'reconciled himself to the consequences' and therefore 'accepts the consequences into the bargain'. 32 Hoctor 33 states that 'the critical consideration for the purposes of criminal liability for harm caused to others is the accused's mental state in respect of such harm to others.' It has been argued, nevertheless, that the conative component adds no value and should be abandoned in favour of the requirement that 'subjective foresight must have existed contemporaneously with the unlawful conduct'. 34 According to this view, dolus eventualis should be established only by subjective foresight of the unlawful conduct. Burchell and Hunt 35 submit that where the accused foresaw harm as a real possibility but nevertheless persisted in their conduct, they consciously took the risk of it happening and therefore possessed dolus eventualis. Paizes 36 states that the conative component is therefore 'a notion without utility' and that an accused who carries on certain conduct 'reconciles himself' to nothing more and nothing less than what he foresaw. In this regard, Whiting 37 states that by acting with foresight of the possibility that a result will ensue one necessarily reconciles oneself to the possibility that it will ensue. It can, therefore, be argued that an accused can never come to the conclusion that harm will ensue where he foresaw it merely as a remote happening and, conversely, an accused cannot legitimately argue that he concluded that death would not ensue when it was foreseen as real, reasonable or substantial.
Burchell and Hunt 38 state that the minimum degree of foresight required is foresight of a substantial or real possibility, which will confine intention 'to a state of mind that can properly be regarded as such and keep the dividing line between intention and negligence clearcut'. Burchell 39 states that where an accused foresaw the possibility of harm materialising as something less than a real possibility, but instead as a remote possibility, conscious negligence rather than dolus eventualis would be present, and the accused would be found guilty of culpable homicide: Where X causes the death of Y which X foresees as a remote possibility his liability for culpable homicide would turn upon whether or not his taking of the risk of Y's death was justified, judged by the objective standard. In deciding this question many factors would be relevant, eg the degree of remoteness of the risk, whether the risk which the accused takes has a social value which outweighs the social harm of the danger inherent in the risk, the urgent and laudable action in which the accused is engaged and whether the precautions may have been so difficult, inconvenient and costly. 40 There is merit in the argument proposing that foresight of a real or substantial possibility be confined to murder, while foresight of a remote possibility be confined to culpable homicide. Negligence is referred to as the 'junior partner' 41 of dolus, because not only does it require a lower level of culpability, but it also involves a lesser sentence. It follows that dolus, involving a higher degree of stigma, harsher sentences and requiring a higher level of culpability, should require a higher degree of foresight than negligence. It is imperative that the definition of murder should ensure that those convicted of murder will be deserving of the stigma associated with it. 42 Therefore, Whiting 43 argues that 'in order to put the law on a sound footing, it will be necessary to reject the notion that the foresight required for dolus eventualis need not be of anything more than a remote possibility as being far too wide'. It has been argued that expanding the scope of murder would not only be unjust, but would also diminish the stigma to which society attaches value. 44 Morkel 45 states that if foresight of a remote possibility constitutes sufficient foresight for dolus, then an accused could be held liable for murder where his conduct did not even fall short of the reasonable person. This could lead to 'anomalous and unjust results' 46 because by extending the scope of foresight to a bare possibility, wrongful convictions could ensue. Burchell and Hunt 47 correctly state that 'if applied, it would lessen confidence in the administration of justice, for it extends the scope of intention to embrace a state of mind which cannot properly be classified as intention at all'.
However, according to Hoctor, 48 the argument that foresight of a remote possibility is never applied in our courts is open to doubt. He refers to S v Nkosi 49 in which the court convicted on the basis of foresight of 'no more than a remote possibility [own emphasis].' In S v Mazibuko, 50 the court refers to the court a quo's decision to convict the accused of murder where 'the death of the deceased was foreseen as no more than a remote possibility [own emphasis].' However, in both cases, the accused's conduct had no social utility and firearms were used, which the accused would have foreseen that if used would pose a real possibility of death. Weldon 51 argues that 'everyone knows that some weapons, such as loaded guns…when used in a dangerous manner, are likely to produce death.' It is contended that where a deadly weapon is utilised or an instrument is used in a deadly manner by the accused, unless it is proved otherwise, inferential reasoning dictates that the accused must have foreseen the real, reasonable or substantial possibility that death could ensue. According to Glanville Williams, 52 and endorsed by Pain, 53 foresight of a bare possibility is sufficient to convict for dolus only if the accused's 'conduct has no social utility, but that the slightest social utility of the conduct will introduce an inquiry into the degree of probability of harm and a balancing of this hazard against its social utility'.

Overview of South African case law
In S v Beukes, 54 the Appellate Division held that it is highly unlikely that an accused will admit to or it will be proved that he foresaw a remote consequence and that it needs to be established that it was reasonably possible that harm would ensue. Hoctor 55 states that this judgment 'by no means excludes foresight of a remote possibility' by referring to where the court states that 'liability for dolus eventualis will normally only follow where the possibility is foreseen as a strong one.' However, Paizes 56 states that the court's use of the word 'normally' covers those 'exceptional cases where foresight of a possibility, however remote, should be viewed as sufficient' such as where the conduct has no social utility or its purpose is to expose the victim to death. Therefore, foresight of a remote possibility should be viewed as the exception, not the norm.
In S v Humphreys, 57 the Supreme Court of Appeal favoured an unqualified degree of foresight. The accused was driving a minibus carrying fourteen schoolchildren when he collided with a train resulting in the death of ten children and injuries to four passengers and himself. On appeal, the court reasoned that every person of normal intelligence would recognise that disregarding the warning signals of an approaching train, and avoiding the boom aimed at stopping vehicles from entering a railway crossing, may result in a fatal accident. 58 The consequence of such a recognition is foresight on the part of 'every right-minded person' that disregarding these safety measures creates the possibility that the foreseen harm may ensue. 59 However, the court found that the possession of foresight alone is insufficient. The court concluded that the appellant foresaw the possibility of a collision occurring, but 'he took a risk which he thought would not materialise.' 60 The court held that because the appellant had previously successfully performed this manoeuvre he believed that he could repeat it without harm, and that such belief constituted negligence. 61 However, this belief could rather be argued on the basis that the accused did not foresee death as a real possibility. 62 The court, further, held that where the accused did not foresee himself being harmed then he cannot be said to have done so with others. 63 The accused's murder conviction was replaced with culpable homicide. Burchell 64 correctly argues that the court neglects to ask whether the accused foresaw death as a real or substantial possibility? Burchell 65 states that the fact that the accused had previously successfully executed such a manoeuvre cannot override the inference that he foresaw that there was a real possibility of failure this time. The accused took a substantial risk in which the social cost outweighed the benefits of the risk and, by doing so, displayed an extreme indifference to the value of his passenger's lives. Hoctor 66 asks, 'should every driver who causes death then be charged with the crime of murder, with the associated heavy sentence and stigma that follows a conviction for murder?' Burchell and Hunt 67 correctly state that, in the event of a fatality from a car accident, 'if in the circumstances he foresaw Y's death as a real possibility, a verdict of murder would be justified'. Whiting 68 states that a verdict of murder would be justified where the driver deliberately took a specific concrete risk. Whiting 69 in this regard provides the following example: A driver who wishes to make a quick getaway drives straight at a person standing in his path, hoping that he will get out of the way but realising that unless he manages to do this he will be hit and perhaps killed. Here the risk to the other person's life which the driver has knowingly taken is of so immediate and concrete a nature that he may well be held to have acted with dolus eventualis.
Associated words with 'concrete' include 'certain', 'real', and 'substantial', 70 and consequently this type of risk taken by the accused indicates that he possessed foresight of a real or substantial possibility. It can, furthermore, be argued, relying on inferential reasoning and the dicta of S v Mini 71 in which 'a trier of fact should try mentally to project himself into the position of that accused at that time', that the accused did in fact accept that possibility into the bargain. Human experience dictates that, for example, an accused who drives straight into a person cannot unreasonably trust that the person will move out of the way -but can merely hope.
In S v Dlamini, 72 the Supreme Court of Appeal held that once it is inferred that the accused subjectively foresaw the real, reasonable or substantial possibility of death, 'credibility is stretched beyond breaking point' where that accused denies that they accepted that death would ensue. In S v Qeqe, 73 the court states that because the accused foresaw death as a real possibility, he can, as a logical inference, be said to have reconciled himself to the death. The court's findings are solely based on the degree of foresight possessed by the accused at the time of committing the crime, and therefore the accused's foresight -which reflects his state of mind -indicates a willingness to kill. The conative component, therefore, consists merely of the accused having proceeded to carry out the risky conduct, despite possessing foresight of something more than the merely possible.

'Bedingter Vorsatz' and 'bewuste Fahrlassigkeit' in Germany
Bedingter Vorsatz, the equivalent of dolus eventualis, exists when the accused foresaw as a possible result of his actions that harm would be caused to another and approved or reconciled himself to that possibility. 74 Therefore, the first leg of the test is concerned with the knowledge and assessment of the possibility of harm by the accused, and the second leg is concerned with the accused's attitude towards the harm. 75 In the Stakic judgment 76 the definition of dolus eventualis in German criminal law was described: 'if the actor engages in life-endangering behaviour [own emphasis], his killing becomes intentional if he reconciles himself or makes peace with the likelihood of death'. Therefore, Taylor 77 states that the crucial question is whether the accused was 'prepared to run the risk, knowing that it might materialise and being reconciled to that possibility?' Germany also recognises conscious negligence, called bewuste Fahrlassigkeit, which contains the same intellectual element as dolus eventualis: the accused is guilty for having carried on conduct, despite realising that such conduct could lead to unlawful consequences. 78 However, dolus eventualis entails the accused having approved of the possible consequences, whereas in the case of conscious negligence, the accused disapproved of them and was confident that such a consequence would not occur. 79 A distinction is drawn between Hoffen (hope) and Vertrauen (reliance): the hope that foreseen consequences will not ensue does not eliminate intent, but reliance on the possibility of avoiding or preventing these consequences, whether rational or not, does eliminate intent. 80 With regard to cases of murder, the inhibition level theory (Hemmschwellentheorie) is applied according to which the intent to kill a person requires the accused to overcome a high inhibition level. 81 This high inhibition level is considered to be overcome when the death of the victim is so likely that 'only a fortunate coincidence could have averted it'. 82 The example put forward is where the accused stabs the victim in the heart. 83 Therefore, German law provides that due to the severity of the crime of murder and how it is treated, the death of the victim cannot be a 'remote' happening. This is because people are generally reluctant to undertake a violent act. 84 According to the 'theory of probability', dolus eventualis exists where the accused foresees the harm as probable, while conscious negligence exists only when the accused foresees the harm as merely possible. 85 The question which has been left unanswered is 'when does foresight reach the level of probability?' One recommendation that has been made to address the problem of vagueness is to reformulate the theory so that intentional conduct occurs when the accused thought that death ensuing from his conduct was more probable than not. 86 Taylor 87 states that the 'theory of probability' is an attempt to define dolus eventualis based on the view that, once a certain level of foresight beyond a remote possibility has been reached, and a person who wanted to avoid causing harm would modify his conduct accordingly or refrain from it altogether, intention has been proved. Taylor 88 goes on to say that proceeding to act in this situation justifies an assumption about the accused's guilt, which seems to be higher than in cases of foresight of the remote possibility of harm ensuing and allows us to dispense with the conative component of dolus eventualis. He argues that 'one who foresees a possible consequence of her actions but goes ahead anyway must approve to some extent of that consequence, or else she would not have gone ahead'. 89 Therefore, no convincing argument has been put forward which justifies the need for a conative component and he validly points out that for dolus indirectus, which requires certain knowledge, no conative component is required. 90 Thus, why does dolus eventualis require a conative component? Morkel 91 demonstrates how prominent German academics such as Schonke, Schroder, and Welzel were on the brink of accepting that dolus eventualis entails that the accused acts 'despite the knowledge that his conduct possesses those inherent qualities that presuppose his culpability -i.e., we blame the accused for having acted despite foreseeing the relevant consequences'. However, Jescheck 92 rejected the test of probability, because he thought that it was possible for the accused to 'trust' that harm will not occur -despite the fact the he foresaw it with a high degree of probability. In response, Morkel 93 correctly states that 'if such a mental state were at all possible, it would be that of a totally unrealistic, irresponsible optimist -and would have nothing to do with real-life experience.' Frisch developed the 'risk-recognition theory' which asks whether a risk existed that was known to the accused and offends the legal system. 94 Frisch 95 states that the reason we impose harsher sentences for intentional conduct than negligent conduct is that the intentional actor has a greater degree of control over their conduct, consciously disobeyed the law or at least took the risk of violating it, and has a greater personal responsibility for the violation than the negligent actor. Therefore, intention cannot be defined by reference only to knowledge that harm may occur, as this is not in line with the reasons as to why intention is punished more harshly. 96 The requirement that the conduct must offend the legal system provides a solution to cases that cannot reasonably be called intentional, such as the overtaking driver, and addresses the problem of the Russian roulette player who believes that there will be a 'happy ending'.

'Extreme indifference' murder in America
American law is also faced with difficulties when trying to demarcate cases of manslaughter, the equivalent of culpable homicide, from 'extreme indifference' murder that occurs under substantially the same circumstances as dolus eventualis. Section 210. [a] person acts recklessly... when he consciously disregards a substantial and unjustifiable risk [own emphasis] that the material element exists or will result from his conduct. The risk must be of such a nature and degree that, considering the nature and purpose of the actor's conduct and the circumstances known to him, its disregard involves a gross deviation [own emphasis] from the standard of conduct that a law-abiding person would observe in the actor's situation. Therefore, with regard to the gravity and utility of the risks, the Code uses the words 'substantial' and 'unjustifiable'. Taking a risk of death does not raise a question of liability unless the risk is substantial and 'the social costs outweigh the benefits of the risk'. 98 In People v Suarez, 99 the New York Court of Appeals held that 'depraved indifference is best understood as an utter disregard for the value of human lifea willingness to act not because one intends harm, but because one simply doesn't care whether grievous harm results'. 100 Abrahmovsky and Edelstein 101 propose that 'extreme indifference' murder should 'require that the defendant consciously disregards a risk that is caused solely [authors' emphasis] by his own conduct' -they provide a hypothetical example to demonstrate this: Someone who drives at a high speed on the sidewalk would create a risk of death to pedestrians that would not otherwise exist. A pedestrian who steps onto a public sidewalk does not ordinarily accept the risk of being struck by a motor vehicle. Therefore, a driver who propels his vehicle onto a sidewalk engages not only in riskdisregarding conduct, but in risk-creating conduct, and might thus be guilty of depraved indifference murder. 102

Conclusion
The research findings demonstrate that the degree of foresight contained in the cognitive component should be qualified. In other words, the accused must have foreseen death as a real, reasonable or substantial possibility, and foresight of anything less will constitute culpable homicide. Due to the high degree of stigma associated with a conviction for murder, the principles of fundamental justice require a level of mens rea that reflects the nature of murder. 103 Therefore, requiring a remote degree of foresight for dolus eventualis does not sufficiently reflect a level of mens rea that is sufficient for murder and the stigma attached. This is why both Germany and America require that there be a high degree of risk for a murder conviction. The courts, in practice, do not find dolus eventualis to be present where the foresight of harm occurring was remote, as a person of normal intelligence cannot accept and therefore intend that death will ensue where they foresaw it as a remote happening. Consequently, it appears that the conative component is rendered redundant, and should be dispensed with. Not only have the courts never delved into explaining the content of the conative component, thereby leaving it a confusing concept, but they also infer this component from cognition by ascertaining whether the accused's acceptance of death can be inferred based on his foresight of the probable result of death.
Therefore, foresight of a probable risk will be taken as proof of the conative component, the result being, as Dubber and Hörnle 104 stated, that the conative component is 'collapsed' into the cognitive component. Furthermore, Taylor 105 validly points out that for dolus indirectus, which requires only the certain foresight of death, no conative component is required. When foresight is of a real, reasonable or substantial possibility, there is no reason why dolus eventualis should contain a conative component merely because there is a reduction in the foresight of the possibility of death. Thus, the arguments put forward that foresight alone does not reflect intention and neglects the accused's state of mind, have no validity. It is for this reason that Loubser and Rabie 106 state that 'dolus eventualis does concern the accused's state of mind, but only in a cognitive sense, in that it requires a conclusion as to whether a harmful result may actually occur in the circumstances'. This conclusion is based on the degree of foresight possessed by the accused at the time of committing the crime, because an accused cannot conclude that a harmful result will ensue when he foresaw it as a remote happening and the courts infer accordingly.
There are, however, cases that cannot reasonably be called cases of intentional acting, such as the Russian roulette player, in which the chance of firing the gun is not probable and the accused therefore believes that there will be a 'happy ending'. The German Risk-Recognition theory in which the accused's conduct must offend the legal system provides a solution to the above-mentioned cases and should therefore be welcomed. This accords with the views that foresight of a remote possibility will suffice when the conduct involved has no social utility, but 'it is the accused's purpose to expose the victim to the risk of death'. 107 Likewise, the slightest social utility of the conduct will introduce an inquiry into the degree of probability of harm and a balancing of this hazard against its social utility'. 108 This can also be said to be in line with Abrahmovsky and Edelstein's reformulation of 'extreme indifference' murder, in which the accused himself created the risk of death because, for example, a car accident is an everyday riskbut when an accused decides to drive on a public sidewalk, it is a risk that would not have been there without the accused's conduct.