Private defence is the civilized remnant of the ancient system of private vengeance as redress for wrong done. The Romans, in permitting self-help formulated the principle moderatio inculpatae (moderation in self-defence) which the European jurists later relied upon to develop a coherent doctrine of private defence. Certain types of intentional killings were no longer regarded as unlawful and therefore are not punished as murder. South African recognizes that killing is justifiable and therefore not murder. Despite the sound rationale underlying the defence, namely the upholding of justice theory where people acting in private defence perform acts where they assist in upholding the legal order, and despite the fact that the defence is established in both criminal law legal theory and practice, there are threshold problems with this rationale which has been subject to much academic criticism. Such criticism must be seen in the context of the wider debate surrounding the circumstances in which battered women kill their abusers - normally in circumstances where the threat is not imminent and therefore the need to uphold justice is not necessary. The purpose of this enquiry is to examine the development and functioning of the defence and more particularly to do so in light of a comparison with the means currently utilized to criminalize conduct falling outside the bounds of self-defence: one of the parent systems of South African law, namely English law and the United States, where battered woman syndrome originated and a profound influence on the way in which the elements of the defence are interpreted in that jurisdiction. For instance, in American law subjective tests for self-defence have been developed such as the particularizing standard. This standard asks whether a reasonable person with the accused’s particular non-universal characteristics would have both perceived the situation as the accused perceived it and would have reacted to that perception by committing the accused’s self-defensive act. If the answer is yes, then the act is considered reasonable. It assumes that individuals freely choose how to perceive and respond to a threatening situation but also acknowledge that certain kinds of nonuniversal characteristics (such as battered woman syndrome) exercise such a powerful causal force on individuals perceptions and actions that it would violate the voluntary act requirement when holding that individual who possess such a characteristic to a standard of conduct that does not take that characteristic into account. The study concludes with an assessment of the form the defence ought to take. In South African law the defence consists of the conditions relating to an attack which includes: an attack, and protected interest and the attack must be unlawful. In respect of the conditions relating to the defence, the defence must be reasonably necessary to avert the attack and the defence must be directed against the attacker. Aspects of these elements have proved to be controversial. In particular, the condition of reasonably necessary to avert the attack has been called into question. Furthermore the requirement of imminence has been rendered especially controversial especially when viewed from the battered woman’s perspective where battered woman syndrome plays a role i.e. the woman’s internal makeup having an influence on the way she views the situation as opposed to an objective test is used to establish if the threat was imminent. While the English and American law elements of the duty to retreat, proportionality and reasonableness approximate the equivalent condition of reasonably necessary to avert the attack, the focal point of this defence in these jurisdictions has similarly been the imminence requirement and the test utilized for self-defence i.e. objective or subjective standard. Prior to evaluating the utility of these elements, the various rationales posited as a justification for the defence will be examined. It is submitted that while various rationales have been posited to form the basis of selfdefence, the autonomy theory (narrowly circumscribed) should be followed in South African law and that the traditional elements for self-defence should remain in force. Regarding the requirement that the attack be reasonably necessary, it is submitted that the traditional mechanism for distinguishing justified from unjustified self-defensive acts should remain an objective test. This is so because by taking account of the knowledge the defender has of her attacker the legal requirements of private defence will eventually be equated with those required for putative self-defence. If putative selfdefence goes to the issue of culpability, which is seen as a particular mental attitude or state of mind - South African law will be evincing a move toward a normative concept of fault. Such an approach has not proved unproblematic in South African law. Both early and modern common law as well as modern case law has expounded a coherent statement of the elements of self-defence which include imminence as a core feature. The problem is that traditional imminence rules do not cater adequately for the battered woman’s situation and for this reason theorists have advocated its abolition. The obvious problem with such a recommendation is that something must stand in its stead to distinguish legitimate cases from illegitimate cases of self-defence. In respect of the imminence requirement, the problems created by this standard cannot be solved by replacing imminence with necessity or by claiming priority for necessity or by demanding that imminence means pacifist rather than the libertarian version of necessity. These positions pose the question but do not answer it. Furthermore, if the imminence question cannot be answered by assuming one side of the necessity debate, then it cannot be answered by referring to the distinction between justification and excuse. It is submitted that “instead of viewing objectivity as not being able to account for battered woman’s situation – the opposite conclusion should be reached – that by rethinking certain situational factors as a set of relatively innocuous and perhaps necessary normative propositions then the abused woman’s situation is consistent with some very standard propositions in the law of self-defence. If the abused women is being attacked and the threat is imminent (in the traditional sense), then she should be able to avail to herself of self-defence, although it should be noted that the court should also consider the fact that the battered women placed herself in this dangerous situation. However, the court would also have to take into consideration the difficulty that the abused woman faced in extricating herself from this position. On the basis of a discussion of the various construals that inform the question of whether proportionality should form a necessary requirement for self-defence, including (i) the liberal aspiration to neutrality, (ii) constitutional norms and (iii) a duty of social solidarity to the state, it is submitted that proportionality should form an integral part of the requirements for self-defence. The test can be set out as follows: not only must the defence be necessary but also the means used by the accused for the purpose of averting the attack must be reasonable in the circumstances. This is in accordance with the autonomy theory. Therefore, would an “ordinary, intelligent and prudent person in the accused’s situation would react to establish if the self-defence claim was justifiable. However, it is submitted that not all the characteristics of the accused should be taken into account. Only those “characteristics which have the most (or direct) bearing on the accused’s situation” should be considered. Despite the rationales underlying self-defence, it has not been entirely clear whether an abused woman is expected to flee. It is submitted that there should be a duty to retreat. In the case of the abused woman, her situation is adequately catered for within the reasonableness neutrality perspective. In respect of the defence of provocation, Roman and Roman-Dutch law did not regard anger, jealousy or other emotions as an excuse for criminal conduct, but only as a factor which might mitigate sentence, if the anger was justified by provocation. South African law with its parent system in Roman-Dutch law might have followed this lead had it not result of the Transkei Penal Code of 1887, it envisaged a type of a partial excuse: even if been for the introduction of the mandatory death penalty for murder in 1917. In 1925 as a killing was intentional, homicide which would otherwise be murder maybe reduced to culpable homicide. The test for provocation was thus an objective one. By 1949 in R v Thibani it was held that provocation was not a defence but a special kind of material from which in association with the rest of the evidence the court should decide whether the accused had acted involuntary or without intent to kill. This introduced a subjective test for provocation. But a number of crucial issues remained unresolved; could intense provocation or emotional stress serve to exclude criminal capacity or voluntary conduct. After the decision in Chretien, the question arose, if severe intoxication could exclude these basic elements of liability then could it not also exclude provocation or emotional stress. At this point, the notion of criminal capacity came to the fore. This notion was unknown in South African common law and was adopted from Continental Legal systems, specifically Germany. The notion took hold with the Rumpff Commission of Inquiry into the Responsibility of Deranged Persons and Related matters, the recommendations of which gave rise to the provision of section 78 (1) of the Criminal Procedure Act. In S v Mahlinza set out that criminal capacity of actor is an essential requirement necessary to establish criminal liability. Criminal capacity consists of cognitive component i.e. ability to distinguish between right and wrong and conative capacity i.e. the ability to act in accordance with the distinction. If either was lacking no liability would ensue. In S v Van Vuuren, the court expressed in unequivocal terms that the accused could not be held liable where failure to comprehend what he is doing is attributable to a combination of factors such as provocation or emotional stress. The very idea of allowing provocation to function as a defence excluding an accused’s criminal liability is inherently controversial. From a moral and ethical perspective people are expected to control themselves, even under provocation or emotional stress. To allow it to function as a complete defence as opposed to mitigating factor means that it gives credence to the belief that retaliation is justified in the eyes of the law and this is the very thing criminal law guards against. Despite the well established nature of the defence of non-pathological incapacity, the law has been thrown into flux by the decision of the Supreme Court of Appeal in S v Eadie which constituted a serious erosion of the notion of criminal capacity, with a concomitant “ripple effect” on other topics within the general principles of criminal law. The question this case has highlighted for South African law of non-pathological incapacity is whether the boundaries of the defence have been inappropriately extended. This is so since the court held not only that there is no distinction between the defence of automatism and nonpathological incapacity, and that it would have to be established that the accused acted involuntarily in order for her defence of lack of capacity to prevail, but furthermore held that the court should assess the accused persons evidence about his state of mind by weighing it against his actions and surrounding circumstances, thereby introducing an objective test. Theorists such as Burchell have considered this move “bold” and “encouraging” for its emphasis on objective norms, and the fact that it brings it into line with both the English and American jurisdictions, where not only is an objective element introduced into the enquiry, but where loss of self-control is not totally excusable since the law assumes that provoked party was not totally incapable of controlling anger. If an accused was unable to control himself, a full excuse would be defensible. The notion of capacity has its approximate equivalent in the English and American law of provocation where the jury must consider the subjective question of whether the accused was actually provoked to lose self-control, the defence requires that a reasonable person in the same circumstances would have lost-self control and acted as the accused did. The South African notion of capacity is examined with reference to the way provocation is treated in these jurisdictions. Should non-pathological incapacity be equated with automatism, the established precedent in provocation and other cases of non-pathological incapacity would have to be revised by implication, and would have serious implications for the principle of legality and restricting the scope of the defence for battered women. Furthermore, it is submitted that a move towards an objective test should not be followed. This is so since such an approach does not extend to encompass the battered woman’s mental and emotional characteristics including recognized psychological disorder symptoms. This results in the court not having any meaningful way to determine whether the battered woman lost self-control and furthermore it will lead to increasing attention being directed at how far the objective test be tailored to fit the capacity of the accused. The problem with the capacity test is that it has created via the Criminal Procedure Act a new element of liability by drawing from both the general physical and the mental liability enquiries. Therefore, by duplicating the voluntary act requirement under mens rea, the courts have asked the same question twice. Once the accused is shown to be acting voluntarily, there will be a measure of goal-directed conduct. Where goal-directed conduct is present, it necessarily implies that here must be a level of capacity present in the case of the defence of non-pathological incapacity. In other words, the question is not whether capacity is present, but to what extent it is present. This point is not acknowledged by our courts: the concept of psychological fault underlying South African law offers no explanation for the fact that culpability is capable of gradation. The effects of battering could be used to support a defence of diminished capacity, which focuses not on mitigating circumstances of the act, but rather on the actor’s inability to form the requisite mens rea for the offence charged. However, the introduction of such a defence could only be achieved by returning to the rules relating to provocation followed in South Africa prior to 1971. According to the specific intent doctrine, policy considerations require that an accused should not be completely acquitted. However, these considerations require that an accused should not be completely acquitted. However, these considerations also require that an accused not be convicted of murder but of culpable homicide. This compromise solution (of culpable homicide) can only be reached by treating provocation as a special defence, one which is not strictly adjudicated in terms of the general principles relating to culpability (mens rea). Furthermore, it is submitted that a subjective test must be applied, since Snyman’s objective-subjective test leads to an illogical confusion between the subjective and objective elements.

Thesis (LL.D.)-University of KwaZulu-Natal, Pietermaritzburg, 2009.