Battered But Not Beaten: Women Who Kill In Self Defence Men and women kill for different reasons and in different circumstances. Men account for the great majority of homicides resulting from intentional violence. They are far more likely than women to kill friends, casual acquaintances or strangers. Many of these homicides result from trivial provoc~tions and trivial quarrels. Many are associated with social activities in "pubs, clubs and other places of entertainment7'.l When pride is at stake, minor assaults and insults can prompt a deadly response. The largest single category of killings however, about a third of the total number of fatal attacks by men, are domestic homicides. Women are far less likely to kill than men. When women do kill, their violence is almost always directed against an immediate family member, most commonly a husband or partner. In the deadly simple arithmetic of intimate homicide, the risk of being killed by a partner in a heterosexual relationship is three or more times greater for women.2 The disparity is even more marked in cases of separation homicide: killings motivated by jealousy and possessiveness during separation or as a consequence of separation. In most of these cases, women are killed by men. It is exceedingly rare for women to kill a male partner during or after the decision to separate.3 In her study of New South Wales homicides spanning the years 1968-1981, Alison Wallace found that separation homicides accounted for close to half of the cases in which men killed their wives or lovers.4 Murder is frequently followed by suicide in these cases. The attack is very often premeditated rather than impulsive.5 * Senior Lecturer in Law, University of Adelaide. 1 Wallace, A, Homicide: The Social Reality (1986) NSW Bureau of Crime Statistics and Research, Research Study No 5 at 151. 2 Id at 84. Similar or higher ratios have been found in studies of intimate homicide in other Australian States and comparable jurisdictions overseas. In addition to the sources to which Wallace makes reference, see Law Reform Commission of Victoria, "Homicide Prosecutions Study" Report No 40; Polk, K and Ranson, D, "The Role of Gender in Intimate Homicide" (1992) 24 ANZ J Crim 15. United States studies discussed by Wallace, ibid, suggest a partial exception to the generalisation: "Amongst blacks, wives were more likely to be the aggressors than the victims. The picture was reversed for whites: white women were three times more likely to be victims than aggressors in spouse killings". 3 Above n l at 98-99: separation was relevant in only three of 79 cases in which women killed their partners. 4 Id at 99. Of her sample of 296 cases of domestic homicide, 217 involved killings by men. Of these, 75 were cases of killing after separation and 23 of killing during separation. In addition to her domestic homicides, Wallace found a further 20 or so cases in which men killed women with whom they had a sexual relationship: id at 150. In about half of these cases too, the homicide was prompted by the woman's decision to terminate the relationship. The murder-suicide link is discussed, id, ch 12. 5 Polk and Ranson, above n2 at 22. Allwood (1975) 18 A Crim R 120 is a typical example. 404 SYDNEY LAW REVIEW [VOL 15: 403 When intimate homicide occurred in a continuing relationship, the disparity between the numbers of male and female victims was less marked, though women were still far more likely to be killed.6 Almost invariably, women who were killed by men did not provoke the fatal assault by violence.7 When women killed men, however, the fatal assault was usually a culminative response to a long history of serious domestic violence.8 Half of the women who killed did so in response to a threat of immediate violence.9 The defences to murder provide structure and expression to the exculpatory effects of fear and resentment. The exculpatory effects of fear find their principal expression in pleas of self defence. But fear, even overwhelming fear, will not always excuse murder. Some minimal level of fortitude in the face of danger is required. And the threat which prompts the fatal response must be directed at some defensible right or interest. A threat of separation, for example, is no excuse for killing though it may, possibly, amount to provocation which will reduce the offence to manslaughter. Resentment of past wrongs is a far more problematic basis for claiming an excuse. Killing to avert a threat of harm can be excused, if not justified, on the ground of necessity. But death inflicted in return for harm done in the past makes the defendant a self appointed executioner who usurps the authority of the state.10 The exculpatory effects of resentment find very limited expression in the partial defence of provocation, a plea traditionally hedged with restrictions, which merely reduces intentional killing to manslaughter.11 The rules which limit and define the exculpatory effects of fear and resentment were elaborated, for the most part, in cases of intentional homicide committed by men. The law of self defence was defined by reference to attacks threatening immediate physical injury and shaped by judicial policies designed to restrict recourse to violent methods of settling quarrels or grudges. Whether self defence or provocation was in issue, the rules defining the defences were meant to ensure conviction for murder when violence was premeditated. A corresponding degree of indulgence was extended to impetuous homicides. In his discussion of the related plea of chance medley, Jeremy Horder suggests that the faults of pride and choleric temper were considered 6 Above n l at 94,98-99.Separation was not a motive in 195 of the sample of 2% intimate homicides. Of these cases: 119 victims were female, 76 male. 7 Id at 97. 8 Ibid: "[Als many as 70 per cent of the husband killings occurred in the context of violence by the husband on the wife". Victorian statistics collected in the Homicide Prosecurions Srudy above n2 Appendix 6: show lower rates of recorded prior violence by male victims of intimate homicide: at 55-58. The Report suggests that the incidence was underestimated. Moreover, the Victorian study, which dealt only with prosecutions for homicide, omitted a substantial number of mur&rs followed by suicide. 9 Polk and Ranson above n2 at 23: "When men kill their women partners, in most cases it can be seen as an act whereby they are exerting their ultimate control over the woman. When women kill, they are most often attempting to protect themselves from the violence that such control involves." 10 Fisher (1837) 8 Car & P 182; [1837] EngR 1115; 173 ER 452: "What a state should we be in if a man ...could be at liberty to take the law into his own hands and inflict vengeance on the offender". 11 Coss, G, "A Brief History of the Doctrine of Provocation in England [1991] SydLawRw 32; (1991) 13 Syd LR 570 follows the divagations of doctrine since its inception. See also Hor&r, J, 'The Duel and the English Law of Homicide" (1992) 12 Oxford J Leg Stud 419. 19931 WOMEN WHO KILL IN SELF DEFENCE 405 too nearly allied to the soldierly virtues for unqualified condemnation.12 In Mawgridge v The Queen,l3 decided at the beginning of the eighteenth century, Holt CJ considered nosepulling and filliping provocation enough to reduce murder to manslaughter. Killings in jealous rage following the discovery of adultery, which Holt CJ called "the highest invasion of propertyW,l4 were also reduced to manslaughter. The mitigation was lost, however, if the facts revealed premeditation or deliberation. There is a long and richly detailed history of the elaborations of doctrine setting the metes and bounds within which jealous rage and the provocations of infidelity can provide a partial excuse for murder. The doctrinal history of self defence as it affected duels and other methods of settling disputes by combat is equally rich in elaboration.15 There are no comparable histories of the development of doctrines allowing exculpatory effect to domestic violence or oppression suffered by women who kill men: the pages are blank.16 We can assume that there is nothing new about women retaliating against domestic violence and the pleas of self defence and provocation were always open to women threatened with serious harm or pushed beyond endurance. The law applied in these cases, however, was the law developed for determining the liability of men accused of homicide. The policies which shaped the modem law of self defence and provocation developed without reference to the very different patterns of homicide when men are killed by women. Mitigation of guilt on the ground that reactions were immediate and impetuous is far less often relevant when women are accused of murder. Even when the act is done in self defence, against threats of immediate violence, the final fatal act is commonly a response to years of violence and abuse. The doctrines of self defence and provocation were geared for discrete episodes sudden quarrels - when hostility erupts into violence. Explicit recognition of the cumulative effects of violence and threats in continuing relationships is rare in the case law on self defence. Provocation was equally restricted by an emphasis on immediate reactions to immediate slights or harms. Only in more recent years have courts begun to recognise, haltingly and uncertainly, the possibility that resentment against continuing cruelty, abuse or exploitation might rank as the equivalent, in exculpatory effect, of jealous rage prompted by real or imagined infidelity.17 12 Horder, ibid. 13 (1707)Kel J 119;84ER 1107at 1114. 14 Id at 1115. 15 Horder, above nl 1 . 16 So far as Australia is concerned, my judgment is impressionistic. In Double Jeopardy: Women Who Kill in VictorianFiction (1990) at 30-39, Virginia Moms suggests that pleas of self defence or provocation were rarely invoked by women in Victorian England. A measure of justice was achieved without recourse to legal doctrine, by sympathetic acquittals and lenient sentencing. Compare the radicaIly different image presented of women's homicides in nineteenth century France, where recognition of the exculpatory effect of jealousy or resentment for infidelity extended to both men and women: see Harris, R, Murders & Madness (1989) chs 6,8. 17 The restrictive approach of English courts on the point is discussed by Wasik, M, "Cumulative Provocation and Domestic Killing" [I9821Crim LR 29 at 30-34. 406 SYDNEY LAW REVIEW [VOL 15: 403 North American comment suggests that doctrinal restrictions on self defence and provocation often compelled women charged with murder to rely on defences of insanity, diminished responsibility or other defences based on mental illness or incapacity.18 Recent Australian homicide surveys are inconclusive on the extent to which the mental incapacity defences have been used for want of fair and adequate criteria governing the pleas of self defence and provocation.19 There is no doubt that women are sometimes persuaded, against their wishes and better judgment, to cast their defences in this form.20 The American tendency to overcome deficiencies in the law of self defence and provocation by invoking mental illness or incapacity is evident in the general acceptance of the theory that women who remain in violent relationships are psychologically disabled. The "battered woman syndrome" is an unhappy hybrid, an evidentiary supplement to existing defences which sometimes takes the guise of a defence in its own right. The theory holds that victims of domestic violence are commonly afflicted by pathological helplessness. Helplessness is said to be learned, as a response to habitual violence, and the state of learned helplessness is said to be linked with other psychological incapacities acquired as a consequence of victimisation. Invented for forensic purposes in North America, the syndrome has begun to make sporadic appearances in jury directions and appellate judgments in Australia and England. This essay commences with a critical evaluation of the battered woman syndrome. It continues with an illustrative biographical fragment, before discussing the polemics of excuse and justification in American criminal law. The concluding section provides an account of the law of self defence and provocation in their application to defensive homicides against attack by an intimate aggressor. The argument of the essay is that Australian common law is capable of discriminating and compassionate justice, in cases of self defence against intimate aggressors and provocation by intimate aggressors. Recourse to the dubious theory that the victims of domestic violence are characterised by helplessness and incapacity is unnecessary. The invention of the battered woman syndrome was, to a very considerable extent, a response to American restrictions on the law of self defence and provocation which have no counterparts in Australian common law.21 18 The assertion is commonly made in the US literature. See, eg, Schneider, E, Jordan, S and Arguedas, C, "Representation of Women who Defend Themselves in Response to Physical or Sexual Assault" in Bochnak, E (ed), Women's SelfDefence Cases: Theory and Practice (1981) at 28-31. Evidence for its truth is rarely provided. The assertion is also made in Schneider, E, "Equal Rights to Trial for Women: Sex Bias in the Law of Self Defence" (1980) 15 Haw CR-CL LR 623 at 638, where it is supported by a reference to a study conducted in New South Wales: Bacon, W and Lansdowne, R, "Women Who Kill Husbands: The Battered Woman on Trial" in O'Donnell, C and Craney, J (eds), Family Violence in Australia (1982). Ruth Hanis explores the nineteenth century origins of the view that women's criminality is an expression of an "underlying pathology of female nature" in above n16 at 302. 19 But see Bacon and Lansdowne, ibid. The authors suggest that defence counsel tended to rely on mental incapacity defences. 20 See Jeffrey [I9671 VR 467, where the accused refused all inducements to rely on the defence of insanity at her trial. The case deserves to be better known. 21 Though the theory is beyond redemption, it has been suggested that judicial recognition of the battered woman syndrome offers a potential stalking horse for the reception of more soundly 19931 WOMEN WHO KILL IN SELF DEFENCE 407 1. The Battered Woman Syndrome Expert evidence of the putative effects of domestic violence was admitted in recent murder trials in New South Wales and in South Australia.22 In both cases women were acquitted of murder on the ground that they were acting in self defence against violence by an intimate aggressor. In both, evidence that they were affected by the battered woman syndrome was offered to support the plea of self defence. Recognition that the effects of domestic violence are an appropriate subject for expert testimony has been sudden and long overdue. North American courts have accepted expert evidence of the characteristic features and effects of domestic violence for a decade or more. Acceptance of the hypothesis that women victimised by domestic violence characteristically suffer from a disorder known as the battered woman syndrome is less welcome. The course of litigation in the United States suggests that recognition of the battered woman syndrome owes far more to forensic necessity than it does to the requirements of therapy for a psychological disorder. The implicit suggestion that women in violent relationships are psychologically disordered in some way23 was used to provide a familiar ground for the admission of evidence by psychologists or psychiatrists.24 That may be a peculiarly American phenomenon. In Australia, psychological or psychiatric evidence is not restricted to diagnoses of the disorders or abnormalities of defendants. Expert evidence is admissible whenever scientific study provides an acceptable basis for testimony which goes beyond common experience and common conceptions or misconceptions.25 There are good reasons for scepticism concerning based expert evidence on the nature and effects of domestic violence. The argument is presented by Sheehy, E, Stubbs, J and Tolmie, J, "Defending Battered Women on Trial: The Battered Woman Syndrome and its Limitations" (1992) 16 Crim W 369 at 387-394. 22 Hickey v The Queen, unreported decision of the Supreme Court of New South Wales, 14 April 1992. Noted, Yeo, S, "Case and Comment" (1992) 16 Crim W 271; R v Kontinnen, unreported decision of the Supreme Court of South Australia, 27 March 1992. Noted, Greene, J, "Case and Comment" (1992) 16 Crim W 360. R v Ahluwalia [I9921 4 All ER 901 presages reception of the theory in England. 23 Walker, L, Terrifying Love (1989) at 48, states that the battered woman syndrome has been "officially accepted ... by experts in the field" as a subcategory of Post Traumatic Stress Disorder (PTSD) which is a diagnostic category within the Diagnostic and Statistical Manual of Mental Disorders (DSM 111 R), the official publication of the American Psychiatric Association. The question whether the battered woman syndrome is a form of mental disorder is discussed in Kinports, K, "Defending Battered Women's Self-Defense Claims" (1988) 67 Oregon LR 393 at 416-417. A related issue is discussed, albeit inconclusively, in Thomson, K, "Post Traumatic Stress Disorder and Criminal Defences" (1991) 21 WALR 279. For a trenchant and critical account of therapeutic perspectives on battered women and their oppressors, see Dobash, R E and Dobash, R P, Women, Violence and Social Change (1992) ch 7. 24 Recognition of the syndrome derives from psychological research, rather than psychiatry. Presumably it is a "psychological" rather than "medical" disorder. Courts which are generally disinclined to accept diagnoses by psychologists of individual disorders (see, eg, R v Peisley (1990)54 A Crim R 42 at 52) can hardly deny their competence in diagnosis of battered woman syndrome. From the defence point of view there may be some advantage in reliance on expert psychological evidence, rather than psychiatric evidence. A more definite boundary can be established behveen the defendant's chosen defence and an unwanted possibility of acquittal on the ground of insanity. For discussion of a comparable problem, see Tarrant, S, "A New Defence in Spouse Murder?" (1992) 17 Alternative W 67. 25 Murphy v The Queen [1989] HCA 28; (1989) 167 CLR 94 at 112, 125-126; discussed Leader-Elliott, I and 408 SYDNEY LAW REVIEW [VOL 15: 403 the validity of the theory behind the battered woman syndrome. It should not be necessary to dress in this dubious disguise evidence which will allow an informed, rational and compassionate determination of cases of self defence against domestic violence.26 In South Australia evidence of the syndrome was admitted as a consequence of the ruling of the Court of Criminal Appeal in R v Runjanjic.27 The defendants in that case, Olga Runjanjic and Eriica Kontinnen, were charged with false imprisonment and various offences of violence against another woman. They pleaded duress and testified that they had been compelled to participate in the offences by a man called Jan Hill, with whom they lived. Both women were victims of appalling violence inflicted over a period of years by Hill. They were convicted and appealed successfully to the Court of Criminal Appeal on the ground that expert evidence of the battered woman syndrome should have been admitted at their trial. The ruling was not restricted in its application to duress. Before her trial for false imprisonment, Eriica Kontinnen turned against Hill and killed him. As a consequence, she was to face a further charge of murder. The Court made its ruling in the knowledge that expert evidence of the battered woman syndrome would be offered in support of a plea of self defence in her pending trial for the murder of Hill. The characteristic features of the syndrome were summarised by King CJ: [Sltudies by trained psychologists of situations of domestic violence have revealed typical patterns of behaviour on the part of the male batterer and the female victim, and typical responses on the part of the female victim ... Repeated acts of violence, alternating very often with phases of kindness and loving behaviour, commonly leave the battered woman in a psychological condition described as ''learned helplessness". She cannot predict or control the occurrence of acute outbreaks of violence and often clings to the hope that the kind and loving phases will become the norm. This is often reinforced by financial dependence, children and feelings of guilt. The battered woman rarely seeks outside help because of fear of further violence. It is not uncommon for such women to experience feelings for their mate which they describe as love. There is often an all pervasive feeling that it is impossible to escape the dominance and violence of the mate. There is a sense of constant fear with a perceived inability to escape the situation.28 At her trial for murder, Kontinnen supported her plea of self defence with expert evidence that she was affected by the battered woman syndrome. She was acquitted. Various grounds of justification have been offered for the reception of such evidence. Expert opinion that the accused displays characteristics common among battered women is meant to provide a context of information and un- Goode, M, "Criminal Law", Annual Survey o Australian Law (1991) 263 at 278-282. f 26 The issues are debated in Stubbs, J, "Battered Woman Syndrome: An Advance for Women or Further Evidence of the Legal System's Inability to Comprehend Women's Experience?' (1991) 3 CICJ 267; Easteal, P, "Battered Woman Syndrome: Misunderstood?" (1992) 3 CICJ 356; Stubbs, J , "The (Un)reasonable Battered Woman? A Response to Easteal" (1992) 3 CICJ 359. See also above n21. 27 [1991] SASC 2951; (1991) 53 A Crim R 362. 28 Id at 366. 19931 WOMEN WHO KILL IN SELF DEFENCE 409 derstanding for defences of provocation, self defence or duress. The most important point of application, in North American case law, is to be found in the use of expert opinion to modify the application of the various standards of normality, reasonableness and proportionality which govern the defences of provocation, self defence, duress and necessity. The standards have all been explained, at one time or another, in anthropomorphic formulae referring to the putative reactions of reasonable or ordinary people in extremis. Australian examples, past and present, include the following cast of representative figures: provocation - could an ordinary person have been driven to the same method and degree of violence as the accused? selfdefence - Earlier formulations of the test supplemented the requirement of necessity with an anthropomorphic standid of proportionality: "the requirement of proportion means that a reasonable man [in the position of the accused] would not have considered that there was disproportion."29 duress - would a "person of ordinary firmness of mind and will, and of the same sex and maturity as the accused"30 have been likely to yield to the threat in the way the accused did?31 The application of standards of reasonableness, ordinariness or capacity for self control does not require reference to these notional actors. One can ask whether the conduct of a particular individual was reasonable without creating a doppleganger for the purposes of comparison. The primary purpose of the anthropomorphic reference is to allow the construction of an imaginary person, "lacking one or more of the defendant's characteristics", who will serve as a standard for comparison. When these standards are applied in an inflexible and biassed fashion, limited by assumptions about the behaviour of reasonable men, the case for admitting expert evidence which will license exceptions for battered women is strengthened. Reliance on a diagnosis of psychological disability can modify the application of unreasonable or unrealistic standards. It is far preferable, of course, to avoid unreasonable, unrealistic or unrepresentative standards in the first place.32 There are other applications for expert opinion on the effects of domestic violence which do not depend on the battered woman syndrome. Women in violent relationships develop an acute awareness of danger signals. Preemptive responses, of one kind or another, may be taken in circumstances where a person less sophisticated in the preliminary nuances of domestic violence would perceive no threat.33 Expert evidence in support of a plea of self de29 R v McKay [I9571 VR 560 at 573; R v Ruiney [I9701 VR 650 at 651. Anthropomorphic references to reasonable men, or substitute figures of variable gender, do not figure in more recent formulations of the defence. On the fluctuating fortunes of the proportionality requirement, see Yeo, S, Compulsion in the Criminal Lclw (1990) at 106-1 12. 30 R v Abusafiah (1991) 24 NSWLR 531 at 545. 31 R v Hurley 119671 VR 526 at 543. 32 For a critical account of attempts to construct specialised versions of reasonable behaviour for battered women, see Schneider, E, "Particularity and Generality: Challenges to Feminist Theory and Practice in Work on Woman-Abuse"(1992) 67 NYLR 520 at 559ff. 33 Breyer, H, 'The Battered Woman Syndrome and the Admissibility of Expert Testimony" 119921 Crim L Bull 99, at 103, 107; Grant, I, "Note, The Battered Woman: When a Woman's 'Place' is in the Courts" (1988) 10 Crirn Just J 273 at 286 ("It is crucial that the 410 SYDNEY LAW REVIEW [VOL 15:403 fence can reinforce the validity of her perceptions of danger. When provocation is in issue, evidence of the characteristic reactions of the victims of domestic violence can reinforce the conclusion that fear or anger led to loss of self control.34 It will serve a similar purpose when the emotional effects of duress are in issue. There is no compelling need for expert evidence to establish the point that prolonged violence breeds fear and resentment of course.35 The more important point of application is to shift common errors concerning the nature of loss of self control. English case law on provocation tends to suggest that the partial defence is available only where there is an immediate and explosive response to provocation.36 But this is only one pattern of loss of self control. Australian courts have displayed a more humane and compassionate understanding of the growth of rage, fear and resentment and the final triggering effects of minor incidents in a protracted history of abuse.37 Though most North American courts insist that the battered woman syndrome is not a defence to crime it may, in time, come to resemble a defence, with its own specialised rules governing the reception of evidence and its applications.38 Law shapes the stories told by defendants and encourages the discovery and elaboration of exculpatory syndromes.39 Psychiatric or psychological theories occasionally amalgamate with legal doctrine to form new variants on old defences.40 There are dangers in these amalgamations. Comjudge and jury understands that a woman may perceive imminent and mortal danger in circumstances where a man would not ... ); Crocker, P, 'The Meaning of Equality for Battered Women Who Kill Men in Self Defence" (1985) 8 Ham Women's LJ 121. 34 Fear no less than anger provides an evidential ground for a defence of provocation: Van Den Hoek v The Queen [1986] HCA 76; (1986) 161 CLR 158. 35 Opinion evidence is inadmissible if it merely relates to matters within the common range of human experience: R v Turner [I9751 QB 834; [I9751 1 All ER 70; R v Runjanjic [1991] SASC 2951; (1991) 53 A Crim R 362 at 368. 36 R v D u f i [I9491 1 All ER 932; R v Thomton [I9921 Crim LR 54; [1992] 1 All ER 306. 37 Parker v The Queen [1963] HCA 14; (1963) 111 CLR 610 (HC); (1964) 11 1 CLR 665 (PC); R (1981) 28 SASR 321. 38 Note, "Criminal Law - Battered Women and Self-Defence - Pennsylvania Allows Expert Evidence on Battered Woman Syndrome as a Basis for Proving Justification in the Use of Deadly Force when Evidence Indicates Defendant is Victim of Abuse: Commonwealth v Stonehouse" (1990) 94 Dick LR 554. Above n21 at 386-387 provide references for the drift towards the status of an independent defence in some United States jurisdictions. Similar pressures could operate in Australia. Compare the denials, commonly made by courts, that automatism, or similar states of brain or mind, are "defences": see eg, R v Hall (1988) 36 A Crim R 368 at 371, per Roden J. The rules governing admission of expert evidence of automatism and the consequential power of the trial judge to withhold consideration of the issue from the jury have developed, nevertheless, formidable legal complexities characteristic of specialised defences: O'Connor, D and Fairall, P, Criminal Defences (2nd edn, 1988). ch 14. 39 See Diamond, S, "Criminal Law: The Justification of Self Defence" (1987) Annual Survey rf Criminal h w 673, at 690-693 on "syndrome proliferation and expert testimony". The battered woman syndrome does not stand alone. It is supplemented by the theory that the victims of domestic violence also suffer from the "Stockholm syndrome", a set of behavioural symptoms said to be characteristic of individuals held as hostages by terrorists: see Graham, D, Rawlings, G and Rimini, N, ''Survivors of Terror: Battered Women, Hostages and the Stockholm Syndrome" in Yllo, K and Bograd, M (eds), Feminist Perspectives on Wife Abuse (1988) 217. 40 For a remarkable example of the injection of legal content into psychiatric diagnosis, see Fairall, P, "Post Stress Syndrome, Automatism and Vietnam Veterans: A Plea for Understanding" [I9871 Australian Current Law Article No 20. 19931 WOMEN WHO KILL IN SELF DEFENCE 41 1 ment on American case law suggests that there is a tendency to make the availability of the plea of self defence depend on a diagnosis of battered wife syndrome.41 That tendency is exacerbated by the comparative rigidity of the law on self defence in most United States jurisdictions.42 Rigidity begets the need for exceptions which will allow an excuse for defensive homicides which do not fit the rules, on the ground that the defendant suffered from some forgivable incapacity to comply with them. Exceptions harden into doctrine and reliance on expert testimony takes the place of reforms which would ensure that standards of expected behaviour are fair and realistic. North American legal theory tends towards dogmatic statements of the view that human life has ultimate value.43 Australian law is less doctrinaire on the point and more accommodating to human variations in response to violent threats. Cathryn Rosen summarises the constraints imposed by theory in her exhaustive essay on the American jurisprudence of defensive homicide: To harmonise the principle that killings in self defence are justified with the principle that human life is the highest value protected by the law, the range of defensive conduct that will be justified must be narrowly circumscribed. The result is a legal environment that is inhospitable to the battered woman's defence.44 In such an environment, expert evidence of the battered woman syndrome tends to couple a plea for compassion with a concession that her use of violence was indefensible in the circumstances. Compassion, in these circumstances, is most readily evoked by evidence which matches, in form at least, a diagnosis of psychological disorder or incapacity. Before we travel too far down this path, in emulation of American models, we should ask whether it is necessary to do so. There are many reasons for scepticism. The concept of "learned helplessness", which is central to the definition of the battered woman syndrome, has the paradoxical effect of transforming an assertive act of self defence into a manifestation of weakness and incapacity.45 The theory looks unconvincing from the outset. Closer analysis 41 Crocker, above n33 at 137. 42 Model Penal Code: Proposed Oflicial Druji (1962) s3.04-3.06, provide a representative set of rules which have been adopted, with greater or lesser degrees of modification, in US jurisdictions. They far exceed in complexity and rigidity anything to be found in the Australian common law, Criminal Codes, legislative restatements such as the Criminal Law Consolidation (Self Defence) Amendment Act 1991 (SA) or in Australian proposals for reform, such as s313 of the Model Criminal Code: General Principles of Criminal Responsibility (Final Report, December 1992) Criminal Law Officers Committee. But see Maguigan, H, "Battered Women and Self Defence: Myths and Misconceptions in Current Reform Proposals" (1991) Pennsylvania LR 379. Maguigan argues that most participants in the debate over defensive homicide by battered women emphasise a b e m t rules and fail to realise or exploit the relative flexibility of the law of self defence in the majority of United Statesjurisdictions. See also Schneider, above n32 at 561ff. 43 The provisions of the Model Penal Code (Proposed Official Draft 1962) ss3.04(2)(b), 3.06(3)(d), 3.07(2)(b) represent the view, conventional in the United States, that the right to use deadly force in self defence is limited to cases where life is threatened. 44 Rosen, C, "The Excuse of Self Defence: Correcting a Historical Accident on Behalf of Battered Women Who Kill" (1986) 36 Amer LR 11 at 27. 45 The point has been made by a number of North American critics. See, eg, Grant, I, "The 'Syndromisation' of Women's Experience" [I9911 BCLR 5152-53; Faigman, D, 'The Battered Woman Syndrome and Self Defence: A Legal and Empirical Dissent" (1986) 72 Vu 412 SYDNEY LAW REVIEW [VOL 15: 403 does not remove the appearance of paradox. The evidence offered in support of the theory weakens rather than supports its veracity. Of course there are cases which do match the contours of the battered woman syndrome. That does not make the condition typical or even common. And even if the stories of women who kill in self defence could be told in this way, we should ask whether this is the way they must be told. Many of the constraints which shaped American case law are absent here. The law of self defence in the common law states of Australia is highly malleable.46 In her critical note on the Australian cases which have accepted opinion evidence on the battered woman syndrome (BWS), Julie Stubbs asks why we should not simply "accept the BWS as a strategy to assist individual women, even if [we are] not convinced by the psychological literature concerning the syndrome?"7 One answer is that it is frequently unnecessary to do so. Stubbs suggests a more fundamental answer as well. Reliance on fictions, however convenient, makes for bad law and imperils women who cannot present a convincing appearance of weakness or incapacity.48 2. Helpless or Cornered? There are just two defining elements of the battered woman syndrome. The first is the subjection of the victim to violence in a repeating three phase pattern: [I] a period of mounting tension, [11] an "acute battering incident" followed by [111] a period of loving contrition.49 Lenore Walker, the clinical psychologist whose work dominates the field, refers to this pattern, which is said to be characteristic of violent domestic relationships, as the Walker Cycle Theory of Violence.50 Violence in the family is often repetitious, patterned and ritualised. It is likely, however, that the patterns display far more variety than Walker allows. The suggestion that the battering phase is usually followed by a phase of loving contrition, may owe more to the expectations generated by theory than empirical study of the evidence. The period of loving contrition is supposed to provide the reward which binds the victim to a continuation of the relationship. No doubt this can happen, particularly in the LR 619 at 641. See also the references in Kinports, K, "Defending Women's Self Defence Claims" (1988) 67 Oregon LR 393 at 453-454, where an attempt is made to answer the point. 46 It is likely that the common law will provide the model which will be followed in the Code States. Current proposals for criminal law reform include adoption of uniform provisions in all jurisdictions codifying the basic elements of criminal liability: see Goode, M, "Codification of the Austnlian Criminal Law" (1992) 16 Cnm LJ 5. Dnft proposals produced by the Review of Commonwealth Law (1990) and subsequent development of those proposals in the Model Criminal Code: General Principles rf Criminal Responsibility (1993) preserve a generalised rather than pmticularised statement of self defence. 47 Stubbs, "Battered Women Syndrome: An Advance for Women or Further Evidence of the Legal System's Inability to Comprehend Women's Experience?'above n26 at 269. 48 Maguigan, above n42 at 443-458, instancing examples of reform which have the pandoxical effect of restricting the scope of self defence for battered women. See too, above n21. 49 Walker, L, Thyfault, R and Browne, A, "Beyond the Juror's Ken" (1982) 7 Vermont LR 1 at 9; Walker, L,The Buttered Woman Syndrome (1984) ch 10. 50 Above n23 at 42 asserts that the Walker Cycle was present in about two thirds of "the sixteen hundred incidents in our studies on battered women". The source is not given. Presumably the reference is to her earlier work, The Battered Womcm Syndrome (1984) which was based on interviews with 400 battered women. 19931 WOMEN WHO KILL IN SELF DEFENCE 413 early stages of violent relationships.51 But the theory suffers from the defect of supposing violent men and battered women to be more stupid and their relationships less complex than is generally the case. Contrition requires an acknowledgement of responsibility, and promises of reform require a show of sincerity. Loving contrition is not a performance which can be repeated or accepted indefinitely. In relationships where violence has become established it is probable that the more usual pattern is refusal, by the aggressor, to acknowledge what has happened.52 When masculine violence is associated with alcohol the refusal to acknowledge responsibility is often coupled with a more or less explicit dissociation from the drunken self who did the harm. The second element in the battered woman syndrome is learned helplessness, a concept developed by extrapolation from "a series of carefully controlled laboratory experiments originally conducted on dogs and later extended to include rats, cats, fish, and humans".53 The best known of these experiments involved caged dogs which were subjected to randomly inflicted electrical shocks. The researchers concluded that subjection to this treatment extinguished the impulse to escape.54 They suggested that there were parallels between the final apathetic state of the dogs and the condition of learned helplessness in human subjects. There is an extensive psychological literature on the condition, exploring the relationship of learned helplessness to a range of human characteristics, behaviour and pathology.55 Research suggests that learned helplessness and clinical depression are closely allied: associated with both are "deficits" of motivation, cognition, affect and self esteem.56 Lenore Walker invokes learned helplessness to explain why the victim of domestic violence does not flee the house, seek help or end the relationship: she does not do these things, or others which uninstructed common sense might suggest, because she cannot. She has been reduced to a condition where action is impossible.57 Expert evidence that the accused was a battered 51 Walker, above n49 at 98-99. 52 Compare Dobash, R E and Dobash, R P, Violence Against Wives (1979) at 116-118 and a paper by the same authors on the results of careful survey of domestic violence in Edinburgh and Glasgow, "The Nature and Antecedents of Violent Events" (1984) 24 Brit J Crim 269 at 280: "after a typical assault the men usually acted as if nothing had happened". By 1984, Walker, above n49 at 96, significantly qualified the assertions made in earlier publications: "phase three could also be characterised by an absence of tension or violence, and no observable loving contrition behaviour, and still be reinforcing for the woman". In Terrifying Love (1989) at 44-45 loving contrition following violence is once more presented as the norm. 53 Garber, J and Seligman, M (eds), H u m Helplessness: Theory and Applications (1980) at xv. 54 Seligman, M and Maier, S, "Failure to Escape Traumatic Shock" (1967) J Experimental Psych 1, reporting the results of experiments conducted by Overmier and Seligman, "Effects of Inescapable Shock on Subsequent Escape and Avoidance Learning" (1967) 63 J Comp Physiology: Psychology at 28: "exposure of dogs to inescapable shock in a Pavlovian harness reliably resulted in interference with subsequent escapelavoidance learning in a shuttle box". 55 Abramson, L,"Leamed Helplessness in Humans: An Attributional Analysis", listing "child development, stomach ulcers, depression, ... death ... intellectual achievement ... crowding ... victimisation ... the coronary prone personality ... and aging", in above n53 at 4. 56 Id at 26-31. Walker, above n49 at 82. 57 The crudity of some subsequent accounts is remarkable. After describing laboratory experi- 414 SYDNEY LAW REVIEW [VOL 15: 403 woman, whose reactions were consistent with the condition of learned helplessness, is meant to provide a counter to the natural tendency to insist that reasonable or ordinary individuals exhaust alternative courses of action, before resorting to the use of deadly force. Learned helplessness is central to the definition of the battered woman syndrome. If the syndrome is taken to be a form of mental disorder, learned helplessness is its only constitutive element. It is the concept of learned helplessness which makes the condition her syndrome rather than his. The other element of the syndrome - subjection to repeated domestic violence -is neither a psychological disorder nor a symptom of psychological disorder.58 The cycle of violence, its rituals and manipulations, is not a manifestation of her abnormality. It is merely the pattern of brutality to which her partner has grown habituated. In view of the crucial role which the concept of learned helplessness has played in discussion and litigation in the United States, it is surprising to find how meagrely it is supported by the evidence in Walker's study of battered women. She tested her theory in a study based on interviews with 400 battered women. The survey made few pretences to scientific rigour.59 Its purpose was to "learn more about intrafamily violence from battered women themselves".~The survey results seem to have surprised the researchers at every turn. They do not support the theory. It was surmised that battered women who stayed with their violent partners would display a constellation of conservative attitudes about women's roles and the values of home and hearth. The women were not conservative: they proved to be more liberal in their attitudes than Walker's control group.61 Equally surprising to the researchers was the finding that the battered women were not lacking in self esteem, one of the four elements supposed to characterise the condition of learned helplessness.62 The theory fared no better on the other indicators of learned helplessness. The women did not perceive themselves to lack control over their lives. On the contrary, they "saw themselves as having a great deal of control over what happens to themW.63 Though battered women were exments on dogs who were too temfied to leave their cages after randomly inflicted shock treatment, Hugh Breyer remarks: "Battered women may suffer from the same psychological condition that these dogs did, learned helplessness". See Breyer, above n33 at 101. 58 The evidence suggests that battered women share no identifying characteristic other than the misfortune of being involved in violent relationships. Walker's survey of battered women led her to conclude "that there are no specific personality traits which would suggest a victim-prone personality for the women, although there may be an identifiable violence-prone personality for the [violent] men", Walker, above n49 at 7. 59 Considered as an example of psychological research, the survey is open to severe criticism: Faigman, above n45 at 636-643. 60 Walker, above n49 at 203. 61 Id at 77-78, "a surprising finding". Compare Russell, D, Rape in Marriage (revised edn, 1990)at 170-171,whose survey results suggest that it is not traditional attitudes which explain failure to escape from abusive relationships but absence of financial and other resources: "the more traditional women (more traditional by virtue of having fewer resources) are more likely to stay". Cf, Wallace, above nl at 88-90.The New South Wales survey results indicate that domestic homicides occur amongst those who have the least material resources for escape. 62 Id at 80: ''It was predicted that battered women's self esteem would be quite low and our results, surprisingly, show the opposite". 63 Id at 78-79. 19931 WOMEN WHO KILL IN SELF DEFENCE 415 pected to score high on indices of depression, they "were not consistent in demonstrating the negative cognitions and moods we would have expected."64 The survey responses, in short, provide no support for Walker's hypothesis that learned helplessness is characteristic of battered women in general.65 Insistence on the hypothesis led Walker to suggest that the women in the survey were deluding themselves. Perhaps they "distort[ed] their feelings about themselves in a positive way as a coping style to continue to function despite chronic depression".66 Their self esteem may have been equally delusive: "The desire to please the interviewer, like the need to please their batterers, may override their ability to accurately know and label their feelingsW.67 Perhaps they suffered from a delusive failure to realise that they were not in control of their lives.68 Perhaps the women who participated in the survey were not typical of battered women generally. Women who were prepared to participate may have been less likely to show signs of helplessness than those who made no effort to do so.@ Perhaps the tests were at fault. Walker suggested that more sophisticated tests might yield a different result.70 There are no published reports, however, of further research conducted by Walker or her associates. When Walker refers to the survey in her more recent study of the battered woman syndrome, the references are rarely specific and the discussion of the results is muffled or anecdotal.71 Though the battered woman syndrome continues to figure in her books, the central hypothesis of learned helplessness can be supported only if the responses of her interview subjects are discounted. Explanations of the battered woman syndrome have shown some development over time. The hypothesis was first elaborated in the late seventies, before completion of Walker's survey of 400 victims of domestic violence. Those earlier accounts suggested a degree of complicity between victim and oppressor. Seduced by the episodes of loving behaviour which alternate with violence, the battered woman "becomes an accomplice to her own battering9'.72In the first of her books on the subject, Walker emphasised the passivity, hopelessness and loss of any sense of self worth supposed to be characteristic of the learned helplessness of battered women.73 This earlier work was criticised on the ground that emphasis on individual pathology merely reinforces existing prejudices and encourages the common tendency to 64 Id at 83. 65 In addition to the material cited, see id, ch 9, "Learned Helplessness and Battered Women". Once again the critical measures of learned helplessness are either ambivalent or inconclusive. 66 Id at 83. "[Tlhere may be a cognitive component to their depression which isn't measured very well by affective means", id at 85. 67 Id at 81. See also at 79, 100, suggesting that the vain hope that the relationship would c h g e for the better led the subjects to a delusive estimate of their capacity to control their lives: "While in a violent relationship, the battered woman is so involved in doing whatever it takes to keep her batterer happy that she perceives this as being in control". 68 Id at 79-80. 69 Id at 87. 70 Id at 114-115. 7 1 Above n23 at 49-53. 72 Walker, L, "Treatment Alternatives for Battered Women" in Chapman, J R and Gates, M (eds), The Victimisation of W a n (1978) at 143, 153 73 Above n49 at 49-50. 416 SYDNEY LAW REVIEW [VOL 15: 403 blame the victims of domestic violence for masochism.74 In the most recent of her books, Walker attempts to present learned helplessness in a more positive light. It no longer seems to be a state of helpless passivity. Learned helplessness, in this later transformation, is a means of coping or adapting to harsh realities. The battered woman chooses those "behavioural responses that will have the highest predictability of an effect within the known, or familiar situation ... she believes the demons she knows well are probably preferable to the demons she does not know at all". The laboratory dogs are invoked once more, to provide a grotesque metaphor. They were not, it now appears, quite so helpless as they seemed: [The dogs] appeared to ... remain entirely passive, sometimes lying in their own excrement, refusing both to leave and to try to avoid the administered electric shocks. However, a closer look revealed that these dogs were not really passive. They had developed coping skills that minimised the pain, lying in their own fecal matter (a good insulator from the electrical impulses) in a part of the electrical grid that received the least amount of electrical stimulation. Seligman found that eventually the dogs would learn to escape after being repeatedly dragged to the cage exit. Once they had learned to escape in this matter, their "learned helplessness response", which was to trade the unpredictability of escape for the more predictable coping strategies,disappeared.75 There is a certain bitter irony in the use of these repellent experiments76 as metaphors to illustrate the effects of oppressive cruelty in the home. Learned helplessness is supposed "to explain why women find it difficult to escape a battering relationship".77 Walker's preoccupation with the theory of learned helplessness obscures the real reasons why they remain. There is no syndrome: the women in Walker's survey did not suffer from a psychological disorder. The experiments with dogs are irrelevant - an inappropriate metaphor. The experiments showed that dogs reduced to a state of learned helplessness made no attempt to escape from pain when escape was possible. The essential point of the experiment was to subject the dogs to maltreatment and present them with a simple and obvious means of escape. Once conditioned by the experience of random and utterly uncontrollable punishment the dogs lost their capacity to respond to further punishment with normal avoidance behaviour. There is no analogy, even on the level of suggestive metaphor, with the dilemmas faced by most women in abusive relationships. Walker concedes that many battered women do escape from their abusive partners.78Dobash and Dobash document the active role characteristically played by women in seeking assistance to end violent relationships.79 In Diana Russell's study of abusive marital relationships, victims of rape or battering were far more 74 Dobash, R E and Dobash, R P,"Research as Social Action: The Struggle for Battered Women" in Yllo, K and Bograd, M (eds), above n39 at 5 1.64. 75 Above n23 at 50. See also Anderson, E and Anderson, A, "Constitutional Dimensions of the Battered Woman Syndrome" (1992) 53 Ohio State W 363 at 371-373. No references are given to original sources dealing with this variation of the dog experiments. 76 Seligman, M, Helplessness: On Depression, Development and Death (1975) at 21-44 provides an accessible account of the animal experiments. 77 Id at 86; above n23 at 47. 78 Walker above n49 at 114: "we found that women do not stay that long in battering as opposed to non battering relationships". 79 Dobash and Dobash, above n23 at 230-235. 19931 WOMEN WHO KILL IN SELF DEFENCE 417 likely to be divorced or separated at the time of the interview than her control group.80 Of course there are women who have not yet reached the point of terminating the relationship and there are some who will never do so. That does not justify the conclusion that those who do remain are irrational or affected by psychological deficits which make them incapable of escape. Walker concedes that there were good and sufficient reasons why many of her battered women remained in violent relationships. Some had been pursued, beaten and recaptured.81 The risk of death or injury when a woman leaves a violent man are well documented.82 Others remained in order to protect their children. Many lacked resources or access to help.83 If a woman does leave, economic and family pressures may compel her to return. There is no need to search for indications of psychological disorder for an explanation of the decision to remain.84 They do so because they have good grounds for supposing that the benefits of the relationship, however meagre, outweigh the available alternatives. When they leave, they do so because the costs can be endured no longer. Though Walker refers at various points to the possibility that data drawn from her survey of battered women might be analysed in this way,85 recognising the real impediments to freedom, the suggestion is relegated to the list of interesting topics for future research. The issue is more immediate however. The hypothesis of learned helplessness is essentially dismissive. The testimony of the women who chose to stay in violent relationships deserved a more responsive hearing. There is another dimension to the issue of escape, which goes beyond any conception of battered women as victims, whether of mental disorder or of social forces beyond their control.86 Assumptions that the choice to continue the relationship is pathological, a consequence of mental disorder, mask the mundane reality of domestic violence. Violence takes its place with economic 80 Russell above n61 at 188-190. 81 Walker, Thyfault and Browne, above n49 at 10, 12: "In the homicide cases the women often felt trapped in a deadly situation. Most of the batterers had warned the women that they would never let them leave alive, or that the batterers would find and punish them if they did escape ... One of the most dangerous times for both partners is at the point, or threat, of separation ..." 82 Above nl at 98-99, 105-108.Wallace surveyed known homicides recorded in New South Wales between 1968 and 1981. She found that close to half of the wives killed by their husbands during this period "had either left or [were] in the process of leaving". On the role of "separation assault" as a "cultural concept, to resolve doctrinal problems in law", see Mahoney, M, "Legal Images of Battered Women: Redefining the Issue of Separation" (1991) 90 Michigan LR 1,71ff. 83 Above n49 at 170-171 who employs a "traditionality" scale to rank impediments to escape. Russell found that women who occupied a dependent position in a marital relationship were no more likely to be beaten or raped than those who had an independent livelihood and no children. They were, however, far more likely to remain in abusive relationships. 84 Discussed in Dobash, R E and Dobash, R P, Violence Against Wives: A Case Against Patriarchy (1980) ch 8. "If we define as masochistic the woman who cannot find a job or provide another home for herself and her children or resolve her mixed feelings about remaining married, then once again we make the error of blaming the woman for being beaten": above n49 at 160. 85 Above n49 at 84,103,116. 86 Two remarkable papers by Mahoney, above n82 and "Exit: Power and the Idea of Leaving in Love, Work, and the Confirmation Hearings" (1992) 65 Sth CalgLR 1283, discuss the issue of escape from abusive relationships. 418 SYDNEY LAW REVIEW [VOL 15: 403 hardship, infidelity, alcoholism and excessive drug use, the competing demands of home and work and the other destructive forces which constitute so large a part of intimate relationships. Many women contend with violence and abuse in order to maintain a home, a marital identity, a commitment, a relationship or a family life for their children. The simplest and most nearly accurate explanation of a woman's choice to remain may be found in her belief that her life in that relationship, and the relationship itself, was still worth defending. There is no reason to assume, even in the rare cases where the violence and abuse culminate in the death of one or other partner, that she stayed because she was psychologically incapable of leaving or that her choice to remain was irrational or pathological. More credit and more humility is due to the courage in adversity, to the ingenuity and to the not infrequent humour which sustains these grossly imperfect, though far from uncommon, relationships. When domestic violence is in issue it is the spectators, rather than the victims, who are likely to engage in unreal speculations and flights from reality.87 3. Utopia Road In Anna Karenina Tolstoy remarks that "all happy families resemble one other, [and] each unhappy family is unhappy in its own way". Cruelty is more ingenious, more inventive and its effects more various, than love or kindness. There are, nevertheless, common patterns of pathology in unhappy families. I know one such pattern, in intimate detail. For seven years, between the ages of nine and sixteen, with intermittent breaks when I was at boarding school, I lived with my mother and the man who eventually became her third husband. Vera and Bill. He was a violent alcoholic. She was his lover, wife, protector and his victim. I choose their story to tell because it is typical and illuminating and because I know it well. Is it necessary to provide further justification? Stories are necessary if we are to understand the criminal law and mould it to human needs.88 It took a good story, in Parker v The Queen,89 told with judicious passion and embellished with references to Othello, to change the law of provocation in Australia. Sir Owen Dixon's account of Frank Parker, his wife Joan and Dan Kelly, her lover, is probably the most effective use of narrative to reform the law in the Australian reports. It is, of course, a story of overwhelmingjealousy and outraged male possessiveness. There are no comparable womens' stories in the canon. 87 Comforting fictions which obscure the reality of domestic violence abound. 'To accept the masochistic explanation of why a woman does not leave a violence relationship is very comforting. It removes the moral outrage over the wife's victimisation and it means the outsiders can quietly ignore the problem without feeling guilty" Dobash and Dobash, above n84. Mahoney deals more particularly with the evasions fostered by resort to the battered woman syndrome in two essays: above n82 at 10-W, above n86 at 1299-1312. Paraphrase or quotation cannot do justice to the extraordinary power of her account. 88 Delgado, R, "Storytelling for Oppositionists and Others: A Plea for Narrative" (1989) 87 Michigan LR 2411. See also the symposium k g a l Storytelling, ibid, in which Richard Delegado's essay appears. 89 (1963) [1963] HCA 14; 111 CLR 610. 19931 WOMEN WHO KILL IN SELF DEFENCE 419 A. Big Ears is Listening They met sometime in 1948, when Bill Adams was a mining engineer, employed in a get rich quick scheme to mine sapphires in northern New South Wales. He was married but his wife had left him, taking their daughter with her. He came to stay at the long since demolished Tattersalls Hotel, in Glen Innes, where Vera was office manager and receptionist. She left her second husband in Melbourne, without divorcing him, four years before she met Bill. That parting was amicable and quite final. Her first husband, my father, had been killed at El Alamein, in the early years of the war. Once she was single we moved often, travelling around Victoria and New South Wales. Many of these journeys were determined by the availability of jobs which allowed her to look after a small child. Restlessness and adventurism probably accounted for other moves: she was a romantic, tense and impatient for life. We stayed a year in Glen Innes, longer than most places. I never saw him in Glen Innes. I had been sent to a provincial boarding school in Warwick. After six months the sapphire adventure failed and he took a job as chief engineer at the coal mines in Lithgow, in the Blue Mountains. Bill and Vera moved to Hermitage House, opposite the old Hermitage mine. Vera came to Warwick to collect me at the end of the year. I do not remember her speaking of him during the long journey to Lithgow. When we arrived she cooked an evening meal for three. It was a celebratory meal. He came in later. He was drunk and jovial. I was nine years old and intensely suspicious. He was a huge man with a scarlet face. He sat me on his knee and said he was to be my father. What was he to call me? Oscar? George? Archibald? Joe? He tried various names in the days following, before settling on Joe. He called my mother Nick or Nicky. Years later I realised that the name was a punning joke name. She was two years older than him and this was her nickname, short for Old Nick, the devil. We never used our own names at home. Their first year at Hermitage House was their alcoholic honeymoon. Though they were often drunk, the violence and abuse which later dominated their life together were rare. Evenings were loud. He made us laugh helplessly. There were bats in the roof which escaped at nights and flew around the dining room. He stood on the oak table, among the dishes, and swatted them with his daughter's tennis racquet. He sang and told stories. He did imitations, many of them repeated over the years without variation, of people he knew or had once known. He was a clever mimic. No one we knew escaped these reiterated displays of their oddities and absurdities in our secret theatre. We spoke a private language in our house, all of his invention. He colonised his world and dominated it by verbal transformation. Household appliances and motor vehicles had pet names. Dogs and cats were named after his relatives. Neighbours, acquaintances and associates all had secret names. The names of common objects and foodstuffs were all changed. The language provided a barrier against outsiders and the outside world was further distanced by ridicule. His wife divorced him and obtained orders for financial support for herself and their daughter. Bill and Vera planned a new start in New Zealand where he could avoid paying maintenance. We left Australia in 1950 and they bought a farm in Northland, the inverted toe of the North Island. The farm was midway between the villages of Paparoa and Maungaturoto, at the end of 420 SYDNEY LAW REVIEW [VOL 15: 403 three miles of winding dirt road. Utopia Road, it was called. Though Vera contributed most of the deposit, the property was bought in his name. She held a second mortgage over the property and later a third mortgage. He named it Kanumbra, after a whistlestop station in Victoria, near the farm at Middle Creek where Vera spent her childhood. Names. Everything named took on a personality. He built the refrigerator: it had a name. The telephone had a name. He became the voice of these inanimate things: he articulated their needs and, through him, they demanded love, respect and obedience. Kanumbra, the most demanding of these presences, consisted of 300 acres of hilly forest and 200 acres of gentle grazing land and dairy paddocks, divided by a stream and the railway line to Whangarei, which ran parallel to the stream. Bank loans financed the purchase of a small milking herd of 40 Jersey cows and 300 aged and decrepit Romney ewes. We lived in this remote and beautiful place, at the end of Utopia Road, for five years. We were isolated by distance and unfamiliarity from the local community. The nearest neighbours were a mile distant. There was no public transport. Mail and small supplies came on the Paparoa primary school bus, which stopped at the other end of Utopia Road. Vera could not drive. Our only vehicle, a '37 flatbed Chevrolet truck, was an extension of his personality. "Chevie" was patched, repainted, overhauled and cosseted. Vera was dissuaded from any attempt to learn to drive by his possessiveness. I said he was huge. He was just under two metres tall, a long legged and rather shapeless, swaybacked man with round shoulders. He wore a moustache and cultivated a slight resemblance to William Powell, star of the Thin Man films. He weighed sixteen stone. Ungainly on land, he swam and surfed with confident grace and immense power. He was a highly competent mining engineer and a skilled draughtsman, carpenter and metal worker. Years later, Vera told me he seemed shy and overpolite when she met him. The contrast between his evident power and his manner was one of the things that once made him so attractive, she said. As the years passed he grew louder, more dominating and more bombastic in company but he was never at ease and he was always acutely vulnerable to hurt. Vera's mother Ada, a woman of indomitable temper, always distrusted and disliked him and referred to him as "The Great I Am". Over the years we lived at Kanumbra, their lives deteriorated. Vera was no match for him. Nothing counted against his sheer muscular bulk. She was a slender woman, not much more than half his weight. She never dared to challenge his physical domination. Her most effective weapon, a cold and biting wit, was not enough to protect her. If he was stung by the things she said he could hit her. He was acutely sensitive to language. Violence was not common however. Threats and abuse were more usual. She was overwhelmed by the inexhaustible tides of his volubility and took refuge in silence. Violence was always associated with alcohol. No liquor was ever kept in our house. It was bought and consumed in binges. The quantity rarely exceeded the requirements of one drunken night. If he wanted to drink an excuse had to be found for the journey to Paparoa or Maungaturoto. When we first settled on Kanumbra, the cycle from binge to binge was about twenty days. Over five years it reduced to ten days. He imposed restraints on himself. We made weekly trips to town for supplies. He rarely bought liquor on these occasions, even when he was approaching the point of the cycle where he was 19931 WOMEN WHO KILL IN SELF DEFENCE 421 going to get drunk. The trips to buy grog were special trips and special excuses were necessary. He discovered he needed to replace a broken tool, or Chevie needed something done at the garage. Vera had a choice. She could adopt the pretence, or reject it and plead with him not to go. His response was never entirely predictable. Occasionally he returned from these expeditions without liquor. Occasionally he gave in to her request and did not go at all.90 To plead with him was dangerous however. Pleas meant that she did not believe him and pleas deprived him of his covering excuse for the trip to town. If she tried and failed to persuade him, he left with a festering sense of outrage and he was likely to return late, already drunk and in a vicious temper. Each alcoholic beverage had its familiar horrors. Beer was the least dangerous. It was a social drink, not serious. He rarely brought beer home. It was consumed during daylight and particularly associated with the occasions he called "Pahi picnics". After the morning milking, Vera made a picnic lunch. We drove the ten miles to Pahi with the dogs barking and leaping on the tray of the truck. Pahi was three houses, a solitary pub and a jetty on the desolate shoreline of a muddy estuary. He drank in the bar of the Pahi pub and Vera spread a blanket in the shade by the truck and waited and read. At midday, he bought a dozen bottles, gathered a small party of regulars from the hotel and brought them back to his Chevie where Vera had arranged sandwiches and fruit on the tray. She presided over this bizarre gathering with rigid embarrassment while he entertained the men. He had names for each of them: BO Plenty, Snow, Bill the Bootbuster, Bluey and Curly. He burst into fragments of song, never more than a verse or two. He told stories they cannot have understood and he told them jokes with no point. They were vacant, friendly men but easily bored and easily hurt. Most of them drifted back to the bar in spite of his efforts to entertain them. Only the most desperate remained. He pretended that these were her friends and that picnics at Pahi were for her benefit. Beer was relatively benign in its effects. The worst of these occasions was that he might go on to something stronger. If he could be humoured through the afternoon, he might go home empty handed. Then the evening would follow its usual routine. Fires would be lit. The cows would be milked by seven. Tea would be prepared and eaten. The negotiation of a safe path to a peaceful end to the day required infinite skill, but it was possible, if she kept her head. When he drank in the evenings he usually drank Johnny Walker whisky. Whisky was masculine. It was associated above all with his father, the Old Man, a ship's engineer, who had been an intermittent visitor in his own home. My stepfather drank whisky from his father's crystal tumblers until they were all broken. He did impressions of the Old Man and told stories of his exemplary rigidity. When he taught me table manners he showed me the scar beside his ring finger. The Old Man impaled his hand with a fork when he reached for food, instead of asking for it to be passed. Johnny Walker induced raucous good humour, in the early part of the evening. He sang. He was a one 90 The respite was usually short lived - a day, two days. Once when she persuaded him to stay, he came back into the house, sat down and drank a bottle of dry ginger ale quite slowly, from a small whisky glass. It took him an hour or more. By the time he had finished, he had drunk himself into a state of simulated intoxication and quite genuine rage which made it possible for him to change his mind and go to Paparoa. 422 SYDNEY LAW REVIEW [VOL 15: 403 man band. He knew very few songs, but he sang them loudly, stamping his feet, clapping and slapping his thighs. He had a dental plate to replace a single central incisor lost in an accident. He could clack this solitary tooth on its vulcanite plate against his lower teeth in fast syncopated rhythm. He could fart at will, it seemed, in time with his song. After the songs came the stories of his father, of school teachers, of fellow students when he studied engineering, of men in the mines. As the stories continued, tension began to gather. If he could be induced to eat, the course of the evening could be controlled. If he ate, he ceased drinking: the two activities were rigidly separated. On the days when he went to town for liquor, Vera usually took particular care over the evening meal. There was a point at which she would offer to serve him and he would decline. The meal was not sufficiently cooked for him or he wanted first to tell another story, sing another song. Occasionally she induced him to eat and terminated the progress of the evening. More often, there was an appallingly sudden end to his joviality: No-one is interested in poor Daddy's stories. Poor Daddy. She placates him, assures him of her interest and offers dinner again. Nobody's interested in the poor old bugger. Poor boring old Daddy. Poor boring old bugger. She replies that she has heard the stories before, she has heard them all, the same stories, so often before. Her voice is too expressive, too theatrical. Say it. Go on, say it. Go on, say Daddy is a boring old bugger. She refuses and he repeats his demand over and over again. She gives in at last to the repetitive, battering demand and repeats his words. Her reply hangs in the air for an interminable moment before the dam wall bursts and the tirade begins. You bitch, Nicky. You filthy stinking bitch. This is what you wanted, isn't it? You're happy now, aren't you? You're off. You're off and running. Into it Nick. Into it. Boots and all. This is what you wanted. She will stay with him for a time making an attempt to placate him, before retreating to the bedroom. Once she leaves him there is no longer any possibility of controlling the course of the night. The nights when he drank gin - any English brand would do - began with reminiscences of his mother. Gin made him weep. It was associated with solitude and sexual loathing when he called Vera a whore and accused her of turning his home into a brothel. More violent were the nights when he drank Jamaica Red Heart OP Rum, which maddened him, inducing megalomania. On these nights, it was too dangerous to stay in the house with him. In summer we sat under the cypresses in the home paddock. She watched the house, smoking cigarettes to calm herself, until it was safe to return. He stood in the back doorway, framed in candlelight from the kitchen, bellowing into the darkness. He wanted a child. He had lost his daughter. All contact ceased when he left Australia. He blamed Vera for separating him from her. Vera became pregnant in 1953. She put aside her fear in the hope that a child of their own would change our lives. She obtained a divorce from her second husband and they married secretly. From the beginning he called the child Zoe, meaning life. His imagined daughter assumed a role and presence in our house. She 19931 WOMEN WHO KILL IN SELF DEFENCE 423 was a source of joy, a light and airy spirit. He built furniture for her and remade a bedroom for her. He did not stop drinking but his binges diminished in severity and they occurred less frequently. He was sometimes exuberant and more often maudlin during these lapses, but never violent, during the pregnancy. In the eighth month something went inexplicably wrong and Vera was delivered of a still born daughter. A new and more terrifying theme entered our lives. Death displaced the child as an imagined character in his monologues. In the beginning it was hardly perceptible. He was deeply depressed, at first, by the loss of the child and he was solicitous. Then he began to get bored. Vera's grief went on too long. She was ill. He was, I believe, tired of the restraints he had endured. The pregnancy had diverted him. Now he wanted excitement again. The drinking pattern quickened and he began to talk of death. At first he restricted himself to mourning, religiosity and self pity. In the ensuing months new routines were elaborated. God was appropriated and appeared as a grotesque stage property in his monologues. God had taken Zoe. Why? He asked the question over and over. Vera gave up trying to answer him. Out of the silence he began to elaborate an answer. God had taken his child because Vera was filled with hatred. You hate me. Why did you do it Nicky? Nick -Old Nick -you hate me, you even hate God. Even poor Jesus. Hate is your middle name. You are rotten with it, rotten with it, rot@ with hatred. God forgive you. God forgive you. This is what you wanted isn't it? Death. There is no life in this house, only death. You are pewing on death, pumng with it. May you rot in Hell ... She had killed his child. And now she was killing him by silence. He had a name for it, the "death silence". They both smoked heavily. Cigarettes were always called "lungbusters". Now her cigarettes became "death sticks". She was reading a book and smoking, ignoring him. She was smoking him to death, reading him to death. His fear of solitude and silence seems to have been quite genuine. Anger kept the demons at bay. He was dying of her. She was death to him. God would not forgive her. She would rot in Hell for what she had done to him. I do not recall him ever making a direct threat to harm her. The threats during this period were always of supernatural retribution. I believe there was an element of creative play in these rhetorical elaborations. For all the battering repetition of his invective, it was in constant evolution. After weeks of quiescence, punctuated by bearable evenings when he drank Johnny Walker whisky and rehearsed old routines, he would return, already drunk, with a bottle of Tanquerays gin or Red Heart rum - stimulants to new flights of rhetoric. Terror drove her from the house. She waited in the bedroom until she heard his footsteps in the hallway and then escaped through the french windows which led to the garden and the paddocks beyond. Each time he got drunk on gin or rum he ventured a little further, raiding the stores of the unsayable, for the final and unforgivable insult or the final and most damning accusation. His reproaches were ludicrous, the bathos ridiculous. His appropriation of the vocabulary of sin, sacrifice, death and damnation was consciously theatrical. By nature he was an atheist. He had absorbed atheism with his education as an engineer in the twenties. His religiosity could be punctured sometimes by sarcasm or laughter but the relief was temporary and a deflating remark might induce violence. That he was often ludicrous in no way diminished her 424 SYDNEY LAW REVIEW [VOL 15: 403 fear. Terror is enhanced by theatre. As the gap between reality and invective widened, he became more unpredictable. Underneath the apparently uncontrolled paroxysms of rage his imagination was at play. He was listening to himself. His imagination had its own autonomous imperatives. He was creating a part for himself. Looking back, I believe he was conscious of his frequent absurdities and relished the abandonment of himself to absurdity. The consciousness of acting a part freed him from responsibility. At the time, it seemed that he was forever approaching the brink of some final and definitive moment when abuse would progress to catastrophe. After the first explosion of rage we go to our rooms. The faintest sounds were audible in our house. We had to listen. The monologue continues in the kitchen. We listened for the scrape of his chair on the lino and his footsteps in the hallway. My door is the first. Are you listening?Can you hear me Big Ears? He hits the door with his fist but does not open it. Then he leaves my door and stands outside the door of their room. Can you hear me Nicky? Big Ears can hear me. Big Ears is listening. Are you listening Nicky? He hammers on the door, accusing her of destroying his life. Yet he rarely enters until the bottle is finished and the e\$ning is over. He returns to the kitchen, sits down and fills his glass again. The tirades outside her door are repeated six times or more in the course of the evening. She was required to make calculations of exquisite refinement. If he reached a pitch of anger where he might enter the room and hit her, she could escape by the french windows. If she left their room, however, he was implacable. He would not sleep without her. If she escaped, the horrors of the night would continue into the next day. He would drive to Paparoa for another bottle. The most constant and bitter of his reproaches was that she condemned him to silence and solitude. Almost always the threats and abuse evaporated without physical attack. He hammered on her door and bellowed in the hallway. He finished the bottle and took his plate of cold, overcooked food from the oven, ate it and went to bed, where she waited for him. On the days following the worst nights he was sick. The house stank. Vera and I left at 6am to milk the cows, leaving him silenced and suffering in their bed. The period of relief which followed these nights was longer than usual. No acknowledgement of what had happened was ever made when he was sober. He was torn between the desire to degrade her and fear that he might succeed in doing so. He reviled her constantly as a drunk. He wanted her to drink with him in order to revile her. He came home often with a bottle of sweet her. If she wine, sauternes, for her. He announced it as a treat, a "prezzy"or refused to drink with him, the abuse began early. To acquiesce and drink with him was even more dangerous. Alcohol made her less careful of her tongue: if she laughed at him, if she was irreverent or if she was sarcastic or careless of her wit he was enraged. If she became drunk he reached depths of loathing which almost deprived him of the capacity to articulate his hatred. You belong in the gutters of Fitzroy, where I found you. In the filth. In the filth. In the slums of Surry Hills. Get back to it. Back to it. Filth. Fitzroy. I believe he was terrified by what he perceived as her loss of control. Drunkenness in others induced his contempt and disgust. No contradiction 19931 WOMEN WHO KILL IN SELF DEFENCE 425 was apparent to him; his own episodic binges were so rigidly ritualised and surrounded by denial. In this mood he would hit her. Mostly his violence was oblique, disguised as an accident and accompanied by self deception. He slammed doors at her, pretending he had not seen her. He engineered collisions in the hallway or he barged through their bedroom door, knowing she might be crouched there, listening for his movements. If she was drunk, however, he hit her deliberately. He hit her with his hands open. He inflicted no lasting physical injuries. When he was done, he left her sobbing in the bedroom. He returned to the kitchen and let the dogs in and spoke to them. Pat, called "The Poose" and my dog Toby, called "Muckhole". I remember her crying loudly in the bedroom. is rage was exhausted and he was ashamed of what he had done. He was searching for a verbal expiation which would transform what had happened into a kind of joke. His voice had a playful tone at first, arch and warm, before drifting into self pity. The monologue went on for half an hour or more. It began with a question to his dog, Pat: Do you know Poose, there is two things poor Daddy did never understand. One of them is boozers.91 Boozers is strange. But strangest of all is wimmins. Wimmings. Do you understand Wimmings Poose? Poor Daddy did never understand them. We do love them though, don't we Poose? We love them more than they can ever know. Even when they don't love us. She doesn't love you, you poor little bugger. She won't let you have any children. Wimmings don't love no boozers. We do still love them though. When poor Daddy is deddibones they will know. Then they will know ... B. If my life should end this night The farm was failing. Bill Adams was no farmer. Most of his considerable energy was devoted to building. He made extensions to the house and built a woolshed, cattleyards and sheepyards. But he could not devote himself to the grinding discipline of improving his stock and increasing farm income. By late 1954 they were approaching a financial crisis. They cultivated the local bank manager and his wife with a round of lunches and picnics. Vera organised these occasions and, for a brief period, enjoyed again the civilities of ordinary social life. She began to play golf with the bank manager's wife. Bill could mimic the bank manager and his wife and revile them after they had gone, but he dared not insult them and he drank in a controlled way when they visited. He was amusing and expansive. Vera said she wished that their financial crisis could be extended and controlled and made to last indefinitely. Later, it became clear that picnics and lunches would not save the farm. She asked her mother Ada, then a woman of considerable wealth, for a loan. Ada came to New Zealand to stay at Kanumbra for a month and the demons were put away during her visit. The money was given, not lent, on condition that Vera secured a third mortgage over the property. Vera's second and third mortgages were a source of bitter resentment. He accused her of forming an alliance against him with lawyers, a hated profession since the maintenance claims arising from his first marriage. Sometime in 91 In our house, dogs were called boozers. The word had no obvious connection in meaning with alcoholic excess. 426 SYDNEY LAW REVIEW [VOL 15: 403 the ensuing year, during a night of unparalleled horror, he burned all her papers in the belief that he could in this way extinguish her legal claims to a share of Kanumbra. When they sold the farm, at the end of 1955, he received the proceeds as his own. We returned to Australia and they bought another farm, a sheep property on the border between New South Wales and Victoria, near Balldale, and settled there. Once again he called it Kanumbra, ill fated name, after the farm at the end of Utopia road. They were more isolated then they had been in New Zealand. He bought another Chevrolet, a large black sedan, which he named and appropriated to himself. I remained in Melbourne. There had been several violent confrontations and it was clear I could no longer live in the same house with my stepfather. I saw them rarely after I left home. Vera wrote frequently and at length and her letters provide a continuing account of their lives, occasionally frank with despair but more often coded and cheerful. A letter full of inconsequential news begun on a Monday and finished in a second instalment on the following Wednesday in April 1957, ends innocuously: Wednesday. Crutching finished today, with the usual celebrations. Such lovely weather -perfect days but cold nights ... Felix the pusscat is curling around me as I write. You wouldn't know him he is just beautiful. Five pages of the letter she never sent were interleaved in her copy of Cowper's Poems, found after her death: Tuesday night - so cold I can hardly write - it has been a thoroughly awful night, with all the trimmings. No need to explain to you ... Joe I am so terribly lonely and sometimes terribly afraid ... It is a dreadful feeling when death assumes the proportion of a release from all care, worry, striving and pain ... and yet I don't consciously want to die, with all its tribulations, life never burned so brightly in me, -why I wonder? Three pages now to destroy, or shall I? The second Kanumbra lasted a shorter time than the first. At the end of 1957 they retreated and rented a small house near the main street in Albury. Vera took a part time secretarialjob and Bill worked as a civil engineer with the local Council. Her life changed. She enjoyed an unprecedented measure of freedom when he was at work. She joined committees and invited friends home during the day. In return for this freedom it became more than ever necessary for her to maintain the boundary around the private squalor of their life together. Heavy drinking was mostly confined to the weekends. On Monday mornings she helped him to wash and dress and pushed him out the door, poisoned and ill with self loathing. It was necessary to placate and conciliate him. He was losing control, ceding responsibility to her and blaming her for his drunken rages. At first she left him alone in the house on the nights when he drank and she walked about the streets until he went to bed. So long as he was conscious, he stood in the back doorway and bellowed down the laneway, in the apparent belief that she would not go beyond earshot. Neighbours complained and the police came. He was warned twice and then arrested to spend a night in the cells. After two or three arrests, she remained with him on the nights when he drank. He was quieter and they reached a tacit accommodation with the neighbours. He abused her, she said, in whispers. She made an amusing story of it, of him sitting with his bottle, hissing at her. 19931 WOMEN WHO KILL IN SELF DEFENCE 427 They sold Kanumbra in 1960. Their efforts to maintain the property, while working in Albury, had become increasingly perfunctory. Then he was offered the job of chief engineer at the King Island Scheelite mine, re-opened during a brief flurry in the world market for tin. They left for the Island, accompanied by the dog Pat, a favourite cat and their black Chevrolet sedan, during Easter 1960. She wrote that the island reminded her of New Zealand, "the proximity of the sea, I suppose, the greenness, the dairy herds and the tea tree, which grows in profusion." The mining company provided a large, plain house in the township of Grassy, on a windswept hillside, overloolung the Strait. Here at last was a chance of achieving social acceptance and respectability. In the farming communities of Paparoa and Balldale they had been exotics. Their relationships with neighbouring families always had an edge of mutual derision. They were city people. They were too feverishly outspoken in company. They drank too much. On King Island their peculiarities were less obvious. The mining community at Grassy was composed, almost entirely, of exiles from the mainland. In this outpost all were exotics and the necessity for company drew them together. His tendency to turn every social occasion into a drinking bout mattered less. There was always someone willing to join him. She was no longer isolated. The Company provided him with a car and Vera was allowed at last to learn to drive the old black Chevrolet: Progressing quite well, except for reversing, which I hate. Chevie is ceasing to be a sort of Juggernaut endowed with malignant life. She cultivated her social position with the other mine wives, joined committees and devoted herself to charities. I wrote accusing her of turning provincial and she replied with faint amazement that she found herself a member of the Kindergarten Committee, the Girl Guides Committee, the Red Cross, the Ladies' Golf Club and vice president of the Combined Church of England and Methodist Ladies Guild: "I wonder did they mean President of Vice?" She wrote in the same self mocking vein of the Ladies' afternoon tea with the Tasmanian Governor and his daughter, when they visited the Island: The weather behaved superbly, their Excellencies were most gracious, the afternoon tea was sumptuous and we all floated home in quite an uplifted frame of mind ... She had less than a year to live. The horror of the nights when she was alone with Bill increased. Worse now because his drinking could not be confined to Friday nights and weekends. They lived on a knife edge. He had recovered his profession. He held as high a position in his professional life as he had ever done and he was approaching the alcoholic climacteric, the inevitable catastrophe when the collusive barriers between his private and public lives would collapse. She nursed him to work in the mornings with black coffee and Veganin tablets. I wrote to her in Spring, asking if she could come to Melbourne. Her letters were less guarded now: I am so tired today that I am afraid my letter won't be a model of either coherence or interest. I haven't had any sleep to speak of the past three nights, and I am one of those unfortunates who can't make it up during the daylight hours, when I have time and opportunity. Trouble in the Balkans - the usual. 428 SYDNEY LAW REVIEW [VOL 15: 403 About my coming to Melbourne, it's as ever problematical. Speaking frankly, it means hell before I go, with or without "permission" and hell for weeks afterwards when I come back. Even so, it would be worth it ...to enjoy for a short space the wonder and glory of being a natural human being, instead of Mrs Engineer Adams, which is rather like being a permanent occupant of the bed of Procrustes, which is never of the same dimensions for two days running. The visit never eventuated. She wrote a week later: My battle was lost, I fear, to wring "permission to depart", so all I've got is the scars thereof, and no victory ... Oh I had such plans, if once I could have got away. After her death I found a note in her fluent cursive, written about this time, interleaved with other notes, prayers and unsent letters: December 5,1960 If my life should end this night, God forgive me my sins and receive my soul. Bill, I wish I could help you, but you have sent me beyond your reach. No sign of this crisis was apparent in her next letter, twenty pages of a joyfully exuberant account of the 1960 Christmas celebrations. There were dinner parties, golfing parties, swimming parties and parties for Christmas and New Year: Don't ever let me hear anyone say Grassy is dull! I suppose it was really the gayest Christmas I've ever had, and if the gaiety rang a little faIse at times, I still did enjoy most of ... Kristmas on King Island. You will be thinking what a frivolous old horror I am, but honestly I enjoyed it. There were people, people and people! Instead of the dreary drinking between ourselves ... Her letters written after this, during the early months of 1961, were unexpectedly dispirited. Something had changed, but the nature of the change was not immediately apparent: It is a lovely day today. I am taking the car over to Currie with a load of ladies on board -CWA meeting ...The Guild are holding a street stall -cakes and produce - next Thursday, so will be baking frantically all day Wednesday, and serving on the damn thing Thursday. There are always meetings to attend, but I don't mind. It fills in time, whilst I'm waiting. For -what? Frankly I don't know, but if it's fulfilment of my ideals, development of my ego, satisfaction of my higher impulses or reward for duty nobly done, it certainly won't be found in this marriage of mine, which seems to be steering closer to the rocks every day. Furthermore, I find I don't greatly care. Things which once used to hurt and infuriate, now only excite contempt and irritation. But that's enough of that ... When I visited the Island in March it became apparent what had happened. Bill was pursuing the wife of S, a tradesman who worked for him. It is unlikely that anything came of his pursuit. Mrs S encouraged him but kept him dangling, meshed in the nets of his own ineptitude where matters of human affection were concerned. It was the first time he sought to be unfaithful to Vera and he was making a public fool of himself, humiliating her. He was absent 19931 WOMEN WHO KILL IN SELF DEFENCE 429 much of the time and my visit passed without the usual recriminations or drunken scenes. After a numb farewell to her at Currie airstrip I returned to Melbourne. A week later she was drowned. I flew to King Island for the funeral, convinced that Bill Adams had killed my mother. I questioned C, another of the tradesmen who worked for him. C and his wife had been on the beach with them that morning. He was a kind man, one of the few men I ever met who were genuinely fond of my stepfather, and he was deeply embarrassed by my hostility and suspicion. He told me it was an accident. Bill and Vera had taken a picnic lunch to the beach with S and C and their wives. It was a brisk day and the sea was boisterous on the unsheltered beach. The others swam for a while and then came in but Vera was slow to follow. Someone called to her. Then, as they watched, she was carried helplessly away. Bill did his best to save her. He reached her but could not hold her in the heavy sea. No man could have done more, C said. He was exhausted and unconscious when rescuers pulled him in: "We almost lost your Dad too". It had been a happy morning C said. Nicky had been happy. He resigned from the mine within weeks of her death and went to Western Australia, taking with him C and his wife. C told me later that Bill's drinking was beyond control in this period and they lost touch with him when he drifted back East. He was briefly married during this period. It was a disastrous union and the woman left him after three months. When I visited him many years later, in 1977, he said nothing of this marriage. He told me that he took up turkey farming for a time in the Riverina, but the market collapsed. Then he married again and settled with his fourth wife near Coonabarabran in New South Wales. She loved him and cared for him devotedly until he died. They lived on a small farm and he achieved modest prosperity as a local builder. Cigarettes killed him in the end. Lungbusters. He died slowly, his huge frame wasted by emphysema, in 1985. There was nothing particularly unusual in the pattern of Bill and Vera's marriage. It was unusually well documented in the later years, however, for Vera was an articulate and copious correspondent. Human choices are always overdetermined. Vera was neither helpless nor incapable of leaving Bill. She stayed with him because she believed she had no better alternatives. He controlled almost everything she had accumulated by years of patient work and saving. She had invested herself and everything she owned in him. She stayed with him because she believed in marriage and hated the thought that she might fail at it for the second time. I have not shown the ordinariness, the normality, of our lives between the nights of violence. Much of that was her achievement. We did the things most families do. And she loved him. He may even have been an exciting lover. Though sexual repression impoverished him, he was naturally a sensuous man. I see him still with a child's eyes and cannot make him loveable. But Vera loved him and stayed with him because there seemed no alternative to loving him. When he was sinking she kept him afloat. Loyalty and pride, the usual combination, led her to conceal the squalor of our lives together, maintaining and defending an appearance of normality. There is no doubt that he was dangerous. Dangerous enough to kill her intentionally? The risk was incalculable, but quite real. Neither of them could 430 SYDNEY LAW REVIEW [VOL 15: 403 know its extent. In this relationship, as in many others, the essence of terrorism was to be found in the unpredictability of violent attack. Unpredictability is heightened by theatrics and enhanced, rather than diminished, by bathos. Domestic terrorism oscillates between reality and absurdity. A missed cue could prompt him to a sudden and unprecedented attack. Though his violence was ritualised, it was not completely within his control. I hesitate to generalise too explicitly from this account. The point of the story was to present, with pitiless clarity, the texture of one particular violent relationship. Two obvious points can be made however. The relational context of domestic violence finds no explicit recognition in the law of self defence. Domestic violence involving oppression and injuries far worse than those which Vera suffered may last for years. Judgments about the seriousness of the threat, its meaning and the measure of excusable force in self defence cannot be divorced from the history of the relationship. One reason for American acceptance of the battered woman syndrome was that it provided a method, however imperfect, of giving context and coherence to the final act of self defence. The second point is related. The degree of harm threatened on any particular occasion is often incalculable. Almost always the aggressor stops himself at some point short of the harm he could inflict. The constant element in the violent relationship is reliance on terror. Subjection to terror is itself a harm which might justify or excuse the use of force in self defence. The law gives no explicit recognition, however, to the possibility that conduct calculated to subject another to terror might amount to a ground for self defence. In American jurisdictions, the battered woman syndrome provided an imperfect means of articulating this element in cases of self defence. Imperfect because the fact of terrorism is obscured, located in her, in her learned helplessness, rather than in his oppression. 4. Justifications, Excuses and Agent Relativity North American legal theorists tend to characterise self defence as a plea which justifies rather than excuses homicide.92 Preoccupation with the distinction between justification and excuses is particularly evident in the literature on defensive homicides by battered women. In her exhaustive survey, Cathryn Rosen argues the dissenting view that self defence is misconceived as a justification. Rosen maintains that self defence is an excuse, capable of accommodating defensive homicides by women in violent relationships.93 George Fletcher may be credited with starting this particular hare in Rethinking Criminal Law94 and other publications of the mid to late seventies, though the historical roots of the distinction lie far deeper in American common law. Elsewhere, Fletcher provides an economical characterisation of justified conduct as the "right and proper thing to do" in the circumstances.95 Killing in self defence is said to be justified if it was the lesser evil 92 Dressier, J, "Justifications and Excuses: A Brief Review of the Concepts and the Literature" (1987) 33 Wayne LR 1155. 93 Above n44 at 25. 94 Fletcher, G, Rethinking Criminal Law (1978) chlO . 95 Fletcher, G, "Should Intolerable Prison Conditions Generate a Justification or an Excuse 19931 WOMEN WHO KILL IN SELF DEFENCE 43 1 in the circumstances. Conduct which is merely excused, on the other hand, is regrettable - the wrong thing to do in the circumstances. Horwitz puts the argument succinctly: "The person on the receiving end of a justified act has no ground for complaint. The person on the receiving end of an excused act does have a legitimate ground for complaint".% If the act cannot be justified, the actor may be excused nevertheless, on the ground of some incapacity for obedience to law. The plea of insanity is frequently instanced as typical of the excuses.97 In Fletcher's view, excuses require evidence of "normative involuntariness" - an absence of the capacity for any real choice in the circumstances: "[pllanning, deliberating, relying on legal precedents -all of these are incompatible with the uncalculating response essential to 'involuntary' conduct.'98 The debate over justification and excuses has generated less enthusiasm in England and Australia.99 Case law generally ignores the difference, referring indifferently to self defence both as a justification and an excuse.loo The potential significance of the distinction is obliterated by reliance on standards of reasonable proportionality, reasonable necessity and the likely or possiblelol reactions of the ordinary or reasonable person. These flexible criteria displace any sharp distinctions between categories of justification and excuse.102 Self for an Escape?" (1979) 26 UCLA LR 1355. See also, Fletcher, G, A Crime of SelfDefence: Bemhard Goetz and the Law on Trial (1988)at 19.To similar effect: Yeo, above n29 at 58. Yeo adds the qualification that conduct may also be justifiable when it is permissible, rather than positively desired. Once made, the qualification tends to be ignored and justified acts are described as "socially desirable". See, eg, Yeo's discussion at 104-106.In the United States, where debate on this particular issue has been lively if inconclusive, Dressier, above n92 at 1163-1165, provides a summary account of divergent views. 96 Honvitz, D, "Justification and Excuse in the Program of the Criminal Law" (1986) 49 Law and Contemporary Problems 109 at 125. 97 See, eg, Fletcher, above n95. Byrd, S, "Wrongdoing and Attribution: Implications Beyond the Justification-Excuse Distinction" (1987) 33 Wayne LR 1289 at 1299. In most instances reliance on the example of insanity is no more than a significant expository device. See,however, above n96 at 117: "duress is plainly in the same category of mental incapacity as insanity". 98 Fletcher, above n94 at 811-812;"Paradoxes in Legal Thought" (1985) 85 Col LR 1263 at 1280. Fletcher assimilates the excuses under the title of moral or "normative involuntariness". The theory has been substantially received into Canadian case law on necessity, discussed by Colvin, E, Principles of Criminal Law (2nd edn, 1991)at 238-253. 99 Among Australian authorities, Yeo, above n29 at 105-112 is one of the few who argue for the importance of the distinction. He draws on Fletcher's work, arguing that self defence is a justification, available only when it can be said that the force used against an aggressor was a lesser evil than infliction of the threatened harm. Fisse, W B, Howard's Criminal Law (5th edn, 1990)refrains from discussing the distinction. Smith, J C, Justifcation and Excuse in the Criminal Law (1989) rejects its significance. Colvin, E, writing from a Canadian perspective in his Principles of Criminal Law (2nd edn, 1991) at 208-211 argues that it has no place in the common law tradition. See also Colvin, E, "Exculpatory Defences in Criminal Law" (1990) 10 Odord Leg Stud 381. 100 Yeo, above n29 at 109-111. An exception must be made, however, for the continuing debate on the question whether self defence is limited to cases of defensive force against unlawful conduct or extends more generally to self defence against lawful conduct as well: see Lawson & Forsythe [I9861 VR 515, in which the historical distinction between justifiable and excusable homicides is brought to bear on this question. The argument is about precedents and the course of case law development, however, rather than an inquiry into conceptual refinements or philosophical distinctions. 101 See R v Abusafiar (1991) 56 A Crim R 424, at 433-435, per Hunt J, on the likelihood and possibility standards in duress and provocation. 102 Colvin above n99 at 388-389. 432 SYDNEY LAW REVIEW [VOL15: 403 defence in Australian common law is an example of an "agent-relative permission".lo3 An acquittal on this ground is perfectly consistent with refusal to approve or endorse the actions of the accused. The accused is guilty of no offence if the use of deadly force in the particular circumstances was acceptable or not beyond the limits of tolerable conduct. Appreciation of the radical flexibility of the Australian common law of self defence is sharpened by acquaintance with the polarities of the American debate. Self defence is narrowly restricted in its operation if it is characterised as a plea which requires the accused to show that killing the aggressor was the right and proper thing to do, a socially desirable thing or the lesser of two evils. Cathryn Rosen's argument that self defence is an excuse was meant to give the plea a more extended operation. Given the American tendency to conflate excuses with pleas of mental incapacity, self defence conceived as an excuse might be based on evidence of psychological disorder and coupled with the concession that the use of deadly force was wrong. American attempts to rationalise the law of self defence in its application to battered women who kill oscillate between these rigid classifications of justification and excuse.104 Professor Fletcher's claim that nothing less than justification will do when self defence is in issue draws on an earlier tradition of United States scholarship, exemplified in classic papers by Joseph Beale Jr.105 Beale was much concerned with the need to limit appeals to personal honour as a ground for the use of deadly force: "The problem now in America is the same as it was three centuries ago in EnglanP.106 His exhaustive restatement of the common law rules requiring retreat rather than retaliation, a restatement intended to restrict the scope of self defence, ends with a plea for compassion: A really honourable man, a man of truly refined and elevated feeling, would perhaps always regret the apparent cowardice of a retreat, but he would regret ten times more, after excitement of the contest was past, the thought that he had the blood of a fellow human being on his hands. It is undoubtedly distasteful to retreat; but it is ten times more distasteful to kill.107 103 Id at 392. Colvin suggests a similar approach, refemng to self defence as a "defence of contextual permission". 104 The most exhaustive of the North American studies in this vein is by Rosen, above n44 at 25 (self defence as excuse rather than justification). Cf Crocker, above n33 (justification, rather than excuse); Taylor, L, "Provoked Reason in Men and Women: Heat-of Passion Manslaughter and Imperfect Self Defence" (1986) 33 UCLA LR 1679 (excuse rather than justification); Acker, J and Toch, H, "Battered Women, Straw Men, and Expert Testimony: A Comment on State v Kelly" (1985) 21 Crim Lmu Bull 125 (justification rather than excuse). 105 Beale, J, "Retreat from Murderous Assault" (1903) 16 Harv LR 567; "Homicide in SelfDefence" (1903) 3 Columbia LR 526. In above n94 at 867, 870, Fletcher draws on the more remote authority of Blackstone, 4 Commentaries 181, at 185. His brief account takes insufficient account of Blackstone's distinction between justifiable homicide (to prevent crime) and excusable homicide. se defendendo: see 183ff. The distinction is discussed in Rosen, C, "On Self Defence, Imminence, and Women Who Kill Their Batteren" (1993) 71 Nth Carolina LR 371,382ff. 106 Beale, id at 580. 107 Id at 581. 19931 WOMEN WHO KILL IN SELF DEFENCE 433 Very similar concerns are evident in Fletcher's monograph on the trial of Bernie Goetz, the "subway vigilante", who shot and wounded four Black youths after one asked him for money in a New York railway carriage.108 Goetz anticipated violence. He went out armed, prepared for it. For him they were "four young muggers", like those who had attacked and injured him some years previously. He took the request for money as the prelude to a mugging. Goetz was acquitted of all charges relating to the severe injuries he inflicted on his victims. He was convicted of the minor offence of possessing a loaded gun in a public place.109 Once it is accepted that self defence is a justification, in the sense which Fletcher gives the word, there is every reason for concern over the question what is to count as a justification for killing another human being. This is a recurring concern in the American jurisprudence of self defence.110 The characterisation of self defence as a justification rather than an excuse is intended to limit the availability of the plea. Though the point is rarely made so bluntly, limiting self defence in this way is meant to secure convictions. Convictions are supposed to vindicate the law and reaffirm the sanctity of human life. There is, in the work of scholars who take the distinction between justification and excuse seriously, a preoccupation with symbolism and the "proper message" conveyed by the outcome of trials where self defence is pleaded.111 That preoccupation was particularly evident in the aftermath of the trial of Bernhard Goetz. Fletcher's study of the Goetz trial is accompanied by the bitter reflection that it represented a "return to absolutist thinking about self defence, a return to the individualist philosophy that ignores the costs of a necessary defence to the aggressor". In Fletcher's analysis, a decision that the shots were fired in self defence entails the conclusion that the shots were justified and Bernhard Goetz represents a model for others to emulate. The effect of the trial was to legitimate anticipatory strikes when threats are perceived in equivocal conduct. The issue bears a deeply felt moral dimension for Fletcher. This is no academic game. Like Joseph Beale, almost a century earlier, Fletcher's most compelling argument is a call for human trust and human kindness: The more Troy Canty, Barry Allen, James Rarnseur and Darrell Cabey come into focus as human beings worthy of our compassion, the more we might expect Goetz to have taken some risks that he could exit from the confrontation on the train without inflicting the human costs that he did.112 No comparable sense of moral urgency animates discussion of the issue in Australia where rules which once set determinate limits to the availability of self defence have been displaced by the accommodating elasticity of general 108 Fletcher, above n95. 109 Id at 71-72 provides a short account of the intricacies of New York gun law. 110 See in particular Beale, above n105. 111 Dressler, above n92 at 1170. 112 Fletcher, above n95 at 37. In preceeding pages, at 30-33, Fletcher adheres to his earlier view that self defence is a justification. His argument is based partly on the history of the defence: in that respect his account is fundamentally at odds with the conclusions drawn from the historical record by Ormiston J in R v Lawson & Forsythe [I9861 VR 515. See also, above n44. 434 SYDNEY LAW REVIEW [VOL 15: 403 principles. There is no paradox in the fact that legal moralists in the United States, a society far more violent than our own, are so concerned with defining limits to the use of self defensive force. The jurors who tried Bernie Goetz were well aware that social intercourse between strangers - a simple request for information or help -can be a prelude to violent attack. When violence is ubiquitous it may begin to seem reasonable to assume the worst when signals are equivocal. It does not follow, however, that pre-emptive force is legitimate because it is reasonable to fear strangers. The more pervasive and reasonable that generalised fear of violence, the greater the need to protect innocents against a mistaken resort to pre-emptive force. One response to ubiquitous violence is to demand a higher level of tolerance of risks and more rather than less in the way of self restraint against giving way to suspicion.113 Reflection on the trial of Bernhard Goetz led Fletcher on to construct a tangled and highly significant analogy. He considers the retaliatory element in Goetz' attack on the youths: Retaliation, as opposed to defence, is a common problem in cases arising from wife battering and domestic violence. The injured wife waits for the first possibility of striking against a distracted or unarmed husband. The man may even be asleep when the wife finally reacts. Goetz's response to the four young blacks was retaliatory so far as he perceived them as "four young muggers" rather than as individuals; he was striking back for having been mugged by the "same type of guys" in 1981 and suffering lasting injuries to his knee and chest. Retaliation is the standard case of "taking the law into one's own hands". There is no way, under the law, to justify killing a wife batterer or a rapist114 as retaliation or revenge, however much sympathy there may be for the wife wreaking retaliation. Private citizens cannot function as judge and jury towards each other.115 Something more than the distinterested passion for abstraction drives this bizarre comparison between Bernhard Goetz and the battered woman who strikes back. In these brief asides Fletcher invokes the threat of bedroom vigilantism, women encouraged to retaliate against masculine violence by overindulgent applications of the law of justified self defence. Alarm at that prospect may account for the strained falsity of the analogy. Goetz was moved by racial prejudice to shoot four strangers. He extended no trust to his victims, assuming the existence of a threat. There is no analogy. The battered woman does not act in ignorance, prompted by prejudice. She knows all too well the reality of the threat represented by her partner's violence. Nothing in Fletcher's preceding argument prepares the ground for this diversion or justifies the crudity of the comparison. It is an afterthought, reflecting a charac- 113 That is not the only response, of course. Goetz enjoyed brief fame and the applause of those who approved his vigilantism. 114 The dead rapist, who reappears at several points in the ensuing discussion, is not meant to be a victim of retaliation or defence by a wife who is raped by her husband. This is evidently a reference to two well publicised cases of defensive homicide against non spousal rape in the mid-seventies. See State v [Inez] Garcia (1976) 54 Cal App 61; Shte v [Joan] Little [Supreme Court, North Carolina 19751; discussed in Schneider, Jordan and Arguedas, above 1118ch 1-5; Walker, Thyfault and Browne, above n49. 115 Above n95 at 20-21. 19931 WOMEN WHO KILL IN SELF DEFENCE 435 teristic preoccupation with the familiar patterns of violence between men and a failure of imaginative identification with the victims of domestic violence. The central element in the restrictive American view of self defence as justification is found in Beale's insistence that "the interests of the state alone are to be regarded in justifying crime1'.l16 For Beale, this was a restriction intended to curb the frontier ethic of duelling on sudden quarrel or retaliating with deadly force for affronts to honour. Fletcher makes essentially the same point when he argues that [Cllaims of justification lend themselves to universalisation. That the doing is objectively right means that anyone is licensed to do it ... the victim has no right to resist, and other persons acquire a right to assist ... "7 It follows that my right to defend myself can rise no higher than the right of the State, or an impartial third party, to intercede on my behalf. On this fundamental point, Australian common law conflicts with North American theory. Earlier I suggested that the Australian common law of self defence rests on an "agent relative permission" to use force against an aggressor. If it is necessary to choose a label, self defence is an excuse rather than a justification. Agent relative permissions are not caught, however, in the net of distinctions which have been drawn between justification and excuse. Several theorists have argued the case for agent-relativity. It is a fugitive theme in the modern literature on self defence.118 Though the central idea is simple, it makes a profound difference to understanding the exculpatory effects of a plea of self defence. It has the additional virtue of relegating much arid speculation over the borderlines of excuse and justification to irrelevancy. Nancy Davis puts it this way: The permissibility of force in self defence is "derived from the value which each person assigns to his or her own life, and not from whatever value may be assigned to its preservation from some impartial point of view."ll9 Kent Greenawalt is less formal: "Because of the threat to his own vital interests the victim might be excused for protecting those interests in a manner that would not generate an excuse for the disinterested intervenor''.l20 Though the terminology is awkward and unfamiliar, agent-relativity is immanent in the Australian law of self defence. Self defence is like this. If I am attacked there may be a choice to make among the possibilities of counter attack, flight and submission.121 Submis116 Beale, above n105 at 582. 117 Fletcher, above n98 at 761-762. The universalisation thesis is the subject of contention. It is discussed by, inter alia, Dressler, J,"New Thoughts About the Concept of Justification in the Criminal Law: A Critique of Fletcher's Thinking and Rethinking" (1984) 32 UCL4 LR 61 at 95-98; Dressler, above n92 at 1173; Byrd, above n97 at 1332-1341; Gur-Arye, M, "Should a Criminal Code Distinguish Between Justification and Excuse?' (1992)5 Cmad i m J Law & Juris at 215, Smith, above n99 at 19-44. 118 In the account which follows, I have drawn on Parfit, D, "Innumerate Ethics" (1978) 7 Philosophy m d Public Affairs 285; Uniacke, S, "Self-Defence and Natural Law" [I9911 Amer J Juris 73; Greenawalt, K , "Violence - Legal Justification and Moral Appraisal" (1983) 32 Emory L 437; Davis, N, "Abortion and Self Defence" in Garfield, J Land HenJ nessy, P (eds), Abortion: Moral and Legal Perspectives (1984). 119 Davis, id at 197-198. 120 Greenawalt, above n118 at 460. 121 Self defence traditionglly includes defence of relatives or intimates: see O'Connor, and 436 SYDNEY LAW REVIEW [VOL 15: 403 sion can include the resignation of property to an aggressor, passive endurance of physical violence and physical abuse or a self abasing surrender and acceptance of humiliation. Sometimes there is no opportunity for counter attack, in which case the issue of self defence will not arise. Sometimes there is no opportunity for flight or submission. The aggressor will kill unless stopped. There may be no alternative to the use of deadly force and no question that deadly force is reasonably necessary. The choices are not always so constricted. There are disputable cases, where some course of action other than the use of deadly force was possible. Imagine a dialogue between a woman whose partner threatens violence and conscience, an imaginary and impartial observer. A dialogue similar to the one Devlin J asked the jury to imagine in his direction to the jury in the trial of Mrs Duffy, who killed her husband after a beating, the last in a prolonged course of marital abuse.122 It does not matter how cruel he was, how much or how little he was to blame, except in s o far as it resulted in the final act of the appellant. What matters is whether this girl had time to say: "Whatever I have suffered, whatever I have endured, I know that Thou shalt not ki11."123 How is she to respond to this imaginary voice? We are to imagine an internalised and impartial observer - the spectatorial Conscience - who asks her to sacrifice her own interests for those of her husband or perhaps for some larger good. Of course there are times when sacrifices should be made. As the costs of sacrifice rise, however, there is a point beyond which she will sacrifice herself no longer. Her own interests, her own integrity, are at stake. She answers Conscience by excusing herself from compliance. She will make no further sacrifices of self. There is a critical point at which she is entitled to do SO, whether or not her choice meets with the approval of the imagined, impartial observer.124 There is a difference between the standards of the impartial observer and those of the individual threatened with violence. The observer might suggest that the sum of human happiness would be less diminished if she submits to the beating instead of responding to his violence with deadly force. How is a beating to be weighed against a life? The American literature on self defence Fairall, above n38 at 171-172. Self defence modulates by insensible degrees into disinterested intervention to prevent crime. References to self defence in the text are meant to include cases of defence of family and other intimates. Women who use force to prevent domestic violence frequently act to protect their children as well as themselves. 122 R v Durn [I9491 1 All ER 932. There is a m r comprehensive account of the facts of the oe Law: Text a d Cures (6th edn, 1989). case in Brett, P, Waller, Land Williams, C R, Cn'rm'~1 123 Id at 932. The full direction is unforgettable in its measured cruelty. Mrs Duffy's defence of provocation was rejected by the jury and she was convicted of murder. In the Court of Criminal Appeal, Lord Goddard CJ described it as a "classical direction ... in a case in which the sympathy of everyone would be with the accused person and against the dead man and it was essential that the judge should see that the jury had an opportunity of vindicating the law, whatever the consequences might be". Though the direction was endorsed in R v Thomton [I9921 1 All ER 307, which also involved retaliation following domestic violence, it appears to have altered in meaning: see text accompanying 1111251-56 below. I24 As Partit, remarks, above 11118at 287: "Most of us ... believe that we may give priority to our own welfare. This priority should not be absolute. Perhaps Y could save his arm rather than X's life; but he ought to save X rather than his own umbrella." 19931 WOMEN WHO KILL IN SELF DEFENCE 437 against intimate aggressors is tormented with these utilitarian calculations.l25 The precise nature of the calculation varies: Are the value of their respective lives to be compared? Is the value of his life to be discounted because he initiated the attack? Does one take into account the risk that condoning her conduct might encourage other women to fatal violence?l26 If self defence is an agent relative defence, these questions are simply irrelevant. Whatever the nature of the criteria supposed to justify the use of deadly force by disinterested intervenors, we have no right to expect anyone subjected to violent attack to judge their own conduct in that way. Each of us is entitled to put a higher value on our own lives and our own integrity than the impartial observer will do. We sometimes excuse ourselves from compliance with legal rules because obedience is simply not required, in the circumstances. The exemption from compliance is not granted because disobedience is right or proper.~ompliancemight be praiseworthy. Nor is the exemption extended on the ground of incapacity for compliance. Obedience may have been possible. The excuse is available because no human agency has the right to demand further sacrifices for the sake of the aggressor. In cases of self defence against domestic violence, a woman is entitled to defend herself, by deadly force if necessary, well short of the break-even point of utilitarian calculation. She is not required to plead or pretend incapacity, weakness or moral t breakdown. She claims. rather. that her decision to ~ u her own interests first should be respected. The law bf homicide, as it apblies in cases of defensive force against domestic violence, is not the place to enforce moralities of selfless sacrifice.127 Imaginary dialogues and calculated choices are bloodless abstractions of coursz. In the real world of domestic conflict, terror and anger prompt the resort to deadly force. It is not terror, anger or other overwhelming emotion which excuses the use of force however. If overwhelming emotion was an excuse there would be nothing to choose, in terms of exculpation, between aggressors and their victims.128 Much unforgivable domestic brutality is the product of routines of explosive emotional release which he -almost always he - has developed over years of domestic oppression. Deadly force used in self defence is excused when it is reasonable to say that no-one - neither the aggressor nor the disinterested observer - is entitled to ask her for a further sacrifice of her own interests to those of the aggressor. If the law demands 125 Most notably above n44. 126 Fletcher, (1979) 26 UCLA LR 1355,above n95 discusses the paradoxes inherent in utilitarian calculations of the consequences of a decision to allow a defence of justification or ex165 cuse. Compare Dressier, above n92 at 1164-1 where the alternatives to utilitarianism are described. For an exhaustive discussion of this and related topics, see Hurd, H, "Justifiably Punishing the Justified* (1992) 90 Michigan LR 2203. 127 See the discussions in various papers by Greenawalt, by far the most perceptive of the American theorists on the plurality of moralities involved in the justification or exculpation of defensive force: above n118; 'The Perplexing Borders of Justification and Excuse" (1984) 84 Columbia LR 1897 at 1904-1907; "Distinguishing Justifications and Excuses" (1986) 49 Lmcr and Contemporary Problem 89 at 97-99. See too, Kadish, S, "Respect for 64 Life and Regard for Rights in the Criminal LawW(1976) CalifLR 871 at 882-883. 128 The arguments are given cogent expression in Singer, R, "The Resurgence of Mens Rea: I - Provocation, Emotional Disturbance and the Model Penal Code " (1986) 27 Boston Coll LR 243. SYDNEY LAW REVIEW [VOL 15: 403 more, requiring unreasonable sacrifices, it forges a compact of complicity with the aggressor.129 There is another consequence of recognising the agent-relativity of self defence worth passing mention. Among the theoretical puzzles which exercise American theorists is the question whether it is justifiable to use deadly force against an aggressor who is insane or otherwise without moral or criminal responsibility for the attack. Once again, Fletcher appears to have provided the impetus for this particular debate.130 The utilitarian calculus seems to provide a reason to doubt whether "any positive benefit ... accrues to anyone other than the killer from the taking of an aggressor's life in self-defence9'.131 How does one measure the comparative worth of lives? Fault on the part of the aggressor tips the utilitarian balance for these theorists: aggression results in a forfeiture of rights.132 If the aggressor is without fault, however, where does one find the justification for defensive force? We may assume that Cathryn Rosen reflects the concerns of current American scholarship when she refers to the diminished responsibility of men who batter women. They too, she writes, "are victims of 'disease' or of their social reality". Their absence of fault or diminished fault, "makes it even more difficult for the legal system to determine that the abuser's life is less valuable than his victim's" life.133 At this point we may have reached the nadir of absurdity in the American debate over self defence, justification and excuse. Her plea of self defence does not require the value of his life to be measured against hers in some imagined set of utilitarian scales. Whether he is to be blamed or not cannot determine her guilt.134 Her liability depends on the question whether or not it is reasonable to conclude that she should have sacrificed her own interests for his sake. So much for theory. I remarked at the outset of this discussion that the concept of agent-relativity was immanent in the Australian law of self defence. Agent-relativity is apparent in the general rule that defensive force is excusable unless the prosecution can prove that force was unreasonable in the circumstances. There is no trace of utilitarian calculation in this enquiry.135 129 Bentham, J, Theory of Legislation (1914) Vol I , 45, The "right of defence is absolutely necessary ... Take away this right ... and you become an accomplice of the base and the mischievous". Fingarette, H, "Victimisation: A Legalist Analysis of Coercion, Deception, Undue Influence and Excusable Prison Escape" (1985) 42 Wmhington & Lee LR 65, utilises a very similar argument in his formulation of criminal law defences. 130 Fletcher, G, "Proportionality and the Psychotic Aggressor: AVignette in Comparative Criminal Theory" (1973) 8 Israel LR 367; Fletcher, G, "Justification and Excuse in the Judaic and Common Law:The Exculpation of a Defendant Charged with Homicide" (1977) 52 NYU LR 599. 131 Aboven44at51. 132 Bentham, above 11129at 45 provides an appropriate source for this particular notion: "were you to slay ten lawless assailants, their death would be a gain to society, whereas the loss of one upright and amiable man is a great evil". For subsequent discussions of forfeiture and aggression, see Kadish, above n127 at 883-885; Dressier, above n92 at 11641165. 133 Above n44 at 51. 134 Fault is not irrelevant of course. Greenawalt, above n118 at 453-454 suggests some of the ways in which respective degrees of fault of the victim and defendant might influence the final assessment of responsibility. 135 Compare Yeo, above n29 at 66: "Whatever the threatened harm ... courts will be prepared to recognise it for the purposes of self-defence with the important consideration being 19931 WOMEN WHO KILL IN SELF DEFENCE 439 Australian common law accepts that deadly force may be reasonable in self defence though the threat involves neither serious nor permanent injury. The nature of the limits on defensive force will be addressed presently. There are more particular grounds for the assertion that agent-relativity is immanent in Australian common law. In the High Court decision in Zecevic v DPF"36 the majority expressed the law of self defence in stark and simple terms: the question "is whether the accused believed on reasonable grounds that it was necessary to do in self defence what he did". The Court accepted the conclusion that "reasonable belief' in this formulation is to be understood in the sense given to the expression in the earlier decision in Viro v The Queen. It refers "not [to] what a reasonable man would have believed, but [to] what the accused himself might reasonably believe in all the circumstances in which he found himself '.I37 The reasonableness of retaliatory action is to be judged from the defendant's point of view, not from the point of view of an impartial observer.138 Those who hold that self defence is a justification rather than an excuse, are committed to the view that the rights of an individual faced with a threat of attack can rise no higher than those of an independent and impartial observer. That is the essential meaning of Beale's claim that "the interests of the state alone are to be regarded in justifying crimeW.l39 is abundantly clear It that a trial direction which restricted the jury's consideration of a plea of self defence in this way would conflict, in the most fundamental way, with prevailing Australian case law. 5. Reasonable Self Defence The law of self defence against physical attack in South Australia, Victoria, New South Wales and the Australian Capital Territory, can be stated with extreme simplicity: no offence is committed if force, even deadly force, was used in the reasonable belief that it was necessary to avert the threatened harm.140 Of the jurisdictions which have codified their criminal law, Tasmania appears to have adopted essentially the same criterion.141 The law in other code jurisdictions is slightly more complex. Queensland, Western Australia and Northern Territory all adopted legislative provisions which supplement the naked reference to "reasonable necessity" with additional rules which whether the force applied by the accused to counter the threatened harm was proportionate to such harm. This approach accords with the justificatory nature of self defence". Requirements of proportionality are endemic in our law. They are, perhaps, loosely consistent with the concept of justification. They imply no positive support for the proposition that self defence is a justificatory defence which is only available to those who choose the lesser evil. 136 [1987] HCA 26; (1987) 25 A Crim R 163; (1987) 162 CLR 645. 137 (1978) 141 CLR 88 at 146, per Mason J, Stephen and Aickin JJ concurring. 138 In Zecevic v DPP [1987] HCA 26; (1987) 25 A Crim R 163 at 183, Deane J, who dissented on another ground, argues at some length that the reasonableness of action taken in self defence is to be judged from the standpoint of the threatened victim. For related formulations in the decisions of English and State courts, see the acute analysis of authorities in Yeo, above n29 at 220-226. 139 Beale, above nl05 at 582. 140 Zecevic above 11138at 170 per Mason CJ; at 176 per Wilson, Dawson, Toohey; at 179, per Brennan J; at 190 per Gaudron J. 141 Criminal Code (Tas) s46. 440 SYDNEY LAW REVIEW [VOL 15: 403 limit the occasion and degree of permissible force.142 In all jurisdictions the plea of self defence can be based on a mistaken belief in the necessity for defensive force. In the majority, however, the mistake must be reasonable: a mistake which was unreasonable in the circumstances will not provide a ground for acquittal.143 For simplicity, the account which follows is limited to South Australian, Victorian and New South Wales law on the use of deadly force144 in self defence. In the common law States, acceptance of the criterion of reasonable necessity as the essence of self defence has reduced the doctrinal content of the law to minimal levels. Criteria which refer to standards of reasonableness can mask more or less determinate rules of course. The development of the law of provocation has been marked by continuing tension between an objective criterion - could an ordinary person have been driven to the same degree of violence as the accused? - and dogmatic insistence on a particular set of limiting rules. The Australian common law of self defence is more loosely structured. There are few determinate rules. The sparseness of the law of self defence is reflected in the leading Australian text on the principles of criminal law which dismisses self defence in less than three pages.145 Provocation, a doctrine of far less practical importance, receives far more detailed treatment. Inarticulate criteria may be preferable on occasion to dogmatic reliance on limiting rules. There are, nevertheless, perils in their use. Biassed or prejudiced conceptions of what can amount to an occasion when deadly force is reasonable or necessary can flourish in the absence of explicit guides or restraints. Self defence may be withheld from consideration for no better reason than dogmatic insistence that no reasonable jury could entertain the possibility that deadly force was a reasonable and necessary response to threatened harm. Directions to juries may fail to provide a balanced and appropriately full account of the circumstances which can make the resort to deadly force reasonable. Recent divisions among appellate judges on the question whether self defence should have gone to the jury indicate a diversity of judicial views on the limits of self defence. The remainder of this essay charts the parameters of reasonableness in the use of deadly defensive force against intimate aggressors. A. Proportionality of Response Contemporary Australian case law appears to have abandoned the dual criteria of necessity and proportion found in earlier authorities.146 Reasonable neces142 The Criminal Code (Qld) ~ ~ 2 7 1 , 2 and Criminal Code (WA) ss248, 249, distinguish be72 tween situations involving self defence against an unprovoked attack and self defence against a provoked attack. These provisions, like s28F of the Northern Tenitory Criminal Code, limit the use of force intended to cause death or grievous bodily harm to situations where the accused was faced with a threat of death or grievous bodily harm. 143 The exceptions are to be found in the Criminal Code (Tas) s46; and the Criminal Law Consolidation (Self Defence) Amendment Act 1991 (SA). 144 Force, that is to say, which is intended to cause death or grievous bodily harm. Recklessness has no apparent role to play in this context: see Zecevic (1987) 25 A Crim R 165 at 176. 145 Fisse, above n99 at 156-159. 146 See, eg, earlier cases such as McKay above n29, per Smith J who derived his formulation from the Report of the Criminal Code Bill Commission of 1879 C2345, English Parliamentary Papers (1879) Vol36, at 1I. See also: Rainey, per Smith J above n29 at 651. 19931 WOMEN WHO KILL IN SELF DEFENCE 441 sity for the use of deadly force is the criterion. The proportionality of the defendant's response to the threatened harm is no more than a factor for consideration in determining whether the response was reasonably necessary. It is not an independent requirement of the defence. In Zecevic the majority said that evidence of disproportion between threat and response is a matter for the jury 147 Australian common law recognises no determinate rules of proportionality, or anything in the nature of a rule requiring equivalence of threat and response. The jury must determine the issue of reasonable necessity on the circumstances of the particular case. English and Australian common law appear to part company in this respect. In Oatridge,l48 a decision of the English Court of Criminal Appeal, a woman killed the man with whom she lived. Death resulted from a single knife wound. There was ample evidence that he was habitually violent when drunk. On the night of the fatality he seized her by the throat. She gave evidence that she stabbed him because she feared he was about to strangle her. Though her throat was bruised, expert evidence for the prosecution said that she "displayed none of the classical symptoms of a serious attempt at strangulation". Gaynor Oatridge was convicted of murder. The emphasis the trial judge gave to proportionality in his direction to the jury was endorsed in the Court of Criminal Appeal: the defendant's response must be "commensurate with the degree of danger created by the attacY.149 The judgment of the Court suggests that nothing short of a life threatening attack could excuse her use of the knife to defend herself. In the event her conviction was quashed. Though the evidence did not suggest a serious attempt at strangulation, the Court inferred a possibility that she may have exaggerated her peril and believed her life to be in danger. The trial judge should have directed the jury that a mistaken belief, whether or not reasonable, can provide the basis for acquittal.150 That holding merely reinforces the suggestion that nothing short of a serious attempt at strangulation could be commensurate with her use of a knife in self defence. Rules requiring a proportionate response, as Glanville Williams remarked, depend "on the view that there are some insults and hurts that one must suffer rather than use extreme forceW.lsl When retreat and resort to lesser force are 147 Zecevic, above n138 at 167-168, per Mason CJ; at 176, per Wilson, Dawson and Toohey JJ; at 174; at 177, per Brennan J. Cf Deane J, id at 182-184, who emphasised the element of proportionality in determining reasonable necessity. The majority view is discussed and criticised in Yeo, above n29 at 107-112. Compare Lee Anthony Train (1985) 18 A Crim R 353 at 356, per McGarvie J: "Of course, the question whether the act of the accused was reasonably proportionate to the believed danger is merely a particular application of the question whether the act was reasonably necessary having regard to the believed danger." Presumably the correct approach is to make specific reference to the element of proportionality in reasonable necessity when the accused responded to a comparatively minor threat with deadly force in circumstances where retreat or submission might have been more appropriate. R v Walden (1986) 19 A Crim R 444 at 446-447. 148 (1992) 94 Cr App R 367 at 369. 149 Id at 370. 150 English law allows unreasonable as well as reasonable beliefs in the existence of a proportionate threat as a sufficient basis for acquittal on the ground of self defence: Beckford v The Queen [I9881 AC 130; Gladstone Williams (1984) 78 Cr App R 276. 151 Textbook of Criminal Law (2nd edn, 1983) at 504 and, to the same effect, Beale, above 442 SYDNEY LAW REVIEW [VOL 15: 403 equally impossible, the victim of aggression must sometimes submit, rather than use deadly force. In such cases, English law appears to require submission to an unlawful attack unless the attack is very serious indeed. Australian common law is different. Proportionality does not require equivalence between threat and response. Though courts have occasionally appeared unwilling to accept the full breadth of the proposition, it is clear that deadly force in response to a threat to "life or ... safety ... from injury, violation or indecent or insulting usage" may be excusable.152 More recent references indicate that the infliction of acute pain153 or incarceration,154 are also sufficient to excuse recourse to deadly force in appropriate circumstances. The width of these statements of law is qualified in a number of cases, however, by an inarticulate and apparently discretionary limitation. The extended direction on the use of deadly force against a threat of indecent or insulting usage is reserved for special cases. In "ordinary cases" the plea of self defence is said to require evidence that the accused feared the infliction of death or serious bodily harm.155 The circumstances which make the case extraordinary or special have never been specified. Sexual assault will certainly do so, even though the assault amounts to no more than a threat of indecent or insulting usage. Defensive homicide against sexual threats will be excused on the ground of reasonable necessity in appropriate circumstances. This small nugget of comparative certitude derives from recurring cases involving the "guardsman's defence" - the supposed reactions of the outraged Australian male faced with a threat of sexual interference or assault by another male.156 The same rule must also apply if women are threatened with indecent or insulting usage of a sexual nature,l57 though there appear to be no reported instances of self defence on this ground. Use of the extended direction is not confined to cases involving overtly sexual threats.158 The essential element which excuses the use of deadly force is the threatened violation of physical integrity. There are good reasons for concluding that the extended direction is appropriate in most cases of defensive homicide by women. Violence between males of the same age, race and culture is commonly ritualised and governed by rules and mutual understandings. Very often, the difference between a deadly attack and an attack limited to inducing surrender, submission or humiliation is quite clear. Men still match themselves in fights governed by informal rules of fair play to settle their differences, much as they did when chance medley was formally rec- n105 at 581-582. 152 R v Howe (1958) 100CLR 448 at 460; R v T i h s (No 2) [I9631 VR 306 at 312; V m v The i Queen (1978) 141 CLR 88 at 128, 134-135, 146, 180. 153 R v h n e [I9831 2 VR 449 at 450 per Lush J. 154 Above n138 at 188 per Deane J: "serious bodily abuse by way of, for example, sexual abuse or prolonged incarceration". 155 Above n138 at 176, per Wilson, Dawson and Toohey JJ; R v Fricker (1986) 23 A Crim R 147 at 141; R v Honeysett (1987) 34 A Crim R 277 at 282. 156 Howe above 11152. And see, R v Hullett [1%9] SASR 141. 157 R v Wulden above 11147 at 446, per Street J: "[Tlhe danger need not necessarily be the prospect of death or grievous bodily harm. Indeed, so to confine the danger would preclude, for example, the woman who kills in self defence in the face of a sexual aggressor." 158 Above 11153. 19931 WOMEN WHO KILL IN SELF DEFENCE 443 ognised in the common law.159 Similarly, the rituals of masculine violence frequently mark a clear distinction between sexual and non sexual assaults. In the majority of cases involving masculine combat, where the nature and limits of threatened harm were reasonably clear to the combatants, it may be appropriate to direct a jury that defensive homicide is excusable only if the accused reasonably feared serious bodily harm. No comparable rituals or understandings govern violence between men and women however. In stranger attacks on women the risks of harm cover a wide and essentially unpredictable spectrum from deprivation of property to sexual attack and murder. Aggression in intimate relationships is often more patterned and predictable. There is a tendency, however, for domestic violence to escalate in frequency and intensity.160 The potential for escalation and the intimate aggressor's reliance on terror as a weapon of domination can remove any ground for assurance that the next attack will be no worse than the last. An attempt by the woman to measure her response, to use force which falls short of potentially deadly force, may result in an escalation of her partner's aggression and increased danger of serious harm.161 Defensive homicides resulting from a prolonged course of domestic violence can be distinguished in another way. Though there is no explicit recognition of the point in reported cases, the use of domestic terrorism as a weapon of subjection may itself provide a ground for the use of defensive force. Earlier I suggested that terrorism is frequently theatrical, depending on illusion and props. For those who occupy a safe point of vantage, it is all too easy to find the performance laughable. Consider the reactions of Rich J to the tale of terror in Bunyan v Jordan,l62 a civil action for damages for nervous shock brought by the plaintiff against her employer. The defendant was undoubtedly ludicrous. Rich J considered his conduct risible: [Plerhaps a female clerk could not be expected to discover the incongruities of the respondent's behaviour, and to discredit the theatrical threats of a man who produced first poison and then a revolver and after the fullest advertisement of his suicidal purposes retreated with the revolver to the public thoroughfare.163 Hindsight blinded him to the horrifying suddenness with which farce can be transformed to tragedy. Terrorism, whether political or domestic, frequently involves ludicrous behaviour. Often enough the behaviour is deliberately ludicrous - an elaborate charade -consciously intended to inflict the additional terrors of uncertainty and the humiliation of fear induced by absurdity. The deliberate employment of terrorism as a weapon of subjection over months or years is no less reprehensible than isolated threats of "indecent or insulting usage". The effects are far more devastating. In the past, courts have tended to concentrate on the nature of the harm threatened by the aggressor, rather than 159 Horder, above n 1 I. 160 Walker, above n49 at 43-44, Dobash and Dobash above n52 ch 7. 161 Dobash and Dobash, id at 109: the majority of the women surveyed told interviewers that resistance to violence resulted in more severe injuries. 162 [1937] HCA 5; (1936-1937) 57 CLR 1. 163 Id at 15. A majority of the Court agreed with his conclusion that a reasonable person would not have suffered nervous shock as a result of the defendact's behaviour. 444 SYDNEY LAW REVIEW [VOL 15: 403 the fact of terrorism itself. That has led in turn to a concentration on the question whether the threat was credible and serious. The prevailing paradigm is the threat, usually made to a stranger, which presents the threatened victim with a choice between compliance with the aggressor's demand or the infliction of some more or less specific harm. Though provision is made for the agony of the moment, courts tend to measure the alternatives by reference to considerations of proportionality between threat and response. Terrorism is different: its essence is uncertainty. There is no measured or proportionate response. If flight is impossible, the victim may face a choice between complete submission to the unknown and the use of deadly force. The principles of the Australian common law on self defence are clear. Noone is required to submit to unlawfull64 injury or indecent or insulting usage and the same principles apply whether harm is threatened by intimates or strangers. Recourse to defensive force is permitted and, if it is clear that nothing short of deadly force will avert the threatened harm, a jury is entitled to conclude that deadly force is excused. A husband or any other intimate aggressor, possesses no privilege restricting a woman's right to defend herself against violent attack. It is more than a century since Lord Halsbury rejected as "quaint and absurd" the suggestion that a husband has a right to beat his wife.165 The landslide of court decisions rejecting the husband's privileged exemption from liability for rape reinforces the conclusion, if reinforcement is necessary, that there is no privilege for the use of force or violence without consent.166 Brennan J stated the law bluntly and forcefully in R v L:l67 marriage "casts upon a husband an obligation to respect a wife's personal integrity and dignity; it does not give the husband a power to violate her personal integrity and destroy her dignity". The fact that the aggressor is intimately known to her can make a difference of course. Knowledge of his conduct in the past can provide a measure of the force necessary for self defence on this occasion. Many couples tolerate a degree of violence, restricted by tacit agreement, in intimate relationships. The physical expression of hostility between them is bounded by unspoken rules and conventions which set limits to their conflict. The test of such a relationship, and they are not uncommon, is whether the physically weaker partner can resort to moderate and limited force to terminate provocation or avoid threatened harm, without risking injury, rape or another disproportionately violent response. If the weaker partner could be confident that moderate defensive force would end the aggression, resort to deadly force is inexcusable. In exceptional cases, however, when relationships are characterised by repetitive cycles of extreme violence or vio- 164 In certain circumstances, involving complexities irrelevant to the the present argument, self defence against "lawful" injury may be excusable: above 11144at 157. 165 R v Jackron [I8911 1 QB 671 at 679. Esher, who went on to add t b t a hurbund has no right to imprison his wife either, refused to believe that a right to violence had ever existed. 166 R v L [1991] HCA 48; (1991) 66 ALJR 36 (H Ct); R v The Queen [I9911 3 WLR 767 (HL); A v M [I9911 3 NZLR 228 at 245-251. In the last of these cases, a civil action for assault against a husband who raped his wife resulted in an award of exemplary damages of $20,000.Spousal immunity extended only to acts of intercourse by the husband. He could be convicted as a principal offender committing the offence through an innocent agent or as an accomplice to rape by another: Re Cogcm & Re Led [I9761 QB 217; R v L id at 43 per Brennan J. 167 R v L i d a t 4 2 . 19931 WOMEN WHO KILL IN SELF DEFENCE 445 lence accompanied by terrorism there is no ground for confidence in moderate measures. Their shared past provides no reassurance that this time will be no worse than the last. Nothing short of force likely to kill or do serious injury will avert the threatened harm. Moderate force may only enrage him, resulting in worse injury. Evidence of previous attacks is admissible to support a claim that deadly force was reasonably necessary in the circumstances to avert continuation of terror, or the infliction of further harm.168 B. Apprehended Threats and Reasonable Responses In Zecevic v DPP169 Deane J emphasised the necessity for estimates of the seriousness of the threat to be made from the defendant's point of view. The fear of impending attack must be reasonable. Reasonableness is not to be judged, however, "in the completely objective sense of what a reasonable man would have believed".l70 It is a qualified objective test, measuring the magnitude of the threat by asking "what the accused himself might reasonably believe in all the circumstances in which he found himself'.l71 The formulation favoured by the majority is consistent with that view: the essential question for the jury "is whether the accused believed on reasonable grounds that it was necessary in self defence to do what he diC.172 Reasonableness, whether relating to perceptions of threatened harm or estimates of the response necessary to avert the harm, is relative to the circumstances and situation in which the threat was perceived. Australian case law presents few clear examples of the difference which might be made by formulating tests of danger or the necessity of deadly force by reference to the agent's perspective. There is one startling example of failure, in the trial direction in R v Collingburn,l73 to follow this injunction. The accused pleaded self defence to a charge of murdering her de facto husband. The jury was directed to measure the actions of the accused by reference to the behaviour of a man facing an attack. The trial judge apparently sought some standard which transcended possible differences between the ways in which men and women might respond to domestic violence. He told the jury, "[wlhere I use the word 'man', I do so because it is easier than saying man or woman or person or he or she and doubling up all the time". A direction in this form is calculated to obscure the agent relative character of the objective test in self defence. The direction was held to be objectionable on other grounds and Collingburn's conviction for manslaughter was quashed on appeal. The New South Wales Court of Criminal Appeal was more faithful to the binding authority of High Court decisions in R v Walden,174 in which the accused shot and killed a neighbouring farmer with whom she had been in- 168 See, eg, R v Howe [I9581 SASR 95; R v Beech (1978) 20 SASR 410; Re Knowles [I9841 VR 751. 169 Above 11138. 170 Id at 182 per Deane J. 171 Ibid. The agent relative orientation of the objective test derives from the formulation in Viro (1978) 141 CLR 88 at 146-147,per Mason J. 172 Id at 174, per Wilson, Dawson and Toohey JJ. 173 (1985) 18 A Crim R 294. 174 (1986) 19 A Crim R 444. 446 SYDNEY LAW REVIEW [VOL 15: 403 volved in a hostile dispute. Mary Walden managed a remote sheep station in outback New South Wales, where she lived with her aged mother. Her neighbour wished to run a telephone line over her land. She refused permission. Some days before the shooting he came to her house. There was a hostile confrontation during which he threatened her with an iron bar. On the day of the shooting she discovered him, with one of his employees, cutting her lines. She remonstrated without effect. He drove a short distance and cut the line again. Mary Walden followed him and alighted from her own vehicle. She was carrying a loaded rifle. The man then walked towards her, calling to his companion as he did so. She killed him with a single shot, fired from a distance of about twelve feet. She was tried for murder and convicted of manslaughter, her plea of self defence being rejected by the jury. The Court of Criminal Appeal quashed her conviction on the ground that it was a misdirection to tell the jury apprehension of grievous bodily harm was necessary, if the use of deadly force in self defence was to be excused. The case is important for another reason however. The nature of the threat and the nature of the action required to avert it cannot be considered apart from the sex and physique of the accused or the circumstances in which she found herself. In the course of its judgment, the Court emphasised her vulnerability as solitary woman of average physique faced with a very large and aggressive man who had recently threatened her with violence. Her physical vulnerability made it more likely that she would be attacked. The overwhelming physical advantages possessed by the deceased left her with few alternatives to the use of the rifle if she was to protect herself. In Walden, the New South Wales Court of Criminal Appeal echoed the more explicit reliance on the principle of agent-relativity in the celebrated American decision in State of Washington v Wanrow.l75 Yvonne Wanrow shot and killed a suspected prowler and child molester after an altercation during which he refused to leave the house where she was staying with another woman and her child. Her conviction for murder was overturned on appeal by the Supreme Court of Washington: The impression created - that a 5'4" woman with a cast on her leg and using a crutch must, under the law, somehow repel an assault by a 6'2" intoxicated man without employing weapons in her defence, unless the jury finds her determination of the degree of danger to be objectively reasonable constitutes a separate and distinct misstatement of the law and ... violates the respondent's right to equal protection of the law ... [She] was entitled to have the jury consider her actions in the light of her own perceptions of the situation, including those perceptions which were the product of our nation's "long and unfortunate history of sex discrimination". In cases like Walden and Wanrow, it is the disparity in physical strength between the aggressor and his victim which provides the ground for excusing the use of deadly force in self defence. The weak or vulnerable are entitled, on occasion, to secure themselves from danger by pre-emptive and unexpected counter attack. There is no privilege for bullies, requiring the weak or vulnerable to incur unreasonable additional perils by warning more powerful aggressors that an unexpected and deadly response is possible. The same principle 175 (1977) 559 Pacific Reporter 2d 548 at 558. The case is often cited to illustrate the potential, but umalised, flexibility of North American law on self defence. 19931 WOMEN WHO KILL IN SELF DEFENCE 447 must be equally applicable when deadly force is used in response to domestic violence.176 Nothing in the association between spouses, lovers or companions entitles the aggressive and violent partner to exemption from the principles of criminal law which apply between strangers. In the past he may have abused her and battered her with impunity. She may have submitted to his violence. Habitual violence does not confer an immunity against counterattack however. He cannot claim some species of prescriptive right to continue his assaults. Nor has he any right to assume that she will submit on this occasion. Whatever the law may have been in the past, marriage or other intimate relationships no longer provide enclaves beyond the reach of criminal laws against violent assault. Australian common law requires evidence of a real threat or reasonable belief that harm is threatened. In this respect, English law is more generous to an accused who relies on self defence. Recent decisions allow the jury to acquit though the accused made a wholly unreasonable mistake, perceiving a non existent threat.177 In South Australia, recent legislation is intended to introduce the same rule.178 Elsewhere, however, Australian courts have not followed the English lead.179 There is much to be said in favour of the view that an honest belief that one is under attack - whether or not that belief is reasonable should be adequate for the purposes of self defence. If a person responds to a threat with force causing death, there can be no conviction for murder unless it is proved that the response was intended to cause death or grievous bodily harm or that those consequences were known to be likely. If the accused failed to realise the risk that the response might cause death or serious injury, it makes no difference whether that failure was reasonable or unreasonable. The prosecution would fail for want of proof of intention or of recklessness with respect to death or serious injury. Yet self defence requires reasonable apprehension of impending harm. The justification for the distinction between unreasonable mistakes about the effects of force used in self defence and unreasonable mistakes about the nature of the threat which prompts the use of defensive force is far from obvious. Glanville Williams argues that consistency requires acquittal of murder in both cases.180 In his view, there is no 176 Two of the conclusions from an analysis of 2% cases of spouse killing in New South Wales above nl are pertinent: (1) (at 97) "Violence or fear of future violence was both the background and the cause of force by women on their husbands. They killed their husbands after they or another family member had been attacked." (2) (at 94) Fatal retaliation by women almost always involved the use of guns or knives: "frequently a woman's use of a weapon is the only way she can protect herself from an attack by her husband. The use of a weapon can be the only means by which she can tilt the physical imbalance between herself and her husband, and thus the only way she can prevent herself from being injured. One consequence of the use of a weapon, however, is that it makes lethal violence more likely ... " 177 Gladstone Wi1lium.v above n150; Beckford v The Queen above 11150; Guynor Outridge (1992) 94 Crim App R 367. New Zealand law is to the same effect. See R v Ranger (1988) 4 CRNZ 6. 178 Criminal Law Consolidution (Self Defence) Amendment Act 1991 (SA). 179 Australian authorities are collected and discussed in Fisse, above n99 at 513-514. 180 Williams above n151 at 138: arguing that no distinction is to be drawn between, "A who shoots his wife, believing her to be a rabbit and ... B who shoots his wife thinking she is a burglar about to attack". The risk of injustice is apparent in cases such as R v Rose (1884) 15 Cox CC 540 (Jury instructed to convict a son of murdering his father if satisfied beyond 448 SYDNEY LAW REVIEW [VOL 15:403 sensible distinction to be drawn between unreasonable mistakes about threats and unreasonable mistakes which go to the mental element of the offence.181 If legal definitions affect outcomes, it should follow that pleas of self defence against ambiguous threats will succeed more often in England or South Australia. The effects of the South Australian reform have yet to be tested. Two factors diminish the significance of differences between English and Australian common law in their approach to unreasonable beliefs about threats. English law has a strong requirement of proportionality between the threat and retaliation. The indulgence extended by English courts to unreasonable beliefs palliates, to some extent, the restrictive effect of the requirement of proportionality. Gaynor Oatridge had her conviction quashed though her response was disproportionate by English standards. It was sufficient that she might have believed, even unreasonably, that she was threatened with death by strangulation. There is less need for a subjective test in Australian common law jurisdictions, which do not require a relationship of near equivalence between the attack and the defensive response. The significance of the difference is further diminished by the English refusal to adopt the subjective stance when the defendant's judgment was affected by voluntary intoxication. English courts hold that self defence is not open if the accused was unreasonably mistaken about the threat as a consequence of intoxication.182 Intoxication is a common precursor to the use of deadly force in domestic killings.183 Australian common law allows a more sympathetic and more realistic approach to intoxication when self defence is in issue. There is an overwhelming body of evidence showing that alcohol exacerbates domestic violence by men.184 His drunkenness is a circumstance to be taken into account when she assesses her danger and the necessity for defensive force. More to the point, her intoxication must be taken into account if it is a factor likely to inflame his aggression, exacerbate his violence or reduce her capacity to take evasive action. When the intoxication of either of them increases her vulnerability to attack, increased force in self defence is excusable on the grounds of reasonable necessity. That conclusion follows as an unavoidable consequence of the criteria for self defence in Viro and Zecevic.185 reasonable doubt that the son lacked reasonable grounds for his belief that his father was about to kill his mother). For a defence of the distinction, see Simester, A, "Mistakes in Defence" (1992) 12 Oxford JLS 295. 181 Arguments for consistency cut both ways of course. At the conclusion of her paper on the issue, Giles, M, "Self Defence and Mistake: A Way Forward" (1990) Modem LR 187 at 199, suggests that adherence to objective tests for the defences might presage attenuation of the subjective criteria determining the mental element of the offence. 182 Patrick Gerald O'Grady (1987) 85 Cr App R 315; R v O'Connor [I9911 Crim LR 135. 183 Above nl at 93: nearly half of the wives who killed husbands or intimate partners had been drinking with them at the time of the final fatal confrontation. 184 In addition above nl at 93-94, see Russell, above n6 1 at 122,156-166. Compare, however, Dobash and Dobash, above n52 at 118. 185 There are close similarities, in this respect, between the objective test in self defence and the objective test in provocation, which allows any personal or idiosyncratic factor affecting the gravity of provocation to be taken into account. See Stingel v The Queen (1990) 171 CLR 171; discussed Leader-Elliott and Goode, "Criminal Law", in Baxt, R, Moore, A and Kerlogue, P, Annual Survey of Australiw Law (1991) 263 at 269-276. 19931 WOMEN WHO KILL IN SELF DEFENCE 449 C. Failure to Retreat Polemical literature in the United States justifies the introduction of expert testimony on the battered woman syndrome as an explanation of her failure to remove herself from danger. The concept of learned helplessness is meant to compel acceptance of the conclusion that victims of domestic violence do not leave because they are psychologically disabled and incapable of leaving. A woman does not lose her right to defend herself if she chooses to remain with her aggressive partner. The question whether deadly force was reasonably necessary does require consideration of the possibility that retreat might have been reasonable in the circumstances. That possibility should not be confused, however, with the different and quite insupportable assumption that a woman owes to her aggressive partner a duty to terminate their relationship so as to avoid any occasion for self defence. Women who kill in response to male violence are almost always attacked in their own home by husbands or other intimates. The bedroom is the most usual location, or the loungeroom.186 Does the law require a woman attacked in her home to retreat rather than use deadly force? Does the answer depend on the further question whether the aggressor is her husband, lover or companion? These questions have been the subject of vigorous debate in the United States. A minority of American States are said to impose an obligation, by way of an explicit rule, requiring a person threatened with harm to retreat rather than use deadly force in self defence.187 In these jurisdictions an exception is usually made if the accused was attacked at home.188 An even smaller minority of States require retreat, however, if the attack is made by another occupier. In State o New Jersey v Ponlery,189 the Supreme Court of New Jerf sey held that a wife was bound to retreat rather than use deadly force when attacked by her husband in their summer home. More extraordinary still was the case of Mrs Grierson who killed Charlie Peabody, her sometime companion and lover, when he attacked her with a knife.190 Peabody was a frequent guest at Mrs Grierson's house and often stayed over. He was a violent man who had beaten her on previous occasions, threatened to kill her and smashed her furniture. On the evening of his death, they had been out together drinking and returned to her house where they fought before she killed him. The Supreme Court of New Hampshire upheld her conviction for unlawful homicide. The trial judge's refusal to instruct the jury that Leah Grierson could stand her ground in her own home was correct, said the Court. It was Peabody's home too: "a man's house is the dwelling place of ... one who is residing [there], however temporarily, as a guest9'.191She should have retreated if she could, rather than respond to his attack with deadly force. This is an example of uncommon buffoonery however. In her survey of 223 appellate decisions involving homicide charges against women who argued that they killed in self -- 186 Above nl at 95; Law Reform Commission of Victoria, above n2 at 39. 187 LaFave, W and Scott, A, Substantive Criminal Law (1986) at 659. 188 The rule and the exception for the home are stated in the Model Penal Code: Proposed Official Draft ALI 1962,Section 3.04. 189 (1955) 117A2d473. 190 State v Grierson (1949) 69 A 2d 851. 191 Id at 854. 450 SYDNEY LAW REVIEW [VOL15: 403 defence, Maguigan concluded that failure to retreat was rarely determinative.192 Though decisions like Grierson have encouraged recourse to expert evidence of the battered woman syndrome, the preferable course is to reform the law of self defence in those few jurisdictions which impose unreasonable requirements of retreat.193 Australian common law does not require retreat in the face of an unlawful attack. Failure to retreat, in cases where retreat was possible, is a factor to be considered by the jury when they determine whether the actions of the accused were reasonably necessary.194 I referred earlier to the risk that bland invocations of reasonable necessity can obscure the reason why retreat might not amount to a reasonable course of action for a woman threatened with domestic violence. In general, there are two reasons for requiring retreat before resort to deadly force. The first has to do with symbolism. Retreat cornmunicates unwillingness to continue the fight. Australian case law is not explicit on the point though it possesses the authority of antiquity and forms an element of modern English case law. In R v Julienl95 the Court of Appeal said that a man who is threatened by another, "must demonstrate by his actions that he does not want to fight". He must, "demonstrate that he is prepared to temporise and disengage and perhaps to make some physical withdrawal". Communicative behaviour, signalling unwillingness to continue hostilities, minimises the risk that action taken in self defence will merely escalate the conflict.l% It follows that a demonstration of unwillingness to continue the fight is unnecessary if the other person is an aggressor who knows that the victim will not go beyond the necessities of self defence. References to the symbolic role of retreat are peculiarly appropriate when there is a tacit or express agreement to settle a dispute by violence. In that case, neither party is entitled to use force in self defence in the absence of disengagement and communication of unwillingness to continue.197 References to the symbolic role of retreat are peculiarly inappropriate, however, when domestic violence is in issue. Policies designed to outlaw mutual combat between contending males198 have no application in the overwhelming majority of domestic conflicts, where there is no mutuality and no agreement to engage in combat. The brutality is onesided and the aggressor needs no symbolic communication of his victim's unwillingness to continue the combat. There is a second and more obvious reason for retreat. Unless there is good reason to the contrary, commonsense and common prudence suggest that one ought to avoid encounters which may require the use of force in self defence. 192 Maguigan, above n42 at 419,436-437. 193 Id at 450-451. 194 Fisse, above n99 at 157. New Zealand law is to the same effect. See R v Savage [I9911 3 NZLR 155. 195 [1%9] 1WLR 839 at 843. 1% Williams, above n151 at 505. Beale accepts the communicativerationale for a requirement of retreat and adds a second consideration. A person who demonstrates unwillingness to continue hostilities shifts responsibility for the subsequent use of defensive violence to the other combatant: See Beale, above n105 at 574-575and at 537-539,respectively. 197 Beale, above n105 at 536-538. 198 Above n l l . Implementation of that policy is perfectly consistent with the provision of qualified defences which reduce the liability of the survivor from murder to manslaughter. 19931 WOMEN WHO KILL IN SELF DEFENCE 45 1 The corollary of this obvious principle is more problematic. Sometimes there is good reason not to avoid or terminate the encounter, though it may lead to the necessity for action in self defence. So, for example, it has always been accepted that robbery may be resisted with whatever degree of force is necessary.199 A person threatened with unlawful dispossession is entitled to refuse to yield and if that refusal leads to escalation of the conflict, to use whatever force may be necessary for self protection against a persistent aggressor. Similar principles appear to govern the right to resist an aggressor in the home. Ownership of the premises is not a prerequisite and English courts have taken a generous view of what amounts to home when self defence and retreat are in question.200 English case law on self defence suggests that a person attacked at home is entitled to stand their ground. It is possible that this is one of the few explicit rules of self defence which have survived. It is also possible, however, that there is no longer any hard and fast rule about retreat and attacks in the home. The question may be governed, like so much else in self defence, by considerations of reasonable necessity.201 Australian authority is even more exiguous. In the few modern cases where Australian courts have considered self defence in the home, there is no suggestion of a hard and fast rule. The old rule was dissolved in the criterion of reasonable necessity. The Victorian Court of Criminal Appeal decision in R v Lane202 emphasised the importance of the fact that the defendant was attacked in his home. Though the case involves a casual homosexual encounter which ended in fatal violence, the situation bears some resemblance to cases of fatal defensive force against threats of domestic violence. There is sufficient similarity, at least, to provide an indication of the way in which the law of self defence would be applied when an abusive domestic relationship culminates in the use of deadly force in self defence. Michael Lane met Leon Parmac in a bar, where they spent the afternoon drinking. At the end of the day they went to Lane's home. Both were drunk. Parmac had taken a substantial number of Mogadon tablets as well. Lane gave varying accounts of Parmac's bizarre behaviour. He threw bottles around the house and smashed furniture and electrical appliances. He apologised. Then he began again, smashing furniture. There was ample evidence, from the devastation of Lane's possessions and from the testimony of neigh- 199 Coke, E, Institutes (1797) at 55. The proposition that no-one is bound to yield to violent threats of dispossession is not affected by growing recognition that it may be unjustifiable to use deadly force, even as a last resort, in order to prevent the escape of a fleeing offender. 200 Israel James Hussey (1924) 18 Cr App R 160. The historical validity of the decision was doubted in Turner, J W C (ed), Russell on Crime (12th edn, 1964) Vol 1 at 442. 201 Williams, above n151 at 519 suggests that the decision in Hussey, ibid, though technically binding on an English court, is obsolete. Williams is primarily concerned, however, with cases in which serious injury is inflicted on an intruder who is known to be acting pursuant to a claim of right. Compare the Model Penal Code provisions forbidding deadly force if the necessity for doing so can be avoided by surrendering personal property to a person who asserts a claim of right: s3.04(b)(ii).The rule allowing defence without retreat may survive when the attack is not pursuant to a claim of right. 202 [I9831 2 VR 449. Compare the converse situation of R v Fricker (1986) 42 SASR 436, where the defendant stabbed his victim in the victim's home in the course of a violent confrontation. He was convicted of murder and the SA Court of Criminal Appeal upheld his conviction. His failure to retreat from his victim's home damned his attempt to rely on self defence. 452 SYDNEY LAW REVIEW [VOL 15: 403 bours, who made no effort to intervene, of a prolonged and violent attack on Lane's possessions. There was little evidence of personal violence however. Lane suffered no injuries, there were no blows struck and no verbal threats. In an unsworn statement at his trial, Lane said that Parmac made one attempt to hit him in the kitchen but slipped and fell. He staggered outside. Lane followed him, carrying a bottle of champagne. When Parmac seemed to threaten him again, Lane hit him with the champagne bottle. He did not stop after the first immobilising blow, but continued to batter Parmac's body with a second bottle. Parmac suffered multiple injuries, some of them inflicted after death. Though it is an extreme case, near the margins of self defence, a majority of the Court held that it was open to a jury to acquit on this ground.203 There was evidence that Lane was terrorised by his companion, who was a younger and stronger man. Should he have called for help? Lane's failure to do so was understandable: "There were problems associated with public disclosure of the presence of the naked deceased in his raving condition and in the applicant's own home." 204 Terror and despair could have driven him to use deadly force to avoid injury. Nor was he bound to retreat, though retreat was possible. A person in their own home, who is "placed in constant danger of serious bodily harm", is entitled to take pre-emptive action, anticipating renewal of an attack by the intruder.205 Self defence must have the same amplitude of application when deadly force is used in response to an attack by a husband or other intimate. The violence in Lane was far less than the violence commonly used against battered women.206 The fact that Lane was threatened by a comparative stranger invited into his home can make no difference to the principle. Unless we are prepared to say that the law provides a licence for violence when a relationship is more than merely casual, the length of an association, or its degree of intimacy, cannot abridge the right of self defence. Nor is a woman bound to retreat from her home when threatened. The scant body of law drawn from the few decisions on self defence, retreat and the home finds its most familiar expression in propositions about men and their castles: he is entitled to "defend his castle against attack without retreating from it, since that would He be to give up the protection of a 'castle' which the law allows himW.207 is "under no duty to take to the fields and the highways, a fugitive from his own home7'.208Michael Lane was the beneficiary of this tradition. It was open to a jury to acquit him on the ground of self defence, though his statements and behaviour suggested that he was primarily concerned for the safety of his house: "I couldn't work him out he kept saying sorry and then smashing more things -well I wasn't going to be the next victim. Bad enough having all my 203 Starke J dissented on the ground that the evidence, together with Lane's unswom statement, failed to provide a credible narrative of sufficient factual particularity to enable rational consideration of the possibility that he killed in self defence. 204 R v Lane, above n153 at 455, per Murphy J. 205 Id at 456. provide descriptive and statistical ac206 Dobash and Dobash, above 1152at 107-108,248-249 counts of the nature and severity of domestic attacks in their sample. See also, Russell, above n61 at 87-101. 207 Beale, above 11105at 574-575. 208 People v Tomlins (1914)213 NY 240, per Cardozo J. 19931 WOMEN WHO KILL IN SELF DEFENCE 453 furniture smashed to pieces9'.209He was defending his castle from an invader. Far more is at stake when a woman defends her place in her home from the enemy within its walls. She is no less entitled to the protection of the law and no less entitled to have her choice to stand her ground respected. D. The Sleeping Tyrant Anticipatory action in self defence is permissible on occasion. A "man about to be attacked does not have to wait for his assailant to strike the first blow or fire the first shot; circumstances may justify a pre-emptive strikeW.210 Existing law possesses some flexibility, sometimes expressed in the proposition that the threats which prompt defensive action can be of "imminent", rather than "immediate" harm. Imminent harm, in this sense, could occur at any time. Though this particular formulation of the proposition lacks authoritative support,211 it is apparent that the right to use deadly force in self defence is not limited to threats of immediate harm.212 The extent of flexibility is uncertain. In what circumstances is a woman's pre-emptive strike against a domestic aggressor excused? Though they appear to constitute no more than a small minority of domestic homicides in self defence,213 cases of women who kill a sleeping husband, lover or companion in order to forestall further attack have attracted much attention and inspired sharp differences of judicial opinion over the borderline between permissible self defence and inexcusable private vengeance. Some courts have condemned the claim that homicide was excused or justified in the circumstances as a "leap into the abyss of anarchy",214 an abandonment of civilized order215 or a "return to the law of the jungleW.2l6 language of judges who have taken this view evokes the ghost The of petit treason, those long since abandoned varieties of homicide committed when a wife killed her husband, a priest his prelate or a servant his master.217 209 Lane above n153 at 453. 210 Becword v The Queen [I9881AC 130at 144.Discussed in Williams, above 11151a 503-504. t 21 1 On North American law, see Maguigan, above 1142at 414,462ff. These alternative formulations were discussed, inconclusively, by Ormiston J in R v Collingburn (1985) 18 A Crim R 294 at 302-303 (Vic CCA). Compare the dictum in R (1981) 4 A Crim R 127 at 131 (SA CCA): threatened harm must be "immediate". New Zealand case law is similarly inconclusive. See R v Wang [I99012 NZLR 529 at 534-540 (NZ Ct of App) and the cases discussed therein. Though New Zealand law on self defence is codified, the statutory criterion in s48 Crimes Act 1961 is indistinguishable, on the imminence/immediacypoint, from Australian common law. 212 The Australian authorities are discussed by Yeo, above 1129at 80-84. For a survey of US law, see Rosen, above 11105. 213 Maguigan, above n42 at 388-401: empirical research suggests that "sleeping man" cases are far outnumbered by confrontation killings. Compare the unsupported assertion in Schneider, above 1118 at 621 at 634: "Typically, the man beats the woman, sometimes threatening to kill her, until he passes out or falls asleep. Fearing that when he awakes he will beat her more severely or act on his threat, she attacks him while he sleeps." 214 State v Stewart (1988)763 P2d 572 at 579, citing Jahnke v State (1984)682 P2d 991, a 997. t Discussed in Willoughby, M J, "Rendering Each Woman Her Due: Can a Battered Woman Claim Self Defence When She Kills Her Sleeping Batterer?'(l989) 38 Kansas LR 169. 215 R (1981)4 A Crim R 127 at 131. 216 R v Wang [I99012 NZLR 529 at 534-535,539:Trial ruling that self defence was not open on the evidence endorsed by the Court of Appeal. 217 Stephen, J, History of the Criminal Law of England Vol 111 (Burt Franklin edn, 1883) at 35. Petit treason was a more serious degree of murder, originally punished with more hor- . 454 SYDNEY LAW REVIEW [VOL 15: 403 Sometimes, as an alternative to the charge of anarchy, the woman is accused of usurping lawful authority, of seeking to act as judge and jury in her own cause.218 In killing the sleeping tyrant, she deprives him of the chance to be heard in his own defence: there is an evident tendency in some American courts to conceive her plea of self defence as an unfair attack on his reputation.219 Consistency with the governing criterion of reasonable necessity should mean that the issue of imminence is a matter of judgment for the jury on the particular facts of the case.220 There are contrary indications, however, that the temporal relationship between threat and response may be governed by more particular rules which define the "occasion" for self defence.221 So, for example, in R222 the South Australian Court of Criminal Appeal suggested that the trial judge should not allow the jury to consider self defence unless there was a threat of present harm. Subsequent cases have been less dogmatic. In Kontinnen,223 there was evidence that the deceased went to sleep with the threat that he would kill the defendant, and others in the household, when he woke in the morning. He was shot in the head as he slept. Self defence was allowed to go to the jury with the usual instruction that they were to acquit unless the prosecution proved that the actions of the accused were not reasonably necessary in the circumstances.224 Though the decision of the New Zealand Court of Criminal Appeal in Wang225 goes the other way, the Court did not reject the possibility that deadly force might be excused though the deceased was asleep at the time. Self defence was withheld in Wang on the ground that no jury could have considered the actions of the accused were reasonably necessary in the circumstances.226 rid methods of inflicting the penalty of death. The offence was abolished by statute in 1828,9Geo IV C31, s2. 218 See, eg, State v Stewart (1988)763 P2d 572 at 592, where the accusations are made in tandem; Fletcher, above n95 at 21: The injured wife is not permitted to "function as judge and jury"; Acker and Toch, above 11104at 151: "Defined as a bad person ... the battering husband [is] perceived as having finally brought about his own homicide. Thus the jury functions as an arbiter of just deserts and not as a finder of facts concerning the events and perceptions which motivated the battered woman to call upon deadly force on the occasion in question." 219 Discussed in Baumann, M, "Expert Testimony on the Battered Wife Syndrome" (1983) 27 St Louis W 407 at 434. In the South Australian case of R v Runjanjic above n27 at 372373, Bollen J rejected any suggestion that testimony supporting a plea of self defence ought to be limited by concern for the reputation of the deceased. 220 Williams, above n151 at 503-504. 221 Much of the remaining technicality of the law on self defence, swept away by the High Court decision in Zecevic v DPP above n 138, was a product of the remarks by Lowe I, in McKay [I9571 VR 560 at 563, on the "occasion [which] wmants action in self defence or for the prevention of felony ...". 222 Dictum in the South Australian decision in R [I9811 4 A Crim R 127 at I31 supports the view that the threat must be immediate: anticipatory self defence is not available because "society" cannot "countenance killing as a means of averting some apprehended harm in the future". 223 Kontinnen above 1122. Trial direction in the South Australian Supreme Court per Legoe J. 224 Eriica Kontinnnen was acquitted of murder and manslaughter, The Advertiser 31 March 1992. 225 Above 11216. 226 Id at 535: "We are satisfied that no ordinary reasonable person who knew that kind of man that the husband was and of his threats to his wife and sister and blackmail of her family, would, while he was unarmed and in a drunken sleep, have believed it necessary to kill 19931 WOMEN WHO KILL IN SELF DEFENCE 455 Other cases, less dramatic than those involving a sleeping victim, support the suggestion that reasonable necessity may excuse the use of deadly force to avert threats of future harm. In Zecevic227 the High Court accepted the possibility of acquittal in a case where the defendant made a pre-emptive strike and killed in the anticipation of a renewed attack by his victim. The decision of the Victorian Court of Criminal Appeal in Lane,228 allowed the possibility of a pre-emptive strike by a householder against an assailant in the home. The right to defend oneself is not restricted to circumstances in which "an actual physical onslaught on the accused was ensuing or immediately threatened.229 The cases on duress provide additional support for this exiguous line of authority. Though duress will excuse only if threats were "present and continuing, imminent and impending9',230 there is no necessity for threats of immediate harm.231 Threats of harm to be inflicted at some indefinite time in the future can excuse present compliance with demands for criminal involvement. Cases involving duress commonly involve relationships of continuing terror between oppressors and their victims.232 The threat tracks the victim. In this sense, duress resembles cases of domestic violence involving continuing terrorism, where the prospect of violence is always pending.233 The conceptual link between duress and self defence is closer than it might at first appear. Though the defences are distinguished in other ways, the rules relating to the imminence of threats must be similar. Suppose a case where duress takes the form of a demand for some criminal act, coupled with a conditional threat to kill or inflict serious physical harm at some indefinite time in the future. The threat is "present and continuing, imminent and impending". It does not induce compliance with the demand however. Instead, the threatened individual turns on the threatener in an unguarded moment and uses force sufficient to disable the threat. In principle, self defence must be available though there was no threat of immediate harm. Encouragement is due to individuals with the courage to resist evil. The law relating to pre-emptive strikes in self dehim". Apart from the estimate of the husband's character, the Court of Appeal appears to have been influenced in its judgment by the fact that the case involved a dispute over money. The husband's threats of physical harm were made for an acquisitive purpose, presumably envisaged further negotiations and lacked the element of urgency characteristic of disputes over matters of more primal concern. 227 In Zecevic above 1-1138at 165, overruling decisions by the trial judge and Victorian Court of Criminal Appeal which witheld self defence from the jury on the ground that there was no reasonable ground for expecting immediate danger from the deceased: Zecevic [I9861 VR 797. 228 Above n153. 229 Id at 456. 230 R v Hurley [1%7]VR 526 at 543, per Smith J. 231 Yeo, above n29 at 84-87 provides a useful discussion of Australian and English case law on the point. 232 It is for this reason that the defence of duress is denied if the person who succumbs to coercion chose to associate with criminal associates who might be expected to compel participation in crime: see Yeo, id at 173ff. 233 In Goddard v Osborne (1978) 18 SASR 481 the appellant's convictions for offences relating to false claims to unemployment benefits were quashed on the ground that her conduct was compelled by her husband's threats of violence. He was absent when she committed the offence and it would have been possible for her seek help without risking immediate harm. The Court of Criminal Appeal recognised the "notorious" fact that an application to the police to protect her from her husband would be ineffective. 456 SYDNEY LAW REVIEW [VOL 15: 403 fence cannot be less generous in its definition of continuing threats than the law of duress. Very little guidance is available on the extent to which pre-emptive strikes can be excused on the grounds of self defence. The preferable view, which has the sanction of the High Court in Zecevic, is that degrees of imminence of threatened harm go to the question of reasonable necessity.234 There is no separate rule which enables the trial judge to withhold the issue from the jury because the threat was not immediate. Cases involving delayed retaliation against aggression or a fatal attack on a sleeping tyrant quite evidently engage the emotions of judges and juries on the deepest and most primitive level. That is all the more reason to ensure that the issue is not avoided or withheld from the jury. E. The Partial Defence of Provocation If the jury is convinced that deadly force was unnecessary, the accused may nevertheless escape conviction for murder on the ground that the fatal act of retaliation was provoked. Abolition of the partial defence of excessive defence by the High Court in Zecevic was accompanied, in the majority judgment, by an assurance that provocation would serve in lieu.235 More recently, the Court has taken the view that provocation is almost always an issue for the jury when the evidence suggests that oppressive or insulting behaviour could have led to a loss of self control and fatal retaliation.236 The course of decisions in the High Court suggests that provocation will generally provide a fallback option, if it is proved that deadly force was neither reasonable nor necessary as a response to domestic violence.237 There is danger of encouraging juries to compromise, of course, in such fallback positions. Provocation, like excessive defence, can operate to deprive an accused of a meritorious plea of self defence.238 234 The argument is also presented in Rosen, above n105. 235 Above n138 at 176. 236 Kolalich v DPP [1991] HCA 47; (1991) 66 ALJR 25 at 27-28. Stingel v The Queen [1990] HCA 61; (1990) 171 CLR 312 and the New Zealand case of R v Erutoe (1990) 5 CRNZ 538; 2 NZLR 28; provide recent examples of cases in which provocation was witheld on the ground that no ordinary person could have been driven to kill by the provocative conduct of the deceased. Both cases involved men who killed in the course of savage attacks prompted by sexual jealousy. 237 Recent cases suggest that self defence and provocation do tend to run in tandem when deadly force is used in response to domestic violence: R v Hi11 (1981) 3 A Crim R 397; R (1981) 4 A Crim R 127; R v Collingburn (1985) 18 A Crim R 294; R v Pita (1989) 4 CRNZ 660, Gaynor Oatridge (1992) 94 Cr App R 372; Kontinnen, above n22. Evidence of domestic violence which is insufficient for the purposes of self defence can nevertheless provide the evidentialbasis for a partial defence of provocation: R v Wang [I9901 2 NZLR 529. 238 Tolmie, J, "Provocation or Self-Defence for Battered Women Who Kill?" in Yeo, S (ed), Parfial Excuses to Murder (1991) at 61 warns against the danger. The decision of the New South Wales Court of Criminal Appeal in Hill (1981) 3 A Crim R 397 exemplifies the problem. The defendant killed her partner after years of violence and abuse. At trial, the jury rejected the defences of self defence and provocation and convicted her of murder. Though the jury had not been misdirected, the Court quashed her conviction on the ground that it was unsafe and substituted a verdict of manslaughter. No justification for the compromise is apparent. If the verdict of the jury was wrong, there is no reason to assume that provocation, rather than self defence, was the appropriate ground for an acquittal of murder. 19931 WOMEN WHO KILL IN SELF DEFENCE 457 Traditional formulations of provocation restrict the application of the defence to cases of overwhelming anger.239 More recent authorities recognise fear, no less than anger, as the spur to excusable loss of self control.240 Confusion over the definition of the partial defence241 and the argument that the criminal law would operate more rationally and more fairly if provocation was abolished,B2 make generalisations risky. Loss of self control is, in itself, a dubious ground for palliating guilt. There is no excuse, not even a partial excuse, for outrageous brutality, no matter how convincing the evidence that the individual lost all self contro1.243 Nor is there any excuse for individuals who brood on past grievances, no matter how real, and nurture their hatred to the point where it finds expression in killing.244 The point of the objective test is clear enough. Ordinarily decent human beings, who are driven by intolerable pressures to exceed the bounds of reasonable necessity and kill their tormentors, should not be convicted of murder. So long as the objective test expresses a standard of misbehaviour which extends no further than a rational exercise of compassion allows, the objective test needs no supplementary rules to restrict its operation. In a number of cases, it is said that the objective test is restricted by a requirement that the provocation be sudden in its effects. The notorious jury direction given by Devlin J in Durn245 required a "sudden and temporary loss of self-control" and an absence of time for reflection during which "reason [could] regain dominion over the mind". This emphasis on the suddenness of the impulse to kill was reinforced by the requirement of a provocative incident, contemporaneous with the fatal response.246 Taken together, the requirements of suddenness and contemporaneity tend to exclude reliance on the defence of provocation in cases where women kill in consequence of protracted domestic oppression.247 The requirement of contemporaneity in Durn 239 See, eg, King C in R above 11211: "The loss of self-control which is essential, is not to be J confused with the emotions of hatred, resentment, fear or revenge." 240 Discussed per Mason J in Vun Den Hoek v The Queen [1986] HCA 76; (1986) 161 CLR 158. The New Zealand Court of Appeal takes the same view in Pita above n237 at 664-665. 241 Confusion which may or may not have been alleviated by the High Court's essay in redefinition in Stingel v The Queen [1990] HCA 61; (1990) 171 CLR 312. 242 Among serious proposals for abolition of the partial defence, see the New Zealand Law Reform Committee Report on Culpable Homicide (1976), Review of Commonwealth Criminal Law, Interim Report: Principles of Criminal Responsibility and Other Matters (1990).The case for abolition is discussed by Goode, "The Abolition of Provocation" in Yeo, above n238 at 51-54 243 The clearest and most thorough exposition of this fundamental element in the definition of provocation is to be found in Singer, above n128 at 304ff. 244 This 1 take to be the morally justifiable central element, mixed with much that cannot be justified, in the very broadly expressed remarks by King C in R above 1121 J 1. 245 (1949) 1 All E R 932. See too,R v Thornton [I9921 1All ER 306 and R v Ahluwalia [I9921 4 All ER 889 cases, like R v Dufy, of fatal violence in response to a long course of domestic violence. These are the most recent English endorsements of the requirement of immediacy. 246 R v Jeffrey [1%7] V R 467 at 476,489; R [I9811 4 A Crim R 127,131, 141, 142; Thornton, id at 312. Amendments in s23(2), (3)(b) of the Crimes Act 1900 (NSW) extend the defence of provocation in that State to cases where the provocative conduct occurred "at any previous time". The factual context of decision of the NSW Court of Criminal Appeal in Maria Morabito (1992) 62 A Crim R 82, an unsuccessful appeal against sentence, suggest that the provocation has a substantially more generous scope in NSW. 247 The issues are discussed in the perceptive Comment, "Provoked Reason in Men and 458 SYDNEY LAW REVIEW [VOL 15: 403 appears to have been the invention of Devlin J. Jeremy Horder argues that it lacked any basis in prior law.248 Even if English courts are bound to accept it,249 Australian courts are not. In New South Wales, legislation removes any requirement of contemporaneity between provocation and response.250 In Parker,25l the leading common law authority on the point, provocation was available as a defence though a significant time had passed between the provocative acts of the victim and Parker's final, fatal assault.252 More recent cases reinforce the conclusion that provocation is available though the last provocative act and the defendant's response were separated in time. Though the Court in R253 said that provocation must be sudden, it held that the defence was open though the victim had been asleep for twenty minutes or more before he was killed. In Kontinnen,254 the most recent of the Australian cases in which a batterer was killed while sleeping there was, once again, a substantial lapse of time between his last provocative act and the shot which killed him. The defence of provocation was allowed to go to the jury which preferred to acquit her on the ground of self defence.255 The requirement of sudden loss of self control remains. It has been detached, however, from the requirement of a contemporaneous provoking act or incident. Australian courts are prepared to allow provocation to reduce murder to manslaughter in cases where the loss of control occurs some considerable time after the provocative conduct of the victim.256 These are not aberrant cases. They express a significant difference between the exculpatory effects of loss of self control from fear and loss of self control Women: Heat of Passion Manslaughter and Imperfect Self Defence" (1986) 33 UCLA LR 1679.See also Edwards, S, "Battered Women Who Kill" [I9901 New L 1380. I 248 Horder, J, Note: "Provocation and Loss of Self Control" (1992) 108 LQR 191.In an earlier article, "Sex Violence, and Sentencing in Domestic Provocation Cases" [I9891 Crim LR 546, Horder surveyed English sentencing practices in cases of manslaughter following "domestic provocation" and advocated a stopgap rule requiring lighter sentences in cases where homicide was provoked by domestic violence. 249 R v Ahluwalia above n245, is the most recent English decision purporting to endorse the Durn direction. Closer inspection of the reasoning of the Court of Appeal suggests that there has been a significant modification of English law. Though the Court continues to insist on a "sudden and temporary loss of self control", as a result of provocative conduct, the defence is not necessarily precluded by a lapse of time between that conduct and the reaction of the accused: id at 8% 'We accept that the subjective element in the defence ... would not at a matter of law be negatived simply because of the delayed reaction ...". 250 CrimesAct 1900(NSW) s23(2),(30(b)) 251 (1964) 111 CLR95. 252 Fisse, above n99 at 94-95 253 (1981) 4 A Crim R 127. Greene, J, "A Provocation Defence for Battered Women Who Kill?" (1989) 12Adelaide LR 145 at 155-156comments on the extreme artificiality of the Court's attempt to fit the facts to the stated requirement of sudden provocation by an act of the deceased. 254 Above n22. 255 Compare above n238 at 61 and 64, Tolmie suggests that such cases fit provocation to "facts which do not seem to fall within the spirit of the defence". She argues that too expansive a definition of provocation will encourage juries to compromise on verdicts of manslaughter. The risk of compromising justice is always present when defences and partial defences run in tandem. The cure is not to be found in restricting provocation however. The better course is to ensure fair and exact application of the existing law of self defence and rigorous appellate scrutiny of trials for murder. 256 In addition to the Australian authorities, see above n249. 19931 WOMEN WHO KILL IN SELF DEFENCE 459 from anger. The difference is rarely remarked because the overwhelming majority of provocation cases involve enraged men who kill from jealousy, outraged possessiveness or insults to their honour, virility or sexual orientation. Women do occasionally kill from jealousy, but reactive killings in response to domestic violence or domestic rape are far more common.257 Anger and fear are physiological kin, as many commentators have remarked, and they are commonly mixed together in reactions to provocation. Some distinctions can be drawn, however, according to which emotion predominates. Anger is fuelled by past grievances and insults. It may be instantaneous or it may mount as the full extent of the hurt sinks in. Unless kept on the boil, however, anger ebbs as the provocative incident recedes into the past. The common run of cases where provocation affords a partial excuse conform to this pattern and the frequency of such cases accounts for the frequent emphasis on the passage of time between provocation and response. Contrast fear, which is characteristically prospective. It mounts with each passing moment, as the threat approaches. It is significant that the decision in Parker,258 which extended the plea of provocation beyond exculpation for instantaneous reactions to irnmediate insult or injury, straddles the middle ground where the emotions of fear and anger are compounded. Parker's wife was leaving him with another man and he wanted her back. Mingled with his anger and hurt was his fear of losing her unless he engineered a confrontation.259 It has been objected that this broadening of the time frame for the operation of provocation will allow the partial defence to operate in unmeritorious cases.260 There is an evident need to discriminate among the possible grounds for provocation. Writing with reference to North American case law, Laurie Taylor puts the argument bluntly: [Clumulative terror should serve as an emotion adequate for heat of passion manslaughter and cumulative rage should not. Affronts to dignity and sexual pride should be coped with over time.261 Taylor accepts, however, that differentiation in terms of moral culpability is possible through the medium of the objective test.262 Provocation is only available when an ordinary individual, similar to the accused in all those characteristics which might have a rational bearing on the degree of culpability, 257 Above nl at 96-97. In her survey of New South Wales homicide cases, Wallace reports that "as many as 70% of the husband killings occurred in the context of violence by the husband on the wife". For a summary of similar reports of husband's violence as victim precipitation in US studies of domestic homicide, see Taylor, L, "Comment: Provoked Reason in Men and Women: Heat of Passion Manslaughter and Imperfect Self Defence" (1986) 33 UCLA LR 1679 at 1720-1726.Wallace reports, id at 100, that the cases where jealousy was a motive were, "almost without exception, wife killings". Polk and Ranson, above n2 at 17-20reach similar conclusions. 258 (1963)111 CLR 610. 259 Id at 628 per Dixon CJ. 260 Above n248 at 553-554. Similar issues are involved in the question whether the defence of provocation is confined to cases in which the provocative acts occurred in the presence of the accused: discussed Lanharn, "Provocation and the Requirement of Presence" (1989) 13 Crim W 133. The cases on presence and provocation are primarily concerned, however, with men who kill from motives of jealousy. 261 Taylor, L, above n257 at 1719. 262 Id at 1732. 460 SYDNEY LAW REVIEW [VOL 15: 403 could have been driven to kill. The more closely the objective test is fitted to the facts of the case, the more likely it is that juries will be able to distinguish between those who deserve the benefits of the partial defence and those who do not. So long as they are within the generously defined bounds of normality, all those characteristics of the accused which tend to excuse the loss of self control and fatal retaliation can be imputed to the ordinary person for the purposes of the test. The test has two purposes. It is meant to preclude the presentation of defences based on mental illness in the guise of provocation263 and it distinguishes between "human frailty", which deserves compassion and "human ferocity", which does not.264 The objective test can be employed, as Taylor suggests, to distinguish the exculpatory effects of cumulative terror from brooding anger or the desire for revenge. 6. Conclusion The Australian common law of self defence is free from many of the restraints which bulk so large in North American case law and analysis. Reasonable necessity governs the availability of the defence and there are few, if any, more particular rules to constrain the operation of that general criterion. The peculiarly American preoccupation with distinctions between justification and excuse has no counterpart here. The fear of opening the floodgates to vigilantes whether in the subway or bedroom, does not exercise so pervasive an effect on the legal imagination. The law of the defences is more accommodating to defensive acts impelled by necessity. There is, accordingly, far less need to emulate North American attempts to extend the operation of the law of self defence by pleading incapacity, helplessness and mental disorder. It is preferable to avoid reliance on special exceptions based on the assumption that women victimised by violent partners are mentally disabled. There is a battle to be fought on the issue of expert evidence. It has nothing to do, however, with the diagnosis of individual pathology. It has everything to do with the need to encourage judicial recognition of the need for expert evidence to inform juries of the appalling realities of domestic violence in our society.265 That is one requisite for justice. The other is rigorous insistence, at trial and appellate levels, on directions which leave the issue of self defence to the jury with adequate instructions on the parameters of reasonable necessity. 263 See R v Enright [I9611 VR 663; R v Falconer [1990] HCA 49; (1990) 171 CLR 30. 264 R v Kirkhnm [1839] EngR 273; (1837) 173 ER 422 at 4W,above n210. The moral character of homicide must be judged of principally by the extent to which the circumstances of the case show, on the one hand, brutal ferocity, whether called into action suddenly or otherwise, or on the other, inability to control natural anger excited by a serious cause."See also, Singer, above 11236 314-322. at 265 Above n21.