Issues on the Appeal
24 The Crown identified two legal errors in his Honour's reasoning which, it submitted, would justify this Court intervening.
25 First, the Crown directed attention to that passage of his Honour's judgment at par [40], quoted in par [20] above, in which he indicated that it was unacceptable for the crime of murder to be "treated as less serious than manslaughter except in the clearest and most exceptional circumstances". The Crown submitted that, although there was a clear hierarchy in offences of homicide, it was not the case that sentences for manslaughter would overlap sentences for murder only in "the clearest and most exceptional circumstances". This proposition is correct. The degree of overlap between sentencing for manslaughter and murder is not able to be stated in terms of identifying "most exceptional circumstances" or words to that effect.
26 The elements of murder include an intention to kill or to inflict grievous bodily harm. Such an intention is, of course, of the greatest significance in the sentencing exercise. Where it is present and a charge of murder is made out, the appropriate sentence will generally be higher than for other forms of homicide. Plainly, other circumstances being equal, the moral culpability of the offender is significantly higher where one person causes the death of another in circumstances where an intention to do so exists. Nevertheless, the multifarious situations in which homicide can occur and the wide range of other considerations that must be taken into account in the exercise of the sentencing discretion are such that it is not correct to conclude that it is only in the "clearest and most exceptional circumstances" that any conviction for manslaughter could lead to a higher sentence than any conviction for murder.
27 Many charges of manslaughter are accompanied by the requisite intent, but the conviction is reduced to manslaughter by reason of provocation or diminished responsibility. In some such cases the extent of provocation or the degree to which responsibility was diminished is not such as to significantly reduce the appropriate penalty below the maximum permissible for manslaughter. On the other hand there are situations in which the charge of murder may be accompanied by circumstances which reduce the objective gravity of the offence, e.g. a case of euthanasia or a suicide pact. Such matters can result in an appropriate sentence being well below the maximum permissible for manslaughter.
28 There is a hierarchy of offences. However the existence of such a hierarchy does not lead to the situation that, where sentences of less than twenty-five years have been imposed for murder, sentences for manslaughter must generally be imposed for a significantly lesser period.
29 Even a case where there is present an intention to kill or maim, which would constitute murder but which is reduced, by reason of provocation or diminished responsibility, to a charge of manslaughter, will not necessarily attract a higher sentence than other forms of manslaughter, including the one relevant here, i.e. killing by an unlawful and dangerous act. As a five judge bench of this Court, including Gleeson CJ, Mason P, Hunt CJ at CL, Simpson and Hidden JJ, said in R v Isaacs (1997) 41 NSWLR 374 at 381:
"The argument for the appellant advanced on this appeal appeared to assume that a case of provocation manslaughter is necessarily, or at least ordinarily, worse than a case of manslaughter by an unlawful and dangerous act. We do not accept that. Each case depends upon its own circumstances. The range of sentencing available in the case of manslaughter is notoriously wide. There have been cases where provocation manslaughter has resulted in non-custodial sentences."
30 Paragraph [40] of his Honour's judgment, set out in par [20] above, is open to the interpretation for which the Crown contends, but does not require such an interpretation. I have no doubt that a judge of the experience of Miles AJ was well aware that sentences for murder and manslaughter can and do overlap. His Honour was referring to the general position that a conviction for murder would ordinarily attract a greater penalty than a conviction for manslaughter. I would reject this basis for alleged legal error.
31 The second error identified by the Crown concerned his Honour's treatment of the sentences imposed in other cases, as summarised in the judgment of Wood CJ at CL in R v Woodland and, particularly, the sentence in R v Ditfort NSWCCA 17 March 1992 (unreported).
32 Although his Honour recognised the wide variety of circumstances in which the crime of manslaughter can occur he did place reliance on what he identified to be an established range in child killing cases. Referring to the cases summarised in Woodland, supra, his Honour accepted at par [37] quoted in par [19] above that: "These cases establish a range of sentences for the unlawful killing of very young children by parents or carers between 5 years imprisonment and … 10 years imprisonment". (See also at [41] quoted in par [4] supra.) Miles AJ adopted a similar analysis in another case of this character which his Honour sentenced on the same day as the present case. (See R v Wilson [2003] NSWSC 1257 at [62]-[68].)
33 His Honour proceeded to treat Ditfort as if it were a benchmark. Indeed at par [49], set out at par [22] above, his Honour found that "the facts and the culpability involved were much the same as in the present case". His Honour took what he understood to be the starting point in Ditfort as the starting point for the sentencing exercise before him.
34 Counsel for the Respondent relied on Ditfort and in particular the references in Ditfort to the sentence in that case being at "the top of the range". However, that is to misunderstand the references in that case to "a range". Those references concerned the appropriate range for sentencing in that specific case, including both its objective and subjective circumstances. They were not references to an appropriate "range" for child killing on a charge of manslaughter.
35 Prior cases and Judicial Commission statistics do not often determine a range appropriate for a particular offence. They reflect what was regarded as appropriate in the wide variety of circumstances of those particular prior cases. Whether or not a sentencing pattern can be said to have emerged requires consideration of the whole body of sentences. It is unlikely that any such pattern can be said to have been established unless there have been a significant number of cases covering a wide variety of objective circumstances. Unless that is so, the cases would not encompass the relevant range of objective criminality.
36 In my opinion this second basis on which the Crown alleges his Honour committed a legal error should be upheld. I do not agree that the cases to which his Honour referred were such as to lead to the conclusion that a recognisable sentencing pattern or range can be said to have emerged with respect to the crime of manslaughter by an unlawful and dangerous act committed by parents or carers on very young children. Nor, in my opinion, is Ditfort so closely analogous as to serve as a benchmark for purposes of the sentencing exercise in this case.
37 Wood CJ at CL did not suggest in Woodland that the six cases which he set out in his judgment indicated anything in the nature of a sentencing pattern. What his Honour said when introducing the list of cases was:
"[28] Notwithstanding the caution which needs to be exercised in seeking guidance from other cases of manslaughter, references to decisions similarly involving the killing by the parent or a carer of a young child, by an unlawful and dangerous act, is of some assistance."
38 Wood CJ at CL identified certain principles appropriate for sentencing in such cases as quoted in par [18] above. His Honour did not suggest that the cases to which he referred had established a range in the sense of a sentencing pattern.
39 Manslaughter may vary, as has been pointed out, from a joke gone wrong to facts just short of murder. (R v Weinman (1987) 49 SASR 248 at 252.) This variation is reflected in sentencing. As Dunn LJ said in R v Boyer (1981) 3 Crim App R (S) 35 at 37:
"The offence of manslaughter attracts the widest band of sentences for any offence known to this court. The sentence can vary from life imprisonment to a conditional discharge. It depends on the precise circumstances of the case. It is impossible to subdivide the offence of manslaughter into different categories and say any particular sentence is appropriate to any particular category of events."
40 Similarly, Gleeson CJ pointed out in R v Blacklidge (NSWCCA unreported 12 December 1995) at [4]:
"It has long been recognised that the circumstances which may give rise to a conviction for manslaughter are so various, and the range of degrees of culpability is so wide, that it is not possible to point to any established sentencing tariff which can be applied to such cases. Of all crimes, manslaughter throws up the greatest variety of circumstances and culpability."
41 It may be possible to identify a distinct category of manslaughter for which variations on a basically similar factual situation can be identified. (See R v Parazisis & Bird (1991) 51 A Crim R 242 at 245.) However, this can only be done if there is a significant number of cases which share the common characteristic and which represent a very broad range of differing circumstances. Child killing by a parent or carer does not occur so frequently to make it possible to deduce a sentencing pattern from past cases. There is no reason to believe that the six cases referred to by Wood CJ at CL in Woodland cover the relevant range of objective criminality.
42 Plainly a violent attack, albeit not one accompanied by an intention to kill or inflict grievous bodily harm, by a person in a position of trust and responsibility upon a defenceless young child, is a circumstance of considerable aggravation. In this sense only can the offence of manslaughter in such cases be regarded as some form of separate category. It is a circumstance of such significance that it is entitled to substantial weight in the exercise of the sentencing discretion. The sense of outrage in the community about such a case is so strong that the element of retribution must play a prominent part in the exercise of the sentencing discretion. However, there is such a wide variety of other factors which may accompany this circumstance, that sentences for offences which share only this characteristic would not be expected to result in an identifiable sentencing range or pattern.
43 In the case of manslaughter, as I noted above, the acts constituting the offence and the circumstances of the offender at the time of the offence, may vary over a very wide range of objective gravity. The actual physical assault leading to death can range from comparatively minor force to a sustained beating over a prolonged period of time, incorporating elements of gratuitous cruelty. The personal culpability of the offender may vary from a carer who acts out of despair or in circumstances bordering on accident, to the vicious acts of a sadist.
44 In this regard child killing does not differ from other cases of manslaughter. When sentencing for the offence of manslaughter, a court must always have regard to the full context in which death was caused. That includes consideration of the extent of violence and the risk of death or serious injury which accompany it. It also includes consideration of any matter impinging on the moral culpability of the offender where there was no intention to inflict such injury, which can vary significantly. Where there was such an intention, but murder was reduced to manslaughter by provocation or mental impairment, the degree of provocation or of impairment, also bearing on moral culpability, can also vary significantly.
45 The fact that child killing is not a distinct subcategory of the crime of manslaughter for purposes of sentencing is suggested by a recent study of the Judicial Commission of New South Wales, published after his Honour's sentence in this case. (Judicial Commission of New South Wales Sentenced Homicides in New South Wales 1994-2000 Monograph Series 23, January 2004.)
46 That research shows that over the period of the study, the median sentence for murder or manslaughter of a child did not diverge significantly from the median sentence for other kinds of murders and manslaughters. In the case of manslaughter, non-child killers had sentences in a range from eighteen months to twenty-two years with a median head sentence of seven years, whereas child killers were sentenced in a range of two years to sixteen years, also with a median sentence of seven years. Non-parole periods for manslaughter ranged from nine months to sixteen years in the case of non-child killers with a median of 4.5 years and, in the case of child killers, from eight months to eleven years, with a median of 4.4 years. I should note that the child killers category included both parents or carers and other offenders who were not in a close relationship with their child victims. This research also revealed a significantly lower pattern of sentencing in the case of parents or carers. This result was affected by the number of offenders in that subcategory who claimed diminished responsibility. This manifests how issues of moral culpability must, and do, temper the need for retribution. (See Judicial Commission of New South Wales, supra, esp at pp43, 47, 48-50.)
47 Neither the research, nor the cases set out in Woodland, indicate anything that could be accurately identified as a sentencing pattern for child killers who are in a parental or carer relationship with the victim. That the end results were of a particular character, and could be stated as if they were in a "range", is not such as could be said to establish a sentencing pattern.
48 As Barr J, with whom Greg James J and Carruthers AJ agreed, said in R v Green [1999] NSWCCA 97:
"[24] Determining a proper sentence for manslaughter is notoriously difficult. Some assistance may be received from a consideration of the facts of other cases and the sentences imposed therein, but those cases do not determine an inflexible range of available sentences or reduce the maximum penalty applicable to a case which falls into the worst category of cases, namely 25 years penal servitude."
49 Even if no sentencing pattern can be identified, previous sentences are of some utility as a guide, by reason of the significance in the administration of criminal justice of the principle of consistency in sentencing. (See R v Whyte (2002) 55 NSWLR 252 at 168 to 189.) The principle was expressed by Gleeson CJ in Wong v The Queen (2001) 207 CLR 584 at [6] as follows:
"All discretionary decision making carries with it the probability of some degree of inconsistency. But there are limits beyond which such inconsistency itself constitutes a forms of injustice. The outcome of discretionary decision making can never be uniform, but it ought to depend as little as possible upon the identity of the judge who happens to hear the case. Like cases should be treated in like manner. The administration of criminal justice works as a system; not merely as a multiplicity of unconnected single instances. It should be systematically fair, and that involves, amongst other things, reasonable consistency."
50 It was for such reasons that Wood CJ at CL in Woodland referred to other decisions for the purpose of "guidance" from which he said he could receive "some assistance" ([28]). He did not refer to them as representing a sentencing range or pattern.
51 In his reasons for judgment Miles AJ regarded the six cases referred to in Woodland as identifying a "range of sentences established by past sentencing practice in this court". In this, in my opinion, his Honour erred. The cases did not establish a sentencing pattern in that sense.
52 Furthermore, in my opinion, his Honour erred in regarding the starting point in Ditfort as in some way an authoritative indication of an appropriate starting point for purposes of the present case. Ditfort bears some similarities with the present case, but there are also significant differences.
53 In Ditfort there were two blows, one a particularly violent blow of considerable force. In the present case it is clear that the assault involved a number of blows which continued over a longer period. Here the fatal blow may have been of a similar force to that in Ditfort. However the overall assault within which that blow occurred could not be classified as a lashing out in a fit of temper. The violence occurred over a period of time and was accompanied by conduct, bordering on torture, of particular heinousness in the application of clamps to the toes of the baby with sufficient force to break the toes.
54 There is a contrast between the impulsiveness of the crime in Ditfort and the more sustained assault which the evidence indicates must have occurred in the present case. There is also an element of cruelty in the present situation which makes Ditfort an inappropriate benchmark to use as identifying the top of the range appropriate to be applied in the present case.
55 The offender in Ditfort had convictions for dishonesty, but no history of violence. The Respondent in this case had convictions for common assault and assault occasioning actual bodily harm. The element of personal deterrence was, accordingly, entitled to greater weight.
56 The subjective circumstances in Ditfort were of greater force. The judge in that case accepted the applicant's contrition as true remorse. The findings in this respect in the present case are much weaker.
57 Furthermore, in this case, unlike Ditfort there were additional offences to be taken into account on a Form 1.
58 Another factor differentiating this present case from Ditfort was the fact that the Respondent was on bail for a stealing offence at the time of the offence presently under consideration.
59 The Crown has made out legal error and, subject to s6(3) of the Criminal Appeal Act 1912, this Court should proceed to exercise the sentencing discretion afresh.