"We have not now to consider what would be malice aforethought to bring a given case within the common law definition of murder; here the statute says that the act must be unlawful and malicious . . . the jury might perhaps have found on this evidence that the act was malicious, because they might have found that the prisoner knew that the natural consequence of his act would be to break the glass, and although that was not his wish, yet he was reckless whether he did it or not; but the jury have not so found . . . "

This decision was distinguished in R. v. Latimer (1886) 17 Q.B.D. 359. Two men quarrelled in a public house. One of them struck at the other with his belt. The glancing blow bounced off and struck the prosecutrix, wounding her severely. The assailant was prosecuted and convicted for having unlawfully and maliciously wounded her, contrary to section 20 of the Offences Against the Person Act 1861. Counsel for the defendant relied on Pembliton.

In his judgment Lord Coleridge C.J. said, at p. 361,

 "It is common knowledge that a man who has an unlawful and malicious intent against another, and, in attempting to carry it out, injures a third person, is guilty of what the law deems malice against the person injured, because the person is doing an unlawful act, and has that which the judges call general malice, and that is enough."

A similar theme is found in the brief judgments of the other members of the court, who were able to distinguish Pembliton which, as Bowen L.J. put the matter: "was founded not upon malice in general but on a particular form of malice, viz., malicious injury to property." The Lord Justice put the case thus:

 "It is quite clear that the act was done by the prisoner with malice in his mind. I use the word 'malice' in the common law sense of the term, viz., a person is deemed malicious when he does an act which he knows will injure either the person or property of another."

    My Lords, I find it hard to base a modern law of murder on these two cases. The court in Latimer was, I believe, entirely justified in finding a distinction between their statutory backgrounds and one can well accept that the answers given, one for acquittal, the other for conviction, would be the same today. But the harking back to a concept of general malice, which amounts to no more than this, that a wrongful act displays a malevolence which can be attached to any adverse consequence, has long been out of date. And to speak of a particular malice which is "transferred" simply disguises the problem by idiomatic language. The defendant's malice is directed at one objective, and when after the event the court treats it as directed at another object it is not recognising a "transfer" but creating a new malice which never existed before. As Dr. Glanville Williams pointed out (Criminal Law, the General Part 2nd Ed. (1961), p. 184) the doctrine is "rather an arbitrary exception to general principles." Like many of its kind this is useful enough to yield rough justice, in particular cases, and it can sensibly be retained notwithstanding its lack of any sound intellectual basis. But it is another matter to build a new rule upon it.

    I pause to distinguish the case of indiscriminate malice from those already discussed, although even now it is sometimes confused with them. The terrorist who hides a bomb in an aircraft provides an example. This is not a case of "general malice" where under the old law any wrongful act sufficed to prove the evil disposition which was taken to supply the necessary intent for homicide. Nor is it transferred malice, for there is no need of a transfer. The intention is already aimed directly at the class of potential victims of which the actual victim forms part. The intent and the actus reus completed by the explosion are joined from the start, even though the identity of the ultimate victim is not yet fixed. So also with the shots fired indiscriminately into a crowd. No ancient fictions are needed to make these cases of murder.

    I turn to deal more briefly with the remaining rules. The third rule, it will be recalled, is that a foetus cannot be the victim of murder. I see no profit in an attempt to treat the medieval origins of this rule. It is sufficient to say that is established beyond doubt for the criminal law, as for the civil law (Burton v. Islington Health Authority [1993] Q.B. 204) that the child en ventre sa mère does not have a distinct human personality, whose extinguishment gives rise to any penalties or liabilities at common law.

    The fourth rule is an exception to the generally accepted principle that actus reus and mens rea must coincide. A continuous act or continuous chain of causes leading to death is treated by the law as if it happened when first initiated. The development of this into the fifth rule, which links an act and intent before birth with a death happening after a live delivery, causes a little more strain, given the incapacity of the foetus to be the object of homicide. If, however, it is possible to interpret the situation as one where the mental element is directed, not to the foetus but to the human being when and if it comes into existence, no fiction is required.

    My Lords, the purpose of this enquiry has been to see whether the existing rules are based on principles sound enough to justify their extension to a case where the defendant acts without an intent to injure either the foetus or the child which it will become. In my opinion they are not. To give an affirmative answer requires a double "transfer" of intent: first from the mother to the foetus and then from the foetus to the child as yet unborn. Then one would have to deploy the fiction (or at least the doctrine) which converts an intention to commit serious harm into the mens rea of murder. For me, this is too much. If one could find any logic in the rules I would follow it from one fiction to another, but whatever grounds there may once have been have long since disappeared. I am willing to follow old laws until they are overturned, but not to make a new law on a basis for which there is no principle.

    Moreover, even on a narrower approach the argument breaks down. The effect of transferred malice, as I understand it, is that the intended victim and the actual victim are treated as if they were one, so that what was intended to happen to the first person (but did not happen) is added to what actually did happen to the second person (but was not intended to happen), with the result that what was intended and what happened are married to make a notionally intended and actually consummated crime. The cases are treated as if the actual victim had been the intended victim from the start. To make any sense of this process there must, as it seems to me, be some compatibility between the original intention and the actual occurrence, and this is, indeed, what one finds in the cases. There is no such compatibility here. The defendant intended to commit and did commit an immediate crime of violence to the mother. He committed no relevant violence to the foetus, which was not a person, either at the time or in the future, and intended no harm to the foetus or to the human person which it would become. If fictions are useful, as they can be, they are only damaged by straining them beyond their limits. I would not overstrain the idea of transferred malice by trying to make it fit the present case.

    Accordingly, I would differ from the thoughtful judgments delivered in R. v. Kwok Chak Ming (1963) H.K.L.R. 226 and 349, and hold that on the presumed facts the judge was right to direct an acquittal on the count of murder.

    Before leaving this part of the appeal I would acknowledge the extensive citation by counsel of passages from learned writers, present and past. There is no space to name them all here, or to recognise the further sources consulted after the close of the hearing. All have proved valuable, even if not all of the opinions expressed have been adopted. Notwithstanding the strong practical character of the criminal law it has over the years gained immeasurably from systematic analysis by scholars who have had an opportunity for research and reflection denied to those immersed in the daily life of the courts. I hope that the practice of drawing on these materials will be continued and enlarged.

    I turn to the question of manslaughter. This has caused me great anxiety. Until I read the speech of my noble and learned friend, Lord Hope of Craighead I had reasoned as follows. The brief discussion by the Court of Appeal assumed that a conviction of manslaughter would be appropriate, and appropriate only, in a case where the defendant intended to cause harm to the mother short of grievous bodily harm. The reasoning was the same as in the case of murder, apart from the difference in intent. Since I have differed from this reasoning in the case of murder I must also reject it for manslaughter. There is, however, another possibility which takes account of the fact that the offence of manslaughter unites a group of crimes which have nothing in common except their name. Two members of this group, often called "unlawful act" and "gross negligence" manslaughter, respectively, share the feature that it is possible to commit them without any violent intent towards the victim. It is obvious that with this type of offence the problems of transferred malice cannot arise. The Court of Appeal had no occasion to consider whether one of these, namely, the "unlawful act" variety, might furnish an alternative route to a verdict of manslaughter, but the trial judge did so and concluded, evidently with some hesitation, that it would not. His reasoning, as I understand it, was that the authorities require that the act in question, before it can be characterised as unlawful for this purpose, must be one which is likely to injure another person; and since the foetus is not a person, the danger to its continued existence could not make the act unlawful.

    My Lords, whilst recognising the logic of this approach I had felt that the learned judge had stopped short of carrying it through, for there was the possibility that the unlawful attack on the mother was a threat not only to the foetus before birth but also to the live child when ultimately born. It appeared to me, therefore, that a defendant in the position of B could in theory be convicted of manslaughter, if the jury was satisfied that a sober and reasonable person would have realised that the attack could harm the child if afterwards born alive, and if causation was proved. I add the qualification "in theory" because the formulation of a direction along these lines in terms which a jury could understand would be difficult, and the prospect that a competently represented defendant would, in fact, be convicted would be small indeed. Nevertheless, that is where the authorities seemed to point, so that a highly qualified affirmative answer should be given to Question 1.1.(ii)(b).

    Study of the speech to be delivered by Lord Hope has, however, persuaded me that this perspective may be too narrow, in directing attention to the foreseeability on the part of the accused that his act would create a risk to the person at whom it was aimed and to a wider class of persons falling within the area of potential danger. My noble and learned friend proposes, however, that the search for this wider class is misconceived. All that it is needed, once causation is established, is an act creating a risk to anyone; and such a risk is obviously established in the case of any violent assault by the risk to the person of the victim herself (or himself). In a case such as the present, therefore, responsibility for manslaughter would automatically be established, once causation has been shown, simply by proving a violent attack even if (which cannot have been the case here) the attacker had no idea that the woman was pregnant. On a broader canvas, the proposition involves that manslaughter can be established against someone who does any wrongful act leading to death, in circumstances where it was foreseeable that it might hurt anyone at all; and that this is so even if the victim does not fall into any category of persons whom a reasonable person in the position of the defendant might have envisaged as being within the area of potential risk. This is strong doctrine, the more so since it might be said with some force that it recognises a concept of general malice (that those who do wrong must suffer the consequences of a resulting death, whether or not the death was intended or could have been foreseen) uncomfortably similar to the one rejected more than 150 years ago by the courts and commentators in the context of murder; and one which, it is proper to add, I have proposed in the first part of this speech, should be rejected once again in that context.

    It is this feature which has caused me to hesitate long in joining the remainder of your Lordships to hold that a verdict of manslaughter can be available in circumstances as broad as this. I am, however, entirely convinced by the speech of Lord Hope that this is the present state of English law. To look for consistency between and within the very different crimes of murder and manslaughter is, I believe, hoping for too much. One can, however, look for a result which does substantial justice, and this is what I believe verdicts that B was not guilty of murder and guilty of manslaughter would have achieved.

    I would, therefore, answer the questions of law in the sense proposed by my noble and learned friend, Lord Hope of Craighead.

    I now turn to a subsidiary issue left over at an early stage which has caused me some concern. It arises from the fact that the life of S came to an end because her constitution was unable to withstand the ordinary perils of infancy. In a narrow perspective, she died of natural causes. This fact is not easily accommodated within the traditional definitions, or indeed the common understanding, of murder as a crime where the victim is killed. Nor is the problem entirely solved by pointing to a situation where a baby is deliberately exposed in bad weather, for there the evil-doer's act causes the child to die, whereas here S's life simply came to an end without anything having been done whilst she was living to shorten its span. Nor is there any useful analogy with the setting of a time-bomb before a birth which explodes afterwards and kills the newborn child, cf. Rex v. Shephard [1919] 2 K.B. 125, at p. 126: for there was in that case an intention that the living child should be killed, which S was not.

    My Lords, if I had thought that the agreed facts could otherwise have founded an indictment for murder I would have wished to look more closely at this problem, although I suspect that in the end it might well have proved to be simply a matter of vocabulary, and that the word "kills" in the traditional definition is wide enough to mean "causing the death of a person by an act or omission but for which the person killed would not have died when he did": Stephen, Digest of the Criminal Law (1877) 138. S. would have lived a longer life but for what happened before she was born.

    In any event, however, this feature does not in my opinion stand in the way of a conviction for manslaughter. The unlawful and dangerous act of B changed the maternal environment of the foetus in such a way that when born the child died when she would otherwise have lived. The requirements of causation and death were thus satisfied, and the four attributes of "unlawful act" manslaughter were complete.

IV. Remaining Issues

    My Lords, it will not be overlooked that this long opinion has so far given answers for only two of the numerous questions raised by the point of law referred: numerous because the point of law depends on three alternative presumptions as to the nature and circumstances of the injury and three presumptions as to B's intention. This makes it apposite to recall section 36(1) of the Criminal Justice Act 1972, by virtue of which the matter is now before the House:

 "36(1) Where a person tried on indictment has been acquitted . . . the Attorney-General may, if he desires the opinion of the Court of Appeal on a point of law which has arisen in the case, refer that point to the court, and the court shall, in accordance with this section, consider the point and give their opinion on it."

I have emphasised certain words which impose an essential restriction on this valuable power. The courts have always firmly resisted attempts to obtain the answer to academic questions, however useful this might appear to be. Normally, where an appeal is brought in the context of an issue between parties, the identification of questions which the court should answer can be performed by considering whether a particular answer to the question of law might affect the outcome of the dispute. The peculiarity of a reference under the Act of 1972 is that it is not a step in a dispute, so that in one sense the questions referred are invariably academic. This peculiarity might, unless limits are observed, enable the Attorney-General, for the best of motives, to use an acquittal on a point of law to set in train a judicial roving commission on a particular branch of the law, with the aim of providing clear, practical and systematic solutions for problems of current interest. This is not the function of the court, and the words emphasised in section 36(1) were in my view designed to keep the proceedings in Attorney-General's references within proper bounds.