COUNSEL: Roger Parker Q.C. and Robert Alexander for the appellants.

David Hirst Q.C., Andrew Bateson Q.C. and M. G. Tugendhat for the respondent.

 

SOLICITORS: Herbert Smith & Co.; Theodore Goddard & Co.

 

DATES:  1971 Nov. 29, 30;

Dec. 1, 2, 6, 7, 8, 9, 13, 14, 15, 16, 20;

1972 Feb. 23

 

JUDGES:  Lord Hailsham of St. Marylebone L.C., Lord Reid, Lord Morris of Borth-y-Gest, Viscount Dilhorne, Lord Wilberforce, Lord Diplock and Lord Kilbrandon.

 

Appeal from the Court of Appeal (Lord Denning M.R., Salmon and Phillimore L.JJ.).

 

February 23, 1972.

 

LORD HAILSHAM OF ST. MARYLEBONE L.C. My Lords,

 

Nature of the proceedings

 

This appeal arises out of two consolidated actions for libel on the publication of a book. The first action was in respect of the 60 proof copies of the book, the second in respect of the principal or hard back edition of the book. We were told that there are separate proceedings still pending in respect of a paper back edition, published under licence by  [*1051]  separate publishers. This paper back edition was mentioned at all stages in the proceedings as being potentially relevant to the question of damages. The House is not otherwise concerned with it.

 

The plaintiff in the action (the first respondent to this appeal) is a retired captain in the Royal Navy of unblemished reputation, who at the time of the matters referred to in the book, held the rank of Commander, and occupied the responsible position of officer commanding the escorts in the ill-fated convoy PQ17. He held active command throughout the war, and ended his wartime naval career with his present rank of captain in command of the battleship Ramillies. The subject matter of the book, and its title, was “The Destruction of Convoy PQ17” which, as is well known, was one of the great naval disasters of the war, in which all but 11 out of over 35 merchant vessels were sunk on their way to the Soviet Union and about 153 merchant seamen killed by enemy action and a vast quantity of war material lost.

 

The defendants in the action were respectively the author of the book, David Irving, who is the second respondent in the appeal, and was not represented before us, and the publishers of the book, Cassell & Co. Ltd., who are the appellants.

 

The result of the trial

 

The trial of the action took, we were told, 17 days before Lawton J. and a jury. In the result, on February 17, 1970, the jury found a verdict for the plaintiff and awarded against both defendants (1) the sum of £1,000 in respect of publication of the proof copies of the book, counsel for the plaintiff having waived any claim to exemplary damages on the proof copies, (2) £14,000 described as “compensatory damages” in respect of the hard back edition, and, (3) in respect of the hard back edition a further sum of £25,000, described as “by way of exemplary damages.” Judgment was entered for the sum of £40,000 against both defendants. The present appeal relates solely to the above sum awarded “by way of exemplary damages” of £25,000.

 

So far as relevant to this appeal, the entire proceedings before Lawton J. were conducted by all the counsel concerned and summed up by the judge to the jury on the basis of the remarks of Lord Devlin on pp. 1220-1233 of the report of Rookes v. Barnard [1964] A.C. 1129, and of the direction following Lord Devlin’s remarks by Widgery J. in Manson v. Associated Newspapers Ltd. [1965] 1 W.L.R. 1038. This was not surprising since all the other members of the House of Lords had expressly concurred in Lord Devlin’s opinion on this point, though without adding reasons of their own, and the opinion in Rookes v. Barnard, which was strictly an intimidation case, though obviously intended to apply generally, had been expressly applied to defamation proceedings by the Court of Appeal in McCarey v. Associated Newspapers Ltd. (No. 2) [1965] 2 Q.B. 86; by Pearson, Willmer and Diplock L.JJ.; in Broadway Approvals Ltd. v. Odhams Press Ltd. (No. 2) [1965] 1 W.L.R. 805, by Sellers, Davies and Russell L.JJ.; in Fielding v. Variety Incorporated [1967] 2 Q.B. 841, by Lord Denning M.R. and Harman and Salmon L.JJ.; and in Mafo v. Adams [1970] 1 Q.B. 548, a case of deceit and other causes of action, the principles enunciated in  [*1052]  Rookes v. Barnard were accepted as applicable where the evidence justified it by Sachs and Widgery L.JJ. and Plowman J.

 

Except for two important passages and one minor passage of which complaint is made, and to which I will come later, Lawton J.’s direction to the jury was unexceptionable as an exposition of the law as it has been declared in the House of Lords by an unanimous House in Rookes v. Barnard and applied by the Master of the Rolls and ten Lords Justices and one puisne judge in the above cases in the Court of Appeal and as it had been expounded by Widgery J. in his direction to the jury in Manson v. Associated Newspapers Ltd.

 

The appeal to the Court of Appeal

 

At the end of the seventeen-day trial the costs of the proceedings which, as between party and party, followed the event, must have already been enormous. Both defendants accepted the verdict on liability. The defendant Irving appealed on all the damages awarded. The present appellants appealed on the award of £25,000 “by way of exemplary damages.” The appeal lasted nine days before the Court of Appeal (Lord Denning M.R., Salmon and Phillimore L.JJ.) and judgment was given on March 4, 1971, dismissing both appeals with costs, which must by this time, with the costs of the trial, even on a party and party basis, have greatly exceeded the amount of the award. Before the Appellate Committee of this House the appeal lasted 13 working days, thus again greatly increasing the sum at stake, though by this time the respondent Irving had given up the struggle.

 

Judgment of the Court of Appeal

 

The Court of Appeal took a somewhat unusual course. On the view which they formed of the matter, which, as will appear, I have come to share though with greater hesitation than they expressed, they were for dismissing the appeal on the grounds that the criticisms of the direction by Lawton J. failed, and that the mere size of the award was not one which, on accepted principles, could be attacked. If they had stopped there, it is possible, and perhaps likely, that the proceedings would have come to an end. It is doubtful if leave to appeal to this House would have been given and if it had not, the two remaining parties would have been spared the costs of the 13 days’ hearing in your Lordships House. Even if leave to appeal had been given in the above circumstances a great deal of the time occupied before us would have been saved.

 

But the Court of Appeal did not stop at dismissing the appeal on these grounds. Whether or not they were encouraged by the zeal of plaintiff’s counsel, they put in the forefront of their judgments the view that Rookes v. Barnard [1964] A.C. 1129 was wrongly decided by the House of Lords and was not binding even on the Court of Appeal. It was, so they said, arrived at per incuriam, and without argument from counsel. It ignored, they claimed, two previous decisions in the House of Lords, Ley v. Hamilton (1935) 153 L.T. 384 and E. Hulton & Co. v. Jones [1910] A.C. 20, which had approved awards of punitive or exemplary damages on lines inconsistent with Lord Devlin’s opinion in Rookes v. Barnard. They felt themselves fortified in this view with the somewhat cool reception in the  [*1053]  Commonwealth of Rookes v. Barnard, particularly in the Australian Supreme Court decision in Uren v. John Fairfax & Sons Pty. Ltd. (1966) 117 C.L.R. 118 which had been affirmed so far as regards Australian law by the Judicial Committee of the Privy Council in the associated case of Australian Consolidated Press Ltd. v. Uren [1969] 1 A.C. 590. Neither Lord Denning M.R. nor Salmon L.J. seem to have been in any way inhibited or embarrassed by the fact that each had been party to at least one of the decisions of the Court of Appeal applying Rookes v. Barnard [1964] A.C. 1129 without question. Not content with all this, all three members of the Court of Appeal went further still and, besides declaring Rookes v. Barnard to have been decided per incuriam and ultra vires, proceeded to say that it was “unworkable,” and in the meantime, therefore [1971] 2 Q.B. 354, 384 “judges should direct juries in accordance with the law as it was understood before Rookes v. Barnard” which the court considered, to use the phrase of Lord Denning M.R. as “settled.”

 

As sent to us by the Court of Appeal, therefore, the appeal before us raised several questions of wide ranging importance. Quite apart from the merits of the respective litigants, these questions include the status of judgments and the relevance of precedent in this House, the circumstances when, if at all, decisions of this House may be questioned by the Court of Appeal, and judges of first instance directed by the Court of Appeal to disregard them. There is also the whole question of exemplary damages as canvassed in Rookes v. Barnard and subsequent decisions. What began as a simple proceeding between a plaintiff and two defendants has assumed, at the expense of two of the litigants, the dimensions of a constitutional question and a general enquiry into one aspect (and perhaps more than one aspect) of the law of damages.

 

The course taken by the Court of Appeal

 

In view of their importance it is unavoidable that before entering into the merits of the appeal I should discuss in a few paragraphs both the propriety and the desirability of the course taken by the Court of Appeal. I desire to do so briefly and with studied moderation.

 

From the point of view of the litigants it is obvious, I would have thought, that, on the view taken by the Court of Appeal, the course taken was unnecessary. Private litigants have been put to immense expense, of which most must be borne by the loser, discussing broad issues of law unnecessary for the disposal of their dispute.

 

If the Court of Appeal felt, as they were well entitled to do, that in the light of the Australian and other Commonwealth decisions Rookes v. Barnard ought to be looked at again by the House of Lords, either generally or under the Practice Declaration of 1966, Practice Statement (Judicial Precedent) [1966] 1 W.L.R. 1234, they were perfectly at liberty to say so. More, they could have suggested that so soon as a case at first instance arose in which the ratio decidendi of Rookes v. Barnard was unavoidably involved, the parties concerned might wish to make use of the so-called “leap-frogging” procedure now available to them under the Administration of Justice Act 1969 and thus avoid one stage in our three-tier system of appeals. But to impose on these litigants, to whom the question was, on the court’s view, unnecessary, the inevitable  [*1054]  burden of further costs after all they had been through up to date was not, in my view, defensible.

 

Moreover, it is necessary to say something of the direction to judges of first instance to ignore Rookes v. Barnard as “unworkable.” As will be seen when I come to examine Rookes v. Barnard in the latter part of this opinion, I am driven to the conclusion that when the Court of Appeal described the decision in Rookes v. Barnard as decided “per incuriam” or “unworkable” they really only meant that they did not agree with it. But, in my view, even if this were not so, it is not open to the Court of Appeal to give gratuitous advice to judges of first instance to ignore decisions of the House of Lords in this way and, if it were open to the Court of Appeal to do so, it would be highly undesirable. The course taken would have put judges of first instance in an embarrassing position, as driving them to take sides in an unedifying dispute between the Court of Appeal or three members of it (for there is no guarantee that other Lords Justices would have followed them and no particular reason why they should) and the House of Lords. But, much worse than this, litigants would not have known where they stood. None could have reached finality short of the House of Lords, and, in the meantime, the task of their professional advisers of advising them either as to their rights, or as to the probable cost of obtaining or defending them, would have been, quite literally, impossible. Whatever the merits, chaos would have reigned until the dispute was settled, and, in legal matters, some degree of certainty is at least as valuable a part of justice as perfection.

 

The fact is, and I hope it will never be necessary to say so again, that, in the hierarchical system of courts which exists in this country, it is necessary for each lower tier, including the Court of Appeal, to accept loyally the decisions of the higher tiers. Where decisions manifestly conflict, the decision in Young v. Bristol Aeroplane Co. Ltd. [1944] K.B. 718 offers guidance to each tier in matters affecting its own decisions. It does not entitle it to question considered decisions in the upper tiers with the same freedom. Even this House, since it has taken freedom to review its own decisions, will do so cautiously. That this is so is apparent from the terms of the declaration of 1966 itself where Lord Gardiner L.C. said [1966] 1 W.L.R. 1234:

 

“Their Lordships regard the use of precedent as an indispensable foundation upon which to decide what is the law and its application to individual cases. It provides at least some degree of certainty upon which individuals can rely in the conduct of their affairs, as well as a basis for orderly development of legal rules.

 

Their Lordships nevertheless recognise that too rigid adherence to precedent may lead to injustice in a particular case and also unduly restrict the proper development of the law. They propose, therefore, to modify their present practice and, while treating former decisions of this House as normally binding, to depart from a previous decision when it appears right to do so.

 

In this connection they will bear in mind the danger of disturbing retrospectively the basis on which contracts, settlements of property and fiscal arrangements have been entered into and also the especial need for certainty as to the criminal law.  [*1055] 

 

This announcement is not intended to affect the use of precedent elsewhere than in this House.”

 

It is also apparent from the recent case of Reg. v. National Insurance Commissioner, Ex parte Hudson [1972] A.C. 944, where the decision in Reg. v. Deputy Industrial Injuries Commissioner, Ex Parte Amalgamated Engineering Union, In re Dowling [1967] 1 A.C. 725 came up for review under the 1966 declaration, that the House will act sparingly and cautiously in the use made of the freedom assumed by this declaration.

 

In addition, the last paragraph of the declaration as quoted above clearly affirms the continued adherence of this House to the doctrine of precedent as it has been hitherto applied to and in the Court of Appeal.

 

The merits of the appeal

 

It is now possible to turn to the merits of the case so far as these were canvassed before us on the assumption of the continued authority of the Rookes v. Barnard decision. Before us the appellants made three contentions. (i) That there was no evidence to be left to the jury that the conditions were fulfilled to bring the case within one of the three “categories” of cases listed by Lord Devlin in Rookes v. Barnard as being appropriate for an award of punitive damages, and in particular the second, which was admittedly the only relevant category. (ii) That, even on the assumption that the first contention was wrong, Lawton J. had misdirected the jury in at least two important matters. (iii) That in any event the award of £25,000 was excessive, and could not be sustained. In order to understand these contentions it is necessary to say something about the facts.

 

The facts on which the book was founded

 

The fate of the PQ17 convoy is one of the most publicised, as well as one of the most tragic, naval operations of World War II. The evidence showed that it had been written about many times, notably by Captain Roskill, R.N., the official naval historian, and by the late Mr. Godfrey Winn, whose book was said to have sold half a million copies. It is unnecessary to recapitulate the facts here. They are graphically described in the judgment of Lord Denning M.R.

 

It is sufficient to say that the primary cause of the disaster flowed from an order to the convoy to scatter, which made the ships in it an easy prey to the aircraft and submarines by which they were attacked. This order to scatter was issued by the Admiralty in Whitehall and was due to a faulty appreciation by the naval staff, in particular, as is now known, by the then First Sea Lord himself, that the German battleship Tirpitz was at sea, and to a decision, also by the then First Sea Lord, to take the responsibility for the order on himself rather than leave the decision to the discretion of the naval officers on the spot. The naval officers on the spot, including Admiral Hamilton in command of the cruiser squadron, and Captain Broome, had no option but to obey, and the convoy was thus left to fan out on individual courses covering a vast area of sea.

 

So far there can be no controversy. But the two naval officers, rightly considering that the order to scatter must denote the approach of a superior  [*1056]  hostile surface force, sailed west in company. Admiral Hamilton was acting under precise orders from the Admiralty. Captain Broome was not. Captain Broome had proposed and Admiral Hamilton accepted that he should put himself under command of the admiral commanding the cruisers. That this decision was courageous there can be no doubt. What has been subsequently disputed was whether it was as wise as it was certainly brave. Some have thought that it was no more than the inevitable reaction of gallant and experienced naval officers to the threat of surface action. Others have thought that its effect was to remove from the area of the convoy the only naval elements, which might have countered the U-boat and air attacks, and thus to contribute to the extent of the convoy’s losses. Which of these two views be correct it is not appropriate here to discuss. But what is relevant to the present appeal is that those who criticised the decision had previously fastened the responsibility on Admiral Hamilton. It was one of the distinctive features of Mr. Irving’s book (which it may have shared with a German work with whose author he had collaborated) that it attempted to place responsibility for the withdrawal of the destroyers entirely or mainly on the shoulders of Captain Broome. This was a difficult thesis to sustain since Captain Broome was the junior officer of the two, and had only “proposed” the course which both forces ultimately pursued. It also involved the propositions, both disputable, that the decision was wrong in the light of the information then available, and that the absence of the destroyers made a significant difference to the loss of life and material.

 

From the start Captain Broome contended that the passages in the book relating to himself which it is not necessary to set out at length were defamatory. In paragraph 5 of his statement of claim he said that they meant and were intended and understood to mean:

 

“that the plaintiff was disobedient, careless, incompetent, indifferent to the fate of the merchant ships and/or by virtue thereof had wrongly withdrawn his destroyer force from the convoy and/or taken it closer to the German airfields than he had been ordered to and had thereby been largely responsible for or contributed extensively to the loss of the aforesaid ships and the effective destruction of more than two-thirds of the Convoy PQ17.”

 

In addition, at the trial it was contended that the ordinary and natural meaning of one of the relevant passages was that Captain Broome was a coward and for this reason “needed no second bidding” to desert the convoy. The defendants both disputed that the book bore any of these meanings, but contended that without them the passages in the book were true. It is evident from their verdict and from the magnitude of the award of damages that the jury rejected the contentions of the defence, though how far and to what extent must be to some extent a matter of speculation.

 

The material before the jury

 

From the commencement of the trial it was contended for Captain Broome that notwithstanding the limitations of Rookes v. Barnard [1964] A.C. 1129, he was entitled to “exemplary” or “punitive” damages. The trial judge ruled (though on this point he was subsequently overruled by the Court of Appeal) that, if so, he was bound to include a plea to this effect in  [*1057]  his statement of claim, and the pleading consequently introduced into the statement of claim by way of reamendment affords a convenient summary of the way the case was then put. The pleader wrote in paragraph 7:

 

“The plaintiff will assert that the defendants and each of them calculated that the money to be made out of the said book containing the passages complained of would probably exceed the damages at risk (if any) and that the plaintiff is consequently entitled to recover exemplary damages.”

 

He then went on to give particulars. If established, the plea clearly puts the case within the second of the three exceptional categories listed by Lord Devlin in Rookes v. Barnard. The question for the judge was whether there was evidence to leave to the jury on which they could find that the case was indeed to be placed in this category. If there was such evidence and if the jury were not misdirected, inclusion within the second category would have entitled (though not compelled) them to make some award on this account.

 

The appellants contended before the Court of Appeal and before us that there was no such evidence. In my opinion, this contention wholly fails. To convince us, they would in practice have to establish that there was no evidence on which a properly directed jury could find that at the time of publication they were fully aware the words bore and were intended and understood to bear the meanings attached to them in the statement of claim, since, if at the time of publication the words were known to bear these meanings, they were false to the knowledge of the appellants and published with that knowledge for profit. In my view, the meanings or most of them are sufficiently obvious from a casual reading of the book, and the inadequate attempts by the author or the publishers to provide an alternative meaning or an escape route by which they could argue the alternative before a jury by small modifications or carefully phrased ambiguities, are less an indication of innocence or na•vetŽ than a clear sighted appreciation of the danger that they faced. Mr. Irving was not represented before us, but his case was strenuously advanced before the Court of Appeal, and in another context (to be discussed later) we had to consider his case when counsel for the appellant expressly accepted as accurate the Master of the Rolls’ colorful account of his behavior. It is abundantly plain from this account that Mr. Irving at least knew, and carefully planned, what he was doing, that he went on with it in spite of repeated warnings from the most authoritative sources, that he conceived the book “as a book with a difference as all men” (that is, including Captain Broome) “were shown to be cowards,” and that he prided himself on being able to say “some pretty near the knuckle things about these people” (he was directly referring to Captain Broome’s threat of proceedings) “but if one says it in a clever enough way, they cannot take action.” The rules of evidence preclude us from taking these admissions of his state of mind as evidence against the appellants. But, in my opinion, the “near the knuckle things” said about Captain Broome in the course of this book, including the allegation that he was a coward, were said sufficiently plainly for an experienced publisher to know perfectly well what their meaning was and the fact that they were said “in a clever enough way” should  [*1058]  have told them plainly that they were said with deliberate intent to convey the meanings without incurring heavy damages.

 

But the case against Cassells does not stop at the obvious meanings to be attached to the passages in the book. Even if, which I could not easily accept, they did not understand the drift of the book at a first reading, they acquired the right to publish and they went on actually to publish in circumstances from which the jury were clearly entitled to infer that they went ahead with the most cold-blooded and clear-sighted appreciation of what they were doing.

 

The appellants were not the first publishers selected by Mr. Irving. His original publishers were William Kimber Ltd., who ultimately refused to publish the book on the ground that the book was “a continuous witch hunt of Captain Broome” having been advised by Captain Roskill, who gave evidence for Captain Broome, and perhaps by others, that “the book reeks of defamation.” In the absence of evidence by either defendant at the trial, it is impossible to say how much of this was known to the appellants. But it is certain that Mr. William Kimber warned the appellants in unmistakable terms that his house had rejected the book precisely on the grounds that it was libellous, amongst others of Captain Broome. The undisputed response of the appellants was either flippant or cynical. Moreover, Captain Broome himself had warned them on several occasions that, if they published the book, as they did, in substantially the form in which he had seen it, they must expect an action for libel from himself. That they took these threats seriously can be seen from their reaction to the latest of them which followed the issue of the proof copies. On receipt of this, the appellants placed a stop on the book in the following terms:

 

“Will you please note that absolutely and positively, not one single copy, on any pretext whatsoever, is to be removed from the house without reference to me.”

 

In attempts to sell the serial rights their efforts were “shot down” by three national Sunday newspapers presumably on the same grounds.

 

What the full explanation of their subsequent publication may have been will never be known, since the appellants did not elect to give evidence. But, in the absence of any explanation, the jury were perfectly entitled to infer that they had calmly calculated that the risks attendant on publication did not outweigh the chances of profit. What is certain is that, in so far as they were aware that the passages complained of could be reasonably understood to bear the meanings attached to them by Captain Broome, including the allegation of cowardice, they published them knowing them in this sense to be false, since no effort was made at any stage to suggest that there was any material on which a reasonable publisher could base the belief that the passages complained of, if they bore these meanings, were true. In his judgment in the Court of Appeal Lord Denning M.R. lists other features of the case against the appellants upon which the jury were entitled to base inferences. With most of these, except the reference to the paperback edition, which, contrary to what he says (perhaps per incuriam), was not published by the appellants but under licence by another publisher, I find myself in agreement. In particular, I concur in what was said in the Court of Appeal about the dust cover of the book,  [*1059]  which, making every allowance for the popular style in such productions, and putting the most favorable interpretation upon every phrase in it, seems, to my mind, in the absence of explanation, to indicate that the publishers were well aware of the full implication of the passages complained of and were prepared to sell the book on this sensational interpretation. In such circumstances, to argue that there was no evidence from which the jury could infer that “the appellants had calculated that the money to be made out of the book containing the passages complained of would probably exceed the damages at risk (if any)” was, to my mind, a somewhat forlorn hope, and nothing which counsel for the appellants said in the course of his strenuous and ably conducted argument has convinced me to the contrary. I will refer to the passage from Lord Devlin’s speech in Rookes v. Barnard [1964] A.C. 1129 relating to the categories later for its proper interpretation, but I cannot see how, on any view, if these facts were proved to the satisfaction of a jury, properly directed, they are not sufficient to enable the jury to base inferences bringing the publication within the second category.

 

The direction on the relation between the two awards

 

There was much more substance in, and I find much greater difficulty in deciding upon, the appellants’ second contention, which was based, not upon Lord Devlin’s three listed categories, but upon his exposition of the general conditions under which exemplary damages may be awarded after the conclusion of the three “considerations” listed on pp. 1227-1228 of the report which, he says, ought always to be borne in mind. At this point, Lord Devlin said, at p. 1228:

 

“Thus a case for exemplary damages must be presented quite differently from one for compensatory damages; and the judge should not allow it to be left to the jury unless he is satisfied that it can be brought within the categories I have specified. But the fact that the two sorts of damage differ essentially does not necessarily mean that there should be two awards. In a case in which exemplary damages are appropriate, a jury should be directed that if, but only if, the sum which they have in mind to award as compensation (which may, of course, be a sum aggravated by the way in which the defendant has behaved to the plaintiff) is inadequate to punish him for his outrageous conduct, to mark their disapproval of such conduct and to deter him from repeating it, then it can award some larger sum” (italics mine). “If a verdict given on such direction has to be reviewed upon appeal, the appellate court will first consider whether the award can be justified as compensation and if it can, there is nothing further to be said. If it cannot, the court must consider whether or not the punishment is, in all the circumstances, excessive. There may be cases in which it is difficult for a judge to say whether or not he ought to leave to the jury a claim for exemplary damages. In such circumstances, and in order to save the possible expense of a new trial, I see no objection to his inviting the jury to say what sum they would fix as compensation and what additional sum, if any, they would award if they were entitled to give exemplary damages. That is the course which he would have  [*1060]  to take in a claim to which the Law Reform (Miscellaneous Provisions) Act, 1934, applied.”

 

In my opinion, this passage contains a most valuable and important contribution to the law of exemplary damages which, prior to Rookes v. Barnard, had not, so far as I am aware, been adequately stressed in any previous case, and which, in my view, would retain, and possibly even increase, its value even if the categories in Rookes v. Barnard were to be wholly rejected.

 

In essence the doctrine is that the award of a punitive element in damages, if it is ever permissible, must also remain discretionary, and, in order to give effect to the second of the three “considerations” listed at p. 1227, the judge should always warn a jury that they need not award anything, and must not do so unless they are satisfied that a purely compensatory award (in a sense which I will explain) is inadequate. It follows that whatever they do award should only be a sum which has taken into account the award of damages already notionally allowed as compensation, including, where appropriate, the “aggravated” element required by a defendant’s bad conduct, and should never exceed the amount by which the required penalty (if that is the right word) exceeds the required compensation.

 

I shall revert to this feature of Rookes v. Barnard later. But what is said in substance by the appellants in this case is that the summing-up failed to give effect to this important and, in my view, vital principle.

 

The learned judge directed the jury over two days and much that he said was irrelevant to the question of exemplary damages. Of what was relevant to exemplary damages, most was a direction to the jury about the second category and the evidence in the case relevant to it. This reflected the balance of argument by counsel during the case and it appears from a remark in the judgment of Phillimore L.J. in the Court of Appeal that, in some sense at least, both counsel agreed that, dependent on the view which the jury took of the facts, Lawton J. should leave the question of exemplary damages to the jury. But there were two passages in the summing-up relevant to the present issue. The first was a passage on the first day of the summing-up when the judge, having directed the jury that punitive damages were in the nature of a fine, went on to give two examples from the criminal law carrying the moral that the punishment must neither be excessive nor inadequate to the gravity of the offence and said:

 

“If you are going to punish a man to show him that libel does not pay, provided, of course, it comes within Widgery J.’s definition” (he was referring to Manson v. Associated Newspapers Ltd. [1965] 1 W.L.R. 1038) “what you do must be reasonable in all the circumstances, bearing in mind that it is a penalty.”

 

The second, and more important, of the passages was on the second day of the summing-up when, after leaving an agreed list of questions to the jury, the learned judge said:

 

“As you will see, the issue of damages has been divided into two questions. The first one is No. 3, ‘What compensatory damages do you award the plaintiff?’ You will remember that compensatory  [*1061]  damages are compensation for something, they are not given to you. When you come to consider that question you must remember that this is a joint publication by Cassell & Co. Ltd., and Mr. Irving. You do not award two different sums. You award one sum and you will leave the lawyers to work out what it means, but it is one sum. Do you all follow that? Then having decided what are the proper additional compensatory damages then you will go on and consider the fourth question, namely, ‘Has the plaintiff proved that he is entitled to exemplary damages?’ It is for him to prove that he is entitled to it, not for the defendants to prove that he is not. This question has got to be divided up into a number of subsidiary questions and the reason for this is problems of law which arise, but you do not have to concern yourselves with those. That is my responsibility. There are two defendants and, as I have been at pains to point out to you during my summing-up, the case against each defendant on the issue of punitive damages is different, so you will have to consider the case against each defendant separately. I suggest you start with Mr. Irving and then go on to Cassell & Co. Ltd. In respect of each of them you will ask yourselves this question: ‘Has the plaintiff proved his entitlement against that defendant?’ If the answer is yes then you will have to go on and assess how much punitive damages should be awarded. If the answer is no he will get no punitive damages. At least that will be your finding. What the law is is another matter, but that will be your finding. Having carried out that operation in relation to Mr. Irving you should carry out exactly a similar operation in relation to Cassell & Co. Remember all the time that letters written by Mr. Irving or to Mr. Irving, other than by Cassells, are not evidence against Cassell & Co. I cannot stress that too much. You will have to ask yourselves: ‘Has he proved that he is entitled to punitive damages against Cassell & Co. Ltd?’ If the answer is no that is that. If the answer is yes you will have to assess the damages. I have put all that into an omnibus lawyers’ series of questions. I could have put it all into one question, but I came to the conclusion that it would probably be better for you. I will read paragraph 4 again. ‘Has the plaintiff proved that he is entitled to exemplary damages? If yes, has he proved his entitlement against one or both of the defendants? If one only, against which one?’ Then you see the last question under this heading, ‘What additional sum should be awarded him by way of exemplary damages?’ Would you be good enough to underline the word ‘additional,’ because I want to know, and learned counsel want to know, if you do decide to award punitive damages, how much more do you award over and above the compensatory damage.”

 

What was said against this passage on behalf of the appellants was that this summing-up was defective in that it did not make it absolutely plain to the jury that before making any punitive award against the defendant they must first take into account and assess the punitive effect of any compensatory award (including any element of “aggravated” damage) and only award such amount (if any) by which the appropriate penalty exceeded such award.  [*1062]  I am bound to say that I have found the greatest difficulty in accepting the summing-up on this point as adequate, and my difficulties were increased by two passages in the final speech of Captain Broome’s counsel which, as counsel for the appellants persuasively argued, seemed to indicate that the respective awards of compensatory and punitive damages were entirely separate assessments and that one should not be balanced against the other. In so far as counsel said this, and he appears to have done so, he was, in my opinion, entirely wrong. In the end, however, I have come to the con-clusion that the judge’s direction was just adequate to convey the impression intended in the passage of Lord Devlin’s speech Which had been accurately read to the jury by counsel for Mr. Irving and that the jury were not in fact misled. In coming to this conclusion I have been impressed, as was the Court of Appeal, by the stress the judge laid on the word “additional” in the passage cited, by the fact that the form of the questions left to the jury (which did not include as it should have done, the words “if any” in that relating to punitive damages) was agreed by counsel and by the fact that the line of the judge’s summing-up was entirely in accord with the case for the appellants as it was put to the jury on their behalf, and that everyone seems to have assumed that the result of the jury’s answers was that which in fact obtained. I desire, however, to say that the direction on this point, if sufficient, as I am constrained to say it was, was only barely sufficient, and that I trust that in future cases of this kind trial judges will stress the matter a good deal more clearly and with greater emphasis than was done here. In the present case I do not think that the judge can be blamed for putting the matter compendiously in a form which seems to have misled no one, which accorded with the way and with the emphasis with which it had been put to the jury on behalf of the appellants, and which, according to Phillimore L.J.’s observation quoted above, had, in some sense, been agreed.

 

A single award or two?

 

Less meritorious, in my view, was the second criticism of the direction put before us. This was in effect that the judge did not correctly direct the jury as to the principles on which a joint award of exemplary damages can be made against two or more defendants guilty of the joint publication of a libel in respect of which their relevant guilt may be different, and their means of different amplitude. With high regard for the judgments of Lord Denning M.R. and of Salmon L.J., I differ from both in what they said on this aspect of the matter, both as to the effect of the judge’s summing-up and to what it ought to be in such cases. Lord Denning M.R. said [1971] 2 Q.B. 354, 383:

 

“There is, of course, a difficulty. How is a jury to assess the one figure against two defendants. Are they to fix it at a high sum which they think the mare blameworthy ought to pay? Or a low sum for the least blameworthy? That must be left to the jury. They may, if they choose, fix a figure in between. The judge can, I think, tell them that they can fix it as against the more blameworthy, expecting him to pay it: and leave the least blameworthy (if he is called upon to pay) to recover contribution. In this case the judge left it to them without any  [*1063]  specific direction. That was, I think quite legitimate; and it is no ground for disturbing the verdict.” (the italics are mine).

 

Lord Denning then added:

 

“In any case, however, I think Cassells are not at liberty to take this point. They did not ask judge or jury to split the damages. The judge told counsel the questions he was going to put to the jury: and asked for their comments. That was the time for counsel to ask for the exemplary damages to be split. As that was not done, it is too late to ask in this court.”

 

Salmon L.J. appears to have thought that the award should reflect the amount due by the most guilty of the tortfeasors and he said at p. 393:

 

“... it is well settled that where there are several defendants who have all committed a joint tort, there can be only one award of one sum of damages against all of them: Greenlands Ltd. v. Wilmshurst and London Association for Protection of Trade [1913] 3 K.B. 507. It may bear hardly on one or more of the defendants. The moral may be that you must be as careful in choosing your companions in tort as you are in choosing your companions when you go out shooting.” (The italics are again mine.)

 

With respect to both judgments which, as will be seen, are arguably not quite consistent with one another, I think the effect of the law is exactly the opposite and that awards of punitive damages in respect of joint publications should reflect only the lowest figure for which any of them can be held liable. This seems to me to flow inexorably both from the principle that only one sum may be awarded in a single proceeding for a joint tort, and from the authorities which were cited to us by Mr. Parker in detail in the course of his argument. Mr. Parker referred us to Heydon’s Case (1612) 11 Co.Rep. 5a; Clark v. Newsam (1847) Exch. 131; Hill v. Goodchild (1771) 5 Burr. 2790; Dawson v. M’Clelland [1899] 2 Ir.R. 486; Greenlands Ltd. v. Wilmshurst and London Association for Protection of Trade [1913] 3 K.B. 507, especially at p. 521; Smith v. Streatfeild [1913] 3 K.B. 764, 769; Chapman v. Lord Ellesmere [1932] 2 K.B. 431, 471 per Slesser L.J.; Dougherty v. Chandler (1946) 46 N.S.W.S.R. 370; Egger v. Viscount Chelmsford [1965] 1 Q.B. 248, 262 and to Gatley on Libel and Slander, 6th ed. (1967), p. 606, para. 1390. I think that the inescapable conclusion to be drawn from these authorities is that only one sum can be awarded by way of exemplary damages where the plaintiff elects to sue more than one defendant in the same action in respect of the same publication, and that this sum must represent the highest common factor, that is, the lowest sum for which any of the defendants can be held liable on this score. Although we were concerned with exemplary damages, I would think that the same principle applies generally and in particular to aggravated damages, and that dicta or apparent dicta to the contrary can be disregarded. As counsel conceded, however, plaintiffs who wish to differentiate between the defendants can do so in various ways, for example, by electing to sue the more guilty only, by commencing separate proceedings against each and then consolidating, or, in the case of a book or newspaper  [*1064]  article, by suing separately in the same proceedings for the publication of the manuscript to the publisher by the author. Defendants, of course, have their ordinary contractual or statutory remedies for contribution or indemnity so far as they may be applicable to the facts of a particular case. But these may be inapplicable to exemplary damages.

 

Having established his principle, counsel for the appellant went on to argue that the judge had misdirected the jury, seeking to encourage us in this belief by the submission that, if he had persuaded at least two members of the Court of Appeal to defend it on one of two possibly inconsistent and erroneous bases, the learned judge might well have succeeded in making the jury accept one of them as the ground of their award.

 

The passage in the summing-up on which the appellants relied for this purpose was as follows. It occurs immediately after the passage already quoted in which the judge directs the jury to regard the exemplary damages as a sum additional to the compensatory award. Lawton J. went on:

 

“You may be saying to yourselves: if we do take the view that both these defendants should pay something by way of punitive damages, should we take into consideration the relative culpability of each one? Again, and I merely say this by way of illustration, and certainly not by way of guidance to you, say, for example you took the view that Mr. Irving was more to blame than Cassell & Co., or to be fair, you took the view that Cassell & Co. being an experienced firm of publishers were more to blame than this young man, Mr. Irving, should you make Cassell & Co. pay a larger sum by way of punitive damages than Mr. Irving? The answer to that is no” (italics mine). Whatever damages, if any, you decide should be awarded by way of punitive damages must be the same sum in respect of both Mr. Irving and Cassell & Co. Ltd., if you find them both liable to pay punitive damages. Have I made that clear?”

 

This direction is in many ways defective as a piece of clear English prose. In particular, it contains an ambiguity, later cured by an exchange in the presence of the jury between counsel and the Bench as to whether the jury is to award a single sum against both defendants or two sums, each against one of the defendants. But on the crucial point as to whether this sum, when awarded, should represent the higher or the lower figure for which the jury found either guilty, I myself find no difficulty in thinking that the jury would have been clear that they were to award the lower. I would hope that on other occasions this would be made even plainer, but I find it difficult to criticise an experienced judge for not being absolutely crystal clear on this point at the end of a two-day direction over a wide range of different topics following a 17-day trial. I would not disturb the verdict on these grounds.

 

I also consider that, having agreed to the form of the questions left to the jury, it was not really open to the appellants to contend, on appeal, that the awards should be split. In any case I am fortified in my view of the matter by the fact that I find the same difficulty as did the Court of Appeal in differentiating in any way between the moral culpability of the two defendants. Mr. Irving may have been the author of the defamatory  [*1065]  matter. But the appellants published it, on the jury’s finding, with their eyes open as to what it contained. It may be that Mr. Irving had fewer means, and, if the jury were looking on the exemplary damages from the point of view of deterring him, they could have awarded a smaller sum. But there seems to have been no evidence concerning the means of either party, and I do not see how at this late date we can properly be invited to speculate. The enterprise was essentially a joint one, and if the appellants had not all the information available to Mr. Irving, they had enough to make sure that they knew exactly what they were doing. It is difficult to know on what principle the jury could have differentiated between the two defendants.

 

Was the award excessive?

 

The final point taken for the appellants was that the award of £25,000 exemplary damages, or, as it was equally properly, and possibly better, put, the total award of £40,000 (which included the exemplary element) was so far excessive of what twelve reasonable men could have awarded that it ought to be set aside and a new trial ordered. I cannot disguise from myself that I found this an extremely difficult point in the case, and have only decided that the verdict should not be disturbed, with great hesitation, because I am very conscious of the fact that I would certainly have awarded far less myself, and possibly, to use a yardstick which some judges have adopted as a rule of thumb, less than half the £25,000.

 

A number of factors lead me, however, to the belief that the verdict should not be disturbed. The first, and paramount, consideration in my mind is that the jury is, where either party desires it, the only legal and constitutional tribunal for deciding libel cases, including the award of damages. I do not think the judiciary at any level should substitute itself for a jury, unless the award is so manifestly too large, as were the verdicts in Lewis v. Daily Telegraph Ltd. [1963] 1 Q.B. 340 or manifestly too small, as in English and Scottish Co-operative Properties Mortgage and Investment Society Ltd. v. Odhams Press Ltd. [1940] 1 K.B. 440, that no sensible jury properly directed could have reached the conclusion. I do not think much depends on the exact formula used to describe the test to be applied, whether the traditional language so large (or small) “that twelve sensible men could not reasonably have given them”: perLord Esher M.R. in Praed v. Graham (1890) 24 Q.B.D. 53, 55 or that of Palles C.B. in M’Grath v. Bourne (1876) Ir.R. 10 C.L. 160, 164 cited by Lord Wright in Mechanical and General Inventors Co. Ltd. and Lehwess v. Austin and Austin Motor Co. Ltd. [1935] A.C. 346, 378, that “no reasonable proportion existed between it and the circumstances of the case.” The point is that the law makes the jury and not the judiciary the constitutional tribunal, and if Parliament had wished the roles to be reversed in any way, Parliament would have said so at the time of the Administration of Justice (Miscellaneous Provisions) Act 1933, since section 6 of that Act expressly excepts defamation actions (otherwise than in a limited class of case) from the general change which it then authorised.

 

In addition to the above cases counsel for the respondent cited Youssoupoff v. Metro-Goldwyn-Mayer Pictures Ltd. (1934) 50 T.L.R. 581,  [*1066]  583, 584; Bocock v. Enfield Rolling Mills Ltd. [1954] 1 W.L.R. 1303; Scott v. Musial [1959] 2 Q.B. 429, 436; Morey v. Woodfield (No. 2) (Note) [1964] 1 W.L.R. 16; McCarey v. Associated Newspapers Ltd. (No. 2) [1955] 2 Q.B. 86 and Broadway Approvals Ltd. v. Odhams Press Ltd. (No. 2) [1965] 1 W.L.R. 805 especially at pp. 818, 820.

 

I do not see anything in the above cases which alters the principle involved, nor am I aware of anything in the nature of exemplary damages to alter it in this limited class of case. It may very well be that, on the whole, judges, and the legal profession in general, would be less generous than juries in the award of damages for defamation. But I know of no principle of reason which would entitle judges, whether of appeal or at first instance, to consider that their own sense of the proprieties is more reasonable than that of a jury, or which would entitle them to arrogate to themselves a constitutional status in this matter which Parliament has deliberately withheld from them, for aught we know, on the very ground that juries can be expected to be more generous on such matters than judges. I speak with the greater conviction because my own view is that the legal profession is right to be callous in such matters and juries are wrong if they can be said to be more generous. But that is not the law and I do not think that judges who hold my view are any more entitled to change the law on this topic than they have been in the past.

 

Counsel very rightly drew our attention to observations of Lord Devlin in Rookes v. Barnard [1964] A.C. 1129, 1227-1228, when he said:

 

“I should not allow the respect which is traditionally paid to an assessment of damages by a jury to prevent me from seeing that the weapon is used with restraint. It may even be that the House may find it necessary to follow the precedent it set for itself in Benham v. Gambling [1941] A.C. 157, and place some arbitrary limit on awards of damages that are made by way of punishment.”

 

I regard Benham v. Gambling as setting an absolutely necessary but wholly arbitrary rule to solve an absolutely insoluble problem, and I do not think it could readily be extended to exemplary damages for libel simply on the ground that judges do not agree with juries on quantum. I do not think the first sentence in Lord Devlin’s observation means more than that the House will use its legitimate powers to interfere with awards by juries with particular regard to the need for preserving liberty, which he was concerned to express, and, if it means that the House was conferring on itself greater powers than it previously possessed, I would have regarded it as an usurpation of the function of the legislature as a whole. We were also referred to the observations of the Court of Appeal in Ward v. James [1966] 1 Q.B. 273, 301. If the passage quoted there means more than that the court, in exercising its undoubted right to interfere with unreasonable verdicts, will have more regard than heretofore to the general level of damages in cases of a similar nature, and particularly personal injury cases, it may need further consideration.

 

The second reason which leads me to decline to interfere with the jury’s verdict in this case is the peculiar gravity of the facts of this case. I share with Phillimore L.J. the view [1971] 2 Q.B. 354, 400 that the jury must have found that  [*1067] 

 

“these were grave libels perpetrated quite deliberately and without regard to their truth by a young man and a firm of publishers interested solely in whether they would gain by the publication of this book. They did not care what distress they caused.”

 

It is true, and I have been constrained to say, that I would have treated this heinous offence against public decency with far less severity than did the jury in this case. But, at the end of the hearing, I found myself as unable to say as were the three eminent judges in the Court of Appeal that no twelve reasonable jurors could have come to a different conclusion from myself. These matters are very highly subjective, and I do not feel myself entitled to substitute my own subjective sense of proportion for that of the constitutional tribunal appointed by law to determine such matters.

 

I should add, lest I be thought to have overlooked the point, that to avoid the expense and anxieties of a new trial counsel on both sides agreed to leave to us, in case the appeal should succeed, the assessment of any sum to be awarded. I doubt myself how satisfactory this would have been but, quite obviously, before we embarked upon such a task we should have to be first satisfied that the original verdict could not stand, and to this preliminary issue the agreement between counsel is necessarily irrelevant.

 

The decision in Rookes v. Barnard

 

These considerations really conclude the result of this appeal. It must, in my view, be dismissed. But, lest other litigants be put to expense and uncertainty comparable to that which the parties to this case have, in my view, unnecessarily suffered, it is now unavoidable that I should deal at length with the wider issues in the law of damages on which the Court of Appeal founded the greater part of its judgment. Before I do so I ought to remark that, though counsel for the appellants took the point that the trial judge should have withdrawn the question of the paper back edition from the jury, I regard the way in which he left it to them as so favorable to the appellants as not to justify a new trial on that ground alone.

 

The judgment of the Court of Appeal was based on the simple proposition that the decision in Rookes v. Barnard [1964] A.C. 1129 so far as it affected punitive or exemplary damages was made per incuriam and without prior argument by counsel and that judges should in future ignore it as unworkable, and that, in directing juries, judges of first instance should return to the status quo ante Rookes v. Barnard as if that case had never been decided at all.

 

I have already said, and will not repeat, what I think about the propriety of the Court of Appeal in doing this at all, and the appropriateness, in view of the consequences to the parties, of their doing it in this case. I now proceed to consider how far their opinions are correct.

 

I make no complaint of their view that Rookes v. Barnard clearly needs reconsideration by this House, if only because of the reception it has received in Australia, Canada and New Zealand. I view with dismay the doctrine that the common law should differ in different parts of the Commonwealth, which is the effect of the decision in Australian Consolidated Press Ltd. v. Uren [1969] 1 A.C. 590, and anything one can do in this case to bring the various strands of thought in different Commonwealth  [*1068]  countries together ought to be done. Moreover, as I shall show, many of Lord Devlin’s statements have been misunderstood, particularly by his critics, and the view of the House may well have suffered to some extent from the fact that its reasons were given in a single speech. Whatever the advantages of a judgment of an undivided court delivered by a single voice, the result may be an unduly fundamentalist approach to the actual language employed. Phrases which were clearly only illustrative or descriptive can be treated in isolation from their context, as being definitive or exhaustive. I am convinced that this has happened here and that to some extent at least, the purpose and nature of Lord Devlin’s exposition has been misunderstood.

 

The law before Rookes v. Barnard

 

Whatever else may be said, the Court of Appeal’s judgment is based on one assumption which is plainly incorrect. This assumption is, to quote its most characteristic expression on the lips of Lord Denning M.R., [1971] 2 Q.B. 354, 379: “Prior to Rookes v. Barnard, the law as to exemplary damages was settled.” In point of fact, it was nothing of the kind. Lord Denning M.R. went on immediately to quote from Mayne and MacGregor on Damages, 12th ed. (1961), the following passage from para. 207:

 

“Such damages are variously called punitive damages, vindictive damages, exemplary damages, and even retributory damages. They can apply only where the conduct of the defendant merits punishment, which is only considered to be so when his conduct is wanton, as where it discloses fraud, malice, violence, cruelty, insolence, or the like, or as it is sometimes put, where he acts in contumelious disregard of the plaintiff’s rights ... Such damages are recognised to be recoverable in appropriate cases of defamation;”

 

If Lord Denning M.R. had gone on to quote from paragraph 212 of the same edition he would have read the following passage, inconsistent with his construction of the foregoing, under the heading “A Double Rationale” which should, I hope, have disabused him of the idea that the law of punitive damages was in fact settled prior to Rookes v. Barnard. The passage is as follows:

 

“3. A Double Rationale

 

Through all these various cases, however, runs another thread, giving a very different explanation of the position. For indeed it cannot be said that English law has committed itself finally and fully to exemplary damages, and many of the above cases point to the rationale not of punishment of the defendant but of extra compensation for the plaintiff for the injury to his feelings and dignity. This is, of course, not exemplary damages at all. It is another head of non-pecuniary loss to the plaintiff.” (The italics are mine).

 

Indeed, in the well-known American textbook on The Law of Damagesby the late Professor Charles T. McCormick, published in 1935 by eke West Publishing Company of Minnesota occurs the following passage to the same effect on p. 278:  [*1069] 

 

“In England, where exemplary damages had their origin, it is still not entirely clear whether the accepted theory is that they are a distinct and strictly punitive element of the recovery, or they are merely a swollen or ‘aggravated’ allowance of compensatory damages permitted in cases of outrage. It is only in America that the cases have clearly separated exemplary from compensatory damages, and it is only here that the doctrine, thus definitely isolated, has been attacked and criticised.”

 

More characteristic than either of these passages and more illustrative of the confusion which reigned before Rookes v. Barnard is the paragraph on the subject in Lord Simonds’ 3rd edition of Halsbury’s Laws of England vol. 11 (1955) title “Damages”; p. 233, para. 391:

 

“Exemplary damages. Where the wounded feeling and injured pride of a plaintiff, or the misconduct of a defendant, may be taken into consideration, the principle of restitutio in integrum no longer applies. Damages are then awarded not merely to recompense the plaintiff for the loss he has sustained by reason of the defendant’s wrongful act, but to punish the defendant in an exemplary manner, and vindicate the distinction between a wilful and an innocent wrongdoer. Such damages are said to be ‘at large,’ and, further, have been called exemplary, vindictive, penal, punitive, aggravated, or retributory.”

 

This passage clearly shows the extraordinary confusion of terminology reflecting differences in thinking and principle which existed up to 1964. Apart from anything else, “aggravated” damages, classed as compensatory by Mayne and MacGregor, and by Professor McCormick, are assimilated to exemplary or punitive damages as such, as is the phrase damages “at large,” — an expression so indefinite in its connotation that counsel for the appellants in argument felt able to include within it (as this passage suggests inappropriately) even the general damages for pain and suffering in a personal injuries case. Clearly, before Rookes v. Barnard, the thinking and the terminology alike called aloud for further investigation and exposition, and, since in such cases it is the classic function of this House to make such reviews, I cannot accept eke simplicity doctrine of the Court of Appeal either that there was no need to make it, or that the only thing to restore clarity is to go back to the state of the law as it was in 1963. In passing, I may say that I do not attach so much importance as did the Court of Appeal to the circumstance that eke two categories mentioned by Lord Devlin had never been discussed in argument by counsel. The cases and textbooks on exemplary damages had been exhaustively read, and when this House undertakes a careful review of the law it is not to be described as acting per incuriam or ultra vires if it identifies and expounds principles not previously apparent to the counsel who addressed it or to the judges and text book writers whose divergent or confusing expressions led to the necessity for the investigation. Of course, in a sense, it would be easy enough to direct a jury under the old law if one simply said to them that any conduct of which they chose on rational grounds to disapprove would give rise to an award of exemplary damages and that any sum they chose to think appropriate as the penalty would be acceptable. But no one in recent years has ever thought this, although it is noteworthy  [*1070]  that as recently as 1912 the author of Sedgwick’s A Treatise on the Measure of Damages, 9th ed., vol. I, para. 349, pp. 688-689 was writing:

 

“Until comparatively recent times juries were as arbitrary judges of the amount of damages as of the facts ... Even as late as the time of Lord Mansfield it was possible for counsel to state the law to be that ‘The court cannot measure the ground on which the jury find damages that may be thought large; they may find upon facts within their own knowledge. ...’ ... The doctrine of exemplary damages is thus seen to hail e originated in a survival in this limited class of cases of the old arbitrary power of the jury.” (Italics mine.)

 

Clearly modern juries must be given adequate professional guidance and the object of Lord Devlin’s opinion in Rookes v. Barnard was to enable them to have it. Speaking for myself, and whatever view I formed of the categories, I would find it impossible to return to the chaos which is euphemistically referred to by Phillimore L.J. [1971] 2 Q.B. 354, 399 as “the law as it was before Rookes v. Barnard.”

 

Before I examine the actual decision in Rookes v. Barnard I would now propose to make two sets of observations of a general character. The first relates to the context in which damages must be awarded, the second to the terminology to be used in particular classes of case.

 

The subjective element in damages

 

Of all the various remedies available at common law, damages are the remedy of most general application at the present day, and they remain the prime remedy in actions for breach of contract and tort. They have been defined as “the pecuniary compensation obtainable by success in an action for a wrong which is either a tort or a breach of contract.” They must normally be expressed in a single sum to take account of all the factors applicable to each cause of action and must of course be expressed in English currency: Mayne and MacGregor on Damages, 12th ed. (1961), para. 1.

 

In almost all actions for breach of contract, and in many actions for tort, the principle of restitutio in integrum is an adequate and fairly easy guide to the estimation of damage, because the damage suffered can be estimated by relation to some material loss. It is true that where loss includes a pre-estimate of future losses, or an estimate of past losses which cannot in the nature of things be exactly computed, some subjective element must enter in. But the estimate is in things commensurable with one another, and convertible at least in principle to the English currency in which all sums of damages must ultimately be expressed.

 

In many torts, however, the subjective element is more difficult. The pain and suffering endured, and the future loss of amenity, in a personal injuries case are not in the nature of things convertible into legal tender. The difficulties arising in the paraplegic cases, or, before Benham v. Gambling [1941] A.C. 157, in estimating the damages for loss of expectation of life in a person who died instantaneously, are only examples of the intrinsically impossible task set judges or juries in such matters. Clearly the £50,000 award upheld in Morey v. Woodfield (No. 2) (Note) [1964] 1 W.L.R. 16 could never compensate the victim of such an accident. Nor, so  [*1071]  far as I can judge, is there any purely rational test by which a judge can calculate what sum, greater or smaller, is appropriate. What is surprising is not that there is difference of opinion about such matters, but that in most cases professional opinion gravitates so closely to a conventional scale. Nevertheless in all actions in which damages, purely compensatory in character, are awarded for suffering, from the purely pecuniary point of view the plaintiff may be better off. The principle of restitutio in integrum, which compels the use of money as its sole instrument for restoring the status quo, necessarily involves a factor larger than any pecuniary loss.

 

In actions of defamation and in any other actions where damages for loss of reputation are involved. the principle of restitutio in integrum has necessarily an even more highly subjective element. Such actions involve a money award which may put the plaintiff in a purely financial sense in a much stronger position than he was before the wrong. Not merely can he recover the estimated sum of his past and future losses, but, in case the libel, driven underground, emerges from its lurking place at some future date, he must be able to point to a sum awarded by a jury sufficient to convince a bystander of the baselessness of the charge. As Windeyer J. well said in Uren v. John Fairfax & Sons Pty. Ltd., 117 C.L.R. 115, 150:

 

“It seems to me that, properly speaking, a man defamed does not get compensation for his damaged reputation. He gets damages because he was injured in his reputation, that is simply because he was publicly defamed. For this reason, compensation by damages operates in two ways — as a vindication of the plaintiff to the public and as consolation to him for a wrong done. Compensation is here a solatium rather than a monetary recompense for harm measurable in money.”

 

This is why it is not necessarily fair to compare awards of damages in this field with damages for personal injuries. Quite obviously, the award must include factors for injury to the feelings, the anxiety and uncertainty undergone in the litigation, the absence of apology, or the reaffirmation of the truth of the matters complained of, or the malice of the defendant. The bad conduct of the plaintiff himself may also enter into the matter, where he has provoked the libel, or where perhaps he has libelled the defendant in reply. What is awarded is thus a figure which cannot be arrived at by any purely objective computation. This is what is meant when the damages in defamation are described as being “at large.” In a sense, too, these damages are of their nature punitive or exemplary in the loose sense in which the terms were used before 1964, because they inflict an added burden on the defendant proportionate to his conduct, just as they can be reduced if the defendant has behaved well — as for instance by a handsome apology — or the plaintiff badly, as for instance by provoking the defendant, or defaming him in return. In all such cases it must be appropriate to say with Lord Esher M.R. in Praed v. Graham, 24 Q.B.D. 53, 55:

 

“... in actions of libel ... the jury in assessing damages are entitled to look at the whole conduct of the defendant” (I would personally add “and of the plaintiff”) “from the time the libel was published down to the time they give their verdict. They may consider what  [*1072]  his conduct has been before action, after action, and in court during the trial.”

 

It is this too which explains the almost indiscriminate use of “at large,” “aggravated,” “exemplary” and “punitive” before Rookes v. Barnard. To quote again from Professor McCormick’s work, it was originally only in America that the distinction between “aggravated” damages (which take into account the defendant’s bad conduct for compensating the plaintiff’s injured feelings) and “punitive” or “exemplary” damage was really drawn. My own view is that in no English case, and perhaps even in no statute, where the word “exemplary” or “punitive” or “aggravated” occurs before 1964 can one be absolutely sure that there is no element of confusion between the two elements in damages. It was not until Lord Devlin’s speech in Rookes v. Barnard that the expressions “aggravated,” on the one hand, and “punitive” or “exemplary,” on the other, acquired separate and mutually exclusive meanings as terms of art in English law.

 

The next point to notice is that it has always been a principle of English law that the award of damages when awarded must be a single lump sum in respect of each separate cause of action. Of course, where part of the damage can be precisely calculated, it is possible to isolate part of it in the same cause of action. It is also possible and desirable to isolate different sums of damages receivable in respect of different torts, as was done here in respect of the proof copies. But I must say I view with some distrust the arbitrary subdivision of different elements of general damages for the same tort as was done in Loudon v. Ryder [1953] 2 Q.B. 202, and even, subject to what I say later, what was expressly approved by Lord Devlin in Rookes v. Barnard [1964] A.C. 1129, 1228 for the laudable purpose of avoiding a new trial. In cases where the award of general damages contains a subjective element, I do not believe it is desirable or even possible simply to add separate sums together for different parts of the subjective element, especially where, as was done by agreement in this case, the subjective element relates under different heads to the same factor, in this case the bad conduct of the defendant. I would think with Lord Atkin in Ley v. Hamilton, 153 L.T. 384, 386: “The ‘punitive’ element is not something which is or can” (italics mine) “be added to some known factor which is non-punitive,” or in the words of Windeyer J. in Uren v. Fairfax & Sons Pty. Ltd., 117 C.L.R. 118, 150:

 

“The variety of the matters which, it has been held, may be considered in assessing damages for defamation must in many cases mean that the amount of a verdict is the product of a mixture of inextricableconsiderations.” (Italics again mine.)

 

In other words the whole process of assessing damages where they are “at large” is essentially a matter of impression and not addition. When exemplary damages are involved, and even though, in theory at least, it may be possible to winnow out the purely punitive element, the dangers of double counting by a jury or a judge are so great that, even to avoid a new trial, I would have thought the dangers usually outweighed the advantages. Indeed, though it must be wholly illegitimate to speculate in such a matter, the thought crossed my mind more than once during the hearing that it may even have happened in this case.  [*1073] 

 

Terminology

 

This brings me to the question of terminology. It has been more than once pointed out the language of damages is more than usually confused. For instance, the term “special damage” is used in more than one sense to denominate actual past losses precisely calculated (as in a personal injuries action), or “material damage actually suffered” as in describing the factor necessary to give rise to the cause of action in cases, including cases of slander, actionable only on proof of “special damage.” If it is not too deeply embedded in our legal language, I would like to see “special damage” dropped as a term of art in its latter sense and some phrase like “material loss” substituted. But a similar ambiguity occurs in actions of defamation, the expressions “at large,” “punitive,” “aggravated,” “re-tributory,” “vindictive” and “exemplary” having been used, as I have pointed out, in inextricable confusion.

 

In my view it is desirable to drop the use of the phrase “vindictive” damages altogether, despite its use by the county court judge in Williams v. Settle [1960] 1 W.L.R. 1072. Even when a purely punitive element is involved, vindictiveness is not a good motive for awarding punishment. In awarding “aggravated” damages the natural indignation of the court at the injury inflicted on the plaintiff is a perfectly legitimate motive in making a generous rather than a more moderate award to provide an adequate solatium. But that is because the injury to the plaintiff is actually greater and, as the result of the conduct exciting the indignation, demands a more generous solatium.

 

Likewise the use of “retributory” is objectionable because it is ambiguous. It can be used to cover both aggravated damages to compensate the plaintiff and punitive or exemplary damages purely to punish the defendant or hold him up as an example.

 

As between “punitive” or “exemplary,” one should, I would suppose, choose one to the exclusion of the other, since it is never wise to use two quite interchangeable terms to denote the same thing. Speaking for myself, I prefer “exemplary,” not because “punitive” is necessarily inaccurate, but “exemplary” better expresses the policy of the law as expressed in the cases. It is intended to teach the defendant and others that “tort does not pay” by demonstrating what consequences the law inflicts rather than simply to make the defendant suffer an extra penalty for what he has done, although that does, of course, precisely describe its effect.

 

The expression “at large” should be used in general to cover all cases where awards of damages may include elements for loss of reputation, injured feelings, bad or good conduct by either party, or punishment, and where in consequence no precise limit can be set in extent. It would be convenient if, as the appellants’ counsel did at the hearing. it could be extended to include damages for pain and suffering or loss of amenity. Lord Devlin uses the term in this sense in Rookes v. Barnard [1964] A.C. 1129, 1221, when he defines the phrase as meaning all cases where “the award is not limited to the pecuniary loss that can be specifically proved.” But I suspect that he was there guilty of a neologism. If I am wrong, it is a convenient use and should be repeated.

 

Finally, it is worth pointing out, though I doubt if a change of terminology is desirable or necessary, that there is danger in hypostatising  [*1074]  “compensatory,” “punitive,” “exemplary” or “aggravated” damages at all. The epithets are all elements or considerations which may, but with the exception of the first need not, be taken into account in assessing a single sum. They are not separate heads to be added mathematically to one another.

 

Analysis of Rookes v. Barnard

 

This being said, it is necessary to analyse the decision in Rookes v. Barnard, a case, it must be remembered, of intimidation and not libel. The only actual decision on damages must be looked for on pp. 1232-1233 where Lord Devlin says:

 

“I doubt whether the facts disclosed in the summing-up show even a case for aggravated damages; a different impression may be obtained when the facts are fully displayed upon a new trial. At present there seems to be no evidence that the respondents were motivated by malevolence or spite against the appellant. They wronged him not primarily to hurt him but so as to achieve their own ends. If that had not been their dominating motive, then what they did would not have been done in furtherance of a trade dispute and the whole case has been fought on the basis that it was. It is said that they persisted in believing that their closed shop position was endangered by the appellant’s conduct even when their official leaders told them that it was not. Be it so; pig-headedness will not do. Again, in so far as disclosed in the summing-up there was no evidence of offensive conduct or of arrogance or insolence. It was, I think, suggested that some impolite observations were made about the appellant, but that is not enough; in a dispute of this sort feelings run high and more than hard words are needed for aggravated damages. Mr. Silkin relied strongly on the flagrant breach of contract with B.O.A.C. and the respondents’ open disregard of their pledges and their lack of consideration. But this was not conduct that affected the appellant. He was no more distressed or humiliated by it than any of B.O.A.C.’s passengers whose convenience, it might be said, and interests were brushed aside by the respondents in their determination to secure their object.”

 

Although, as will be seen, I prefer much of what Lord Devlin said on the subject of exemplary damages to what has been said by his subsequent critics, and propose to follow it, the decision in Rookes v. Barnard must be viewed in the light of these conclusions. It is not verbally inspired. But it is a careful and valuable decision not lightly to be set aside.

 

The passages in the report which have given rise to criticism and discussion go from p. 1220 to the top of p. 1231 of the Law Report and can be divided conveniently into the following parts.

 

The first part consists in exposition of the authorities and principles which is contained in pp. 1220-1225 where Lord Devlin begins to draw his conclusions.

 

These conclusions, which form the second portion of his opinion, include the three “alleged categories” (pp. 1225-1227), the three “considerations” (pp. 1227-1230) and finally pp. 1230-1231 the commentary and exposition of  [*1075]  the consequences of what he has said and these occupy the rest of the passage under discussion.

 

Was the decision per incuriam?

 

Now, I think I must protest at the outset at the theory that Lord Devlin (or those members of the House who agreed with him) was speaking “per incuriam.” I have already dealt with the argument that his conclusions did not follow the actual submissions of counsel on either side.

 

Lord Devlin was, of course, perfectly well aware that, in drawing these conclusions from the authorities, he was making new law in the sense in which new law is always made when an important new precedent is established. Thus, he said, at p. 1226:

 

“I am well aware that what I am about to say will, if accepted, impose limits not hitherto expressed on such awards and that there is powerful, though not compelling, authority for allowing them a wider range. I shall not, therefore, conclude what I have to say on the general principles of law without returning to the authorities and making it clear to what extent I have rejected the guidance they may be said to afford.”

 

But a judge is always entitled to do this when the exact limits, rationale, and the extent of a principle is being discussed, and when those limits, rationale, and extent have never been authoritatively defined.

 

Nor can it be said fairly that he had ignored Ley v. Hamilton, 153 L.T. 384. In fact he quoted from it at length and treated it, making allowance for the confusion in the legal terminology at the time to which I have already drawn attention, as a case of “aggravated” damages. I think he was right in so doing; although I also think Salmon L.J. was almost certainly right in thinking that the inverted commas in which Lord Atkin puts “punitive” are not a guide to its meaning. The word is in inverted commas for the same reason that “real” in the earlier passage is in inverted commas. They are quotation marks and Lord Atkin was quoting the actual words in the judgment of Maugham L.J. which he was criticising.

 

It is a fairer criticism of Lord Devlin to say that he did not mention E. Hulton & Co. v. Jones [1910] A.C. 20. Both Mr. Hewart in argument in that case and Lord Loreburn L.C. in his speech, at p. 24, which may have been extempore, reflect a view of the law of damages for libel apparently at variance with the law as Lord Devlin has now declared it to be. But, as I shall show, the difference is more apparent than real. It is difficult to square either Mr. Hewart’s argument or the passage of Lord Loreburn L.C.’s speech with the explicit admission made in the Court of Appeal and repeated in the facts stated on p. 20 of the report that the use of the name “Artemus Jones” by the editor and author was innocent, and it is on this basis that the case is normally cited as an authority. Judging the use made of the case in the Court of Appeal by their own criteria of Lord Devlin, the case is certainly not a binding authority on the law of exemplary damages. It was never argued as such, although the observations of Lord Loreburn L.C. can be fairly used as testimony, and even as persuasive authority, for the state of legal thinking at the time. In law, however, if Lord Devlin be right, the law of exemplary damages was still evolving,  [*1076]  and E. Hulton & Co. v. Jones made no pretence at altering or defining it, nor did either counsel in the case argue the case in terms which raised the question in its present form.

 

Did Rookes v. Barnard extend exemplary damages to fresh torts?

 

Having rejected the theory that Lord Devlin’s speech can be pushed aside as having been delivered per incuriam, I hope I may now equally dispose of another misconception. I do not think that he was under the impression either that he had completely rationalised the law of exemplary damages, nor by listing the “categories” was he intending, I would think, to add to the number of torts for which exemplary damages can be awarded. Thus I disagree with the dictum of Widgery L.J. in Mafo v. Adams [1970] 1 Q.B. 548, 558 (which, for this purpose, can be treated as an action for deceit) when he said:

 

“As I understand Lord Devlin’s speech, the circumstances in which exemplary damages may be obtained have been drastically reduced, but the range of offences in respect of which they may be granted has been increased, and I see no reason since Rookes v. Barnard [1964] A.C. 1129 why, when considering a claim for exemplary damages, one should regard the nature of the tort as excluding the claim.”

 

This would be a perfectly logical inference if Lord Devlin imagined that he was substituting a completely rational code by enumerating the categories and stating the considerations. It is true, of course, that actions for deceit could well come within the purview of the second category. But I can see no reason for thinking that Lord Devlin intended to extend the category to deceit, and counsel on both sides before us were constrained to say that, though it may be paradoxical, they were unable to find a single case where either exemplary or aggravated damages had been awarded for deceit, despite the fact that contumelious, outrageous, oppressive, or dishonest conduct on the part of the defendant is almost inherently associated with it. The explanation may lie in the close connection that the action has always had with breach of contract: see the discussion in Mayne and MacGregor, chapter 41, especially at para. 968.

 

Where solatium is enough

 

The true explanation of Rookes v. Barnard is to be found in the fact that, where damages for loss of reputation are concerned, or where a simple outrage to the individual or to property is concerned, aggravated damages in the sense I have explained can, and should in every case lying outside the categories, take care of the exemplary element, and the jury should neither be encouraged nor allowed to look beyond as generous a solatium as is required for the injury simply in order to give effect to feelings of indignation. It is not that the exemplary element is excluded in such cases. It is precisely because in the nature of things it is, and should be, included in every such case that the jury should neither be encouraged nor allowed to look for it outside the solatium and then to add to the sum awarded another sum by way of penalty additional to the solatium. To do so would be to inflict a double penalty for the same offence.

 

The surprising thing about Rookes v. Barnard is not that Lord Devlin  [*1077]  restricted the award of exemplary damages viewed as an addition to or substitution for damages by way of solatium to the three so-called categories, but that he allowed the three so-called categories to exist by way of exception to the general rule. That he did this is due at least in part to the fact that he felt himself bound by authority to do so, but partly also because he thought that there were cases where, over and above the figure awarded for loss of reputation, for injured feelings, for outraged morality, and to enable a plaintiff to protect himself against future calumny or outrage of a similar kind, an additional sum was needed to vindicate the strength of the law and act as a supplement to its strictly penal provisions (cf. what he says at pp. 1226, 1230 of the report).

 

Is Rookes v. Barnard unworkable?

 

I confess I am quite unable to see why such a view of the matter is “unworkable.” As I have already pointed out, it has been worked in fact for nearly eight years. On the contrary, by insisting on a single sum being awarded for outrageous behavior in nearly every case of tort, and allowing the jury full vent to their legitimate feelings within the proportions set by the injury involved, it seems to me that judge and jury are set an inherently less difficult task than if they were told first to take into account the aggravating factors, and then to impose an additional “fine” for the size of which they have neither the qualifications nor any measure by which they can limit their discretion, particularly since neither counsel nor the judge can mention particular figures which can have any relevance to the actual case. The difficulty consists, not in working the system of aggravated and purely compensatory damages, where they apply, as they do in almost every case of contumelious conduct under Lord Devlin’s opinion, but in working a system of punitive damages alongside the system of aggravated and compensatory damage. This difficulty exists whether Lord Devlin’s limitation to the categories be right or wrong and, if it were wrong, would exist in every case, and not only in a small minority of cases. The difficulty resides in the fact that the thinking underlying the two systems is as incompatible as oil and vinegar, the one based on what the plaintiff ought to receive, the other based on what 12 reasonable, but otherwise uninstructed, men and women think the defendant ought to pay.

 

The meaning of the categories

 

As regards the meaning of the particular categories, I have come to the conclusion that what Lord Devlin said was never intended to be treated as if his words were verbally inspired, and much of the criticism of them which has succeeded reports of the case has been based on interpretations which are false to the whole context and unduly literal even when taken in isolation from it.

 

The only category exhaustively discussed before us was the second, since the first could obviously have no application to the instant case. But I desire to say of the first that I would be surprised if it included only servants of the Government in the strict sense of the word. It would, in my view, obviously apply to the police, despite Attorney-General for New South Wales v. Perpetual Trustee Co. Ltd. [1955] A.C. 457, and almost  [*107\8]  as certainly to local and other officials exercising improperly rights of search or arrest without warrant, and it may be that in the future it will be held to include other abuses of power without warrant by persons purporting to exercise legal authority. What it will not include is the simple bully, not because the bully ought not to be punished in damages, for he manifestly ought, but because an adequate award of compensatory damages by way of solatium will necessarily have punished him. I am not prepared to say without further consideration that a private individual misusing legal powers of private prosecution or arrest as in Leith v. Pope (1779) 2 Wm.Bl. 1327, where the defendant had the plaintiff arrested and tried on a capital charge, might not at some future date be assimilated into the first category. I am not prepared to make an exhaustive list of the emanations of government which might or might not be included. But I see no reason to extend it beyond this field, to simple outrage, malice or contumelious behavior. In such cases a properly directed jury will not find it necessary to differentiate between what the plaintiff ought to receive and what the defendant ought to pay, since the former will always include the latter to the extent necessary to vindicate the strength of the law.

 

When one comes to the second category, we reach a field which was more exhaustively discussed in the case before us. It soon became apparent that a broad rather than a narrow interpretation of Lord Devlin’s words was absolutely essential, and that attempts to narrow the second category by a quotation out of context of one sentence from the passage wherein it is defined simply will not do. Lord Devlin founded his second category on a sequence of cases beginning with Bell v. Midland Railway Co. (1861) 10 C.B.N.S. 287, and on the judgment of Maule J. in Williams v. Currie (1845) 1 C.B. 841, 848 and the dictum of Martin B. in Crouch v. Great Northern Railway (1856) 11 Exch. 742, 759. None of these were examples of precise calculation of the balance sheet type. Then he said, at p. 1227:

 

“It” (that is the motive of making a profit) “is a factor also that is taken into account in damages for libel; one man should not be allowed to sell another man’s reputation for profit. Where a defendant with a cynical disregard for a plaintiff’s rights has calculated that the money to be made out of his wrongdoing will probably exceed the damages at risk, it is necessary for the law to show that it cannot be broken with impunity. This category is not cantoned to moneymaking in the strict sense. It extends to cases in which the defendant is seeking to gain at the expense of the plaintiff some object — perhaps some property which he covets — which either he could not obtain at all or not obtain except at a price greater than he wants to put down. Exemplary damages can properly be awarded whenever it is necessary to teach a wrongdoer that tort does not pay.” (Italics mine.)

 

Even a casual reading of the above passage shows that the sentence: “Where a defendant with a cynical disregard for a plaintiff’s rights has calculated that the money to be made out of his wrongdoing will probably exceed the damages at risk, it is necessary for the law to show that it cannot be broken with impunity” is not intended to be exhaustive but illustrative, and is not intended to be limited to the kind of mathematical calculations  [*1079]  to be found on a balance sheet. The sentence must be read in its context. The context occurs immediately after the sentence: “one man should not be allowed to sell another man’s reputation for profit,” where the word “calculation” does not occur. The context also includes the final sentence: “Exemplary damages can properly be awarded whenever it is necessary to teach a wrongdoer that tort does not pay.” The whole passage must be read sensibly as a whole, together with the authorities on which it is based.

 

It is true, of course, as was well pointed out by Widgery J. in Manson v. Associated Newspapers Ltd. [1965] 1 W.L.R. 1038, 1045, that the mere fact that a tort, and particularly a libel, is committed in the course of a business carried on for profit is not sufficient to bring a case within the second category. Nearly all newspapers, and most books, are published for profit. What is necessary in addition is (i) knowledge that what is proposed to be done is against the law or a reckless disregard whether what is proposed to be done is illegal or legal, and (ii) a decision to carry on doing it because the prospects of material advantage outweigh the prospects of material loss. It is not necessary that the defendant calculates that the plaintiff’s damages if he sues to judgment will be smaller than the defendant’s profit. This is simply one example of the principle. The defendant may calculate that the plaintiff will not sue at all because he has not the money (I suppose the plaintiff in a contested libel action like the present must be prepared nowadays to put at least £30,000 at some risk), or because he may be physically or otherwise intimidated. What is necessary is that the tortuous act must be done with guilty knowledge for the motive that the chances of economic advantage outweigh the chances of economic, or perhaps physical, penalty.

 

At this stage one must examine some of the counter arguments which found favour in the Court of Appeal. How, it may be asked, about the late Mr. Rachman, who is alleged to have used hired bullies to intimidate statutory tenants by violence or threats of violence into giving vacant possession of their residences and so placing a valuable asset in the hands of the landlord? My answer must be that if this is not a cynical calculation of profit and cold-blooded disregard of a plaintiff’s rights, I do not know what is. It is also argued that the second category does not take care of the case of a man who pursues a potential plaintiff to ruin out of sheer hatred and malice. The answer is that it does not do so because this is already taken care of in the full compensation or solatium for the injury involved in which the jury can give full rein to their feeling of legitimate indignation without going outside the bounds of compensatory damages in the sense in which I have explained the phrase, that is, damages of sufficient size to enable the plaintiff to point to the size of the award to indicate the baselessness of the false charge, and damages for the outrage inflicted in exact proportion as it was unprovoked, unatoned for, or malicious. I would have thought the second category was ample to cover any form of injury committed within the scope of those torts for which aggravated and exemplary damages may be awarded where the motive was material advantage. Mafo v. Adams [1970] 1 Q.B. 548 is not really an authority to the contrary, although I would have thought that the damages there awarded for inconvenience, breach of covenant, and loss of a regulated tenancy were perhaps at present-day values too small for the wrong committed. What  [*1080]  was at issue in Mafo v. Adams was the award of exemplary damages in an action for deceit (see Sachs L.J., at p. 555) and this, in the event, was never decided. What was decided in that case was that the plaintiff had not discharged the onus of proof that the defendant’s motives were such as to bring the case within the second category. This is clear from the fact that both Sachs L.J. and Widgery L.J. based their judgments on a passage from the decision of the county court judge, where he said: “The defendant’s reasons for his actions are obscure,”: see per Sachs L.J., at p. 556, and per Widgery L.J., at p. 559. I am far from saying that, in so far as it could have been shown that the defendant was actuated by hope of gain, and if the action had been one of trespass, exemplary damages could not have been awarded under the second category, and even though in the absence of authority I am of opinion that exemplary damages cannot be awarded in an action for deceit, I cannot claim that that matter has been finally determined.

 

The main criticisms of Lord Devlin’s speech are thus shown to have been unfounded. That he went beyond the existing law he had no doubt, and nor have I. But, as I have shown, he was entitled to do so. It may very well be that, in deciding in favour of the two exceptional categories, he was making an unnecessary concession to tradition. But he made the concession after a careful analysis of the authorities, and, speaking for myself, and given the cautious approach indicated in Lord Gardiner L.C.’s practice declaration [1966] 1 W.L.R. 1234 and by a majority of this House in Reg. v. National Insurance Commissioner, Ex parte Hudson [1972] A.C. 944, I do not think there is any reason for disturbing them. I regard the Australian cases, and in particular Uren v. John Fairfax & Sons Pty. Ltd., 117 C.L.R. 118, as deciding no more than that on the particular facts of that case the award of exemplary damages was not acceptable. In so far as they claim to establish that exemplary damages can be awarded for any contumelious disregard of the plaintiff’s rights I may not, of course, comment so far as regards the law of Australia, but, so far as regards the law of England, I would say that an adequate award of compensatory damages in such a case must of necessity include, and perhaps more than include, any punitive or exemplary element. The proposition, as a proposition, would have been perfectly acceptable so long as the looser terminology prevalent before Rookes v. Barnard was in use. So far as regards the more strict terminology now to be employed, the proposition is not to be treated as acceptable in the English courts.

 

Before turning to the so-called “considerations” I desire to say a word concerning the decisions in Williams v. Settle [1960] 1 W.L.R. 1072 and Loudon v. Ryder [1953] 2 Q.B. 202, upon which Lord Devlin also commented. Williams v. Settle was a case under section 17 (3) of the Copyright Act 1956. I agree with Lord Devlin that it is for consideration in the light of subsequent cases whether that section, which does not use the phrase “exemplary damages,” does in fact give a right to damages which are exemplary in the narrower sense used since Rookes v. Barnard [1964] A.C. 1129. If it does, the case should be regarded as a second category case, since the defendant’s motive was profit. If it does not, and if it is to be regarded as still authoritative, Williams v. Settle can only be regarded as an extreme example of aggravated damages, though the language of the  [*1081]  county court judge was so strong as to lead me to think that I would not myself have been prepared to make so large an award.

 

Loudon v. Ryder [1953] 2 Q.B. 202 is the earliest instance which I have been able to find where a split award was made of exemplary and compensatory damages for the same tort, and the split was made in circumstances which are not altogether plain from the report, after an award of a lump sum had been announced. What would have happened if Devlin J. (as he was) had summed up to the jury in favour of a generous award of aggravated damages on the lines of his later speech in Rookes v. Barnard is, of course, a question which no one can possibly answer. The answer might well have been, substituting “trespass” for “defamation,” what Windeyer J. said in Uren v. John Fairfax & Sons Pty. Ltd., 117 C.L.R. 118, 152:

 

“Telling the jury in a defamation action that compensation is to be measured having regard to aggravating circumstances the result of the defendant’s conduct might not result in a verdict different from that which they would return if they were told that because of that conduct they could give damages by way of example.”

 

What is certain is that the summing-up by Devlin J. in that case could not, as Lord Devlin himself surmised, now survive the analysis by Lord Devlin in Rookes v. Barnard of the theoretical basis of exemplary damages in the sense in which the term should now be employed.

 

The “considerations”

 

I turn now to Lord Devlin’s three “considerations.” It is worth pointing out that neither the Court of Appeal nor any of the counsel who appeared before us attacked these as such. Nor, so far as I am aware, have these been attacked in the cases in which Commonwealth judges have felt constrained to criticise Rookes v. Barnard [1964] A.C. 1129. This alone would be a good reason against a simple return to the status quo ante proposed by the Court of Appeal, because the first and second “considerations” coupled with the passage from which I have already quoted on p. 1225 are themselves, and quite independently of the “categories,” an important, and I think original, contribution to the law on exemplary damages. Whilst, as I have indicated, I cannot myself follow what Lord Devlin says on the second category so far as regards the right of appellate courts to interfere with jury awards on principles different from the traditional nor, I think, with the proposal that Benham v. Gambling [1941] A.C. 157 offers a precedent for arbitrary limits imposed by the judiciary in defamation cases, I regard it as extremely important that, for the future, judges should make sure in their direction to juries that the jury is fully aware of the danger of an excessive award. A judge should first rule whether evidence exists which entitles a jury to find facts bringing a case within the relevant categories, and, if it does not, the question of exemplary damages should be withdrawn from the jury’s consideration. Even if it is not withdrawn from the jury, the judge’s task is not complete. He should remind the jury: (i) that the burden of proof rests on the plaintiff to establish the facts necessary to bring the case within the categories. (ii) That the mere fact that the case falls within the  [*1082]  categories does not of itself entitle the jury to award damages purely exemplary in character. They can and should award nothing unless (iii) they are satisfied that the punitive or exemplary element is not sufficiently met within the figure which they have arrived at for the plaintiff’s solatium in the sense I have explained and (iv) that, in assessing the total sum which the defendant should pay, the total figure awarded should be in substitution for and not in addition to the smaller figure which would have been treated as adequate solatium, that is to say, should be a round sum larger than the latter and satisfying the jury’s idea of what the defendant ought to pay. (v) I would also deprecate, as did Lord Atkin in Ley v. Hamilton, 153 L.T. 384 the use of the word “fine” in connection with the punitive or exemplary element in damages, where it is appropriate. Damages remain a civil, not a criminal, remedy, even where an exemplary award is appropriate, and juries should not be encouraged to lose sight of the fact that in making such an award they are putting money into a plaintiff’s pocket, and not contributing to the rates, or to the revenues of central government.

 

If this be correct, the agreed list of questions submitted to the jury in the present case is not the ideal procedure for ensuring that the jury keep their verdict within bounds. They should normally be asked to award a single sum whether as solatium or as exemplary damages. If, in order to avoid a second trial, they are asked a second question, they should be asked, in the event of their awarding exemplary damages, what smaller sum they would have awarded if they had confined themselves to solatium in the sense explained.

 

It follows from what I have said that I am not prepared to follow the Court of Appeal in its criticisms of Rookes v. Barnard, which I regard as having imposed valuable limits on the doctrine of exemplary damages as they had hitherto been understood in English law and clarified important questions which had previously been undiscussed or left confused. From one point of view, there is much to be said for the interpretation put upon Lord Devlin’s speech by Windeyer J. in Uren v. John Fairfax & Son Pty. Ltd., 117 C.L.R. 118, 152 immediately before the passage I have just quoted:

 

“What the House of Lords has now done is, as I read what was said, to produce a more distinct terminology. Limiting the scope of terms that often were not distinguished in application makes possible an apparently firm distinction between aggravated compensatory damages and exemplary or punitive damages.”

 

But it is not to be inferred from this that the ruling in Rookes v. Barnard is a pure question of semantics. It may well be true that in most individual cases the precise terminology in which the question is asked of the jury may not make much difference to the amount of the award. Both Windeyer J. in the passage just cited and Lord Devlin at p. 1230 were evidently of this view. But the following positive advantages can be gained from adhering to the rules he laid down, if properly interpreted: (1) The danger of double counting, of adding a pure “fine” to what has already been awarded as solatium without regarding the deterrent or punitive effect of the latter, has been eliminated, or at least reduced to a minimum. (2) In all cases where the categories do not apply, the jury  [*1083]  must be told to confine the punitive or deterrent element in their thinking within the limits of a fair solatium. In other words, to borrow the language, though not the sentiments, expressed in Forsdike v. Stone (1868) L.R. 3 C.P. 607, 611, the jury must be told to consider only what the plaintiff should receive after giving full allowance to the need to re-establish his reputation and for the outrage inflicted upon him, and not what the defendant should pay independently of this consideration. (3) In cases where the categories do apply, juries can be given directions a little more informative and regulatory than was the case up to and including the new analysis.

 

Rookes v. Barnard has not perhaps proved quite the definitive statement of the law which was hoped when it was decided. This is often the case. I remember with suitably mixed feelings of filial piety and inherited caution that in his judgment in Robert Addie & Sons (Collieries) Ltd. v. Dumbreck [1929] A.C. 358 my father believed he was putting a final end to doubts about the limits of occupiers’ liability to trespassers, licensees and invitees. But the way forward lies through a considered precedent and not backwards from it. I would hope very much that, in the light of observations made on Rookes v. Barnard in this case, Commonwealth courts might see fit to modify some of their criticisms of it. I do not know how far it can be of value in the United States of America where it seems to me that the decisions of the Supreme Court have been influenced greatly by the terms of the First Amendment to the Constitution, and by the unsatisfactory rules prevalent in American courts as to the recovery of costs. However that may be, we cannot depart from Rookes v. Barnard here. It was decided neither per incuriam nor ultra vires this House; we could only depart from it by tearing up the doctrine of precedent, and this was not the object of this House in assuming the powers adopted by the Practice Statement (Judicial Precedent) [1966] 1 W.L.R. 1234.

 

Lest I should have been thought to have forgotten it, I would observe that the Court of Appeal overruled the decision of Lawton J. that a claim for exemplary damages should be pleaded. I am content to accept their view on the basis of the present practice. But in the light of the decision of this House in the instant case I propose to refer to the Rule Committee the question whether in the light of Rookes v. Barnard and the present decision the present practice should not be altered. There is much to be said for the view that a defendant against whom a claim of this kind is made ought not to be taken by surprise.

 

My Lords, it follows from what I have said in my opinion this appeal should be dismissed and that costs should follow the event.

 

LORD REID. My Lords, the appellants published a book, “The Destruction of Convoy PQ17” which according to their advertisement on the dust jacket was the result of five intensive years of meticulous research by the author. It contained many statements about the conduct of the respondent who was the naval officer in command of the convoy. He sued the appellants and the author for damages for libel. After a trial which lasted for some 17 days a number of questions were left to the jury. They found that the words complained of were defamatory of the respondent and were not true in substance and in fact. They were  [*1084]  asked what compensatory damages they awarded, and they awarded £15,000. Then they were asked “Has the plaintiff proved that he is entitled to exemplary damages?” Their answer was yes against both defendants. Next they were asked “What additional sum should be awarded him by way of exemplary damages?” Their answer was £25,000. So judgment was entered against both defendants for £40,000.

 

Others of your Lordships have dealt in detail with these statements and I do not think it necessary to say more than that in my opinion the jury were well entitled to find that they conveyed imputations of the utmost gravity against the character and conduct of the respondent as a naval officer. Indeed the appellants do not now seek to disturb the award of £15,000 as “compensatory damages.” Their contention before your Lordships is twofold: first, that the jury were not entitled to award any exemplary damages and, secondly, that the amount awarded under this head was much too great. As no objection was taken at the time to the form of the question there cannot now be any objection to the jury having been asked in this case to consider separately compensatory and exemplary damages.

 

The whole matter of exemplary damages was dealt with in this House in Rookes v. Barnard [1964] A.C. 1129 in a speech by Lord Devlin with which all who sat with him, including myself, concurred. The Court of Appeal dealing with the present case held that if they applied the law as laid down in Rookes v. Barnard the appellants’ appeal must fail and the jury’s verdict must stand. They could have stopped there, but they chose to go on and attack the decision of this House as bad law. They were quite entitled to state their views and reasons for reaching that conclusion but very unfortunately Lord Denning M.R., apparently with the concurrence of his two colleagues, went on to say [1971] 2 Q.B. 354, 384:

 

“This case may, or may not, go on appeal to the House of Lords. I must say a word, however, for the guidance of judges who will be trying cases in the meantime. I think the difficulties presented by Rookes v. Barnard are so great that the judges should direct the juries in accordance with the law as it was understood before Rookes v. Barnard. Any attempt to follow Rookes v. Barnard is bound to lead to confusion.”

 

It seems to me obvious that the Court of Appeal failed to understand Lord Devlin’s speech, but, whether they did or not, I would have expected them to know that they had no power to give any such direction and to realise the impossible position in which they were seeking to put those judges in advising or directing them to disregard a decision of this House.

 

That aberration of the Court of Appeal has made it necessary to reexamine the whole subject and incidentally has greatly increased the expense to which the parties to this case have been put.

 

The very full argument which we have had in this case has not caused me to change the views which I held when Rookes v. Barnard was decided or to disagree with any of Lord Devlin’s main conclusions. But it has convinced me that I and my colleagues made a mistake in simply concurring with Lord Devlin’s speech. With the passage of time I have come more and more firmly to the conclusion that it is never wise to have only one speech in this House dealing with an important question of law.  [*1085] 

 

My main reason is that experience has shown that those who have to apply the decision to other cases and still more those who wish to criticise it seem to find it difficult to avoid treating sentences and phrases in a single speech as if they were provisions in an Act of Parliament. They do not seem to realise that it is not the function of noble and learned Lords or indeed of any judges to frame definitions or to lay down hard and fast rules. It is their function to enunciate principles and much that they say is intended to be illustrative or explanatory and not to be definitive. When there are two or more speeches they must be read together and then it is generally much easier to see what are the principles involved and what are merely illustrations of it.

 

I am bound to say that, in reading the various criticisms of Lord Devlin’s speech to which we have been referred, I have been very surprised at the failure of its critics to realise that it was intended to state principles and not to lay down rules. But I suppose that those of us who merely concurred with him ought to have foreseen that this might happen and to have taken steps to prevent it. So I shall try to repair my omission by stating now in a different way the principles which I, and I believe also Lord Devlin, had in mind. I do not think that he would have disagreed with any important part of what I am now about to say.

 

Damages for any tort are or ought to be fixed at a sum which will compensate the plaintiff, so far as money can do it, for all the injury which he has suffered. Where the injury is material and has been ascertained it is generally possible to assess damages with some precision. But that is not so where he has been caused mental distress or when his reputation has been attacked — where, to use the traditional phrase, he has been held up to hatred, ridicule or contempt. Not only is it impossible to ascertain how far other people’s minds have been affected, it is almost impossible to equate the damage to a sum of money. Any one person trying to fix a sum as compensation will probably find in his mind a wide bracket within which any sum could be regarded by him as not unreasonable — and different people will come to different conclusions. So in the end there will probably be a wide gap between the sum which on an objective view could be regarded as the least and the sum which could be regarded as the most to which the plaintiff is entitled as compensation.

 

It has long been recognised that in determining what sum within that bracket should be awarded, a jury, or other tribunal, is entitled to have regard to the conduct of the defendant. He may have behaved in a highhanded, malicious, insulting or oppressive manner in committing the tort or he or his counsel may at the trial have aggravated the injury by what they there said. That would justify going to the top of the bracket and awarding as damages the largest sum that could fairly be regarded as compensation.

 

Frequently in cases before Rookes v. Barnard, when damages were increased in that way but were still within the limit of what could properly be regarded as compensation to the plaintiff, it was said that punitive, vindictive or exemplary damages were being awarded. As a mere matter of language that was true enough. The defendant was being punished or an example was being made of him by making him pay more than he would have had to pay if his conduct had not been outrageous. But the  [*1086]  damages though called punitive were still truly compensatory: the plaintiff was not being given more than his due.

 

On the other hand when we came to examine the old cases we found a number which could not be explained in that way. The sums awarded as damages were more — sometimes much more — than could on any view be justified as compensatory, and courts, perhaps without fully realising what they were doing, appeared to have permitted damages to be measured not by what the plaintiff was fairly entitled to receive but by what the defendant ought to be made to pay as punishment for his outrageous conduct.

 

That meant that the plaintiff, by being given more than on any view could be justified as compensation, was being given a pure and undeserved windfall at the expense of the defendant, and that in so far as the defendant was being required to pay more than could possibly be regarded as compensation he was being subjected to pure punishment.

 

I thought and still think that that is highly anomalous. It is confusing the function of the civil law which is to compensate with the function of the criminal law which is to inflict deterrent and punitive penalties. Some objection has been taken to the use of the word “fine” to denote the amount by which punitive or exemplary damages exceed anything justly due to the plaintiff. In my view the word “fine” is an entirely accurate description of that part of any award which goes beyond anything justly due to the plaintiff and is purely punitive.

 

Those of us who sat in Rookes v. Barnard thought that the loose and confused use of words like punitive and exemplary and the failure to recognise the difference between damages which are compensatory and damages which go beyond that and are purely punitive had led to serious abuses, so we took what we thought was the best course open to us to limit those abuses.

 

Theoretically we might have held that as purely punitive damages had never been sanctioned by any decision of this House (as to which I shall say more later) there was no right under English law to award them. But that would have been going beyond the proper function of this House. There are many well-established doctrines of the law which have not been the subject of any decision by this House. We thought we had to recognise that it had become an established custom in certain classes of case to permit awards of damages which could not be justified as compensatory, and that that must remain the law. But we thought and I still think it well within the province of this House to say that that undesirable anomaly should not be permitted in any class of case where its use was not covered by authority.

 

In order to determine the classes of case in which this anomaly had become established it was of little use to look merely at the words which had been used by judges because, as I have said, words like punitive and- exemplary were often used with regard to damages which were truly compensatory. We had to take a broad view of the whole circumstances.

 

I must now deal with those parts of Lord Devlin’s speech which have given rise to difficulties. He set out two categories of cases which in our opinion comprised all or virtually all the reported cases in which it was clear that the court had approved of an award of a larger sum of damages  [*1087]  than could be justified as compensatory. Critics appear to have thought that he was inventing something new. That was not my understanding. We were confronted with an undesirable anomaly. We could not abolish it. We had to choose between confining it strictly to classes of cases where it was firmly established, although that produced an illogical result, or permitting it to be extended so as to produce a logical result. In my view it is better in such cases to be content with an illogical result than to allow any extension.

 

It will be seen that I do not agree with Lord Devlin’s view that in certain classes of case exemplary damages serve a useful purpose in vindicating the strength of the law. That view did not form an essential step in his argument. Concurrence with the speech of a colleague does not mean acceptance of every word which he has said. If it did there would be far fewer concurrences than there are. So I did not regard disagreement on this side issue as preventing me from giving my concurrence.

 

I think that the objections to allowing juries to go beyond compensatory damages are overwhelming. To allow pure punishment in this way contravenes almost every principle which has been evolved for the protection of offenders. There is no definition of the offence except that the conduct punished must be oppressive, high-handed, malicious, wanton or its like — terms far too vague to be admitted to any criminal code worthy of the name. There is no limit to the punishment except that it must not be unreasonable. The punishment is not inflicted by a judge who has experience and at least tries not to be influenced by emotion: it is inflicted by a jury without experience of law or punishment and often swayed by considerations which every judge would put out of his mind. And there is no effective appeal against sentence. All that a reviewing court can do is to quash the jury’s decision if it thinks the punishment awarded is more than any twelve reasonable men could award. The court cannot substitute its own award. The punishment must then be decided by another jury and if they too award heavy punishment the court is virtually powerless. It is no excuse to say that we need not waste sympathy on people who behave outrageously. Are we wasting sympathy on vicious criminals when we insist on proper legal safeguards for them? The right to give punitive damages in certain cases is so firmly embedded in our law that only Parliament can remove it. But I must say that I am surprised by the enthusiasm of Lord Devlin’s critics in supporting this form of palm tree justice.

 

Lord Devlin’s first category is set out on p. 1226. He said:

 

“The first category is oppressive, arbitrary or unconstitutional action by the servants of the government. I should not extend this category — I say this with particular reference to the facts of this case — to oppressive action by private corporations or individuals.”

 

This distinction has been attacked on two grounds: first, that it only includes Crown servants and excludes others like the police who exercise governmental functions but are not Crown servants and, secondly, that it is illogical since both the harm to the plaintiff and the blameworthiness of the defendant may be at least equally great where the offender is a powerful private individual. With regard to the first I think that the  [*1088]  context shows that the category was never intended to be limited to Crown servants. The contrast is between “the government” and private individuals. Local government is as much government as national government, and the police and many other persons are exercising governmental functions. It was unnecessary in Rookes v. Barnard to define the exact limits of the category. I should certainly read it as extending to all those who by common law or statute are exercising functions of a governmental character.

 

The second criticism is I think misconceived. I freely admit that the distinction is illogical. The real reason for the distinction was, in my view, that the cases showed that it was firmly established with regard to servants of “the government” that damages could be awarded against them beyond any sum justified as compensation, whereas there was no case except one that was overruled where damages had been awarded against a private bully or oppressor to an amount that could not fairly be regarded as compensatory, giving to that word the meaning which I have already discussed. I thought that this House was therefore free to say that no more than that was to be awarded in future.

 

We are particularly concerned in the present case with the second category. With the benefit of hindsight I think I can say without disrespect to Lord Devlin that it is not happily phrased. But I think the meaning is clear enough. An ill disposed person could not infrequently deliberately commit a tort in contumelious disregard of another’s rights in order to obtain an advantage which would outweigh any compensatory damages likely to be obtained by his victim. Such a case is within this category. But then it is said, suppose he commits the tort not for gain but simply out of malice, why should he not also be punished? Again I freely admit there is no logical reason. The reason for excluding such a case from the category is simply that firmly established authority required us to accept this category however little we might like it, but did not require us to go farther. If logic is to be preferred to the desirability of cutting down the scope for punitive damages to the greatest extent that will not conflict with established authority then this category must be widened. But as I have already said I would, logic or no logic, refuse to extend the right to inflict exemplary damages to any class of case which is not already clearly covered by authority. On that basis I support this category.

 

In my opinion, the conduct of both defendants in this case was such that the jury were clearly entitled, if properly directed, to hold that it brought them within the second category. Again, I do not intend to cover ground already covered by my noble and learned friends. So I say no more than that the jury were fully entitled to hold that the appellants knew when they committed this tort that passages in this book were highly defamatory of the respondent and could not be justified as true and that it could properly be inferred that they thought that it would pay them to publish the book and risk the consequences of any action the respondent might take. It matters not whether they thought that they could escape with moderate damages or that the enormous expense involved in fighting an action of this kind would prevent the respondent from pressing his claim.

 

It was argued that to allow punitive damages in this case would hamper  [*1089]  other publishers or limit their freedom to conduct their business because it can always be inferred that publishers publish any book because they expect to profit from it. But punitive damages could not be given unless it was proved that they knew that passages in the book were libellous and could not be justified or at least deliberately shut their eyes to the truth. I would hope that no publisher would publish in such circumstances. There is no question of curtailing the freedom of a reputable publisher.

 

The next passage in Lord Devlin’s speech which has caused some difficulty is what has been called the “if, but only if,” paragraph on p. 1228. I see no difficulty in it but again I shall set out the substance of it in my own words. The difference between compensatory and punitive damages is that in assessing the former the jury or other tribunal must consider how much the plaintiff ought to receive, whereas in assessing the latter they must consider how much the defendant ought to pay. It can only cause confusion if they consider both questions at the same time. The only practical way to proceed is first to look at the case from the point of view of compensating the plaintiff. He must not only be compensated for proved actual loss but also for any injury to his feelings and for having had to suffer insults, indignities and the like. And where the defendant laws behaved outrageously very full compensation may be proper for that. So the tribunal will fix in their minds what sum would be proper as compensatory damages. Then if it has been determined that the case is a proper one for punitive damages the tribunal must turn its attention to the defendant and ask itself whether the sum which it has already fixed as compensatory damages is or is not adequate to serve the second purpose of punishment or deterrence. If they think that that sum is adequate for the second purpose as well as for the first they must not add anything to it. It is sufficient both as compensatory and as punitive damages. But if they think that sum is insufficient as a punishment then they must add to it enough to bring it up to a sum sufficient as punishment. The one thing which they must not do is to fix sums as compensatory and as punitive damages and add them together. They must realise that the compensatory damages are always part of the total punishment.

 

It was argued that the jury were not properly directed by the trial judge on this matter. I agree with your Lordships that that argument must fail. A judge’s direction to a jury is not to be considered in vacuo. It must be read in light of all the circumstances as they then existed and I cannot believe that the jury were left in any doubt as to how they must deal with this matter.

 

Next there are questions arising from the fact that there were two defendants. When dealing with compensatory damages the law is quite clear. There was one tort of which both defendants were guilty. So one sum is fixed as compensation, and judgment is given for that sum against both defendants, leaving it to the plaintiff to sue whichever he chooses and then leaving it to the defendant who has paid to recover a contribution if be can from the other.

 

But when we come to punitive damages the position is different. Although the tort was committed by both, only one may have been guilty of the outrageous conduct or, if two or more are so guilty they may be  [*1090]  guilty in different degrees or, owing to one being rich and another poor, punishment proper for the former may be too heavy for the latter.

 

Unless we are to abandon all pretence of justice, means must be found to prevent more being recovered by way of punitive damages from the least guilty than he ought to pay. We cannot rely on his being able to recover some contribution from the other. Suppose printer, author and publisher of a libel are all sued. The printer will probably be guiltless of any outrageous conduct but the others may deserve punishment beyond compensatory damages. If there has to be one judgment against all three, then it would be very wrong to allow any element of punitive damages at all to be included because very likely the printer would have to pay the whole and the others might not be worth suing for a contribution.

 

The only logical way to deal with the matter would be first to have a judgment against all the defendants for the compensatory damages and then to have a separate judgment against each of the defendants for such additional sum as he should pay as punitive damages. I would agree that that is impracticable. The fact that it is impracticable to do full justice appears to me to afford another illustration of how anomalous and indefensible is the whole doctrine of punitive damages. But, as I have said before, we must accept it and make the best we can of it.

 

So, in my opinion, the jury should be directed that, when they come to consider what if any addition is to be made to the compensatory damages by way of punitive damages, they must consider each defendant separately. If any one of the defendants does not deserve punishment or if the compensatory damages are in themselves sufficient punishment for any one of the defendants, then they must not make any addition to the compensatory damages. If each of the defendants deserves more punishment than is involved in payment of the compensatory damages then they must determine which deserves the least punishment and only add to the compensatory damages such additional sum as that defendant ought to pay by way of punishment.

 

I do not pretend that that achieves full justice but it is the best we can do without separate awards against each defendant.

 

It was argued that here again there was misdirection of the jury because all that was not made plain to them. But again I agree with your Lordships that in the whole circumstances we ought not to hold the direction of the learned trial judge to be inadequate. Again the jury can have been in no doubt as to what was required of them.

 

There remains what is perhaps the most dim cult question in this case — whether the additional award of £25,000 as punitive damages is so excessive that we can interfere. I think it was much too large, but that is not the test. I would like to be able to hold that the court has more control over an award of punitive damages than it has over an award of compensatory damages. As regards the latter it is quite clear that a court can only interfere if satisfied that no twelve reasonable men could have awarded so large a sum and the reason for that is plain. The court has no power to substitute its own assessment for the verdict of a jury. If it interferes it can only send the matter back to another jury. So before it can interfere it must be well satisfied that no other jury would award so large a sum. I do not see how this House could arrogate to itself any wider power with  [*1091]  regard to punitive damages. We could not deprive the plaintiff of his right to a new trial so we must adhere to the established test. Any diminution or abolition of the functions of a jury in libel cases can only come from Parliament. If this case brings nearer the day when Parliament does take action I for one shall not be sorry.

 

Whether or not we can interfere with this award is a matter which is not capable of much elaboration. In considering how far twelve reasonable men might go, acting as jurors commonly do act, one has to bear in mind how little guidance the court is entitled to give them. All they can be told is that they must not award a sum which is unreasonable. In answer to questions whether anything more definite could properly be said neither counsel in this case was able to make any suggestion and I have none to offer. The evidence in this case is such that the jury could take an extremely unfavorable view of the conduct of both defendants. I do not say that they ought to have done so, but they were entitled to do so. And they must have done so. I find it impossible to say that no jury of reasonable men, inexperienced but doing their best with virtually no guidance, could reach the sum of £25,000. Or, to put it in another way, I would feel no confidence that if the matter were submitted to another jury they must reach a substantially different result. So with considerable regret I must hold that it would be contrary to our existing law and practice if this House refused to uphold this verdict.

 

It is true that in this case the parties agreed that if the verdict for £25,000 were quashed they would leave it to this House to substitute another figure. But that agreement cannot justify us in doing other vise than we would have done if the parties had stood on their legal rights. The obvious reason for that agreement was a common desire to avoid the enormous expense of a new trial. This is not the first occasion on which I have felt bound to express my concern about the undue prolixity and expense of libel actions. I would not blame any individuals. It may arise from the conduct of a trial before a jury being more expensive than a trial before a judge. If so that is an additional argument for taking these cases away from juries. Or it may be that it suits wealthy publishers of newspapers, books and periodicals that the cost of fighting a libel action is so great that none but a person with large financial backing can sue them effectively. Whatever be the reason the costs of this case have already reached a figure which many laymen would call scandalous. I think that those in a position to take effective action might take note.

 

Finally, I must say something about a strange misconception which appears in the judgments of the Court of Appeal in this case. Somehow they reached the conclusion that the decision of this House in Rookes v. Barnard was made per incuriam was ultra vires, and had produced an unworkable position. It must be noted that in at least three earlier cases the Court of Appeal were able without difficulty or question to apply that decision: McCarey v. Associated Newspapers, Ltd. (No. 2) [1965] 2 Q.B. 86; Broadway Approvals Ltd. v. Odhams Press Ltd. (No. 2) [1965] 1 W.L.R. 805 and Fielding v. Variety Incorporated [1967] 2 Q.B. 841. What has caused their change of mind does not appear but I must deal with their new view. As regards the present position being unworkable, of course many difficulties remain in this branch of the law, but these difficulties are an  [*1092]  inheritance from the confusion of the past. I have dealt fairly fully with the proper interpretation of Rookes v. Barnard [1964] A.C. 1129 and it appears to me that that decision removes many old difficulties and creates few if any new ones.

 

I need not deal separately with the novel idea that a decision of this House can be ultra vires because that charge appears to be consequential on the charge that this House acted per incuriam in reaching its decision. It is perfectly legitimate to think and say that we were wrong but how anyone could say we acted per incuriam in face of the passage on page 1230 I fail to understand.

 

This charge is really based on what appears to me to be a misreading by the Court of Appeal of two decisions of this House, E. Hulton & Co. v. Jones [1910] A.C. 20 and Ley v. Hamilton, 153 L.T. 384. Hulton’s case has always been regarded as the leading authority for the proposition that a defamatory description intended to apply to a fictional person may in fact be a libel on a real person and therefore a subject for damages. I see nothing in the speeches in this House to indicate that punitive damages in the modern sense were being considered. It was said that there was an element of recklessness in the failure of the defendants to realise that there was a real Artemus Jones and that this justified a rather high sum of damages but I see nothing to indicate any view that the damages went beyond anything that could be justified as compensation and could only be justified as being punitive in the modern sense.

 

Ley v. Hamilton requires rather fuller consideration. But again I see nothing to indicate that this House held that the damages went beyond compensation or that there had been outrageous conduct justifying a punitive award which went beyond compensation. The majority in the Court of Appeal certainly held that the £5,000 damages award was punitive in the modern sense. They held that the real damage was trifling and the rest punishment. Greer L.J. said (1934) 151 L.T. 360, 369 that if Hamilton had been prosecuted for criminal libel it was inconceivable that he would have been fined £5,000. Maugham L.J. said, at p. 374, that the damages could not be described as a fair and reasonable compensation but were in the nature of a fine.

 

In this House only Lord Atkin delivered a speech. I read it as intended to show that elements properly included in compensatory damages were far wider than the majority in the Court of Appeal had thought and that the whole of this £5,000 was in fact justified as being compensatory. He said, 153 L.T. 384, 386:

 

“The fact is that the criticism with great respect seems based upon an incorrect view of the assessment of damages for defamation. They are not arrived at as the Lord Justice seems to assume by determining the ‘real’ damage and adding to that a sum by way of vindictive or punitive damages. It is precisely because the ‘real’ damage cannot be ascertained and established that the damages are at large. It is impossible to track the scandal, to know what quarters the poison may reach: it is impossible to weigh at all closely the compensation which will recompense a man or a woman for the insult offered or the pain of a false accusation. No doubt in newspaper libels juries take into account the vast circulations which are justly claimed in present  [*1093]  times. The ‘punitive’ element is not something which is or can be added to some known factor which is non-punitive. In particular it appears to present no analogy to punishment by fine for the criminal offence of publishing a defamatory libel.”

 

By saying that compensation for insult or the pain of a false accusation cannot be weighed at all closely and that there was nothing here analogous to punishment by fine, he was to my mind making it as clear as words can make it that the whole of this £5,000 was truly compensatory in character. So I think that Lord Devlin was perfectly right in saying that there is no decision of this House which recognises punitive damages in the modern sense of something which goes beyond compensation. Where the Court of Appeal went wrong was in failing to realise that in the older cases damages were frequently referred to as exemplary or punitive although they were in reality compensatory.

 

On the whole matter I would dismiss this appeal.

 

LORD MORRIS OF BORTH-Y-GEST. My Lords, at the trial of this action questions arose as to whether if the plaintiff succeeded, he was entitled to recover exemplary damages in addition to compensatory damages. The law relating to exemplary damages was considered in your Lordships’ House in 1964 and was laid down in the decision in Rookes v. Barnard [1964] A.C. 1129. That decision bound the learned judge. It bound the Court of Appeal. It continues to be binding authority in all courts unless and until it appears to your Lordships to be right to depart from it.

 

In presiding at the trial the learned judge set himself loyally and faithfully to follow the binding authority of the decision. His directions to the jury followed the approach laid down in the decision though it is contended that in regard to one or two matters there was faulty exposition which was sufficiently serious to vitiate the award made by the jury of exemplary damages. These matters call for separate consideration. If the contentions concerning them do not succeed there remains an issue as to whether the award of the jury was excessive and should be set aside. If it is held that there was nothing amiss at the trial and that the law as laid down in your Lordships’ House was properly applied by the learned judge it would be an unhappy conclusion if it were now held that the trial had in fact been conducted on wrong or at least on unnecessary lines but that this had only been so because the law which had to be followed had been wrongly laid down. If that were the conclusion it is by no means certain that it would be possible to avoid ordering a new trial which would then be conducted on the basis of the law as newly laid down. But a result so lamentable (and for the parties so calamitous) must be contemplated as at least a possibility if it is decided that the law was wrongly declared in 1964 and must now be changed or changed back again.

 

Before considering this aspect of the matter further I must express my view in regard to the main contentions which are raised by the appellants. They for their part do not in any way question the validity of Rookes v. Barnard. Their appeal relates only to the award of exemplary damages. The jury found that the words complained of in the hard back edition were defamatory of the plaintiff and that the words were not true in substance  [*1094]  or in fact. They found similarly in regard to the proof copies. They awarded compensatory sums respectively of £14,000 and £1,000. No challenge as to such results is made. No criticism is advanced in regard to the very careful summing-up of the learned judge dealing with the facts and with the issues as to liability. No suggestion is made that the awards of compensation can be attacked as being excessive or unreasonable.

 

The learned judge left three questions to the jury on the issue of exemplary damages. First they were asked whether the plaintiff had proved that he was entitled to exemplary damages. Here the learned judge was carefully following Rookes v. Barnard. There may be exemplary damages if a defendant has formed and been guided by the view that, though he may have to pay some damages or compensation because of what he intends to do, yet he will in some way gain (for the category is not confined to moneymaking in the strict sense) or may make money out of it, to an extent which he hopes and expects will be worth his while. I do not think that the word “calculated” was used to denote some precise balancing process. The situation contemplated is where someone faces up to the possibility of having to pay damages for doing something which may be held to have been wrong but where nevertheless he deliberately carries out his plan because he thinks that it will work out satisfactorily for him. He is prepared to hurt somebody because he thinks that he may well gain by so doing even allowing for the risk that he may be made to pay damages. As the learned judge put it in reference to defamation there may be exemplary damages in cases where someone wilfully or knowingly or recklessly peddles untruths for profit. There must be evidence fit to be left to the jury but, if there is, then it is for the jury to decide whether there is entitlement to exemplary damages on the basis to which I have referred.

 

It was contended on behalf of the appellants that there was no evidence fit to be left to the jury in this case on this issue. In my view this contention wholly fails. There was ample evidence. It was painstakingly recounted in the summing up of the learned judge. It is helpfully referred to and summarised in the judgment of the learned Master of the Rolls. It is reviewed in the speech of Lord Hailsham of St. Marylebone L.C. which I have had the advantage of reading in advance.

 

Similar considerations apply to the question which was put to the jury and which they answered by saying that entitlement to exemplary damages was proved against both defendants.

 

It is in regard to the next question and answer that the greatest doubts and difficulties in my view arise. Being asked — What additional sum should be awarded him by way of exemplary damages? the answer of the jury was £25,000. So there were three awards: one being (for the hard back edition) the compensatory figure of £14,000: another being the exemplary damages figure of £25,000. For the total of £40,000 judgment was entered.

 

I must confess that for my part I should greatly regret it if the practice became general of having a separate award of exemplary damages in this manner (I will return to this question later). But the learned judge was only following the guidance specifically given in Rookes v. Barnard. There it was said, at p. 1228, that the fact that the two sorts of damage differ essentially does not necessarily mean that there should be two awards.  [*1095] 

 

But it was said that there may be cases in which it is difficult for a judge to say whether he ought or ought not to leave a claim for exemplary damages to the jury. I can quite see that in such a case it will be easier for an appellate court (where an issue is raised whether there was evidence which could justify an award of exemplary damages) if there are two awards. The award of exemplary damages could be set aside without the necessity for a new trial if the appellate court considered that the evidence was not such as to have been fit for the consideration of the jury so as to entitle them to award exemplary damages. For this reason it was stated in Rookes v. Barnard, at p. 1228, that if a judge is in doubt whether he ought to leave a claim for exemplary damages to a jury, then he could invite them to say “what sum they would fix as compensation and what additional sum, if any, they would award if they were entitled to give exemplary damages.” It was this course that the learned judge followed in the present case. But if this course is followed the words “if any” become of importance. They were not included in the question which was put to the jury.

 

There are three very important issues which arise. (1) Did the learned judge give an adequate direction to the jury to ensure that they understood that they should only award an “additional” sum if they were satisfied that the amount they were awarding as compensatory damage was in itself not enough to punish the defendants? (2) Did the learned judge give an adequate direction to meet the situation where (as in this case) there are two defendants? (3) In any event is the sum of £40,000 excessive as an award of exemplary damages and a figure which no reasonable jury could award — with the result that though the purely compensatory part £15,000 is not challenged the award of an additional £25,000 must be set aside?

 

(1) The relevant sentences in the summing up have been referred to in the speech of Lord Hailsham of St. Marylebone L.C. and I need not set them out. I would have been happier if the direction on this point (which came towards the end of what I venture to think was a masterly review of the case) had been ampler and more explicit than it was. But the learned judge did emphasise the word “additional.” He asked the jury to underline it. He said that they should underline it because both the court and counsel would want to know “if you do decide to award punitive damages how much more do you award over and above the compensatory damage.” Even so it would have been better to have made it abundantly clear that the punitive element is not to be considered in isolation: an enforced obligation to pay a large sum by way of compensation has itself a punitive impact. So a jury ought fully to understand that only if a sum awarded as com pensation is inadequate as a punishment should any larger sum be awarded.

 

Much earlier in his summing-up the learned judge had dealt with this matter in an introductory way. He told the jury that they were being asked

 

“not only to give Captain Broome compensatory damages that is a reasonable sum for the injury to his reputation and the exacerbation of his feelings: but in addition to fine Cassells and Mr. Irving for having done what they have done. The money which you decide — if you do decide — to award by way of punitive damages will not go  [*1096]  into the national Exchequer. It will have to go into Captain Broome’s pocket.”

 

Here again there was an omission to emphasise that an award of compensation must always and inevitably be a part of the “fine” in cases where the imposition of a “fine” is warranted.

 

Though a study of the shorthand note of what was said has led me to the view that there should have been amplification in the way to which I have referred, the important question now is whether it should be held that the jury were misled with the result that their award cannot stand. The emphasis placed upon the word “additional” could not have been lost sight of by the jury. Additional to what? Quite clearly, additional to the amount of compensation awarded. The jury were asked “how much more” they would award. The “more” was to be “over and above” the compensation. It surely must have been clear to the jury that any “more” that they decided upon or any “additional” sum would have to be paid by those against whom they awarded it on top of the sum that they were first awarding. Here was a jury that listened to the case over a period of 17 days. They deliberated for nearly five hours. They awarded a sum of £25,000 to be “additional” to their award of £15,000. They knew that the total was £40,000. Thereafter they heard both counsel agree that there should be a single judgment for that amount. No suggestion was made (or I think could possibly have been made) that the £25,000 included the £15,000. I would find it difficult to accept that at the stage in their deliberations when they were considering whether Cassells and Mr. Irving should be punished by being made to pay money they should at that stage have left out of account one part of the money that they themselves were awarding. If, having decided that it was a case for punishment, the jury were considering the monetary sum which, as such punishment, should be paid, the point would surely have been raised by one member if not by all members of the jury: Are we not punishing them enough by saying that they must pay £15,000? They could have recorded that as their view had they entertained it. I am not prepared to assume that something which at that stage must really have been quite obvious was overlooked by the jury.

 

(2) There is nothing in regard to this question which I could usefully add to what Lord Hailsham of St. Marylebone has said in reviewing the authorities and in formulating his conclusion. I express my concurrence.

 

(3) The approach which should be followed by an appellate court in considering whether an award of damages made by a jury should be assailed on the ground that the sum awarded is excessive has been clearly defined in authoritative decisions. They are referred to in the speech of Lord Hailsham of St. Marylebone. I am bound to say that the figure of £40,000 appears to me to be a high figure. Certainly it must be a very unusual case in which on a correct application of the law as laid down in Rookes v. Barnard the amount which defendants must pay should so greatly exceed the amount which is reasonably to be received by the plaintiff by way of compensation. It is this disparity between the £40,000 and the £15,000 that has caused disquiet as to whether the jury may have been caused or allowed to be under a misunderstanding. But if the conclusion is reached  [*1097]  that the jury knew what they were about and chose their figures advisedly then I do not think that I ought to conclude that their “additional” figure of £25,000 was so high that no reasonable jury could award it. To translate injury to and attack upon reputation into monetary terms is at all times a difficult exercise. But it was the same jury that fixed the “additional” figure of £25,000 that also — without being impeached for so doing — fixed the compensatory figure of £15,000. If they did not go wide when fixing the latter why should it be determined that they went wide in fixing the former? The conclusion which I think can be drawn is that the jury took a very serious view of the conduct and attitude of the defendants. If, after hearing all the relevant features of the case probed and examined over a period of seventeen days and hearing the evidence of such of the parties as decided to call or give evidence, the jury did take a very serious view, there was evidence which entitled them to do so. They may have regarded the conduct and attitude of each of the defendants with equally sharp disfavor. If it was their considered collective view that the defamation was grave, and that publication was deliberately undertaken by those who had regard for their own advantage but none for the honour and renown of one whom they traduced, then the jury were warranted in deciding that such conduct should be heavily penalised. Whatever might have been my personal assessment had I been on the jury I have not been persuaded that it must be decided that the penalty imposed was beyond the limit to which a reasonable jury could go. Nor can it be said with any assurance that an estimation of a figure by a learned judge would necessarily have superior validity. A learned judge has experience and knowledge of other cases but in a matter so elusive as fixing in monetary terms a reflection of feelings of disapproval there is no norm. It may be difficult to give guidance but a judge should be able to express to a jury the same guidance as he would give to himself.

 

For the reasons which I have given I consider that the appeal should be dismissed. As I have indicated, the appellants in no way sought to impugn the decision in Rookes v. Barnard [1964] A.C. 1129. Such ardour in criticism as may have been evinced in the Court of Appeal by counsel for the respondent became tempered and modified by the reflection that an assault upon Rookes v. Barnard was not essential for his success in this appeal and that the overturning of Rookes v. Barnard might at least possibly, involve the jettisoning of all the proceedings to date and a complete new trial on a fresh basis. But as so much was said about Rookes v. Barnard and because in the printed case of the respondent the first reason set out was that your Lordships’ House should depart from its decision in Rookes v. Barnard (in so far as that decision altered the law on exemplary damages generally or at least in defamation cases) I must record my opinion.

 

In Rookes v. Barnard one submission that was made was that exemplary damages could not be awarded in that case. Other submissions led to a somewhat general consideration of the law relating to exemplary damages. The report of the arguments (from p. 1158 to p. 1164) shows that certain authorities and certain text books were referred to and were examined. There were citations of some thirty cases. In the result the House examined and reviewed the law and came to certain conclusions.  [*1098]  The House was not bound to limit those conclusions within any formulation which counsel had thought fit to formulate.

 

It would be idle to deny that a very considerable pruning operation was decided upon. It may be that there are some who would not have pruned so much and so drastically. It may be that there are some who would have pruned more severely. What was done was done in the hope of removing from the law “a source of confusion between aggravated and exemplary damages.” It may be that there are some who feel that though the previous law (built up, as the common law is, as a result of particular decisions given in particular sets of circumstances) was in very many respects imprecise and even illogical, yet it was somehow found in practice to work and to be no serious cause of confusion. It may be that there are some who consider that manifest variations and divergencies in terminology did not reflect any really fundamental differences of approach: that, for example, when in Mediana (Owners) v. Comet (Owners) [1900] A.C. 113, 118 Lord Halsbury L.C. made a reference, though only a passing and incidental one, to punitive damages:

 

“I put aside cases of trespass where a high-handed procedure or insolent behavior has been held in law to be a subject of aggravated damages, and the jury might give what are called punitive damages”

 

he had much the same conception in mind as had Lord Atkinson when in Addis v. Gramophone Co. Ltd. [1909] A.C. 488, 496, he made an incidental reference to circumstances of malice, fraud, defamation or violence which would sustain an action of tort in which a person might no doubt “recover exemplary damages, or what is sometimes styled vindictive damages”; or as had Lord Loreburn L.C. when he spoke in E. Hulton & Co. v. Jones [1910] A.C. 20, 25:

 

“In the second place the jury were entitled to say this kind of article is to be condemned. There is no tribunal more fitted to decide in regard to publications, especially publications in the newspaper press, whether they bear a stamp and character which ought to enlist sympathy and to secure protection. If they think that the licence is not fairly used and that the tone and style of the libel is reprehensible and ought to be checked, it is for the jury to say so;”

 

But, even if some of the thoughts above referred to are in fact entertained, do they give warrant for re-opening now the debate that led to the decision in Rookes v. Barnard? I do not think so. I do not think that the power that was referred to in the statement of July 26, 1966, [1966] 1 W.L.R. 1234 was intended to encourage a tendency periodically to chop and change the law. In branches of the law where clarification becomes necessary there may well be decisions which as a matter of policy are not universally welcome or where some may think that some variant of the decision one way or the other would have been more acceptable. But this does not mean that decisions of this House should readily be reviewed whenever a case presents itself which is covered by a decision. There must be something much more.

 

In his book Principles of the Law of Damages (1962) Professor Street poses the question whether awards of exemplary damages are ever justified.  [*1099] 

 

He outlines seven arguments against them and with mathematical impartiality seven arguments in their favour concluding that one cannot say whether or not exemplary damages are desirable. Whatever general views may be entertained or whatever inclination there may be in different personal views, I see no advantage in refusing at this juncture to recognise that a deliberate pronouncement was made in Rookes v. Barnard.

 

Though I consider that no reason has been shown for denying to that pronouncement the authority of a decision of this House it is not inconsistent with this approach to express the hope that a necessity for a separate and isolated assessment of exemplary damages will be rare. In the search for authority only one case was found prior to Rookes v. Barnard in which there was such a result. That was Loudon v. Ryder [1953] 2 Q.B. 202 now overruled. The present case is, I think, the first one subsequent to Rookes v. Barnard in which such a separate award has actually been made.

 

In the older cases the “vindictive” or “exemplary” or “punitive” aspect merely became one element in a composite whole. Thus the law as it was in 1877 was summarised in the 3rd edition of Mayne’s Treatise on Damages. He pointed (see p. 37) to the difference between damages in cases of contract (where they were only a compensation) and in cases of tort. In the latter “when there were no circumstances of aggravation they are generally the same.” But he said at p. 37:

 

“Where the injury is to the person, or character, or feelings, and the facts disclose fraud, malice, violence, cruelty, or the like, they operate as a punishment, for the benefit of the community, and as a restraint to the transgressor.”

 

In the various cases cited (see pp. 36, 37, 514, 516) one amount only of damages was assessed. For a later general summary of the law (as it was in 1895) reference may be made to Sir Frederick Pollock’s 4th edition of The Law of Torts. He refers (see p. 174) to cases where there is great injury without the possibility of measuring compensation by any numerical rule. In such cases he said:

 

“... juries have been not only allowed but encouraged to give damages that express indignation at the defendant’s wrong rather than a value set upon the plaintiff’s loss. Damages awarded on this principle are called exemplary or vindictive.” He went on to explain that: “The kind of wrongs to which they are applicable are those which, besides the violation of a right or the actual damage, import insult or outrage,”

 

The cases cited, to which I need not refer in detail, again appear to me to be cases in which only one figure of damages was assessed.

 

When juries came to award damages in such cases of tort they did there fore give and indeed were “encouraged” to give a sum which marked displeasure or indignation or which was to serve as a deterrent or as an example or which vindicated the law or which was a way of punishing the defendant. But juries were not invited to isolate such element as was purely punitive. I do not expect that they did in practice. In some cases their displeasure or indignation would operate as a kind of topping-up process. But if the process by which they had arrived at a figure could have been analysed (which normally it could not have been), while it would probably have been  [*1100]  found that there had been nothing in the nature of a mathematical addition of separate sums, yet it would have been recognised that some (wholly unascertainable) part of the whole must have been purely punitive. Stated otherwise such (unascertained) part was a fine. Logical analysis forces the conclusion therefore that in the result there would in a civil action have been punishment for conduct not particularised in any criminal code and that such punishment had taken the form of a fine not receivable by the state but as a sort of bonus by a private individual who would, apart from it, be solaced for the wrong done to him. There may be much to be said for making it permissible in a criminal court to order in certain cases that a convicted person should pay compensation. There is much to be said against a system under which a fine becomes payable in a civil court without any of the safeguards which protect those charged with crimes. If therefore the working of the law before Rookes v. Barnard is exposed to a relentless logical examination it has to be conceded that some features of it were not in principle acceptable. Yet it may be that no serious injustice resulted. And indeed as we have been told the life of the law often lies not in logic but in experience. It would however be an unfortunate and bizarre result if a wholly laudable attempt to rationalise the law had brought it about that the element which it was most sought to suppress was so brought into sharp relief that it attained a significance never before exhibited.

 

I would regard the present case as exceptional in the sense that the jury must have considered that the conduct of the defendants merited very special condemnation. In other than an exceptional case where exemplary damages are to be awarded I would hope that a jury would be unlikely to award a total sum which exceeded its purely compensatory component element to an extent in any way comparable to that which is revealed in the present case.

 

I would dismiss the appeal.

 

VISCOUNT DILHORNE. My Lords, the main issues to be determined in this appeal are (1) whether what was said by my noble and learned friend, Lord Devlin, in Rookes v. Barnard [1964] A.C. 1129 with regard to exemplary damages, and with which all the other members of the House then sitting agreed, correctly states the law: (2) if it does, whether Lawton J. erred in leaving the question of exemplary damages to the jury: (3) having left it to them, whether he misdirected them with regard thereto: and (4) whether the sum of £40,000 awarded by them, of which £25,000 was exemplary damages, was so excessive that that verdict cannot be allowed to stand.

 

I propose to consider the first of these questions last. Although Rookes v. Barnard was not concerned with damages for libel, I consider the other questions on the assumption that what was said in that case is not to be regarded as obiter in relation to libel cases and is to be regarded as binding on all inferior courts.

 

Lord Devlin expressed the view that there were only three categories of cases in which exemplary damages could be awarded, namely: (1) where there had been oppressive, arbitrary or unconstitutional action by servants of the government; (2) where the defendant’s conduct had been calculated by him to make a profit for himself which might well exceed  [*1101]  rhe compensation payable to the plaintiff; and (3) where exemplary damages are expressly authorised by statute.

 

The appellants contended that this case did not come within the second category. They called no evidence at the trial and the question whether it should have been left to the jury to consider exemplary damages, depends on whether there was evidence given or adduced on behalf of the plaintiff on which the jury were entitled to infer and conclude that the defendants’ conduct was of that character.

 

I do not think that Lord Devlin ever envisaged that, to bring a case within the second category, the plaintiff would have to show that there had been something in the nature of a mathematical calculation by the defendant, an assessment of the profit likely to ensue from the publication of defamatory matter and an estimation of the risk of being sued and the damages likely to be awarded if an action was brought. If a plaintiff had to prove that, it would be seldom that he would be in a position to do so.

 

Newspapers and books are usually published for profit and that fact does not by itself make the publisher liable to pay exemplary damages.

 

I think that Widgery J., as he then was, was right when he said in Manson v. Associated Newspapers Ltd. [1965] 1 W.L.R. 1038, 1040-1041:

 

“... it is perfectly clear, from those authorities,” McCarey v. Associated Newspapers Ltd. [1965] 2 Q.B. 86 and Broadway Approvals Ltd. v. Odhams Press Ltd. (No. 2) [1965] 1 W.L.R. 805 “that in a case in which a newspaper quite deliberately publishes a statement which it either knows to be false or which it publishes recklessly, careless whether it be true or false, and on the calculated basis that any damages likely to be paid as a result of litigation will be less than the profits which the publication of that matter will give, then Lord Devlin’s conditions are satisfied and exemplary damages are permissible.”

 

He went on to say at p. 1041 that he proposed to tell the jury that they could consider exemplary damages:

 

“... if, having considered what material there is before them, they are driven to the inference that this was an article published by the defendants conscious of the fact that it had no solid foundation and with the cynical and calculated intention to use it for what it was worth, on the footing that it would produce more profit than any possible penalty in damages was likely to be.”

 

I think too that Lawton J. put the matter correctly when he said in the course of his summing-up: “A man is liable to pay damages on a punitive basis if he wilfully and knowingly, or recklessly peddles untruths for profit.”

 

In my opinion, there was ample evidence on which the jury was entitled to come to the conclusion that the case came within the second category. On December 9, 1966, Mr. Irving the author, sent the manuscript of the book to Cassells with a letter in which he said that Captain Broome had threatened legal action if the manuscript was published, and on December 23 he sent them a long letter in which he quoted an extract from a letter he had received from Kimbers the publishers to whom he had first submitted  [*1102]  the manuscript. That extract stated: “... if the book goes to a legal man as it is, he could only tell you that half is libellous. We could not possibly publish the book as it is ...”

 

The manuscript submitted to Cassells was identical with that which Kimbers had seen. Perusal of it by any intelligent publisher must, even without the advantage of having the views of another publisher, have led to the conclusion that it contained many very grave and serious libels on Captain Broome and the jury were fully entitled to conclude that Cassells realised this. Mr. Kimber gave evidence that about March 8, 1967, he had telephoned Mr. Parker, a director of Cassells and told him that they had had one or two threats of libel actions if they published the book; to which Mr. Parker’s response was: “In that case we will tighten up the indemnity clause in Mr. Irving’s agreement.”

 

On December 27, 1967, Captain Broome wrote to Cassells saying that the manuscript was “unquestionably libellous.” They replied saying that in the light of his comments “drastic revisions” had been made. In fact, as Cassells must have known, the revisions that were made did not materially affect the passages defamatory of Captain Broome. On February 16, 1968, the business director of Cassells circulated a memorandum in the following terms, to all concerned:

 

“It is anticipated that early copies of The Destruction of Convoy PQ17 will start coming into the house on March 5. Will you please note that absolutely and positively not one single copy, on any pretext whatsoever, is to be removed from the house without reference to me.

 

Mr. Mitchell: Would you please notify the printer that this book is to be treated on a maximum security basis and ensure that not one single copy slips through their net.”

 

Shortly thereafter Cassells circulated proof copies of the book. Why they did so after the circulation of this memorandum is not known for no evidence was given for them. In the absence of any explanation the jury were, in my view, entitled to draw the inference that they had decided to publish the book, despite Captain Broome’s threats of action, knowing that passages in the book were libellous of Captain Broome and not caring whether those passages were true or false and on the footing that it was worth their while to run the risk of an action being brought by him and of his obtaining damages in order to make a profit on the book.

 

On March 5, 1967, Captain Broome issued a writ for libel. On April 29, 1968, his statement of claim was delivered. Cassells then knew, if they were in any doubt before, of what passages he was complaining. On June 14, 1968, they delivered their defence. They pleaded that the words complained of were true in substance and in fact in their natural and ordinary meaning. They did not seek to justify the meaning which the statement of claim alleged the words complained of bore, inter alia, that Captain Broome had been disobedient, careless, incompetent, indifferent to the fate of the merchant ships and had been largely responsible for or contributed extensively to the loss of two-thirds of the ships of the convoy.

 

Despite the issue of this writ, Cassells went on and published a hard back  [*1103]  edition of the book. That led to another writ being issued by Captain Broome.

 

Again in their defence to this statement of claim Cassells pleaded that the words complained of were in their natural and ordinary meaning true in substance and in fact but did not seek to justify the meanings which in the statement of claim it was alleged they bore. The jury by their verdict rejected the plea of justification and must have accepted that the passages complained of bore the meanings alleged by the plaintiff.

 

I do not propose to set out what those passages were. Suffice it to say that they clearly alleged that Captain Broome had been disobedient, careless, incompetent, indifferent to the fate of the merchant ships, that he had wrongly withdrawn his destroyer force from the convoy, that he had taken it closer to the German airfields than he had been ordered to do and that he had been responsible for the loss of two-thirds of the ships in the convoy. He was in fact accused of cowardice.

 

That Cassells did not appreciate that the passages complained of could be understood to have these meanings, is hard to accept. Yet after publication of the proof copies, after receipt of the writ and the statement of claim in respect of that publication, and when they knew the meanings which it was alleged the passages bore, they went on and published the hard back edition, and at the trial persisted in their plea of justification.

 

In these circumstances if Lawton J. had ruled at the end of the plaintiff’s case, as he was asked to do, that there was no evidence from which the injury could infer that the case came within the second category, he would in my opinion have erred. I therefore reject this contention of the appellants.

 

After specifying the three categories of cases in which in his view exemplary damages might be awarded, Lord Devlin in Rookes v. Barnard said that there were three considerations which must always be borne in mind and then went on to say [1964] A.C. 1129, 1228:

 

“In a case in which exemplary damages are appropriate, a jury should be directed that if, but only if, the sum which they have in mind to award as compensation (which may, of course, be a sum aggravated by the way in which the defendant has behaved to the plaintiff) is inadequate to punish him for his outrageous conduct, to mark their disapproval of such conduct and to deter him from repeating it, then it can award some larger sum.”

 

Complaint is made that Lawton J. gave no such direction to the jury. With the agreement of counsel, he asked them to answer seven questions. The first was whether, in respect of the hard back edition, the words complained of were defamatory of the plaintiff; the second, were they true in substance and in fact. Their answer to the first question was, Yes and to the second, No. The third question was: “What compensatory damages do you award the plaintiff?” Their answer was £14,000. Then in answer to the fourth and fifth questions they said that he was entitled to exemplary damages against both defendants. The sixth question was “What additional sum should be awarded him by way of exemplary damages?” Their answer was £25,000.

 

After the questions had been handed to the jury in the course of the summing-up, Lawton J. told them that, after considering what were the  [*1104]  compensatory damages if they found for the plaintiff, they should go on to consider whether he was entitled to exemplary damages. As to that, he told them to consider the case against each defendant separately, saying:

 

“In respect of each of them you will ask yourselves this question: ‘Has the plaintiff proved his entitlement against that defendant?’ If the answer is yes, then you will have to go on and assess how much punitive damages should be awarded.”

 

In the next paragraph of his summing-up, he repeated this, saying:

 

“You will have to ask yourselves: ‘Has he proved that he is entitled to punitive damages against Cassell & Co. Ltd.?’ If the answer is no, that is that. If the answer is yes, you will have to assess the damages.”

 

And then he asked the jury to underline the word “additional” in the sixth question as he and learned counsel wanted to know: “If you do decide to award punitive damages, how much more do you award over and above the compensatory damage?”

 

The jury were thus clearly told that if they found that the plaintiff was entitled to punitive damages, they must then assess what punitive damages should be awarded. They were never told that in considering whether any sum should be so awarded, they must have regard to the sum they awarded for compensatory damages, and if, and only if, that sum was inadequate to punish the defendants, should they add to it by awarding a sum for exemplary damages.

 

The failure to give such a direction, I regret that I cannot but regard as a most serious omission. It is one of the most important features of Lord Devlin’s speech that a direction on the lines he stated should be given. It was not, and instead the jury were told twice that, if they held that Captain Broome was entitled to exemplary damages, they must assess them. The jury’s verdict shows that they thought that £15,000 compensatory damages was insufficient, but if they had been told that they must, in assessing exemplary damages, take into account the sum awarded in compensation, it is possible that they would have awarded not £25,000 but only £10,000 as exemplary damages, that is to say, that they would have deducted from the £25,000 the £15,000 compensatory damages.

 

I regret having to come to this conclusion but I see no escape from it. After a trial lasting 17 days and lengthy hearings in the Court of Appeal and in this House, one feels some reluctance to say that the jury’s verdict should not stand. If all the counsel engaged in the case had told the jury that a sum should only be awarded for exemplary damages if the amount of the compensatory damages was insufficient punishment, then it might be possible to say that despite the omission in the summing-up, the jury can have been in no doubt as to what they were required to do. Unfortunately all counsel did not tell them that. One counsel told the jury in his final address that they must consider exemplary damages quite separately from compensatory damages. He told them: “... they are completely unconnected with each other and in no sense does the one head fall to be balanced against the other” and: “The two sums are so different that there is no propriety in any sense in balancing them up.”  [*1105] 

 

He thus indicated that account should not be taken of the amount of compensatory damages when deciding what, if any, sum should be awarded for exemplary damages. Counsel for Cassells did not refer to the matter but Mr. Colin Duncan in his final address for the defendant Irving read to the jury the “if, but only if,” passage of Lord Devlin’s speech.

 

As the case was presented to the jury, I can see no ground for the conclusion that they must, despite the omission in the summing-up, have been aware that they had to take into account the compensatory damages when deciding, if they held that there was entitlement to exemplary damages, what sum, if any, should be awarded on that account. On the contrary, the passages I have cited from the summing-up show that they were told that, if they found entitlement, they must then assess an amount for exemplary damages.

 

I have regretfully come to the conclusion that in consequence of this omission, the verdict should not be upheld.

 

Another criticism made of the summing-up was that the jury were not told on what basis they should assess the exemplary damages if they found that the plaintiff was entitled to them from both defendants and if, in their opinion, the degree of guilt of the defendants differed.

 

In the Court of Appeal there was considerable divergence of view as to the proper direction to be given on this. While there is ample authority for the proposition that against joint tortfeasors there can only be one verdict and one judgment for a joint tort, there is not a great deal of authority on this question. Such as there is points to the conclusion that the plaintiff can only recover the amount which all the defendants should pay and that the amount to be awarded should not be increased to a sum thought adequate to punish the most guilty defendant: see Dawson v. M’Clelland [1899] 2 Ir.R. 486, 490 per Andrews J.: per Boyd J. at p. 493 and per FitzGibbon L.J., at p. 499; Smith v. Streatfeild [1913] 3 K.B. 764, 769 per Bankes J. and Gatley on Libel and Slander, 6th ed. (1967), para. 1389. If that were not the case an innocent party or a less guilty party might have to pay a sum far in excess of that which he ought to pay. The result of this conclusion appears to be that if three defendants are sued for writing, printing and publishing a libel, if the publisher and author are held liable to pay exemplary damages and the printer is not, the plaintiff will not be awarded exemplary damages and the publisher and author will avoid liability for such damages.

 

The summing-up contained this passage:

 

“... say, for example, you took the view that Mr. Irving was more to blame than Cassell & Co., or to be fair, you took the view that Cassell & Co., being an experienced firm of publishers were more to blame than this young man, Mr. Irving, should you make Cassell & Co. pay a larger sum by way of punitive damages than Mr. Irving? The answer to that is No. Whatever damages, if any, you decide should be awarded by way of punitive damages must be the same sum in respect of both Mr. Irving and Cassell & Co. Ltd., if you find them both liable to pay punitive damages.”

 

Later in response to an intervention by counsel, the judge made it clear  [*1106]  hat this did not mean awarding one sum against each defendant but one sum against both.

 

While it can be said that the direction on this might have been more clearly expressed, I think it suffices for this passage did indicate to the jury that they should award a sum which was appropriate to the less guilty of the two. It may, of course, be the case that the jury did not find that one was more guilty than the other.

 

I now turn to the question whether the damages awarded were so excessive that the verdict cannot be allowed to stand.

 

In Rookes v. Barnard [1964] A.C. 1129, 1228 Lord Devlin recognised that where there was entitlement to exemplary damages, that did not necessarily mean that there must be two awards though he expressed the view that where there was doubt about entitlement to such damages, to avoid the risk of a new trial, it might be convenient to have separate awards.

 

One consequence of there being two awards, one for compensatory damages and one for exemplary, is that the jury’s verdict is more open to attack. If £15,000 was sufficient to compensate the plaintiff for the injury indicted on him, what justification can there be for an award of a further £25,000 as exemplary damages?

 

Lawton J. very clearly told the jury that they were being asked to fine Cassells and Mr. Irving for what they had done. He told them that they were “really in the position of a judge or magistrate trying a criminal case” and that punitive damages “must be reasonable in all the circumstances.”

 

An appellate court should only interfere with a jury’s verdict as to damages if it is such as to show that the jury has failed to perform its duty: Mechanical and General Inventions Co. Ltd. v. Austin and Austin Motor Co. Ltd. [1935] A.C. 346 per Lord Wright at p. 375: Bocock v. Enfield Rolling Mills Ltd. [1954] 1 W.L.R. 1303: Scott v. Musial [1959] 2 Q.B. 429: Lewis v. Daily Telegraph Ltd. [1963] 1 Q.B. 340 and other cases. To be set aside, the verdict must be out of all proportion to the facts.

 

The award of £25,000 for exemplary damages, as a fine and despite the direction given by Lawton J. to which I have referred, in addition to the award of £15,000 compensatory damages is, in my opinion, out of all proportion to the facts and suffices to show that they failed to perform their duty. Their award was, in my view, far in excess of the most that twelve reasonable men could be expected to give. If they had appreciated that they had to take into account the compensatory damages, then as I have said perhaps they might have awarded an additional £10,000 as exemplary damages. I would myself have assessed a considerably lower figure. Perhaps, one does not know, they may have thought that the judge had power to set off one against the other. However that may be, I think that the highest figure that could have been awarded by a jury performing its duty for exemplary damages would have been £10,000 in which case judgment would have been given not for £40,000 but for £25,000.

 

On this ground, too, in my opinion the verdict cannot stand.

 

I turn now to the first question. Does Rookes v. Barnard [1964] A.C. 1129 correctly state the law with regard to exemplary damages?

 

The Court of Appeal held that it did not. It was said that it was a decision given per incuriam. The Court of Appeal refused to follow it and  [*1107]  judges were told to direct juries in accordance with the law as understood before that case.

 

Decisions of this House are binding on all inferior courts and must be followed by them. There are, I think, two grounds on which the Court of Appeal can justifiably refuse to follow what has been said in this House. The first is that what was said was obiter. While it might be argued that the observations made with regard to exemplary damages in so far as they related to libel actions were obiter as no question with regard to them arose in Rookes v. Barnard where the question was, could such damages be given for intimidation, the Court of Appeal did not base their action on this ground. The second is where there are two clearly inconsistent decisions of this House, and the Court of Appeal has then to choose which to follow. In the Court of Appeal it was asserted that what was said in Rookes v. Barnard was in conflict with two previous decisions of this House, E. Hulton & Co. v. Jones [1910] A.C. 20 and Ley v. Hamilton, 153 L.T. 384 but, as I read the judgments, the Court of Appeal did not proceed upon this ground.

 

To say that a decision of this House was given per incuriam is, to say the least, unusual and could be taken, though I cannot believe it was so intended, as of a somewhat offensive character. While I regret the use of this expression, I doubt if it was intended to mean more than that the questions involved deserved more consideration in relation, among other things, to libel actions. If that is what was meant, it is, I must confess, a view with which I have considerable sympathy.

 

As I understand the judicial functions of this House, although they involve applying well established principles to new situations, they do not involve adjusting the common law to what are thought to be the social norms of the time. They do not include bowing to the wind of change. We have to declare what the law is, not what we think it should be. If it is clearly established that in certain circumstances there is a right to exemplary damages, this House should not, when sitting judicially, and indeed, in my view, cannot properly abolish or restrict that right. This, indeed, was recognised by Lord Devlin when he said, at p. 1226, that it was not open to this House to “arrive at a determination that refused altogether to recognise the exemplary principle.” If the power to award such damages is to be abolished or restricted, that is the task of the legislature.

 

One criticism that can be made of Lord Devlin’s speech is that while recognising that a refusal altogether to recognise the exemplary principle was not possible, he nevertheless restricted the power to award such damages so that they ceased to be obtainable in cases where prior to Rookes v. Barnard they might have been given.

 

I agree with Lord Denning M.R. that the pre-Rookes v. Barnard law was well stated in Mayne and McGregor on Damages, 12th ed. (1961), p. 196, para. 207 where it is said that such damages can only be given

 

“where the conduct of the defendant merits punishment, which is only considered to be so where his conduct is wanton, as where it discloses fraud, malice, violence, cruelty, insolence or the like, or, as it is sometimes put, where he acts in contumelious disregard of the plaintiff’s rights.”  [*1108] 

 

A similar statement is to be found in Mayne on Damages, 11th ed. (1946), p. 41.

 

I do not think that this statement of the law is to be questioned because at p. 200 in paragraph 212 of the 12th edition it is said that: “it cannot be said that English law has committed itself finally and fully to exemplary damages,” (a view which conflicts with the opinion of Lord Devlin to which I have referred), “and many of the ... cases point to the rationale not of punishment of the defendant but of extra compensation for the plaintiff for the injury to his feelings and dignity. This is, of course, not exemplary damages at all. It is another head of non-pecuniary loss to the plaintiff.”

 

This passage in paragraph 212 did not appear in the earlier editions. I am not concerned with the rationale but with what was recognised to be the law before Rookes v. Barnard. And I am reinforced in my view by the fact that what was said in paragraph 207 appears to accord with Australian law. In this field there does not appear to have been any difference between Australian and English law prior to Rookes v. Barnard.

 

In Uren v. John Fairfax & Sons Pty. Ltd., 117 C.L.R. 118 the High Court of Australia refused to follow Rookes v. Barnard and held that exemplary damages might be awarded if it appears that the defendant’s conduct in committing the wrong exhibited a contumelious disregard of the plaintiff’s rights, McTiernan J. saying, at p. 122, that the law of exemplary damages was “compendiously stated” in the passage I have cited from Mayne and McGregor.

 

Lord Devlin’s first category “oppressive, arbitrary or unconstitutional action by servants of the government,” a category which he said he would not extend to oppressive action by private corporations or individuals, was subjected to serious criticism by Taylor J. in Uren v. John Fairfax & Sons Pty. Ltd., 117 C.L.R. 118. He pointed out that in none of the three old cases on which this category was apparently based, did the decisions turn on the fact that the defendants had acted for the government. Surely it is conduct, not status, that should determine liability. Power to award exemplary damages may be an anomaly, but I doubt whether it is beneficial to the law to seek to reduce the area of that anomaly at the price of creating other anomalies and illogicalities. Surely it is anomalous if a person guilty of oppressive conduct should only be liable to exemplary damages if a servant of the government. In these days there are others than the government who can be guilty of oppressive conduct. Why should they be treated differently? I can find nothing in the three cases to indicate that if the conduct complained of had been by persons other than servants of the government, liability to exemplary damages would have been acceded.

 

Just as the definition of this category might be said to have been obiter to the decision in Rookes v. Barnard [1964] A.C. 1129, so might consideration of it be regarded in this case. Nevertheless as Rookes v. Barnard has to be considered in this appeal in consequence of the action taken by the Court of Appeal, I feel I should express my opinion which is that this narrow definition does not appear to me to be justified by the authorities on which it was based.

 

It may also be contended that Lord Devlin’s second category is also too  [*1109]   narrowly drawn, for why should conduct lead to exemplary damages if inspired by the profit motive or some material interest, and similar conduct due to other motives not do so? But the substantial criticism that can be made is that by his categorisation, the previously existing and recognised power to award exemplary damages is restricted. Lord Devlin indeed appreciated the novelty of what he was doing when he said that acceptance of his views would “impose limits not hitherto expressed on such awards” (p. 1226). I do not think that this should have or could properly be done. It should have been left to the legislature.

 

This conclusion does not, however, mean that the jury’s verdict as to liability must be interfered with. It was urged that Cassells’ decision to call no evidence was based on the assumption that Rookes v. Barnard applied and that the issue was, did the case come within the second category? While it may be that the plaintiff would have presented his case differently but for what was said in Rookes v. Barnard, the defendants had to meet the case as presented whether or not Rookes v. Barnard applied, and it was in relation to that case that they decided to call no evidence. As the case presented would prior to Rookes v. Barnard, if established, have justified the award of exemplary damages, I cannot accept that the defendants might have reached a different decision about calling evidence on the case as presented if Rookes v. Barnard had not been followed.

 

I now turn to the passage in Lord Devlin’s speech dealing with the assessment of damages, a passage which, save in the respect to which I have referred, was closely followed by Lawton J. in his summing-up.

 

I think that Salmon L.J., as he then was, correctly summarised the pre-Rookes v. Barnard practice when he said [1971] 2 Q.B. 354, 387-388:

 

“Judges used to direct juries in libel actions that, if they found in favour of the plaintiff, they should award him a sum which would make it plain to the world that there was no truth in the libel and which, as far as money could do so, would compensate him for the distress, humiliation and annoyance which the libel had caused him. They were also told in appropriate cases that they could take the whole of the defendant’s conduct into account down to the moment they returned their verdict, and that if they came to the conclusion that he had behaved outrageously they might, as a deterrent, reflect their disapproval of the defendant’s conduct in the amount of the damages which they awarded. At the same time they were always warned to be fair and reasonable and not to allow themselves to be inflamed against the defendant but to decide dispassionately what in all the circumstances would be a reasonable sum to award.”

 

The summing-up in Loudon v. Ryder [1953] 2 Q.B. 202 which was previewed by the Court of Appeal also recognised that outrageous conduct was a ground for exemplary damages. That appears to be the first case in which a jury was asked to award separate sums for exemplary and for compensatory damages and in which it was suggested that the amount awarded for exemplary damages was to be regarded as the imposition of a fine.

 

In Ley v. Hamilton, 151 L.T. 360 the Court of Appeal by a majority (Greer and Maugham L.JJ.; Scrutton L.J. dissenting) allowed an appeal  [*1110]  from a jury’s verdict awarding £5,000 damages for libel, one ground for the decision being that the damages awarded were excessive; Maugham L.J. saying, at p. 374, that the sum could not be described “as a fair and reasonable compensation for the damages which the plaintiff” had suffered, that the verdict could only be justified on the view that the jury were exercising the right to give vindictive or punitive damages, and that

 

“when the damages in question are really not compensation for an injury sustained by the plaintiff but in the nettler of a fine inflicted on a defendant,”

 

the Court of Appeal would be compelled to interfere.

 

In this House, 153 L.T. 384 Maugham L.J.’s approach was rejected by Lord Atkin in a speech with which Lords Tomlin, Thankerton, Macmillan and Wright agreed. Part of the relevant passages of Lord Atkin’s speech were cited by Lord Devlin but two sentences which I italicise and which I regard as important were omitted. The full passage is as follows, at p. 386:

 

“The fact is that the criticism” (Maugham L.J.’s) “with great respect seems based upon an incorrect view of the assessment of damages for defamation. They are not arrived at as the Lord Justice seems to assume by determining the ‘real’ damage and adding to that a sum by way of vindictive or punitive damages. It is precisely because the ‘real’ damage cannot be ascertained and established that the damages are at large. It is impossible to track the scandal, to know what quarters the poison may reach: it is impossible to weigh at all closely the compensation which will recompense a man or a woman for the insult offered or the pain of a false accusation. No doubt in newspaper libels juries take into account the vast circulations which are justly claimed in present times. The ‘punitive’ element is not something which is or can be added to some known factor which is non-punitive. In particular it appears to present no analogy to punishment by fine for the criminal offence of publishing a defamatory libel.”

 

Maugham L.J. did not in his judgment refer to “real” damage. I think it is clear that by “real” damage Lord Atkin meant the damage which the plaintiff had suffered.

 

Yet is not the very process condemned in Ley v. Hamilton that which it was said in Rookes v. Barnard should be followed and that which, pursuant to Rookes v. Barnard, was followed in this case? Lord Atkin said that for the reasons he gave “real” damage i.e., compensatory damage, could not be ascertained and established. Under Rookes v. Barnard a jury is to be directed that that which Lord Atkin said could not be done, is to be done and “compensatory” damages assessed first. The punitive element is not something that can be added. Yet in Rookes v. Barnard it is said that it should be added if, but only if, the compensatory damages are insufficient. Lord Atkin said that there was no analogy to punishment by a fine for a criminal libel, yet following Rookes v. Barnard, juries are to be told that punitive damages amount to a fine. I must confess my  [*1111]  inability to reconcile the views of this House as expressed in Ley v. Hamilton with those expressed in Rookes v. Barnard.

 

Before Rookes v. Barnard the words “aggravated,” “punitive,” “exemplary” and “retributory” were used indiscriminately to indicate that the damages awarded might be enhanced and might contain a punitive element. By Rookes v. Barnard precise meanings were attached to the words “aggravated” and “exemplary.” Lord Devlin recognised, at p. 1221, that the jury could take into account the motives and conduct of the defendant where they aggravate the injury to the plaintiff. He said:

 

“There may be malevolence or spite or the manner of committing the wrong may be such as to injure the plaintiff’s proper feelings of dignity and pride. These are matters which the jury can take into account in assessing the appropriate compensation.”

 

So where the injury is aggravated, an addition can be made to the compensatory damages.

 

While in some cases it may be evident that malice or misconduct has added to the injury, there may be other cases where, although it is clear that there has been malice and misconduct, it cannot be said that the injury inflicted is any greater than it would have been if there had been no malice or misconduct. In such cases it would seem from Rookes v. Barnard that the compensatory damages should not be increased. Nor, in such cases would it seem that exemplary damages as there defined could always be awarded for they are only to be awarded if the sum given in compensation is “inadequate to punish for outrageous conduct, to mark the jury’s disapproval of such conduct, and to deter a repetition.” The existence of malice may not make the defendant’s conduct outrageous, and yet it is, I think, established beyond all doubt that before Rookes v. Barnard a jury was always entitled to award larger damages than they otherwise would have given if satisfied that the libel was actuated by malice.

 

All the members of the Court of Appeal thought that the Rookes v. Barnard approach was wrong and in conflict with the views expressed in this House in Ley v. Hamilton. I can find no escape from that conclusion and if the choice now lies between following one or the other of those decisions, I would myself choose to follow the simpler and more flexible approach in Ley v. Hamilton. The Court of Appeal also thought that there was a conflict with the decision of this House in E. Hulton & Co. v. Jones [1910] A.C. 20. While there are some passages in the report of that case which afford some ground for that contention, I do not think that they suffice to establish that that is so with any degree of certainty.

 

While, if the views I have expressed prevailed, it would not be necessary to disturb the jury’s verdict as to liability, I cannot regard a direction to assess damages in accordance with Rookes v. Barnard as a proper direction in accordance with the pre-Rookes v. Barnard practice and as complying with Ley v. Hamilton. So, if my view were to prevail, the verdict given in this case could not be sustained and there would, if there had not been agreement by counsel that this House should in that event assess the damages, have to be a new trial limited to the assessment of damages. As my view does not prevail, it is not necessary to express an opinion on what that sum should be if this House had to assess it.

 

For the reasons I have stated, I would allow the appeal.  [*1112] 

 

LORD WILBERFORCE. My Lords, this case must be accounted, as in many respects, an unhappy one. After a trial of seventeen days before a judge and jury, in which the defendants called no evidence, the plaintiff, Captain Broome, was awarded against author and publishers jointly £40,000 damages in respect of libels contained in the book “The Destruction of Convoy PQ17.” This total sum was awarded by the jury as to £15,000 as “compensatory” damages and as to £25,000 as “punitive” damages. Captain Broome was awarded his costs of the trial.

 

An appeal was taken to the Court of Appeal by both defendants. The substantial points for argument were two: (1) whether the summing-up was defective as regards the circumstances in which punitive damages may be given in addition to compensatory damages (2) whether the damages awarded were excessive. There was also a question as to whether a separate award should have been made against each defendant. Since the passages in the book principally complained of reflected upon the conduct of officers of the Royal Navy, in combat conditions, there was an obvious danger that the jury may have become inflamed. This made it particularly necessary that there should be a dispassionate and cool review of the sums awarded and of the summing-up in the Court of Appeal.

 

If matters had taken their proper and normal course, these matters should have been disposed of within a few disobey dismissal of the appeal or by an order for a new trial, and no question of appeal to this House would have arisen.

 

This did not happen. The trial had been conducted properly, and inevitably upon the basis that the law to be applied as regards any claim for punitive damages was that stated by this House in Rookes v. Barnard [1964] A.C. 1129. The learned judge considered that he was bound by what was said in this House, as he clearly was. But in the Court of Appeal argument was admitted to the effect that Rookes v. Barnard, on punitive damages, was wrong and should not be followed: the Court of Appeal so decided, and three judgments, separate exercises in forceful advocacy, were delivered giving their reasons.

 

The course permitted and taken was doubly surprising. First, there was nothing new about Rookes v. Barnard. It was decided in 1964: it had been followed and applied in England by the Court of Appeal itself three times since then in, amongst others, libel cases without difficulty or protest by any of the Lords Justices involved. Secondly, it was, on the view of the facts which the Court of Appeal took, unnecessary for the decision of the appeal to decide whether Rookes v. Barnard on punitive damages was right or wrong. The Court of Appeal, having held that it was wrong, still dismissed the appeal, and in an alternative passage held that the same result followed if it was right.

 

The consequences for the present litigants have been heavy. An appeal has been brought here and argued for thirteen days. Counsel for the appellants were forced into the necessity of arguing at length that Rookes v. Barnard is right, and this argument was answered on the respondent’s side. A mountain of costs has piled up and it is as well that the size of this should be understood: it is open on the record. As shown by the order of the Court of Appeal, the plaintiff’s costs at the trial have been taxed at £22,000. His costs as assessed in the Court of Appeal are £7,000. His costs in this  [*1113]  House must exceed this figure. The taxed costs of the defendants are unlikely to be less: there will be further solicitor and own client costs on either side. It may not be unfair to put the aggregate bill, which an unsuccessful party may have to bear, at more than £60,000. It would be entirely unfair to suggest that the whole, or even half this sum, is due to the course taken in the Court of Appeal — the greater part flows from the inherent nature of our system. But it is necessary to say that in a legal system so extravagant and punitive as to costs as ours is in civil cases, and particularly libel actions, the addition of further burdens, and here they were certainly considerable, carries the result further into an unacceptable area of injustice. England has not the equivalent of the New South Wales Suitors Fund Act 1951 nor of the Victoria Appeal Costs Fund Act 1964, so when the machinery creaks it is the private litigants who pay. I have felt deep concern about this throughout the hearing.

 

My Lords, observations have already been made on other constitutional aspects of the Court of Appeal’s judgments. I concur entirely with what has been said, and the fact that for reasons of space I abstain from using my own words does not mean that my concurrence is any the less wholehearted.

 

I proceed to the principal task we have, which is to decide the present appeal. Before examining the summing-up, on which the jury’s verdict was based, it is necessary to establish the law. This involves some reexamination of those parts of the decision in Rookes v. Barnard [1964] A.C. 1129 which relate to punitive damages.

 

I shall consider Rookes v. Barnard under three heads. First, as to the analysis it contains of damages in tort cases: secondly, as to defamation actions in relation to Lord Devlin’s second category — both of these being directly relevant to the present case: thirdly, and briefly, as to the first and second categories, their inclusions and exclusions.

 

I deal first with that portion of the judgment which analyses damages in tort cases into “compensatory” damages, a subhead of which is said to be “aggravated” damages and punitive damages, because I think that this has been largely misunderstood — a misunderstanding which has fatally entered into the present case.

 

The judgment points out that in the reported English authorities, over some two hundred years, there is no clear terminology used; aggravated, exemplary, punitive, vindictive, retributory being adjectives which have been used, singly or in combination, without distinction or difference. Then it is suggested that in future there should be a clear and conscious distinction between compensatory/aggravated and punitive (or exemplary) damages, the former reflecting what the plaintiff has suffered materially or in wounded feelings, the latter the jury’s (or judge’s) views of the defendant’s conduct. The statement of categories, in which alone punitive damages may be given, follows from this.

 

This analysis is powerful and illuminating and undoubtedly represents a valuable contribution to English judicial thought on the subject* but it has its dangers in practical application, as the present case only too well shows.

 

* cf. in the United States Fay v. Parker (1873) 53 N.H. 342-397 per Foster J.: and as to textbook discussion, Mayne and Macgregor on Damages, 12th ed. (1961): Street’s Principles of the Law of Damages (1962)  [*1114]  English law does not work in an analytical fashion; it has simply entrusted the fixing of damages to juries upon the basis of sensible, untheoretical directions by the judge with the residual check of appeals in the case of exorbitant verdicts. That is why the terminology used is empirical and not scientific. And there is more than merely practical justification for this attitude. For particularly over the range of torts for which punitive damages may be given (trespass to person or property, false imprisonment and defamation being the commonest) there is much to be said before one can safely assert that the true or basic principle of the law of damages in tort is compensation, or, if it is, what the compensation is for (if one says that a plaintiff is given compensation because he has been injured, one is really denying the word its true meaning) or, if there is compensation, whether there is not in all cases, or at least in some, of which defamation may be an example, also a delictual element which contemplates some penalty for the defendant. It cannot lightly be taken for granted, even as a matter of theory, that the purpose of the law of tort is compensation, still less that it ought to be, an issue of large social import, or that there is something inappropriate or illogical or anomalous (a question-begging word) in including a punitive element in civil damages, or, conversely, that the criminal law, rather than the civil law, is in these cases the better instrument for conveying social disapproval, or for redressing a wrong to the social fabric, or that damages in any case can be broken down into the two separate elements. As a matter of practice English law has not committed itself to any of these theories: it may have been wiser than it knew.

 

This is not the place to argue out the general case for or against punitive damages in English law. The existence of the principle has its convinced opponents, particularly, I understand, in Scotland. The arguments against it — that it is an “anomaly,” that it brings a criminal element into the civil law without adequate safeguards, that it leads to excessive awards, an unmerited windfall for the plaintiff; these and others are by now well known: they, and the counter-arguments are well summed up in Professor Street’s Principles of the Law of Damages (1962), pp. 33-34. Perhaps the opponents have, marginally, the best of it in logic, but logic in excess has never been the vice of English law and I am impressed, as I think was Lord Devlin, with the fact that the principle has shown, and continues to show, its vitality not only in England but in Australia, Canada and New Zealand, as well (though there are special considerations there) as in the United States of America. This is shown not only by reported cases. of which Canadian Provinces, Australian States and New Zealand provide a number of modern examples *2, but in the daily unreported practice of the courts. Its place in the law has been endorsed by many eminent judges in terms which clearly recognise the punitive element. The principle of punitive damages has been recognised by the High Court of Australia on five occasions, by the Supreme Court of Canada and by the Supreme Court of the United States of America.

 

To my mind the strongest argument against it is that English law already contains a heavy, indeed exorbitant, punitive element in its costs

 

FN*2 See, as to Canada, 48 Canadian Bar Review (1970) p. 373.  [*1115] 

 

system; contrast the United States where it is the absence of this (advocate’s costs not being normally recoverable) which is invoked as a justification for punitive damages. One or other must clearly be reformed, and it is Parliament alone that can do it.

 

I take the discussion one step further, because the point is very relevant here. In Lord Devlin’s opinion the distinction is made between aggravated damages and punitive damages; it is said that many of the authorities are really cases of aggravated damages though other words are used, that apart from the exceptional cases included in the three categories, aggravated damages are the appropriate and sufficient remedy. Although I doubt very much whether all the cases can be explained in this way — to do so seems to attribute a high degree of confusion of thought or inaccuracy of expression to judges of eminence — there is attraction in the distinction. It has the advantage, to some minds, of reducing the area of “punitive” damages, and of bringing the remedy nearer to “compensation.”

 

But closer examination causes one to doubt whether the separation, otherwise than in analysis, of compensatory from punitive damages does not involve some real danger in practice. As Windeyer J. said in Uren v. John Fairfax & Son Pty., 117 C.L.R. 118, 152:

 

“What the House of Lords has now done is, ... to produce a more distinct terminology. Limiting the scope of terms that often were not distinguished in application makes possible an apparently firm distinctionbetween aggravated compensatory damages and exemplary or punitive damages. How far the different labels denote concepts really different in effect may be debatable. I suspect that in seeking to preserve the distinction we shall sometimes find ourselves dealing more in words than ideas.”

 

(cf. Salmond on Torts, 15th ed. (1969), which maintains the old “confusion.”) The distinction does not in my belief greatly correspond to what happens in reality. Take a common case: a man is assaulted, or his land is trespassed upon, with accompanying circumstances of insolence or contumely. He decides to bring an action for damages, he need not further specify the claim. Is he suing for compensation, for injury to his feelings, to teach his opponent a lesson, to vindicate his rights, or “the strength of the law,” or for a mixture of these things? Most men would not ask themselves such questions, many men could not answer them. If they could answer them, they might give different answers. The reaction to a libel may be anything from “how outrageous” to “he has delivered himself into my hands.” The fact is that the plaintiff sues for damages, inviting the court to take all the facts into consideration, and, if he wins, he may ascribe his victory to all or any of the ingredients.

 

As, again Windeyer J. has said, at p. 150, “the amount of a verdict is the product of a mixture of inextricable considerations.” Sedgwick (Measure of Damages, 3rd ed. (1858)) said:

 

“Where either of these elements [sc. malice, oppression etc.] mingle in the controversy, the law, instead of adhering to the system, or even the language of compensation adopts a wholly different rule. It permits the jury to give what it terms punitive, vindictive or exemplary damages, in other words, it blends together the interests of society and  [*1116]  of the aggrieved individual and gives damages not only to recompense the sufferer but to punish the offender. This rule ... seems settled in England and in the general jurisprudence of [the U.S.A.].”

 

Lord Atkin said just this in Ley v. Hamilton, 153 L.T. 384 in a passage (cited in other opinions, vide that of Viscount Dilhorne) which, if any in modern times, is clear and authoritative. Dixon C.J. endorsed the principle — see citation below — as did the key passage in Halsbury’s Laws of England,vol. 11, p. 223, cited by Lord Hailsham of St. Marylebone L.C. To segregate the punitive element is to split the indivisible and to invite the stock criticism (vide Street loc. cit.) that civil courts have no business to impose fines.

 

This is of critical importance in practice. If the separation of damages into compensatory/aggravated and punitive is carried through into the instruction to the jury, there is the greatest possible risk of excessive awards, through counting twice what is but a different facet of the same bad conduct. Lord Devlin himself clearly understood this; the careful passage on p. 1232, containing the “if, but only if,” prescription, provided his antidote — an effective one if judges can administer it in a timely and effective way.

 

My Lords, I think there was much merit in what I understand was the older system, before Rookes v. Barnard [1964] A.C. 1129. I agree with the Court of Appeal that in substance, though not perhaps philosophically or linguistically, this was clear and as explained above I doubt if there was any confusion as to what the jury should do. It was to direct the jury in general terms to give a single sum taking the various elements, or such of them as might exist in the case, into account including the wounded feelings of the plaintiff and the conduct of the defendant, but warning them not to double count and to be moderate. A formula on these lines commended itself to Dixon J. in Smith’s Newspapers Ltd. v. Becker (1932) 47 C.L.R. 279, 300. What amount of damages, he asked, “is enough to serve at once as a solatium, vindication and compensation to him and a requital to the wrongdoer.” An earlier example is the direction of Abbott J. in Sears v. Lyons (1818) 2 Stark. 317: as evidence that modern practice corresponds I could not desire more than the passage, based on considerable experience, in the judgment of Salmon L.J. in this case [1971] 2 Q.B. 354, 387-388 cited in full by Viscount Dilhorne and which I need not repeat. If judges were to act in this way, and direct substantially as Salmon L.J. describes, I would see no basis for ascribing to them any error in law. If, on the other hand, use were to be made of the aggravated-punitive distinction, I would think that it is even more necessary that the jury should be directed to give a single sum (Lord Devlin’s exception to avoid a new trial is entirely laudable, but, I respectfully think, risky). The direction to give a single sum should mean (the necessity to say this illustrates again the dangers of the terminology) not merely producing a single figure by way of verdict, but arriving in their discussion at a single sum. It would be wrong, and a novelty in the law, that they should, in the jury room, find separately the various elements — pure compensation, aggravated compensation and penalty and add them up to a total. In no previous cited case, except in Loudon v. Ryder [1953] 2 Q.B. 202 (overruled by Lord Devlin himself), was this done; it was directly discountenanced by Lord Atkin in Ley v. Hamilton, 153 L.T. 384.  [*1117] 

 

I regret that this rather lengthy analysis has been necessary before I deal with the present appeal: but in my view it is fundamental to a consideration of the summing-up.

 

The full account of the trial which has been given in previous opinions enables me to summarise. The critical stages were these (page references are to the appeal record appendix, part II) of the printed case in the House of Lords:

 

(1) The jury were told that there were two aspects of damages, compensatory and punitive. They were asked first to consider compensatory damages. They had read to them a passage from the judgment of Pearson L.J. in McCarey v. Associated Newspapers Ltd. (No. 2) [1965] 2 Q.B. 86 in which it was said in clear terms that if there had been any high-handed, oppressive or contumelious behavior which increased the mental pain and suffering caused by the defamation, this might be taken into account (pp 91-93).

 

(2) The judge then pointed out that Captain Broome had suffered no actual pecuniary loss: that he had not been shunned by his comrades: that the trial had been conducted without exacerbation: but that what was said in the book might be very wounding to his feelings (pp. 93-98).

 

(3) The learned judge then dealt with punitive damages by reference to the second category in Rookes v. Barnard, cited the words of Widgery J. in Manson v. Associated Newspapers Ltd. [1965] 1 W.L.R. 1038 and said:

 

“You are being asked here not only to give Captain Broome compensatory damages, that is, a reasonable sum for the injury to his reputation and the exacerbation of his feelings; but in addition to fine Cassells and Mr. Irving for having done what they have done ... You are really in the position of a judge or a magistrate trying a criminal case; you have got, so to speak, to fine the defendant” (p. 101) (emphasis supplied)

 

and he gives examples of reasonable and unreasonable fines. Later he gives lengthy directions relevant to the second category (was there a calculation to profit etc.) and on the next day returns finally to damages (p. 137).

 

(4) The final direction as to damages consisted of the statement of questions for the jury and explanation of them. The first question (No. 3) is “What compensatory damages do you award the plaintiff?” The summing-up continues:

 

“Then having decided what are the proper additional (sic) compensatory damages you will go on and consider the fourth question, namely, ‘Has the plaintiff proved that he is entitled to exemplary damages?’”

 

and directs the jury to consider this in relation separately to each defendant. Lastly there is this passage:

 

“Then you see the last question under this heading, ‘What additional sum should be awarded him by way of exemplary damages?’ Would you be good enough to underline the word ‘additional,’ because I want to know, and learned counsel want to know, if you do decide to award punitive damages, how much more do you award over and above the compensatory damage.”  [*1118] 

 

The result of this was an award of £15,000 compensatory damages and £25,000 as an additional sum for exemplary damages.

 

My Lords, I regret to have reached the conclusion that this verdict ought not to stand. Apart from the reasons given by my noble and learned friend, Lord Diplock, with which I respectfully agree, I think for myself that the separation of the element of compensatory damages from that of punitive damages, brought about through the interpretation placed on the second category and the application of it, involving, as it did, the need to fix compensation (plus aggravation) first, see if the case came within the category, and then fix a separate punitive sum, is fundamentally wrong. It has brought about precisely the result which was to be feared from breaking down the indivisible whole, namely, of fixing a compensation figure swollen by aggravation and then adding a fine on top — a fine in this case exceeding greatly the aggravated compensation. If the matter rested on the figures alone, I should find the greatest difficulty in supporting. even with all the inhibitions properly felt against substituting a judicial opinion for that of the jury, so large a punitive element, particularly in a case such as this where the libel was considered to be (I say nothing as to my own opinion) of a most wounding character, so that the “compensatory” damages must necessarily include a large “punitive” element. But when it is seen how the jury were directed to calculate, and the direction was certainly clear and certainly and visibly acted on, their figures become impossible to accept.

 

In argument the issue was put in the form whether the judge’s direction complied with Lord Devlin’s “if, but only if,” advice (p. 1228). I think that it certainly did not. The dangers of separating the compartments (compensatory damages and punitive damages) in so watertight a way are so great, as I have tried to explain, indeed, in my opinion, so wrong in principle that I doubt very much whether any instructions, in a difficult case, could avoid them. That is why I think that any interpretation placed on Rookes v. Barnard which requires this separation, or authorises it, and the introduction of the profit gateway which almost compels it, ought to be discarded. But however that may be, the directions given fall far short of what was necessary — I say this without any criticism of the learned judge who was merely following Rookes v. Barnard as previously applied by the Court of Appeal. When all is said, the warning to the jury against the danger was contained in the word “additional” in question 4. I think this was not enough, for they had been told that they could inflict a fine.

 

For these reasons, without committing myself to any particular figure if we were called upon to substitute one, I agree with the conclusion of my noble and learned friend, Lord Diplock, as to the necessity for a new trial on the question of punitive damages.

 

I must add one other point. This is the question of a joint award of damages against two wrongdoers, publisher and author. There is no doubt that the existing law is ill adapted to deal fairly with a case where “guilt” of joint defendants is unequal. But it is clear enough what the law is: it is stated by Lord Hailsham of St. Marylebone L.C. in terms which I need not repeat. In the Court of Appeal Lord Denning M.R. said that the jury were free to decide whether to fix punitive damages at the highest figure, the lowest figure, or at a figure between the two and I fear that the jury may well have proceeded on this somewhat libertarian view of the law. One may  [*1119]  escape from the conclusion that this vitiates the verdict by assuming that the two defendants were equally “guilty,” but I am not prepared to make this assumption or to ascribe a view to that effect to the jury. I think that the jury must have been, at best, confused, at worst misled by the direction, and I cannot accept that acquiescence by counsel validates the defect.

 

I must now deal as briefly as I can with other aspects of the judgment in Rookes v. Barnard. I deal first with its effect on the law of damages for defamation.

 

I am far from convinced that Lord Devlin ever intended to alter the law as to damages for defamation or intended to limit punitive damages in defamation actions to cases where a “profit motive” is shown. (I use this compendiously for the formula in his second category.) I summarise the reasons:

 

(a) Defamation is normally thought of as par excellence the tort when punitive damages may be claimed. It was so presented in argument by counsel for the respondent (arguing against punitive damages) and he was an acknowledged expert in the subject. Every practitioner and every judge would take this view.

 

(b) Lord Devlin’s passage where he sets up his second category does not refer to any defamation case, but to three other miscellaneous cases which he illuminatingly bases in the profit motive. He makes merely an incidental reference to libel where he says the profit motive is always a factor, not, it should be observed, a condition.

 

(c) It is difficult to believe that Lord Devlin was intending to limit the scope of punitive damages in defamation actions so as to exclude highly malicious or irresponsible libels. At least if he intended to do so at a time when the media of communication are more powerful than they have ever been and certainly not motivated only by a desire to make money, and since elsewhere the judgment shows him conscious of the need to sanction the irresponsible, malicious or oppressive use of power, I would have expected some reasons to be given.

 

If we cannot interpret his judgment as leaving libel outside category 2 as a separate case, well known to everyone, in which punitive damages may be given in familiar circumstances and as stating category 2 as a qualification for other cases, hitherto not explained or rationalised, then since the disposal of defamation actions was there dealt with briefly, I would say incidentally, and obiter, I consider that in this case where we are directly concerned with such an action we should disagree with it.

 

This would leave the law as I understand it to be in Australia and Canada, countries where, in this respect, there is not known to me to be any such difference in “social conditions” as to call for the recognition, by this House, judicially, of a divergent law. If changes are to be made, they should be made, after proper investigation, by Parliament.

 

I would add, with reference to this point, that the present case well illustrates the irrationality of the supposed new principle. For if the profit motive is essential for the recovery of punitive damages, one would expect the damages given to bear some relation to the supposed profit and/or to the means of the offender: the idea (if there is any logic in the requirement) must be to take the profit out of wrongdoing. Yet there was not, and in many such cases cannot be, any real consideration of the likely profit or of  [*1120]  the offender’s means. There was no evidence what these might be and the jury were given no guidance. How, then, could the punitive £25,000 be other than an arbitrary guess? If one replies that it represents the jury’s view of the defendants’ conduct (as it probably did) what purpose is served by introducing the profit motive gateway?

 

Finally, as to other torts as to which, before Rookes v. Barnard, punitive damages could be given but on which some restriction is evidently intended to be placed by the judgment. That this House, as a matter of law, or of legal policy, was entitled to restrict the scope of punitive damages I have, with all respect to the Court of Appeal, no doubt and, whatever my own reservations as to the wisdom of the policy, I should feel myself obliged to accept a new statement of principle if it were clear, consistent and workable and intelligibly related to the main stream of authority. That it was not entirely clear, appears well enough from the opinions in the present case: and I cannot entirely blame the Court of Appeal for attempting to escape from it, just as one may sympathise with a customer when he finds his new suit almost at once requiring alteration, or patching, for putting it aside and reaching for his old tweeds. There is not perhaps much difficulty about category 1: it is well based on the cases and on a principle stated in 1703 — “if public officers will infringe men’s rights, they ought to pay greater damages than other men to deter and hinder others from the like offences”: Ashby v. White (1703) 2 Ld.Raym. 938, 956 per Holt C.J. Excessive and insolent use of power is certainly something against which citizens require as much protection today: a wide interpretation of “government” which I understand your Lordships to endorse would correspond with Holt C.J.’s “public officers” and would partly correspond with modern needs. There would remain, even on the most liberal interpretation, a number of difficulties and inconsistencies as pointed out by Taylor J. in Uren v. John Fairfax & Sons Pty. Ltd., 117 C.L.R. 118.

 

I have more difficulty with the commonplace types of trespass or assault accompanied by insult or contumely, which, even more than “first category” cases touch the life of ordinary men and occupy the county courts. Although Lord Devlin studiously refrains from overruling earlier cases (other than Loudon v. Ryder [1963] 2 Q.B. 202) which undoubtedly proceeded on, or contained, a punitive element, his opinion has been understood as laying down that in future such cases cannot, unless the “profit motive” is present, be treated as cases for punitive damages but only as cases for aggravated damages. The phrase used has been “aggravated damages can do the work of punitive damages.”

 

I understand that a majority of your Lordships, for possibly differing reasons, are satisfied with this, so it will remain the law in this country. But, if only in fairness to the Court of Appeal with whose approach to this matter I agree, I must state very briefly why I feel some difficulty.

 

I am far from clear how juries, or judges, are intended to act in the future. Are they to take it that the law has been changed, so that (absent a profit motive) only “compensatory” damages can be given, plus an element for “aggravation” if that is proved? I fear that there will be difficulty in seeing how far earlier cases, or Commonwealth cases, are now authority and that there will be much argument whether a particular case was one of “aggravated” or “punitive” damages or of both. Alternatively,  [*1121]  if “aggravated damages” are “to do the work of punitive damages” and if it is to be supposed that juries, or judges, will continue giving damages much as before, then nothing has been gained by changing the label and we are indulging in make belief and encouraging fictional pleading. The whole point is well brought out by Pearson L.J. in McCarey v. Associated Newspapers Ltd. (No. 2) [1965] 2 Q.B. 86, 105. If the compensatory principle is accepted, punitive damages must not be allowed “to creep back into the assessment in some other guise.” I must confess to sympathy with the Court of Appeal’s preference for the older system and with the objections to the new stated by Taylor J. in Uren’s case, the weight of which clearly impressed the Privy Council. Their validity has been endorsed by cases post-Rookes v. Barnard in Australia, Canada and New Zealand. I share their doubt whether we have yet arrived at a viable substitute.

 

But I note with satisfaction and agreement the opinion expressed by the noble and learned lord on the Woolsack that the relevant passage in Lord Devlin’s judgment, which he cites, should be read sensibly as a whole together with the authorities on which it is based. This may provide a sound basis for redevelopment of the law.

 

My Lords, on all other points not expressly dealt with in this opinion I wish to express my concurrence with that of Lord Hailsham of St. Marylebone L.C. I regret to differ from him in thinking that the appeal should be allowed on the grounds I have stated.

 

LORD DIPLOCK. My Lords, the trial of this action proceeded, correctly, on the basis that as respects the measure of the damages which the jury might award, the judge was bound to direct them in accordance with the law as laid down by this House in Lord Devlin’s speech in Rookes v. Barnard [1964] A.C. 1129.

 

I agree with all your Lordships that there was material upon which the jury were entitled to find that the conduct of each of the defendants brought the case within Lord Devlin’s second category of cases in which exemplary or, as I would have preferred to call them, punitive damages may be awarded. The jury did so find by special verdicts. That part of the judge’s summing-up in which he directed them as to the matters for their consideration in arriving at their findings on this issue as respects each of the defendants cannot be faulted.

 

It was, however, also incumbent upon the judge to instruct the jury as to the measure of the damages which they might award if they reached the conclusion that the case as against each of the defendants was one in which they were not precluded from awarding punitive damages. On this aspect of the case there were two principles of law which should have been stated clearly to the jury. Neither was self-evident.

 

The first was that, even if the jury found that the case came within Lord Devlin’s second category and that the defendants’ conduct merited punishment, it did not necessarily follow that they must award as damages to the plaintiff a greater sum than was sufficient to compensate him for all the harm and humiliation that he had suffered as a consequence of the defendants’ tortuous acts. They should take into account as part of the punishment inflicted on the defendants any sum (in the result £15,000) which they were minded to award to the plaintiff as compensatory damages; and only if they  [*1122]  thought that sum to be inadequate in itself to constitute sufficient punishment were they to award such additional sum as would, when added to the compensatory damages, amount to an appropriate penalty for the defendants’ improper conduct.

 

The second was that if the jury thought that the conduct of one of the joint defendants deserved to be penalised by a lesser sum than the conduct of the other, the most that the jury were entitled to award against the defendants was that lesser sum, if it were to exceed the amount which they were minded to award as compensatory damages.

 

I have the misfortune to differ from the majority of your Lordships in that I find it impossible to discover in the language of the summing-up any clear statement of either of these principles. At best I think that when the jury retired they must have been confused as to how the punitive damages, if any, were to be assessed. At worst I think that they may well have thought that they were to arrive at a sum which they thought was an appropriate penalty for the defendants’ conduct and to add it to any sum awarded as compensatory damages.

 

My Lords, I do not think that on this vital question of the assessment of exemplary damages the jury were adequately directed. I am fortified in this view by my conviction that, if properly directed, no reasonable jury could possibly have reached the conclusion that the appropriate penalty to inflict on the less culpable of the defendants was £40,000 for publishing a libel of which the victim was in their view adequately recompensed at £15,000 for all the harm and humiliation that it had caused to him.

 

A penalty of £40,000 is, I believe, very much larger than any of your Lordships would have thought it appropriate to inflict upon the defendants. I doubt if any of your Lordships would have hesitated to interfere with it if it had been awarded by a judge sitting alone. He would have been vulnerable because he would have given his reasons. Shibboleths apart, there survive to-day two valid reasons why an appellate court should be more reluctant to disturb an assessment of damages by a jury than an assessment by a judge. The first is applicable to all kinds of actions. It is that a judge articulates his findings on the evidence and his reasoning, whereas a jury state the result of their findings and their reasoning but otherwise are dumb. In considering whether an award of damages by a jury is excessive an appellate court cannot do other than assume that the jury made every finding of fact and drew every inference that was open to it on the evidence as favorably as possible to the plaintiff and as adversely as possible to the defendant. In the instant case, however, this handicap to an appellate court’s ability to do justice is palliated by the facts: that there was no conflict of evidence for them to resolve — for the defendants called none, and that the jury were given a partial gift of speech. By their special verdict this House has been told that they considered that the plaintiff would be fully compensated by £15,000. The second reason for reluctance to interfere with a jury’s award of damages applies particularly to actions for defamation. It is that, unless the parties otherwise agree, the consequence of setting aside the jury’s verdict must be a new trial before another jury. This involves the parties, through no fault of their own, in greatly increased costs which, particularly in libel actions, are, to the discredit of our legal system, out of all proportion even to the large compensatory damages  [*1123]  awarded in the instant case. For my part, I should not be deflected from setting aside a jury’s verdict as unreasonable by the fear, sometimes expressed by appellate judges, that another unreasonable jury might make a similar unreasonable award of damages on the new trial. So far as I know this has never happened yet. But the consideration of the costs involved is one which it would be unrealistic and unjust to ignore. In the instant case, however, the parties agreed that this House should assess the damages in the event of the jury’s verdict being set aside. No more costs would be incurred if the appeal were allowed than if it were dismissed — though the incidence of them on the parties might be different.

 

It may be said, and not implausibly, that there is nothing in the training or experience of a judge which makes him fitter than a jury to determine the pecuniary compensation which a plaintiff should receive for a reputation that is damaged or feelings that are hurt. And there may be safety in numbers. But it runs counter to the basis of our criminal law, in which the jury determine guilt and the judge determines the appropriate punishment, to treat the jury as better qualified than a judge to assess the pecuniary penalty which a defendant ought to pay for conduct which merits punishment. On an appeal from the jury’s award of £40,000 which I know to be compensatory to the extent of £15,000 only, I should approach it in the same way as I should approach a fine of £40,000 imposed by a judge in a criminal prosecution. Even if I thought the jury had been given an adequate direction by the judge, I would have set the award aside and substituted an award of £20,000.

 

I have thought it right to express my own minority opinion as to what the result of this appeal should be. It is that with which the parties are primarily concerned — and it is they who are paying for it. It is, however, inherent in our legal system that, owing to the manner in which the Court of Appeal dealt with the instant case, the unsuccessful party is also paying for the ruling of this House upon two questions of law of much more general importance. The first is as to the effect of the decision in Rookes v. Barnard [1964] A.C. 1129 on the assessment of damages for defamation and whether that decision ought to be followed. The second is as to the propriety of the manner in which the Court of Appeal, as an intermediate appellate court, dealt with the decision of this House in Rookes v. Barnard. To these two topics I now turn.

 

In Rookes v. Barnard the plaintiff’s claim was for damages for the tort of intimidation At the trial the judge had summed up to the jury in terms which left it open to them to award exemplary damages. There was a cross-appeal against the amount of damages. upon which this House heard separate and lengthy argument. It was necessary as a matter of decision of the cross-appeal for this House to determine whether the facts in Rookes v. Barnard brought it within a category of cases in which exemplary damages were recoverable at common law. This House determined that they did not and ordered a new trial.

 

There were two different processes of reasoning by which it would have been possible to reach this conclusion of law. One, which was not adopted by this House, was to hold that the particular tort of intimidation was one in which the common law did not permit of exemplary damages. The other, which was adopted by this House, was to state the categories of cases in  [*1124]  which alone exemplary damages might be awarded at common law and to determine whether the facts in Rookes v. Barnard brought it within one of these categories.

 

Lord Devlin’s speech upon the cross-appeal in Rookes v. Barnard, in which all the five members who heard the appeal explicitly concurred, was a deliberate attempt by this House to do two things. (A) As a matter of legal exposition, to formulate the rationale of the assessment of damages for torts in which damages are “at large.” (B) As a matter of legal policy, to restrict the categories of cases in which damages can be awarded against a defendant in order to punish him, to those in which this method of inflicting punishment still serves some rational social purpose today.

 

Lord Devlin’s speech, however, does not follow the simple arrangement of exposition followed by choice of policy. He starts by formulating three heads of damages. The purpose of two of them is to compensate the plaintiff; that of the third is to punish the defendant. This formulation is followed by an analysis of the previous authorities. These authorities lead to the policy decision to accept two categories of cases in which exemplary damages may be recovered and, proleptically, to reject other categories of cases in which it had previously been thought that damages might be awarded in order to punish the defendant. He then reverts to exposition of some considerations which follow from the purpose served by exemplary damages and which should be borne in mind when awards of exemplary damages are made. Finally he reverts to an analysis of the previous authorities for the purpose of completing the policy decision by over-ruling those which were authority for the award of exemplary damages where the injury to the plaintiff had been aggravated by malice or by the manner of doing the injury, that is, the insolence or arrogance by which it was accompanied.

 

It is, however, convenient for the purposes of the instant appeal to deal with exposition and with policy separately.

 

The three heads under which damages are recoverable for those torts for which damages are “at large” are classified under three heads. (1) Compensation for the harm caused to the plaintiff by the wrongful physical act of the defendant in respect of which the action is brought. In addition to any pecuniary loss specifically proved the assessment of this compensation may itself involve putting a money value upon physical hurt, as in assault, upon curtailment of liberty, as in false imprisonment or malicious prosecution, upon injury to reputation, as in defamation, false imprisonment and malicious prosecution, upon inconvenience or disturbance of the even tenor of life, as in many torts, including intimidation. (2) Additional compensation for the injured feelings of the plaintiff where his sense of injury resulting from the wrongful physical act is justifiably heightened by the manner in which or motive for which the defendant did it. This Lord Devlin calls “aggravated damages.” (3) Punishment of the defendant for his anti-social behavior to the plaintiff This Lord Devlin calls “exemplary damages.” I should have preferred the alternative expression “punitive damages” to emphasise the fact that their object is not to compensate the plaintiff but to punish the defendant and to deter him, and perhaps others, from committing similar torts. To avoid confusion I have,  [*1125]  however accepted the lead of Lord Hailsham of St. Marylebone L.C. in adhering to Lord Devlin’s adjective “exemplary.”

 

It may seem remarkable that there had not previously been any judicial analysis, even as elementary as this, of the constituent elements of the compound “damages at large.” But it has to be remembered that at common law the assessment of damages was the exclusive function of a jury, and, despite growing exceptions from the mid-nineteenth century onwards, nearly all actions for torts in which damages were at large were tried by jury until after 1933. The assessment of damages was an arcanum of the jury box into which judges hesitated to peer; and it does not appear to have been their practice to give any direction to the jury as to how they should arrive at the amount of damages they should award, beyond some general exhortation to do their best in a matter which was peculiarly within their sphere.

 

What is disclosed by an examination of previous judgments since the eighteenth century, given upon applications for a new trial on the grounds that the award of a jury was too large or too small, is a confusion of language and consequently of thought as to what were the constituent elements in an award of damages at large. In particular there is a complete failure to distinguish between aggravated and exemplary damages in cases where the malice of the defendant or the manner in which he did the wrongful act had both increased the injury to the plaintiff’s feelings and aroused the indignation of the jury themselves.

 

In addition to the cases specifically referred to by Lord Devlin in Rookes v. Barnard, your Lordships have been referred to many others in the course of the argument in the instant appeal. They serve but to confirm the confused state of the law upon this subject before 1964.

 

The tort of defamation, to which Lord Devlin made only a passing reference in Rookes v. Barnard, has special characteristics which may make it difficult to allocate compensatory damages between head (1) and head (2). The harm caused to the plaintiff by the publication of a libel upon him often lies more in his own feelings, what he thinks other people are thinking of him, than in any actual change made manifest in their attitude towards him. A solatium for injured feelings, however innocent the publication by the defendant may have been, forms a large element in the damages under head (1) itself even in cases in which there are no grounds for “aggravated damages” under head (2). Again the harm done by the publication, for which damages are recoverable under head (1) does not come to an end when the publication is made. As Lord Atkin said in Ley v. Hamilton, 153 L.T. 384, 386: “It is impossible to track the scandal, to know what quarters the poison may reach.” So long as its withdrawal is not communicated to all those whom it has reached it may continue to spread. I venture to think that this is the rationale of the undoubted rule that persistence by the defendant in a plea of justification or a repetition of the original libel by him at the trial can increase the damages. By doing so he prolongs the period in which the damage from the original publication continues to spread and by giving to it further publicity at the trial, as in Ley v. Hamilton, extends the quarters that the poison reaches. The defendant’s conduct between the date of publication and the conclusion of the trial may thus increase the damages under head (1).  [*1126] 

 

In this sense it may be said to “aggravate” the damages recoverable as, conversely, the publication of an apology may “mitigate” them. But this is not “aggravated damages” in the sense that that expression was used by Lord Devlin in head (2). On the other hand the defendant’s conduct after the publication may also afford cogent evidence of his malice in the original publication of the libel and thus evidence upon which “aggravated damages” may be awarded under head (2) in addition to damages under head (1). But although considerations such as these may blur the edges of the boundary between compensatory damages under head (1) and compensatory damages under head (2) in the case of defamation, they do not affect the clear distinction between the concept of compensatory damages and the concept of exemplary damages under head (3).

 

My Lords, the major clarification of legal reasoning to be found in the expository part of Lord Devlin’s speech in Rookes v. Barnard was the recognition, first, that the award of a single sum of money as damages for tort, while it must always perform the function of giving to the plaintiff what he deserves to receive to compensate him fully for the harm done to him by the defendant, may in appropriate cases also perform the quite different function of fining the defendant what he deserves to pay by way of punishment; and secondly, that even in those appropriate cases, it is only if what the defendant deserves to pay as punishment exceeds what the plaintiff deserves to receive as compensation, that the plaintiff can be also awarded the amount of the excess. This is a windfall which he receives because the case happens to be one in which exemplary damages may be awarded.

 

It is not necessary to dwell upon the three considerations which Lord Devlin referred to as arising from the nature and function of punitive damages. The first consideration qualifies the categories of cases in which exemplary damages may be awarded. The plaintiff must himself have been the victim of the conduct of the defendant which merits punishment: he can only profit from the windfall if the wind was blowing his way. The second consideration is relevant to the attitude of an appellate court to a jury’s assessment of exemplary damages. I have already taken it into account in forming my conclusion that the jury’s award of £40,000 ought to be set aside. The third conclusion relates to the relevance of the defendant’s means to any assessment of punitive damages in excess of the amount required to compensate the plaintiff.

 

These three considerations are followed by the crucial exposition of the way in which a jury should be directed in a case in which it is open to them to award punitive damages. I have already dealt with this in the first criticism which I have made of the summing-up at the trial in the instant case.

 

It should perhaps be pointed out that Lord Devlin did not suggest that in a case which clearly came within a category which justified an award of exemplary damages the jury should be invited to make separate awards in respect of the compensatory and the punitive element, although no doubt a judge sitting alone should do so. It was only in cases where it might be doubtful whether exemplary damages were permissible that he suggested that special verdicts splitting the total award might serve a useful purpose in avoiding the necessity of a new trial in the event of appeal.

 

It has not been contended that those parts of Lord Devlin’s speech  [*1127]  which expounded the rationale of the award and the assessment of exemplary damages in those cases in which they could be recovered did not serve a useful purpose which lay well within the functions of this House in its judicial capacity. It brought some order out of chaos, some light and reason into what was previously a dark and emotive branch of the common law. What has been criticised is the decision of legal policy to restrict the categories of cases in which exemplary damages may be awarded.

 

If the common law stood still while mankind moved on, your Lordships might still be awarding bot and wer to litigants whose kinsmen thought the feud to be outmoded — though you could not have done so to the plaintiff in the instant appeal, because defamation would never have become a cause of action. The common law would not have survived in any of those countries which have adopted it, if it did not reflect the changing norms of the particular society of which it is the basic legal system. It has survived because the common law subsumes a power in judges to adapt its rules to the changing needs of contemporary society — to discard those which have outlived their usefulness, to develop new rules to meet new situations. As the supreme appellate tribunal of England, your Lordships have the duty, when occasion offers, to supervise the exercise of this power by English courts. Other supreme appellate tribunals exercise a similar function in other countries which have inherited the English common law at various times in the past. Despite the unifying effect of that inheritance upon the concept of man’s legal duty to his neighbour, it does not follow that the development of the social norms in each of the inheritor countries has been identical or will become so. I do not think that your Lordships should be deflected from your function of developing the common law of England and discarding judge-made rules which have outlived their purpose and are contrary to contemporary concepts of penal justice in England, by the consideration that other courts in other countries do not yet regard an identical development as appropriate to the particular society in which they perform a corresponding function. The fact that the courts of Australia, of New Zealand and of several of the common law provinces of Canada have failed to adopt the same policy decision on exemplary damages as this House did for England in Rookes v, Barnard affords a cogent reason for re-examining it; but not for rejecting it if, as I think to be the case, re-examination confirms that the decision was a step in the right direction — though it may not have gone as far as could be justified.

 

The award of damages as the remedy for all civil wrongs was in England the creature of the common law. It is a field of law in which there has been but little intervention by Parliament. It is judge-made law par excellence. Its original purpose in cases of trespass was to discourage private revenge in a primitive society inadequately policed, at least as much as it was to compensate the victim for the material harm occasioned to him. Even as late as 1814 Heath J. felt able to say: “It goes to prevent the practice of duelling, if juries are permitted to punish insult by exemplary damages”: Merest v. Harvey (1814) 5 Taunt. 442, 444.

 

No one would today suggest this as a justification for rewarding the victim of a tort for refraining from unlawful vengeance on the wrong-doer. Conversely, the punishment of wrong-doers today is regarded as the function of the state to be exercised subject to safeguards for the accused  [*1128]  assured to him by the procedure of the criminal law and with the appropriate punishment assessed by a dispassionate judge and not by a jury roused to indignation by partisan advocacy. One of the most significant and humane developments in English law over the past century and a half has been the increasing protection accorded to the accused under our system of criminal justice. As my noble and learned friend, Lord Reid has pointed out no similar protection is available to a defendant as a party to a civil action.

 

So the survival into the latter half of the twentieth century of the power of a jury in a civil trial to impose a penalty on a defendant simply to punish him had become an anomaly which it lay within the power of this House in its judicial capacity to restrict or to remove; though it would have been anticipating by two years the recent change in the practice of this House if to have done so would have involved overruling one of its own previous decisions.

 

Lord Devlin’s analysis of previous decisions disclosed three kinds of cases in which the courts had recognised the right of a jury to award damages by way of punishment of the defendant in excess of what was sufficient to compensate the plaintiff for all the harm occasioned to him. The categorisation was new. Its purpose has, I think, been misunderstood. No one suggests that judges, when approving awards of exemplary damages in particular cases in the past consciously differentiated between one kind of case in which exemplary damages could be awarded and another. They dealt with them all as falling within a single nebulous class of cases in which the defendant’s conduct was such as to merit punishment. The purpose of Lord Devlin’s division of them into three categories was in order to distinguish between factual situations in which there was some special reason still relevant in modern social conditions for retaining the power to award exemplary damages, and factual situations in which no such special reason still survived.

 

With this end in view Lord Devlin extracted from the single nebulous class which appeared to be all that had been consciously recognised as justifying an award of exemplary damages at common law, two categories of cases in which this House decided that there were special reasons why the power to award exemplary damages should be retained. These two (apart from cases where exemplary damages are authorised by statute) are generally referred to as “the categories.” But there is also to be found in the previous cases a third category, consisting of the remainder of the single nebulous class in which this House decided that the anomalous practice of awarding exemplary damages in civil proceedings ought to be discontinued.

 

The first category comprised cases of abuse of an official position of authority. This would seem to be analogous to the civil law concept of dŽtournement de paver, with the imitation that it must involve the commission of an act which would be tortuous if done by a private individual. The cases cited are two hundred years old. It would not appear that the actual conduct of the defendant himself need justify an award of aggravated damages. In Huckle v. Money (1763) 2 Wils. 205 the defendant appears to have treated the plaintiff with courtesy and consideration. The servant was the whipping-boy for the political head of the government.  [*1129]  Nor need he have known that his act was wrongful. Money, a mere subordinate official, can hardly have been expected to know that general warrants issued by the Secretary of State were illegal. In Wilkes v. Wood (1763) Lofft. 1, however, it was said that a belief that the act impugned was lawful could be pleaded in mitigation of damages.

 

The second category was of cases where an act known to be tortuous was committed in the belief that the material advantages to be gained by doing so would outweigh any compensatory damages which the defendant would be likely to have to pay to the plaintiff. This would seem to be analogous to the civil law concept of enrichessement indue subject to a similar limitation that the act resulting in enrichment must be tortuous. The cases cited by Lord Devlin do not include underground trespass to minerals, which provide the classic examples in the nineteenth century of this category of tort. There is high authority both in this House (Livingstone v. Rawyards Coal Co. (1880) 5 App.Cas. 25) and in the Privy Council (Bulli Coal Mining Co. v. Osborne [1899] A.C. 351) that in the case of wilful clandestine trespass to minerals the damages may be assessed at the market value of the minerals without deduction for the cost of working an award which would exceed both the loss to the plaintiff and the profit to the defendant from his wrongful act. The excess is punishment.

 

The third — and rejected — category is numerically by far the largest. It consists of cases in which the manner in which the tort has been committed has attracted a whole gamut of dyslogistic judicial epithets such as wilful, wanton, high-handed, oppressive, malicious, outrageous; particularly those where the defendant’s manner of doing the tortuous act has been characterised by arrogance or insolence or, in the preferred Australian phrase, a contumelious disregard of the plaintiff’s rights. These are nearly all cases in which “aggravated damages” by way of compensation apart from punishment can be awarded and much of the previous confusion about exemplary damages stems from this.

 

Apart from this confusion or perhaps because of it, I do not doubt that it was the general understanding of English judges and of those who practiced in the English courts that exempLary damages by way of punishment of the defendant as well as aggravated damages by way of compensation of the plaintiff could be awarded in cases which fall within the third category. Lord Devlin’s speech in Rookes v. Barnard explicity acknowledges this. It was an understanding which he himself had shared. He had given effect to it in his own summing-up in Loudon v. Ryder [1953] 2 Q.B. 202.

 

The decision of legal policy which this House made in Rookes v. Barnard was to retain the first two categories and to discard the third as obsolete.

 

In describing the two categories retained I have deliberately departed from the ipsissima verb a of Lord Devlin’s description of them. His statement of the categories was not intended as a definition to be construed as if it were enacted law. They were retained because this House considered that there were circumstances in which a power to award exemplary damages still served a useful social purpose and the descriptive words must be understood in the light of the social purpose which they were designed to serve.

 

My Lords, had I been party to the decision in Rookes v. Barnard I doubt  [*1130]  if I should have considered it still necessary to retain the first category. The common law weapons to curb abuse of power by the executive had not been forged by the mid-eighteenth century. In view of the developments, particularly in the last twenty years, in adapting the old remedies by prerogative writ and declaratory action to check unlawful abuse of power by the executive, the award of exemplary damages in civil actions for tort against individual government servants seems a blunt instrument to use for this purpose to-day. But if it is to be rotunda question which cannot arise in the instant appeal — the reasoning which supports its retention would not confine it to torts committed by servants of central government alone. It would embrace all persons purporting to exercise powers of government, central or local, conferred upon them by statute or at common law by virtue of the official status or employment which they held.

 

I have no similar doubts about the retention of the second category. It too may be a blunt instrument to prevent unjust enrichment by unlawful acts. But to restrict the damages recoverable to the actual gain made by the defendant if it exceeded the loss caused to the plaintiff, would leave a defendant contemplating an unlawful act with the certainty that he had nothing to lose to balance against the chance that the plaintiff might never sue him or, if he did, might fail in the hazards of litigation. It is only if there is a prospect that the damages may exceed the defendant’s gain that the social purpose of this category is achieved — to teach a wrong-doer that tort does not pay.

 

To bring a case within this category it must be proved that the defendant, at the time that he committed the tortuous act, knew that it was unlawful or suspecting it be unlawful deliberately refrained from taking obvious steps which, if taken, would have turned suspicion into certainty. While, of course, it is not necessary to prove that the defendant made an arithmetical calculation of the pecuniary profit he would make from the tortuous act and of the compensatory damages and costs to which he would render himself liable, with appropriate discount for the chances that he might get away with it without being sued or might settle the action for some lower figure, it must be a reasonable inference from the evidence that he did direct his mind to the material advantages to be gained by committing the tort and came to the conclusion that they were worth the risk of having to compensate the plaintiff if he should bring an action.

 

I see no reason for restoring to English law the anomaly of awarding exemplary damages in the third category of cases. If malice with which a wrongful act is done or insolence or arrogance with which it is accompanied renders it more distressing to the plaintiff his injured feelings can still be soothed by aggravated damages which are compensatory. I share the scepticism expressed by Windeyer J. in Uren v. John Fairfax & Sons Pty. Ltd., 117 C.L.R. 118, 151-152 as to whether what was in the defendant’s mind at the time he committed the tort really increases the injury to the plaintiff’s feeling. I think too that an evanescent sense of grievance at the defendant’s conduct is often grossly over-valued in comparison with a lifelong deprivation due to physical injuries caused by negligence. But my own equable temperament may be idiosyncratic and the law of “aggravated damages” does not call for closer examination in the instant appeal.

 

Finally, on this aspect of the case I would express my agreement with the  [*1131]  view that Rookes v. Barnard was not intended to extend the power to award exemplary or aggravated damages to particular torts for which they had not previously been awarded; such as negligence and deceit. Its express purpose was to restrict, not to expand, the anomaly of exemplary damages.

 

My Lords, there is little that I should wish to add to what Lord Hailsham of St. Marylebone L.C. and my noble and learned friend, Lord Reid, have already said about the way the instant case was treated in the Court of Appeal. It is inevitable in a hierarchical system of courts that there are decisions of the supreme appellate tribunal which do not attract the unanimous approval of all members of the judiciary. When I sat in the Court of Appeal I sometimes thought the House of Lords was wrong in over-ruling me. Even since that time there have been occasions, of which the instant appeal itself is one, when, alone or in company, I have dissented from a decision of the majority of this House. But the judicial system only works if someone is allowed to have the last word and if that last word, once spoken, is loyally accepted.

 

The Court of Appeal found themselves able to disregard the decision of this House in Rookes v. Barnard by applying to it the label per incuriam. That label is relevant only to the right of an appellate court to decline to follow one of its own previous decisions, not to its right to disregard a decision of a higher appellate court or to the right of a judge of the High Court to disregard a decision of the Court of Appeal. Even if the jurisdiction of the Court of Appeal had been co-ordinate with the jurisdiction of this House and not inferior to it the label per incuriam would have been misused. The reasons for applying it were said to be: first, that Lord Devlin had overlooked two previous decisions of this House in E. Hulton & Co. v. Jones [1910] A.C. 20 and Ley v. Hamilton, 153 L.T. 384; secondly, that the “two categories” selected as those in which the power to award exemplary damages should be retained had not been previously suggested by counsel in the course of their arguments.

 

I find the suggestion that E. Hulton & Co. v. Jones, the leading case on innocent defamation, is to be regarded as an authority for an award of exemplary damages, quite unacceptable. Ley v. Hamilton was discussed at some length in Lord Devlin’s speech. I myself agree with his interpretation of Lord Atkin’s speech. The Court of Appeal did not and in this they now have the powerful support of my noble and learned friend, Viscount Dilhorne. But, however wrong they may have thought Lord Devlin was, they cannot have thought that he had overlooked Ley v. Hamilton.

 

The second reason I find equally unconvincing. On matters of law no court is restricted in its decision to following the submissions made to it by counsel for one or other of the parties. After listening to a lengthy argument which embraced a full examination of a large and representative selection of the relevant previous authorities this House was fully entitled to come to a conclusion of law and legal policy different from that which any individual counsel had propounded.

 

With regard to the amount of exemplary (and also aggravated) damages which may be awarded where the plaintiff elects to sue defendants jointly for a single tort, I agree with Lord Hailsham of St. Marylebone L.C. that the Court of Appeal got it wrong. Where I differ from him is in thinking that  [*1132]  the trial judge got it right. I am fortified in this view by the fact that Lord Denning M.R. understood the summing-up as leaving to the jury a choice whether to award a sum appropriate as a punishment of the more blameworthy of the defendants or the less blameworthy or something in between the two sums. Salmon L.J. appears to have taken the same view. Both thought that this was a correct statement of the law. In this I think that they were mistaken as to the law, but right as to what the jury would have understood the summing-up to mean.

 

On the wider aspects of the course adopted by the Court of Appeal it is best that I should content myself with expressing my concurrence with all that Lord Hailsham of St. Marylebone L.C. has said.

 

LORD KILBRANDON. My Lords, there are several reasons which induce me to be as brief as I can. First, the case in its important general aspects is concerned with doctrines, and to some extent with procedures, with which I am not familiar. Secondly, those general aspects have been examined in great detail and in an authoritative manner by your Lordships who have preceded me. Thirdly, since it is unlikely that any contribution of mine would be regarded as of value in clarifying the law of England, I may at least wind up the consideration of a disastrous case with economy, the lack of which, especially in this class of litigation, is, as others of your Lordships have observed, a notoriously discreditable feature of our jurisprudence. In short, having had the advantage of reading the speeches prepared by my noble and learned friends, Lord Hailsham of St. Marylebone L.C., Lord Reid and Lord Morris of Borth-y-Gest, I agree with them.

 

It is conceded by the appellants that they libelled the respondent, and they do not attack as excessive the sum awarded by the jury as compensation for the damage they did to his feelings and his reputation. It is also conceded that, if there was evidence upon which a properly directed jury could find that the appellants had calculated that they might make a profit from publication which might exceed the compensation payable to the plaintiff, then, since “one man should not be allowed to sell another man’s reputation for profit,” and since it may be “necessary to teach a wrongdoer that tort does not pay,” the jury were entitled to award punitive damages, on the authority of Rookes v. Barnard [1964] A.C. 1129, 1227. The first question, and one which from first to last occupied a very great deal of time in your Lordships’ House, was whether there was such evidence.

 

I have no doubt on this point at all, and I do not rehearse the evidence. The jury had before them the state of the appellants’ knowledge before publication — that the respondent had warned them that he regarded certain passages as libellous, that professional naval opinion was to the same effect, and, above all, that another reputable publisher had refused to handle the book because of its defamatory character. The appellants’ attitude is demonstrated by their written references to libel actions as affording “first class publicity,” and to “tightening up the indemnity clause.” No doubt there was an element of the jocular in these remarks, but they do show that the appellants were going ahead with their eyes open as to consequences, and they must have thought it would be worth their while.

 

Counsel for the appellants pointed out, and I for one agree, that since  [*1133]  all commercial publication is undertaken for profit, one must be watchful against holding the profit-motive to be sufficient to justify punitive damages: to do so would be seriously to hamper what must be regarded, at least since the European Convention was ratified, as a constitutional right to free speech. I can see that it could be in the public interest that publication should not be stopped merely because the publisher knows that his material is defamatory; it may well be in the public interest that matter injurious to others be disseminated. But if it were suggested that this freedom should also be enjoyed when the publisher either knows that, or does not care whether, his material is libellous — which means not only defamatory but also entreat would seem that the scale is being weighted too heavily against the protection of individuals from attacks by media of communication.

 

The conduct of the appellants, accordingly, is in my view brought within the principle of the rule laid down in Rookes v. Barnard to which I have just referred. If a publisher knows, or has reason to believe, that the act of publication will subject him to compensatory damages, it must be that, since he is actuated by the profit-motive, he is confident that by that publication he will not be the loser. Some deterrent, over and above compensatory damages, may in these circumstances be called for.

 

This leads me to the little I have to say on the doctrine of punitive damages. I do not propose to discuss its merits or demerits, because I agree with Lord Devlin, not only that it forms part of the law of England, but also that its abolition would not be within the judicial functions of this House. I will, however, add that I am not convinced that any statutory example of the recognition of the doctrine is to be found. By the Law Reform (Miscellaneous Provisions) Act 1934 it is provided that where a cause of action survives for the benefit of the estate of a deceased person, the damages recoverable for the benefit of that estate shall not include any exemplary damages. In the previous subsection provision had also been made, per contra, for causes of action subsisting against the estates of deceased persons. Since punitive damages are punitive or deterrent against the author of them, it would have been understandable if the statute had refused to allow them against a dead man. But, instead, they have been disallowed when they are claimed in respect of an injury to a dead man. This leads me to suppose that by the phrase “exemplary damages” Parliament was here referring to what are usually called “aggravated” damages; the estate of a dead man must pay them in order to indemnify the living, but the estate of a dead man, whose feelings, post mortem, have become irrelevant, does not receive them.

 

In the same sense I would interpret section 13 (2) of the Reserve and Auxiliary Forces (Protection of Civil Interests) Act 1951 which provides for the award, in certain circumstances, of “exemplary damages.” Section 13 (2) applies, by virtue of section 13 (6), to Scotland, and since I can hardly believe that this Act introduced for the first time, as it were by a side-wind, the doctrine of punitive damages into the law of Scotland, I conclude again that “exemplary” really means “aggravated.” Aggravated damages, in the English sense, are available to pursuers in defamation cases in Scotland, subject to this qualification, that the conduct of counsel (cf. Greenlands Ltd. v. Wilmshurst and London Association for Protection of Trade [1913] 3 K.B. 507) is not accepted as an aggravation unless that conduct has been  [*1134]  on the express instructions, or with the privity, of counsel’s client — see James v. Baird, 1916 S.C. 510. Finally, Lord Devlin, at p. 1225, doubted whether section 17 (3) of the Copyright Act 1956 authorised an award of exemplary damages: in my opinion it did not.

 

I do not suppose that anyone now sitting down to draft a civil code would include an article providing for punitive damages. But the doctrine exists, and in my respectful opinion the rationale of it is explained, by illustrations as apt as one could find, in the speech of Lord Devlin. The doctrine proceeds upon the footing, whether sound or not, that in some torts, and in some circumstances, there is an element of public interest to be protected. The only way in which that can be done may be by awarding to a plaintiff a sum of damages which he does not deserve, being in excess of any loss or injury he has suffered; that sum includes an element calculated to deter the defendant, and other like-minded persons, from committing similar offences. One example, which is Lord Devlin’s second category, I have already noticed — the publisher who does not mind paying compensatory, even aggravated, damages for libel, because he will still have a profit after paying them. It is not in the public interest, especially as the publishing agencies become more and more monolithic, that such conduct should go unchecked, and no remedial measures other than punitive damages seem to be open. A second example — Lord Devlin’s first category — is in the sphere of public authority. While, as some of your Lordships have pointed out, the illustration may have been too narrowly drawn, the rationale is clear, and is the same. An example might be an outrageous excess of official authority without any aggravating circumstances (cf. Huckle v. Money, 2 Wils. 205) resulting in the wrongful imprisonment of a person of bad character. False imprisonment is primarily actionable as an injury to reputation. If the plaintiff has none to lose, the amount of his compensatory damages may be inadequate to deter, in the public interest, flagrant injustices of this character.

 

The exclusion of the “common bully” category, and the consequent overruling of Loudon v. Ryder [1953] 2 Q.B. 202, are entirely consistent with this principle. Very large compensatory damages, which should be adequate deterrent, are proper in such cases, and in most of them the criminal law can also take care of the public interest.

 

I accordingly accept that the case of Rookes v. Barnard, as it has now been expounded by my noble and learned friend, Lord Reid, correctly states the law of England. It cannot be said, and it does not purport, to state the law of Scotland; it may be that in other parts of the Commonwealth also it is not, for what may be very different reasons, acceptable. Nevertheless it appears to me to give content to the doctrine of punitive damages, and to set proper limits upon it.

 

The trial having been correctly and inevitably conducted upon the basis of Rookes v. Barnard as then understood, the question now arises whether the learned judge gave the jury adequate and accurate directions in law on that basis. First, did he fail to make it clear to the jury that, if they had made an award of compensatory damages, any additional award by way of punitive damages could be made “if, but only if,” the amount of the compensatory damages did not itself constitute a sufficient deterrent? The second objection was that the learned judge gave an inadequate  [*1135]  direction as to the course to be followed by the jury should they find punitive damages due, but a different degree of culpability in the two defendants. I think it is sufficient for me to say that I agree with those of your Lordships who are of opinion that the directions, in both matters, were adequate.

 

The aspect of the case which has given me the greatest difficulty is the question whether the total amount of the damages awarded is so excessive that the verdict cannot stand. That it is excessive I do not doubt, but that is not a sufficient reason for the award to be set aside. The assessment of damages in such cases as this is not, in our law, a judicial function. In so far as compensatory damages are concerned, it may well be right that that should be so. If he were called on to estimate the sum appropriate to repair the injured feelings and damaged reputation of a citizen who had been defamed, a judge would be making not a legal, but something more like a social, assessment: there is no reason to suppose that his estimate would more probably be correct than would that made collectively by any twelve sensible men and women. So when one looks at a jury’s award in such a case one has to ask, whether it could have been made by sensible people acting reasonably, or whether it must have been arrived at capriciously, unconscionably or irrationally. On that test I think the present award must stand. Moreover, it is not unprecedented. For example, in a case in which the libel was in some ways less wounding than the present — Youssoupoff v. Metro-Goldwyn-Mayer Pictures Ltd., 50 T.L.R. 581 — an award, adjusted for the change in money values, of well over twice as large as this was upheld by experienced judges.

 

The same test, as the law now stands, must be applied to a jury’s award of punitive damages. Whether this should be so is another matter; it is arguable that the assessment of punishment is not properly a jury’s function, and ought more readily to be challengeable on appeal to a judicial authority. It is obvious that, as counsel for the appellants forcibly pointed out, a defendant against whom punitive damages is sought stands to a great extent stripped of the constitutional safeguards which would be his right were he arraigned before a criminal court. One of those safeguards is a calm judicial determination of the penalty appropriate to his offence. Perhaps, if the doctrine of punitive damages is to be retained, it ought to be made a condition precedent of their being asked for that the plaintiff forgo his right to have the case tried by a jury; it is not likely that a defendant would wish to stand on his own right in that respect.

 

So, although I would myself have assessed the damages at a much smaller sum, I cannot say that the award, on the principles under which we now operate, ought not to stand, or that, were a new trial to be ordered, the result would, in my confident opinion, be substantially different.

 

Finally, I do not consider it necessary for me to say anything on the issue of the relations between this House and the Court of Appeal, except that I entirely agree with what has fallen from my noble and learned friend the Lord Chancellor on this topic.

 

I would, accordingly, dismiss this appeal.

 

Appeal dismissed.