Mitigation of damages and the Burstein principle
32 In Warren v Random House Group Limited [2009] QB 600, the Court of Appeal of England and Wales explained the Burstein principle in the following terms (at [78]):
The decision of this court in Burstein v Times Newspapers Ltd [2001] 1 WLR 579, cited above, established two important interlocking propositions. (a) In relation to the court's assessment of damages for libel it is open to a defendant to seek to rely upon such facts as fall within the "directly relevant background context" to the defamatory publication. See in particular the judgment of May LJ, at para 42. (b) It is illogical and undesirable that a defendant can seek to rely upon such facts in relation to such assessment only if he has presented them as part of a substantive defence to liability, in particular within a plea of justification of the publication. He can rely upon them as freestanding matters pleaded as relevant only to the assessment of damages: see in particular the judgment of May LJ, at para 47.
33 That rather simple statement of the propositions flowing from Burstein somewhat belies the uncertainty that has arisen concerning the application of those propositions. That uncertainty mainly revolves around the vague and unhelpful expression "directly relevant background context". Judges are often rightly sceptical when the tender of evidence is sought to be justified on the basis that it provides "background" or "context". Those words often conceal or obscure, rather than explain, whether or why the evidence is relevant. Careful attention to what was actually decided in Burstein, however, tends to remove some of the uncertainty.
34 In Burstein, the claimant claimed that a publication in "The Times" was defamatory of him because it stated that he "used to organise bands of hecklers to go about wrecking performances of modern atonal music". He bought libel proceedings against the publisher. The publisher relied on a defence of fair comment, though the trial judge proceeded to strike out that defence. There was, therefore, "no effective issue but that the words complained of were defamatory". In relation to the assessment of damages, the trial judge refused to allow the publisher to rely, in reduction of damages, on facts it had pleaded in support of its defence of fair comment. Those facts, the particulars of which are set out at [10] of the judgment of May LJ, all involved the claimant's conduct as a co-founder of a "group of militant campaigners against modernist atonal music which styled itself 'the Hecklers'". The Court of Appeal of England and Wales found that the trial judge erred in excluding evidence of the particularised facts.
35 May LJ (at [19]) noted that the evidence which is properly admissible in reduction of damages for defamation includes evidence of the claimant's bad reputation, evidence properly before the court on some other issue and evidence of the claimant's own conduct. The admissibility of evidence of the claimant's reputation centred, May LJ noted (at [20]) on the decision of the Queens Bench Divisional Court in Scott v Sampson (1882) 8 QBD 491. That case established that evidence of general bad reputation was admissible in reduction of damages, but that "evidence of rumours that the plaintiff had done what was charged in the libel and evidence of particular acts of misconduct on the part of the plaintiff tending to show his character and disposition were inadmissible".
36 In relation to the second category of evidence, evidence properly before the court on some other issue, May LJ referred (at [21]) to Pamplin v Express Newspapers Ltd (Note) [1988] 1 WLR 116 in which Neill LJ stated:
But a defendant is also entitled to rely in mitigation of damages on any other evidence which is properly before the court and jury. This other evidence can include evidence which has been primarily directed to, for example, a plea of justification or fair comment … There may be many cases, however, where a defendant who puts forward a defence of justification will be unable to prove sufficient facts to establish the defence at common law and will also be unable to bring himself within the statutory extension of the defence contained in section 5 of the Defamation Act 1952. Nevertheless the defendant may be able to rely on such facts as he has proved to reduce the damages, perhaps almost to vanishing point. Thus a defence of partial justification, though it may not prevent the plaintiff from succeeding on the issue of liability, may be of great importance on the issue of damages.
37 This "apparently straightforward state of law and practice" had, according to May LJ, led to "unsatisfactory tactical manoeuvring" (at [22]). In short, defendants had pleaded weak cases of justification so as to be able to use the facts pleaded in support of that unsuccessful defence in reduction of damages.
38 May LJ referred at length (at [29]-[34]) to Speidel v Plato Films Ltd [1961] AC 1090, a case in which the House of Lords gave detailed consideration to Scott v Sampson. The House of Lords confirmed that evidence as to the "circumstances under which the alleged libel was published" was inadmissible and that evidence of particular acts on the part of the plaintiff could not be given in reduction of damages where the defendants had failed to justify the libel complained of. At [35], May LJ said:
I have quoted at length from the speeches in Speidel v Plato Films Ltd to show that a main concern was to prevent libel trials from becoming roving inquiries into the plaintiff's reputation, character or disposition; that what was held to be inadmissible was particular facts said to be relevant to the plaintiff's general reputation or disposition; and that the case does not decide that particular facts directly relevant to the context in which a defamatory publication came to be made are inadmissible. Indeed, two of the speeches by inference accept that evidence of the circumstances surrounding the publication are admissible.
39 To understand the reference, in that passage, to "facts directly relevant to the context in which a defamatory publication came to be made", it is necessary to go back to the speech of Lord Denning in Speidel, which was one of the two speeches that May LJ was plainly referring to. In Speidel, Lord Denning had given an example of a publication that a man had been convicted six times for dishonesty, when in fact he had only been convicted twice. According to Lord Denning, in such a case, the defendant could adduce evidence of the two convictions "in partial justification" because, otherwise "the plaintiff would recover damages for a character which he did not possess or deserve; and this the law will not permit".
40 It was in that context that May LJ said (at [42]):
In my view, permitting the defendants to rely on the directly relevant background context in the way in which I have described would not offend anything said in Scott v Sampson 8 QBD 491 or Speidel v Plato Films Ltd [1961] AC 1090. The material to which I have referred as directly relevant background context was, as I have indicated, recognised in Speidel v Plato Films Ltd as being admissible as the circumstances in which the publication came to be made. In the present case, those circumstances are not sensibly limited to the concert in memory of John Smith and the fact that the claimant's music was played at it. For practical purposes, every publication has a contextual background, even if the publication is substantially untrue. In addition, the evidence which Scott v Sampson excludes is particular evidence of general reputation, character or disposition which is not directly connected with the subject matter of the defamatory publication. It does not exclude evidence of directly relevant background context. To the extent that evidence of this kind may also be characterised as evidence of the claimant's reputation, it is admissible because it is directly relevant to the damage which he claims has been caused by the defamatory publication.
41 May LJ then referred to some authorities relevant to what had earlier been referred to as "unsatisfactory tactical manoeuvring" and said (at [47]):
In my view, these authorities show that it is not permissible to advance an unsustainable defence of justification and thereby, under the guise of particulars of justification, seek to rely on particulars which Scott v Sampson 8 QBD 491 and Speidel v Plato Films Ltd [1961] AC 1090 would not permit. That, however, does not prevent a defendant from frankly accepting that there is no proper plea of justification, but seeking to rely in reduction of damages on particulars which Scott v Sampson and Speidel v Plato Films Ltd do not exclude. If this were not so, there is a danger that the jury would be required to assess damages in blinkers, in ignorance of background context directly relevant to the damage which the claimant claims has been caused by the defamatory publication. This is consistent with the sense of what Lord Denning said in Speidel v Plato Films Ltd, where his conclusion was that it was permissible to adduce the evidence in question. Thus, in my view, a defendant is not prevented from taking the sensible course of accepting that the publication means what it says and that it is not on the facts justified, and yet putting in evidence directly relevant facts which in other circumstances might have been ingredients of a defence of justification. It would, I think, be illogical and unfair if this were not so. Evidence in support of a plea of justification which fails is admissible in reduction of damages. But the very same evidence would not be admissible to a sensible defendant who acknowledges that it will not support a plea of justification. What is not permissible is to plead a defence of partial justification which in truth is no defence at all.
42 It is tolerably clear, when paragraphs [42] and [47] of the judgment of May LJ are read in context, that the expression "directly relevant background context" is a reference to evidence of misconduct on the part of the claimant which, while not sufficient to make out a defence of justification, nevertheless was conduct which was in the same sector of the claimant's life as the defamatory publication and was therefore directly relevant to his or her reputation. The example given by Lord Denning in Speidel is a classic example of such evidence. "Directly relevant background context" does not include rumours or mere allegations that the claimant had done what was charged in the libel. That type of evidence is properly excluded by what was said in Scott v Sampson. Nothing said by May LJ in Burstein suggested that Scott v Sampson did not continue to apply in the case of such evidence.
43 Notwithstanding this, the Burstein principle has been productive of some uncertainty, mainly because the expression "directly relevant background context", if not properly considered and understood in context, is apt to obscure exactly what sort of evidence can be admitted pursuant to the principle. That point was made in Turner v News Group Newspapers Ltd [2006] 1 WLR 3469, where Keene LJ (with whom Moses LJ and Pill LJ agreed) said (at [56]):
I accept the point made in argument that it is somewhat repetitive to use the words "background" and "context" in the phrase "directly relevant background context", but that in itself does not produce obscurity. It is in any event inevitable that cases will occur where it is not easy to determine whether the test in Burstein's case is met or not. That does not mean that the test is an inappropriate one, any more than is that propounded in Scott v Sampson: as Viscount Simonds recognised in Speidel's case, the line between evidence of general bad reputation and evidence of specific conduct giving rise to such a reputation is not easy to draw. What constitutes the directly relevant background will vary from case to case, but I would myself accept the need for the courts to proceed, as Mr Browne advocates, with some caution in applying Burstein's case, given that it represents a modification of the long-standing rule in Scott v Sampson. As Eady J put it in Polanski v Condé Nast Publications Ltd (unreported) 21 October 2003, one should guard against extending too creatively the concept of "directly relevant background". The Court of Appeal in Burstein's case was concerned to avoid jurors having to assess damages while wearing blinkers. If evidence is to qualify under the principle spelt out in Burstein's case, it has to be evidence which is so clearly relevant to the subject matter of the libel or to the claimant's reputation or sensitivity in that part of his life that there would be a real risk of the jury assessing damages on a false basis if they were kept in ignorance of the facts to which the evidence relates.
44 Moses LJ explained the difficulties in the following way (at [87]-[90]):
The purpose of the approach taken by the Court of Appeal in Burstein's case is clear; it was to ensure that the claimant was properly vindicated and fairly compensated. Exclusion of the facts identified by Sir Christopher Slade, at paras 59 and 60, would not have achieved that result. The problem, as in so many branches of the law, lies in the expressions used to identify and describe the principle deployed to admit that evidence. May LJ referred to "relevant background context": para 41 or "background context directly relevant to damage": para 47. Ever forgetful of the need not to read judgments as if they were statutes, the words "directly relevant background context" have been used as a label to be stuck on by a defendant or ripped off by a claimant. But the words themselves are no substitute for the reasoning of the court in Burstein's case.
The difficulty may lie, I suggest, in the fact that the label, which it has been sought to apply in subsequent cases, does not itself greatly assist in identification. Once detached, it does not, pace Keene LJ at para 56, provide a test, since it assumes a meaning to the words of description it employs. To describe a fact as "directly relevant" has no meaning without identification of the issue to which it is alleged to be relevant. The word "context" may itself be misleading. It is accepted that, following Burstein's case, facts may be admitted, notwithstanding that they did not themselves cause or provoke the publication of the defamatory material; that must follow from the decision in Burstein's case itself. Further, there is no requirement that the facts should have been known to the publisher at the time of publication; there is no logic in such a requirement to achieve that which the principle seeks to achieve, a fair measure of damages. In those circumstances to ask whether the facts form part of the context in which the defamatory material was published does not seem to me to be of help to anyone save the lawyers who may gain much by endless debate as to what the label signifies. Nor does it assist to speak of the jury or judge being kept "in blinkers", unless it is clear what is to be hidden from their eyes.
Keene LJ refers to "directly relevant background" and Sir Christopher Slade similarly refers to "directly relevant background facts"; the reference to "background" draws a distinction from those facts which relate directly to the defamatory words, which, ex hypothesi, after an offer of amends, cannot be used in justification or to deploy the defence of fair comment. After an offer of amends, only those facts which are directly relevant to the existence and extent of the alleged damage may be admitted, unless a claimant seeks to rebut the presumption in section 4(3) of the Defamation Act 1996. The key lies in identifying that which is directly relevant to the issue of damages. That lies, in my view, in recalling the reference by Cave J in Scott v Sampson to "uniform propriety" cited by Keene LJ at para 29. A claimant's life may appropriately be considered in sectors: see Lord Denning in Speidel v Plato Films Ltd [1961] AC 1090, 1140 and Goody v Odhams Press Ltd [1967] 1 QB 333, 341a. A defendant may seek to reduce the damages by adducing evidence which is directly relevant to a claimant's conduct or reputation in the particular sector to which the defamatory material relates for the purpose of mitigating damage.
Direct relevance to a particular sector of the claimant's life should be assessed by starting with a careful identification of the sector of the claimant's life to which the defamatory material relates; too broad an approach, such as his private life, will defeat the function of the reference to "direct relevance". Further restriction may be achieved by scrutiny of the purpose for which it is sought to adduce the facts in issue. It should be recalled that the purpose of the rule in Scott v Sampson was by no means only a desire to keep a case within manageable proportion. The rule serves the important function of preventing a claimant from being terrorised into submission. It is worth recalling that, alongside its failed attempt to abrogate the rule in Scott v Sampson, the Neill Committee advised that a defendant making an offer of amends should not be permitted to attack a claimant's character. Such an attack, once the rule had been abolished, should only be made before a jury: see the Supreme Court Committee Report on Practice and Procedure in Defamation (July 1991), p 79, para VII.28. But that advice has gone, along with the recommendation to abolish the rule.
45 Those passages from Turner support the proposition that the facts which are able to be pleaded and proved in the mitigation of damages pursuant to the Burstein principle must concern specific conduct that is directly relevant to either the subject matter of the alleged defamatory statement, or the claimant's reputation in the part of his or her life the subject of the defamatory publication. The rationale for permitting evidence of such facts being led is that otherwise damages may be assessed on a false basis. It is equally clear that courts, including this Court, must proceed with caution in applying Burstein, should guard against "extending too creatively" the concept of "directly relevant background", and should subject the proposal to adduce facts under the Burstein principle to careful scrutiny. Mere resort to the label "directly relevant background context" will not suffice.
46 The Burstein principle has been accepted in Australia: Australian Broadcasting Corporation v McBride (2001) 53 NSWLR 430; Channel Seven Sydney Pty Ltd v Mahommed (2010) 278 ALR 232 at [262]-[265]); Holt v TCN Channel Nine Pty Ltd (2014) 86 NSWLR 96; Gacic v John Fairfax Publications Pty Ltd (2015) 89 NSWLR 538; Coxon v Wilson [2016] WASCA 48; and, most recently, Fairfax Digital Australia & New Zealand Pty Ltd v Kazal [2017] NSWCA 77.