Two high-profile celebrity libel cases have recently ended with a substantial financial settlement in the Claimant’s favour: undisclosed, following a ruling on meaning and liability, in the case of Sally Bercow and Lord McAlpine; and a £125,000 damages payment in the case of the “Celebrity Psychic” Sally Morgan versus the Daily Mail, reached privately with the matter already due for trial.
Each case has raised eyebrows, but both drive home the pivotal role played by the framing of meaning – as pleaded by the parties or as handed down by the judge – in determining the odds of any libel claim. As the Defamation Act 2013 switches for good the presumption in libel trials against juries, so it seems likely (following Tugendhat J’s lead in McAlpine v Bercow) that a preliminary judicial ruling on meaning will also become the norm.
These developments both derive from the laudable intention of cost-saving, and (for better or worse) it seems likely cases will frequently be settled in private off the back of these preliminary hearings. Of course there is the initial question of whether or not the words are even defamatory. But even beyond this, each party will seek to frame the meaning of the words according to the strength or weakness of their own case evidentially: so a shifting of the territory on which the claim is to be fought can be instantly fatal to the case of one or other party.
It is important first to consider the test of meaning, and the key elements. The current test is explained by Sir Anthony Clarke MR in Jeynes v News Magazines Limited ( EWCA Civ 130):
“The legal principles relevant to meaning have been summarised many times and are not in dispute…. They may be summarised in this way: (1) The governing principle is reasonableness. (2) The hypothetical reasonable reader is not naïve but he is not unduly suspicious … (3) Over-elaborate analysis is best avoided. (4) The intention of the publisher is irrelevant. (5) The article must be read as a whole, and any “bane and antidote” taken together. (6) The hypothetical reader is taken to be representative of those who would read the publication in question.”
Whilst the intention of the author and/or publisher is irrelevant to meaning, to be determined by the hypothetical reader, it may impact upon defences and damages. Hence it is understandable why some Defendants might wish to maintain a stoical “innocent face” about what they meant. All the same, a pragmatic approach to meaning tends to be more productive than a stubbornly optimistic one.
Lawyers love to read in aggressive and perhaps even ‘over-elaborate’ meanings to words complained of, especially – with no firms mentioned – when firing off the initial broadsides; it is part of the heat and noise of a letter of claim. But a realistic approach is necessary in order to meet the governing test of reasonableness; hence the parties’ ground is likely to drift somewhat, even before a matter reaches judicial scrutiny.
Case 1: McAlpine v Bercow
In reasonableness terms, the Defendant in McAlpine v Bercow ( EWHC 1342 (QB)) appeared on a hiding to nothing once her case rested on a completely literal view of what was, and was found to be, an arch and knowing tweet (“Why is Lord McAlpine trending? *Innocent face*”).
Given the elaborate stage wink, it was never credible to expect a representative follower of Ms Bercow – a politically-minded consumer of news – to believe that she was simply asking a naïve question unconnected to any other allegations. This meaning relied perhaps too much on the Defence’s case on timing: that the rumours were not quite so rife at the time of the tweet’s publication as they later became. The more intriguing question, though now academic, is whether arguing for such a brazenly literal meaning was a necessary, or helpful, position for her to take.
Where Ms Bercow most clearly came aground was a further rule in libel: that, as Tugendhat J put it, “a defendant who repeats a defamatory allegation made by another is treated as if he had made the allegation himself”. This is trite law, and its practical intent – to avoid a defence of idle gossip – is clear. However, the facts of this case would have made it an interesting one on which to test the strict effect of the so-called ‘Repetition Rule’, if that was how the defence was run.
From the common law position on neutral reportage, and its inclusion in the 2013 Act, it may be inferred that the repetition rule is not intended to be absolute – in the sense of always importing liability on the same terms as the original publication. Indeed, Tugendhat J considered this point (at 45: “The harshness of that [repetition] rule is tempered by another rule, known as reportage”), but applied the two rules consecutively: one as a strict rule of publication (namely that the repetition carried the full original defamatory meaning), the other as a possible basis of defence or rebuttal.
This is certainly one view of Lord Phillips’ dictum in Flood v Times Newspapers (at 34): “In a case of reportage qualified privilege enables the defendant to avoid the consequences of the repetition rule.” What was not fully pleaded, or considered, in McAlpine v Bercow is how far the reportage elements in the repeat publication might at least act to dilute the gravest meaning conveyed by the original allegations.
The reportage argument was raised in submissions but not formally pursued by Ms Bercow’s team; it is far from clear that Ms Bercow’s ‘account’ of what one might broadly term a ‘dispute’ between Lord McAlpine and a third party was sufficiently neutral, bearing as it did so much ironical weight. However, the principle remains: that repetition by way of mere observation, even casually done (where the matter is of public interest), does not always ‘mean’ full endorsement. It is notable that Tugendhat J flagged this point, whilst remarking in so many words that it was not open to him to apply it (he did go on to imply the odds were against such an argument actually succeeding)(see ).
The Defence’s case did include a submission that if, contrary to primary submissions, the tweet was found to be defamatory, it should still be construed as being a less serious allegation than the highest Newsnight allegations. But the Defendant’s primary case felt like an all-or-nothing meaning, and it was hard to climb down from this position ‘in the alternative’. Ms Bercow (who, it should be noted, had sought terms of settlement before the hearing) was left snookered by a meaning which required, in effect, a finding of pure naivety; not her own (intention being irrelevant) so much that of as her readers, who would have had to be incapable of drawing the relevant inference.
But was there any more attractive meaning open to the Defendant to argue – a *knowing face*, perhaps, but not an imputation of guilt; one which acknowledged the reality of the tweet’s tone, but fell short of making the allegation at its highest? In short, could she have successfully persuaded the Court that the tweet may have linked Lord McAlpine to the Twitter gossip of the day, but did not automatically ‘endorse’ in full the BBC allegations which she was found in effect to repeat?
A more credible meaning for Ms Bercow’s tweet, and (but for the strict application of the repetition rule) a more defensible one, might have been along the following lines: “Based purely on Twitter’s trending topics, it seems Lord McAlpine is being widely linked to a major news story; but [by *innocent face*] no inference should be drawn as to the accuracy of the general speculation, or the news story, or my own views.”
It is accepted that pleading the meaning above would be a high-risk tactic and, as with making arch use of the word ‘allegedly’, still at risk of incurring some liability in any event; but no more so than by playing completely dumb (as indeed it proved). Accordingly it is worth a moment’s consideration, if nothing else, on the issue of seriousness.
On the facts, it must be admitted the Newsnight story was overwhelmingly the likely candidate for the reference; even if it was not the only story of the moment, Lord McAlpine fit the description of its central protagonist. Under the repetition rule, if the reader were to make the connection to her tweet, Ms Bercow could not normally expect to wash her hands of all the false and serious allegations made in that programme. The unusual position here – also the damaging aspect of Ms Bercow’s publication to her 56,400 followers – was that Lord McAlpine had not been named in the unsafe BBC investigation. So Ms Bercow’s chief mistake was to draw attention the trending name.
This cannot be said to be the only basis of liability. Though a range of tweets naming Lord McAlpine were litigated upon, in many cases being settled on lesser terms, it seems by no means certain that every tweet which named Lord McAlpine that morning – even coming from a person of a certain political prominence – would have been found to impute a meaning that he was a paedophile and serial abuser.
It is not hard to construct tweets which could credibly carry a meaning similar to Ms Bercow’s which Lord McAlpine might have hesitated to pursue to Court, even if they identified him more directly with the very serious allegations. This applies equally to the more guileless and transparent (“Twitter users might be naming Lord McAlpine as a paedophile, but I am not so sure” … “Is Twitter trying to tell me that Lord McAlpine is the man in the BBC Newsnight programme?”) as to the more subtly ironical (“I’ve just been reminded of Lord McAlpine’s spotless political career by Twitter”).
Ms Bercow’s may not have been entirely innocent, but it was not far removed in meaning from these hypothetical examples. These other tweets could be argued to contain elements of doubt or denial – or at least ‘bane and antidote’ together – but there was an element of judicial discretion at play in electing the highest meaning. The faux-naivety of Ms Bercow’s tweet might not stand up as being wholly literal, nor wholly neutral; but the judge did not have to conclude, as suggested by the Claimant, that it actively pushed the reader, ‘avid for scandal’, to draw the gravest possible conclusion. It might simply have encouraged the reader to do some research of their own.
Furthermore, although Twitter has become a go-to forum for speculation, sometimes accurate, as to ‘secret’ celebrity identities (often those otherwise protected by injunctions or Court procedure), the site – and its users – are also not known as being consistently reliable. Many names get tossed into the mix, and sometimes competing rumours will ‘trend’ at the same time. A reasonable representative Twitter follower of Ms Bercow would surely have understood this, and not necessarily taken the tweet to mean there was good reason to believe Lord McAlpine was actually a paedophile, and guilty. (It should be noted that Tugendhat considered the case law on this point ( to ) concluding that the discretion in the matter was his own).
What Ms Bercow was so recklessly commenting on, and arguably what she was repeating, seems to have been not so much the sting of the programme itself but more the frenzied online speculation. Admittedly, in repeating the libel of a fellow Twitter user – or a significant number of them, as the name was trending – Ms Bercow was infected by their liability. She became part of the casual finger-pointing. It was not practical for the Court to consider taking a poll or identifying a ‘mean’ or modal meaning of what other Twitter users were saying about Lord McAlpine; the judge had to consider the words complained of themselves. But hers was a tweet which at least in one credible reading addressed the audience quizzically, not affirmatively; as if to say, “Come off it Twitter users, I am not naïve, I know what this is about. What IS going on these Lord McAlpine allegations, then?”
Tugendhat J’s judgement left no doubt as to how he viewed the matter. He applied the rule strictly and with no room for a lesser meaning:
“the effect of the repetition rule is that the Defendant, as the writer of the Tweet, is treated as if she had made, with the addition of the Claimant’s name, the allegation in the Newsnight and other media reports which had previously been made without his name. It is an allegation of guilt. I see no room on these facts for any less serious meaning.” [emphasis added]
But – insofar as it is legitimate to speculate on such things – we cannot now tell how the judge’s instinct on the merits of the argument, particular as to the seriousness of the allegation () would have been informed by the unappealing path chosen by the Defendant as to meaning. The brazen literalism of what may be termed the “innocent face defence” left little room for mitigation or grey area. Perhaps the knowing tone of Ms Bercow’s tweet, coupled with the perception that she was a Westminster ‘insider’, gave rise to a damning inference that she knew more than she was saying. Or perhaps the judge simply preferred the Claimant’s evidence. Either way, the judgment was certainly more damning for the fact that the Defendant’s arguments on meaning were not held to be credible.
Whilst the ruling will no doubt be considered a landmark in Twitter libel, it does not set out a clear precedent in respect of reportage, nor its relationship with meaning. Perhaps consideration on both fronts ought to be given to the limitations on a user’s power of expressing themselves in a clear and balanced way in 140 characters, or the built-in function to ‘Retweet’ uncritically. The prescribed maximum length of a Tweet is not a great deal to play with: some users put little or no thought into how they use the allocation; others put considerable thought into the medium, as if composing a haiku or telegram. So issues of context and repetition, of inference and innuendo will always be present; what is left unsaid will be as important as what is said.
On such basis, it is still reasonable to think (notwithstanding point (3) of the Jeynes test) that the more elaborate, drawn-out meaning still has a place in modern libel actions. This is especially so where a meaning which is ‘pure’– in either sense of the word – simply does not wash.
Case 2: Morgan v Associated Newspapers Limited
The case of Morgan v Associated Newspapers Limited has also been controversial, not least among ordinary members of the public (albeit guided by the media – the traditional media, rather than the plural of ‘medium’). This is principally because of an assumption that, essentially, a performing psychic accused of a fraud or scam upon their audience could hardly evidence the good faith of her professional activities without proving the supernatural.
Of course, this misunderstands the prevailing burden of evidence: once the meaning is found to be defamatory, it is for the Defendant to justify their libel. With this in mind, imputations as to the Claimant’s lack of good faith – which is never easy to prove (almost on the level of witchcraft, in the view of some libel reformers) – would certainly have been incorporated into her pleaded meaning.
Equally important was a specific, central allegation about Ms Morgan’s stage act in the Daily Mail article: that she was receiving secret audio instructions from accomplices via an earpiece. If the Defendant newspaper could not justify this – as became clear it could not (the ‘stooges’ in fact being theatre staff unconnected to the Claimant) – then it could not hope to defend any meaning which was contingent on these facts, nor rely on fair comment in relation to them.
Not uncommonly, even where a Claimant has obtained an unreserved apology and retraction, the joint Statement in Open Court (here) appears to disclose a tension between the pleaded meanings of each party. For the Defendant’s part, its statement, retraction and apology (which following convention is much the shorter part) was restricted entirely to the specific allegation about the use of messages to cheat the audience, which it accepts was untrue. No mention is made of the wider thrust of the article (in the Claimant’s case, that psychics were “charlatans”), nor of any word such as ‘scam’ or ‘fraud’.
It would seem however, from the drafting of the statement, that the Claimant understood the article more generally to mean she had “deliberately and dishonestly perpetrated a scam on [and later, “cheated”] her audience in Dublin”. Clearly included are the elements of intent, which would have presented the Defence with a considerable obstacle: however, such a meaning (isolated from the false allegations of audio prompts) would still have been open to interesting arguments in cross-examination. What is it the audience expected from a psychic show: pure entertainment? A form of communion and comfort? Actual communication with the dead? And to what extent did Ms Morgan believe in the literal truth and accuracy of her powers, and how was she satisfied of them?
In short, the Defendant only formally retracted and regretted the specific error that was the article’s sting. There hangs thereby the suggestion that the case might have been taken to trial if it had become (as Ms Morgan told press it was not intended to be) a wider ‘audit’ of the world of psychics generally. However, it is clear from both parties’ statements that the wider meaning (of dishonesty) was materially linked to the allegations about the secret audio instructions, “whereby” the alleged “scam” on her audience was said to have been perpetrated.
The test of clairvoyance must wait for another day, and most likely a new statutory regime. In terms of how the Court will approach that regime – well, the job of reading minds and predicting the future will be once more in the hands of legal advisors. The test of what meaning a reasonable reader reads is an objective one, even if at times it seems akin to a psychic power. But with more preliminary hearings on meaning expected, it seems that more than ever that this will be a key battleground of the action: an area where the merits of any defences hang in the judicial balance, and where the personal instinct of an individual judge will play a considerable part.
Owen O’Rorke is an Associate at M Law LLP