Riley v Murray

This is a libel action concerning the following Tweet, posted by the Defendant at

At the time of the Tweet, the Defendant was employed as the Stakeholder Manager in the office of Jeremy Corbyn, then the leader of the Labour Party. The Claimant complains that she has been defamed by the Tweet. In her Particulars of Claim, the Claimant contends that the natural and ordinary meaning of the Tweet (“the Claimant’s Meaning”) was:

“The Claimant had publicly supported a violent attack upon Jeremy Corbyn at a mosque by saying that he deserved it. She has shown herself to be a dangerous person who incites unlawful violence and thuggery and is therefore so beyond the pale that people should boycott her and her tweets.”

In the alternative, the Claimant contended that this meaning was conveyed by innuendo. The Particulars of Innuendo relied upon were:

“The words complained of were published on Twitter. Users of Twitter use the word “engage” to mean to interact with another user, whether by reading their (sic) tweets, liking their tweets, retweeting their tweets and so on. Therefore all or a substantial number of the publishees would have understood the words complained of to bear [the Claimant’s Meaning].”

The Claim was commenced on 31 May 2019. The Defendant has not yet filed a

The Claimant’s case is that, at the date of publication of the Tweet, the Defendant had 7,245 followers on Twitter. The Tweet was retweeted 1,544 times and liked by 4,738 people. The Claimant contends that the Tweet would also have been republished via email, WhatsApp and other forms of communication outside Twitter. It is unlikely that it will be possible to identify the full extent of publication or to identify every one of the publishees. Although not on the scale of a national newspaper publication, the Tweet was published generally, rather than to a limited and identifiable group.

When ordering the trial of the Preliminary Issues, Master Yoxall also gave further directions, including:

Although the Claimant complained of an innuendo meaning, the factual issue to which it gives rise is of a limited compass - whether publishees of the Tweet were aware of the special meaning attributed to “engage” – it appears that the directions for disclosure and witness statements were made because the Defendant wished to put before the Court factual matters beyond the Tweet itself.

The Defendant’s Case was duly filed on 29 October 2019 and included the following:

The Defendant has filed a witness statement dated 20 December 2019. The Claimant has filed no evidence. The Defendant’s witness statement contains sections headed “Background” – setting out her current and previous roles in the Labour Party; “The Claimant’s social media activity”; “Owen Jones’ tweet of 10 January 2019”; “The attack on Jeremy Corbyn and the subsequent tweets” and “Use of the word ‘engage’”.

It is unusual for the court to be asked to consider evidence when determining the natural and ordinary meaning of a publication and whether the hypothetical ordinary reasonable reader would have understood the words to be making an allegation of fact and/or expression of opinion. The reason it is unusual is that evidence is not admissible on the issue of what readers understood an allegedly defamatory publication to mean. The assessment of the single natural and ordinary meaning of words is wholly objective. Neither the meaning the publisher intended to convey, nor the meaning the publishees actually understood the publication to bear is relevant. From all the authorities that could be cited for these principles, it probably suffices to refer to Charleston -v- News Group Newspapers Ltd [1995] 2 AC 65, 70 and Slim v- Daily Telegraph Ltd[1968] 2 QB 157, 171-172, 173, 174. What publishees actually understood a statement to mean can sometimes become relevant if a defendant seeks to establish that, notwithstanding publication of an objectively defamatory single meaning, the claimant’s reputation was not actually harmed seriously (a point of potential relevance under s.1 Defamation Act 2013 and in any assessment of damages): Monir -v- Wood [2018] EWHC 3525 (QB) [196(iv) and (v)]. Such an exercise can only usually be expected to be fruitful in cases where the individual publishees can be identified. It might be thought that these are fundamental principles of defamation law that have been clearly established over many decades, across the Commonwealth. However, there is a creeping tendency, under the guise of alleged “context”, to attempt to adduce evidence extrinsic to the words complained of on the issue of the natural and ordinary meaning. The submissions made in this case are the latest, but it must be said one of the more ambitious, efforts to rely upon

With the consent of the parties, no hearing took place to hear the submissions of the parties. Instead, I have considered the written submissions of the parties on the issues to be determined. In accordance with the practice I outlined in Hewson -v- Times Newspapers Ltd [2019] EWHC 650 (QB) [25], copies of the parties’ written submissions will be made available with copies of this judgment.

Most of the principles relevant to determining these preliminary issues are not controversial.

There is equally no dispute that the context of the publication is very important when ascertaining meaning (and fact/opinion): Jeynes -v- News Magazines Ltd [2008] EWCA Civ 130 [14]; Bukovsky -v- CPS [2018] 4 WLR 13 [13]; Koutsogiannis [12(ix)]; Tinkler -v- Ferguson [2019] EWCA Civ 819 [15]. Mr Hudson QC is correct to submit that the importance of assessing a publication in its proper context received clear endorsement from the Supreme Court in Stocker -v- Stocker [2019] 2

In Stocker, the Supreme Court was dealing with the meaning of a posting on Facebook. Reference must be made to the Court of Appeal judgment to see the immediate context in which the words, “he tried to strangle me”, were published: [2018] EMLR 15 [11]. Stocker was an important restatement of existing principles of defamation law in relation to modern methods of communication. It re-emphasised the importance of taking the hypothetical ordinary reasonable reader to be a person who would read, and react to, a publication “in a way that reflected the circumstances in which it was made” [39] and that “the way in which the words are presented is relevant to the interpretation of their meaning” [40]. By so doing, the Supreme Court was not overthrowing the principles governing the admissibility of extrinsic evidence on the natural and ordinary meaning of words in defamation actions. Indeed, in [42],

In Monroe -v- Hopkins, Warby J went on to explain the limits of what could be admissible as “context”:

The underlined passages establish that the following material can be taken into account when assessing the natural and ordinary meaning of a publication:

The fundamental principle is that it is impermissible to seek to rely on material, as “context”, which could not reasonably be expected to be known (or read) by all the publishees. To do so is to “erode the rather important and principled distinction between natural and ordinary meanings and innuendos”: Monroe -v- Hopkins [40]. When I considered this principle very recently, I explained that the distinction was between “material that would have been known (or read) by all readers and material that would have been known (or read) by only some of them. The former is legitimately admissible as context in determining the natural and ordinary meaning; the latter is relevant only to an innuendo meaning (if relied upon)” (emphasis in original): Hijazi -v- Yaxley-Lennon [2020] EWHC 934 (QB) [14]. As Warby J noted in Monroe -v- Hopkins [38], the second principle is influenced by the test for whether two publications are to be read together and treated as a single publication for the purposes of ascertaining their meaning.

Applying these principles can raise sometimes fine questions of judgment – for example the extent to which the ordinary reasonable reader is taken to read hyperlinks (see Falter -v- Atzmon [2018] EWHC 1728 (QB) [12]-[13]; Poulter -v- Times

For the Claimant, Mr Bennett QC makes his submission in a single paragraph of his written submissions. The ordinary reasonable reader would have read the Tweet, standing alone, in one go (before moving further down his or her timeline) and absorbed its very simple message. The Tweet was self-contained. Nothing in it suggested that the reader needed to bear in mind or refer to any other information in

Mr Hudson QC for the Defendant introduces his submissions by setting out what he contends are the following “material facts” (these being largely drawn from the Defendant’s witness statement):

Mr Hudson QC submits that whether read in isolation or “in its proper context”, the Tweet would be understood by the ordinary reasonable readers to contain the following statements:

The first sentence, he contends, is a statement of fact. On reading the second sentence, he argues, the ordinary reasonable reader would understand it to mean that the Claimant had posted a tweet following the attack on Mr Corbyn which was concerned with and commenting on the attack on Mr Corbyn. The reader would appreciate that the Defendant was not reproducing the Claimant’s tweet by, for example, “quotetweeting” it or by setting it out in the Tweet. The reader would realise that the Defendant was describing or summarising the meaning and effect of the Tweet and would understand that it was setting out the Defendant’s impression of what the Tweet was conveying. Mr Hudson QC submits that the hypothetical ordinary reasonable reader: “would understand [the Tweet] to mean that the Claimant had posted a tweet (following the attack on Jeremy Corbyn) which (in the Defendant’s view) meant (or conveyed) that Jeremy Corbyn deserves to be violently attacked because he is a Nazi”. The third, fourth and fifth sentences of the Tweet set out the Defendant’s reaction to, and views on, the conduct of the Claimant in posting the tweet she did about the attack on Mr Corbyn. Readers would understand the Defendant to be saying that, in posting such a tweet, the Claimant’s behaviour was dangerous and stupid and that, as a result, the Defendant’s followers should not engage with the Claimant by replying or responding to the Claimant’s tweets on such matters. The word ‘Ever’ (on its own) would be understood simply to be emphasising the point that her followers should not engage with the Claimant.

To support this construction of the Tweet, Mr Hudson QC then advances a detailed argument, over several pages of his written submissions, as to the “relevant context”, relying principally on the matters I have identified above. However, he argues that, “the meaning that an ordinary reasonable reader would derive from [the Tweet]… is simply reinforced when [it] is read in its proper context. It does not lead to a different meaning.”

In answer to Mr Hudson QC’s case on “context”, Mr Bennett QC contends that they have no, or no sufficient, connection with the Tweet to justify being treated as having been read together with it. The Defendant cannot rely upon information which the reader might discover after reading the Tweet. More generally, Mr Bennett QC asks why the Court should assume that the hypothetical reader read a further tweet A, but not tweets B, C, D and so on. If the Defendant’s argument on “context” is correct, then the Court would have to assess every tweet published relevant to the publication complained of by way of context.

This is a straightforward case. My findings are these:

I have removed some parts of the Claimant’s meaning which appear to me either unjustifiably to gloss the words complained of, or to add nothing to the defamatory sting of the meaning. I agree with Mr Hudson QC that “engage” is an ordinary English word. It needs no further explanation in the natural and ordinary meaning. I address below the issue of whether it has an extended innuendo meaning to certain readers (see [30]-[32] below).

I reject Mr Hudson QC’s argument that the hypothetical ordinary reasonable reader would understand the second sentence of the Tweet to be a statement of opinion. This argument is untenable – with or without the elaborate argument based on “context”. The second sentence was a simple factual statement and would be understood as such; it provided a summary of the content of a tweet alleged to have been posted by the Claimant.

In any event, I reject the argument on context. None of the conditions for the material relied upon by Mr Hudson QC to be admissible on the question of the natural and ordinary meaning of the Tweet (see [16] above) is met.

An imputation that a person had publicly supported a violent attack on someone is plainly defamatory at common law; it is conduct which would substantially affect, in an adverse manner, the attitude of other people towards the Claimant or have a tendency so to do. Had it stood alone, the description of the Claimant as “dangerous” and “stupid” would also have been defamatory, but the gravity of the defamatory meaning is largely supplied by the allegation of fact rather than the expression of opinion based upon it.

Given my findings in relation to the natural and ordinary meaning of the Tweet, the innuendo meaning adds nothing, even if the supporting facts could be established. All that the Claimant has relied upon by way of evidence is a collection of printouts from various websites including from Twitter. I have been provided, for example, with pages titled “What fuels a Tweet’s engagement?” and “Introducing Twitter Engage”, which appears to relate to what is described as “a new companion app for Twitter”, providing “real-time data and insights, allowing you to quickly understand, engage, and grow your audiences… Twitter Engage surfaces (sic) the most important follows and @mentions from influencers and loyal fans, providing an effortless way to stay plugged into Twitter”. There is a printout of a page from the Twitter website containing “Definitions”. “Engagements” is defined as “total number of times a user interacted with a Tweet. Clicks anywhere on the Tweet, including Reteweets, replies, follows, likes, links, cards, hashtags, embedded media, username, profile picture, or Tweet expansion”. Also included in the hearing bundle – without any explanation that conventionally might be expected to be found in a witness statement – are pages from a blog, from Ben Brown, on a website www.bitcatcha.com. In it, Mr Brown tells his readers, “How I got 10,000 Twitter Followers in 3 months (and a 13% engagement rate)”.

In Mr Bennett QC’s written submissions as to the innuendo meaning, he set out the following:

“The Claimant's evidence regarding the way ‘engage’ is understood by Twitter users is [in the hearing bundle]. Engage means to engage with a tweet by retweeting it, liking it, replying to it etc. It connotes some sort of interaction beyond merely reading it. See in particular the definition of engagements... The Particulars of Claim state … that engagement means reading a tweet. This is not wholly correct. Whilst an engagement is evidence that someone has read a tweet, it is possible to read a tweet without engaging by not interacting with the tweet by enlarging it, liking it etc.”

I will not spend time trying to decipher the “evidence” that the Claimant has relied upon. I doubt that it provides any real evidence, or any evidence of any value. If there exists a category of reader who – submerged in the lexicon of Twitter – understands the word “engage” in some sort of Twitter-specific way, the meaning that s/he would understand is not materially different from the natural and ordinary meaning of the word as it appeared in the Tweet.