Defamation on Twitter – A Multi-Dimensional Conversation

Warby J’s decision in Monroe v Hopkins [2017] EWHC 433 (QB) has broad implications for users in the ‘wild west’ of social media. Its detailed analysis of the mechanics of Twitter – including the much talked about five-page “How Twitter Works” appendix – represents the courts’ very latest understanding of defamation online.

Factual Background

The Claimant and Defendant are both well-known public figures, and the general factual matrix has already been widely covered in the press. On 18 May 2015 D posted a number of tweets, including the two complained of:

  1. At 7.20pm: “@MsJackMonroe scrawled on any memorials recently? Vandalised the memory of those who fought for your freedom. Grandma got any more medals?”; and
  2. At 9.47pm: “Can someone explain to me – in 10 words or less – the difference between irritant @PennyRed and social anthrax @MsJackMonroe.

Prior to D sending the second tweet, C responded with three posts to the effect that the first tweet was a lie. D deleted the first tweet before posting the second tweet. No full apology was published, despite C’s solicitors sending proposed wording to D’s management on 21 May 2015. Proceedings were issued in December 2015.

Legal Principles

Warby J took the established approach to deciding liability in defamation by addressing: (1) the meaning(s) of the tweets; (2) whether the agreed meaning(s) had a defamatory tendency; (3) the extent of publication; and (4) whether the serious harm requirement was satisfied – s1 Defamation Act 2013.

In relation to (1), Warby J took into consideration the possibility that extraneous facts might affect the way that an ordinary reasonable person would understand a particular statement – i.e. that there could be an “innuendo” meaning (as per Fullam v Newcastle Chronicle & Journal [1977] and McAlpine v Bercow [2013]). By these means an otherwise innocent statement may be defamatory, or an otherwise defamatory statement innocent, in the eyes of readers aware of the innuendo facts [23(3)].

In the present case the innuendo facts relied upon by C included the fact of the vandalisation of a war memorial on Whitehall; the resulting outrage in the press and online; and D’s two tweets on 18 May 2015.


Given the conversational medium of social media, an “impressionistic approach” rather than an “elaborate analysis of a 140-character tweet” was said to be appropriate [35]. Warby J found that, taking a natural and ordinary view (rather than on a literal reading), the first tweet would be understood to mean that C “condoned and approved of scrawling on war memorials, vandalising monuments commemorating those who fought for her freedom” [43]. The second tweet was, in view of the innuendo fact that the first tweet had been published shortly beforehand, of equivalent meaning.

The tweets reached 20,000 and 100,000 readers respectively. Warby J accepted that, as a result, C was caused to feel “anxious and upset, and had difficulty sleeping”, though he noted that injury to feelings alone is insufficient to establish serious harm [64]-[67].

Although D posted the first tweet in error, both were held to have meanings with a defamatory tendency. Publication did cause real and substantial distress and harm to C’s reputation (albeit not serious or very grave). C was awarded compensation of £24,000, plus costs to be assessed. (This is a lower award than the £90,000 damages in Cairns v Modi – a Twitter case involving very grave allegations).

What Next?

A number of notable points arise from the judgment. Social media users would do well to bear in mind the below when posting messages online:

(1) Transience: The length of time that a post is online may be indicative of serious harm but it is not determinative – [71(2)]. Warby J drew comparison with the powerful impact of a live TV broadcast; even though the viewer may only see the live images on screen once they could still have a lasting effect. Print copies of newspapers are also rarely read more than once, yet they still have the capability to inflict serious harm. It is the impact of the post, not how long it is online, that is important.

(2) Credibility: Warby J did not attach weight to D’s submission that a tweet would prima facie carry less credibility in the eyes of publishees simply because the publisher was not as authoritative as e.g. a main stream news organisation/established institution – [71(3)]. Such ‘authority’ will turn on the facts, but a defendant may struggle to conclude that a reader would discount their post(s) simply because of the particular person (rather than publisher/company) from whom they originate.

(3) What Followers Actually Believe: The absence of evidence that an allegation was believed by C’s followers, i.e. the lack of any evidence that readers/followers changed their position as a result of D’s post, is not determinative [71(4)]. Warby J was not persuaded by D that the absence of such evidence meant that no harm was caused. (That is not to say that such evidence might not necessarily be helpful in future cases – it was just absent on C’s case here). Therefore, don’t expect a claimant to necessarily have to stand up serious harm by reference to how their followers react.

(4) Existing Reputation: Warby J was not convinced by D’s argument that those who abused C in the aftermath of D’s tweets were people who were already making the same or similar comments about C before D’s tweets – [71(7)]. He said that it would not be safe to infer that a claimant’s reputation has not been harmed by a specific defamatory allegation just because a person who makes rude remarks about the claimant after publication also made rude remarks about him/her before.

(5) A Shortcut To Serious Harm?: Absent the development of any complex factual/legal issues, Warby J confirmed that an allegation that has (1) a seriously defamatory tendency, and (2) is widely published, may give the basis for an inference that serious harm was actually caused. This means that in some cases it may be possible to establish serious harm without engaging “in a detailed forensic examination of the precise factual picture” [69].


Defamation on social media is still a relatively new area of law and this decision goes some way to clarifying the extent to which the courts will approach novel cases – no matter how transient or incredible the posting(s). Warby J also made a number of observations that could be useful to parties involved in such litigation in future:

  • Criticisms of C

A claimant who refers to a “torrent” of abuse suffered as a result of an allegedly defamatory posting must have strong supporting evidence. There was a factual dispute in the present case, and Warby J found some of C’s evidence unreliable. No serious distortion of the facts by C was evident, but he found ‘torrent’ to be something of an overstatement (and in any event issues of causation were also identified) [72]-[73].

C’s tweets posted subsequent to the first and second tweets by D on 18 May 2015 – in effect mocking D and “revelling in the prospect of suing her” – were found to be unhelpful to C’s case. However, Warby J did say that “the response of the accused is inherently unlikely to undo the damage caused initially” – [71(9)].

  • Criticisms of D

A number of issues were raised by Counsel for D in relation to disproving serious harm, most of which have been given short shrift in the judgment. Whilst Warby J generally accepted C’s evidence as to serious harm, he noted that in any event he would have found it hard to disregard in the absence of any positive case from D – who did not give oral evidence/fully explain her position in written statements produced by her solicitors. The Judge said that “if she [Hopkins] had done so, and persuasively rebutted what C says about these matters, I might have disregarded part of C’s evidence” – [66].

D did not plead to any innuendo. Warby J comments at [41] that, if a claimant has put forward an innuendo case, “there is nothing inherently unfair or wrong about the defendant adopting and relying on the same matters”. That is to say that where a claimant attributes meaning by reference to contextual factors, there is nothing in principle wrong with the defendant providing a detailed case about context by way of rebuttal. In the present case, D did not plead any such ‘defendant innuendo’ and Warby J’s statement that “a case should be stated, regardless of whether it is considered to be one of innuendo or context” at [41] appears to be critical of D.

Emphasis was placed by Counsel for C on the “defiant” nature of D’s second tweet [49]. The lack of any full public apology was also taken into consideration for the purposes of assessing serious harm, and Warby J found that its absence “allowed the claimant’s injured feelings to remain raw” [65]. The injury to C’s feelings was increased by D’s behaviour, and some reasonable mitigation by D could have ultimately prevented the onset of costly proceedings.