Determining meaning on social media – context is key, but how much?

May 20, 2022
laptop and phone

Two recent decisions from the Media & Communications list highlight the tricky nature of determining the single meaning of words complained of when those words are published on social media.

Wilson v Mendelsohn concerned a Facebook post about the claimant published by an unconnected third party, in which Mr Wilson was called a “weirdo” and a “freak”, and accused of harassing a mother on the school run and taking pictures of her, her car and her daughter [1].  The Facebook post was then re-published on Twitter by the second and third defendants in the course of an argument with the claimant.  

Sivananthan v Vasikaran concerned posts in a WhatsApp Group of around 20-30 political Tamil activists who were all well known to each other [2].  The posts accused Mr Sivananthan of misleading or misinforming Boris Johnson in the course of his 2019 Conservative Party leadership campaign in a manner that undermined the hard work of the activists, and of taking credit for the campaigning work of other Tamil activist groups.

In each case, the offending post was found to be defamatory at common law.

Hearings on meaning

The Wilson hearing was held before Mr Richard Spearman QC sitting as a Deputy Judge.

The Sivananthan hearing on meaning was held before the honourable Mrs Justice Collins Rice.

General principles and recent case law

In each case, the judge was guided in their analysis by a similar set of precedents that will be familiar to practitioners in this area, namely:

  • Koutsogiannis v Random House Group [2020] 4 WLR 25 (at paragraphs 10 to 12);
  • Stocker v Stocker [2019] 2 WLR 1033 (in particular, paragraphs 41 to 47);
  • Riley v Murray [2020] EWHC 977 (QB) (in particular, paragraphs 13 to 18);
  • Vardy v Rooney [2020] EWHC 3156 (QB);
  • Monroe v Hopkins [2017] 4 WLR 68; and
  • Monir v Wood [2018] EWHC 3525 (QB).

To summarise, when looking at the meaning of social media posts, a judge must keep in mind the way in which such posts are usually made and read (Stocker, paragraph 41).  Twitter and Facebook are considered to be fast-moving media, where people scroll through messages relatively quickly (Monir, paragraph 90) and do not undertake a detailed analysis of each of the words used in a given post. 

Of course, these propositions may be of less relevance in a case where it would be clear to the ordinary reader from the outset that the post was meant to convey a message of some importance (Vardy, paragraph 25). 

For instance, in Wilson, the claimant argued that the Facebook post operated as the 21-century equivalent of a “Wanted” poster, which was only merited by some sort of serious wrongdoing.  Mr Spearman QC alluded to this observation at paragraph 48 of the judgment, even though he did not ultimately find that it supported the most serious imputation sought by the claimant (i.e. that the Facebook post carried the innuendo meaning that he was a paedophile, as considered further below).

In order to determine the natural and ordinary meaning of the statement of which the claimant complains, it is necessary to take into account the context in which it appeared and the mode of publication (Koutsogiannis, paragraph 12).  

The question in social media cases will be how much of the surrounding contextual material should the judge take into account when determining the meaning of the words.

  • On the one hand, a tweet may well need to be read as part of a series of tweets which the ordinary reader will have seen at the same time as the tweet complained of, or beforehand.  A matter can be treated as part of the context in which an offending tweet is to be read if it is on Twitter and sufficiently closely connected in time, content or otherwise such that it is likely to have been in the hypothetical reader's view, or in their mind, at the time when they read the words complained of.  This would clearly include an earlier tweet or reply that was available to view on the same page as the offending material. It could include earlier material, if sufficiently closely connected.  But it is not necessarily the case that it would include tweets from days beforehand (Monroe, paragraphs 34, 38 and 39). 
  • On the other hand, as was noted by Nicklin J in Riley v Murray at paragraph 10, "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".

Sivananthan summary of the judicial approach 

The approach to be taken by the court in determining meaning was stated in clear terms by Mrs Justice Collins Rice in Sivananthan at paragraphs 5 and 36 of her judgment: 

“I first read the posts complained of, without knowing what either party wanted to say they meant. I formed and noted some provisional views.  I then read the pleadings and the skeleton arguments.  I heard oral submissions and reserved judgment.

I remind myself that my task is not to agree or side with one party or another about meaning but, having formed some initial quick views about these few posts, to consider what the parties say about them, look at such context as appears to me to be proper, reasonable and relevant, and test my initial views, before doing what I am required to do by law: find a single natural and ordinary meaning.  I must look at the posts as they would have been read, in such of their immediate conversational context as I have. I remind myself that social media comment is fast-moving, consumed quickly and reacted to impressionistically and briefly, but that the letter which features in this particular conversation needs to be properly understood for the light it sheds on it.”

Application of principles in Wilson and Sivananthan

In both Wilson and Sivananthan, the meaning of the offending posts was found to be defamatory at common law. 


In Wilson, the offending posts formed part of a larger argument on Twitter between the parties.  Both parties had submitted their own evidence as to the context in which the tweets were posted, and there was some dispute between them as to how certain threads had started.  As users of Twitter will know, such arguments can arise from long “threads” with different overlapping conversations and commentary taking place in the “replies” to individual posts within that thread.  How those tweets appear on the screen may vary depending on the individual message that the user has highlighted.  

That did not ultimately affect the meaning of the words complained of, as the judge found that the defendants’ tweets added nothing to the meaning of the Facebook post that they had each republished, and they were found to have adopted the entirety of the Facebook post in their republication.

The judge found that the natural and ordinary meaning of the Facebook post was that: (a) the claimant objected to a mother leaving her car engine running while dropping her daughter off at junior school, banged on her car window, was very rude to her, and took pictures of her, her car and her daughter, which he retained on his phone; and (b) that conduct was unwarranted and worrying, was the conduct of a weirdo and a freak, and amounted to harassment.

In the judge’s view, both of those statements were defamatory at common law.


In Sivananthan, the posts formed part of a WhatsApp group conversation which had been set up over a year previously.  The offending posts themselves were several months apart, and the claimant and defendant admitted to having a disputatious relationship.  Curiously, however, neither side had provided the surrounding messages from the WhatsApp conversation to provide contextual evidence for the meaning of the offending posts, and both were criticised in the judgment for their failure to do so. 

One of those WhatsApp posts also specifically directed that its recipient should “CAREFULLY read Boris’s letter” – referring to a letter from Boris Johnson that had been posted in the group by Mr Sivananthan a few days before the first offending post.  The letter was seemingly sent to the claimant by Mr Johnson after they met at an event for sympathetic focus groups.  The letter spoke of “deepening ties with countries such as Sri Lanka” with trade deals after Brexit, and praised the claimant’s group (the British Tamil Conservatives) for their campaigning.  

That did not sit well with the defendant, as a foreign policy of deepening trade ties with that regime is the opposite of what their campaigning was seeking to achieve.  On the contrary, there had been a concerted effort to encourage the imposition of trade sanctions. The defendant also said that Mr Johnson’s praise for the BTC’s campaigning arose out of a trip to the UN Human Rights Council, which had actually been organised by another group, and so the claimant must have been taking credit for other people’s efforts.  

Given the importance of the letter to the context of the WhatsApp posts, and given that both sides accepted all the people in the WhatsApp group were likely to have read the letter, Collins Rice J considered it was necessary to take the letter into consideration as part of the context in which the WhatsApp posts were made when reaching a decision on meaning.

Accordingly, the judge concluded that the natural and ordinary meaning of the words complained of was that: (a) Mr Sivananthan misled or misinformed Mr Johnson, causing Mr Johnson to refer in his letter to “deepening ties” with Sri Lanka, in turn having the effect of undermining or reversing the hard work, and wasting the resources, others in the Tamil community have put in to promoting a very different foreign policy towards Sri Lanka and/or attracting support for the Conservative Party; (b) Mr Sivananthan, actively or by omission, took the credit given in the letter to the BTC for taking parliamentarians to the UNHRC in Geneva, when it was the BTF that did so (which, with other behaviours, also undermines others in the Tamil community); and (c) Mr Sivananthan acted in breach of the BTC constitution in order to advance his own interests.

The judge considered that the meaning of those allegations (whether allegations of fact or opinion) was defamatory in tendency at common law.


Each decision was a preliminary ruling on meaning, and each claimant would still need to establish that the posts complained of also pass the threshold set out in section 1 of the Defamation Act 2013 – i.e. that they have caused, or are likely to cause, serious harm to the claimant’s reputation.  As Collins Rice J noted, that would require looking beyond the intrinsic meaning of the words and considering the facts and evidence about their impact.  Then, if serious harm can be shown, the defendant would need to establish that the factual allegations are substantially true, the opinions are justifiable, or that some other available defence applies.

While the exercise to be taken by the court in determining meaning may not be straightforward in every case, there is already a rapidly developing body of case law to act as guidance.  It is abundantly clear that judges will expect to see at least some of the material showing the context in which social media posts are made, and practitioners should include a reasonable sample of the contextual material in their evidence.

[1] Wilson v Mendelsohn & Ors [2022] EWHC 715 (QB).

[2] Sivananthan v Vasikaran [2022] EWHC 837 (QB).

Hugo MasonHugo Mason
Hugo Mason
Hugo Mason

News & Insights