QB-2022-001397 - [2025] EWHC 2193 (KB)

It can be seen that the word “harass” appears within the meaning of the first article; “harassment” is part of the meaning of the eighth article; the meanings of the second, third, fourth, fifth and sixth articles incorporate the term “sexual harassment”; while the meaning of the seventh article includes the term “sexually harassing”.

In his closing submissions, the Claimant contended that “sexual harassment” ought to be defined by reference to s.26(2) of the Equality Act 2010 (‘the 2010 Act’). Several weeks after the conclusion of the trial, the Claimant filed further submissions retracting reliance on that provision and his “concession” that a single incident can be enough to constitute harassment (para ‎59 below). The Claimant seeks to rely instead on the Protection from Harassment Act 1997 (‘the 1997 Act’), which he contends is more appropriate in a “non-employment context”, and in particular the requirement that to be guilty of an offence of harassment, or the statutory tort, a person must be shown to have committed “a course of conduct”.

S.7(3) of the 1997 Act states:

“A ‘course of conduct’ must involve –

(a)

in the case of conduct in relation to a single person (see section 1(1)), conduct on at least two occasions in relation to that person”.

In his post-hearing submissions, the Claimant also drew attention to a definition of harassment, given in the Oxford English Dictionary, which he submits implies that repeated conduct is required.

I agree with the Defendant that the statutory definitions are irrelevant. The Claimant’s assumption appears to be that because the meanings found by Johnson J contain words which are defined in statutes, the court should apply one of those statutory definitions in determining the truth defence. That is wrong. Johnson J did not find, in relation to any of the Articles, that the words complained of alleged strong grounds to believe that the Claimant had committed an offence of harassment or the statutory tort of harassment, or that he had harassed female colleagues within the meaning of the 1997 Act (or the 2010 Act). Indeed, at the meaning trial, neither party contended that the statements complained of would have been understood by ordinary, reasonable readers as conveying a technical legal meaning.

In principle, it is possible for the words complained of to convey a meaning that the claimant has committed a particular legally defined wrong or offence, as was the case in Bukovsky v Crown Prosecution Service [2017] EWCA Civ 1529, [2018] 4 WLR 13. But when determining meaning, as the authors of Gatley on Libel and Slander (13^th ed., 2022) observe at 3-017:

“The question is what would the words convey to the mind of the ordinary, reasonable, fair-minded reader? The natural and ordinary meaning of words is not determined by any meaning that the law may attach to the same words. Words may acquire a specific meaning in law that an ordinary person would not understand the words to bear and consequently the same words could be found to have a meaning in a libel claim that is different from their technical legal meaning. Whether they do so is a question of fact that turns on the application of the principles of interpretation.”

It is clear that the meanings found by Johnson J are not linked to any statutory definition. In determining whether the Guardian has made out the defence of truth, the focus has to be on the meanings found by Johnson J, without any gloss or qualification.

In Stocker v Stocker [2020] AC 593 the Supreme Court held that it was not appropriate to have regard to dictionary definitions of words when deciding the meaning of a statement. That is consistent with the well-established principle that no evidence, beyond the statement complained of, is admissible in determining the meaning. Lord Kerr JSC (with whom all members of the Court agreed) observed:

“25.

Therein lies the danger of the use of dictionary definitions to provide a guide to the meaning of an alleged defamatory statement. That meaning is to be determined according to how it would be understood by the ordinary reasonable reader. It is not fixed by technical, linguistically precise dictionary definitions, divorced from the context in which the statement was made.”

The dictionary definition on which the Claimant seeks to rely would have been inadmissible at the meaning trial. It is even more objectionable to seek to rely after the conclusion of the trial on a dictionary definition to superimpose a gloss on the determined meaning.

The Claimant’s contentions as to the meanings of the court-determined meanings are misconceived. But it would in any event be unfair to allow the Claimant to advance arguments in closing submissions, or after the trial has concluded, which – if acceded to – would have the effect of shifting the target that the Defendant must hit to defend the claim.

The importance of meaning in a libel case can hardly be overstated. For present purposes, the key point is that it is the imputation conveyed by the statement complained of that a defendant must prove is substantially true in order to establish a defence under s.2 of the 2013 Act. Early resolution of any dispute as to meaning fixes, in advance of the trial, the single meaning which is of cardinal importance in evaluating the truth defence. If any argument about the meaning of the meanings was to be raised, it should have been raised long before trial. That is not to encourage arguments about “the meaning of meaning” which, as Warby LJ observed in Riley v Murray [2022] EWCA Civ 1146, [2023] EMLR 3, are “rarely productive”, given that “Judges are generally very well aware of the need for clarity and precision in their determinations of meaning” ([30]). That is true here: the meanings determined by Johnson J were clear and precise. But in any event, the Guardian prepared and presented its case based on the meanings determined by the court. It is far too late, in closing submissions or after the trial has concluded, to raise arguments as to the interpretation of those meanings.

The Data Protection Claim

On the first day of the trial the Claimant conceded that the meanings of the data in the Articles should be the same for his data protection claim as those found by Johnson J in respect of the libel claim, rather than as pleaded in the Amended Particulars of Claim: see my order of 5 March 2025.

The Guardian submitted in opening that the data protection claim was duplicative of the libel claim, and added nothing. Following the conclusion of the evidence, on Friday 4 April 2025, leading Counsel for Mr Clarke, Mr Williams, agreed that the causes of action would stand or fall together. In response to Mr Williams’ suggestion that his closing submissions would be unlikely to touch much on the data protection claim, I indicated that the Claimant’s closing submissions would need to address the data protection claim, and respond to the Guardian’s detailed opening submissions on that cause of action, if the Claimant wished to maintain his data protection claim. I asked the Claimant to confirm his position regarding the data protection claim by the following Monday, so that both parties would know whether it would be necessary to address it in their closing submissions.

On 7 April 2025, the Claimant confirmed that as the data protection claim “does not add anything of substance” he did not seek a ruling on it. Accordingly, it has been withdrawn and so this judgment addresses the libel claim only.

Anonymity, Reporting Restrictions and Confidential Journalistic Sources

There is an anonymity and reporting restriction order (‘the RRO’) in place in respect of:

For source protection reasons, the Guardian’s journalists have referred to a number of confidential journalistic sources (‘CJSs’) by cyphers (e.g. CJS5) or, in instances where the individual was given a pseudonym in one or more of the Articles, by that pseudonym denoted with an asterisk (e.g. Leila*). In this judgment, the names of those whose identities are subject to the RRO are shown in single quotation marks, whereas pseudonyms given to sources by the Guardian are denoted by an asterisk. In some instances, on the basis of the evidence, I have worked out the identity of the CJS – albeit that is neither confirmed nor denied by the Guardian – and so, where relevant, I may make reference to the alternative pseudonym (e.g. ‘Maya’/Kim*) or cypher (e.g. Davie Fairbanks/CJS3).

To protect the anonymity of those individuals who are subject to the RRO, I have identified five productions only as ‘Production A’ to ‘Production E’. In a confidential schedule to this judgment I have given the name of each production.

The History of the Proceedings