England and Wales High Court (Queen's Bench Division) Decisions

IN THE HIGH COURT OF JUSTICE
QUEEN'S BENCH DIVISION
MEDIA AND COMMUNICATIONS LIST

		Royal Courts of Justice Strand, London, WC2A 2LL
		17 May 2022

This judgment was handed down by the Judge remotely by circulation to the parties' representatives by email and release to The National Archives. The date and time for hand down is deemed to be 10:30am on Tuesday 17 May 2022.

His Honour Judge Lewis:

1. In these proceedings the claimant seeks damages and an injunction for libel in respect of a tweet posted by the defendant on 17 March 2019 ("the Tweet").
2. The defendant applies on notice for summary judgment pursuant to CPR rule 24.2. He says the claimant has no real prospect of succeeding in proving that the publication of the tweet caused him 'serious harm', and there is no other reason to allow the case to proceed. The claimant opposes the application.

Background

3. The claim arises in relation to a controversy about who was the developer of Bitcoin, which is a cryptocurrency. In October 2008, an academic paper was published under the name Satoshi Nakamoto which described how the electronic cash system operated. Since then, Satoshi Nakamoto has come to be regarded as the pseudonym for the person or persons who developed Bitcoin.
4. The claimant is a computer scientist with an interest in cryptocurrencies. He claims that he is Satoshi Nakamoto, the author of the 2008 academic paper and the developer of Bitcoin ("the Bitcoin Claim").
5. The defendant is a citizen of Norway, resident in Oslo. He has tweeted on various technology issues, including cryptocurrencies, and has an interest in Bitcoin and its development. He considers the Bitcoin Claim to be false.
6. The claimant says that he has never wanted people to know that he is Satoshi Nakamoto. He says he was shocked in 2015 when he was approached by journalists looking into the story. In December 2015, some articles were published making the Bitcoin Claim, but a few days later the same journals published articles casting doubt on whether the claim was, in fact, true.
7. This sparked a debate on the subject within the Bitcoin community. As part of this, the claimant provided private demonstrations of Bitcoin 'key signings' to four people - a leading figure in the Bitcoin community, a board member of the Bitcoin Foundation and journalists from the BBC and the Economist. The claimant is adamant that he made clear to these four observers that his use of the private keys would not in itself provide conclusive verification of the Bitcoin Claim. It would, however, form part of the matrix of evidence needed to establish the identity of Satoshi Nakamoto.
8. The claimant provided separate demonstrations to each of the four observers between 23 March and 26 April 2016. In each instance, he says the observer applied the 'public key' associated with the 'private key' that the claimant was using, and then verified that the claimant had signed the messages with the correct 'private key'. As a result, the claimant says that the two observers from the Bitcoin sector (not the journalists) confirmed publicly that the claimant had demonstrated that he was likely to be Satoshi Nakamoto.
9. On 26 April 2016, the claimant also met a journalist from GQ magazine, with a view to undertaking a further demonstration. There is a dispute about what happened at the meeting. The claimant says that the demonstration did not go ahead. The defendant says that there was a row after the journalist attended with a cryptography expert who did not find the claimant's attempt at verification to be convincing.
10. On 2 May 2016, the results of the private demonstrations were published. The BBC included the Bitcoin Claim on the Today programme, and on its website, and confirmed that evidence had been produced to back this up. The Economist published a report which is said by the defendant to be in more sceptical terms, asking for the claimant to produce better evidence.
11. On the same day, 2 May 2016, a lengthy post was published on the claimant's blog. The claimant says that whilst this was based on a document he had written previously, it was edited and published on the site by a third party without his knowledge or consent. The purpose of the article appears to have been to explain the process of verifying a set of cryptographic keys.
12. On 3 May 2016 a further post was published on the claimant's blog entitled "Extraordinary claims require extraordinary proof". The article gives the impression that it has been written by the claimant – again, he denies being the author and says

that at the time his website was controlled and managed by a third party. He also says that by this time he was extremely upset by the media furore surrounding the Bitcoin Claim, was in a state of mental collapse and was not fit to take any decisions about what should be published.

13. The author of the piece said that he was able to prove access to the early Bitcoin keys. He said that this in itself would not be sufficient to prove the Bitcoin Claim, and so over the coming days he would post a series of pieces that would lay the foundation for the claim, and then post independently-verifiable documents and evidence addressing some of the false allegations. He also confirmed that he would also be transferring Bitcoin from an early book.
14. The next day, the defendant says that the BBC arranged for one of its journalists and two others to send small amounts of Bitcoin to the public address used in the first ever Bitcoin transaction. The defendant says that the plan was that the claimant would then send the Bitcoin back from that address, using the Satoshi Nakamoto private key, but he did not do so. The claimant says that he was not involved in making any arrangements for this demonstration. By this stage he says his mental state was such that he was barely functioning.
15. On 5 May 2016, another post was published on the claimant's blog. Again, he says that this was nothing to do with him, and the message was posted by a third party who was running his site. The author of the blog posting said the following:

"I believed that I could do this. I believed that I could put the years of anonymity and hiding behind me. But, as the events of this week unfolded and I prepared to publish the proof of access to the earliest keys, I broke. I do not have the courage. I cannot.

When the rumors began, my qualifications and character were attacked. When those allegations were proven false, new allegations have already begun. I now know that I am not strong enough for this.

I know that this weakness will cause great damage to those that have supported me, and particularly to Jon Matonis and Gavin Andresen. I can only hope that their honour and credibility is not irreparably tainted by my actions. They were not deceived, but I know that the world will never believe that now. I can only say I'm sorry.

And goodbye."

16. There is a significant level of disagreement between the parties about the events that have just been described. In the Defence, twenty pages are taken up with the particulars relied upon in support of the plea of truth. The claimant does not accept that he promised to prove the Bitcoin Claim publicly using the private keys. He says that some of the criticisms made of him arise out of a misunderstanding of the workings of Bitcoin or are made by those with a vendetta against him. As noted above, he also disputes being the author of much of the material posted on his blog.

The Tweet

17. The defendant has tweeted about the claimant for some years, focussed mainly on the fact that he does not believe the Bitcoin Claim and thinks the claimant is a fraud.
18. On 17 March 2019, the defendant decided to launch a campaign against the claimant, tweeting as follows:

"As a tribute to Craig Wright being a fraud, I'm going to make next week "Craig Wright is a fraud week", and tag all my tweets with #CraigWrightIsAFraud Feel free to join the celebration [praying emoji]"

19. Later that day – possibly around lunchtime in the UK - the defendant posted the Tweet:

"The forensics to CSW's first attempt to fraudulently 'prove' he is Satoshi. Enabled by @gavinandresen. Never forget. #CraigWrightIsAFraud."

20. This is the only message sued upon within these proceedings. The claimant puts forward an innuendo meaning that the Tweet meant that he "had fraudulently claimed to be Satoshi Nakamoto, that is to say the person, or one of the group of people who developed the cryptocurrency Bitcoin".
21. The defendant's campaign appears to have taken off quite quickly, with the defendant tweeting later the same day to acknowledge the response on Twitter. Three of those tweets were in the following terms:

"Eric says giving attention to frauds is wrong. Which is often correct. But sometimes that can also end up enabling their scams. Making people aware that this man is a fraud is important, many new people coming in unaware. Also, I think it triggers him, which is a bonus".

"#CraigWrightIsAFraud The chain goes strong".

"The fact that Twitter agrees #CraigWrightIsAFraud must surely be causing a serious meltdown as we speak. Long on popcorn for the next couple of days".

22. The next day, 18 March 2019, the tweets posted by the defendant included:

"Happy #CraigWrightIsAFraud week everyone!"

"#CraigWrightIsAFraud week gets off to a flying start with the fraud himself disappearing from twitter on day 1. Double taco rations tonight!"

23. A letter of claim was sent to the defendant via Twitter on 29 March 2019 in respect of nine tweets. The claimant's solicitors sought an undertaking not to repeat the allegations and to delete the offending tweets, a statement in open court and the publication on Twitter of an apology as follows: "APOLOGY TO DR CRAIG WRIGHT. I was wrong to allege Craig Wright fraudulently claimed to be Satoshi. I accept he is Satoshi. I am sorry Dr Wright. I will not repeat this libel". The letter said he would forgo costs and damages if the matter was resolved promptly.
24. The same day, the claimant says that this campaign was picked by a Mr McCormack, who is a UK-based podcaster and blogger specialising in content about Bitcoin and cryptocurrencies. He tweeted at 8.17pm that the claimant had started filing lawsuits against "those falsely denying he is Satoshi".
25. The defendant says that on 9 April 2019 he deleted all nine of the tweets complained of (including the Tweet). He then deactivated his Twitter account, temporarily.
26. On 19 May 2019, the defendant brought proceedings in the Oslo District Court seeking a declaratory judgment that the offending tweets were lawful. The claimant has brought a counterclaim in those proceedings for libel. I am told that the final hearing in those proceedings has been adjourned until later this year.
27. On 26 June 2019, the claimant issued these libel proceedings. They were served in early August. The claimant has confined his claim to the Tweet and now only seeks to recover damages for the harm caused by publication within England and Wales.
28. The defendant sought to challenge this court's jurisdiction, on the basis that he had already started proceedings in Norway. He was successful at first instance in getting the UK claim dismissed, but this was overturned on appeal, see Wright v Granath [2020] EWHC 51 (QB) and Wright v Granath [2021] EWCA Civ 28.
29. The defendant has now served a 35-page defence: (i) he admits that the Tweet identified the claimant; (ii) he denies that the Tweet was defamatory of the claimant on the basis that he says it did not cause, nor was it likely to cause, serious harm to the claimant's reputation; (iii) he asserts that the claim is an abuse of process; (iv) he pursues a defence of truth, seeking to meet the claimant's pleaded meaning; and (v) he pursues a defence that publication was on a matter of public interest. In response, the claimant has served a 39-page reply.
30. The claimant has also brought proceedings against others.
31. On 17 April 2019, the claimant sued Mr McCormack, the person referred to above. Initially, he sued over the publication of ten tweets posted on or after 29 March 2019, but the scope of the claim has since widened. The claimant says that the defamatory meaning of the publications is that he "had fraudulently claimed to be Satoshi Nakamoto, that is to say the person, or one of the group of people who developed Bitcoin". Mr McCormack responded with pleaded defences of truth, public interest and abuse of process, but these have since been withdrawn. On 8 October 2021, Julian Knowles J handed down a very detailed judgment dealing with various case management and pleading issues: Wright v McCormack [2021] EWHC 2671 (QB). The defendant's application for permission to appeal was refused on paper by Warby LJ on 27 January 2022. The claim is listed for trial this year, the focus of which will be on whether each of the publications complained of caused the claimant serious harm.
32. On 2 May 2019, the claimant sued a Mr Ver in respect of a YouTube video and two tweets. Again, the claimant said that the defamatory meaning of these publications was that he "had fraudulently claimed to be Satoshi Nakamoto, that is to say the person, or one of the group of people who developed Bitcoin". On 31 July 2019, Nicklin J struck out the claim, finding that the court had no jurisdiction to hear and determine the action, England and Wales not being "clearly the most appropriate place" in which to bring the claim pursuant to s.9 Defamation Act 2013 ("the Act"). This decision was upheld by the Court of Appeal: see Wright v Ver [2019] EWHC 2094 (QB) and Wright v Ver [2020] EWCA Civ 672.

Summary judgment

33. CPR rule 24.2 provides that the court may give summary judgment against a claimant or defendant on the whole of a claim or on a particular issue if:

"(a) it considers that—

(i) that claimant has no real prospect of succeeding on the claim or issue; or

(ii) that defendant has no real prospect of successfully defending the claim or issue; and

(b) there is no other compelling reason why the case or issue should be disposed of at a trial".

34. The approach to be taken when considering a defendant's application for summary judgment applications was summarised by Lewison J in Easyair Limited (Trading As Openair) v Opal Telecom Limited [2009] EWHC 339 (Ch) at [15] (citations removed):

"i) The court must consider whether the claimant has a "realistic" as opposed to a "fanciful" prospect of success.

ii) A "realistic" claim is one that carries some degree of conviction. This means a claim that is more than merely arguable.

iii) In reaching its conclusion the court must not conduct a "mini-trial".

iv) This does not mean that the court must take at face value and without analysis everything that a claimant says in his statements before the court. In some cases it may be clear that there is no real substance in factual assertions made, particularly if contradicted by contemporaneous documents.

v) However, in reaching its conclusion the court must take into account not only the evidence actually placed before it on the application for summary judgment, but also the evidence that can reasonably be expected to be available at trial.

vi) Although a case may turn out at trial not to be really complicated, it does not follow that it should be decided without the fuller investigation into the facts at trial than is possible or permissible on summary judgment. Thus the court should hesitate about making a final decision without a trial, even where there is no obvious conflict of fact at the time of the application, where reasonable grounds exist for believing that a fuller investigation into the facts of the case would add to or alter the evidence available to a trial judge and so affect the outcome of the case.

vii) On the other hand it is not uncommon for an application under Part 24 to give rise to a short point of law or construction and, if the court is satisfied that it has before it all the evidence necessary for the proper determination of the question and that the parties have had an adequate opportunity to address it in argument, it should grasp the nettle and decide it. The reason is quite simple: if the respondent's case is bad in law, he will in truth have no real prospect of succeeding on his claim or successfully defending the claim against him, as the case may be. Similarly, if the applicant's case is bad in law, the sooner that is determined, the better. If it is possible to show by evidence that although material in the form of documents or oral evidence that would put the documents in another light is not currently before the court, such material is likely to exist and can be expected to be available at trial, it would be wrong to give summary judgment because there would be a real, as opposed to a fanciful, prospect of success. However, it is not enough simply to argue that the case should be allowed to go to trial because something may turn up which would have a bearing on the question of construction."

Serious Harm

35. Section 1(1) of the Defamation Act 2013 provides that: "A statement is not defamatory unless its publication has caused or is likely to cause serious harm to the reputation of the claimant"
36. The current approach to be taken when considering issues of serious harm was summarised by Nicklin J in Turley v Unite the Union and another [2019] EWHC 3547(QB)

"107. This provision was considered by the Supreme Court in Lachaux-v-Independent Print Ltd [2019] 3 WLR 18. Although, the Supreme Court agreed with the ultimate decision of the Court of Appeal dismissing the defendant's appeal ([2018] QB 594), it disagreed with its reasoning and held that Warby J's analysis of the law, at first instance ([2016] QB 402), was "coherent and correct, for substantially the reasons he gave" [20] per Lord Sumption. The Supreme Court held:

i) s.1 raised the threshold of seriousness above the tendency of defamatory words to cause damage to reputation; the application of the test of serious harm must be determined "by reference to actual facts about its impact and not just to the meaning of the words" [12]-[13].

ii) Reference to the situation where the statement "has caused" serious harm is to the consequences of publication, and not the publication itself [14]: "It points to some historic harm, which is shown to have actually occurred. This is a proposition of fact which can be established only by reference to the impact which the statement is shown actually to have had. It depends on a combination of the inherent tendency of the words and their actual impact on those to whom they were communicated."

iii) Reference to the situation where the statement "is likely to cause" serious harm was not the synonym of "liable to cause" in the sense of the inherent tendency of defamatory words to cause damage to reputation: [14].

iv) The conditions under s.1 must be established as facts [14] and "necessarily calls for an investigation of the actual impact of the statement": [15]; a claimant must demonstrate as a fact that the harm caused by the publication complained of was serious [21].

v) If serious harm could be demonstrated simply by the inherent tendency of statements to damage reputation, little substantive change would have been effected by the Act [16]: "The main reason why harm which was less than 'serious' had given rise to liability before the Act was that damage to reputation was presumed from the words alone and might therefore be very different from any damage which could be established in fact. If, as Ms Page submits, the presumption still works in that way, then this anomaly has been carried through into the Act. Suppose that the words amount to a grave allegation against the claimant, but they are published to a small number of people, or to people none of whom believe it, or possibly to people among whom the claimant had no reputation to be harmed. The law's traditional answer is that these matters may mitigate damages but do not affect the defamatory character of the words. Yet it is plain that section 1 was intended to make them part of the test of the defamatory character of the statement."

vi) A claimant may produce evidence from publishees of the statement complained of about its impact on them, but his/her case does not necessarily fail for want of such evidence; inferences of fact as to the seriousness of harm done to reputation may be drawn from the evidence as a whole [21].

vii) In Mr Lachaux's case, the finding that serious harm had been proved was based on a combination of (a) the meaning of the words; (b) the situation of the claimant; (c) the circumstances of publication; and (d) the inherent probabilities.

viii) A judge's task is to evaluate the material before him/her and arrive at a conclusion, recognising that this is an issue on which precision will rarely be possible [21].

ix) The judge can consider the impact of the publication upon people who do not presently know the claimant but might get to know him/her in the future [25].

108. At first instance in Lachaux, Warby J expressed his conclusion on s.1 as follows:

[65] In summary, my conclusion is that by section 1(1) of the 2013 Act Parliament intended to and did provide that a statement is not defamatory of a person unless it has caused or will probably cause serious harm to that person's reputation, these being matters that must be proved by the claimant on the balance of probabilities. The court is not confined, when deciding this question, to considering only the defamatory meaning of the words and the harmful tendency of that meaning. It may have regard to all the relevant circumstances, including evidence of what has actually happened after publication. Serious harm may be proved by inference, but the evidence may or may not justify such an inference.

109. Finally, and consistently with Lord Sumption's analysis in Lachaux, there are three further relevant principles:

i) In an appropriate case, a Claimant can also rely upon the likely 'percolation' or 'grapevine effect' of defamatory publications, which has been "immeasurably enhanced" by social media and modern methods of electronic communication: Cairns -v-Modi [2013] 1 WLR 1015 [26] per Lord Judge LCJ. In the memorable words of Bingham LJ in Slipper -v-British Broadcasting Corporation[1991] 1 QB 283, 300: "... the law would part company with the realities of life if it held that the damage caused by publication of a libel began and ended with publication to the original publishee. Defamatory statements are objectionable not least because of their propensity to percolate through underground channels and contaminate hidden springs."

ii) It is well-recognised that a claimant may struggle to identify, or to produce evidence from, all those to whom an article was published and in whose eyes the claimant's reputation was damaged: Doyle -v-Smith [2019] EMLR 15[122(iv)]; Sobrinho-v-Impresa Publishing SA [2016] EMLR 12[48]; Ames -v-Spamhaus [2015] 1 WLR 3409 [55].

iii) Assessment of harm to reputation has never been just a 'numbers game': "one well-directed arrow [may] hit the bull's eye of reputation" and cause more damage than indiscriminate firing: King -v-Grundon[2012] EWHC 2719 [40] per Sharp J. Very serious harm to reputation can be caused by publication to a relatively small number of publishees: Sobrinho[47]; Dhir-v-Sadler [2018] EWHC 2935 (QB) [55(i)]; Monir -v-Wood[2018] EWHC 3525 (QB)[196].

The application

37. The parties agree that people who read the defendant's tweets are likely to be persons with a special interest in, and knowledge of, Bitcoin and cryptocurrency.
38. The claimant's amended particulars of claim put the case on serious harm as follows:

a. The defamatory allegation was very grave.

b. The Tweet was published to at least the defendant's 8,878 followers. It can be inferred that the Tweet was published very extensively within this jurisdiction, including to an audience way beyond the defendant's twitter followers.

i. An earlier tweet sent 90 minutes beforehand had (by 28 March) already been retweeted 62 times and liked 569 times. The number of readers of that tweet is likely to very substantially larger than the number of engagements.

ii. It is inferred that a substantial number, if not all, of the publishees of the earlier tweet would have read the Tweet either because they were followers of the defendant, or they accessed it by clicking on the defendant's username or the hashtag, which would have taken them to the Tweet.

iii. The defendant became extremely well known in the aftermath of publication of the Tweet, particularly within the cryptocurrency communities. It is inferred that this celebrity as a campaigner would have prompted Twitter users who were not followers of the defendant to go to the defendant's feed and read the Tweet.

iv. Mr McCormack (see above) had 57,000 followers on his twitter account. On 29 March, he started a tireless campaign to publicise threats made by the claimant to sue the defendant over his tweets. It is to be inferred that Mr McCormack drove a substantial number of his twitter followers to the defendant's profile, and subsequently to read the Tweet.

39. The defendant seeks summary judgment on the basis that he says the claimant has no real prospect of establishing that the Tweet caused serious harm to his reputation. The defendant's primary argument is that the claimant has not demonstrated that he has suffered serious harm. Mr Tomlinson says that the case on serious harm is wholly inferential, based on the inherent tendency of the words complained of. Issue is taken with the fact that no readers of the Tweet have been identified, nor any re-publications. It is said that the claimant has not advanced a case on the actual impact of the words complained of on those who read it.
40. In respect of the claimant's pleaded case, the defendant also says that the allegation is not "very grave" when seen in context given that the readership of the Tweet was familiar with the claimant's notoriety as a fraud. It is said that at the time of publication, the claimant had a general reputation as a dishonest and fraudulent person both in relation to the Bitcoin Claim but also in a more general sense, as a result of adverse judicial and quasi-judicial findings. It is said that the claimant's own conduct was also a prominent cause of this general bad reputation. Mr Tomlinson also points out that the claimant's pleaded case on identification and meaning relies on innuendo. He says those with the required special knowledge would already know something about the background and the claimant's attempts to prove the Bitcoin Claim, and so will have formed a view of what has been going on.
41. The defendant also takes issue with the scale of the publication, which Mr Tomlinson says was would have been very small. He says that it is highly improbable that the Tweet will have made any impact on the claimant's reputation.
42. The claimant says that this case is about a blogger in the cryptocurrency sphere waging an unpleasant attack against the claimant, launching a "fraud week" and trying to get a Twitter "pile-on" going. The Tweet formed part of this campaign, and what was said in the Tweet went to the core of the claimant's reputation. It is said that by trial, the claimant will be able to show that the Tweet was likely read by significant numbers in this jurisdiction. Mr Wolanski says that it is impermissible for the defendant to say that because others have said similar things in the past about the events of 2016, the defendant can now do so with impunity. He points out that allegations made in other articles about the Bitcoin Claim and the events in 2016 are not "evidence" that the claimant had a general bad reputation, particularly given that there are significant disputes of fact to be determined. The claimant says that these are not issues to be determined summarily, the claimant has a real prospect of success and the matter should be allowed to proceed to trial.

Scale of publication

43. A significant part of the defendant's evidence in support of this application focusses on what he says would have been the extremely limited publication of the Tweet.
44. When the Tweet was deleted, so was the data held about it by Twitter, or at least such data as Twitter makes available to its users.
45. With traditional media – such as magazines, newspapers, television and radio – it has never been possible to identify the precise number of readers, viewers or listeners for any given article or broadcast item. Courts have taken a broad view based on available evidence, including for example programme ratings and audited publication figures.
46. With some electronic publications, more precise data is available – for example the number of hits on a specific page of a website, the number of times a programme has been viewed from a TV streaming service or the number of times that an online video has been played.
47. With Twitter and some other social media platforms, the position is less straightforward. In Munroe v Hopkins [2017] 4 WLR 68, Warby J prepared an explanatory note on how Twitter works, and the types of data available to assist a court in determining the likely readership of a tweet. The platform works in a similar way today, although there have been some changes. Although not material to the application before me today, I note in respect of paragraph 6 of Warby J's note that Twitter users now have the option of having additional material included in their timeline from accounts they do not follow – material selected by Twitter based on the accounts and topics that the user follows.
48. The number of "impressions" that a tweet has received is particularly important, defined by Twitter as being "the number of times a user is served a Tweet in timeline or search results".
49. The parties have brought my attention some reported decisions in which the court has needed to decide at trial on the likely readership of a tweet:

a. Cairns v Modi [2012] EWHC 756 (QB) (Bean J). This decision pre-dates the introduction of the 'serious harm' test, with the court considering the extent of publication in the context of damages. Both parties had expert evidence on the issue, with the parties agreeing that the court should 'split the difference' between the experts and proceed on a readership of 65 for the more serious of two publications sued upon, and 1,000 for the second.

b. Munroe v Hopkins [2017] 4 WLR 68 (Warby J). The defendant had around 570,000 followers. Data was not available for one of the two tweets complained of, because it had been deleted. Instead, the court looked at other data available from Twitter Analytics, including for other tweets sent the same day (looking at figures for impressions, engagements and re-tweets), and the overall number of impressions for the defendant's account over the month of publication, and the month preceding it (5.7m per month). It was noted that precision would be impossible, but also not necessary providing that the court could make a sound assessment of the overall scale of publication. Warby J was satisfied that the claimant's estimate of around 20,000 for the publication of the first tweet was entirely reasonable, and more likely to be an under-estimate than an over-estimate.

c. Riley v Murray [2022] EMLR 8 (Nicklin J). This was a claim over a single tweet. The defendant had 7,252 followers. She too deleted the tweet, and suspended her twitter account, and so Twitter Analytics were unavailable. In the absence of this data, the court indicated that it would seek to draw sensible inferences from any evidence available that it accepts as reliable. This included posts on social media responding to the tweets sued upon, screen shots of the tweet which showed that it had been re-tweeted 1,585 times, liked by 4,932 people and provoked 736 responses. Taking this all into account, and the likely grapevine/percolation effect, the court estimated that the tweet was published to 10,000-15,000 people.

50. The defendant's case and analysis in respect of the number of people who would have read the Tweet is as follows:

a. The defendant's twitter account had 8,878 followers in April 2019.

b. A "tweepsmap" obtained by the defendant on 17 June 2019 shows that 7.5% of the defendant's followers are from the UK, which equates to 675 people. I note that in addition to this there will of course be some people who have not provided geolocation data, and 9.7% of the followers shown on the map do not appear to have been linked to a specific country.

c. Approximately 89% of the UK's population is within the jurisdiction of the court (England and Wales), which leaves 601 followers.

d. It is unlikely that all or even most of these 601 followers will have read the Tweet.

e. The defendant has undertaken a search on Google and could not find evidence that the Tweet had been re-published, although this is perhaps unsurprising given that the Tweet was deleted some time ago.

f. Analytics data for September 2018 – March 2019 for undeleted tweets shows that an average tweet over that period had 130 impressions in England and Wales.

51. At first, the defendant's solicitor confirmed in a witness statement dated 22.09.2021 that there is no additional data or disclosure that would allow the claimant to provide more evidence on publication, and that the court now has all available information. This proved to be incorrect. In response to requests from the claimant's solicitors, the defendant's solicitor provided a further statement confirming that, in fact, on 6 April 2019, Twitter had provided the defendant with an archive of his twitter data. It is unclear why this had not been confirmed sooner or referred to in the defendant's evidence in support of this application.
52. The defendant has still not provided the claimant with access to this data. A solicitor for the defendant has instead reviewed the files and produced print outs of the parts that she considers relevant, but the claimant has not been able to check this.
53. The defendant has also not provided his data for the period from 1 April 2019. This is needed for two reasons. Firstly, the Tweet remained online during part of this period - the claimant says it was only on 29 March that Mr McCormack joined in, which was twelve or so days before the Tweet was deleted. Secondly, the information is needed to properly evaluate the data that we have already and to understand the level of activity on this Twitter account over the relevant period.
54. The data that has been provided does, however, paint a picture that is somewhat different from the impression given by the defendant. As far as we know, the data for March 2019 does not include statistics in respect of the nine tweets that the defendant deleted, including the Tweet. From this information provided it appears that:

a. The defendant was a popular and prolific tweeter.

b. There was a sharp rise in the level of interest in the defendant's Twitter account in the months leading up to the publication of the Tweet. We see the number of impressions per month on his account rising from 1.09m in September 2018 to 4.46m in February 2019. In March 2019, the data provided shows 3.99m impressions, but this excludes the nine deleted tweets.

c. During this time, the defendant received a lot of new followers, increasing by 777 in January, 3,829 in February and 1,103 in March 2019.

d. The data shows that the top follower of the defendant for each month was someone clearly linked to Bitcoin or the cryptocurrency sector. These "top followers" had significant audiences: the top followers for December 2018, January 2019, February 2019 and March 2019 had 168k, 1.32m, 5.9m and 2.34m followers respectively.

e. The number of daily impressions varies. In February 2019, the defendant's account was receiving 159,400 average daily impressions, but we can see this varied significantly – on four days the daily figure was over 200,000, and on one it was over 400,000. We also know that the top twenty tweets that month each had between 15,649 and 78,607 impressions.

f. The data from March is less straightforward, as it most likely does not include the data from the nine deleted tweets. The account was receiving 128,600 average daily impressions. Again, this varied significantly with four days having over 200,000 and on one day it was around 400,000. The top twenty recorded tweets (ie excluding any that had been deleted, each had between 20,433 and 110,412 impressions.

g. There is also evidence on "engagements". We know that only a small proportion of readers will engage with a tweet, and so this data is probably less useful, but still of relevance.

55. On the day of the hearing, the claimant's legal team produced a screenshot that had been taken of the Tweet on 28 March 2019, eleven days after first publication, and the day before Mr McCormack's tweet. It is unclear why this was not disclosed sooner. The screenshot shows that on this date the Tweet had 15 worldwide "engagements" (see below), which the defendant says most likely equates to just one within this jurisdiction.
56. The claimant's case is that the day after this screenshot was taken, Mr McCormack started his "pile on", and it is to be inferred that more people will have read the Tweet. Even focussing on the screenshot data, it is said that the median twitter engagement rate is 0.466%, so if there were 15 engagements, this would equate to more than 3,000 impressions. If one says that 7% will have been in this jurisdiction, which Mr Wolanski does not accept, this will still mean 230 impressions. It is said that this alone puts the case way outside any arguable case that it is bound to fail.
57. Mr Wolanski takes issue with the data that has been provided by the defendant. He notes that the claimant is based in England, which he says may mean that a higher proportion of the impressions may have been in the UK. He also says that the court cannot simply look at the numbers of followers since there would have been other readers, including those who clicked on the hashtag which was used to promote the defendant's campaign.
58. Of course, all these figures are for worldwide publication. They do however provide a very different picture to that set out by the defendant. One striking example of the data being presented on a skewed basis is the claim that the defendant received an average of 130 impressions per tweet. This average was for some reason taken across a seven-month period, even though it is only the last month of this period that is relevant to this case. By including data for the earlier months, in which the number of impressions were relatively low, the average reduced from 241 impressions per tweet (assuming 7% of readers were in the UK, which is not accepted by the claimant), to the 130 impressions cited. The data also highlights the problems using averages, which potentially distorts the results because it ignores the impact of the inevitable variations in the data.
59. The court will not usually require expert evidence when ascertaining the extent of a publication on Twitter, nor expect lawyers to know about complex statistical modelling. As we have seen in Riley and Munroe, the court will seek to reach the best decision that it can on the evidence that is available. For this to be able to happen, it is important that parties present data fairly, and in a way that will assist the court. How this is done will vary from case to case and will depend on what data is available. In most cases, it is likely to be helpful for parties to include a brief explanation of the reasons why the data has been presented in a certain way, if not obvious, to assist the court's understanding.

Alleged "bad reputation" and misconduct

60. In Barron v Vines [2016] EWHC 1226 (QB) at [21], Warby J sought to summarise the approach taken by the court when assessing damages, relevant parts of which include the following:

"(2) The existence and scale of any harm to reputation may be established by evidence or inferred. Often, the process is one of inference, but evidence that tends to show that as a matter of fact a person was shunned, avoided, or taunted will be relevant. So may evidence that a person was treated as well or better by others after the libel than before it.

(3) The impact of a libel on a person's reputation can be affected by:

a) Their role in society. The libel of Esther Rantzen was more damaging because she was a prominent child protection campaigner.

b) The extent to which the publisher(s) of the defamatory imputation are authoritative and credible. The person making the allegations may be someone apparently well-placed to know the facts, or they may appear to be an unreliable source.

c) The identities of the publishees. Publication of a libel to family, friends or work colleagues may be more harmful and hurtful than if it is circulated amongst strangers. On the other hand, those close to a claimant may have knowledge or viewpoints that make them less likely to believe what is alleged.

d) The propensity of defamatory statements to percolate through underground channels and contaminate hidden springs, a problem made worse by the internet and social networking sites, particularly for claimants in the public eye: C v MGN Ltd (reported with Cairns v Modi at [2013] 1 WLR 1051) [27]….

(4) ….

(5) A person who has been libelled is compensated only for injury to the reputation they actually had at the time of publication. If it is shown that the person already had a bad reputation in the relevant sector of their life, that will reduce the harm, and therefore moderate any damages. But it is not permissible to seek, in mitigation of damages, to prove specific acts of misconduct by the claimant, or rumours or reports to the effect that he has done the things alleged in the libel complained of: Scott v Sampson (1882) QBD 491, on which I will expand a little. Attempts to achieve this may aggravate damages, in line with factor (d) in Sir Thomas Bingham's list.

(6) Factors other than bad reputation that may moderate or mitigate damages, on some of which I will also elaborate below, include the following:

a) "Directly relevant background context" within the meaning of Burstein v Times Newspapers Ltd [2001] 1 WLR 579 and subsequent authorities. This may qualify the rules at (5) above.

b) Publications by others to the same effect as the libel complained of if (but only if) the claimants have sued over these in another defamation claim, or if it is necessary to consider them in order to isolate the damage caused by the publication complained of. (…)"

61. Burstein particulars are directly relevant background context which go to the claimant's reputation in the relevant sector of his or her life. If evidence is to qualify under the principle spelt out in Burstein's case, it has to be evidence which is so clearly relevant to the subject matter of the libel or to the claimant's reputation or sensitivity in that part of his life that there would be a real risk of the jury assessing damages on a false basis if they were kept in ignorance of the facts to which the evidence relates, Turner v News Group Newspapers [2006] EWCA Civ 540 per Keene LJ at [56].
62. There is also what is known as the rule in Dingle (from Dingle v Associated Newspapers Limited [1964] AC 371). This is a rule of evidence or case management. Its ratio is that "whilst the defendant to a claim in defamation may prove, in mitigation, that the claimant had a pre-existing general bad reputation, this may not be done by relying on other publications to the same or similar effect", per Warby J in Sicri v Associated Newspapers Limited [2020] EWHC 3541 (QB) at [178(6)].
63. The rule in Dingle is not confined to what material is admissible in mitigation of damages, but also extends to the admissibility of other articles on the question of whether the claimant can establish serious harm to reputation: Lachaux v Independent Print Limited [2016] QB 402. The parties agree with the summary of relevant principles as set out by Julian Knowles J in Wright v McCormack [2021] EWHC 2671 at [167]:

"…it is worth reiterating for the purposes of the issues arising in this case that Dingle and Lachaux demonstrate that: (a) evidence of the claimant's general bad reputation is admissible in relation to mitigation of damages, but such must be proved in a specific way by calling persons who know him and who have had dealings with him and who can speak to his bad reputation. Subject to exceptions, evidence of specific conduct is not admissible. Exceptions to that general prohibition include a previous conviction or possibly (per Warby J in Lachaux at [74]), a previous notorious incident, and 'judicial strictures in previous civil litigation' (Turner v News Group Newspapers Ltd [2006] 1 WLR 3469, [48]). Other than that, such evidence is generally only relevant to a plea of justification; (b) it is not permissible for a defendant to prove, in mitigation of damages, that, previously to his publication, there were reports and rumours in circulation to the same effect as the libel; (c) nor can a defendant rely on such publications to show that his publications could not have caused the claimant serious harm for the purposes of s 1 of the DA 2013 because other, similar, publications had already harmed him."

64. The defendant's arguments in respect of the claimant's alleged bad behaviour and misconduct have three main strands.

a. The defendant seeks to rely on "evidence of previous convictions and adverse judicial findings", which he says is admissible under the rules in Turner and also under Burstein as directly relevant background context.

b. The defendant says that it was the claimant's own wrongful actions that provided the key context for the publication of the Tweet, which he says is admissible as evidence of directly relevant background context. Mr Tomlinson clarified that this relates to the events of 2016. Examples relied upon include the claimant's sessions with journalists and his blog posts, in which it is said he promised and then failed to provide satisfactory proof of the Bitcoin Claim. It is said that the claimant's actions had, at the time of the Tweet, given rise to a widespread belief that the claimant was a fraud and that the Bitcoin Claim was false; and

c. The defendant seeks to rely on what he says were an exceptionally high number of tweets and other public statements, indicating that the Bitcoin Claim was dishonest and/or fraudulent. Mr Tomlinson says this evidence is admissible on the question of causation. The evidence served in support of the defendant's application also seeks to rely on these other tweets and publications as "directly relevant background context".

Strand one - "evidence of previous convictions and adverse judicial findings"

65. The defendant relies on three events.

a. Firstly, the fact that the claimant was committed for contempt of court in New South Wales in 2004 for breach of an undertaking, receiving a suspended sentence with 250 hours of community service. The claimant says that this is irrelevant to the present case, having occurred over fourteen years before the Tweet was published, in another jurisdiction. The claimant also says that this is not something that falls within the relevant sphere of the claimant's reputation, being unrelated to the allegation of fraud in this case.

b. Secondly, the fact that in March 2016, the Australian tax office published a 54-page report following a tax audit of a company connected to the claimant. The report made adverse findings about the claimant's honesty, considering that he had provided false information in order to deceive the tax authorities. The claimant says that it is improbable that anyone would have known about this when reading the Tweet: the report was not available on the internet until a year after the Tweet was published, and then only as part of a dense 54-page report.

c. Two decisions of the court in Florida, but these post-date the publication of the Tweet by five or so months and it is accepted by Mr Tomlinson that these are not admissible.

Strand two – the defendant's own wrongful actions

66. Mr Tomlinson made very clear in submissions that he is not relying here on other publications. He is relying on the conduct of the claimant in 2016, in particular he says the claimant saying that he would prove the Bitcoin Claim and failing to do so, and the material published on his blog. He says that the suggestion that readers would not know about the 'events' of 2016 is wholly fanciful. It is said that as a result of the claimant's own actions, there was widespread public reaction and public claims he was a fraud. Mr Tomlinson says that the court would be "putting on blinkers" if it does not take into account this material: it would have been well known to people who understand the innuendo and therefore impacts on how the court assesses the claimant's reputation.
67. Mr Wolanski says that the defendant's argument appears to be based on a version of events that is not true. The claimant does not accept he made promises that failed. He does not accept that readers in 2016 would understand that he made promises that failed. He says the issue the court should focus on is whether at this very early stage it can determine that in 2016 the claimant in fact failed to do what he promised to do. He says that unless the defendant can persuade the court that the case is so clear cut on that issue, then any alleged "notoriety" becomes irrelevant, because it is not based on an actual incident that took place, but on inaccurate reporting of the incident.

Strand three – other publications

68. In support of the defendant's application, his solicitors prepared statements to which they have exhibited a raft of articles and detailed schedules of 6,778 tweets and retweets filling 228 pages from period up to 14.3.2019. Most are from 2018 and are described by the defendant's solicitor as being "on the theme of the Claimant's frauds".
69. In her statement, the claimant's solicitor is critical of the defendant's schedules, noting the absence of any clear methodology and that the defendant managed to serve two completely different versions of the same spreadsheet. The solicitor also notes that 23% of the tweets identified were re-tweets, 21% were either not about the claimant or the Bitcoin Claim, 10% were supportive of the claimant and 10% were duplicates. She estimates that only 3,916 go to the issues raised, "which if averaged over the period shown would equate to 25 tweets a week". The solicitor also undertook her own investigations for period 13 – 29 March – which showed that in the 2.5 weeks after publication of the Tweet, 1,045 tweets about fraud or a scam were published, which she says equates to 418 per week and a significant increase in the numbers pre-publication.
70. The defendant accepts that the rule in Dingle applies in this case. Mr Tomlinson said that he does not seek to rely on other publications in mitigation of serious harm, and so he is not seeking to go behind the rules in Dingle or Scott v Sampson. He says he is primarily relying on the claimant's own conduct in 2016. He says the other publications are, however, relevant on the issue of causation of harm because the claimant must prove that the Tweet caused serious harm to his reputation, and not the other widespread reporting which harmed him. As noted above, the claimant's evidence in support of this application also says that this material is relevant under Burstein as directly relevant background context around what the claimant did in April 2016.
71. The claimant says that this is a flagrant attempt to circumvent the rule in Dingle. It is said by Mr Wolanski that there are cases where publications by others to the same effect are admissible, but as noted by Warby J in Barron (Ref), this is where it is necessary to isolate the damage caused by the publication complained of. Mr Wolanski explained that the issues around isolation typically arise where you have a case in which it is said that there have been specific consequences of publication, which the court needs to consider, for example

a. In Napag Trading Limited v Gedi Gruppo Editoriale spa [2020] EWHC 3034 (QB) at [51-57], where the court was considering a claim for special damages, where it was said that the events giving rise to any damage pre-dated the publication of the words complained of and so, as a matter of fact, the damage could not have been caused by anything done by the defendants. At [57] Jay J said: "These causation problems may arise… where a claimant seeks to ascribe a specific consequence to a particular publication, or where an examination of the claim for special damage demonstrates that the harm in question could not have been caused by the publication at issue. Thus, if a claimant says that X happened because of publication Y, or if it is clear to the court that the reason X happened was because of publication Y, it is no use the claimant suing publisher Z in respect of that consequence."

b. In Wright v McCormack (supra) – the court considered facts almost identical to those in this case, involving the same claimant. Julian Knowles J said: "Thus, publications to the same effect as that sued on may be admissible, for example, where a claimant sues publication X, claiming that a particular damaging consequence occurred because of something written by publication X. It would in principle be permissible for publication X to plead and seek to prove that it was not its publication which caused the particular adverse event harmful to the claimant, but it was a story in a different publication to the same effect which caused the event. This is sometimes known as the 'rule of isolation'." In refusing permission to appeal, Warby LJ acknowledged that the judge "plainly recognised that Dingle does not always exclude consideration of rival causal factors; where a claimant identifies some specific item of harm as consequent on the alleged libel the defendant may in principle rely on specific alternative causes."

72. These issues were also considered in some detail in Sicri v Associated Newspapers Limited (supra) at [178]. The court identified the different approach to be taken when considering causation issues around specific items of loss, or particular events that are relied on as evidence of damage. An example given is that where a claimant proves he was taunted or abused or shunned or avoided by people who formerly enjoyed his company, the court must review causation to determine whether to compensate the claimant on the basis that such taunts etc were a consequence of the defendant's tortious behaviour.
73. Mr Wolanski says that these issues do not arise in this case, because of the way that the case has been pleaded.

Discussion

74. This is an application for summary judgment, not trial. It is for the defendant to establish that the claim does not have a real prospect of success, and there is no other compelling reason why the case or issue should be disposed of at trial.
75. Whilst meaning has not been determined, I am proceeding on the basis that the claimant is able to establish his meaning at trial. What was said about the claimant went to the core of his professional reputation in this field.
76. When considering the test of "serious harm" the court is not, of course, just looking at the meaning of the words. The court needs to determine the matter by reference to actual facts about their impact. Inferences of fact as to the seriousness of harm done to reputation may of course be drawn from the evidence as a whole – by which I mean the evidence that is reasonably likely to be available at trial.
77. Whilst the claimant has not pleaded examples of actual harm, he has set out a case that draws on the seriousness of the words complained of, together with inferences relating to the level of publication based on evidence that will be available at trial.
78. The evidence does not at this stage appear to support the defendant's arguments in respect of the level of publication of the Tweet.
79. The defendant needs to provide disclosure and inspection of the full data that is already in his possession, and other data within his control that he can request from Twitter, for example the Analytics data for April and possibly May 2019. As a matter of procedural fairness, the claimant needs an opportunity to consider this and evaluate the accuracy of what the defendant says.
80. The parties then need to give careful thought to a proportionate, coherent, sensible and fair way of interpreting this data, and any other information that might assist a court in reaching a reasonable view on the likely extent of publication. As already noted, the analysis provided by the defendant in support of this application is somewhat lacking in that it is obviously incomplete, does not appear to represent the underlying data as fairly as it might, and fails to explain the reasons for interpreting that data in a certain way.
81. Whilst it seems unlikely that the Tweet would have been read by many thousands, this is not simply a numbers game, and it is the "quality of the publishees, not their quantity, that is likely to determine the issue of serious harm in cases involving relatively small-scale publication": Dhir v Saddler [2017] EWHC 3155 (QB). In this case, both parties accept that the likely readership would have been within the cryptocurrency community. Again, these are issues for trial.
82. Turning to the defendant's secondary arguments, Mr Tomlinson was clear that he is not saying that any of these on their own would be a "knock out blow", but when added together he says the court can be satisfied that the claimant does not have a real prospect of success.
83. The defendant seeks to rely on the 2004 contempt finding by the Australian courts and the 2016 conclusions in the audit report prepare by the Australian tax authorities. I am not satisfied that either falls to be considered as evidence of general bad reputation.
84. As we have seen, convictions for serious criminal offences within the relevant sphere of a claimant's reputation can sometimes be admitted as evidence of general bad reputation (Goody v Odhams Press [1967] 1 QB 333). This principle is also relevant when considering serious harm under s.1: Ahmed v Express Newspapers [2017] EWHC 1845 (QB). It is less clear whether court judgments containing findings adverse to a claimant are admissible, see for example the observations of Nicklin J in Alsaifi v Trinity Mirror and another [2017] EWHC 2873 (QB) at [105](iii).
85. I am doubtful that a finding of civil contempt from another jurisdiction would be admissible as evidence of general bad reputation under the principles just outlined, but I have not heard argument on this point. Assuming that evidence of this type can be admitted for this purpose, it does not seem to me that the 2004 finding has any particular relevance to the present case: it was over fourteen years ago, and not something that was in the relevant sphere of the claimant's reputation.
86. The audit report of the tax office is not the equivalent of a conviction for a criminal offence. Nor in my view is it even the equivalent of a judicial stricture in civil proceedings: it appears to be an administrative decision of the tax authorities. Even if it was equivalent, I do not consider it would be admissible as evidence of general bad reputation. It was only published online a year or so after the Tweet, and so would not have been common knowledge, certainly not in this jurisdiction.
87. The defendant also seeks to rely on the audit report under the principles in Burstein, as "directly relevant background context". I think more needs to be known about the status of this report, the basis upon which evidence was gathered and any subsequent follow up, before a decision can be taken on this. It is a lengthy report. Much of it appears wholly irrelevant to this case, although there are aspects that might be in the relevant sphere of the claimant's reputation – although not to the extent that it will have a material impact on this summary judgment application.
88. Turning to the second limb, the defendant seeks to rely on the claimant's conduct in 2016 and events said to be notorious. The main problem with this is that there is not an agreed factual narrative of what happened in 2016, or indeed the extent to which the claimant was involved in the material posted onto his website. There is an extensive plea of justification, and whilst parts of this are admitted, a significant number of central issues remain. The defendant accepts that these matters cannot be dealt with now, will need to be determined at trial. This must be correct.
89. In respect of the third limb, great care has been taken by Mr Tomlinson to try and put a case that does not contravene the rule in Dingle. The defendant relies on what has been said in other publications, on the basis that these are relevant to causation, or can be relied upon under Burstein.
90. The issues raised in respect of causation have nothing to do with the rule in Dingle. I have already set out the circumstances where the court will need to consider causation issues in respect of specific items of loss, which is different to the approach taken generally when considering general damages. This is not a case that has been pleaded in a way that makes it necessary to consider other third-party publications in order to isolate the damage caused by the Tweet. The claimant does not, for example, identify some specific item of harm as consequent on the alleged libel, allowing a defendant in principle to rely on specific alternative causes.
91. In respect of Burstein, there does not appear to be any principled basis for allowing the defendant to rely on a selection of other publications as directly relevant background context. The material makes the same or similar defamatory allegations, the truth of which is heavily disputed.
92. Taking all this together, the defendant has not been able to show that the claim does not have a real prospect of success. Whilst the claimant does not rely on evidence of actual harm, he has put forward a realistic case, which will need to be evaluated at trial once important information has been obtained (and shared) in respect of likely publication. In respect of the defendant's second limb, the defendant has in my view fallen short of establishing that in this jurisdiction the claimant's reputation was so tarnished that a serious allegation in relation to the Bitcoin Claim could not cause him serious harm.
93. In the circumstances, there is no need for me to consider the second limb of the summary judgment test, namely whether there is a compelling reason why the case or issue should be disposed at trial.