Internet Libel: Archive, Context and Snippets

March 29, 2010

In Budu v BBC [2010] EWHC 616 (QB) the claimant alleged that the BBC libelled him in an article written in 2004 which suggested that he had been rejected for a job with the Cambridgeshire Police because he was an illegal immigrant. There were further articles which gave the claimant’s side of the story, including a clear denial of the status of illegal immigrant. The claimant was not named in the first article but it was clear from the linking of the articles that he was the person referred to in the first piece. Budu delayed any action and was confined to complaining about readers of the articles post-May 2008, by which time the only means of accessing the first article was by links from the subsequent articles. There were also difficulties in producing evidence that any person accessed the articles during the relevant period. 

Giving judgment, Mrs Justice Sharp DBE said that attention had to be paid to the ‘composite meaning’ of the articles and that it was fanciful to suggest that a reader would somehow come across the first article and know that it referred to the claimant, but not go on to read the later articles. In essence, she considered that what was left of the claim was so thin and of so little import that further proceedings and further expense were not justified. 

The main point of interest in this seriously flawed claim is the suggestion that the BBC was liable by virtue of republication for the search snippet on Google that read as follows:

‘BBC NEWS| England | Cambridgeshire | Candidate denies illegal status

Ghanaian Sam Budu, Ipswich and Suffolk Council racial equality director, was later rejected as diversity manager after security vetting…’

Mrs Justice Sharp DBE considered this issue at some length (at [59]-[77]). She concluded that ‘it would not be appropriate or just in my view to make the publisher of the original webpage responsible in law for a snippet which makes a defamatory allegation (for example, because it detaches certain words from their context) not made in the original webpage itself’. However, her reasoning bears detailed examination and is set out below: 

  1. This [snippet] is described by [counsel for the claimant] as “a defamatory digest of the second article” for which the BBC is liable under the principles relating to republication set out in McManus v Beckham [2002] 1 WLR 2982. Moreover, it is said, since it has been held that Google itself is not liable as a publisher for defamatory snippets in search results, (in [Metropolitan International Schools Ltd v Designtechnica Corporation & Ors [2009] EWHC 1765 (QB)]) a fortiori the Claimant must be able to sue the BBC for the Google compilation of the second article.
  1. The BBC describes the Google snippet as a “highly selective and distorted summary” of the second article, for which the BBC should not be held liable on ordinary principles of republication. [Counsel for the BBC] makes the following additional submissions. An internet search page whose sole purpose is to direct the searcher to an underlying article should for the purposes of meaning be read with that article (applying by analogy) the well known principle in Charleston v News Group Newspapers Limited [1995] 2 AC 65 – and certainly if it is the law that the author of the article is potentially liable for the search result.
  1. As a matter of reality, he suggests it is implausible that someone would do such a search, read the link and not then read the underlying second article, or indeed the third to which it provides a link, both of which it is said put a wholly different complexion on the words complained of. Thus, whether one applies the Charleston principle, or has regard to the underlying articles as context, the result is the same. Alternatively if, he suggests a searcher is sufficiently disinterested not to go to the underlying entry, he either knows its terms already or is so disinterested in the Google search entry itself so as not to be truly a publishee.
  1. In Tarpley, Clerk v Blabey (1836) 2 Bingham New Cases 437, it was held that evidence of a letter written by a defendant was properly admitted against him at the trial of a libel action against him for the publication of it in a newspaper, even though the editor had deleted certain of the more offensive passages from the letter before publishing it. Tindal C.J. said

“If the Defendant authorised the printing of the libel in its larger and more offensive form, he gave the minor authority to print the less offensive parts of it. The case would be different if any qualifying expressions had been left out.”

  1. In Slipper v BBC [1991] 1 Q.B. 283 C.A. the Court of Appeal upheld the refusal of the lower courts to strike out a claim that the defendant was responsible for the damage caused by the publication in various newspapers of reviews of a film it had allowed to be previewed since they seemed to contain the sting of the libel which was relied on. It is clear the court proceeded on the assumption that the plaintiff would be in a position to prove at trial that the passages from the reviews repeated the defamatory sting of the film, and that the defendant could have reasonably foreseen that they would.
  1. In McManus the Court of Appeal moved away from the traditional formulations of natural and probable consequence and reasonable foreseeability when considering the basis for concluding a claimant could recover for the damage suffered by reason of its repetition (in that case, by independent third parties).
  1. Waller LJ said this:

“12. I accept that if the press articles were not publishing any part of the sting of the alleged slander and/or had no causative link with the alleged slander, it would be wrong to allow the paragraphs to be pleaded. But, the argument, so far as the articles are concerned, was that they should be struck out because the articles were not repeating the full sting of the alleged slander. The distinction being drawn was between an allegation that the claimants sold fakes generally on an habitual basis, and an allegation asserting only a part of that whole sting i.e. that the claimants habitually sold fake David Beckham autographed memorabilia. It seems to me that when one is not concerned with separate causes of action but is concerned with whether damage flowed from the original publication, even a partial publication of the original sting can be causative of damage. Furthermore, damage could flow from a report such as that which appeared in the News Shopper following the claimant (Mr. McManus) being asked about what happened. Thus it seems to me that on the first aspect the approach of the judge was over-analytical and he was wrong to strike out the paragraphs simply on the basis that they did not ‘repeat’ the slander. On that basis no question of amendment would seem to arise.”

  1. He then said this:

“34. What the law is striving to achieve in this area is a just and reasonable result by reference to the position of a reasonable person in the position of the defendant. If a defendant is actually aware (1) that what she says or does is likely to be reported, and (2) that if she slanders someone that slander is likely to be repeated in whole or in part, there is no injustice in her being held responsible for the damage that the slander causes via that publication. I would suggest further that if a jury were to conclude that a reasonable person in the position of the defendant should have appreciated that there was a significant risk that what she said would be repeated in whole or in part in the press and that that would increase the damage caused by the slander, it is not unjust that the defendant should be liable for it. Thus I would suggest a direction along the above lines rather than by reference to ‘foreseeability’.

  1. Laws LJ said this:

“39. It might be thought that the ascertainment of a causal relation between an act and a result is always a question of fact; and a decision on pure fact is, so to speak, always value-free. In very many ordinary cases, that will be quite right. But where the court has to decide whether D should be responsible to C for the effects of what was done or omitted by a third agency X, the court’s task is not purely one of ascertaining fact, and is certainly not value-free. In every such case D’s act may credibly be called a cause of the damage which flows after X has done whatever he has done. If it were otherwise, if the consequences of X’s part in the story simply had no perceptible connection with D or with the consequences of what D had done, the case would admit of a very short answer indeed: D would not be liable upon any rational approach to causation and legal responsibility. The issue for the court is not, therefore, purely one of factual causation. The true nature of the exercise does not consist in an ever closer examination of the facts to find some feature which one might at first have missed. The reality is that the court has to decide whether, on the facts before it, it is just to hold D responsible for the loss in question. Perhaps I might repeat a few words from the judgment I gave in Rahman v Arearose Ltd [2001] QB 351 at 366, 367–368, [2000] 3 WLR 1184 at 1198, 1199–1200 (a personal injury case involving successive torts), with which Henry and Schiemann LJJ agreed:

’29 … The law is that every tortfeasor should compensate the injured claimant in respect of that loss and damage for which he should justly be held responsible. To make that principle good, it is important that the elusive conception of causation should not be frozen into constricting rules …

’33 So in all these cases the real question is, what is the damage for which the defendant under consideration should be held responsible … Novus actus interveniens, the eggshell skull, and (in the case of multiple torts) the concept of concurrent tortfeasors are all no more and no less than tools or mechanisms which the law has developed to articulate in practice the extent of any liable defendant’s responsibility for the loss and damage which the claimant has suffered.’

40. The problem of a second cause or novus actus interveniens should have no more absolutist or metaphysical overtones for the law of defamation than it does for the law of negligence. Of course the conception of a duty of care has no analogue in defamation. But that, if anything, serves to simplify the approach to be taken in defamation cases to the task of ascertaining in any given case the extent of any liable defendant’s responsibility for the loss and damage which the claimant has suffered where a potential novus actus is involved.

41. The defamation cases have over time been girt about with unhelpful complexities. Thus (1) Ward v Weeks (1830) 7 Bing 211, 131 ER 81 gave rise to the suggestion that there was a rule of law to the effect that a defendant could in no circumstances be held responsible to the claimant for the consequences of an effective repetition of a slander perpetrated by him. I think that was plainly wrong. (2) A distinction was drawn, having no readily perceptible basis in principle, between the effects of a repeat publication in the case where a slander was actionable per se (the repeat publication might in some circumstances be relied on) and the case where it was not so actionable (the repeat publication might never be relied on): see Parkins v Scott (1862) 1 H & C 153, 158 ER 839. (3) The circumstances in which repeat publication might be relied on became ossified into categories: see Speight v Gosnay (1891) 60 LJQB 231. (4) The courts’ perception of causation was not advanced by an uncomfortable oscillation between the old language of ‘natural and probable cause’ and the later formulation ‘reasonable foreseeability’: this can be seen, with great respect, as late as this court’s judgment in Slipper v BBC [1991] 1 All ER 165, [1991] 1 QB 283, to which Waller LJ has referred.

42.The law needs to be simplified. The root question is whether D, who has slandered C, should justly be held responsible for damage which has been occasioned, or directly occasioned, by a further publication by X. I think it plain that there will be cases where that will be entirely just. The observation of Bingham LJ as he then was in Slipper‘s case [1991] 1 All ER 165 at 179, [1991] 1 QB 283 at 300 that ‘[d]efamatory statements are objectionable not least because of their propensity to percolate through underground channels and contaminate hidden springs’ states an ancient and persistent truth, long ago vividly described in Virgil’s account of Aeneas and Dido, Queen of Carthage (see Aeneid iv 173–188).

43. It will not however in my judgment be enough to show that D’s slander is a cause of X’s further publication: for such a cause might exist although D could have no reason to know of it; and then to hold D responsible would not be just. This is why the old formula, ‘natural and probable cause’, is inapt even as a figurative description of the relationship that needs to be shown between D’s slander and the further publication if D is to be held liable for the latter. It must rather be demonstrated that D foresaw that the further publication would probably take place, or that D (or a reasonable person in D’s position) should have so foreseen and that in consequence increased damage to C would ensue.

44. Such an approach, I hope, may go some modest distance to demythologise the law of defamation. I make it clear that I intend what I have said to be in conformity with Waller LJ’s suggestion at the end of his judgment as to how a jury might be directed, though with deference to him I think that any avoidance of the term ‘foreseeability’ is commended by the need for clarity rather than adherence to principle, for in principle the approach he proposes, and for what it is worth my own reasoning, require that the damage in question flowing from X’s act be foreseen or foreseeable by D, or the reasonable person in D’s position.”

  1. In this case, the BBC is sued for the Google snippet as a separate republication. Thus a more restrictive test for liability applies than if it were sued for the damage flowing from the repetition of the original publication (see Gatley on Libel and Slander, 11th edition, para 6.36 and following). But for an original publisher to be liable whether for a republication (sued on as a separate cause of action) or for damage suffered by the repetition of the original publication, the secondary publication must, as a minimum, repeat the whole, or at least part of the sting of the original libel. This in my judgment accords with long-standing authority, with good sense, and what is just. Since I have taken the view that the second article is incapable of bearing the “security risk” meaning, which is the only defamatory meaning attributed to the Google snippet, it follows that the BBC cannot be liable, even arguably, for its “republication” as the Claimant contends. The Google snippet simply extracts a sentence from the second article, divorced from the context which (on the Claimant’s case) bears a meaning which in my judgment the article which it is said to republish, is not capable of bearing.
  1. That is sufficient to dispose of the Claimant’s case on republication. Indeed [Counsel for the claimant] expressly accepted during the course of argument, that if the second and third articles do not contain the “security risk” allegation, it would follow the BBC could not be liable for the Google snippet, on ordinary principles of causation.
  1. In Metropolitan, Eady J decided that the operator of a search engine could not be liable for the publication of a snippet which it played no part in formulating; and given it had no knowing involvement in the publication of the relevant words, it could not be characterised as a publisher at common law of the search result, but was a mere facilitator. [Counsel for the claimant] at one point also complained that to hold the BBC could not be liable for the snippet would result in a potential injustice, since a person in the position of the Claimant cannot then sue either Google or the original publisher of the underlying article for the snippet.
  1. However, liability for publication cannot accrue by default, but can only attach on principled grounds. If the real complaint is in respect of the original article/webpage from which the snippet is, or is said to be extracted then a claimant genuinely interested in compensation or vindication can pursue a claim, including for injunctive relief, against the original publisher of that webpage.
  1. Such a claimant might not necessarily be shut out from taking steps in respect of a snippet. As was contemplated by Eady J In Metropolitan a company such as Google might at some point become liable, if the publication of a defamatory search result (or a “signpost to a conduit” as he described it) continued after notification of the specific URL from which the words complained of originated (though not, as Eady J said, while efforts were made by the operator of the search engine to achieve a “take down”).
  1. It should also be borne in mind in my view when considering what is said to be this potential injustice, and indeed, possibly, the issue of liability at all that this form of snippet is a robot produced piece of text, created by the search terms entered by the person doing the search, which simply points to the underlying text, accessed via the hyperlink displayed. Search engines after all are not used by accident. In the 21st century, they are accessed and used by computer literate individuals as a method of finding and accessing underlying information: and would in my view be read – or glanced at – in that light. It might also be thought that those who use Google search engines are well aware that such a snippet is merely a fragment of a larger whole (the underlying publication); by analogy, a tiny extract torn at random from a page to which no human publisher has attached any particular significance.
  1. One approach might be to regard the underlying article as necessary context, or actually as part of the same publication as [Counsel for the BBC] suggested, so that the snippet could not then be read in isolation from the underlying article to which it provides a direct link. This in itself would not necessarily be a problem free approach, however, as similar practical problems to those considered by Eady J in Metropolitan in relation to injunctive relief and continuing liability might then arise.
  1. Be that as it may, it would not be appropriate or just in my view to make the publisher of the original webpage responsible in law for a snippet which makes a defamatory allegation (for example, because it detaches certain words from their context) not made in the original webpage itself.