Tamiz v Google: the Court of Appeal Verdict

March 19, 2013

The Court of Appeal has delivered its Tamiz v Google judgment on Google’s potential defamation liability for comments posted to a blog hosted on its Blogger platform.  

Although the Court ultimately decided that the claim against Google was an abuse of process, the judgment is of most interest for what is said about Google’s position as a publisher at common law and its possible defences as a secondary publisher under the Defamation Act 1996. 

When analysing the decision, it is as well to remember that this was a jurisdiction challenge by Google Inc, the US provider of the Blogger platform.  For Google to succeed on the issues under appeal the court would have had to be satisfied that there was no good arguable case against Google – on the face of it a high hurdle. 

At first instance Eady J decided that Google could not be characterised as a publisher at common law, even after being notified of the defamatory statement – so no case could be made against it.  To reverse that decision, the Court of Appeal did not have to decide that Google was a publisher, only that there was an arguable case that it was; and then that there was an arguable case that Google was not covered by any available defences.

Eady J’s judgment laid down a bright line precluding defamation liability for at least some types of intermediary.  The position after the Court of Appeal’s decision is fuzzier.   While confirming that pre-notification Google was not a publisher, the Court of Appeal decided that at some point after notification it was arguably a publisher. 

Also arguably, on the facts, Google was not able to avail itself of defences under the Defamation Act 1996.  The Court found it unnecessary to address the possibility of a hosting defence under the E-commerce Directive.

The finding of abuse of process was founded on the short amount of time for which the comments were available in the period after Google could be considered liable as a publisher and before they were removed. Jurisdiction was declined on that basis. 

So what can we take from the Court of Appeal’s decision? The three key points are that Google was:

·         not a publisher pre-notification.

·         arguably a publisher at some point post-notification

·         arguably not protected by the Defamation Act 1996, s 1 at some point after notification.


Not a publisher pre-notification

Strictly this aspect of the Court of Appeal’s judgment was obiter, since Tamiz’s claim covered only the period after Google was notified of his complaint. Nevertheless, like Eady J at first instance, the Court of Appeal considered the position pre-notification.

At common law there are two categories of publishers: primary publishers, who are strictly liable for statements they publish; and secondary publishers, who, while they still are regarded as having participated in the publication, have at common law an ‘innocent dissemination’ defence available to them.  Someone who merely facilitates publication is not a publisher at all.

The Court of Appeal found that Google’s provision of the Blogger service plainly facilitated publication of the blog and comments. However, its involvement was not such as to make it a primary publisher. It was not in a position comparable to that of an author, an editor or the corporate proprietor of a newspaper. Bloggers were plainly independent of Google and did not act in any sense on its behalf or in its name.

The Court was less certain that Google’s role was not that of a secondary publisher, expressing itself only ‘very doubtful’ that it was so. The Court’s reasoning on this point, on the face of it, conflates what one would think should be three separate issues:

·         the degree of involvement necessary to be regarded as a publisher (facilitation versus participation),

·         the distinction between primary and secondary publisher, and

·         the conditions under which a secondary publisher can escape strict liability.

However, the same confusion has bedevilled the innocent dissemination cases since at least Vizetelly v Mudie’s Select Library, a century-old case in which Romer LJ and Vaughan Williams LJ got in a tangle about whether innocent dissemination precluded publication or absolved the disseminator from liability for publication.

In any event the Court of Appeal in Tamiz held that the argument that Google was a secondary publisher could get nowhere for the pre-notification period since it could not be said that Google either knew or ought reasonably to have known of the defamatory comments.

Although the Court of Appeal did not address the point directly, there remains a fundamental difficulty in reconciling a finding that Google was not a publisher (whether primary or secondary) with the broad ground on which Morland J considered in 1999 in Godfrey v Demon Internet that Demon Internet was a publisher of Usenet newsgroups hosted by it: ie whenever there was a transmission of a defamatory posting from the storage of the defendant’s news server, the defendant was a publisher of that posting.

As I suggested in my article on the first instance decision in Tamiz v Google, this looks like the death throes of that aspect of Godfrey v Demon Internet.

A publisher post-notification

At first instance, Eady J held that Google was not a publisher even after notification of the existence of a defamatory comment to a blog that it hosted. He likened Google to the owner of a wall on which people chose to inscribe graffiti, for which the owner was not responsible.

The Court of Appeal preferred to think of Google as more like the provider of a notice board: one for which, moreover, Google provided a blogger with design tools for his part of the notice board and facilities for displaying advertisements alongside notices; and, according to the Court of Appeal most importantly, Google provided the notice board on terms of its choice and could readily remove or block access to any notice which did not comply with those terms.

As such Google’s blog platform fell within the scope of the reasoning of Byrne v Deane. In that case from the 1930s, an anonymous defamatory notice was affixed to a wall in a golf club and remained there after it was drawn to the attention of the golf club’s secretary. The golf club proprietor and the secretary were held to have published the notice after notification, on the basis that by failing to remove it they had made themselves responsible for its continuing presence.

The Court of Appeal held that:

‘if Google Inc allows defamatory material to remain on a Blogger blog after it has been notified of the presence of that material, it might be inferred to have associated itself with, or to have made itself responsible for, the continued presence of that material on the blog and thereby to have become a publisher of the material.’

Google argued that the vast difference in scale between the Blogger set-up and the small club-room in Byrne v Deane made such an inference unrealistic and that nobody would view a comment on a blog as something with which Google had associated itself or for which it had made itself responsible by taking no action to remove it after notification of a complaint. 

While the Court acknowledged that those were matters for argument, they were not decisive in Google’s favour where the court was concerned only with whether there was an arguable case of publication.

One issue that is left uncertain when applying Byrne v Deane is whether it is relevant to consider if the defendant platform has knowledge of the words complained of and their illegality or potential illegality. Eady J at first instance in Tamiz thought these might be relevant. The Court of Appeal said that no inference of publication should be drawn until Google had had a reasonable time in which to act to remove the defamatory comments; and that it was arguable that five weeks was sufficiently long for an adverse inference to be drawn against Google. 

However, a reasonable time in which to act to remove the defamatory comments is not the same thing as a reasonable time in which to determine actual or potential illegality. The judgments in Byrne v Deane itself strongly suggest that the trigger is responsibility not for the defamation, but for the continued publication. There was no suggestion in Byrne v Deane of any need to show that the defendant knew the statement was defamatory or unlawful, or of any grace period in which to determine unlawfulness. 

At any rate a single poem written on a piece of paper stuck a to a golf club wall in August 1935 does seem a shallow foundation on which to construct an entire edifice of answers for a 21st century internet platform that hosts blogs and associated comments amounting to, according to Eady J at first instance, half a trillion words and 250,000 new words every minute.

Defence under the Defamation Act 1996, s 1

In the event that it was held to have published the statement, Google invoked the statutory ‘secondary publisher’ defence under the Defamation Act 1996, s 1. This defence has three elements:

(a)    That the defendant was not the author, editor or publisher of the statement complained of. Each of these has specifically defined meanings. In particular, a ‘publisher’ is someone whose business is issuing material to the public. This is completely different from a publisher at common law as discussed above.

(b)   That the defendant took reasonable care in relation to its publication. This provision creates the notorious ‘Catch 22’: how can one take reasonable care without becoming an editor under s 1(a)?

(c)    That the defendant did not know, and had no reason to believe, that what he did caused or contributed to the publication of a defamatory statement. 

Eady J had held that Google was not a publisher under these provisions. The Court of Appeal agreed. It could not sensibly be said to have ‘issued’ the defamatory comments. Its involvement was of a kind analogous to the example given in s 1(3)(e) of someone who should not be regarded as an author, editor or publisher:

‘only involved … as the operator of or provider of access to a communications system by means of which the statement is transmitted, or made available, by a person over whom he has no effective control.’

Eady J considered that by ‘effective control’ it was likely that the draftsman had in mind effective day-to-day control rather than the possibility of intervention in reliance on a contractual term about the permitted content of a web page.  

These comments may be compared with the rulings of Gray J in MORI v BBC (unreported, 17 June 1999 – not cited by either Eady J or the Court of Appeal) concerning the meaning of ‘effective control’ in the live broadcast exception under s 1(3)(d). Although Eady J’s approach was generally similar, they appear to diverge on the possible relevance of a contractual relationship. 

The Court of Appeal agreed with Eady J. The existence of a contractual term about the content of blogs was not sufficient to give Google effective control over the person who posted the defamatory comments. 

Both Eady J and the Court of Appeal concentrated on s 1(3)(e) as the closest analogous example to the Blogger host. However, s 1(3)(e) is on the face of it aimed more at transmission-type operations. Section 1(3)(c), which does not contain a ‘no effective control’ limitation, appears to be more obviously aimed at operations in the nature of hosting. It applies to any person ‘only involved … in operating or providing any equipment, system or service by means of which the statement is retrieved, copied, distributed or made available in electronic form.’

While the distinction between subsections 1(3)(c) and 1(3)(e) is by no means clear and they may possibly overlap, the focus on s 1(3)(e) for a blogging platform is mystifying. 

1996 Act defence pre-notification

For the period prior to notification, the Court of Appeal agreed with Eady J that Google would if necessary have a defence under s 1.  There was no basis for concluding that Google had not taken reasonable care in relation to the publication (s 1(1)(b)), or that it knew or had reason to believe that what he did caused or contributed to the publication of a defamatory statement (s 1(1)(c)).

It is noteworthy that in McGrath v Dawkins, a case not discussed by the Court of Appeal, HHJ Moloney QC speculated that it might be reasonable for a wholly automated large web site such as Amazon to take no pre-publication steps at all in relation to user-submitted book reviews. That also seems to be the import of the Court of Appeal’s conclusion in Tamiz as regards the pre-notification period.

1996 Act defence post-notification

For the period after notification, the Court of Appeal agreed with Eady J that Google had taken reasonable care in relation to the continued publication of the comments. Unlike in Godfrey v Demon Internet, where Demon Internet had taken no steps at all after notification, Google had passed on the complaint about the comment to the blogger.  

However the Court of Appeal disagreed with Eady J about s 1(1)(c). For the same reasons as applied to its publication analysis under Byrne v Deane, the Court held that it was arguable that at some point following notification it knew or had reason to believe that its continued hosting of the material in question caused or contributed to the publication of a defamatory statement. For this reason, the Court was not satisfied that Google had an unassailable defence under s 1 of the 1996 Act. 

This was the approach of HHJ Parkes QC in Davison v Habeeb, which the Court of Appeal preferred to that of Eady J at first instance.

However, Davison v Habeeb concerned statements made in posts by the actual blogger, not in comments to the blogger’s posts. In Tamiz the blogger was itself, as regards comments, an intermediary who had the ability to remove comments. It is questionable whether, when the blog platform has notified the complaint to a more proximate intermediary, s 1(1)(c) should be applied to the underlying blog platform as strictly as when the complaint is made about postings by the blogger himself.

In any event, the Court of Appeal found only that the point on s 1(1)(c) was arguable against Google – not that Google would be unable to rely on the defence.


As regards common-law publication and s 1 of the 1996 Act, the upshot of the Court of Appeal’s decision is that for a blog platform to pass a defamation complaint back to the blogger is not necessarily sufficient to avoid liability. It may still be at risk of becoming a publisher at common law and of being unable to rely on the 1996 Act.  

This leaves the E-commerce Directive, for damages, as potentially a stronger line of defence for a hosting platform. It has always been thought to be so, because to preclude reliance on the defence requires actual or constructive knowledge not just of the defamatory nature of the statement but of illegality. That requires some information about the claim and possible defences (see Bunt v Tilley). 

However, the Directive does not provide a complete defence to proceedings, since it does not prevent a claim for an injunction. A platform that is keen to remove itself completely from initiated proceedings will still, therefore, have an incentive to remove allegedly defamatory material even if it is not fixed with knowledge of illegality.

Graham Smith is a partner in Bird & Bird LLP and specialises in IT, internet and intellectual property law. His Cyberleagle blog is at www.cyberleagle.com.