SUPREME COURT OF CANADA
19 OCTOBER 2011
(with whom Binnie, LeBel, Charron, Rothstein and Cromwell JJ joined)
To succeed in an action for defamation, the plaintiff must prove on a balance of probabilities that the defamatory words were published, that is, that they were “communicated to at least one person other than the plaintiff” (Grant v Torstar Corp., 2009 SCC 61,  3 S.C.R. 640, at para. 28).
A hyperlink is a device routinely used in articles on the Internet whereby a word or phrase is identified, often with underlining, as being a portal to additional, related information. Clicking on the hyperlink connects the reader to that information.
The legal issue in this appeal is whether hyperlinks that connect to allegedly defamatory material can be said to “publish” that material.
Wayne Crookes is the President and sole shareholder of West Coast Title Search Ltd. He brought a series of lawsuits against those he claimed were responsible for allegedly defamatory articles published on a number of websites, arguing that the articles represented a “smear campaign” against him and other members of the Green Party of Canada.
Jon Newton owns and operates a website in British Columbia containing commentary about various issues, including free speech and the Internet. One of the articles he posted on it was called “Free Speech in Canada”. The article contained hyperlinks to other websites, which in turn contained information about Mr. Crookes.
Mr. Crookes sued Mr. Newton on the basis that two of the hyperlinks he created connected to defamatory material, and that by using those hyperlinks, Mr. Newton was publishing the defamatory information. One was a “shallow” hyperlink, which takes the reader to a webpage where articles are posted, and the other was a “deep” hyperlink, which takes the reader directly to an article (Matthew Collins, The Law of Defamation and the Internet (3rd ed. 2010), at para. 2.43). Both shallow and deep hyperlinks require the reader to click on the link in order to be taken to the content.
The two hyperlinks, identified by underlining, were in the following excerpt from Mr. Newton’s posting:
Under new developments, .... I’ve just met Michael Pilling, who runs OpenPolitics.ca. Based in Toronto, he, too, is being sued for defamation. This time by politician Wayne Crookes.
We’ve decided to pool some of our resources to focus more attention on the appalling state of Canada’s ancient and decrepit defamation laws and tomorrow, p2pnet will run a post from Mike [Pilling] on his troubles. He and I will also be releasing a joint press statement in the very near future. [A.R., at p. 125]
OpenPolitics.ca was hyperlinked to the Open Politics website where several articles were posted and were said by Mr. Crookes to be defamatory. Wayne Crookes was hyperlinked to an allegedly defamatory article called “Wayne Crookes”, published anonymously on the website www.USGovernetics.com.
Mr. Crookes wrote to Mr. Newton asking him to remove the two hyperlinks. When he got no response, Mr. Crookes’ lawyer wrote to Mr. Newton, repeating the request. Mr. Newton refused to remove the hyperlinks.
Mr. Crookes sued Mr. Newton for defamation in British Columbia. He did not allege that anything on Mr. Newton’s webpage was itself defamatory. Rather, he argued that by creating hyperlinks to the allegedly defamatory articles, or by refusing to remove those hyperlinks when told of their defamatory character, Mr. Newton himself became a publisher of the articles. By then, Mr. Newton’s article had been “viewed” 1,788 times. There is no information in the record about whether, or how many times, the hyperlinks themselves had been clicked on or followed.
At trial, Kelleher J. concluded that the mere creation of a hyperlink in a website does not lead to a presumption that someone actually used the hyperlink to access the impugned words (2008 BCSC 1424, 88 B.C.L.R. (4th) 395). He agreed with Mr. Newton’s submission that hyperlinks were analogous to footnotes since they only refer to another source without repeating it. Since there was no repetition, there was no publication. And in the absence of evidence that anyone other than Mr. Crookes used the links and read the words to which they linked, there could not be a finding of publication.
In the Court of Appeal, Saunders J.A., with whom Bauman J.A. concurred, held that the appeal should be dismissed (2009 BCCA 392, 96 B.C.L.R. (4th) 315). Agreeing with the trial judge, she found that reference to an article containing defamatory comments without repetition of the comments themselves is analogous to a footnote or a card index in a library and should not be found to constitute republication of the defamation. While some words in an article may suggest that a particular hyperlink is an “encouragement or invitation” to view the impugned site, she saw no such encouragement or invitation in this case. She also refused to accept that the number of “hits” on the article itself was a sufficient basis for drawing an inference that a third party had read the defamatory words (paras. 89 and 92).
Prowse J.A. dissented. While she agreed that the mere fact that Mr. Newton had created hyperlinks to the impugned sites did not make him a publisher of the material found at the hyperlinked sites, she did not accept the “footnote analogy” as being dispositive of the publication issue (para. 60). In her view, the fact that Mr. Newton’s website had been viewed 1,788 times made it “unlikely” that no one had followed the hyperlinks and read the impugned articles (para. 70). Moreover, the context of Mr. Newton’s article suggests that readers were in fact encouraged or invited to click on the links. In her view, therefore, there was publication.
In British Columbia, pursuant to the Libel and Slander Act, R.S.B.C. 1996, c. 263, publication is deemed to have occurred in certain situations. There is, however, no such presumption in relation to material published on the Internet. Nonetheless, Mr. Crookes argued that when a hyperlink has been inserted on a webpage, it should be presumed that the content to which the hyperlink connects has been brought to the knowledge of a third party and has therefore been published. For the reasons that follow, I would not only reject such a presumption, I would conclude that a hyperlink, by itself, should never be seen as “publication” of the content to which it refers.
Mr. Crookes also complains that the Court of Appeal imposed too high a burden of proof, essentially requiring direct evidence that a third party followed the hyperlink to the allegedly defamatory content. This, he claims, deprives him of the ability to rely on an inference that at least one person followed one of the impugned hyperlinks to the allegedly defamatory content, and that the defamatory meaning has therefore been published. (See Gaskin v Retail Credit Co.,  S.C.R. 297.) In view of my conclusion that hyperlinking is not, in and of itself, publication, there is no need to address this argument.
To prove the publication element of defamation, a plaintiff must establish that the defendant has, by any act, conveyed defamatory meaning to a single third party who has received it (McNichol v Grandy,  S.C.R. 696, at p. 699). Traditionally, the form the defendant’s act takes and the manner in which it assists in causing the defamatory content to reach the third party are irrelevant:
There are no limitations on the manner in which defamatory matter may be published. Any act which has the effect of transferring the defamatory information to a third person constitutes a publication.
(Stanley v Shaw, 2006 BCCA 467, 231 B.C.A.C. 186, at para. 5, citing Raymond E. Brown, The Law of Defamation in Canada (2nd ed.), vol. 1, at No. 7.3.)
See also Hiltz and Seamone Co. v Nova Scotia (Attorney General) (1997), 164 N.S.R. (2d) 161 (S.C.), at para. 21, aff’d in part (1999), 173 N.S.R. (2d) 341 (C.A.); Grant, at para. 119; “Truth” (N.Z.) Ltd. v Holloway,  1 W.L.R. 997 (P.C.); Lambert v Thomson,  O.R. 341 (C.A.), per Rowell C.J.O.; Pullman v Walter Hill & Co.,  1 Q.B. 524 (C.A.), at p. 527, per Lord Esher M.R.
Mr. Crookes argues that, under this definition, a person who includes a hyperlink on a webpage has “published” any defamatory remarks to which the hyperlink leads, because that person has done an act which “has the effect of transferring the defamatory information” to any third person who clicks on the link.
Under this sole disseminator/sole reader paradigm, the breadth of activity captured by the traditional publication rule is vast. In R. v Clerk (1728), 1 Barn. K.B. 304, 94 E.R. 207, for example, a printer’s servant, whose only role in an act of publication was to “clap down” the printing press, was found responsible for the libels contained in that publication, despite the fact that he was not aware of the contents (p. 207). In Hird v Wood (1894), 38 S.J. 234 (C.A.), pointing at a sign displaying defamatory words was held to be evidence of publication. Other cases have also held that acts merely facilitating communication can amount to publication: see, e.g., Buchanan v Jennings,  UKPC 36,  1 A.C. 115; Polson v Davis, 635 F.Supp. 1130 (D. Kan. 1986), aff’d 895 F.2d 705 (10th Cir. 1990); Crain v Lightner, 364 S.E.2d 778 (W. Va. 1987), at p. 785; and Spike v Golding (1895), 27 N.S.R. 370 (S.C. in banco). And in McNichol v Grandy, the defendant was found to be liable when he raised his voice and made defamatory statements that were overheard by someone in another room.
The publication rule has also captured the following range of conduct:
[The defamatory meaning] may be communicated directly by the defendant either orally, or in some written or printed form, or by way of a symbolic ceremony, dramatic pantomime, mime, brochure, gesture, handbill, letter, photograph, placard, poster, sign, or cartoon. It may be inscribed on a blackboard, posted on a mirror or a telephone pole, or placed on the wall of a building or the gable wall of the defendant’s property, or on the front of a cheque, or entered in a database, or accessed on or downloaded from a website on the internet. It may appear on an ariel banner flown behind an airplane, or someone’s attention may be drawn by the defendant to a poster, or a defamatory writing already in circulation. A third person may be given access to defamatory material, or defamatory matter may be left in a place where others can see it, or the defendant may request others to go to a place where the defamatory information is available to see and read it, or it may be set into motion as a result of the defendant’s death. In each case there is a publication.
(Raymond E. Brown, Brown on Defamation (2nd ed. (loose-leaf)), at para. 7.3)
Defendants obtained some relief from the rule’s significant breadth with the development of the “innocent dissemination” defence, which protects “those who play a secondary role in the distribution system, such as news agents, booksellers, and libraries”: Allen M. Linden and Bruce Feldthusen, Canadian Tort Law (8th ed. 2006), at pp. 783-84; see also Society of Composers, Authors and Music Publishers of Canada v Canadian Assn. of Internet Providers, 2004 SCC 45,  2 S.C.R. 427 (“SOCAN”), at para. 89; Philip H. Osborne, The Law of Torts (4th ed. 2011), at p. 411. Such “subordinate distributors” may escape liability by showing that they “have no actual knowledge of an alleged libel, are aware of no circumstances to put them on notice to suspect a libel, and committed no negligence in failing to find out about the libel” (SOCAN, at para. 89; Vizetelly v Mudie’s Select Library, Ltd.,  2 Q.B. 170 (C.A.), at p. 180; Brown, at para. 7.12(6)(c); and also Sun Life Assurance Co. of Canada v W. H. Smith and Son Ltd. (1934), 150 L.T. 211 (C.A.), at pp. 212-14).
Recently, jurisprudence has emerged suggesting that some acts are so passive that they should not be held to be publication. In Bunt v Tilley,  EWHC 407,  3 All E.R. 336 (Q.B.), considering the potential liability of an Internet service provider, the court held that in order to hold someone liable as a publisher, “[i]t is not enough that a person merely plays a passive instrumental role in the process”; there must be “knowing involvement in the process of publication of the relevant words” (para. 23 (emphasis in original); see also Metropolitan International Schools Ltd. v Designtechnica Corpn.,  EWHC 1765,  1 W.L.R. 1743 (Q.B.)).
Acknowledging these developments, the question on this appeal is whether a simple reference – like a hyperlink – to defamatory information is the type of act that can constitute publication. Some helpful guidance on this point is available in two American cases. In Klein v Biben, 296 N.Y. 638 (1946), the New York Court of Appeals decided that a statement saying “For more details about [the plaintiff], see the Washington News Letter in The American Hebrew, May 12, 1944” (p. 639) was not a republication of the May 12 libel.
And in MacFadden v Anthony, 117 N.Y.S.2d 520 (Sup. Ct. 1952), a complaint of defamation was dismissed in a case where a radio host “called attention to [an allegedly defamatory] article in Collier’s Magazine” (p. 521). Relying on Klein, the court concluded that referring to the article was neither a republication nor a publication of the libel.
These cases were relied on in Carter v B.C. Federation of Foster Parents Assn., 2005 BCCA 398, 42 B.C.L.R. (4th) 1, where the plaintiff alleged that, by mentioning the Internet address of an online discussion forum, the publisher of a newsletter was responsible for republishing defamatory comments published on that site. Relying on MacFadden and Klein for the proposition that “reference to an article containing defamatory comment without repetition of the comment itself should not be found to be a republication of such defamatory comment” (para. 12), Hall J.A. held that there was no publication.
I agree with this approach. It avoids a formalistic application of the traditional publication rule and recognizes the importance of the communicative and expressive function in referring to other sources. Applying such a rule to hyperlinks, as the reasons of Justice Deschamps demonstrate, has the effect of creating a presumption of liability for all hyperlinkers, an untenable situation in my view.
A reference to other content is fundamentally different from other acts involved in publication. Referencing on its own does not involve exerting control over the content. Communicating something is very different from merely communicating that something exists or where it exists. The former involves dissemination of the content, and suggests control over both the content and whether the content will reach an audience at all, while the latter does not. Even where the goal of the person referring to a defamatory publication is to expand that publication’s audience, his or her participation is merely ancillary to that of the initial publisher: with or without the reference, the allegedly defamatory information has already been made available to the public by the initial publisher or publishers’ acts. These features of references distinguish them from acts in the publication process like creating or posting the defamatory publication, and from repetition.
Hyperlinks are, in essence, references. By clicking on the link, readers are directed to other sources. Hyperlinks may be inserted with or without the knowledge of the operator of the site containing the secondary article. Because the content of the secondary article is often produced by someone other than the person who inserted the hyperlink in the primary article, the content on the other end of the link can be changed at any time by whoever controls the secondary page. Although the primary author controls whether there is a hyperlink and what article that word or phrase is linked to, inserting a hyperlink gives the primary author no control over the content in the secondary article to which he or she has linked. (See David Lindsay, Liability for the Publication of Defamatory Material via the Internet (2000), at pp. 14 and 78-79; Collins, at paras. 2.42 to 2.43 and 5.42.)
These features – that a person who refers to other content generally does not participate in its creation or development – serve to insulate from liability those involved in Internet communications in the United States: see Communications Decency Act of 1996, 47 U.S.C. §230 (1996); see also Jack M. Balkin, “The Future of Free Expression in a Digital Age” (2009), 36 Pepp. L. Rev. 427, at pp. 433-34; Zeran v America Online, Inc., 129 F.3d 327 (4th Cir. 1997); Barrett v Rosenthal, 146 P.3d 510 (Cal. 2006); Fair Housing Council of San Fernando Valley v Roommates.Com, LLC, 521 F.3d 1157 (9th Cir. 2008).
Although the person selecting the content to which he or she wants to link might facilitate the transfer of information (a traditional hallmark of publication), it is equally clear that when a person follows a link they are leaving one source and moving to another. In my view, then, it is the actual creator or poster of the defamatory words in the secondary material who is publishing the libel when a person follows a hyperlink to that content. The ease with which the referenced content can be accessed does not change the fact that, by hyperlinking, an individual is referring the reader to other content. (See Dell Computer Corp. v Union des consommateurs, 2007 SCC 34,  2 S.C.R. 801, at paras. 97-102.)
Hyperlinks thus share the same relationship with the content to which they refer as do references. Both communicate that something exists, but do not, by themselves, communicate its content. And they both require some act on the part of a third party before he or she gains access to the content. The fact that access to that content is far easier with hyperlinks than with footnotes does not change the reality that a hyperlink, by itself, is content-neutral – it expresses no opinion, nor does it have any control over, the content to which it refers.
This interpretation of the publication rule better accords with our Court’s recent jurisprudence on defamation law. This Court has recognized that what is at stake in an action for defamation is not only an individual’s interest in protecting his or her reputation, but also the public’s interest in protecting freedom of expression: Hill v Church of Scientology of Toronto,  2 S.C.R. 1130.
Pre-Charter approaches to defamation law in Canada largely leaned towards protecting reputation. That began to change when the Court modified the “honest belief” element to the fair comment defence in WIC Radio Ltd. v Simpson, 2008 SCC 40,  2 S.C.R. 420, and when, in Grant, the Court developed a defence of responsible communication on matters of public interest. These cases recognize the importance of achieving a proper balance between protecting an individual’s reputation and the foundational role of freedom of expression in the development of democratic institutions and values (Grant, at para. 1; Hill, at para. 101).
Interpreting the publication rule to exclude mere references not only accords with a more sophisticated appreciation of Charter values, but also with the dramatic transformation in the technology of communications. See June Ross, “The Common Law of Defamation Fails to Enter the Age of the Charter” (1996), 35 Alta. L. Rev. 117; see also Jeremy Streeter, “The ‘Deception Exception’: A New Approach to Section 2(b) Values and Its Impact on Defamation Law” (2003), 61 U.T. Fac. L. Rev. 79; Denis W. Boivin, “Accommodating Freedom of Expression and Reputation in the Common Law of Defamation” (1996-1997), 22 Queen’s L.J. 229; Lewis N. Klar, Tort Law (4th ed. 2008), at pp. 746-47; Robert Danay, “The Medium is not the Message: Reconciling Reputation and Free Expression in Cases of Internet Defamation” (2010), 56 McGill L.J. 1; the Hon. Frank Iacobucci, “Recent Developments Concerning Freedom of Speech and Privacy in the Context of Global Communications Technology” (1999), 48 U.N.B.L.J. 189; Reno v American Civil Liberties Union, 521 U.S. 844 (1997), at p. 870.
The Internet’s capacity to disseminate information has been described by this Court as “one of the great innovations of the information age” whose “use should be facilitated rather than discouraged” (SOCAN, at para. 40, per Binnie J.). Hyperlinks, in particular, are an indispensable part of its operation. As Matthew Collins explains, at para. 5.42:
Hyperlinks are the synapses connecting different parts of the world wide web. Without hyperlinks, the web would be like a library without a catalogue: full of information, but with no sure means of finding it.
(See also Lindsay, at pp. 78-79; Mark Sableman, “Link Law Revisited: Internet Linking Law at Five Years” (2001), 16 Berkeley Tech. L.J. 1273, at p. 1276.)
The centrality of the role of hyperlinks in facilitating access to information on the Internet was also compellingly explained by Anjali Dalal in “Protecting Hyperlinks and Preserving First Amendment Values on the Internet” (2011), 13 U.Pa.J. Const. L. 1017 (pp. 1019 & 1022):
Hyperlinks have long been understood to be critical to communication because they facilitate access to information. They provide visitors on one website a way to navigate to internally referenced words, phrases, arguments, and ideas. Under this view, if the Internet is an endless expanse of information where “any person .... can become a pamphleteer” then “[h]yperlinks are the paths among websites, creating the bustling street corners for distribution of those pamphlets and inviting passersby to engage more deeply with the issues raised.”
.... While the concerns motivating cases brought against hyperlinks are often legitimate, limiting the use of links poses a significant danger to communication and future innovation.
[emphasis in original; footnotes omitted]
The Internet cannot, in short, provide access to information without hyperlinks. Limiting their usefulness by subjecting them to the traditional publication rule would have the effect of seriously restricting the flow of information and, as a result, freedom of expression. The potential “chill” in how the Internet functions could be devastating, since primary article authors would unlikely want to risk liability for linking to another article over whose changeable content they have no control. Given the core significance of the role of hyperlinking to the Internet, we risk impairing its whole functioning. Strict application of the publication rule in these circumstances would be like trying to fit a square archaic peg into the hexagonal hole of modernity.
I do not for a moment wish to minimize the potentially harmful impacts of defamatory speech on the Internet. Nor do I resile from asserting that individuals’ reputations are entitled to vigorous protection from defamatory comments. It is clear that “the right to free expression does not confer a licence to ruin reputations” (Grant, at para. 58). Because the Internet is a powerful medium for all kinds of expression, it is also a potentially powerful vehicle for expression that is defamatory. In Barrick Gold Corp. v Lopehandia (2004), 71 O.R. (3d) 416 (C.A.), at para. 32, Blair J.A. recognized the Internet’s “tremendous power” to harm reputation, citing with approval the following excerpt from Lyrissa Barnett Lidsky, “Silencing John Doe: Defamation & Discourse in Cyberspace” (2000), 49 Duke L.J. 855, at pp. 863-64:
Although Internet communications may have the ephemeral qualities of gossip with regard to accuracy, they are communicated through a medium more pervasive than print, and for this reason they have tremendous power to harm reputation. Once a message enters cyberspace, millions of people worldwide can gain access to it. Even if the message is posted in a discussion forum frequented by only a handful of people, any one of them can republish the message by printing it or, as is more likely, by forwarding it instantly to a different discussion forum. And if the message is sufficiently provocative, it may be republished again and again. The extraordinary capacity of the Internet to replicate almost endlessly any defamatory message lends credence to the notion that “the truth rarely catches up with a lie”. The problem for libel law, then, is how to protect reputation without squelching the potential of the Internet as a medium of public discourse.
[Blair J.A.’s emphasis deleted]
New activities on the Internet and the greater potential for anonymity amplify even further the ease with which a reputation can be harmed online:
The rapid expansion of the Internet coupled with the surging popularity of social networking services like Facebook and Twitter has created a situation where everyone is a potential publisher, including those unfamiliar with defamation law. A reputation can be destroyed in the click of a mouse, an anonymous email or an ill-timed Tweet.
(Bryan G. Baynham, Q.C., and Daniel J. Reid, “The Modern-Day Soapbox: Defamation in the Age of the Internet”, in Defamation Law: Materials prepared for the Continuing Legal Education seminar, Defamation Law 2010 (2010), at p. 3.1.1)
But I am not persuaded that exposing mere hyperlinks to the traditional publication rule ultimately protects reputation. A publication is defamatory if it both refers to the plaintiff and conveys a defamatory meaning: Grant, at para. 28. These inquiries depend, respectively, on whether the words used or “the circumstances attending the publication are such as would lead reasonable persons to understand that it was the plaintiff to whom the defendant referred” (Brown, at para. 6.1), and whether the words would “ten[d] to lower a person in the estimation of right-thinking members of society” (Botiuk v Toronto Free Press Publications Ltd.,  3 S.C.R. 3, at para. 62). Defamatory meaning in the words may be discerned from “all the circumstances of the case, including any reasonable implications the words may bear, the context in which the words are used, the audience to whom they were published and the manner in which they were presented” (Botiuk, at para. 62, citing Brown (2nd ed. 1994), at p. 1-15). (See Brown, at paras. 5.2, 5.4(1)(a) and 6.1; Knupffer v London Express Newspaper, Ltd.,  A.C. 116 (H.L.); Butler v Southam Inc., 2001 NSCA 121, 197 N.S.R. (2d) 97; Bou Malhab v Diffusion Métromédia CMR inc., 2011 SCC 9,  1 S.C.R. 214, at paras. 63 and 112.)
Where a defendant uses a reference in a manner that in itself conveys defamatory meaning about the plaintiff, the plaintiff’s ability to vindicate his or her reputation depends on having access to a remedy against that defendant. In this way, individuals may attract liability for hyperlinking if the manner in which they have referred to content conveys defamatory meaning; not because they have created a reference, but because, understood in context, they have actually expressed something defamatory (Collins, at paras. 7.06 to 7.08 and 8.20 to 8.21). This might be found to occur, for example, where a person places a reference in a text that repeats defamatory content from a secondary source (Carter, at para. 12).
Preventing plaintiffs from suing those who have merely referred their readers to other sources that may contain defamatory content and not expressed defamatory meaning about the plaintiffs will not leave them unable to vindicate their reputations. As previously noted, when a hyperlinker creates a link, he or she gains no control over the content linked to. If a plaintiff wishes to prevent further publications of the defamatory content, his or her most effective remedy lies with the person who actually created and controls the content.
Making reference to the existence and/or location of content by hyperlink or otherwise, without more, is not publication of that content. Only when a hyperlinker presents content from the hyperlinked material in a way that actually repeats the defamatory content, should that content be considered to be “published” by the hyperlinker. Such an approach promotes expression and respects the realities of the Internet, while creating little or no limitations to a plaintiff’s ability to vindicate his or her reputation. While a mere reference to another source should not fall under the wide breadth of the traditional publication rule, the rule itself and the limits of the one writer/any act/one reader paradigm may deserve further scrutiny in the future.
I am aware that distinctions can be drawn between hyperlinks, such as the deep and shallow hyperlinks at issue in this case, and links that automatically display other content. The reality of the Internet means that we are dealing with the inherent and inexorable fluidity of evolving technologies. As a result, it strikes me as unwise in these reasons to attempt to anticipate, let alone comprehensively address, the legal implications of the varieties of links that are or may become available. Embedded or automatic links, for example, may well prove to be of consequence in future cases, but these differences were not argued in this case or addressed in the courts below, and therefore need not be addressed here.
Nothing on Mr. Newton’s page is itself alleged to be defamatory. The impugned conduct in this case is Mr. Newton’s insertion of hyperlinks on his webpage. Mr. Crookes’ argument is that by linking to webpages and websites containing allegedly defamatory content, Mr. Newton has published that defamatory content. Since in my view the use of a hyperlink cannot, by itself, amount to publication even if the hyperlink is followed and the defamatory content is accessed, Mr. Crookes’ action against Mr. Newton cannot succeed. Moreover, even if Mr. Crookes had alleged that Mr. Newton should be understood, in context, to have expressed defamatory meaning, I would agree with the trial judge and the majority of the Court of Appeal that the statements containing the impugned hyperlinks on Mr. Newton’s page could not be understood, even in context with the hyperlinked documents, to express any opinion – defamatory or otherwise – on Mr. Crookes or the hyperlinked content.
I would dismiss the appeal with costs.
McLachlin CJ & Fish J
 We have read the reasons of Deschamps J. and Abella J. While we agree in large part with the reasons of Abella J., we respectfully propose a different formulation of the test for when a hyperlink reference in a text constitutes publication of defamatory matter to which it links.
The question, in legal terms, is when inclusion of a hyperlink in a text constitutes publication of a defamation in the hyperlinked material. Abella J. states that “a hyperlink, by itself, should never be seen as ‘publication’ of the content to which it refers” (para. 14). As justification, she notes that the hyperlinker has no control over the content referred to; the hyperlinker is not the creator of the content, and the content of the page linked to may change at any time (paras. 26-27). A hyperlink, therefore, is a reference and references are by definition “content-neutral” (para. 30).
Abella J. concludes that “[o]nly when a hyperlinker presents content from the hyperlinked material in a way that actually repeats the defamatory content, should that content be considered to be ‘published’ by the hyperlinker” (para. 42). In our view, the combined text and hyperlink may amount to publication of defamatory material in the hyperlink in some circumstances. Publication of a defamatory statement via a hyperlink should be found if the text indicates adoption or endorsement of the content of the hyperlinked text. If the text communicates agreement with the content linked to, then the hyperlinker should be liable for the defamatory content. The defendant must adopt or endorse the defamatory words or material; a mere general reference to a web site is not enough. Thus, defendants linking approvingly to an innocent Web site that later becomes defamatory would not be liable.
Finding publication in adoption or endorsement of the defamatory material in a Web site is consistent with the general law of defamation. In Hill v Church of Scientology of Toronto,  2 S.C.R. 1130, at para. 176, this Court held:
If one person writes a libel, another repeats it, and a third approves what is written, they all have made the defamatory libel. Both the person who originally utters the defamatory statement, and the individual who expresses agreement with it, are liable for the injury.
In sum, in our view, a hyperlink should constitute publication if, read contextually, the text that includes the hyperlink constitutes adoption or endorsement of the specific content it links to.
It is true that the traditional publication rule does not require the publisher to approve of the material published; he or she must merely communicate that material to a third party. However, the proposed adoption or endorsement standard for references is conceptually different. A mere reference without any adoption or endorsement remains that – a content-neutral reference. Adoption or endorsement of the content accessible by a link in the text can be understood to actually incorporate the defamatory content into the text. Thus, the content of the text comes to include the defamatory content accessed via hyperlink. The hyperlink, combined with the surrounding words and context, ceases to be a mere reference and the content to which it refers becomes part of the published text itself.
We add a final comment, with an eye to future technological changes. Abella J., as noted, states that “a hyperlink, by itself, should never be seen as ‘publication’ of the content to which it refers” (para. 14). So long as it is necessary to click on a hyperlink to access its content, this may be correct. What, however, of features in which a hyperlink projects content on the page automatically, or in a separate frame, with little or no prompting from the reader? Would inclusion of such a hyperlink, by itself, amount to publication? Like the issue of embedded hyperlinks, this question is not before us and should not be taken to have been decided in this case. Like Abella J. (at para. 43), we would leave issues concerning hyperlinks of this sort to be dealt with if and when they arise.
We agree with Abella J. that this appeal should be dismissed with costs.
This appeal offers yet another opportunity for the Court to consider the proper balance in the common law of defamation between the protection of reputation and the promotion of freedom of expression. In WIC Radio Ltd. v Simpson, 2008 SCC 40,  2 S.C.R. 420, and Grant v Torstar Corp., 2009 SCC 61,  3 S.C.R. 640, the Court took incremental steps to improve this balance by recognizing and modernizing defences to liability for defamation. In the instant case, the Court must determine whether, notwithstanding these defences, the scope of conduct that may attract prima facie liability for defamation is itself too broad and in need of adjustment to further promote freedom of expression.
Proof of publication is necessary in order to establish liability for defamation. “Publication” has an established meaning in the law of defamation. It refers to the communication of defamatory information in such a way that it is “made known to a third party”: Gaskin v Retail Credit Co.,  S.C.R. 297, at p. 299. Professor Brown explains that “[i]t is a bilateral act by which the publisher makes available to a reader, listener or observer in a comprehensible form the defamatory information” (R. E. Brown, Brown on Defamation (2nd ed. (loose-leaf)), at para. 7.2). Thus, publication has two components:
an act that makes the defamatory information available to a third party in a comprehensible form, and
the receipt of the information by a third party in such a way that it is understood.
The question is whether the first component of publication needs to be reconsidered owing to the impact of new forms of communications media. In answering this question, it will of course be necessary to bear in mind the particular technological feature – the Internet hyperlink – at issue in the case at bar. At the same time, however, the answer must be adaptable to other modes of communication and to future technological change.
I have read the reasons of my colleague Abella J. It is her view that the concept of publication as understood at common law needs to be altered so as to exclude references, including hyperlinks, from its scope. No longer must a reference, like any other act, simply make the defamatory information available in a comprehensible form: “Only when a hyperlinker presents content from the hyperlinked material in a way that actually repeats the defamatory content, should that content be considered to be ‘published’ by the hyperlinker” (para. 42). With respect, I disagree with this approach. I have also read the reasons of the Chief Justice and Fish J. They purport to approve Abella J.’s approach but, in effect, find that there are circumstances in which hyperlinked information can lead to a finding of publication. In so doing, they depart from the bright-line rule proposed by Abella J. In order to give guidance, I would prefer to outline a rule that is consistent with the common law and the civil law of defamation and that will also accommodate future developments in Internet law.
To create a specifically Canadian exception for references, which has the effect of excluding hyperlinks from the scope of the publication rule, is in my view an inadequate solution to the novel issues raised by the Internet. On the one hand, this blanket exclusion exaggerates the difference between references and other acts of publication. On the other hand, it treats all references, from footnotes to hyperlinks, alike. In so doing, it disregards the fact that references vary greatly in how they make defamatory information available to third parties and, consequently, in the harm they can cause to people’s reputations.
A more nuanced approach to revising the publication rule, and one that can be applied effectively to new media, would be for the Court to hold that in Canadian law, a reference to defamatory content can satisfy the requirements of the first component of publication if it makes the defamatory information readily available to a third party in a comprehensible form. In addition, the Court should make it clear that not every act, but only deliberate acts, can lead to liability for defamation.
Freedom of expression must be reconciled with the “equally important” right to reputation: Hill v Church of Scientology of Toronto,  2 S.C.R. 1130, at para. 121. Unlike my colleague Abella J.’s approach, the one I propose gives due weight to freedom of expression yet at the same time is grounded in a recognition that even simple references to defamatory information can significantly harm a person’s reputation. It would enable individuals whose reputations are so harmed to seek relief from those who, through deliberate acts, make defamatory material readily available to third parties.
If the person who made the reference was unaware that the information referred to was defamatory, the defence of innocent dissemination may be available: Society of Composers, Authors and Music Publishers of Canada v Canadian Assn. of Internet Providers, 2004 SCC 45,  2 S.C.R. 427 (“SOCAN”), at para. 89. In addition, as the Court mentioned in WIC and Grant, defences such as fair comment and reasonable communication on matters of public interest are also available. With respect, a shortcoming in my colleague’s approach is that she fails to consider the law of defamation generally, and instead focuses narrowly on one aspect of the rules governing publication. The result is problematically one-sided, as individuals who suffer harm to their reputations are left with no recourse against those who perpetuate defamatory information.
A reference, devoid of context, has never amounted in law to publication of the information to which it directs the third party. To so hold would be to disregard the bilateral nature of publication. Publication is not complete until someone other than the person referred to receives and understands the defamatory information. Thus, “to shout aloud defamatory words on a desert moor where no one hears them, is not a publication” (Gambrill v Schooley, 93 Md. 48 (1901), at p. 60). In the context of the Internet, a simple reference, absent evidence that someone actually viewed and understood the defamatory information to which it directs third parties, is not publication of that content.
In the instant case, the respondent, Jon Newton, admits that he hyperlinked to information that the appellants, Wayne Crookes and West Coast Title Search Ltd., allege to be defamatory. In the circumstances, Mr. Newton acted as more than a mere conduit in making the hyperlinked information available. His action was deliberate. Inasmuch as one link made the allegedly defamatory material readily available, the requirements of the first component of publication are satisfied for that link. However, publication is not complete unless the plaintiff adduces evidence that satisfies, on a balance of probabilities, the requirements of the second component of publication: that a third party received and understood the information to which reference is made. I agree with the majority of the Court of Appeal that the evidence adduced with respect to the second component of publication is insufficient in this case. I would therefore dismiss the appeal with costs and add that Mr. Newton might also have been able to raise defences to liability for defamation.
Two broad types of hyperlinks can be found on Web pages. The first and most common is an ordinary link. It is always user-activated: the Internet user clicks on a link on a Web page and is transferred to another page. The second type, which is often automatic but can also be user-activated, is created by a process referred to as “framing”. Unlike with ordinary links, in the case of framing, the Internet user does not leave the original Web page: information from another Web page appears in a “frame” on the page already accessed by the user. Where the framing is automatic, the content of other pages appears simultaneously in a frame when the user accesses the primary page. Moreover, a hyperlink can be either a “shallow” link to a site’s home page or a “deep” link to a page located on that site or another site. Both types of hyperlinks, as well as the distinction between shallow and deep links, have previously been referred to by the Court (SOCAN; Dell Computer Corp. v Union des consommateurs, 2007 SCC 34,  2 S.C.R. 801, at para. 4; and see, generally, M. Collins, The Law of Defamation and the Internet (3rd ed. 2010), at para. 2.43). In addition to Web pages, a variety of other forms of Internet communications can contain hyperlinks, including e-mail messages and on-line fora. In these reasons, the terms “hyperlink” and “link” will, unless otherwise specified, be used to refer to user-activated hyperlinks.
On July 18, 2006, Mr. Newton published an article on his Web site. It was entitled “Free Speech in Canada” and contained the following passage [A.R., at p. 125]:
Under new developments, thanks to the lawsuit, I’ve just met Michael Pilling, who runs OpenPolitics.ca. Based in Toronto, he, too, is being sued for defamation. This time by politician Wayne Crookes.
We’ve decided to pool some of our resources to focus more attention on the appalling state of Canada’s ancient and decrepit defamation laws and tomorrow, p2pnet will run a post from Mike on his troubles. He and I will also be releasing a joint press statement in the very near future.
The underlined text in this passage constituted hyperlinks, which Mr. Newton admits he created. The evidence indicates that “OpenPolitics.ca” was a shallow link to the Web site in question, on which, Mr. Crookes alleges, several interlinked articles defaming him could be found. The evidence further indicates that “Wayne Crookes” was a deep link to an article on another Web site, www.USGovernetics.com, that Mr. Crookes also alleges to be defamatory. These impugned hyperlinks were among seven different links contained in Mr. Newton’s article and were both preceded and followed by other hyperlinks. All the hyperlinks and the information to which they referred the user could be accessed by the public on line without restrictions.
On August 18, 2006, Mr. Crookes wrote to Mr. Newton and demanded that the hyperlinks “OpenPolitics.ca” and “Wayne Crookes” be removed. He received no response. On October 31, 2006, Mr. Crookes’ lawyer wrote to Mr. Newton and demanded that the hyperlinks be removed. On November 9, 2006, Mr. Newton refused to do so. In response to interrogatories that were subsequently addressed to him, Mr. Newton offered the following explanation for his refusal to remove the hyperlinks: “I saw no need. It was merely a hyperlink” (A.R., at p. 186).
As of February 1, 2008, the article “Free Speech in Canada” had been accessed 1,788 times. There is no information in the record on how many people – if any – read the allegedly defamatory material, or on whether any person who may have viewed the material accessed it by clicking on the two hyperlinks or by other means. The evidence is also silent on the behaviour of Internet surfers with respect to hyperlinks, and on the jurisdiction in which any individuals who may have read the hyperlinked material reside.
Mr. Crookes and his company, West Coast Title Search Ltd., sued Mr. Newton for defamation in British Columbia. The parties brought applications for summary trial and judgment under Rule 18A of the British Columbia Supreme Court Rules, B.C. Reg. 221/90. Kelleher J. granted Mr. Newton’s application and dismissed the action, finding that there had been no publication (2008 BCSC 1424, 88 B.C.L.R. (4th) 395).
Kelleher J. rejected Mr. Crookes’ contention that publication by Mr. Newton of the allegedly defamatory material should be presumed from the fact that he had created the hyperlinks. In the trial judge’s view, the plaintiffs’ failure to adduce any evidence that people had actually clicked on the hyperlinks and read the information was fatal to their position (paras. 20 and 24). Kelleher J. also concluded that the circumstances of the case did not support a finding of publication, drawing an analogy between a hyperlink and a footnote or a reference to a Web site in printed material (para. 29). Although the hyperlinks provided “immediate access to material published on another website” (para. 30), they did not make Mr. Newton a publisher of what readers would find if they chose to click on the link. Kelleher J. endorsed the proposition of the British Columbia Court of Appeal in Carter v B.C. Federation of Foster Parents Assn., 2005 BCCA 398, 42 B.C.L.R. (4th) 1, that “reference to an article containing defamatory comment without repetition of the comment itself should not be found to be a republication of such defamatory [comment]” (para. 33). He cautioned, however, that this did not mean that a hyperlink can never lead to liability for defamation. If Mr. Newton had written “the truth about Wayne Crookes is found here” and the word “here” was a hyperlink to defamatory information, Kelleher J. might have concluded differently (para. 34).
A majority of the British Columbia Court of Appeal dismissed the appeal (2009 BCCA 392, 96 B.C.L.R. (4th) 315). The judges were unanimous in rejecting Mr. Crookes’ argument that the creation of a hyperlink leads to a presumption of publication of the information to which the link points. Prowse J.A., who spoke for the court on this issue, noted that the legislature had provided for a presumption of publication in the case of broadcasts and newspapers, but not of the Internet (paras. 32-33). She was not prepared to create such a presumption on the basis of the record before her, and she expressed the opinion that this was a matter that might more appropriately be determined by the legislature (para. 41).
Saunders J.A. (Bauman J.A. concurring) wrote the reasons of the majority on the issue of whether publication could be inferred from the circumstances of the case. In her view, it could not. Accepting the bilateral nature of publication, she found that the hyperlinks did not satisfy the requirements of the first component of the definition of that term. She relied on the same jurisprudence as Kelleher J. in support of the proposition that merely referring to defamatory information without repeating it does not constitute publication (para. 81, citing Carter; MacFadden v Anthony, 117 N.Y.S.2d 520 (Sup. Ct. 1952); Klein v Biben, 296 N.Y. 638 (1946)). However, the circumstances of a case can demonstrate that a particular hyperlink invited or encouraged a reader to view the hyperlinked site, or that the information on that site was endorsed (para. 84). The majority held that there were no such circumstances in the instant case and that the way Mr. Newton presented the hyperlinks to potential users was “most comparable to a footnote for a reader, or a card index in a library” (para. 89). On the second component of publication, Saunders J.A. expressed the opinion that the bare number of 1,788 hits on Mr. Newton’s article was insufficient to support an inference that at least one person other than Mr. Crookes had clicked on the hyperlinks and read the allegedly defamatory information (para. 92).
Prowse J.A., dissenting, agreed with the majority on the first component of publication, but held that the circumstances of the case supported an inference that readers had been actively encouraged by Mr. Newton to click on the hyperlinks. In other words, the hyperlinks were being presented in such a way that they operated as more than simple footnotes. Furthermore, on the second aspect of publication, Prowse J.A. was prepared to infer that at least one person in British Columbia had clicked on the hyperlinks and read the impugned information. As a result, she found that there had been publication and stated that she would have allowed the appeal.
II. Positions of the Parties
Mr. Crookes argues that there is a presumption of publication whenever there are facts from which it can reasonably be inferred that the allegedly defamatory information was brought to the knowledge of some third person. Applying this proposition to hyperlinks, he contends that given the “deliberateness, immediacy and facilitation of access [to information]” that characterizes such links, the hyperlinked information should be understood to have been incorporated into the Web page on which the links were embedded. As a result, “[i]t can be reasonably inferred that embedding a hyperlink in the primary article brings the contents of the hyperlinked material to the knowledge of a third person [who accesses that article]” (A.F., at paras. 61-62).
Building on these arguments, Mr. Crookes also submits that the circumstances of this case would support an inference of publication. He disputes the position of the majority of the Court of Appeal that Mr. Newton’s hyperlink was comparable to a footnote, arguing that “[i]n the present case the barrier to accessing the defamatory comment is attenuated. All that is required is a keystroke” (A.F., at para. 75). He endorses Prowse J.A.’s approach, supporting her finding that the context of Mr. Newton’s hyperlinks, together with the 1,788 hits on Mr. Newton’s article, was sufficient to support an inference that at least one of those who read the article had clicked on the hyperlinks and read the allegedly defamatory information.
As for Mr. Newton, he rejects the view that publication should be presumed or inferred in the case of hyperlinks. Furthermore, he supports the finding of Kelleher J. and of the majority of the Court of Appeal that the evidence of 1,788 hits on his article was insufficient to support a conclusion that readers of the article had actually clicked on the hyperlink and read the allegedly defamatory information (R.F., at para. 35).
More fundamentally, Mr. Newton’s position is that hyperlinking does not constitute publication: a user-activated hyperlink “involves no transmission, copying or presentation of the alleged defamation, no control over its content, and no control over whether a reader makes the choice to follow the hyperlink” (R.F., at para. 58). He disputes the view that the fact that a hyperlink facilitates access to the linked information should justify a finding of publication, since publication “only occurs when a message is both posted and read” (para. 57). He also submits that, since any user-activated hyperlink could to some degree be considered an “invitation” to click on the link, it would be wrong to find that there was publication on the basis of such an invitation. This submission, too, is based on the principle that publication is not complete until the allegedly defamatory information is read. Finally, the fact that the creation of a hyperlink was deliberate cannot in itself be the basis for a finding of publication, since the question is “whether the bilateral posting/reading test for publication is met, and the insertion of a hyperlink alone does not meet either part of the test” (para. 85).
III. Positions of the Interveners
The interveners in this appeal take divergent approaches to when, if ever, a hyperlink could be said to constitute publication of the information to which it refers. The approach of my colleague Abella J. most closely resembles that of the Canadian Civil Liberties Association, which argues that publication “should be reserved for those situations where an individual actually communicates the allegedly defamatory words” (Factum, at para. 52). The result my colleague reaches is also consistent with the position taken by a group of interveners composed of print, broadcast and Internet media organizations. They argue that hyperlinks should never be equated with publication, because that would threaten the ability of its members to compete and fulfil their mandates in the Internet age (Media Coalition Factum, at paras. 42-43).
Certain other interveners – the British Columbia Civil Liberties Association (“BCCLA”), the Samuelson-Glushko Canadian Internet Policy and Public Interest Clinic (“CIPPIC”), and NetCoalition – accept that hyperlinking may constitute publication in appropriate circumstances, but seek to restrict those circumstances to instances involving variants of knowledge or explicit endorsement or adoption of linked information by the hyperlinker (e.g., BCCLA Factum, at paras. 22 and 24; CIPPIC Factum, at para. 2; NetCoalition Factum, at para. 4). CIPPIC takes this argument a step further, submitting that a plaintiff must prove publication on “a convincing evidentiary basis” (Factum, at para. 4). Finally, NetCoalition submits that, where a hyperlinker has not explicitly endorsed defamatory information or where there is no nexus between the hyperlink and such an endorsement, a defence of innocent dissemination should be available (Factum, at para. 34).
When a plaintiff seeks to establish prima facie liability for defamation, the court must not only consider whether the impugned information can reasonably be said to be defamatory and whether it in fact refers to the plaintiff, but must also consider the question of publication to a third party, bearing in mind the bilateral nature of publication. To be published, defamatory words must be “communicated” (Grant, at para. 28). “Communication” means that a message is both sent in a comprehensible form, and received and understood. Publication does not occur until “the defamatory matter is brought by the defendant or his agent to the knowledge and understanding of some person other than the plaintiff” (McNichol v Grandy,  S.C.R. 696, at p. 704, per Duff J. (as he then was); Brown, at paras. 7.2 and 7.8; Gatley on Libel and Slander (11th ed. 2008), at p. 164).
To determine whether an act can form the basis for a finding of publication, it is first necessary to consider the nature of the act itself.
Acts that may form the basis for a finding of publication and which may ultimately attract liability for defamation have never been limited to acts of the author or creator of the defamatory information. Rather, “[a]ll those jointly responsible for the publication are liable. Liability extends to all those who take part in the publication of the defamatory material, including those who merely cause or procure it to be published” (Brown, at para. 7.4 (footnotes omitted); Gatley on Libel and Slander, at p. 169; Lambert v Thomson,  O.R. 341 (C.A.), at p. 344: “.... where the libel is published in a newspaper or book, everyone who takes part in publishing it or in procuring its publication is prima facie liable”). As this Court noted in Botiuk v Toronto Free Press Publications Ltd.,  3 S.C.R. 3, at para. 75, “the design or agreement of persons to participate in acts which are tortious” is sufficient for them to be found liable for defamation as joint concurrent tortfeasors even if “they did not realize they were committing a tort”.
Historically, it has been understood that the precise method employed to make information available is immaterial: “There are no limitations on the manner in which defamatory matter may be published” (Brown, at para. 7.3). In Day v Bream (1837), 2 M. & Rob. 54, 174 E.R. 212, for example, a person who merely delivered parcels was held prima facie liable for putting them into publication; and in R. v Clerk (1728), 1 Barn. K.B. 304, 94 E.R. 207, a case referred to by my colleague Abella J., a printer’s servant whose sole role was to “clap down” the printing press was found to be responsible for publication. However, the law has changed in the centuries that have elapsed since Day v Bream and R. v Clerk.
The courts have begun incrementally to impose limitations on the nature and types of actions that can attract liability for defamation at common law. To understand this evolution, it will be helpful to look broadly at cases from both the United States and other common law jurisdictions, such as England and Australia. I note that in considering the U.S. cases, we must be mindful of the impact of the First Amendment on the protection of expression in the United States, and of certain significant statutory limits on liability (see comments to this effect: Godfrey v Demon Internet Ltd.,  4 All E.R. 342 (Q.B.), at pp. 343-44; Dow Jones & Co. v Gutnick,  HCA 56, 210 C.L.R. 575, at para. 52).
There appears to be an emerging consensus among the courts and commentators that only deliberate acts can meet the first component of the bilateral conception of publication. According to Prof. Brown, “a person must knowingly be involved in the process of publishing the relevant words” (para. 7.4 (emphasis added)). In Stanley v Shaw, 2006 BCCA 467, 231 B.C.A.C. 186, pleading that the defendants “said and did nothing” (para. 7) was held to be insufficient to support a finding of publication, because no tortious act had been alleged in relation to their silence (see also Smith v Matsqui (Dist.) (1986), 4 B.C.L.R. (2d) 342 (S.C.), at p. 355; Wilson v Meyer, 126 P.3d 276 (Colo. Ct. App. 2005), at p. 281 (“[a] plaintiff cannot establish [publication] by showing that the defendant silently adopted a defamatory statement”); Pond v General Electric Co., 256 F.2d 824 (9th Cir. 1958), at p. 827 (“[s]ilence is not libel”); Brown, at para. 7.3). In Scott v Hull, 259 N.E.2d 160 (Ohio Ct. App. 1970), at p. 162, a U.S. court held that “liability to respond in damages for the publication of a libel must be predicated on a positive act, on something done by the person sought to be charged”. I agree with this view.
A deliberate act may occur in a variety of circumstances. In Byrne v Deane,  1 K.B. 818 (C.A.), the defendants, proprietors of a golf club, were found to have published the words contained on a piece of paper that was posted on premises over which they held complete control. The defendants admitted to having seen the paper, but denied having written it or put it there. Although the words were ultimately found not to be defamatory, Greene L.J., concurring on the issue of publication, concluded that there are circumstances in which, by refraining from removing or obliterating defamatory information, a person might in fact be publishing it (at p. 838):
The test it appears to me is this: having regard to all the facts of the case is the proper inference that by not removing the defamatory matter the defendant really made himself responsible for its continued presence in the place where it had been put?
(See also Hellar v Bianco, 244 P.2d 757 (Cal. Dist. Ct. App. 1952); Tacket v General Motors Corp., 836 F.2d 1042 (7th Cir. 1987); Urbanchich v Drummoyne Municipal Council (1991), Aust. Torts Rep. ¶81-127 (N.S.W.S.C.).)
Byrne and its progeny are consistent with the requirement that any finding of publication be grounded in a deliberate act. If a defendant was made aware (or had reason to be aware) of defamatory information over which he or she had sufficient control but decided to do nothing about it, this nonfeasance might amount to a deliberate act of approval, adoption, promotion, or ratification of the defamatory information (see, e.g., Frawley v State of New South Wales,  NSWSC 1379 (AustLII)). The inference is not automatic, but will depend on an assessment of the totality of the circumstances. In Underhill v Corser,  EWHC 1195 (BAILII) (Q.B.), the defendant, who was treasurer and a board member of a charity, had been aware that an editorial levelling accusations against the plaintiff would be published, but took no action and gave no further thought to the matter. Although Tugendhat J. found that the defendant could have prevented the publication of the editorial, which the plaintiff alleged to be defamatory, he distinguished Byrne, concluding that the defendant’s role as a board member was different from that of the proprietors in Byrne (para. 110).
This requirement of a deliberate act has already been applied in the context of the Internet. In Godfrey, the defendant Internet service provider (“ISP”) had received a “posting”, which it stored on its news server. The plaintiff had notified the ISP that the posting was defamatory and requested that it be removed, but the defendant had allowed it to remain on its servers until it automatically expired 10 days later. The court held that the ISP’s failure to act, once it had become aware of the defamatory information over which it had control, constituted an act of publication once an Internet subscriber had accessed the posting. In the circumstances of the case, the ISP’s failure to act amounted to a deliberate act of approval, adoption, promotion or ratification of the defamatory information.
In Bunt v Tilley,  EWHC 407,  3 All E.R. 336 (Q.B.), the defendant ISPs were found not to be publishers because, even though they provided services, their role in the publication process was a passive one. This aspect of the decision in Bunt is a welcome development and should be incorporated into the Canadian common law. Those whose services are used to facilitate the communication of defamatory information may be shielded from responsibility for publication if they played a “passive instrumental role” in that process (para. 23).
A case that illustrates the nature of activities engaged in by a person who plays the role of a mere conduit is Metropolitan International Schools Ltd. v Designtechnica Corpn.,  EWHC 1765,  1 W.L.R. 1743 (Q.B.). In that case, the question was whether Google Inc. could be found to be the publisher of the search result “snippets” that were generated automatically by its search engine in response to search terms entered by users. Eady J. found that there was no publication, because Google Inc. played a passive instrumental role in facilitating the appearance of the snippets on the users’ screens (paras. 50-51).
It should be plain that not every act that makes the defamatory information available to a third party in a comprehensible form might ultimately constitute publication. The plaintiff must show that the act is deliberate. This requires showing that the defendant played more than a passive instrumental role in making the information available.
The question that remains is whether, even with these emerging limits on the common law principle, further refinements are necessary. While I agree that improvements can be made, I do not share the view of my colleague Abella J. that the solution is to exclude references, including hyperlinks, from the scope of the publication rule. In my view, the proper approach is (1) to explicitly recognize the requirement of a deliberate act as part of the Canadian common law publication rule, and (2) to continue developing the rule incrementally in order to circumscribe the manner in which a deliberate act must make defamatory information available if it is to result in a finding of publication.
More specifically, only where the plaintiff can establish on a balance of probabilities that the defendant performed a deliberate act that made defamatory information readily available to a third party in a comprehensible form will the requirements of the first component of publication be satisfied. Of course, before the court will make a finding of publication, the plaintiff must also satisfy the requirements of the second component of publication on a balance of probabilities, namely, that the “defamatory matter [was] brought by the defendant or his agent to the knowledge and understanding of some person other than the plaintiff” (McNichol, at p. 704).
Whether defamatory information is readily available is a question of fact. A court asked to infer that information has been made readily available should consider all circumstances related to the ease with which a third party would be able to gain access to it. Defamatory information is readily available if, in the circumstances, it can be immediately accessed. In other words, there must be no meaningful barrier that would prevent a third party from receiving it. In practice, this will not effect a significant change to the common law, since, for the vast majority of acts that presently make information available, there is nothing to prevent the information in question from being received by third parties; they need only use their eyes to read or view it, or their ears to hear it. Of course, if those third parties cannot understand it, that is a separate concern, one that would on its own bar a finding of publication.
But the requirement that defamatory information be readily available does result in an incremental change to the common law in those cases where the act complained of refers third parties to defamatory information. It narrows the circumstances in which the requirements of the first component of publication will be satisfied. In such instances, the totality of circumstances will reveal whether a third party can have immediate access – without any meaningful barriers – to the information to which the reference directs him or her.
What should be clear from this is that not all forms of references are the same as regards the extent to which they facilitate access to the information in question. While my colleague’s statement that “[h]yperlinks are, in essence, references” (para. 27) is superficially correct, it is inaccurate to equate a hyperlink with, for example, a footnote in a book. A footnote that does not actually reproduce the information to which the reader is being referred does not make that information readily available. The reader has to locate and obtain the document the footnote refers to and then find the information within the document. In contrast, an automatic hyperlink requires no action whatsoever, while an embedded deep hyperlink requires only the tap of a finger to gain access to the information. The effort involved is even less than that of turning a page in a book. Although it is of course true that hyperlinks are a form of reference, the extent to which they facilitate access and their ubiquity on the Internet cannot be overlooked.
This Court has previously recognized the ease with which information can be accessed by means of a hyperlink. In Dell, a hyperlink to an arbitration clause was considered in relation to a provision of the Civil Code of Québec, S.Q. 1991, c. 64, designed to deal with contractual stipulations – known as external clauses – that are physically separate from the contract itself. Adopting a contextual approach, the Court held that the hyperlinked arbitration clause should not be considered physically separate from the main contract document (at para. 97):
.... it is difficult to accept that the need for a single command by the user [i.e. clicking on a single hyperlink] would be sufficient for a finding that the provision governing external clauses is applicable. Such an interpretation would be inconsistent with the reality of the Internet environment, where no real distinction is made between scrolling through a document and using a hyperlink. Analogously to paper documents, some Web documents contain several pages that can be accessed only by means of hyperlinks, whereas others can be viewed by scrolling down them on the computer’s screen. There is no reason to favour one configuration over the other.
The arbitration clause in Dell was found to be reasonably accessible, since clicking on one hyperlink took the consumer directly to a page containing the terms and conditions of the sale, including the arbitration clause. Thus, there was no material distinction between having access to the clause by hyperlink and actually having a paper copy of it (at para. 100):
.... the clause was no more difficult for the consumer to access than would have been the case had he or she been given a paper copy of the entire contract on which the terms and conditions of sale appeared on the back of the first page.
What can be inferred from the circumstances in Dell is that the hyperlinked arbitration clause was, in fact, readily available: a link to it appeared on every Web page the consumer accessed, and all the consumer had to do to view the clause was to click on the link once (para. 100).
Because the inquiry into availability is essentially factual, it would be neither prudent nor desirable to attempt to adopt a bright-line rule indicating the exact time when something becomes “readily” available. Such an approach could hinder the evolution of the common law. However, given the context of this appeal, a few specific observations about hyperlinks are in order. In determining whether hyperlinked information was readily available, a court should consider a number of factors, including whether the hyperlink was user-activated or automatic, whether it was a shallow or a deep link, and whether the linked information was available to the general public (as opposed to being restricted). This list of factors is by no means exhaustive. Any matter that has a bearing on the ease with which the referenced information could be accessed will be relevant to the inquiry.
The effect of this approach is not to create a “presumption of liability for all hyperlinkers”, as Abella J. suggests (para. 25). In order to satisfy the requirements of the first component of publication, the plaintiff must establish, on a balance of probabilities, that the hyperlinker performed a deliberate act that made defamatory information readily available to a third party in a comprehensible form. Moreover, the plaintiff will also have to satisfy the requirements of the second component of publication on a balance of probabilities, namely that a third party received and understood the defamatory information.
My colleague Abella J. states that “[r]eferencing on its own does not involve exerting control over the content” (para. 26 (emphasis in original)). Yet the concept of publication in the common law of defamation has never involved a rigid requirement of control. Instead, the inquiry has always been contextual: Did the defendant act knowingly and what were the consequences of his actions? (See Brown, at para. 7.3.) Although a formal distinction can of course be drawn between references and other acts of publication, this distinction evades the questions that are at the heart of the law of defamation. Where a person deliberately makes defamatory information readily available through the creation of a hyperlink, the very rationale for the tort of defamation comes into play.
Because of the fluidity that characterizes the Internet and of the variety in types of hyperlinks, the Court should be particularly reluctant to fashion a bright-line rule. The approach it adopts must ensure that the law is properly attuned to how hyperlinks function in practice and how they may evolve in the future. Merely excluding hyperlinks from the scope of the publication rule will hardly make it possible to adapt the law of defamation to technological change.
By exempting ISPs from liability for their passive conduct, U.S. legislators have, in the Communications Decency Act of 1996, 47 U.S.C. §230 (1996), relied on the same criterion of deliberateness that exists in the common law of defamation. Nowhere does the American legislation suggest that those who deliberately create hyperlinks to defamatory material should be protected from liability. The cases cited by my colleague (at para. 28 of her reasons) all deal with the immunity enjoyed by ISPs and Web site operators with respect to defamatory content posted on line (Zeran v America Online, Inc., 129 F.3d 327 (4th Cir. 1997); Barrett v Rosenthal, 146 P.3d 510 (Cal. 2006); Fair Housing Council of San Fernando Valley v Roommates.Com, LLC, 521 F.3d 1157 (9th Cir. 2008)).
In the English case of Islam Expo Ltd. v The Spectator (1828) Ltd.,  EWHC 2011 (BAILII) (Q.B.), the issue was whether the words complained of were capable of referring to the claimant. The allegedly defamatory information in that case was an on-line text containing four hyperlinks. In concluding that the words complained of were capable of referring to the claimant, the court regarded the hyperlinked information as if it was incorporated into the message containing the hyperlinks. At a conceptual level, I see no difference between looking at the hyperlinked information to identify the defamed person and looking at it to find the defamatory words. I agree with Tugendhat J.’s statement in Islam Expo that the principles are the same.
Finally, if hyperlinks are excluded from the traditional conception of publication, there is a risk that freedom of expression will be favoured over reputational interests despite this Court’s pronouncement that they are “equally important” (Hill, at para. 121). Although I agree with my colleague that the most effective remedy for someone who has been defamed on line is to sue the person who created the defamatory material (para. 41), it may not always be possible to do so in the context of the Internet. A hyperlink can make public what was originally intended for only a select audience. Moreover, anonymity is easy to achieve on line, and who created the defamatory material may not always be known. If no remedy exists against “mere” hyperlinkers, persons defamed on line may in many cases not be able to protect their reputations.
In sum, an approach that focuses on how a hyperlink makes defamatory information available offers a more contextual and more nuanced response to developments in communications media than merely excluding all hyperlinks from the scope of the publication rule.
Of course, proving that the hyperlinker deliberately made the defamatory content readily available will satisfy only the requirements of the first component of publication. To satisfy the requirements of the second component, the plaintiff must demonstrate that a third party received and understood the defamatory information (McNichol, at p. 704; Brown, at paras. 7.2 and 7.8; Gatley on Libel and Slander, at p. 164).
I agree with the Court of Appeal in the instant case that a legal presumption should not be created in this regard. The legislatures of certain provinces, including British Columbia, have created presumptions of publication that apply to particular communications media. Under the Libel and Slander Act, R.S.B.C. 1996, c. 263, there is a presumption of publication in respect of “broadcast[s]” but not in respect of hyperlinks (s. 2). Since the legislature has seen fit to create a specific presumption of publication, the Court should refrain from creating a new one.
Absent a presumption of publication, the requirements of the second component of publication can be satisfied either by adducing direct evidence or, and this is more likely, by asking the court to draw an inference. Inferring that defamatory information has been read and understood is not new. In Gaskin, the Court, quoting Gatley on Libel and Slander, observed (at p. 300):
It is not necessary for the plaintiff in every case to prove directly that the words complained of were brought to the actual knowledge of some third person. If he proves facts from which it can reasonably be inferred that the words were brought to the knowledge of some third person, he will establish a prima facie case.
In the context of a hyperlink, an inference that a third party clicked on the link and read and understood the linked information will depend on a variety of factors, some of which may overlap with those considered when determining whether the linked content was made “readily available”. Some, but not necessarily all, of the factors a court can consider are: whether the link was user-activated or automatic; whether it was a deep or a shallow link; whether the page contained more than one hyperlink and, if so, where the impugned link was located in relation to others; the context in which the link was presented to users; the number of hits on the page containing the hyperlink; the number of hits on the page containing the linked information (both before and after the page containing the link was posted); whether access to the Web sites in question was general or restricted; whether changes were made to the linked information and, if so, how they correlate with the number of hits on the page containing that information; and evidence concerning the behaviour of Internet users.
Abella J. indicates that merely referring to defamatory material does not constitute publication of it (para. 42). Insofar as this means that a simple reference should not attract liability for defamation absent evidence that satisfies the requirements of both components of publication on a balance of probabilities, I agree. But if it means that a simple reference cannot form the basis for a finding of publication without actually expressing something defamatory, I cannot agree.
As the Court held in Grant, at para. 28, a plaintiff in an action for defamation is required to prove three things to succeed:
that the impugned words were defamatory,
that the words referred to the plaintiff and
that they were published.
Once the plaintiff establishes prima facie liability for defamation, the onus shifts to the defendant to raise any available defences. There are some defences that are worth discussing briefly in the context of hyperlinking.
The Court has in the past alluded to the existence of an “innocent dissemination” defence: SOCAN. In that case, this defence was described as follows (at para. 89):
.... the defence of innocent dissemination [is] sometimes available to bookstores, libraries, news vendors, and the like who, generally speaking, have no actual knowledge of an alleged libel, are aware of no circumstances to put them on notice to suspect a libel, and committed no negligence in failing to find out about the libel ....
In my view, the innocent dissemination defence should be made available to hyperlinkers, provided that the criteria mentioned in SOCAN are satisfied. There are good reasons for so holding. Hyperlinkers have not created or developed the information to which they link. Moreover, although the creation of a hyperlink may often be a deliberate act that makes the creator’s role more than a passive one, the hyperlinker often has little control, if any, over the linked information. This last point is significant given the dynamic nature of the Internet. A hyperlink may originally be created to link to information that is not defamatory, but the linked information may subsequently be altered so as to become defamatory without the hyperlinker’s knowledge. In such circumstances, provided that the criteria mentioned in SOCAN are satisfied, the innocent dissemination defence should operate until the moment the hyperlinker becomes aware that he or she is now linking to allegedly defamatory information. If this defence were widely available, it should dissuade overeager litigants from having a chilling effect on hyperlinking on the Internet.
Even when the innocent dissemination defence is not available to hyperlinkers, other defences may apply, such as those of fair comment (WIC) and of responsible communication on matters of public interest (Grant). In Grant, the Court, quoting Jameel v Wall Street Journal Europe SPRL,  UKHL 44,  1 A.C. 359, at para. 54, made the latter defence “available to anyone who publishes material of public interest in any medium” (para. 96). Moreover, it defined the concept of “public interest” expansively (at para. 106):
Public interest is not confined to publications on government and political matters, as it is in Australia and New Zealand. Nor is it necessary that the plaintiff be a “public figure”, as in the American jurisprudence since Sullivan [New York Times Co. v Sullivan, 376 U.S. 254 (1964)]. Both qualifications cast the public interest too narrowly. The public has a genuine stake in knowing about many matters, ranging from science and the arts to the environment, religion and morality. The democratic interest in such wide-ranging public debate must be reflected in the jurisprudence.
These defences, when available, would enable the hyperlinker to justify his or her conduct and avoid liability.
It should be evident that the law of defamation, viewed as a whole, is the result of a sustained effort to maintain a balance between the protection of reputation and freedom of expression. An approach that limits the discussion to the fact that what is in issue is a reference and disregards the context, including the nature of the reference and the various aspects of the law of defamation, is too narrow. In my view, the question of publication must be approached from a wider perspective, bearing in mind the incremental adjustments made to the law of defamation in recent years (e.g., WIC and Grant). Moreover, any further adjustments to defamation law should also be made incrementally, not by way of a sweeping declaration that treats all forms of reference alike and rules out the possibility of a reference to defamatory material supporting a finding of publication.
Mr. Newton admits that he created the two hyperlinks which refer readers to the allegedly defamatory information.
Mr. Newton’s involvement in creating the hyperlinks cannot be described as a passive instrumental role equivalent to that of a mere conduit, such as an ISP or a telephone carrier. His acts were deliberate. Moreover, it is common ground that the allegedly defamatory information to which Mr. Newton linked was available to Internet users in a comprehensible form, and without restrictions. The question is whether the information was readily available.
The first link, “OpenPolitics.ca”, was a shallow link to the Open Politics Web site. The evidence indicates that to consult the allegedly defamatory articles on that Web site, Internet users who initially clicked on the shallow link would then have had to take further action (i.e. click on other links) to locate them.
The second link, “Wayne Crookes”, was a deep link to an article on www.USGovernetics.com that Mr. Crookes also alleges to be defamatory. Clicking on the link would take a user directly to the allegedly defamatory information.
Both of these hyperlinks appeared in an article entitled “Free Speech in Canada”, which contained five other hyperlinks. The article dealt with free speech and defamation law, and referred to lawsuits concerning these issues. There were 1,788 hits on it between July 18, 2006, when it was created, and February 1, 2008.
There is nothing in the record that indicates whether anyone other than Mr. Crookes clicked on any of the links. Nor is there any information, from any point in time, on the number of hits on the Web pages containing the allegedly defamatory statements. And no evidence has been presented regarding the behaviour of Internet users.
Having regard to the totality of the circumstances, I am not prepared to infer in this case that the link to the Open Politics Web site made the defamatory content readily available. Because this was a shallow link, the reader would have to take further action in order to find the defamatory material. Exactly what action would have been necessary is unclear, but the evidence indicates that the various articles were not placed on the site’s home page and that they had separate addresses. This constituted a meaningful barrier to the receipt, by a third party, of the linked information. Insofar as Mr. Crookes’ defamation action against Mr. Newton is premised on this hyperlink, it cannot succeed.
However, I am prepared to infer that the “Wayne Crookes” link to the allegedly defamatory article on www.USGovernetics.com did make the content of that article readily available. It was a deep link. All the reader had to do to gain access to the article was to click on the link. The effort required to do this is not significantly different from the effort that would be required to move up or down the same page to read the article. A single click does not constitute a barrier to the availability of the material. Thus, Mr. Crookes has satisfied the requirements of the first component of publication on a balance of probabilities where this link is concerned.
On the second component of publication, I agree with the Court of Appeal, which was unanimous on this point, that the courts should not create a presumption of publication with respect to hyperlinks. As I mentioned above, and as Prowse J.A. explained (at para. 41), the creation of a new presumption in this field is a matter that would be better left to the legislature.
The determinative question is therefore whether Mr. Crookes has proven facts from which it can reasonably be inferred that the allegedly defamatory information was brought to the knowledge of some third person. On this point, I agree with the majority of the Court of Appeal that the evidence in this case does not support such an inference. I come to this conclusion mindful of the Internet context, the nature of Mr. Newton’s article, the way the various links were presented, the facts that the “Wayne Crookes” link was the third of seven links contained in Mr. Newton’s article and that at least one shallow link appeared before it, and the number of hits on the article between July 18, 2006, and February 1, 2008. Numbers take on a different meaning in the context of the Internet. The number of hits on Mr. Newton’s article – 1,788 – in the 18-month period does not provide a sufficient basis for concluding that out of the seven hyperlinks, someone has both clicked on the only link that made the defamatory information readily available and read that information.
In Internet law, as it presently stands, a bare number of hits will tell the trier of fact very little. As the majority of the Court of Appeal noted (at para. 92):
In the context of internet life, we have no way to assess the volume of “hits” here compared to the norm, the usual behaviour of internet readers or “surfers”, or the jurisdiction in which they reside. The conclusion drawn by my colleague is, with respect, tantamount to a presumption that in the case of a website accessed to any significant extent, there has been communication of the offensive material .... There may be cases in which more is known supporting such an inference, but such is not the case here where all that is before us is the bald number of hits. In my view there is an insufficient basis upon which to make such an inference ....
Because Mr. Crookes has not established facts supporting the second component of publication, he cannot succeed in his defamation action with respect to either of the impugned hyperlinks.
As a final comment, I would point out that, even if Mr. Newton had been found prima facie to be liable for defamation, he could likely, having regard to the record before the Court, have raised one of the existing defences, given that his article concerned matters that were arguably of public interest. I conclude that the hyperlinks at issue in this case did not result in publication of the information to which they referred the reader, and I would therefore dismiss the appeal with costs.
Communications Decency Act of 1996, 47 U.S.C. §230 (1996)
Libel and Slander Act, R.S.B.C. 1996, c. 263: s.2
Supreme Court Rules, B.C. Reg. 221/90: r.18A
Authors and other references
Balkin, Jack M. “The Future of Free Expression in a Digital Age” (2009), 36 Pepp. L. Rev. 427.
Baynham, Bryan G., and Daniel J. Reid. “The Modern-Day Soapbox: Defamation in the Age of the Internet”, in Defamation Law: Materials prepared for the Continuing Legal Education seminar, Defamation Law 2010. Vancouver: Continuing Legal Education Society of British Columbia, 2010.
Boivin, Denis W. “Accommodating Freedom of Expression and Reputation in the Common Law of Defamation” (1996-1997), 22 Queen’s L.J. 229.
Brown, Raymond E. Brown on Defamation: Canada, United Kingdom, Australia, New Zealand, United States, vols. 1 and 2, 2nd ed. Toronto: Carswell, 1994 (loose-leaf updated 2011, release 3).
Collins, Matthew. The Law of Defamation and the Internet, 3rd ed. New York: Oxford University Press, 2010.
Dalal, Anjali. “Protecting Hyperlinks and Preserving First Amendment Values on the Internet” (2011), 13 U. Pa. J. Const. L. 1017.
Danay, Robert. “The Medium is not the Message: Reconciling Reputation and Free Expression in Cases of Internet Defamation” (2010), 56 McGill L.J. 1.
Gatley on Libel and Slander, 11th ed. by Patrick Milmo and W. V. H. Rogers. London: Sweet & Maxwell, 2008.
Iacobucci, Frank. “Recent Developments Concerning Freedom of Speech and Privacy in the Context of Global Communications Technology” (1999), 48 U.N.B.L.J. 189.
Klar, Lewis N. Tort Law, 4th ed. Toronto: Thomson Carswell, 2008.
Lidsky, Lyrissa Barnett. “Silencing John Doe: Defamation & Discourse in Cyberspace” (2000), 49 Duke L.J. 855.
Linden, Allen M., and Bruce Feldthusen. Canadian Tort Law, 8th ed. Markham, Ont.: LexisNexis Butterworths, 2006.
Lindsay, David. Liability for the Publication of Defamatory Material via the Internet, Research Paper No. 10. Melbourne: University of Melbourne, Centre for Media, Communications and Information Technology Law, 2000.
Osborne, Philip H. The Law of Torts, 4th ed. Toronto: Irwin Law, 2011.
Ross, June. “The Common Law of Defamation Fails to Enter the Age of the Charter” (1996), 35 Alta. L. Rev. 117.
Sableman, Mark. “Link Law Revisited: Internet Linking Law at Five Years” (2001), 16 Berkeley Tech. L.J. 1273.
Streeter, Jeremy. “The ‘Deception Exception’: A New Approach to Section 2(b) Values and Its Impact on Defamation Law” (2003), 61 U.T. Fac. L. Rev. 79.
Donald J. Jordan, Q.C., and Robert A. Kasting (Taylor Jordan Chafetz, Vancouver; Stewart, Aulinger & Company, Vancouver), for the appellants.
Daniel W. Burnett and Harvey S. Delaney (Owen Bird Law Corporation, Vancouver), for the respondent.
all rights reserved