Corps de l’article


It is rare for the tabloids to produce anything of interest to legal theorists. And yet it seems to be the fate of privacy law that many significant developments have been motivated by the prurient practices of celebrity gossip journalism. For example, in 2004 the House of Lords ruled that Naomi Campbell could recover damages for breach of confidentiality on account that her privacy had been violated by a British tabloid for detailing her alleged drug treatment, accompanied by a photograph of her purportedly leaving a Narcotics Anonymous meeting. The headline read: “Naomi: I am a drug addict.”[1] In the same year the New Zealand Court of Appeal ruled that it was willing to recognize an independent tort of publication of private facts, even though it would not extend its protection to the children of celebrity couple Mr. and Mrs. Hosking who were photographed while being pushed in a stroller by their mother.[2]

These cases recall Samuel Warren and Lewis Brandeis’s concern regarding the practices of yellow journalism and their claim, in 1890, that

[i]nstantaneous photographs and newspaper enterprise have invaded the sacred precincts of private and domestic life; and numerous mechanical devices threaten to make good the prediction that “what is whispered in the closet shall be proclaimed from the house-tops.”[3]

The remedy, they argued, was for judges to recognize a common law right to privacy, a right that American courts have subsequently protected through the law of torts.[4] This paper argues, however, that even if tabloid journalism has inadvertently forged a consensus regarding the need to respect privacy, it is still not clear whether, in what manner, and to what extent private law should be enlisted in this endeavour. One of the difficulties regarding private law protection for privacy is that this emerging consensus regarding the need for legal protection is nonetheless fragile and masks deep tensions. As I argue in this paper, while there is a clear “privacy impulse” emerging across several common law jurisdictions, there is also a “containment anxiety” at play in the case law whereby judges seek to contain the legal protection for privacy. This containment anxiety takes different forms and is responsible for the tensions and differences now evident in private law protection for privacy.

There are different ways to understand this containment anxiety. The simplest explanation is that it is motivated by the definitional difficulties surrounding privacy; as has been well catalogued in the literature, there is little consensus regarding what privacy is and when it has been violated.[5] However, while the definitional dilemma might be a familiar starting point for privacy theorists, this paper adopts a different approach. I argue that the containment anxiety at issue is better explained in terms of a justificatory dilemma rather than a definitional one.

All legal rights raise the question of scope, and this question is intimately linked to issues of value and justification: what values underlie the right claimed, and how do these values connect to justifications for legal liability? Traditional liberal strategies for justifying private law rights focus on either autonomy or harm. Using Mill, I analyze what privacy understood as insulation from harm looks like and using Kant, I analyze what privacy understood as the protection of autonomy looks like. My argument is that if we understand privacy as protecting either the value of autonomy or freedom from harm, then we can justify only a fairly narrow legal right to privacy. Although this can provide a reasonably precise explanation for the specific ways in which courts have sought to limit the scope of privacy claims, it severely undermines the growing recognition of the importance of privacy. Therefore, instead of delineating a narrow privacy right, I argue that this focus on justification shows that privacy is in fact not well described at all in terms of either harm or autonomy, and that a legal right based on either of these justifications will be unduly narrow.

The alternative account that I put forward at the end of this paper is that we should view privacy as protecting one’s identity. Indeed, identity is a value upon which a number of privacy theorists are converging, although academic work on the relationship between privacy and identity is more prominent outside legal scholarship.[6] I invoke identity in order to argue that privacy protects one’s capacity for identity formation. This can provide both an account of the privacy interests at stake in cases like Campbell and Hosking, as well as a justification for a more robust privacy right than that offered by an understanding of privacy rooted in either autonomy or harm. Attending to questions of justification in this way can therefore also help to illuminate the nature of what I have been calling the emerging privacy impulse in a manner that can help develop a more coherent jurisprudence.

I. Unravelling the Tort Claim

Although their precise legal formulations vary, many of the major common law jurisdictions are beginning to converge on recognizing a private law right to privacy. For example, while courts in the United States have recognized a tort of invasion of privacy since 1904,[7] courts in New Zealand now also recognize this tort, at least as it applies to the publication of private facts,[8] and courts in the United Kingdom have held that the law of confidentiality protects privacy in similar circumstances.[9] Canadian courts have made numerous encouraging statements regarding common law protection for privacy, although most of these have arisen in very different contexts from the publication of private facts situation in both Campbell and Hosking.[10] Canadian commentators have also called for the general recognition of a tort of invasion of privacy.[11] Finally, a number of lower-court decisions in Australia have also endorsed some version of the tort.[12]

For the purposes of comparison, this paper focuses on tort liability associated with the publication of private facts as it has developed in the United States, the United Kingdom, and New Zealand. Because other jurisdictions do not have similar leading decisions regarding the publication of private facts, I do not include a discussion of their developing privacy jurisprudence.[13] I argue that this strong judicial privacy impulse—an intuition that privacy is a fundamental value worthy of legal protection—is being undercut by a judicial containment anxiety. This anxiety pertains to the consequences of providing the legal protection that seems called for by that privacy impulse and is leading to a number of important areas of disagreement and confusion in the case law. The following two sections outline in detail some of the inconsistencies and disagreements between jurisdictions as to the ambit of privacy rights in order to identify the key tensions and questions regarding private law protection for privacy. The first section reveals these tensions in the context of the question of whether to recognize a tort of invasion of privacy; the second section considers the substance of the action. Subsequent sections of this paper argue that the precise contours of this containment anxiety are actually quite instructive in relation to our understanding of privacy and can be best understood in terms of a justificatory dilemma in relation to a private law right to privacy.

A. Recognition of the Tort

This section outlines the differences between the United States, the United Kingdom, and New Zealand in relation to the issue of whether to recognize a tort of invasion of privacy. Courts in the United States have long recognized a tort of invasion of privacy. In a well-known 1960 article, William Prosser surveyed the jurisprudence that had emerged after Warren and Brandeis’s influential article, and classified the cases into four branches: publication of private facts, false-light advertising, misappropriation of name or likeness, and intrusions upon seclusion.[14] Prosser argued that these four branches in fact protected very different interests—including mental, reputational, and proprietary interests—and should be treated separately by the courts. His classification was subsequently adopted by the American Restatement of Torts and now forms the basis of American privacy jurisprudence.[15]

Prosser’s fracturing of the tort has also strongly influenced the development of the law of privacy in other jurisdictions. Outside the United States, most jurisdictions have been unwilling to recognize what Lord Hoffman, in Wainwright, called a “high-level principle” of invasion of privacy justifying a general tort of invasion of privacy.[16] Indeed, the difficulties that U.S. courts have faced in delineating the content of the tort are a central reason for the hesitation on the part of the House of Lords to follow the American example. For example, in Wainwright, Lord Hoffmann stated:

The need in the United States to break down the concept of “invasion of privacy” into a number of loosely-linked torts must cast doubt upon the value of any high-level generalization which can perform a useful function in enabling one to deduce the rule to be applied in a concrete case. ... There are a number of common law and statutory remedies of which it may be said that one at least of the underlying values they project is a right of privacy. ... But there are gaps; cases in which the courts have considered that an invasion of privacy deserves a remedy which the existing law does not offer. Sometimes the perceived gap can be filled by judicious development of an existing principle. The law of breach of confidence has in recent years undergone such a process.[17]

In the United Kingdom, therefore, privacy is protected by private law only insofar as courts can fashion discrete developments within existing legal categories.

This need to frame one’s claim within an existing cause of action has narrowed the scope of protection available to potential claimants. For example, in Wainwright the House of Lords denied Mrs. Wainwright a remedy for having been strip searched while visiting her son in prison, despite the court’s finding that the search was gratuitous and distressing. Nonetheless, one year later the House of Lords was willing to grant Naomi Campbell a remedy for the Daily Mirror article and photos. Because Ms. Campbell had made public statements that she was not a drug addict, the House of Lords unanimously held that the newspaper was fully entitled to “put the record straight” by reporting that she was a drug addict and was receiving treatment.[18] However, the House of Lords split on its assessment of the facts of the case. Three of the five Law Lords held that the newspaper violated Ms. Campbell’s privacy when it further reported

the fact that the treatment which she was receiving was provided by Narcotics Anonymous; ... details of the treatment—for how long, how frequently and at what times of day she had been receiving it, the nature of it and extent of her commitment to the process; and ... a visual portrayal by means of photographs of her when she was leaving the place where treatment had been taking place.[19]

The reason for the difference in treatment between Mrs. Wainwright and Ms. Campbell was that in Campbell the House of Lords could invoke the breach of confidence action to protect her privacy, instead of some “high-level principle.”[20] On its face this result seems anomalous in permitting serious violations of privacy while punishing much more minor ones, making the broad sweep of the American tort appear preferable despite its many difficulties.

The New Zealand Court of Appeal explicitly disagreed with the House of Lords’s adaptation of the law of breach of confidence in order to protect privacy:

Privacy and confidence are different concepts. To press every case calling for a remedy for unwarranted exposure of information about the private lives of individuals into a cause of action having as its foundation trust and confidence will be to confuse those concepts.[21]

They were particularly concerned about the implications of such an expansion on more traditional areas of the law of confidentiality; however, the court also thought that confidentiality provided too limited a protection for privacy. For example, Peckv. United Kingdom highlighted for the court the limitations of the law of confidence.[22] In that case, the European Court of Human Rights held that the law of confidentiality could not plausibly catch the broadcast of footage of an attempted suicide captured by a closed-circuit television camera in a public place, even though this broadcast was a violation of privacy. Because of these concerns, the New Zealand Court of Appeal rejected the House of Lords’s reliance upon confidentiality and instead recognized a tort of invasion of privacy—at least as it applied to the publication of private facts. This rejection raises a certain irony, given the subsequent U.K. decision in Murray v. Express Newspapers, where the Court of Appeal of England and Wales rejected the outcome in the New Zealand decision of Hosking and, purporting to follow Campbell, affirmed that the child of a famous parent could have a privacy right to restrain the publication of photos taken of him in a public place.[23]

Despite this difference in terminology—and the substantive concerns underlying this difference—the New Zealand position follows that of the United Kingdom in not recognizing a general tort of invasion of privacy as is found in American jurisprudence. Perhaps this fact, and the overlap in concerns detailed below, led the New Zealand court to indicate that it believed its approach to be “very close” to that of the House of Lords.[24]

Like the House of Lords decision in Campbell, Hosking was a split decision (three-to-two) with two strong dissents that reveal a number of further tensions surrounding the recognition of the tort of invasion of privacy. Justice Keith wrote in dissent that no cause of action for the publication of private facts should exist in the common law of New Zealand. For him, the reasons for this included the perceived lack of need, given the fact that there were already specific protections in place—such as the law of confidentiality—as well as a concern for the effect of such a tort on freedom of expression.[25] Justice Anderson agreed, holding that “this new liability, created in a side wind, is amorphous, unnecessary, a disproportionate response to rare, almost hypothetical circumstances and falls manifestly short of justifying its limitation on the right to freedom of expression.”[26]

Ironically, this dissenting inclination not to recognize any tort of invasion of privacy out of concern for freedom of expression is more in line with actual judicial practice in the United States where, as has been outlined, a much broader tort is recognized. A particularly striking example of privacy interests being overridden by freedom of expression is Florida Star v. B.J.F.[27] In that case, the Supreme Court of the United States overturned a lower-court decision imposing damages on a newspaper for publishing the name of a rape victim. Despite the fact that such publication was in violation of a Florida statute, as well as the newspaper’s policy, the court ruled that damages would violate the First Amendment. As Justice White held in dissent,

[a]t issue in this case is whether there is any information about people, which—though true—may not be published in the press. By holding that only “a state interest of the highest order” permits the State to penalize the publication of truthful information, and by holding that protecting a rape victim’s right to privacy is not among those state interests of the highest order, the Court accepts the appellant’s invitation ... to obliterate one of the most noteworthy legal inventions of the 20th century: the tort of the publication of private facts. ... Even if the Court’s opinion does not say as much today, such obliteration will follow inevitably from the Court’s conclusion here. If the First Amendment prohibits wholly private persons (such as B.J.F.) from recovering for the publication of the fact that she was raped, I doubt that there remain any “private facts” which persons may assume will not be published in the newspapers or broadcast on television.[28]

Justice White’s concern has been echoed by a number of commentators.[29] Therefore, the privacy tort in the United States covers a broader set of potential privacy invasions than publication of private facts, including intrusions upon seclusion, false-light advertising, and the appropriation of name or likeness. However, when the focus is on publication of private facts, American common law in fact provides protection to a far narrower class of publications than other jurisdictions due to the pre-eminence given to freedom of expression.

Judicial reluctance to recognize a tort of invasion of privacy therefore remains strong and can be summarized through reference to a number of interrelated concerns that show judicial anxiety regarding the ability to contain legal liability arising out of privacy violations. There are concerns that privacy is too vague a concept upon which to create a general tort, and that freedom of expression is important and should not be easily limited by such vague concerns. Many of these anxieties also influence the scope of the tort once it is recognized: if the common law develops to protect privacy, it should do so by expanding upon existing causes of action or at least by proceeding incrementally, rather than by establishing a broad and general principle and by recognizing a robust countervailing interest in freedom of expression that limits liability. As the following section outlines, this containment anxiety is also present in disputes regarding the substantive elements of the action.

B. Substance of the Action

Although the United States, the United Kingdom, and New Zealand all recognize some form of common law protection for the publication of private acts, there are important substantive differences between their approaches. This section outlines these differences as well as how they may be cast as variations of a judicial containment anxiety with respect to privacy. The first variation of this anxiety—found in both approaches of the United States and of New Zealand—involves limiting the scope of privacy through reference to a threshold of “offensiveness”.[30] The second variation—found in the United Kingdom but with echoes of it in all jurisdictions—is effectively to deny any difference between “private” and “confidential” information.

As already mentioned, American courts recognize four different branches to the tort of invasion of privacy: publication of private facts, intrusions upon seclusion, false-light advertising, and the appropriation of name or likeness. In contrast, the New Zealand Court of Appeal has restricted its recognition of the tort to the publication of private facts. While American courts have signalled an unwillingness to let privacy concerns outweigh freedom of expression, the New Zealand Court of Appeal has indicated that both values are worthy of protection and that one should not be subsumed by the other.[31] Nonetheless, apart from these differences, the New Zealand tort of publication of private facts is quite similar to the “publicity given to private life” branch of the American tort of invasion of privacy.

The central point of agreement is that the publication of private facts is not enough to establish liability; this publicity must also be “highly offensive”.[32] According to Justices Gault and Blanchard, there are two requirements for a successful claim of interference with privacy in New Zealand:

  1. The existence of facts in respect of which there is a reasonable expectation of privacy; and

  2. Publicity given to those private facts that would be considered highly offensive to an objective reasonable person.[33]

This is very similar to the American position, which provides that:

One who gives publicity to a matter concerning the private life of another is subject to liability to the other for invasion of his privacy, if the matter publicized is of a kind that

  1. would be highly offensive to a reasonable person, and

  2. is not of legitimate concern to the public.[34]

The New Zealand Court of Appeal did not make absence of legitimate concern an element of the tort, but rather held that a legitimate public concern in the information was a defence to the tort.

The court in Hosking spoke explicitly of its reasons for requiring offensiveness as a means to contain liability. Justices Gault and Blanchard argued,

In theory, a rights-based cause of action would be made out by proof of breach of the right irrespective of the seriousness of the breach. However, it is quite unrealistic to contemplate legal liability for all publications of all private information. It would be absurd, for example, to consider actionable merely informing a neighbour that one’s spouse has a cold. By living in communities individuals necessarily give up seclusion and expectations of complete privacy. The concern of the law, so far as we are presently concerned, is with widespread publicity of very personal and private matters. Publication in the technical sense, for example as applies in defamation, is not in issue.

Similarly publicity, even extensive publicity, of matters which, although private, are not really sensitive should not give rise to legal liability. The concern is with publicity that is truly humiliating and distressful or otherwise harmful to the individual concerned. The right of action, therefore, should be only in respect of publicity determined objectively, by reference to its extent and nature, to be offensive by causing real hurt or harm. ...

We consider that the test of highly offensive to the reasonable person is appropriate. It relates, of course, to the publicity and is not part of the test of whether the information is private.

We do not see personal injury or economic loss as necessary elements of the action. The harm to be protected against is in the nature of humiliation and distress.[35]

In this way, the court endeavoured to ask first, whether the information at issue was private, and second, whether it met a certain kind of threshold of seriousness in order to engage legal protection. This second, threshold question is separate from the issue of balancing privacy against other values such as freedom of expression, since the court also recognized a defence of legitimate public concern, which can take into account freedom of expression.[36] Rather, the threshold of seriousness is meant to identify the harm involved, which is understood in terms of “humiliation and distress.”[37]

Apart from requiring a threshold of seriousness, both the American and New Zealand versions of the tort require a prior determination that the information published be “private”. And this leads to a second strategy of containment, although it is usually not explicitly understood as such. In determining whether the published facts are indeed “private”, courts often revert to ideas of confidentiality. For example, in denying a reasonable expectation of privacy in the photographs of the Hosking children, the New Zealand Court of Appeal ruled,

The photographs ... do not disclose anything more than could have been observed by any member of the public in Newmarket on that particular day. They do not show where the children live, or disclose any information that might be useful to someone with ill intent. The existence of the twins, their age and the fact that their parents are separated are already matters of public record.[38]

American courts would have likely come to the same conclusion.[39] However, this idea that there can be no privacy interest in information that has already in some sense been revealed in public comes directly out of a paradigm of confidentiality.[40] As will become clearer in the discussion of confidentiality in connection with the Campbell decision below, information that is publicly available generally fails the test for confidentiality. However, even if this is quite a familiar conclusion regarding confidentiality, privacy-focused analysis can lead to a different result. Indeed, many commentators have indicated that there can be a privacy interest in “public” information, and their view is also supported by some judicial opinion.[41] It is thus not clear why the court in Hosking felt the need to deny a reasonable expectation of privacy in public, when it could have invoked its test of offensiveness to achieve the same result.

The relationship between privacy and confidentiality is more explicitly addressed in Campbell, although many issues remain unresolved. Despite a three-to-two split, the House of Lords in Campbell understood itself to agree that the recognition of breach of confidentiality is a cause of action capable of protecting privacy, and to split instead on its application to the facts of the case. However, there was also significant disagreement regarding whether the court was simply applying the traditional breach of confidentiality action or creating a new, extended version of the action. Implicit in this disagreement are differing views regarding the relationships between “confidential” and “private” information. A brief overview of the law of confidentiality is necessary to distinguish these concepts.

Breach of confidentiality was originally understood as an equitable doctrine.[42] Until Saltman Engineering Coy. Ld. v. Campbell Engineering Coy., Ld.,[43]

the main cases on [breach of] confidentiality [were seen to be] significant for their lack of any uniform jurisprudential basis. The confidences were mostly commercial and arose generally in the context of contracts, such as contracts of partnership, employment, agency or sale.[44]

Saltman is significant in that it affirmed that there could be an obligation of confidence without a contractual relationship. Saltman also defined when the information at issue had the necessary quality of confidence about it. Lord Greene M.R. stated, “The information, to be confidential, must, I apprehend, apart from contract, have the necessary quality of confidence about it, namely, it must not be something which is public property and public knowledge.”[45] Confidential information therefore refers to the fact that the information is kept secret, rather than any quality intrinsic to the information itself. Notably, public information is by definition nonconfidential.

The classic statement of breach of confidence, now understood to stand independently of contract, is found in Coco v. A.N. Clark (Engineers) Ltd., where Justice Megarry, as he then was, identified three elements for a successful breach of confidence claim:

First, the information ... must “have the necessary quality of confidence about it.” Secondly, that information must have been imparted in circumstances importing an obligation of confidence. Thirdly, there must be an unauthorized use or disclosure of that information to the detriment of the party communicating it.[46]

The second element has generated significant interest, alongside inquiry into the circumstances that import an obligation of confidence. More specifically, can there be an obligation where the parties are not in a relationship of confidence?

It was the perceived need for a relationship of confidence that led Warren and Brandeis to reject breach of confidentiality as a means of protecting privacy, and to call for the recognition of a more general tort. As they argued, such a doctrine

may have satisfied the demands of society at a time when the abuse to be guarded against could rarely have arisen without violating a contract or a special confidence; but now that modern devices afford abundant opportunities for the perpetration of such wrongs without any participation by the injured party, the protection granted by the law must be placed upon a broader foundation.[47]

When the only way to have a photograph taken was to pose in a studio, the law of contract or confidentiality could control its dissemination. However, with modern photographic technology this is no longer the case: the photographer can take a subject’s picture without there being any relationship of confidentiality involved.

However, the law of confidentiality did evolve to catch cases where there is no relationship of confidence between parties.[48] The first development along these lines was to find an obligation where a third party discloses information he has received from someone who he knows is under a duty of confidence.[49] The second development can be traced to Lord Goff’s holding in Guardian that a duty of confidence can arise independently of a relationship of confidence “where an obviously confidential document is wafted by an electric fan out of a window into a crowded street, or where an obviously confidential document, such as a private diary, is dropped in a public place and is then picked up by a passer-by.”[50] The salient question is whether this evolution properly catches the kinds of privacy cases that Warren and Brandeis were concerned about when they rejected breach of confidentiality. And this depends upon whether disclosure of an obviously confidential document catches the same cases as does the publication of private facts.

The argument that it does not is an argument that an action for breach of confidentiality, even when not explicitly requiring the breach of a confidential relationship, is still ultimately concerned with protecting confidential relationships. The reasons for protecting confidences are numerous. As several of the judgments in the House of Lords pointed out, the details of Ms. Campbell’s treatment needed to be protected in order to protect the therapeutic relationship. Lord Hope stated,

The therapy is at risk of being damaged if the duty of confidence which the participants owe to each other is breached by making details of the therapy, such as where, when and how often it is being undertaken, public.[51]

In other words, even if the press learned of Ms. Campbell’s therapy without there being a violation of the confidential relationship between the participants in that therapy, publication of this information can nonetheless damage that relationship.[52]

But there may be information that is “private” where its publication does not threaten any kind of confidence or relationship of trust.[53] If the House of Lords is protecting “private” information rather than “confidential” information, then it is shifting the breach of confidence action away from where Lord Goff placed it with his willingness to impose a duty of confidence regarding obviously confidential documents. The clearest statements of such a shift are found in the judgments of Lords Nicholls and Hoffman. Lord Nicholls argued that

[t]he continuing use of the phrase “duty of confidence” and the description of the information as “confidential” is not altogether comfortable. Information about an individual’s private life would not, in ordinary usage, be called “confidential”. The more natural description today is that such information is private. The essence of the tort is better encapsulated now as misuse of private information.[54]

Lord Hoffmann also felt that the law of confidentiality had shifted in a fundamental manner. In addition to no longer requiring the violation of a confidential relationship, there had been

the acceptance, under the influence of human rights instruments such as article 8 of the [European Convention for the Protection of Human Rights and Fundamental Freedoms], of the privacy of personal information as something worthy of protection in its own right.[55]


[w]hat human rights law has done is to identify private information as something worth protecting as an aspect of human autonomy and dignity. ...

The result of these developments has been a shift in the centre of gravity of the action for breach of confidence when it is used as a remedy for the unjustified publication of personal information. It recognises that the incremental changes to which I have referred do not merely extend the duties arising traditionally from a relationship of trust and confidence to a wider range of people. ... Instead of the cause of action being based upon the duty of good faith applicable to confidential personal information and trade secrets alike, it focuses upon the protection of human autonomy and dignity—the right to control the dissemination of information about one’s private life and the right to the esteem and respect of other people.[56]

The shift in breach of confidentiality, although stated differently in two judgments, concerns a shift in the nature and status of the protected information. The traditional breach of confidence action concerns confidential information, which is information that is kept secret rather than public, and is protected because of the importance of protecting confidences and relationships of trust. For Lord Nicholls, the need to protect confidences no longer adequately describes the information at issue. Lord Hoffman goes further, holding that the protection of “private” information is necessary for human autonomy and dignity.

As other Law Lords use confidentiality and privacy interchangeably, they do not provide a clear view of the shift toward the protection of private information. Still, others such as Lord Hope actively resist it:

The questions that I have just described seem to me to be essentially questions of fact and degree and not to raise any new issues of principle. ...

The language has changed following the coming into operation of the Human Rights Act 1998 ... I doubt whether the result is that the centre of gravity ... has shifted.[57]

On this view, the boundaries of the protection of privacy are clearly marked by the principles of confidentiality, no less when the language of privacy is adopted.

Because the facts of Campbell fit within both the traditional breach of confidence action and its purported extension into an action for misuse of private information, the disputed relationship between privacy and confidentiality was left unresolved. But it raises a number of important questions. For example, the breach of confidence model suggests that all that is required is an unauthorized use or disclosure of the confidential or private information—not widespread publicity, as required by both the New Zealand and American versions of the tort of invasion of privacy. Similarly, the breach of confidence model suggests that this misuse of private information results in “detriment” but the New Zealand and American versions of the tort instead require that the publication be highly offensive. Indeed, the idea that private information is protected only from publicity that is highly offensive was rejected in Campbell. At most, the U.K. position is that the test for publicity that is highly offensive is relevant only to help identify information as private when it is not easily identified as such.[58] It is not an additional requirement that narrows tort protection to private information, the publication of which is highly offensive.

One can see that the scope of the action is considerably narrowed by maintaining an emphasis on the protection of confidential information. If this does indeed shift to a broader category of information, then the scope of the action will likewise be broadened. Liability would be imposed for unauthorized disclosure of private information to an individual’s detriment—subject to other considerations such as freedom of expression—unless another variation of containment anxiety modified this extended action even further.

Subsequent case law in the United Kingdom has done little to clarify the relationship between confidentiality and privacy.[59]McKennitt concerned the publication of details of singer Loreena McKennitt’s life by a former friend and work associate. Mosley concerned the publication of allegations that the claimant, FIA (Fédération Internationale d’Automobile) President Max Mosley, had participated in a “Nazi sex orgy”. Although these cases refer to Campbell and the newly extended form of the breach of confidence action, both the High Court of England and Wales and the House of Lords nonetheless viewed the matter before them as an “old-fashioned breach of confidence.”[60] At the same time, there are some indications that the emphasis of the extended action is indeed shifting toward a more expansive notion of privacy. The most notable example in this regard is Murray, where the Court of Appeal of England and Wales held that the child of a famous parent had a “reasonable expectation that he or she will not be targeted in order to obtain photographs in a public place for publication,” where the photographer knew that it was likely that consent would not be given.[61] As this case dealt with whether the claim should be struck because there was no reasonable expectation of privacy at issue, it remains unclear how this decision will affect the jurisprudence regarding the subsequent question of how privacy is to be balanced against competing interests.[62] Moreover, this case and others are increasingly influenced by jurisprudence regarding article 8 of the European Convention on Human Rights,[63] which will potentially shift the focus further away from traditional ideas of confidentiality and import a very different idea of privacy.[64]

In sum, what I have been calling the privacy impulse is responsible for the recognition of some form of the tort of invasion of privacy in jurisdictions such as the United Kingdom and New Zealand, at least insofar as it applies to the context of the publication of private facts. In the United Kingdom, it is also responsible, in the breach of confidentiality action, for a shift away from a concern for confidential information and toward the protection of private information, although this shift is far from settled law. Nonetheless, this privacy impulse is often curtailed by what I have called containment anxiety. This takes several forms in the case law. In the United Kingdom, it pushes some judges to resist the shift from confidential information to the potentially broader category of private information, understood to protect interests other than confidences. Indeed, even in jurisdictions that explicitly recognize privacy as the protected interest, judges often implicitly appeal to confidentiality norms and thus narrow the scope of protection for privacy. In the United States and New Zealand, the containment anxiety also pushes the courts to require that the publication of private facts be highly offensive. In some jurisdictions, such as the United States, it motivates courts to privilege other values, such as freedom of expression, over privacy.

II. The Justificatory Dilemma

One response to the variations of containment anxiety outlined above is to see them simply as part of the general drama of tort law: actions are recognized and courts struggle to contain the implications of liability within reasonable limits. Yet this is insufficient, as legal theory is also concerned with this task, which can be rephrased in theoretical terms as the question of justification. That is, the question of the nature and scope of tort liability can be answered by reference to the underlying justification for the imposition of legal liability in the private law context. Liberal justifications for such liability have traditionally adopted one of two strategies: a focus on harm, or a focus on autonomy. In other words, the state is justified in assisting an individual who has been harmed or coerced by another. Therefore, if privacy is the kind of interest that merits legal protection—rather than simply moral censure—then the situations in which we hold there to be a violation must correspond to some cognizable harm or interference with an individual’s autonomy. Moreover, this harm or coercion can help to determine the merited scope of the protection.

I argue in this section that if we do this analysis—if we seek to understand the harm or coercion at stake in cases like Campbell and Hosking—a surprising conclusion emerges: it is in fact extremely difficult to cast the “wrong” at issue in these cases in terms of harm or violation of autonomy. Moreover, the precise ways in which it is difficult can account for many of the specific features of the containment anxiety outlined previously, such as the tendency to revert to confidentiality when determining the scope of the privacy tort, the seeking of a threshold of “offensiveness”, and the seeming engulfment of the privacy tort by freedom of expression in the United States.

The justificatory dilemma thus does not shed light on general dilemmas in tort law, but rather illuminates the particular nature of privacy. It does so in two ways. First, it is descriptive, helping to account for the state of the jurisprudence as we find it. Second, it is normative, helping us to understand the kind of value that privacy should be best understood to protect. On this latter point, the justificatory dilemma indicates that privacy is not well understood in terms of either protecting individuals from harm, or promoting autonomy.

It is important to distinguish the justificatory dilemma from what might be termed a “definitional dilemma”. One could argue that the best explanation for the containment anxiety found in the cases is not a justificatory dilemma, but rather a definitional dilemma. The literature on privacy is replete with the observation that privacy is notoriously difficult to define.[65] Indeed, this has become an axiomatic starting point of many privacy discussions. However, the definitional dilemma suggests a different structure of analysis from what I am proposing here. If the focus of analysis is on the definition of privacy, one must first define privacy, then use this definition to identify a violation of privacy in a particular context, and finally ask whether such a violation should be protected by law.[66] In practice, these latter two stages of analysis are often conflated and the determination of a violation of privacy is thought to justify a legal remedy. In contrast, if the focus of analysis is on the justification for a legal right to privacy, one must first ask what kind of behaviour is thought to constitute a legal wrong. Then one must ask whether this behaviour is captured by theories of legal wrongs such as harm or coercion—that is, does it crystallize into either harm or coercion of a nature that may justify a legal right? In this way, the definitional quagmire is bypassed by asking the question of legal justification directly in relation to the behaviour at issue, and not in relation to a particular definition of privacy.

Since the justificatory approach takes as its point of departure a particular context of behaviour rather than a particular definition of privacy, the following sections do not interrogate a “high-level principle” of privacy.[67] Nor do they focus primarily upon deciphering the meaning of “publication of private facts” because of the risk that a focus on what constitutes “private facts” will return us to the definitional dilemma that I want to sidestep. Instead, the following sections focus on the type of behaviour at issue in cases like Campbell and Hosking. In terms that do not prejudge the privacy question, one could say that the publication of photos in tabloids exposes the subject to increased curiosity and spreads information about him or her similar to other practices of gossip. The question at issue is therefore not “What is privacy?” but rather “When does the gossip and curiosity of others constitute a wrong of the type that justifies legal sanctions?”

There are two points to note in relation to this latter question. The first is that the practices of gossip and curiosity at stake in the celebrity publicity cases already discussed are not so different from the original concerns that motivated Warren and Brandeis’s argument that the law should recognize and protect a right to privacy. As previously discussed, they too were concerned with the practice of gossip and its transformation by advances in photography and newspaper publication.[68] Indeed, gossip and curiosity capture many of our contemporary concerns regarding the availability of personal information on the internet. John Perry Barlow, co-founder of the Electronic Frontier Foundation, argues that one implication of the internet is that everyone leads “as public a life as I do at home” in a small town where

[a]nyone in Pinedale who is interested in me or my doings can get most of the information he might seek in the Wrangler Café. Between them, any five customers could probably produce all that is known locally about me, including a [sic] quite a number of items which were well known but not true.[69]

Therefore a focus on the practices of gossip and curiosity captures important intuitions we have regarding privacy in some contexts, while remaining focused on the behaviour at issue and not a particular definition of privacy.

The second important point is that there is a persistent intuition, also reflected in the case law, that the technological augmentation or transformation of the social practices of gossip and curiosity have results that merit legal attention.[70] Therefore, to focus the justificatory question further, we must inquire why and when gossip and curiosity—and in particular their transformation through modern technology such as publication in mass media—justify the imposition of legal liability.

The following sections treat this question in detail and conclude as follows. First, a harm analysis justifies legal liability in the form of the American tort of publicity given to private facts, which imposes a threshold of harm (offensiveness) on the type of publicity that is actionable. Second, a coercion analysis justifies legal liability in the form of a breach of confidentiality action but resists its extension beyond traditional contours of confidentiality. Third, both accounts show that limiting freedom of expression through privacy rights can itself lead to either harm or coercion. I thus argue that the most important conclusion to draw from this discussion of justifications is that the wrong arising out of practices of gossip and curiosity is not appropriately described in terms of protection from harm or coercion. The alternative that I propose is that liability for such practices is better understood in terms of protecting the capacity for identity formation, upon which I elaborate in the last section of the paper.

A. The Harm Argument

A focus on harm can help to articulate some scenarios in which gossip can be deemed wrong. For example, gossip can clearly harm an individual’s reputation. However, harm to one’s reputation cannot ground a right to privacy. Such harm is already protected in private law through the law of defamation, and the law of defamation protects only against the dissemination of information that is false (under American law) or that cannot be proven true (in Commonwealth jurisdictions such as in the United Kingdom and Canada). For the purposes of defamation law, true but “private” information does not receive protection, for one is not entitled to the protection of a good reputation gained through the suppression of truth.

A focus on reputational harm also does not capture what some see as decisive with respect to a claim for privacy. For example, Warren and Brandeis argued that the legal right to privacy involves the right to determine when and how one’s thoughts, sentiments, and emotions will be communicated to others. Importantly, Warren and Brandeis thought that protection for the communication of one’s thoughts, sentiments, and emotions had a “spiritual” rather than material basis, which led them to reject any analogy between privacy and defamation.[71] The common law of defamation provided an inappropriate conceptual basis from which to derive a principle of privacy, they argued, because

[i]t deals only with damage to reputation, with the injury done to the individual in his external relations to the community, by lowering him in the estimation of his fellows. The matter published of him, however widely circulated, and however unsuited to publicity, must, in order to be actionable, have a direct tendency to injure him in his intercourse with others ... the effect of the publication upon his estimate of himself and upon his own feelings not forming an essential element in the cause of action.[72]

In calling for the protection of one’s thoughts, sentiments, and emotions from unauthorized publication, therefore, Warren and Brandeis sought to protect an individual’s sense of self and not his relationship with others in the community as protected by his reputation.

Apart from harm to reputation, there is another way to think about the harm involved in gossip and curiosity. There is some information about ourselves that, if known by others, would leave us vulnerable to harassment, discrimination, or other types of abuse. One could identify the types of information that generate such vulnerabilities and protect them from collection and dissemination through a right to privacy. This information would not necessarily be sensitive or intimate in a traditional sense, but rather would simply have to be linked to the problematic practices listed above. However, it is important to understand that in such cases privacy functions more like an anticipatory remedy than a description of the wrong.[73] By granting a privacy right over the information, we protect an individual against wrongs that are not themselves best described as invasions of privacy. For example, if an individual’s HIV status is considered private in order to provide some protection against discrimination, then the wrong at issue is the consequent discrimination and not the invasion of privacy. Nonetheless, by protecting against disclosure of the information, the discrimination can be prevented, providing a kind of remedy in anticipation of the harm.

Although this account of privacy as an anticipatory remedy can capture some important elements of harm that might arise through gossip and curiosity, it still does not capture the intuition that seems to be at play in many discussions of privacy where privacy, in insulating us from gossip and curiosity, also insulates us from forms of social pressure that do not crystallize into other, independent harms. For example, Ferdinand Schoeman argues that privacy norms “define the presumptive boundaries within which it is permissible to apply direct social pressure of accountability and of threatened social disgrace.”[74] The question then is how to define the circumstances under which such social pressure might constitute a harm sufficient to justify a legal right to privacy.

One reason for concern about such pressure is its inhibiting effect. For example, Gavison has argued that privacy promotes liberty by permitting “individuals to do what they would not do without it for fear of an unpleasant or hostile reaction from others.”[75] These unpleasant reactions need not themselves constitute independent wrongs such as harassment and discrimination; they may simply impede individual experimentation out of the fear of social sanction for unpopular ideas and actions. Instead, privacy in this sense protects our ability to act and think in unpopular ways. Privacy protects individuality understood in terms of our ability to be eccentric.[76]

Here, John Stuart Mill’s classic articulation of the harm principle within liberal theory can help outline why it is in fact difficult to ground a legal right to privacy on this approach, even if from a sociological perspective it seems to capture something important about the relationship between gossip and privacy.[77] Mill was acutely sensitive to the conditions of modern life and the rise of the social sphere. He focused on the social sphere as distinct from either the “public” understood as the state or the “social” understood as the interactions between discrete individuals. Mill argued that society understood collectively can practice “a social tyranny more formidable than many kinds of political oppression” and that there should be

a limit to the legitimate interference of collective opinion with individual independence; and to find that limit, and maintain it against encroachment, is as indispensable to a good condition of human affairs as protection against political despotism.[78]

Mill’s general strategy with respect to the tyranny of the social sphere is to draw a line between that which concerns only the individual—matters for which the individual is not accountable to society—and that which concerns one’s relations with others. An individual is accountable to society for conduct that concerns others, and in particular, conduct that harms others. Therefore society is not justified in judging another for conduct that concerns only the individual.[79] Based upon the foregoing, we could argue that conduct that concerns only the individual should be protected by a right to privacy and thus shielded from the gossip and curiosity of others.

However, Mill’s own analysis would not support such a conclusion.[80] For example, Mill suggests that individuals do not require insulation from the effects of being spoken about. In his discussion of the freedom of religious belief and opinion, Mill agrees that individuals should not be deprived of a livelihood because of their beliefs and the social stigma that ensues from their profession of opinions not shared by the majority.[81] Nonetheless, he draws a line between social coercion aimed at depriving an individual of his livelihood, and social censure that simply leads to an individual being ill-thought of, a result which “ought not to require a very heroic mold to enable them to bear.”[82] Put another way, being ill-thought of is not an unjustified interference with individual autonomy. This supports privacy understood in terms of an anticipatory remedy that can help protect individuals from harms like discrimination, but it does not support privacy understood in terms of insulation from social pressure that does not rise to this level of harm.

Indeed, Mill’s analysis allows for a wide scope of social censure. He argues that even with respect to that which concerns only the individual (which is his boundary demarcating individual liberty from social interference), society can express its disapprobation through “[a]dvice, instruction, persuasion, and avoidance by other people, if thought necessary by them for their own good.”[83] He elaborates:

In these various modes a person may suffer very severe penalties at the hands of others for faults which directly concern only himself; but he suffers these penalties only in so far as they are natural and, as it were, the spontaneous consequences of the faults themselves, not because they are purposely inflicted on him for the sake of punishment.[84]

Therefore others have a right to express their own individuality, as well as their own beliefs, which will inevitably involve talking about others in a negative manner. Given this, individuals may talk about others so long as this “gossip” does not take the form of intentional punishment or go so far as to deprive an individual of his or her livelihood. Even if the gossip is about individual concerns and not those of society, an individual has no right to be insulated from it. Such gossip is understood to be the “spontaneous consequences” of the faults themselves, or part of the right of other individuals to express their opinions. Some of these concerns might be accommodated through the requirements, such as in American and New Zealand tort law, that publication be extensive and offensive, making the expression about another more like the “parading” or “punishing” that Mill states is not acceptable.

However, there is still another major source of difficulty for Mill in articulating a concern for privacy: freedom of expression. A legal right to privacy looks like a legal restraint on speech and for Mill, the chief protection of individual liberty and diversity in the face of social tyranny is not privacy but freedom of speech. To recognize a right not to be talked about would be to stifle the speech of others. In fact, Mill seems to recognize a tension between the intersubjective elements of speech and his division between conduct that concerns an individual and conduct that concerns others:

The liberty of expressing and publishing opinions may seem to fall under a different principle, since it belongs to that part of the conduct of an individual which concerns other people, but, being almost of as much importance as the liberty of thought itself and resting in great part on the same reasons, is practically inseparable from it.[85]

Mill can then be interpreted as coming down in favour of treating speech like thought, which concerns only an individual, rather than as an action that also concerns others. In doing so, speech takes primacy over any kind of concern that might ground a right to privacy.[86] This is reflected in the American experience with the tort of invasion of privacy. As discussed earlier, cases like Florida Star call into question whether a privacy interest can ever withstand a claim to freedom of expression, and in particular, freedom of the press.[87] Although courts in both the United Kingdom and New Zealand have disagreed with this primacy of freedom of expression, we can see that the justification for this objection faces some obstacles.

For all of these reasons, grounding a legal right to privacy in an analysis of the harm of gossip and curiosity faces some significant hurdles in moving beyond an understanding of privacy as an anticipatory remedy for other harms. Claims to privacy look like claims to insulation from the social pressure resulting from others’ inquisitiveness into our lives and their gossip about us—a social pressure that promotes a kind of social conformity. It is difficult to justify legal coercion to vindicate this interest unless it causes some kind of injury. This explains the popularity of the test that the publicity of private facts be highly offensive, since this requirement singles out cases where there is significant “humiliation and distress.”[88] Importantly, whatever harms arise from this pressure need to be viewed in light of the speech interest of others. It is difficult to articulate why diffuse social pressure that does not rise to the level of more defined and tangible harms should take precedence over a right to speech, especially when a right to speech can itself promote individual diversity. Through a justificatory lens, a Millian, harm-based argument for a private law right to privacy is thus left looking like the American version of the tort of publicity given to private life.

B. The Coercion Argument

Apart from a focus on harms, the second major liberal strategy for grounding legal rights is a focus on autonomy or freedom from the coercion of others where that coercion is understood in relation to individual freedom rather than individual harm. Of course, there are diverse accounts of autonomy and of its relevance to questions of legal rights. One dominant account of autonomy is that of Kant and this section outlines the significant hurdles it also faces in articulating why gossip and curiosity could amount to a legal, rather than an ethical wrong.

Kant’s division between external and internal freedom illuminates the hurdles faced by this account of autonomy. The doctrine of right, concerning what Kant refers to as external lawgiving, explicates the conditions of external freedom. Kant describes this as a matter of the reciprocal relation of choice in which the wishes, needs, and ends of the other are irrelevant. He writes, “Right is ... the sum of the conditions under which the choice of one can be united with the choice of another in accordance with a universal law of freedom.”[89] In contrast, Kant’s doctrine of virtue concerns duties of inner freedom.[90] Inner freedom pertains to the motivation for action, which must be self-chosen and self-legislated rather than imposed. Consequently, the “duties of virtue are duties for which there is no external lawgiving.”[91] Therefore, for Kant a legal right is one that protects an individual’s external freedom—his or her agency—from violation by others.

The boundaries of what is considered an impermissible violation of external freedom are set by the idea of the equal agency of all. A Kantian account of a legal right to privacy would therefore have to show that the gossip and curiosity of others can in some instances violate external freedom. In this regard it is important to note that the violation of external freedom need not result in an injury to an individual. For example, if another takes what is mine and uses it for his or her own purposes, then I am wronged. This is the classic articulation of a trespass to land: you cannot take and use my property without my consent even if such use causes me no injury.[92] A Kantian account of privacy therefore has the potential to reorient the debate of privacy away from a focus on the effects and harms of gossip and curiosity. Indeed, it could ground a claim for a privacy invasion where someone was surreptitiously monitored and never found out about it.

The key to a Kantian justification of privacy is therefore to ground a sense in which information about me is “mine”, such that in collecting it and using it for your own purposes you wrong me. However, as outlined below, such an approach faces a number of significant hurdles in justifying the “mineness” of such information. These hurdles can help to explain why courts face a constant temptation to move back toward a confidentiality paradigm, why concerns regarding harm continue to intrude upon the analysis, and why freedom of expression is such a potent countervailing interest.

For Kant, my reputation is “mine”. As he argues, “a good reputation is an innate external belonging, though an ideal one only, which clings to the subject as a person.”[93] Gossip that affects another’s reputation would therefore violate that individual’s external freedom. However, as noted in the previous discussion regarding harm, this would ground a right against the spreading of only false information, not true information, and so would not protect an individual’s desire to have some true facts kept from further circulation. In other words, it grounds the legal doctrine of defamation but not privacy.

What, then, about the many things that I endeavour to keep private? Here Kant’s distinction between physical possession and rightful possession, as outlined in his discussion of property, is helpful. My innate right to my body means that I have a right to an apple while I am holding it; to interfere with this possession is to interfere with my right to occupy a particular space. However, if I put the apple down then you wrong me only if I can additionally claim that the apple is my property—that it is something over which I have rightful, and not merely actual, possession.[94] Similarly, if I endeavour to keep my thoughts, emotions, and sensations private, then your interference with the steps I take to do so is an interference with my innate right to personality. However, once I choose to communicate them then there must be a different basis for my entitlement to my thoughts, emotions, and sensations. If a right to privacy is to be more expansive than something like secrecy—which seems to be at stake in the move from confidentiality to privacy—then what is needed is an understanding of privacy analogous to the rightful possession of the apple.

One way to create such an account is to focus on the relationship between the individual who communicates her thoughts and the individual who then makes some further use of this communication. Through invoking the norms of contract or confidentiality, a Kantian account could allow for some of these uses to be wrongful. For example, an individual who receives letters from you on the understanding that these are confidential wrongs you in publishing them. But these doctrines would not catch the case of a stranger seeking to publish my private letters, unless this piggybacks upon the initial relationship, such as when a stranger understands that the information at issue has been obtained through a breach of confidence. As discussed, this still does not catch private information per se.

Since technology permits access to an individual’s thoughts, emotions, and sensations without this access violating any special relationship, Warren and Brandeis argued that what is required is a right against the world such as that afforded by property, not by confidentiality or contract. By this they did not mean the traditional doctrine of property such as one’s property in the home. Historically, the protection afforded by trespass went a long way to protect the privacy of the home. But such protection is vulnerable to the critique that technology makes it possible to intrude on others without actually violating any property right. Instead, Warren and Brandeis grounded their argument in common law cases regarding copyright protection for unpublished works, which granted individuals a right to prevent the publication of original expression. They argued that such cases were better understood as applications of a more general right to privacy than as intellectual property cases and should therefore be expanded to protect one’s thoughts, emotions, and sensations regardless of form: “whether expressed in writing, or in conduct, in conversation, in attitudes, or in facial expression.”[95]

Would Kant recognize such protection for private letters and would he allow for the expansion of such protection in the manner sought by Warren and Brandeis? Kant clearly recognizes a right against the copying of books. A book, according to Kant, “represents a discourse that someone delivers to the public by visible linguistic signs.”[96] No one else may attach his name to another’s work, for this would be to pass oneself off as the author. Similarly, if an author speaks in his own name through a book, a publisher speaks in the name of the author and may rightfully do so only with the permission of the author.[97] However, the wrong involved in publishing without the author’s permission is the wrong “of stealing the profits from the publisher who was appointed by the author.”[98] It is not clear how this wrong would extend to someone who is publishing previously unpublished material without the author’s permission. It is also not clear how it would extend to someone who wanted to comment on the private communication of another and so was more in the position of a journalist than a publisher who could be said to be speaking in the name of the author.[99]

Furthermore, it is not clear that Kant would extend his analysis of copyright in the manner argued for by Warren and Brandeis, to include protection from the unauthorized dissemination of one’s thoughts, emotions, and sensations in whatever form. Kant distinguishes writing from other kinds of works such as “an etching which represents a certain person in a portrait, or a work in plaster that is a bust” for these, unlike writing, are immediate signs of a concept.[100] To interfere with another’s use of such signs would presumably interfere with an individual’s freedom to communicate his own thoughts. Similarly, to grant an individual the control over disseminating details of one’s personal appearance, acts, and relationships would be to interfere with the ability to express one’s opinion about another.

In fact, there is a strong basis upon which to argue that for Kant, speaking about others is generally something that one does as a matter of right rather than something that violates rights. This is because the principle of innate freedom implicitly involves the authorization

to do to others anything that does not in itself diminish what is theirs, so long as they do not want to accept it—such things as merely communicating his thoughts to them, telling or promising them something, whether what he says is true and sincere or untrue and insincere (veriloquium aut falsiloquium); for it is entirely up to them whether they want to believe him or not.[101]

In a footnote to this statement, Kant elaborates that intentionally telling an untruth is considered a lie “in the sense of bearing upon rights” when it deprives another of what is his, such as “the false allegation that a contract has been concluded with someone.”[102] Therefore for Kant, communication to others is presumptively authorized even if it involves insincere falsehoods so long as such communication does not affect another’s external freedom. As has been argued, it is difficult to determine how such communication could affect another’s external freedom apart from the limited protection afforded by defamation, contract, confidentiality, and copyright.

Indeed, for Kant, a general concern for how others talk about an individual falls more directly within the realm of ethics than law. Consider Kant’s discussion of what he calls defamation, which is closer to the sense of gossip that we are concerned with. Kant argues:

The intentional spreading (propalatio) of something that detracts from another’s honor—even if it is not a matter of public justice, and even if what is said is true—diminishes respect for humanity as such, so as finally to cast a shadow of worthlessness over our race itself, making misanthropy (shying away from human beings) or contempt the prevalent cast of mind, or to dull one’s moral feeling by repeatedly exposing one to the sight of such things and accustoming one to it. It is, therefore, a duty of virtue not to take malicious pleasure in exposing the faults of others so that one will be thought as good as, or at least not worse than, others, but rather to throw the veil of benevolence over their faults, not merely by softening our judgments but also by keeping these judgments to ourselves; for examples of respect that we give others can arouse their striving to deserve it.—For this reason, a mania for spying on the morals of others (allotrio-episcopia) is by itself already an offensive inquisitiveness on the part of anthropology, which everyone can resist with right as a violation of the respect due him.[103]

Kant therefore classifies defamation, understood here as malicious gossip, as a vice that violates the duty of respect that we owe to other human beings. However, such a violation concerns the violator’s inner motivation: his failure is a failure to treat another as an end in itself. This is why the violation is a matter of the doctrine of virtue and not the doctrine of right. In other words, one can lack respect for another without violating his or her external freedom.[104]

Kant does argue, however, that one of the duties of right is the duty of rightful honour, which he expresses as, “Do not make yourself a mere means for others but be at the same time an end for them.”[105] One could argue that this duty would allow an individual, as a matter of right, to resist being treated as a means even if this treatment falls short of constituting a violation of external freedom. That is, one might have a privilege to resist treatment that involves a lack of respect even if one does not have a right to bring a claim against the disrespectful perpetrator. This privilege appears to be the basis for Kant’s statement that we can resist the offensive inquisitiveness of others “with right”: I can resist the prying of others and have no obligation to co-operate in their inquisitiveness. But if someone finds out something about me through some other means and spreads this gossip, I have no legal claim against them.

Does something particular to mass media make a difference such that we could make sense of the requirement of “publicity”? It may be that the significance of mass media, and other contemporary forms of information and communications technology, is that they make the prying of others more effective and individual efforts at resistance less effective. For example, suppose that I enter a restaurant and see my neighbour there with a man who is not her husband. Apart from taking steps to avoid detection, my neighbour has no legal claim to prevent me from telling others about it the next day. Suppose that instead I publish this information on a community bulletin board on the internet. This is even a greater interference with my neighbour’s attempts to avoid detection, as it dramatically expands the number of people who can now “observe” her. Nonetheless, unless these practices somehow affect the nature of the relationship between me and my neighbour, then it is difficult to see how an individual’s external freedom is violated. She has a privilege to resist; I do not wrong her by prying, even if in doing so I act unethically.

A Kantian approach to privacy therefore faces a number of significant hurdles in articulating an account of privacy. Moreover, many of these difficulties are similar to those that arose in the previous discussion of harm. First, it is difficult to isolate a general concern for the wrong of gossip and curiosity apart from more specific wrongs such as defamation or breach of confidentiality. Second, any attempt to do so quickly encounters conflict with freedom of speech, and a strong entitlement to freedom of speech is easier to ground than a strong entitlement to privacy. The result is that any kind of strong censure against the practices of gossip and curiosity, even when enhanced by mass media, must be a matter of moral but not legal concern. At most, a Kantian approach would endorse grounding privacy claims in breach of confidentiality. This review of a Kantian approach to privacy therefore gives an account of why one recurring version of the judicial containment anxiety regarding privacy is either to revert to a confidentiality paradigm, or to resist a shift out of it in the first place.

III. Alternatives: Identity versus Authenticity

The justificatory dilemma, as outlined in the previous section, leaves several options open. The first option is to follow Mill and create a justification for the American version of the tort of publicity given to private life, with its requirement of harm and its threatened engulfment by freedom of expression. The second is to follow Kant and provide protection only for breach of confidentiality, with its focus on the protection of confidential relationships rather than a particular category of information. Each of the first two options has costs, because each considerably narrows the range of privacy claims that the law is able to protect. The third option is to reject both horns of the dilemma. Rather than using the justificatory dilemma as a means to narrow privacy protection, we can use it to shed light on the nature of what I have been calling the emerging privacy impulse.

The most important thing that the justificatory dilemma tells us is that the value protected by a robust private law right to privacy is neither autonomy nor freedom from certain kinds of harms. If we can understand the nature of the wrong involved in the practices of gossip and curiosity—and their technological augmentation or transformation—in different terms, then perhaps we can answer the justificatory dilemma without narrowing the scope of legal protection of privacy.

One way to re-envision the question of what kind of value the right to privacy protects is to argue that privacy is an important condition for autonomy, but is nonetheless conceptually distinct from the individual liberties that are seen to secure freedom in a liberal democracy. Thus, for example, Beate Rössler argues:

If we consider the telos of freedom to be autonomy and thus the possibility of asking oneself what sort of person one wants to be and how one wants to live, and if civil liberties guarantee just this (within the well-known limits), it is clear that—when it comes to the question of how one would like to live—violations of privacy restrict or tie a person down in a way that contradicts the spirit of civil liberties demanded and secured, but that for a variety of reasons is not or cannot be prevented by the civil liberties themselves.[106]

On such a view, privacy protects individuals in their ability to reflect on the kind of person they would like to be and on how they would like to live.[107] In this sense, privacy protects aspects of one’s self-relationship: one’s ability to assess and appraise personal choices and to live an authentic life. Rössler’s approach is promising, as it provides us with a way of understanding the significance of being insulated from the gossip and curiosity of others—and its attendant social pressure—in a manner that relates directly to the value of autonomy, without the need to establish that violations of privacy are themselves direct violations of autonomy.

Although generally helpful at outlining one of the values protected by privacy, this approach nonetheless faces its own difficulties in justifying a private law right to privacy, particularly of the kind at issue in cases of publication of private facts.[108] For example, Rössler does not deal with this particular legal question but does deal with the analogous issue of surreptitious surveillance, which she views as a paradigmatic violation of informational privacy.[109] One could argue that the publication of private facts is akin to placing someone under surreptitious surveillance and then placing the results of this surveillance in public view. The problem with such practices, on her account, is that they result in an individual acting upon false assumptions regarding what others know about her, compromising her ability to engage in authentic behaviour toward others. However, even if one accepts this account, it does not show why the publication of private facts is problematic: prior to their publication, one acts under the true assumption that such facts are not generally known and after their publication, one acts under the true assumption that such facts are now generally known. Indeed, the problem of false assumptions can be generally remedied though widespread publication and dissemination of “private” information as much as through respect for privacy norms. Once the results of surreptitious surveillance are publicized, there is no problem of false assumptions and therefore no problem regarding authenticity.

Rössler has a slightly different rationale for why known surveillance is problematic. When an individual knows that she is being observed, she adapts her behaviour accordingly. This adaptation, according to Rössler, also interferes with authenticity: one acts in ways one would not have acted if unobserved. The problem is that instead of having one’s behaviour issue from one’s own beliefs and values, one instead acts in reaction to the reality of observation—observation that therefore inhibits an authentic relation with one’s self. However, even if one accepts this account, it still does not show why the publication of private facts is problematic. In the typical fact scenario of such a case, one does not know that one is being observed until after publication. When Naomi Campbell attended Narcotics Anonymous, she did not know that she was being observed and so did not change her behaviour in response to this observation. The subsequent publication of the details of her treatment does not change the character of her original action. To say, “had I known this would be in the papers I would not have engaged in this behaviour” does not mean that the original behaviour was inauthentic—it just means that you did not want others to know of it. Part of the difficulty is that if privacy protects self-relation with respect to some behaviour, then it is unclear how publicity that is unanticipated and that occurs after-the-fact could retroactively alter this self-relation, thereby rendering the behaviour inauthentic. The only case that Rössler’s view of known surveillance could account for is when publication, because of its predictability or prevalence, effectively amounts to a kind of anticipated surveillance that would have the inhibiting effects she describes. This might include instances where certain celebrities are routinely followed and hounded by the press.[110]

A similar version of the authenticity thesis is one that I have put forward in the past: in presenting yourself to others (e.g., when you are in public), your presentation is governed by the norms applicable to that particular context. We require some respite from this context if we are to fully affirm those aspects of our thoughts and identity as in fact ours and not a simple response to the desire to conform to these norms of presentation.[111] Although this accounts for some of our intuitions regarding why privacy is valuable, and shows why a society of widespread surveillance is problematic, it also faces difficulties in accounting for the legal right to privacy as we find it in tort law. It is one thing to argue that we need some respite from the public gaze in order to forge an authentic self, and quite another to assert that the particular gaze of a particular other in fact interferes with this project of authenticity.

A more promising account of privacy views the value of privacy as protecting one’s identity rather than one’s authenticity, autonomy, or freedom from harm. Before I outline why an identity-based account avoids the difficulties inherent in an authenticity account, let me outline why it also provides a route out of the justificatory dilemma.

Liberal justifications for rights based upon notions of either harm or coercion remain neutral with respect to the particular ends of individuals affected. Indeed, this is consistent with a general view of the common law, and the role of the courts, whereby the common law protects a realm of individual freedom rather than a particular vision of the good. Because of this, private law discussions often proceed in abstraction from the particularities of any individual’s actual identity, including their beliefs, desires, and needs.[112] Nonetheless, this does not mean that the law should be unconcerned with an individual’s capacity to construct his or her identity. Indeed, one could argue that liberal accounts of rights either presuppose or at least are strongly compatible with the idea that individuals have an identity and that individuals have the capacity to participate in its construction.

The idea of identity at stake here is not my own idea of myself resulting from a process of self-reflection and does not rely upon any idea of a “true”, “inner”, or authentic self. Instead, the idea of identity that I want to invoke is that of the “self” that one presents to others—one’s “public” persona, or “social self”.[113] This “self” that is presented may or may not differ in relation to different “others”, and may or may not vary over time, across contexts and social roles in contradictory ways.[114] In short, this is a very “thin” notion of identity that focuses simply on one’s ability to present oneself to others. It is therefore quite different from ideas of identity that have been invoked by other legal theorists. For example, some outline the connection between privacy and identity in terms of “personhood”, which itself is variously understood in terms of the protection of choices that are seen as intrinsic to one’s identity,[115] the “integrity” of one’s identity,[116] and the conditions under which a society recognizes that an individual’s “existence is his own.”[117] Others argue that privacy protects the capacity of individuals to withdraw from the demands of social interaction in order to protect the development of individuality or even authenticity.[118] In contrast, one can be concerned about an individual’s capacity for identity formation in the sense of “self-presentation” while remaining indifferent to these other questions regarding the particular nature and quality of one’s identity.

Although identity, on this understanding, is always constructed in relation with and to others, it is problematic from the perspective of liberalism to posit that this identity is fully constructed by others. Therefore, just as the claim that one has complete self-sovereignty over identity is implausible, so too is the claim that one has no role in its creation. There must be some capacity to participate in the construction of one’s identity—some ability to determine whether some elements are disclosed to others, to whom they are disclosed, and in what circumstances.

Although a full account of the relationship between this idea of one’s capacity for identity formation and justifications for legal rights is beyond the scope of this paper, this argument at least establishes the plausibility of this claim. Identity therefore provides a potentially fruitful way out of privacy’s justificatory dilemma by positing an alternative to autonomy or harm as the justification for the imposition of legal liability. Distinct from values such as autonomy, identity still fits within a recognizable family of liberal interests that underpin much of our liberal legal tradition and provide the justification for legal liability. One could even argue that to be harmed or to act autonomously presupposes an individual who at least has this bare capacity for identity formation, however poorly it may be exercised. Nonetheless, it remains to be seen whether identity can provide a compelling basis for understanding privacy. What does an identity account of privacy look like and how does it differ from other accounts?

Once we turn to identity as our central focus, a privacy claim must be conceived in terms of protecting the conditions of self-presentation rather than as a claim to insulate us from the social pressure that might result from a particular self-presentation. Basic to this idea of self-presentation is the requirement that you know the “who” to whom you present yourself. It is not the case that you present yourself to the world in general, and then particular others may or may not think well of this presentation for a myriad of reasons. The claim here is that the very idea of self-presentation includes the idea of presentation to someone, to an audience. It is not a kind of bare self-assertion but an act of social communication—an act that depends upon knowing the “who” to whom one is communicating. Knowledge of this “who” therefore does not go simply to one’s effectiveness in creating an identity but to one’s basic capacity for doing so. Privacy violations involve disruptions in this communication—for example, by changing the identity of one’s audience through surreptitious surveillance or unexpected publication.

Consider the following two examples of information disclosure, one of which is a privacy violation, and one of which is not. First, suppose that you tell another academic colleague, “X was invited to visit Yale.” You have disseminated information about X to another colleague, but few of us would consider this gossip. Contrast this with a second scenario, in which you instead tell your colleague, “X is having an affair with Y.” Now you have disseminated information about X to another colleague and most of us would consider this to be gossip. The difference between the two scenarios is that in the second you have passed along information that X would likely not want other professional colleagues to know about. It is in this sense “private” where “private” simply refers to information that one would not want to be part of one’s self-presentation to a particular audience. In contrast, in the first example it is unlikely that X did not want her work colleagues to know that she had been invited to visit Yale even if she had not yet told them. The difference between the two examples—what makes the second gossip and the first merely the dissemination of social knowledge—is the fact that one involves “private” information whereas the other does not. Moreover, we can understand what “private” means through reference to the idea of self-presentation rather than some inherent features of the information itself or the delineation of a “private sphere”. For example, the things that we conventionally label “private” in relation to our professional lives are the things that most of us would not want to be part of our self-presentation in a professional context. In this way, conventional markers are helpful in that they alert us to situations where individuals might not have wanted information to be disclosed to a particular audience. (For example, most of us would think it is inappropriate for employers and colleagues to find out about one’s sexual history.) But they are a proxy for the individual’s determination regarding the “self” that she wants to present to a particular audience.

This analysis of privacy in terms of self-presentation highlights an important difference between an identity account of privacy and a harm account of privacy. Importantly, the focus is on the self-presentation aspect, and not on how this presentation is received or dealt with by the intended audience, or on the social pressure that others may bring in relation to particular presentations. Therefore, on this account, the United States and New Zealand’s limitation of liability to publications that are “highly offensive” is mistaken. Instead, this approach would endorse the House of Lords’s position that described the test for offensiveness as helpful to the identification of a privacy violation in cases where it was unclear.[119]

However, an identity-based account of privacy would also call into question some of the features of the U.K. position regarding confidentiality. First, while a focus on self-presentation retains a relational component to privacy that is analogous to confidentiality, it is nonetheless a much broader notion of relationship. One’s capacity to present oneself to others does not depend on a confidential relationship with others even if it does depend on knowing who the others are. Second, the U.K. position on confidentiality left open the question of whether publication was required for liability, or whether mere disclosure of “private” information is sufficient. As I outline below, a violation of privacy understood in terms of a capacity for self-presentation will usually require publication.

The mere dissemination of private information does not necessarily imply a violation of privacy. If privacy is a claim to protect the conditions of self-presentation, then a violation of privacy will occur only when the dissemination of private information also undermines one’s capacity for self-presentation. It is not the case that simply passing along information about X that she would not wish to pass along undermines her own self-presentation to others. Schoeman helpfully observes,

We surely invade a colleague’s privacy if we announce at a meeting that she and our secretary are having an affair. Norms of privacy make this sort of disclosure unconscionable. Norms of privacy, however, do not make it seem as serious, or even at all serious, if we simply privately relate the same information to each person in the department. What we mean by privacy, then, or invading a person’s privacy, is not the fact of disclosing personal information to a variety of people without the consent of the object of discourse, but the means by which the private information is distributed. Thus, characterizations of gossip as revelation are incomplete or misleading. We must differentiate dissemination from publication. Publication means dissemination plus the conversion of a matter that is personal into a matter that is “open” or acknowledged as a “public fact.”[120]

For Schoeman, the significance of “public facts” is that they bring with them a set of norms whereby it is appropriate for others to hold you accountable for such facts and to use social pressure to do so. Since gossip operates “behind the back” and does not deal with public facts, it “permits a person to maintain a public face” insulated from this pressure.[121] Placed in terms of an account of privacy as protecting the conditions of self-presentation irrespective of any attendant social pressure, we can say that the dissemination of “private” information affects one’s self-presentation only when it becomes a “public fact”. For example, upon “publication” that X is having an affair with Y, X has little choice but to make this information part of her self-presentation, or to revise some element of her self-presentation to reflect this. Unlike in the case of the colleague who knows about the affair but does not let on that he knows, she can no longer “save face” or maintain her original self-presentation as if nobody knows.[122] This can account for why American privacy law has generally not imposed liability for gossip unless the information is published.[123]

We can see the difference that the “publication” requirement makes by re-examining Campbell. Naomi Campbell did not want to disclose to the public any information, either of a general or specific nature, regarding her treatment for drug addiction. For this reason, we can say that the Daily Mirror disseminated “private” information. The salient distinction is not between public and private spheres of life, but rather between the types of audiences to which Campbell did and did not want to present this aspect of her self. However, although this information is private in relation to the general public, its dissemination is a violation of privacy only upon publication. The significance of publication is that once published, Campbell has little choice but to make this information part of her self-presentation, or to revise some element of her self-presentation to reflect or contest this. She cannot “save face” and carry on as if nothing happened.

There may of course be justifications for this violation of her privacy and, indeed, in this case all of the Law Lords agreed that the Daily Mirror was entitled to set the record straight regarding the fact that Campbell was addicted to drugs since she had publicly claimed that she was not. One could argue that if we decide to present some aspect of ourselves to a particular audience then that audience is entitled to “pry” in order to determine if those aspects are in fact accurate. The point is that we can understand the nature of the privacy violation at issue without reference to a general idea of a “private” sphere distinct from the public, or ideas of confidentiality, or ideas of harm.

This analysis of privacy as protecting the conditions of self-presentation can also show why there was a privacy violation at stake in the Hosking case and why the New Zealand Court of Appeal should not have so easily dismissed the idea of privacy in “public”. The capacity for self-presentation of children as young as the Hosking twins is, of course, limited. However, we can accept the parents’ presentation of them as a reasonable proxy. We can agree that in pushing the children in a stroller in public, the mother presents her children to the people who are in that public place. However, she does not necessarily choose the audience; because it is a public place, others are entitled to be physically present and their observation is incidental to this.[124] Therefore, from this descriptive fact of being in “public” one cannot infer the normative implication that Mrs. Hosking would have chosen to present her children in this way to the additional and quite different audience of a particular publication who might view a photo of them. Because it forces a new audience upon her, it undermines her capacity for her own and her children’s self-presentation. Moreover, unlike the people she meets while in a public place, the audience of a particular publication lacks an immediate justification for their observation of her and her children. Given this, the approach of the English Court of Appeal in Murray—which on similar facts accepted that the child had a legally significant privacy interest—is much more defensible.[125]

This is not to say that there are no compelling justifications for the dissemination of private information through “publication”. There may be many justifications for what would otherwise be privacy violations. In particular, it seems reasonably clear that if you have to interact with a person in a particular context, you are entitled to information about the person that is relevant to that interaction, even if that person would not wish you to have it. The demands of others can and do limit individual claims to control over self-presentation. We might therefore claim: we all have the ability to talk about other people behind their backs and to talk in front of their faces about things that are relevant to our relationship. Obviously much more work would need to be done to fully delineate the relationship between freedom of expression and privacy understood in terms of self-presentation. Nonetheless, the brief account here suggests several possible conclusions regarding this relationship. First, freedom of expression does not obviously trump privacy and vice versa. Second, even though these rights might be in tension in particular factual circumstances, they are not in fundamental opposition: both freedom of expression and privacy understood in terms of one’s capacity for self-presentation protect expressive activity. The “balancing” approach endorsed by courts in both the United Kingdom and New Zealand is therefore more easily justified than that of the United States, where freedom of expression has essentially engulfed the tort of publication of private facts.

Even if “publication” is generally necessary, is it always necessary in order to find a violation of privacy? For example, it might be that your self-presentation is affected by the dissemination of private information even in the absence of publication if your reputation is altered to such an extent that people interact with you as if you are a different person from the “self” you present. You then have no capacity to present yourself as other than this person even though in some sense this gossip retains its behind-the-back quality. The difficulty for ascribing tort liability in such a situation would lie in identifying a wrongdoer rather than in identifying a wrong. That is, each individual who participates in the gossip does not violate privacy merely through the dissemination of “private” information although the cumulative effect of this might be to undermine one’s capacity for self-presentation. Perhaps if the information is of such a nature that one could reasonably foresee such effects of further dissemination, then it is a privacy violation even without meeting the “publication” requirement. The American tort of giving publicity to private facts does not foreclose this possibility.[126]

Privacy understood as a claim to protect the conditions for self-presentation is therefore promising for a number of reasons. First, because it does not rely on what an audience does, we do not have to identify a threshold of harm or form of coercion in order to impose legal liability. Second, it provides a compelling justification for why legal liability should follow from the publication of what would otherwise be gossip. Third, it provides a plausible basis for asserting that freedom of expression needs to be balanced against the protection of one’s capacity for identity formation rather than engulfing it. Finally, this capacity is capable of being defined—and much more needs to be done than is presented here—and therefore provides the basis for differentiating privacy violations from other types of legal wrongs in a manner that can potentially address concerns regarding the vagueness of privacy claims. In this way it provides a promising approach to understanding the nature of the wrong at issue in privacy rights, and therefore a way of developing a privacy tort that responds to the justificatory dilemma without retreating to an overly narrow view of liability.


Although there is an emerging consensus regarding the need for private law protection for privacy, at least in the context of the publication of private facts, this paper has argued that the resulting jurisprudence exhibits a number of important tensions. Many of these tensions revolve around the questions of whether harm is required to give rise to liability for infringement of privacy, whether “private” refers to a category distinct from “confidential”, and the question of the role played by freedom of expression. This paper has argued that the tensions in the case law are not best understood in terms of the privacy impulse that underpins this emerging consensus but rather the containment anxiety that motivates courts to seek strong limits upon the scope of protection provided. In turn, this containment anxiety is best explained through reference to the courts’ search for either the kind of harm or the kind of coercion that could justify the imposition of a legal obligation to respect another’s privacy. In this way, the privacy impulse surfacing in case law is undercut by a justificatory dilemma.

In response to this justificatory dilemma, I have argued that the right to privacy does not protect one against harm or coercion. Instead, it protects one’s identity where this is understood in terms of one’s capacity for self-presentation. Not only does this approach provide a compelling justification for the imposition of legal liability in contexts such as the publication of private facts cases, but it also provides a clear account of the nature of the wrong at issue in such cases and so illuminates something important about the nature of privacy.

Such an account could potentially help courts to understand the connections between various types of privacy claims, providing more credence to the idea of a “high-level principle” of privacy that does in fact have an accessible conceptual core. Indeed, an account of privacy as protecting the conditions for self-presentation has the potential to unite the four branches of the American tort of invasion of privacy. Apart from the publication of private facts branch, the law also provides relief against false-light advertising, misappropriation of name or likeness, and intrusions upon seclusion. Despite Prosser’s influential argument that these four branches in fact protect four separate interests, one could argue that the four branches in fact protect against four different kinds of violations of one’s capacity for self-presentation.[127] For example, publishing an advertisement that indicates an endorsement of a product puts words in the plaintiff’s mouth that he or she has not chosen. Surely this is the most basic sense in which one has a capacity for self-presentation—to choose what one wishes to express and to whom. This capacity is also undermined when another “appropriates” one’s identity and uses it without consent. Similarly, intrusions upon seclusion change the audience of one’s self-presentation against one’s wishes. Such an account might also provide the basis upon which to relate the tort of invasion of privacy to statutory “privacy” regimes modelled upon fair information practices and which generally provide individuals with rights relating to the collection, use, and disclosure of their personal information—a set of interests that commentators increasingly discuss in terms of protecting an individual’s “digital persona”.[128]

An understanding of privacy in terms of the protection of the conditions for self-presentation, and its relationship with various types of legal regimes purporting to protect privacy, obviously requires much more elaboration. However, the claim here is that such an account not only is compatible with liberal justifications for legal rights, but also may help us understand a wide range of privacy claims in a manner that is difficult if we remain focused on definitional difficulties, or traditional analyses of harm or coercion. In this way, examining what I have been calling the justificatory dilemma provides not just a more compelling analysis of the containment anxiety present in the cases, but also a shape to the privacy impulse that is emerging.