Jurisdiction and personality rights – in which Member State should harmful online content be assessed?

1. Introduction ¹

Legal responses to online infringements are complex for many reasons. Law is put into action in courts, most of which are national. The territorial boundaries of their jurisdiction are ill-suited to deal with the borderless nature of the internet. This article concentrates on the difficulties of deciding jurisdiction when harm happens online. Specifically, it deals with infringements of personality rights. When an alleged infringement occurs on a webpage, in an online newspaper or on a social media platform, the question arises as to which Member State has jurisdiction. ²

In this article, I will discuss the case law from the CJEU on jurisdiction concerning the protection of personality rights, such as privacy or personal reputation. The crucial provision is Article 7(2) of the Brussels Ia Regulation (henceforth ‘the Regulation’). Disputes concerning national jurisdiction are generally resolved with reference to the principle of predictability. This article asks, in essence, what it means – and what it should mean – in an online environment.

There are three particularly relevant judgments that guide interpretation when assessing the infringement of personality rights online: eDate, ³ Bolagsupplysningen ⁴ and the very recent Mittelbayerischer Verlag. ⁵ The latter will be of particular interest for our analysis because of the unusual facts of the case, which highlight certain problematic aspects of personality rights. ⁶ The case illustrates, among other things, the discrepancies that arise from the fact that jurisdiction is harmonized while personality rights are not.

This recent ruling can also function as a foundation on which to start developing new ways of thinking. Both the judgment, delivered in June 2021, and the opinion of Advocate General Bobek from February 2021 provide inspiration for rethinking the problems pertaining to online issues. ⁷ Hence, I will discuss jurisdiction by drawing on these three central cases but focus especially on Mittelbayerischer Verlag.

Considering that the internet is becoming the most relevant site where these kinds of infringements occur, the law would need urgent clarification. Unfortunately, the latest major legislative endeavour that concerns liabilities online, the Digital Services Act (DSA) proposed by the Commission in 2020, does not seem to promise improvement in this regard. ⁸ There is no definition of the interpretation of personality rights nor of the ways of determining jurisdiction in the proposal, even though the Act specifically concerns responsibilities online. The Commission's proposal does refer to the Brussels Ia Regulation, but does not offer anything more tangible that could resolve questions of jurisdiction.

As there is no help to be had from the legislators, the focus turns to legal practice. The article demonstrates that the CJEU's interpretation practices in cases of personality infringements online have been problematic. They cause increasing uncertainties pertaining to the interpretation of Article 7(2) of the Regulation. The Mittelbayerischer Verlag case offered a chance to clarify the meaning of the Article, but unfortunately, the Court did not seize that opportunity.

What this article shows, in essence, is that there are two separate issues that become intertwined in these cases. Firstly, the definition of personality rights and the lawful protection of them varies in different Member States. Secondly, the rules for determining the forum in cases of infringement of personality rights are unclear. This leads to the situation where the rules of international jurisdiction are ill-suited to accommodating the different understandings of personality rights.

Ultimately, the issue of protecting personality rights online has to do with the difficulty of balancing freedom of information and the protection of certain rights. The indeterminacy of the law in the area of jurisdiction is one part of a larger picture that causes real-life problems. If the media and other actors publishing content online do not know where they might be sued and for what possible infringements, this could undoubtedly lead to limitations in freedom of information, that is, the chilling effect. On the other hand, added clarity on the possibilities of injured parties to take legal action is also needed for the sake of access to justice. As the situation stands now, the European rules governing these situations do not serve anyone very well. ⁹

In the absence of clarifying legislation, I propose that the legal assessment of personality rights infringements online develops in a direction where the so-called mosaic approach is abandoned. The place of harm should be assessed on the basis of the centre of gravity of the case, which is considerably clearer. The centre of gravity of the case would include elements concerning the applicant but also elements concerning the publication. It would strive at an objective assessment, not focusing solely on the individual whose personality rights have been infringed. This approach will not resolve all of the problems but it can nevertheless serve as an appropriate instrument for tackling some of the pressing issues.

2. What are personality rights?

Personality rights do not have a clear definition in EU law, ¹⁰ which is where they differ from personal data protection, for instance, which has been codified in Article 16 TFEU, Article 8 of the Charter and much secondary law. Personal data duly has a specific definition, on which there is also settled case law from the Court. ¹¹ The meaning and scope of personality rights, on the other hand, are much less clear. No easy competence rules can be found, on the basis of which the protection of personality rights could be harmonized. ¹² In effect, personality rights tend to refer to a cluster of rights, such as privacy, identity and dignity.

‘Personality rights’ is an overarching term, and the protection of these rights derives from Member States’ legal systems. These rights have largely been protected in EU law through reference to national legal sources. However, definitions can also be found in the case law of the European Court of Human Rights (ECtHR). In its decisions, the Court has included in the protection of private life, according to Article 8(1) ECHR, several aspects relating to personality and personal identity, such as a person's reputation, name or picture. ¹³

The fact that the definition of personality rights rests upon Member States’ law becomes important when analyses about jurisdiction are made. When considering the jurisdiction of a Member State court to hear a case on personality rights infringements, the court will naturally base its assessment on the definition of personality rights in that specific Member State. ¹⁴ This is one reason why jurisdiction becomes so important for all parties involved: each Member State may have different rules on what personality rights are and how they are protected. Member States use these rights to protect very different things.

Even though personality rights have not been harmonized, the CJEU has pronounced several times its view on what they might be. Case law indicates that at least the following can be classified as personality rights: the right to privacy, ¹⁵ the right to one's own image, ¹⁶ the prohibition of defamation ¹⁷ and the protection of good name and reputation. ¹⁸ In the context of jurisdiction disputes, a distinction is sometimes made between violations of privacy and (other) rights relating to personality, but this is not always the case. ¹⁹ Furthermore, the case law on infringements of personality rights covers a wide range of infringements, be they the tort of defamation in the sense usually attributed to this type of harm in continental legal systems, or the protection of privacy typical of common law systems. ²⁰ Moreover, the EU Court's jurisdiction shows that personality rights may belong to a natural person or, as was the case in Bolagsupplysningen, to a legal person. This way, the scope of these rights is wide and their content varies.

3. The rules on jurisdiction

Instruments of EU private international law do not distinguish between cases that take place offline and those that happen online. Both are governed by the same rules. ²¹ The relevant ones on jurisdiction are included in the Regulation on jurisdiction and enforcement in civil and commercial matters, the so-called Brussels Ia Regulation. ²² This recast was adopted in 2012 and has been applicable since 2015. It is the successor to the Brussels I Regulation and the Brussels Convention, which originated in 1968. ²³

The current Regulation from 2012 does not include any special rules on jurisdiction in the online environment. It is up to the CJEU to decide how the rules apply to situations where damage occurs through – or on – the internet. ²⁴ The most obvious problem is inherent in the nature of the net, which eludes all attempts at spatial definition. It is neither a place nor a space, nor is it located anywhere. All rules on jurisdiction in the Regulation, however, are conceptually linked to a place. Herein lies the core of the problem.

The Regulation aims at establishing clear and uniform rules for civil cross-border litigation within the EU. The functioning of the internal market is the key purpose from the Union's point of view. ²⁵ Difficulties in enforcing civil claims in other Member States tend to discourage persons from establishing cross-border trade relations and hamper the sound operation of the internal market. The Regulation is based on the principle of mutual trust, which includes the expectation that each Member State is willing to obey the provisions of the Regulation and to enforce them in national courts. ²⁶

Another objective of the Regulation is to secure the principle of legal certainty. The provisions seek to make the choice of jurisdiction predictable. ²⁷ A reasonable defendant should be able to foresee in which courts in the EU they could be sued. It is not a matter of discretion for a national court whether to entertain a suit or not. The court is bound by the provisions of the Regulation. ²⁸

As the Regulation does not include any specific rules on jurisdiction in online cases, there are several hurdles to overcome before it can be established which court(s) have jurisdiction. The interpretation of EU law is in general highly dependent on teleological reasoning, which the Court often employs. When interpreting the Regulation, the Court has pronounced that interpretation needs to proceed in a systematic and teleological fashion. ²⁹ Hence, it is always necessary to study what the purpose and aim of the relevant rules are. This Regulation seeks to do several things, but its overall purpose is to unify the rules of conflict in civil and commercial matters in a way that makes the choice of jurisdiction predictable.

The main rule is that jurisdiction lies with the courts of the Member State in which the defendant is domiciled. This is expressed in Article 4. There are, however, rules on special jurisdiction laid down in, inter alia, Article 7. In some instances, the claimant does not have to go to court where the defendant is domiciled but can choose some other jurisdiction that may be closer, easier or cheaper. According to the Court, however, the rules on special jurisdiction constitute derogations from the main rule and should be interpreted restrictively. ³⁰ An applicant may decide if they want to rely on the rules of special jurisdiction or not.

According to recital 16, special jurisdiction is called for in some situations:

In addition to the defendant's domicile, there should be alternative grounds of jurisdiction based on a close connection between the court and the action or in order to facilitate the sound administration of justice. The existence of a close connection should ensure legal certainty and avoid the possibility of the defendant being sued in a court of a Member State which he could not reasonably have foreseen. This is important, particularly in disputes concerning non-contractual obligations arising out of violations of privacy and rights relating to personality, including defamation.

Recital 16 maintains therefore that there has to be some link between the case and the jurisdiction, in order for the applicant to be able to go to court in that jurisdiction. Alternatively, the needs of the sound administration of justice may, in some situations, require special jurisdiction.

The rules on special jurisdiction are found in Article 7. According to Article 7(2), in matters relating to tort, delict or quasi-delict, a person domiciled in a Member State may be sued in another Member State where the harmful event occurred or may occur. The concept of ‘matters relating to tort, delict or quasi-delict’ covers all actions that seek to establish the liability of a defendant and that do not concern matters relating to a contract. ³¹ In these cases, the place of harm is the decisive factor.

It is important to note that different rules apply to matters relating to tort, delict or quasi-delict and matters relating to a contract. When harmful content is published online, it is usually done without reference to any contract. This is not to say that such a scenario could not occur. If a person makes an agreement with a newspaper on the publication of an interview, for instance, then the matter is naturally related to a contract and the rules for jurisdiction in the case of harm may need to be considered along those lines. ³² However, in the following I will only discuss issues where no contract exists and that therefore fall under Article 7(2): matter relating to tort, delict or quasi-delict.

The place of harm indicates the place of jurisdiction. When assessing online content, the place of harm is the problem. Where do online harms occur? Content can be accessed anywhere and harm may occur simultaneously in several locations. As the interpretation of the Regulation is based on the principle of predictability, it needs to be taken into account in the analysis of online infringements as well. It is a guiding principle of the whole Regulation and receives its expression in the sum of all of the articles, but from a systematic point of view, it is also a principle that guides the interpretation of the Regulation. ³³ Nevertheless, it may be that the predictability of jurisdiction can only be utilized when there is sufficient case law on the online context from the CJEU to lead the way.

The Court does not have an easy task developing consistent case law. The various rules and principles of the Regulation combined with the lack of rules on online scenarios make it tricky to form coherent criteria for application of the Regulation.

4. Doctrine: The mosaic approach and the centre of interests test

To deal with the issue of jurisdiction in cases concerning infringements of personality rights online, the CJEU has developed a two-part doctrine, which causes a certain amount of confusion. There is the mosaic approach and the centre of interests test. ³⁴

In Shevill, which concerned printed publications, the Court associated the place where the harmful event occurred with any Member State ‘in which the [allegedly harmful] publication was distributed and where the victim claims to have suffered injury to his reputation’. However, the courts of this Member State had jurisdiction only in respect of the harm caused in the State of the court seized. ³⁵

The idea is that a defamatory publication inflicts harmful effects upon a victim in all places where the publication is distributed if the victim is known in those places. Therefore, the courts in these Member States are territorially best placed to assess the libel committed in that State and to determine the extent of the corresponding damage. ³⁶ This solution produces a mosaic effect because many different Member States’ courts can have jurisdiction on different parts of the overall harm. It does presuppose, however, that the publication is distributed in that Member State in order for its courts to have jurisdiction.

Consequently, the Shevill judgment allowed two alternative jurisdictions from which the applicant could choose. One option would be the State of domicile of the defendant, where the victim may bring a claim in respect of the whole of the damage suffered. This is the main rule of the Regulation. The other option would be the State in which the publication is distributed and the victim is known, where a claim may be brought only in respect of damage caused in that State. ³⁷ This is called the mosaic approach to special jurisdiction.

The problem moving from a print publication to newer forms of media is that the ‘area of publication’ becomes global. It does not seem feasible for many Member States to have jurisdiction in the same case. In the eDate judgment, this issue became significant and the Court acknowledged the difficulties. The injury took place online. The Court put forward a new interpretation of Article 7(2), developing the centre of interests test. It stated that a person who has suffered an infringement of a personality right by means of the internet may bring an action in one forum in respect of all of the damage caused. The proper forum should be determined by the State in which the victim's centre of interests lies. This is because the impact that material placed online is liable to have on an individual's personality rights is best assessed by the court of the place where the victim has his centre of interests. ³⁸

Since eDate, these two ways of determining jurisdiction have both been applied. Accordingly, the victim of an online infringement of personality rights may choose between (a) the courts of one or several Member States in respect of the part of the damage that has been caused there, or (b) the court of the Member State where his or her centre of interests is based in respect of all of the damage caused. ³⁹ Of course, the victim may also utilize the main rule of the Regulation, according to which the courts of the Member State in which the defendant is domiciled have jurisdiction.

In online situations, the existence of two different approaches to special jurisdiction causes several problems, both practical and theoretical. The principle of predictability is particularly prone to suffer. Which approach should therefore be preferred? Let us first consider the mosaic approach. As AG Bobek has stressed, this approach makes it very difficult for the defendant to predict where he or she might be sued. ⁴⁰ In addition, it can also make it impossible for the claimant to obtain compensation for their entire damage anywhere other than at the defendant's domicile. If they wanted to avoid the courts in this country and take advantage of the rules of special jurisdiction, they would need to potentially bring actions in all other Member States to gain full compensation. ⁴¹

Moreover, the mosaic approach is also complicated from the point of view of courts. Even if the claimant only seeks monetary compensation, the courts face a real problem in assessing which part of the damage has been caused by the accessibility of the content in their jurisdiction. ⁴²

Summing up the argument so far, we notice that the mosaic approach can produce unpredictable and fragmented jurisdiction in online cases. ⁴³ This can lead to the chilling effect on publications on the one hand, but also to problems pertaining to access to justice from the point of view of the claimant.

However, the centre of interests test has its problems as well. It is difficult to ascertain what a centre of interests means in practice. The term is vague and its application unclear. The domicile of the person has been the most important indicating factor, but cannot be the only one. ⁴⁴ Therefore, it is hard to see how a defendant can predict jurisdiction on the basis of the centre of interests test either.

Nevertheless, the Court has maintained its dual approach to special jurisdiction. It has added some restrictions, however, most notably in Bolagsupplysningen. Only the courts of the Member State in which the claimant's centre of interests is located may hear claims for the rectification or removal of the allegedly harmful online content. Such a claim can only be brought before a court with jurisdiction to rule on the entirety of the claim. ⁴⁵ This is logical in itself but does not solve the main issue of two doctrines being applicable side by side.

5. Mittelbayerischer Verlag: An opening towards something new?

The Mittelbayerischer Verlag case offered the Court an opportunity to resolve some of the problems pertaining to jurisdiction. The dispute, however, was not an easy one to solve. What is particularly interesting is that the infringement in this case has no explicit link to a person. This poses a challenge for the centre of interests test. I argue, however, that because of the specific facts of this case, it proves very useful for rethinking special jurisdiction.

The case concerned a request for a preliminary ruling from Poland. The applicant, a Polish national, had been a former prisoner in Auschwitz. He brought a civil claim against a German newspaper before the Polish courts. The paper had used the expression ‘Polish extermination camp of Treblinka’ in an online article to refer to a Nazi extermination camp built on the territory of occupied Poland during the Second World War. The expression remained on the website for a few hours and was subsequently replaced with the following: ‘was murdered by the Nazis in the German Nazi extermination camp of Treblinka in occupied Poland’. A footnote to the article contains a short explanation that the phrase ‘Polish extermination camp of Treblinka’ was originally used in the text but had been corrected. The applicant maintained that the online publication harmed his national identity and dignity. He sought monetary compensation as well as other remedies including a court order prohibiting the publisher from using the expression ‘Polish extermination camp’ in the future.

The central questions were as follows. First, does the applicant enjoy protection of personality rights when the article does not mention him? Secondly, do Polish courts have jurisdiction to hear his claim considering that the newspaper is German and the article was published in Germany in German? The first question concerned the merits of the case and was not as such for the CJEU to decide, but, as we shall see in the discussion below, it is difficult to separate from the question of jurisdiction. This is because the centre of interests doctrine necessarily requires at least some features of the applicant to be taken into account.

This case differs from many others not least because the person alleging the infringement was not named in the publication at issue. Nevertheless, it appears that in Polish law, the personality rights of Polish nationals include the protection of their national identity, national dignity, as well as respect for the truth about the history of the Polish nation. Considering the national legislation, then, it would seem that personality rights of Polish survivors of Nazi extermination camps can be affected. This has been confirmed in the case of former prisoners of Nazi extermination camps, who can invoke an infringement of their personality rights if these camps are described as being ‘Polish’. Hence, the matter seems to be rather clear from a Polish point of view: personality rights can be infringed upon even when an individual is not mentioned by name. As EU law on personality rights is not harmonized but relies on the different meanings found in national legal systems, this becomes important in the case.

The defendant and the Commission nevertheless argued that the centre of interests test requires that the victim of an alleged infringement of personality rights has been named in the publication in question. Both saw it as a dangerous expansion of international jurisdiction if the centre of interests jurisdiction were available for undefined members of national, ethnic, religious, or other groups merely referred to indirectly in an online publication. ⁴⁶

The newspaper drew attention to its regional profile and readership. Its reporting covers the Oberpfalz (Upper Palatinate, Bavaria, Germany) and focuses primarily on regional news. By way of example, the heading ‘Germany and the World’ is only in fourth place on the page menu. The defendant also stressed that the website exists solely in German. ⁴⁷

The referring court wondered what the requirement of predictability of jurisdiction implies in the case. The expression ‘Polish extermination camp’ is likely to elicit a negative reaction in Poland. Could the defendant have reasonably foreseen that an action for the protection of personality rights might be brought against it before a Polish court even though the applicant was not described, directly or indirectly, in the text?

For the centre of interests doctrine, this is a particularly contentious case. What is such a centre for a claimant who was not individually identified by the publication? The centre of interests has usually been equated with the claimant's domicile, assuming that this is the place where their reputation will be most affected. In this way, the concept of the centre of interests has drawn on the subjective situation of the alleged victim. ⁴⁸

One very difficult question in this particular case is who exactly may reasonably claim to be a victim. This is primarily a matter of merits of the claim, set out by national applicable law. AG Bobek stressed that it is not an issue of jurisdiction under Article 7(2) of the Regulation. He therefore suggested that the Court adopt a narrow and minimalist approach in the case. ⁴⁹

The AG's stance is understandable, but it sidesteps the fact that in these cases, the national definition of personality rights will always have an impact on the standing of the parties, and therefore on the assessment of harm, as well as on the place of harm. The national definition of the right causes problems for a harmonized definition of jurisdiction. The two issues are linked, and as long as this remains so, it will be very difficult for a Court to completely disregard the merits of the case.

As for predictability, should the German publisher be surprised that the expression ‘Polish extermination camp’ might be received negatively in Poland? An excursion to ECHR case law may prove helpful here. In Lewit v. Austria, ⁵⁰ the ECtHR found that there had been a violation of Article 8. The case concerned a former prisoner of Mauthausen and a publisher who had published an article suggesting that the prisoners in that camp were criminals and not victims of the German occupation. The ECtHR found that, although the article did not mention the applicant's first and last name, he could nevertheless feel humiliated by the publication. The Court found that the Austrian courts had failed to protect the applicant's rights because they had never dealt with the central issue of his claim: that he had been defamed by an article that had used terms like ‘mass murderers’, ‘criminals’ and ‘a plague’ to describe people like him. Instead, the Austrian courts had concluded that he had no standing in the case as the number of people liberated had been so large that he could not have been personally affected by statements in which he had not been named. The lack of standing meant to the ECtHR that the applicant's rights had not been protected.

The issue of who has standing in Mittelbayerischer Verlag is crucially connected to the question of centre of interests. Therefore, the analysis of jurisdiction keeps slipping into a discussion that touches upon the merits of the case. Leaning on the jurisdiction developed by the ECtHR, the statement made in the German publication does concern the applicant, even if this is not directly apparent. It can be argued that the statement, because of its content, affects the applicant's personality rights. The Court did not share this view, but took the AG's advice and followed the minimalist approach.

6. The judgment

The Court did justify its judgment by utilizing the centre of interests test. The mosaic approach was not used. The Court ruled that the Polish court does not have jurisdiction to hear the dispute under Article 7(2) of the Regulation. This is because a court has jurisdiction to hear an action for damages brought by the applicant only if the content contains objective and verifiable elements that make it possible to identify, directly or indirectly, the applicant as an individual.

Hence, the Court put a lot of emphasis on the fact that the applicant was not named in the publication. It argued that the centre of interests doctrine is ill-suited to situations where a person is neither mentioned by name nor indirectly identified as an individual in the content. This may be so, but one can wonder whether the doctrine should therefore be amended.

Predictability became the leading principle on which the Court decided the case. It pointed out that the German publisher cannot have reasonably foreseen being sued before a Polish court, since the publisher was not, at the time when they put the article online, in a position to know the centres of interest of persons who were not in any way targeted by the content. ⁵¹

What is perhaps surprising is that the Court makes a specific reference to the person being identified as an individual. This solution echoes data protection law – which only protects identified or identifiable individuals – even though this is not a data protection case. ⁵² In the Court's own case law, personality rights have been understood as quite different from the right to personal data. Therefore, personality rights need not be interpreted in such individualist ways. The emphasis on an identifiable individual seems slightly out of place in this judgment.

The Court restated the main idea of the Regulation, according to which derogations from the principle of jurisdiction of the defendant's forum must be exceptional in nature and interpreted strictly. ⁵³ In a situation like the one in this case, it becomes relevant whether there exists a particularly close link between the dispute and the courts. This is intended to ensure legal certainty and to prevent the alleged infringer of personality rights from being sued before a court of a Member State that he could not reasonably have foreseen. ⁵⁴ The fact that a person belongs to a large identifiable group does not prove a close enough link. ⁵⁵

Even though the Court acknowledges in the judgment that the applicant considers that the expression which he contests constitutes an infringement of his personal rights, there is not a close enough link to the Polish court. Hence, there is no particularly close connection between the court of the place where the centre of interests of the applicant invoking his personal rights is situated and the dispute concerned. Therefore, the Polish court does not have jurisdiction. ⁵⁶

The Court's argumentation is unfortunately not very extensive, so the judgment does not provide quite as much clarity as one would have hoped. Many questions remain open, especially concerning the nature of internet publications as stretching beyond any location, as well as the exact criteria for when a person is identified as an individual in situations where a publication refers to a group. On these points, the ruling can be criticized.

What is left open is also the future of the dual nature of special jurisdiction. The Court does not utilize the mosaic approach – which is a good thing considering predictability and clarity – but does this mean that the approach is abandoned? In this case, it would have required a discussion of which different courts could hear different parts of the claim. It is evident that in cases like these, such a mosaic of jurisdictions is not prone to offer any practical solution either from the point of view of the newspaper or the applicant.

The more profound critique, however, should be directed at the Court for taking the individualist view. This move is not argued for in any detail, which is a shame. It does, however, take the broad traditional understanding of personality rights into new directions, ones that would need closer scrutiny. Regardless of the minimalist aspirations, it means that the Court does go into the merits of the case because it effectively disagrees with the Polish view of personality rights where a person does not have to be named in the publication. I argue that a more objective stance – one that would not only pay attention to the close link of the individual and the place – could be more fruitful in the future.

7. The objective approach

Both AG Bobek and AG Cruz Villalón ⁵⁷ have emphasized the importance of foreseeability, proposing an objective assessment of it. They envisage a kind of foreseeability that is ensured by identifying the centre of gravity of the dispute itself, not only the alleged victim's centre of interests. Such a centre of gravity would be composed of two cumulative elements, one focusing on the claimant and the other on the nature of the information at issue. The courts of a Member State would have jurisdiction only if that were the place of the claimant's centre of interests and if the information at issue was expressed in such a way that it may reasonably be predicted that the information is objectively relevant in that Member State. ⁵⁸

AG Bobek concludes in his opinion in Mittelbayersicher Verlag that on the basis of the facts, it is difficult to suggest that it would have been wholly unforeseeable for a publisher in Germany, posting the phrase ‘the Polish extermination camp of Treblinka’ online, that somebody in Poland could take issue with such a statement. He argues that, while it is ultimately for the national court to examine these issues, it is difficult to see how special jurisdiction under Article 7(2) of the Regulation could be axiomatically excluded. ⁵⁹ This argument is persuasive.

Therefore, making an objective analysis of the dispute itself and its centre of gravity, it could have been argued that the Polish court has jurisdiction to hear the case. This conclusion has two specific merits. Firstly, it does not make jurisdiction dependent on whether an individual has been explicitly named in the article or not. If such a requirement were introduced, then many potential infringements would be excluded from any court's jurisdiction, which would make the future assessment of such cases highly problematic. Secondly, focusing on the dispute itself, taking into consideration both the alleged victim and the nature of the content, the conclusion does include a move away from analysis of only the applicant's centre of interests, which usually points to domicile.

Even though the Court does agree with the general idea of an objective approach, it draws different conclusions. It argues that predictability requires a close link. According to the Court, the link cannot be evaluated solely by subjective elements, and it cannot be dependent on a person's individual sensibilities, but on objective and verifiable elements. These elements should make it possible to identify, directly or indirectly, a person as an individual. ⁶⁰ It could be argued that in this case, however, the person could be identified indirectly as a survivor of the camps. The content was expressed in such a way that it is plausible for a Polish survivor to be affected. This was not, however, enough for the Court.

Even though the Court claims to be striving at an objective assessment, its argumentation lacks attentiveness to the centre of gravity of the dispute itself. It is unfortunate, in this regard, that the Court did not take the road proposed by the Advocates General. The Court's view that stresses the need to identify an individual is a solution that does not focus on the dispute nor the content of the publication.

An assessment of the centre of gravity of the dispute could be a better way to go because it would not only focus on the applicant's situation, but also the nature of the publication and its content. This solution would undoubtedly imply that a degree of uncertainty would remain, but the legal assessment of special jurisdiction would include analysis of more factors than merely the centre of interests of the applicant.

Taking the centre of gravity of the dispute as a starting point would increase foreseeability, at least to some extent. However, the victim of an online infringement of a personality right would still be able to choose between several places of jurisdiction. They may always resort to the main rule of the Regulation and seize the courts of the Member State in which the defendant is domiciled. In addition, the victim could go to court in the Member State where the centre of gravity of the dispute is located. This court would have jurisdiction to rule on the entirety of the claim.

The problem with an objective approach such as this one is admittedly that when considering personality rights, some subjective elements are always likely to sneak into the assessment. There would of course not be any claim to begin with unless somebody took offence, and offence taken is always subjective. Different things offend different people. For instance, expectations of privacy vary and are culturally dependent.

Moreover, even though the assessment would be made on an objective level considering both the applicant and the nature of the publication, uncertainty will always remain about jurisdiction as long as there is no definition of personality rights. Predictability may suffer for this reason. The defendant cannot reasonably foresee what counts as an infringement of a personality right and which persons can have standing in the case as long as interpretations on personality rights depend on national laws. This problem becomes very clear in Mittelbayerischer Verlag. It is, however, extremely difficult to resolve in EU law because the Union does not have any clear competence on the basis of which it could harmonize personality rights. Therefore, the Court's decision to push the interpretation of personality rights in the direction of personal data rights in Mittlebayerischer Verlag is not an optimal one.

8. Conclusions

Questions concerning special jurisdiction are generally resolved by reference to the principle of predictability. This article has asked what this principle entails in an online environment. What is the predictable place of harm for online infringements of personality rights when the internet is limitless?

The existence of two ways of determining special jurisdiction, the mosaic approach and the centre of interests approach, causes a number of problems from the point of view of predictability. In the current situation, the claimant can seize the court of the Member State where the defendant is domiciled (the main rule), or some other court either by applying the mosaic approach or the centre of interests test. This is not ideal because the number of jurisdictions to choose from becomes large, possibly encompassing all Member States. This can lead to the chilling effect on publications on the one hand, but it also causes problems pertaining to access to justice for the claimant.

In the above, I have sought to address these issues. Firstly, when considering special jurisdiction in online cases, it is time to move away from the mosaic approach, even though the Court does not explicitly refuse it in Mittelbayerischer Verlag. The mosaic makes it difficult for the defendant to predict where they might be sued because many different Member States’ courts can have jurisdiction on different parts of the overall harm. The mosaic approach can also render it impossible for the claimant to obtain compensation for their entire damage anywhere else than at the defendant's domicile. In addition, the mosaic approach causes a headache for national courts. They would need to calculate which part of the damage has been caused by the accessibility of the content in their jurisdiction. Hence, in order to improve predictability, I argue that the mosaic approach should be abandoned.

The centre of interests test is better suited to define special jurisdiction. However, in line with the arguments by the Advocates General discussed above, the test should be developed towards an assessment of the centre of gravity of the case, taking into account both the applicant's situation and the nature of the publication. A centre of gravity test can work in a satisfactory way even when a person is not named or otherwise directly identified in the publication. It can only work, however, if it includes a holistic analysis of the content of the publication as well as the situation of the claimant.

The result of such an approach is, however, that there is a risk that factors pertaining to the merits of the case infiltrate parts of the analysis. This is unfortunate but cannot always be avoided in these types of cases regardless of which test is used.

A Member State court will naturally base its assessment of jurisdiction at least partly on the way personality rights are defined in that specific Member State. Each Member State can have different rules on what personality rights are, how they are protected, and whom they protect. Therefore, it is even more crucial, as I see it, to clarify the application of the Brussels Ia Regulation on online infringements. The rules on international jurisdiction may sometimes be at odds with different understandings of personality rights, which becomes clear in the Mittelbayerischer Verlag case. Here, an individualized view on jurisdiction does not fully accommodate views on personality rights that are less individual-centred.