Reputation in European trade mark law: a re-examination

Under the harmonised European trade mark regime marks with a reputation enjoy expanded protection. This article casts doubt on whether this ‘reputational trigger’ can be justified. It then explores some difficult operational questions about the way the reputation threshold works in cases where the mark enjoys fame only in niche markets or in a limited geographical area, the aim being to illustrate further why reputation is an unsatisfactory trigger for a different type of trade mark protection. Finally, it looks at some of the evidential difficulties involved in adjudicating disputes in which expanded protection is being claimed. It concludes by suggesting that if the evidential problems we identify were tackled the reputation threshold could be abandoned.


Protection of marks with a reputation
The central aim of European trade mark law is to prevent consumers being confused or deceived as to the origin of goods or services. Consequently, a trade mark will be deemed to have been infringed in a case where consumers are likely to be misled into believing that the defendant's goods or services were produced or provided by the trade mark owner. This reflects the traditional understanding of the nature of trade mark rights. Importantly, however, European law provides an additional layer of protection for trade marks with 'a reputation'. This was the case under the original instruments that set up the European trade mark regime, 1 and is continued in the new framework created by Directive (EU) 2015/2436 2 and Regulation (EU) 2015/2424. 3 Once a mark has acquired a reputation, the owner is afforded protection against three additional types of harm that are sometimes grouped under the single rubric of 'dilution'. Specifically, the owner is afforded protection against so-called 'blurring', 'tarnishment' and 'free-riding'. Protection against 'blurring' is aimed at preventing the dispersal of the identity of the mark as might occur if a well-known mark began to be used in relation to unrelated goods or services. In the language of the EU trade mark regime this is referred to as use that is 'detrimental to the distinctive character of the mark'. Protection against tarnishment serves to prevent a mark from being linked with goods or services that are in some sense unwholesome or that might create unfortunate associations in the minds of consumers. This is referred to under EU law as use that is 'detrimental to the repute of the mark'. Protection against free-riding is designed to prevent the use of a mark in a manner that takes advantage of the investment that a trade mark owner makes in developing its brand. This is reflected in the provision of EU law that prevents the use of a sign that 'takes unfair advantage of the distinctive character or the repute of the trade mark'. These extended protections are afforded to owners of marks with a reputation in two contexts: first, where the owner is claiming that its mark has been infringed; 4 and second, as grounds that serve to block the registration of later, similar, marks. 5 As Advocate General Jääskinen has noted, the consequence of the extended protections against dilution is to transform trade mark law from a 'deception-based' 1 4 1989 Directive, Art. 5 (2) and 2015 Directive, Art. 10(2)(c); CTMR, Art. 9(1)(c) and EUTMR, Art. 9(2)(c). 5 1989 Directive, Art. 4(4)(a) and 2015 Directive, Art. 5(3)(a); CTMR, Art. 8 (5) and EUTMR, Art. 8 (5). regime to a 'property-based' regime. 6 These extended protections have been controversial, not only in Europe, 7 but also in other jurisdictions where some of these harms have been recognised. 8 The question that has been asked much less frequently, however, is why 'reputation' serves as the trigger for the shift towards a property-based regime. That question forms the focus of this article. We start by problematising the idea that it is only marks that enjoy a reputation that warrant expanded protection. This is not to suggest that we would wish to make it easier for trade mark owners to rely on the expanded form of protection, but if we are to achieve intellectual coherence within the trade mark system, difficulties associated with the reputation trigger must be confronted. We then turn to address some particularly difficult operational questions about the way the reputation threshold works in cases where the mark enjoys fame only in niche markets or across a limited geographical territory, with the aim of further illustrating why reputation is an unsatisfactory trigger for a different type of trade mark protection. Finally, we ask questions about how reputation maps on to the ways in which consumers respond to trade signs and, relatedly, some of the evidential difficulties involved in adjudicating disputes in which expanded protection is being claimed. We conclude by suggesting that if the evidential problems we identify were tackled the reputation threshold could be abandoned.

Reputation and functions analysis
In order to understand why the idea that it is only marks with a reputation that warrant expanded protection might be problematic, it is first necessary to appreciate the model of trade mark protection that the CJEU has developed. Consistent with the traditional understanding of the nature of trade mark rights, the CJEU has said repeatedly that the 'essential function' of the European trade mark regime is to safeguard the communication of accurate information about the trade origin of goods or services. 9 This explains why protecting trade marks against uses that might cause consumers to be confused as to the trade source of the products in question remains at the centre of EU trade mark law. The Court has, however, also long emphasised that other social and commercial functions that trade marks might perform enjoy a degree of legal protection. Specifically, the Court has held that the functions of 'guaranteeing the quality of the goods or services in question and those of communication, investment or advertising' are also protected. 10 The nature of these other functions has been articulated almost exclusively in the context of so-called 'double identity' cases. In the infringement context, these are cases where the defendant is using an identical mark to the claimant's registered mark in relation to identical goods or services. 11 Perhaps counter-intuitively, in these cases the articulation of these other functions serves to limit the scope of trade mark rights. This is because the double identity provisions appear to set up a form of absolute protection for trade mark owners. Subject to the operation of any defence (for example, to allow comparative advertising), liability appears to be strict, in that there is no need for the claimant to demonstrate that it has suffered harm of any description. However, starting with its decision in L'Oréal v Bellure, the CJEU has held that defendants will escape liability in cases where although they have used an identical mark in relation to identical goods or services, the use is not of a type that is likely to harm any one of the protected functions. 12 By way of example, in its 2010 decision in Google France, the Court indicated that unauthorised use of a trade mark as a keyword in online advertising would not give rise to liability unless it adversely affected one of the functions. 13 The CJEU's functions analysis has been the subject of sustained criticism from both trade mark scholars 14 and national courts. 15 One of these criticisms is that the Court has failed to articulate the way in which the legally protected functions map on to the 'dilution' provisions. The only clear guidance that the Court has provided is that 'free-riding' has no place within the double identity provision. 16 Nevertheless, 'blurring' and 'tarnishment' must be understood as serving to protect (albeit in some way that has yet to be fully articulated) the other functions. 17 Indeed, it has been pointed out that for the purposes of the double identity provision, it is unclear when, if at all, the functions would be damaged other than through confusion, blurring or tarnishment. 18 The link between the functions and dilution harms is significant because when one turns to approach issues of, say, blurring and tarnishment by reference to the functions, it is not at all clear that the subsistence of 'a reputation' can sensibly serve as a pre-condition for the establishment of liability. Taking the advertising function by way of example, the CJEU has told us that this function may be harmed by any use that 'adversely affects the proprietor's use of its mark as a factor in sales promotion 11 See 1989 Directive, Art. 5(1)(a) and 2015 Directive, Art. 10(2)(a); CTMR, Art. 9(1)(a) and EUTMR, Art. 9(2)(a). 12  or as an instrument of commercial strategy'. 19 This points towards a broad sphere of protection, and it is not at all obvious why such protection should be confined to marks with a reputation, other than in double identity situations. Indeed, the very fact that the advertising function is protected for all trade marks in cases of double identity must tell us that it is not only marks with a reputation that can be used as instruments of commercial strategy. Similarly, if one turns to the investment function, there are situations where a trade mark owner's investment might be significantly compromised even in the absence of the mark enjoying a reputation among the public. Admittedly, any attempt to analyse the investment function is complicated by the fact that this function is particularly opaque. Nevertheless, the most convincing explanation for this function is that provided by Advocate General Poiares Maduro in Google France, who indicated that the investment function serves to promote 'innovation and commercial investment'. 20 On this reading of the function at least, there are unquestionably situations where a third party's use might compromise brand investment even in the absence of the brand having a reputation. Take, for instance, a defendant that deliberately sets out to undermine a competitor's planned product launch (in which there is likely to have been significant investment) by using an identical mark on dissimilar, possibly unwholesome goods. Such a use would undermine the mark owner's investment but would not be actionable under the dilution provisions of EU trade mark law. The incoherence of linking extended protection for the investment function to reputation is, of course, further reinforced by the fact that for the purposes of the double identity provisions, the subsistence of a reputation is not a pre-condition.
One objection to our analysis above is that it is based on an interpretation of the relationship between the functions and the harms of dilution that has not yet been confirmed by the CJEU. This does not, however, mean that the 'reputation' threshold is unproblematic, even if the harms of dilution are approached on a stand-alone basis without any reference to the functions they might be said to protect. In other words, it is not at all clear that antidilution harms should be confined to marks with a reputation, even if one sets functions analysis aside. We maintain that this is true for both blurring and tarnishment, but is most easily illustrated by reference to blurring.
To elaborate on the immediately preceding point, protection against blurring is said to prevent the dispersal of the identity of the mark, as might occur if a wellknown mark began to be used in relation to unrelated goods or services. In the famous example developed by Frank Schechter, the first advocate of antidilution protection, protection against blurring is necessary because the unchecked use of a famous mark, such as Rolls-Royce, on unrelated goods such as restaurants, cafeterias and clothing, will lead to the eventual destruction of the Rolls-Royce mark as an effective commercial identifier. 21 If, however, one investigates the harm that is being posited in this type of example, one can see that the underlying concern is that use of a similar mark on dissimilar goods and services can substantially inhibit the ability of the trade mark to 'reach out' to consumers. To our mind, however, it is counter-intuitive to suggest that this type of harm should only be a cause for concern in cases where the mark enjoys a reputation. Such marks are, by definition, already firmly fixed in the public mind and the risk that third party uses will diminish consumers' ability to recall the mark and its associations must be relatively limited. In contrast, in cases where the mark has not yet established a reputation, the presence of unauthorised uses of the mark on unrelated goods might have a much more significant impact on consumer recognition of the mark. Moreover, one might bear in mind that the principal harm that blurring seeks to prevent is often understood as being 'progressive' in nature. It is not so much that the single use of, say, 'Rolls-Royce Café' will harm the brand, but more that many such uses might collectively lead to the eventual destruction of the mark. Thus, there is a significant forward-looking component to blurring cases-we are worried about a particular defendant's behaviour because, although the harm is still distant, it is thought to be inevitable if the defendant's conduct were to be replicated and left unchecked. If all blurring cases have this forward-looking component at their core, the case for protecting marks without a reputation at the time of the action is still more compelling. We return to consider some of these issues when addressing the evidential difficulties associated with reputation and antidilution protection.

Problems associated with limited reputation
The suggestion in the previous section that reputation is not a coherent device that can be used to trigger stronger trade mark rights is, admittedly, controversial. There are, however, much more widely acknowledged concerns about how the reputation threshold operates in practice in EU trade mark law. Two important concerns relate to circumstances in which the owner can only demonstrate a reputation amongst a limited group of consumers (sometimes known as 'niche fame'), or where a Community trade mark has a reputation only in a limited geographical area. Analysis of these issues helps reinforce the general problems with the reputation trigger discussed above.
In order to understand the two specific issues with which were are concerned in this section, it is first necessary to say something about how the CJEU has approached the issue of what trade mark owners need to prove in order to establish reputation in a mark generally. The leading authority remains the Court's 1999 decision in General Motors v Yplon, where it was said that a decision-maker must take into account 'the market share held by the trade mark, the intensity, geographical extent and duration of its use, and the size of the investment made . . . in promoting' the mark. 22 Significantly, in considering the level of reputation required, the decision-maker must consider whether the mark is known by 'the public concerned by the products or services' covered by the registration. It is only amongst this group that the mark must be recognised by a 'significant' proportion of consumers. 23 The consequence of reputation being approached in this way is that antidilution protection is not confined, as it is now in the United States, to marks that enjoy a reputation among the general consuming public. 24

Niche fame
Recognising the possibility of 'niche fame' raises the question of how a tribunal ought to proceed in a case where a trade mark is very well known but for a product where the total market is very limited. The sort of case we have in mind can be illustrated by reference to the 2014 decision of the Australian Trade Marks Office in Argenta Ltd v Argenta Discovery 2009 Ltd. 25 In that case the mark owner could only establish a reputation among a few hundred people in Australia. However, given the highly specialised nature of the services in question (consultancy and research services in the field of animal health products and veterinary pharmaceuticals), those with knowledge of the opponent's mark formed the entire market for the services. There can be little question that under the Yplon formulation, such a mark would be deemed to have 'a reputation' under European law. However, there seems to be a consensus that it would be undesirable to confer any significant degree of antidilution protection on such marks. The reasons for taking this view have not always been properly articulated, but it is possible to discern two, rather different concerns at play.
The first concern is that it would seem somewhat arbitrary to confer antidilution protection on marks that enjoy significant niche fame while denying such protection to marks that have a larger, but more diffuse customer base. For example, it might seem unfair to protect a mark known only to a few hundred consumers of animal health products and veterinary pharmaceuticals, whilst not affording protection to the owner of a mark enjoying a larger number of total consumers, but for goods such as clothing, where the relevant market is the general consuming public. 26 Admittedly, this problem is mitigated by the fact Yplon sets up a relatively low threshold for reputation. But, nevertheless, a degree of tension remains. The second concern is that antidilution protection is more palatable if the reputation of the earlier mark is taken to put the junior user on notice of the existence of the senior user's rights. One of the advantages of the trade mark registration system is that the register affords traders the opportunity to determine whether it is safe to use a trade sign by consulting the register. While there is reason to be sceptical, in general terms, about the effectiveness of the register in providing useful information to traders, 27  creates an additional set of concerns in that the register will never provide an indication of whether a mark enjoys a reputation, such that the mark is entitled to the higher level of protection. This may be thought to be a matter of limited concern if the reputation of the mark in question is sufficient to displace the need for a publicly accessible record of the mark's scope of protection. For example, a trader ought hardly to need to look at the trade marks register in order to determine that it would be a risky business strategy to use NIKE as a mark for ashtrays. Whether this is a convincing response to the notice problem is, in any event, open to doubt, given the low standard established by Yplon. But the 'reputation as adequate notice' argument is clearly unconvincing, when one turns to marks with very niche fame.
Consequently, it is therefore understandable that tribunals and commentators have been keen to seek to find ways of avoiding the outcome that marks with very niche fame attract the benefits of broader legal protection. Drawing on the CJEU's decision in Intel v CPM, 28 Ilanah Simon Fhima has argued in her leading comparative text on trade mark dilution that a way around the niche fame problem can be found in the requirement that consumers draw a 'link' between the senior and junior marks. 29 The argument is that such a link is unlikely to be made where there is little, if any overlap, between consumers of the senior user and junior user's goods. Subsequent decisions of OHIM also point in this direction. 30 There are situations where the requirement of a link might convincingly explain why, in practice, owners of marks with niche fame would struggle to make out dilution. By way of example, if the mark ARGENTA only enjoys fame amongst researchers into veterinary pharmaceuticals, use of ARGENTA on products targeted at dentists would be unlikely to cause a link to be made with the senior ARGENTA mark. In other cases, however, the suggestion that the 'link' requirement produces the desired outcome seems much less convincing. Here, it is important to be wary of a slip that can occur between how we understand harm in cases where confusion is being alleged and how we conceptualise harm in antidilution cases. In cases where confusion is being alleged, we are concerned with the mental state of consumers of the junior user's goods. For example, assume that the mark MATTONI is registered for mineral water and has been extensively used on such goods for more than a century. If a junior user were to start selling MATTONI confectionery, liability would turn on whether purchasers of the confectionery would think that these goods were made by, or affiliated with, the mineral water producer, taking account of the reputation of the senior MATTONI mark. But in cases where dilution is alleged, to focus on the responses of consumers of the junior user's product would be misplaced. The harms against which antidilution laws are designed to protect are that consumers of the senior user's goods would think differently about the senior user's mark. Logically, therefore, this must mean that the focus should be on the senior user's consumers' responses when they come into contact with the junior use. This can be most readily illustrated in the context of cases in which tarnishment 28 Case C-252/07 Intel Corporation Inc v CPM United Kingdom Ltd [2008] ECR I-8823. 29 Fhima [6], pp. 30-31. 30 See, e.g., Case R724/2009-4 Repsol YPF SA v Resinas Olot SA (Fourth Board of Appeal, 29 April 2010) (finding no link would be made between a 'Repsol' device mark, which had a reputation only in relation to the distribution of energy services, and a 'Resol' device mark for goods such as indoor and outdoor furniture, textiles, carpets, and leather goods). is being alleged. Consider, for example, the mark TIFFANY'S used on condoms. In such a case, when determining whether the defendant had caused detriment to the repute of the TIFFANY'S mark, the focus would fall on whether consumers of Tiffany's jewellery would, when confronted by the condoms, draw a link with the senior user's mark 'in such a way that the trade mark's power of attraction is reduced'. 31 Similarly, if ARGENTA condoms were to be produced, it would be the impact on veterinary pharmaceutical researchers who encounter the condoms that would be relevant, and not the attitude of purchasers of condoms in general.
To build on a point we made in Part 1, the above should not be taken to indicate that we would be in favour of giving owners of marks with niche fame more expansive protection. Rather, our aim in considering niche fame is to show that it gives rise to problems that further illustrate the conceptually unsatisfactory way in which 'reputation' operates as a trigger for a different form of legal protection for trade marks.

Limited geographical reach
A second feature of the Yplon test for the subsistence of reputation is that the reputation must exist in a 'substantial part' of the territory. 32 In the case of Community trade marks, it was confirmed by the CJEU in PAGO v Tirolmilch that the reputation must subsist in a substantial part of the EU. 33 As with cases involving fame in a limited product market, the application of the PAGO standard is widely understood to raise some difficult questions of interpretation. One issue that has attracted particular attention is what happens in a case where reputation is confined to a single Member State. In PAGO itself, the Court held that there was no requirement of a cross-border reputation, and that on the facts of the case, a reputation in Austria alone was sufficient. While it has not been conclusively determined that reputation in a single Member State is always sufficient, it is doubtful that the Court would ever deem the territory of a Member State to be less than a substantial part of the EU.
The other question that has attracted significant attention, and the one that is more important for present purposes, concerns the relationship between the relief available and the extent of the claimant's reputation. The starting point is that the unitary nature of Community trade mark rights means that 'a prohibition against further infringement or threatened infringement issued by a competent Community trade mark court must therefore, as a rule, extend to the entire area of the European Union'. 34 However, in DHL v Chronopost the CJEU held that if the claimant's reputation is confined to a particular geographical area, it might be appropriate to limit the scope of an injunction to the areas in which the reputation subsists. 35  a given territory are not familiar with a trade mark there can be no risk of confusion, and nor will any other function be implicated. Significantly, the Court appears to have sounded a partial retreat from some of the possible implications of this position. The Court has now held in its 2015 decision in Iron & Smith v Unilever, a case involving the scope of the antidilution provision of the Community Trade Mark Regulation, that it would be enough that 'a commercially significant part' of the consuming public in a Member State in which the mark owner did not have a reputation might recognise the mark-there was no need to demonstrate as a threshold matter a reputation among 'the relevant public' in that Member State. 36 In other words, although the Court did not seek to limit the effect of DHL by suggesting there was no need to demonstrate harm to one of the functions in order to secure injunctive relief, it nevertheless made it easier to secure such relief by lowering the reputational bar. The consequence of this aspect of Iron & Smith, which appears to have gone unnoticed, is that the Court has now acknowledged that antidilution harms are capable of arising in circumstances where the mark owner enjoys less than the level of reputation needed to satisfy the Yplon threshold. To our mind, this is the strongest illustration thus far of the difficulties and pitfalls of attempting to limit antidilution harms to marks with a reputation. The Court seemed to be animated by a recognition that a trade mark owner can be harmed by the fact that it might in the future seek to enter a market where the distinctive character of its mark has been compromised in advance. This comes close to judicial acknowledgement of the proposition that the type of harm against which blurring is designed to prevent is not confined to marks with a reputation.

Reputation, consumer responses and evidence thereof
The final two, related issues we wish to consider concern, first, how consumers interact with trade marks that enjoy a reputation and, second, how the law affords weight to evidence of likely consumer responses to such marks.
The legal starting point is the CJEU's decision in Adidas-Salomon v Fitnessworld, where it was held that a pre-condition for liability to arise under the antidilution provisions of the European trade marks regime is that the relevant section of the public establish a 'link' between the senior and junior mark. 37 As Bently and Sherman explain, 'without the link, there can be none of the "cross-pollination" needed to transfer value to the . . . defendant'. 38 The subsequent focus of courts and commentators has been on the various factors that need to be taken into account in determining whether such a link exists. For example, in Intel, the CJEU devoted much of its focus to identifying a set of factors that could be used to identify when consumers might make the necessary connection. 39 However, there has been relatively little engagement with the states of mind of consumers confronted with the impugned junior uses, and no serious attempt to bring back into the analysis how the reputation of the senior mark might impact on these states of mind, even though reputation is the trigger for potential liability.
The complexity of any serious assessment of the relationship between reputation and consumer responses is best illustrated by considering how the subsistence of a reputation impacts on the likelihood of confusion. In Europe, the position appears to be that the presence of a reputation invariably makes it easier for the mark owner to establish confusion. 40 This means, for example, that Nike's prospects of bringing a successful action against a retailer of MIKE sporting goods would be increased because of the fame of the NIKE mark. This conclusion does not, however, rest even on a convincing explanation of how consumers would respond to the MIKE goods, let alone on hard empirical evidence. Importantly, moreover, other jurisdictions have rejected the idea of a unidirectional relationship between reputation and consumer confusion. This is best illustrated by the position in Australia, where it has been said that '[r]eputation in a trade mark can . . . be a double-edged sword and may either reinforce or mitigate the effect of differences'. 41 Where the mark has a substantial reputation, this has been held to work against the owner insofar as the junior user's goods or services are the same as those provided by the trade mark owner and there are clear and obvious differences between the marks. Consumers will immediately notice the differences between the marks, with reputation serving to displace any danger of the senior mark being misremembered. Provided the goods are essentially the same they will be understood to be competing products rather than a brand extension. The implications of this approach were most fully realised in a case in which Mars, the owner of a famous 'Maltesers' device mark for confectionery, was unable to prevent the sale of a similar product called 'Malt Balls' in colours and packaging reminiscent of that used by Mars. The judge accepted that 'consumers are so familiar with Maltesers that they could not possibly be confused by the Malt Balls packaging-more formally, there is no likelihood of imperfect recollection by them of the Maltesers mark leading to confusion'. 42 Under Australian law, reputation will only work in favour of a finding of confusion where the junior user's goods or services might be treated as an extension of the trade mark owner's business. In particular, tribunals have recognised that owners will sometimes adopt a brand variant when extending their product lines and, consequently, that consumers may well be confused when they are reminded of an earlier famous mark on products that could conceivably be an extension of the owner's business. 43 While the Australian approach also lacks a firm empirical base, it at least rests on an intuitively convincing explanation of how the fame of a mark is likely to impact on consumer decision-making (that is, as has been seen, that the fame of the mark will displace any danger of confusion that might otherwise be caused by imperfect recollection of the senior mark). No-Tox products as complementary in the same way as "coke" and "diet coke" might be seen as complementary products').
Moving to antidilution cases directly, insofar as empirical evidence is available, this suggests that the relationship between the subsistence of reputation and the drawing of a link between marks is highly complex. For example, in Maureen Morrin and Jacob Jacoby's well-known study, in which they sought to establish whether dilution by blurring can occur, it was discovered that in the case of some very famous marks, such as 'Hyatt' hotels, the introduction of a potentially dilutive stimulus, such as 'Hyatt' legal services, actually served to reinforce recognition of 'Hyatt' as a hotel brand. 44 Furthermore, empirical work on authorised brand extensions suggests that consumers do a much better job than might be expected in confining any negative reactions they might have to problematic extensions. 45 For example, Henrik Sjödin & Fredrik Törn have argued that although a brand extension that is thought to be consistent with the values of the core brand can result in positive feedback effects among consumers, where an authorised brand extension is perceived as being incongruent with core brand beliefs, any negative evaluation of the extension will not affect consumers' reactions to the core brand. 46 If this holds true for authorised brand extensions, this must seriously call into question whether consumers' attitudes towards famous marks such as 'Tiffany's' and 'Coca-Cola' will be affected by being confronted with these marks or variants used in relation to dissimilar or unwholesome goods.
More recent research into consumer psychology has emphasised the importance of a background level consumer misunderstanding created merely by having healthy competition for the goods and services in question. The entry of a new player into a product market will cause consumers to mix up products and draw associations between entirely dissimilar brands. Some consumers will simply misremember a Visa advertisement as an advertisement for MasterCard, leading to some transfer of investment, and potentially reputation or prestige, in circumstances of which the law could not possibly take cognisance. 47 A thorough analysis of how evidence of reputation impacts on the sorts of links consumers make would also need to take into account historical examples that suggest that the types of consumer reaction assumed by antidilution advocates are not inevitable if, indeed, they ever arise. Writing in this journal, Simon Malynicz has pointed out that two unrelated entities have been using the mark 'Rolls-Royce' for more than forty years without the Rolls-Royce name suffering any obvious loss of prestige or distinctive character. 48 Also relevant is the strange history of Danish pop group Aqua's famous song 'Barbie Girl'. Shortly after the song's release in the US, Mattel brought an action against Aqua's record label alleging that the song diluted the 'Barbie' mark. In the Court of Appeals for the Ninth Circuit, Judge Kozinski held that this was 'a classic blurring injury' (in that consumers would now associate 'Barbie' with two different products), 49 albeit that, on the facts, Mattel was eventually un- 44 Morrin/Jacoby [16]. 45 See generally Tushnet [22], pp. 543-544; McKenna [15], pp. 104-105. 46 Sjödin/Törn [21]. 47 Humphreys et al. [10]. 48 Malynicz [14], p. 54. 49 Mattel, Inc v MCA Records, Inc, 296 F 3d 894, 904 (9th Cir, 2002). successful in its action. But whatever we may think of Judge Kozinski's acceptance at the time of this being a case in which blurring was likely, this conclusion now looks highly suspect. In an ironic twist, Mattel, recognising that for a sizeable group of consumers the Aqua song actually served to promote the Barbie brand (the sexualised and satirical lyrics apparently having been overlooked), have now produced a licensed, sanitised version of the song. 50 To be fair to the CJEU, it must be acknowledged that in its judgment in Intel the court seem to acknowledge the evidential difficulties that can arise in antidilution cases. Subsequent decisions have, however, been much less consistent in confronting these difficulties. Moreover, at no point has the Court begun to grapple with the broader and still more complex question of how reputation impacts on consumer reaction to brands and the difficulty of leading probative evidence of when consumers will draw an association or 'a link' between marks. Here it is necessary to say something about the 'free-riding' limb. It is in relation to this head of damage in particular that there is real uncertainty about the evidential requirements, and the cases have too often proceeded on the basis that some advantage from use of similar sign can be presumed. But this glosses over the nature of consumer response and distracts attention from both the question of when the defendant will gain an advantage from creating a link in the minds of consumers and the question of when any such advantage is unfair, as the European legislative framework requires. If one were to approach the unfair advantage question on the basis of a proper analysis of likely consumer response a rather different picture might emerge. Take, for example, the facts of L'Oréal v Bellure. The defendant's aim in using the claimant's trade marks in comparison tables was to communicate that it was producing perfumes that smelt similar to the L'Oréal originals. The information provided in the tables was accurate and one of the most concerning aspects of the case is that the CJEU provided no indication of the nature of the advantage conferred on the defendant, let alone how any advantage might be said to be 'unfair'. Yet there is clearly a need to identify a reaction that is different from, say, legitimate comparative advertising. If the response of consumers to seeing L'Oréal comparison lists is much the same as when they encounter a comparative advertisement-if the owner cannot demonstrate that some additional or alternative mental process is at play-then a finding of unfair advantage ought never to be appropriate.

Conclusion
The problems we have identified are not confined to the harmonised European trade mark regime-other jurisdictions are also struggling with how to make protection against dilution work in an intellectually coherent manner. However, this is perhaps in no small part because other jurisdictions also provide for a recognition threshold. Irrespective of whether this threshold is expressed in terms of the mark having a reputation, as in Europe; or the mark being famous, as in the United States; or the mark being well-known, as in New Zealand, 51 the problem remains that the harms 50 See further Burrell/Gangjee [4] and the sources cited therein. that antidilution actions are intended to protect against are not confined to marks with a reputation. For anyone who is concerned about the potential overreach of trade mark law there is, of course, a significant political risk to problematising the reputation threshold, in that this would seem to point to the need for stronger protection for all marks. To our mind, however, this is only a matter for concern if we set aside what we know about consumer psychology-the available evidence seems to be that blurring and tarnishing harms are rare. While a good deal of uncertainty remains as regards the operation of free-riding, if one takes the notion of unfair advantage seriously there is also reason to conclude that the rejection of the of the reputation threshold would not lead to a dramatic expansion in the number of successful antidilution actions, provided the evidential requirements are taken seriously.
Indeed, it is at least arguable that by confining antidilution protection to marks that enjoy public recognition we are actually facilitating antidilution actions, including in situations where none of the harms that such actions are intended to prevent are remotely likely. This is because antidilution cases can, as matters stand, turn on the level of recognition that the mark enjoys. By opening antidilution actions up to all marks the emphasis would have to fall on the more appropriate question of how consumers are likely to respond to the specifics of the junior use. This would require a much greater willingness to engage with questions of consumer psychology. There can be no question that such engagement poses its own challenges, but is essential for the development of a more coherent trade mark system.
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.