Professor Weatherall’s thought-provoking critique of the selective resistance to empiricism in trade mark law is a significant and welcome intervention. Trade mark law allocates and enforces property rights that are contingent upon consumer perception. For instance, are consumers likely to view the sign being claimed as a source indicating signal (subsistence)? Is one producer’s mark close enough to another’s to cause consumer confusion (infringement)? Taking the most frequently used infringement test as its focus, her article investigates whether – in the context of the likelihood of confusion – trade mark law cares ‘about real world consumers, or imagined ones’ (at page 57).
The answer which emerges is that trade mark law selectively ‘blows hot and cold’ towards empirical evidence measuring consumer responses to signs in the marketplace. In the subsistence context, trade mark registrars and judges are far more amenable to empirical evidence and consumer surveys in particular. Thus surveys assist courts in assessing whether a shape has become a mark. Do a sufficient proportion of consumers perceive that the shape of a four-fingered chocolate wafer bar has acquired source-indicating significance over time? Yet when it comes to the heavily utilised likelihood of confusion test, Anglo-Australian courts have blown cold, if not positively arctic, towards surveys in recent years. Similar scepticism is evident in other jurisdictions as well. Judges have raised the threshold for admissibility while downplaying the probative value of consumer surveys, restricting them to very specific situations. It is argued that this ambivalence to surveys is symptomatic of ‘a deeper disjunct between trade mark law and scientific research into consumer decision-making’ (at page 59).
This comment will (i) sketch the outlines of Professor Weatherall’s argument and (ii) while broadly endorsing its conclusions, identify some of the limitations of any argument that calls for a greater empirical infusion in trade mark law. While trade mark law has grown increasingly abstract and de-contextualised, the existence of certain structural features suggests that only a qualified turn to empiricism is possible. Before doing so, the originality and value of her contribution is worth emphasising, by contrasting it with two existing bodies of research. First, a practitioner-driven literature has emerged in response to the judicial scepticism of surveys. However it primarily focuses on how to do surveys better, within the narrow confines of contemporary doctrinal parameters. Second, a body of academic scholarship has begun to engage with the question of ‘whether trade mark law ought to be framed in a way that takes better account of scientific insights regarding consumer capacities and limitations, decisions and responses’; however this literature focuses on doctrine and its improvement: ‘on how statutes, directives, and court decisions conceptualise the identity, features, and behaviour of consumers’ (at page 59). It seeks to modify aspects of trade mark law in order to accommodate empirical insights. It is largely prescriptive, ‘law reform’ style scholarship.
While Professor Weatherall’s article also seeks to put the insights of cognitive science to work, in order to improve aspects of trade mark law – there are lessons indicating how one could sensibly integrate empirical evidence into the existing test for confusion – its agenda is more overtly epistemological. It ‘asks a legal audience to flip their perspective, and think about how representatives of cognitive psychology or related sciences, asked to provide empirical tests of legal concepts, might view both trade mark lawyers’ conceptions of consumer decision-making, and the way empirical evidence is critiqued by courts’ (at page 60). Weatherall is clear that she is not advocating the fundamental reshaping of trade mark doctrine in accordance with scientific findings about human cognition. Law need not necessarily march to the beat of Science. However she does aim to promote cross disciplinary communication between cognitive scientists and trade mark practitioners. She sets about making worlds (or at least word views) collide in this article and does so very productively. The take home message could be summarised thus: at the very least, those applying trade mark law should be more open to (laboratory or field based) experimental evidence, which can (i) illuminate specific stages or waypoints in the consumer confusion enquiry – for instance, are consumers likely to perceive two signs as being similar – or (ii) provide a better informed overarching framework for understanding consumer perception and decision making in the real world.
Weatherall’s argument has three substantial components. First, the article sets out the basis for judicial discomfort with empirical evidence in the context of infringement. The likelihood of confusion test involves an assessment of ‘the risk that the public might believe that the goods and services in question come from the same undertaking or, as the case may be, from economically linked undertakings’. Litigants have attempted to use consumer surveys as a means of establishing this confusion. Increasing judicial resistance to this form of evidence rests on five principal grounds: (i) the most significant reason is that surveys generate disproportionate costs, which generally do not outweigh their evidentiary benefits. These costs can subsequently be shifted to the losing party. However ‘[m]odern rules of procedure require courts to consider the cost and resource implications of decisions to admit or reject evidence’ (at page 65). Therefore the present UK approach is to only admit surveys where they add ‘real value’ i.e. where their benefits outweigh their costs. Yet Weatherall notes that one driver of costs is the very demanding standards that surveys have to live up to, as contained in the Whitford guidelines, described as ‘combination of the indisputable, the impracticable and the inscrutable’ (at page 62). Cognitive science research suggests that some of this rigour is unnecessary or irrelevant. The judiciary may therefore be complicit in driving up costs. (ii) Surveys are dismissed on the basis of being tainted by leading questions. By ‘asking “who made this shoe”, [such questions] are said to cause consumers to speculate about an association between a claimant’s mark and an alleged infringement’ (at pages 65-66). (iii) These exercises are also accused of being ‘non-ecological’, taking place in an artificial bubble as opposed to the real world context of actual purchases. (iv) Courts have cautioned against the seductive power of numbers, observing that pre-determined percentages of survey results should not displace judicial assessment. (v) Finally, courts should assess the consumer response within a broader normative framing. Would a hypothetical, reasonably observant and circumspect consumer – similar to the reasonable person of tort law – be confused? As part of this exercise, courts proactively attribute a degree of attention and care to this hypothetical subject in order to avoid the pitfalls of paternalism. As surveys have only a limited role to play in this broader balancing exercise, is the game worth the candle?
In the second movement, the argument ‘shifts perspective to consider how someone trained in cognitive psychology or related disciplines might view questions of trade mark law and proof’ (at page 66). Here the relevant teachings from cognitive psychology are identified: prominently insights from ‘thinking fast and slow’ research, which distinguishes unconscious mental processing of information from conscious, deliberative processing. Our unconscious auto-pilot may be entirely responsible for a certain range of purchasing behaviours while also laying the groundwork for the conscious decision making which follows in other situations. By contrast trade mark law presumes conscious engagement as the default. Weatherall subsequently draws on established experimental methods in the cognitive sciences to emphasise that judges are rejecting tools and techniques which cognitive psychologists are willing to work with. For instance, there is qualified support for the ‘wisdom of crowds’. Multiple ‘independent measurements, or assessments, will often be closer to the “truth” than any single measurement or assessment’. Therefore taking ‘even five or ten measures and averaging them provides a markedly better estimate of the true value than taking a single measure’ (at page 70). This implies that even a small scale experimental test investigating how consumers would react to certain signs would be better than relying on the impression of a single judge. Along these lines, the significance of ‘coverage error’, or whether the right people have been included in the survey, varies. Therefore even if the surveyed population does not precisely match the legally relevant group of consumers, there are ways in which ex post adjustments can be made without rejecting the findings in their entirety. Finally, cognitive scientists accept that ‘confidence in a proposition is increased when it is supported by convergent evidence’ (at page 71). So convergent evidence provided by an array of experiments generating empirical results – for example, measuring response times after exposure to certain stimuli or memory tasks – ought to be more acceptable to courts as well.
However trade mark law has its own model of the consumer, with its distinctive quirks and features. Adjudicators are required to adopt the vantage point of a hypothetical consumer who is reasonably informed, observant and circumspect. This vantage point can accommodate normative inflections, since the scope of proprietary entitlements are defined via the consumer protection argument. After all, it is rival trade mark owners who bring claims, not aggrieved consumers. This is an issue which we return to below. Trade mark law also presumes a degree of background knowledge and memory attributes on the part of consumers. For instance, if the sign has not yet been used across the full range of goods applied for, a notional usage and corresponding consumer response will be assumed. Finally, legally relevant confusion is not congruent with empirically measurable confusion, since experimental research indicates ‘that even reasonable, rational consumers sometimes make mistaken choices’ while unconscious processes drive decision making to a large degree. Trade mark law therefore appears to be asking whether a hypothetical average consumer ought to be confused, while taking surveys to task for not being able to map on to this entirely constructed set of standards. Notably, other sources of evidence are not held up to such high standards. Courts ‘require perfection in surveys but skate over the many problems which beset other sources of proof of confusion’ (at page 80). Curiously, judges are willing to entertain witness evidence of actual confusion in the real world, even though the witnesses may not be representative and anyone can have an unreasonable moment.
Having identified the divergences and dissonance, the final section reflects on where we might go from here. Professor Weatherall rejects a ‘shrug and do nothing’ approach as well as one advocating the refashioning of trade mark law in the image of cognitive psychology. Trade mark law outcomes cannot be deterministically settled by empirical findings because trade mark infringement involves normative decisions, since proprietary entitlements are allocated during the course of the confusion analysis. If someone has claimed a weakly distinctive or largely descriptive sign as a trade mark, there are good reasons for presuming that consumers will not be readily confused when a competitor also uses those descriptive elements of the sign. Having rejected this binary choice, Weatherall concludes by suggesting two promising avenues. First, the manner in which empirically informed cognitive sciences expertise can assist indirectly, at the level of generalised propositions about consumer information processing and decision making (Option A). Cognitive science can help establish ‘how people behave in the market generally’. These generalised response patterns can always be judiciously filtered since they may not apply in the specific dispute before an adjudicator. For instance, determining whether the spelling or the sound of words are more important when consumers are assessing their similarity, is an empirically testable proposition. Scientific evidence of this kind should inform the existing default rules within trade mark manuals and established judicial tests. Second, in the context of dispute-specific empirical evidence, better engagement is possible (Option B). Since the perfect has become the enemy of the good (or good enough) the alternative to excessively rigorous surveys is ‘a series of smaller, less expensive tests, using a variety of methods, [that] might well be able to generate greater confidence in a result’ (at page 85). Here convergent evidence from a range of ‘quick and dirty’ yet scientifically valid tests could help establish a likelihood of confusion.
Weatherall’s argument is ambitious, provocative and draws on a larger research project involving close collaboration with researchers from the cognitive sciences. She reconsiders trade mark doctrine through their eyes and this adds real value. Her argument is therefore best understood in epistemological and methodological terms; it is neither overtly pro-claimant nor pro-defendant, which are the default categories for much contemporary trade mark scholarship. The paper also chooses a legitimate target to critique, since trade mark law has steadily been distanced from the context of an actual marketplace over the past century. This is attributable to several developments, including:
- The historic transition from case specific unfair competition or tort protection, with its fact-intensive analysis and requirement of damage, to registration-based proprietary regimes which operated at the national and then regional (i.e. European) level. This compelled courts to think in terms of consumer responses in the abstract, across vast national or even continental markets.
- Registration, by way of ‘reserving a space’, can precede actual commercial use by up to 5 years. Adjudicators have to make educated guesses during this grace period as to how consumers might respond if a third party uses a similar mark on similar goods.
- The likelihood of confusion test is run as a paper-based or formal exercise when it comes to relative grounds oppositions, without any requirement of actual confusion. This generates a considerable and influential body of doctrine applying the confusion test purely in notional space.
- In remedial terms, the availability of quia timet relief allows courts to grant injunctions before an infringing act has actually taken place.
- The ‘likelihood’ in likelihood of confusion lends itself to speculative or probabilistic assessments, as does the requirement that a ‘substantial proportion’ of consumers needed to be confused (which could amount to a modest minority). So cumulatively, the test merely requires that it is more likely than not that some consumers might be confused.
Trade mark doctrine is therefore increasingly abstract, if not overtly formalistic. This is a drift which deserves to be challenged. By highlighting the resistance to empirical evidence about real world consumer perception and information processing abilities, Weatherall’s argument is a challenge to those who advocate the ratcheting up of trade mark protection in the name of protecting consumer interests or efficiently functioning markets.
However there are three implications of the argument which are less appealing. The first relates to the extent to which the frosty reception to survey evidence is emblematic of a broader disjuncture. Trade mark judges have engaged with empirical evidence – and experimental evidence in particular – when it comes to another family of infringement claims relating to the ‘dilution’ of a reputed trade mark. Dilution refers to a set of non-confusing yet wrongful mental associations between a claimant and defendant’s marks. In the US, wrongfulness stems from harm or injury to the plaintiff’s reputed mark where there is blurring or tarnishment. In the EU, it additionally extends to a third type of claim: an unfair advantage or benefit obtained by the defendant where there is free riding on the reputed mark. Dilution is therefore a form of enhanced protection reserved for marks with a reputation, which has historically targeted unauthorised use on non-competing goods. It seemingly favours the interests of proprietors in protecting their brand identity and image, exposing the turn to brands within trade mark law. Establishing the first stage, the likelihood of a mental association or link between two marks in the minds of consumers, is relatively straightforward. It is the second stage, the consequences flowing from that association, which poses problems. An illustration of a harm-based variant – blurring or detriment to distinctiveness – is provided by the ‘Barbie Girl’ case, where Mattel, the manufacturer of Barbie dolls, claimed that the unauthorised use of that trade marked name by the music group Aqua in their song ‘Barbie Girl’ would erode its ability to indicate a doll by Mattel. Measuring impairment of the source-indicating ability of a mark clearly falls within the province of empirical testing and US courts have been open to such evidence. Experiments have been designed to evaluate whether consumers experience increased internal search costs or greater recall time after exposure to the junior user’s identical or similar mark. However reliance upon such evidence throws up a number of challenges for trade mark law. It bolsters claims advocating the propertisation of consumer attention and mental associations. If trade mark owners can lay claim to preserving these mental connections, there will be undesirable consequences for competition and free speech, since both processes inevitably set up dissonant associations. Additionally, is what is experimentally measurable (the disambiguation time in the Barbie case might amount to milliseconds) legally meaningful (does it change real world purchasing behaviour and cause actionable harm)? Therefore dilution doctrine, the other category of trade mark infringement claims, has begun to engage more closely with empirical evidence as well as many of its challenges.
This leads to the second concern. Weatherall’s research fits within the broader engagement between ‘Law and Science’. Debates which traverse similar cognitive and behavioural terrain have emerged in other areas of law. When allocating responsibility in the law of civil and criminal wrongs, how does neuroscientific or behavioural psychology evidence interact with the proof requirements, doctrine and theory of legal liability premised upon free will? As we have seen in the preceding paragraph, in the trade mark infringement context an adjudicator would have to balance the competing demands of consumer interests in preserving the clarity of a trade mark signal, property interests of trade mark owners, the interests of competitors in competitively aligning themselves and the public interest in being allowed to use a trade mark as a broader register of meaning. Any empirical evidence relating to consumer perception would have to be carefully integrated within this balancing act. A related concern has to do with how scientific evidence is marshalled for the purposes of legal enquiries. With the predictable oversimplification disclaimer, we might associate the culture of science with facts and the culture of law with pursuing the ends of social or political values. Consequently, the two cultures approach truth quite differently: ‘science generally approximates truths by hypothesis-testing, whereas the legal system frequently approximates truths by evaluating what happens when two highly incentivized teams shoulder a legally imposed duty to gather evidence and to argue in favour of two directly opposite propositions.’
This brings us to the final concern. Trade mark rules on standing exclude the consumer, who can never claim she is confused or deceived. Trade mark law is in the business of allocating property rights over signs. It is trade mark proprietors and their competitors – those with the relevant commercial stake – who have the necessary standing to commence litigation. The consumer construct is a litmus test for assessing when property rights have been infringed, unlike consumer protection or labelling laws which directly protect consumer interests. Let’s presume that the empirical evidence reveals that our autopilots that help us continually interpret the world – fast thinking systems – may lead us astray during the initial stages of the purchase process, or some of us will always be confused at least some of the time. Should such findings be used to enhance property rights by encouraging the recognition of doctrines such as initial interest confusion? Also referred to as pre-transaction confusion, this doctrine ‘allows for a finding of liability where a plaintiff can demonstrate that a consumer was confused by a defendant’s conduct at the time of interest in a product or service, even if that initial confusion is corrected by the time of purchase’. The danger with casting the trade mark confusion net so wide is that it threatens to encompass mere diversion of consumer interest and impede desirable competitive processes. Then there is the issue of disaggregating consumer responses, given the ambiguity surrounding when a substantial proportion of consumers are confused. If a minority of consumers find an unauthorised use of a mark – say, in a comparative advertisement – confusing, what about the majority of consumers who don’t and would actually benefit from the information provided by the comparison? Or consider the argument that low level, modest confusion which is confirmed by empirical data actually helps consumers to be more discriminating over time and improves their search capacities. While Weatherall’s argument does not preclude these questions, it stops shy of engaging with them. Therefore, of the two recommendations proposed in her concluding section, my preference is for Option A, whereby courts engage with empirical evidence establishing generalised propositions about consumer information processing and decision making. Option B would encourage the submission of dispute-specific empirical evidence, including cheaper, ‘quick and dirty’ experimental evidence. Given the epistemic foundations of litigation (legal truth benchmarks often being different from scientific truth) and the tendency for cross examination and questioning of any evidence to drive costs up, Option B runs the risk of circling back to the same concerns currently associated with surveys.
It only remains to conclude on a note of appreciation. Weatherall’s argument is creative, rigorous and provocative. It deserves to be engaged with and will reward both courts and scholars who do so.