Session 1, Continued
Mid-point summary: Rebecca Tushnet
Might be interesting to go back to KP Permanent as a key moment in revisionist history of TM: as I recall, the position of Lasting Impression relied very heavily on the statutory history with respect to incontestability, explaining why incontestability could have been seen as the equivalent of double identity, thus making likely confusion plausibly and coherently part of the descriptive fair use defense as a defense put in the statute specifically to deal with the new concept of incontestability. Its analysis, while historically rich, came off as all but incomprehensible from the presentist narrative you see in the Supreme Court opinion.
eBay and this discussion: in irreparable harm, the magic happens all over again. What is the difference between goodwill and lost sales? Even courts that demand an eBay analysis in infringement cases think that goodwill losses automatically result from confusion and result in irreparable harm, meaning that injunction is automatic. Need to start questioning that.
Search costs and confusion: search costs may have offsetting benefits. Richard Craswell is excellent at explaining that eliminating costs also eliminates benefits in some cases, and the question is balancing. Failure to recognize benefits—often a failure to recognize pleasure. Courts are aware of information reduction costs—descriptive fair use and nominative fair use exist to protect against lost information. They’re much more uncomfortable with having fun. Charbucks court has no idea what might be enjoyable about consuming Charbucks coffee, and even though the Chewy Vuiton court got the result right it described the effect as parody, but parody is often thought to require a message; what Chewy Vuiton is really about is the value of enjoyment, of play. Without recognizing that we’ll get cost-benefit analysis wrong.
Bently’s question: why are we making it easier on deceptive marks? Because the abandonment of producer rhetoric coincided with the reconfiguring of law to favor producers. Same results obtain in the US, as I’ve argued—only instance in which the PTO doesn’t need to show materiality to refuse registration based on deceptiveness is when the deceptiveness is about source/affiliation. Structural opportunity in US given that false advertising and infringement of unregistered marks are subject to the same statutory provision and nearly identical language—can find justification for materiality and certain kinds of empiricism in existing Lanham Act doctrine.
Search costs can work if we do cost-benefit analysis about competition. To the extent TMs confer information about the social status of the user, we in this room have discomfort with endorsing robust protection for a single owner. We should be cautious about promoting the protection of this kind of information, but we have to confront the reality that consumers derive value from associating themselves with prestige products. Posner & Landes say that if people buy Tylenol because it makes them feel better, that’s not a terrible thing. Need to deal with the questions of reduction/increase in search costs and the overall value and quality of information in markets.
Discussion of harms should take place in context of benefits to potential users of marks or marks similar to TMs. We tend to talk about overexpansion of TM either in the abstract or with respect to certain crazy cases. But it’s worth asking whether, when there’s a potential info harm, there is a plausible, compelling interest on the other side in access to the mark. McGeveran has identified many such interests, and Tushnet also in the value of play. But critique in the abstract plus a few jarring cases can be misleading. Harm to TM owners exists on a spectrum, as does benefit to users, and they need to be married better.
TM appears overinclusive with respect to materiality; but if we accept the role of accurate information in efficient markets, then unfair competition might appear underinclusive. This depends on how we define materiality and the risks of false positives/false negatives.
McKenna: confusion is imperfectly related to the idea of misinformation. Misinformation might be something that’s not objectively false. Most cases involve a bunch of things at the same thing, good and bad information—you can’t just say we want to deal with misinformation. Comparative advertising: raises search costs; might be worth it because the long run effect of competition is worth it, but that means accepting some level of confusion, e.g., in similarity of cereal boxes imitating national brands. Normatively, courts say there’s no confusion, which means there’s no confusion we care about. Costs and benefits can’t be cashed out in search costs, but in ultimate effects. Hard because we tend to include a lot of things in “effect on competition.” Competition on price, quality, product characteristics; but in the modern brand world, competition can also be between brands—from one perspective there’s nothing anticompetitive about saying “compete not on price or quality but by creating your own brand.”
Lemley: Search costs analysis is part of a broader analysis, in the service of a well-functioning market. TM can do that can enforcing rights against people who raise search costs by deceiving, but not always. Search costs per se don’t have much to say about functionality, for example, which is about market competition more directly. (McKenna says it’s a doctrine that means that we have to swallow some search costs.) That’s inconsistent only insofar as we think it means that people have internalized the idea that this useful product can only come from one maker. Likely to be true only in a vanishingly small set of cases. (RT: But there will be a spectrum of consumers.)
Search cost theory often argues for limiting TM. Doesn’t think that comparative advertising causes confusion that has to be tolerated; he thinks comparative advertising on balance reduces search costs. Genericide is the same—people may be confused on both sides—but search costs tell us to look closely at TM claims that interfere with a free market for information. Eric Goldman’s paper on search relevancy is useful here—if people get useful info from internet searches, regulating the way they do searches on TM grounds is problematic.
Dogan: the effect on search costs should be just one consideration in determining whether a use of a mark makes a market function better. Even if you accept that comparative advertising increases search costs, search costs are just one vehicle for competitive markets—it forces the consumer to pause and think about the characteristics of the product, which is good for competition overall (McKenna: in price/quality terms).
Eric Goldman: so much of TM assumes a single model of how consumers behave, and that model doesn’t work for him. Different products entail different search strategies/engagement with choice. Different consumers of the same product have different choice strategies. More semantic ambiguity than we acknowledge. Empiricism: you usually find subcommunities, some with greater confusion than others.
Bone: We need to avoid making the inquiry so complicated that it’s nonfunctional. We need to pitch our solutions at an administrable level of generality. Search costs is the wrong box—we need to sort via costs balanced with benefits, and that requires various kinds of tailoring. Saying it’s just about competition is difficult because lots of things affect competition, and many bodies of law regulate it, including antitrust. How do we fit TM into that vast landscape?
Dogan: certainly TM can’t guarantee competitive markets. But you agree it can facilitate them?
Bone: TM’s role is to protect quality of information offered to marketplace. In some abstract sense, it’s to keep markets competitive, but there’s lots of ways we do that—something special about supplying info through marks to the marketplace, keeping that info from being tainted.
Beebe: search costs v. information costs—what is the distinction between those?
Bone: search costs—what we often mean is not that consumers will search, but that consumers won’t bother and will buy the product and not get the product they want. Maybe we want to measure that harm in other ways. If judges were thinking sensibly about this, it shouldn’t make a difference. But search costs label tends to push us in a direction where we suddenly worry about whether dilution increases search costs. That’s ridiculous; why worry about extra time thinking? Point of TM is not to keep those costs low, but to keep the quality of info TMs are conveying as appropriate as possible (costs and benefits both).
Lemley: What things go into cost and benefit buckets? One of benefits of thinking about TMs in competitive market is not just broader framing, but that when TM in the supposed service of one aspect of competition ends up interfering with the market economy, that has to count in the analysis. This won’t happen if we think just about confusion and ignore, for example, functionality. Cost benefit analysis might have the implication that some of the ways in which we bias the confusion inquiry are wrong. If we think it’s costless, then stopping 7% confusion might be great. If we think there are costs to TM enforcement, then we must weigh confusion against costs of stopping. In genericness cases, we require a substantial majority to see the marks as generic, because it’s obvious that there’s injury on both sides—people who think Kleenex is generic are injured by exclusivity, and people who think it’s a mark are injured by nonexclusivity. This leads to the question of rules v. standards. When do we short-circuit the inquiry in any particular case? He likes standards but recognizes the appropriateness of presumptions sometimes.
McKenna: everyone agrees in evaluating benefits. Shares Bone’s sense that just saying it’s about promoting competitive markets isn’t enough, because it’s not helpful to courts in any particular case. Falls to us to specify at a class level what kinds of things you should take into account. Not sure “quality information” gets much better. Quality info depends on perspective: mark owner thinks any implication of involvement should be stamped out and that’s quality info. In a lot of cases, like Chewy Vuiton, there’s not info in the sense of product characteristics—using the mark for some other purpose.
Dinwoodie: Lemley & Dogan offer a totalizing theory allowing you to balance harms and gains. Problem is more acute—many of the benefits of third party uses are not commensurable in those terms. Competitive market analysis is the same as search costs in this way—it’s attractive for certain things like functionality, but TMs do much more than serve an economic function.
Burrell: difficulties of modeling how TMs really relate to market functioning—how do you model enjoyment? How do we deal with consumer variability? What about rational ignorance?
RT: I think McKenna’s conclusion that Chewy Vuiton doesn’t involve product characteristic information is both a natural reaction to the use of law & economics lingo and wrong—we should identify pleasure as an effect of info if we are very serious about taking preferences as exogenous. But this is a classic problem of the inability of this type of analysis to truly accommodate preferences (especially since they are not exogenous but, very noticeably in Chewy Vuiton and similar cases, shaped by the market).
McKenna: we have a sense that uses that affect meaning are different because we’re used to thinking about price/quality characteristics. But Posner of course thinks that the “value” or quality includes meaning. [RT: And so Posner punts to a property concept which turns effects on meaning—nonexogenous—from benefits to costs.]
Michael Grynberg: question of benefits of an incomplete theory. A good theory of search costs could be used to screen out some problematic uses, and then we’d have leftover uses like Chewy Vuiton. Would the ability to screen out some uses make it easier to generate new defenses, like First Amendment defenses, for the problematic things that were left over? Do we need a theory to handle everything?
Lemley: another way to think about utilitarianism: a functioning market should allow people to do what they want with their preferences. If you want to spend twice as much on Advil as on generic, knock yourself out. But what do we do with Chewy Vuiton? In a market economy, that’s a spillover effect, and we don’t care. Propertization says we have to internalize all externalities to provide incentives, and this has had a lot of traction in copyright and has crept in through free riding to TM, justifying the argument that a preference for a Chewy Vuiton toy comes from Louis Vuitton, that must be an externality to internalize for LV. As a matter of economics, that’s just not right. Spillovers enhance growth/innovation.
McKenna: Chewy Vuiton is not just uncaptured positive spillover, but it also has a potential effect on Louis Vuitton, which can be articulated as harm to the mark owner—changing how people feel about the brand—not just an uncaptured externality but a change to preferences. His argument: construct consumers as capable of managing preferences themselves. Courts should be symmetrical. If consumers can make up their minds for themselves about whether they care about image, then to the extent that all defendant’s conduct does is affect image, it should be left alone by the TM system.
Lemley: but if it distorts purchase of plaintiff’s product then there’s a case for action.
Me: what’s distortion? [Change in preference is change, in the absence of deception about what it is you’re buying, not distortion.]
Goldman: Communities are related to the question of play. How do brand communities exist, and can we recognize them as independent of the brand owner?
Bone: Complex, pluralistic consumer world out there. Makes cost-benefit seem very hard to do. The key word is: is there systematic distortion of the market? He’s attracted to rules rather than standards. We have to keep the rules relatively narrow, recognizing that TM is not all-purpose tool for avoiding harm to consumers. Places where unauthorized TM use is harmful exist, and we should have core rules for those (without even presumptions), and resist the impulse to create extensions to expand that core.
McGeveran: We’re talking infringement, dilution, unfair competition together. If the topic is confusion, then even if confusion is meant to be the core of infringement we still have satellite areas of law doing other things.
Beebe: argues that most of the empirical studies out there are not very good. 70 undergraduates are tested in front of a computer; results used to intervene in TM doctrine are based on terrible protocols (as I pointed out in Gone in 60 Milliseconds).
McKenna: it’s true that methodology leaves a lot to be desired. Worth noting that those are done by marketing departments to tell brands how to act. If anything, they overstate the case for brands. But companies rely on them. To the extent they nonetheless don’t support TM doctrine in favor of brands, then it’s fair to hold marketers to that.
Goldman: So how do we engage marketing researchers better to help them understand TM? Maybe we should be inviting them.
McGeveran: marketing literature is not about quality—it’s about telling the brand story, personality, play. Talking about confusion without acknowledging that courts are protecting what they see as an investment in brand personality is swimming upstream.
Dinwoodie: given discussion of biases, how would things change if surveys were inadmissible, as they are in some countries?
One of the problems of drive for totalizing theory is that we’ve taken different causes of action pursuing similar but distinct policy goals and collapsed them. Historical urge among scholars/courts to find totalizing theories for all causes of action?
RT: I agree but am more interested in the return of the repressed: we collapse the causes of action into one and then we realize that different situations require different responses so we get a million defenses and also this contributes to the rise of the multifactor confusion test. This is a greater judicial culture problem: the standard Supreme Court case now articulates five things for the court of appeals to consider on remand. (See also the development of the “standing” test in false advertising, in which multifactor analysis substitutes for a basic test of whether there’s harm.)
McKenna: before the great shift in confusion for noncompeting goods there were different proof frameworks for different harms, and then we wound up with the same proof framework for all different kinds of harms including harms where presumptions were not appropriate as they were for pure market substitution.
Bone: disagrees with McKenna some about the role of property concepts and when types of harm became conceived of as the same. But, per McKenna, agrees that the multifactor test emerged after harm from noncompeting uses had been deemed to be all the same thing as harm from competing uses. Causes of action do get unified under that harm theory and blended together. [Similar story in copyright, I think—the question has often been “did we exclude that from copyright’s scope because we weren’t thinking broadly enough or did we exclude it because we didn’t think copyright should extend that far?”]
McKenna: a narrow right is less likely to collide with other interests; once the right broadens, its underlying theory becomes much more important so that you can figure out where the right should end and give way to other considerations.
Lemley: if we disaggregate, a logical breakpoint would be to have different rules or standards for different types of confusion—double identity should get one set of rules, source confusion another, sponsorship confusion/affiliation confusion a different set.
McKenna: courts always say when they’re breaking new ground that confusion about anything counts, so we’d go a long way by saying that they have to specify what the confusion is about.
Heymann: would also help us specify what harm we’re balancing, rather than assuming that it’s always the same interest.
Bonus: a caricature of Mark McKenna smashing TM law, by Bill McGeveran: any effect on his brand?