Commentator: Rebecca Tushnet
International element to the question of how to limit trademark to protect free speech. TRIPS doesn’t impose insurmountable barriers to rights-protective legislation, but international obligations will be raised by high-protectionists as reasons to avoid reform and so it’s worth thinking about what international law does and doesn’t constrain. Nations and free speech advocates should therefore focus their efforts on developing specific, speech-protective model rules for trademark law to provide guidance to decision-makers, similar to WIPO’s Joint Recommendation Concerning Provisions on the Protection of Well-Known Marks and the recent agreement on a treaty on access to texts for the visually impaired.
Advocates for soft law over mandatory defenses, allowing states flexibility. One question for further discussion is how well flexibility in allowed defenses has worked out, not just in broad international treaties but European harmonization. Are there examples of success?
Paper starts with summarizing the conflict between aggressive trademark claims and free speech, then turns to principles for defenses. Ramsey argues that nations should pay more attention to what they deem the aims of trademark to be, and that identification of those aims will naturally suggest appropriate defenses. (One risk, it seems to me, is that nations will decide “we want trademark owners to have rights to protect the dignity and reputation of their marks, however insulted or harmed”; having done so, that’s about it for free speech defenses.)
Many nations protect free speech, as do significant international agreements. But the US is an outlier in articulation, if not in practice when it comes to IP, and it’s generally recognized that free speech can be limited in the service of other appropriate goals. And many nations recognize property as a constitutional or basic right, and then consider trademarks to be property. Syllogistically, this suggests that protection of trademark interests can trump free speech rights, though then the question is, given that trademarks are intangible, is there a scope of these property rights that is entailed by the interface between property claims and free speech claims. With tangible goods, it is generally unobjectionable to say that a speaker can’t appropriate another person’s printing press to communicate her own message. But the analogy breaks down when the boundaries of the property right are what’s up for contestation.
As Ramsey points out, if US courts applied what she calls real First Amendment doctrine—doctrine applied to things that aren’t IP—then much of trademark would be constitutionally suspect. Let me say in passing that I think we have to accept that IP’s First Amendment is real First Amendment doctrine too. It may be bad, and it is blatantly conflicting with other First Amendment doctrine, but I believe in it the way I believe in infant baptism: I’ve seen it. I don’t think we can expect harmonization of IP and non-IP cases, certainly not harmonization in a speech-favorable direction, any time soon, and this is what makes US courts not as significant outliers with respect to speech based limits on trademark as they are with respect to limits on the regulation of hate speech et cetera.
Anyway, given this situation, Ramsey points to Bill McGeveran’s work on internal limitations to trademark. By internalizing doctrines as part of the structure of trademark itself, it’s possible to protect speech interests without explicit reference to First Amendment limits, using safe harbors and affirmative defenses. This flexibility can be accomplished by explicit exceptions created by legislators as well as by judicial interpretation.
Ramsey thus suggests having a menu “including exclusions from protectable subject matter, limitations on the scope of trademark rights, defenses, and remedies” from which nations can select.
As she points out, the ability of nations to deny registration to nondistinctive terms, descriptive terms, and generic terms is an important protector of free speech. She suggests that nations might want to deny protection for descriptive terms even if they’ve acquired secondary meaning, a suggestion that will probably find a fair amount of resistance but one that builds on her previous work on descriptiveness. More mildly, she proposes stepped-up distinctiveness requirements for protection of descriptive marks, the adoption of a descriptive fair use defense, and a more limited scope of a mark if it’s descriptive.
She also suggests limiting protection for “cultural heritage marks” to prevent the privatization of culture. (Here it would be interesting to engage with the opposite claim that cultural heritage should be privatized and put into the hands of particular non state groups; it’s been argued that trademark-style rights, even expanded rights, would allow representatives of the culture to best preserve and benefit from it. Denying rights might mean that anyone could use the culturally significant terms, objects, etc. which might prove unsatisfactory.)
Ramsey also suggests a limited form of a trademark use defense for slogans and symbols on standard merchandise, like T-shirts, as a way to protect common media of personal communication. She points to attempts to register marks that seem to appropriate common cultural touchstones, like Disney’s application for Day of the Dead/Dio de las Muertas and various attempts to register phrases like “Let’s Roll”; “Don’t Mess with Texas”; and “If You See Something, Say Something.” The latter raise interesting questions of government control, though I’m not sure people who aren’t steeped in the First Amendment will object to government control of symbols in the same way—after all, most nations have copyright in government works; what’s different about trademark?
Ramsey also takes up the question of whether it’s better to have exclusions from protection for marks like these or limitations on scope. Strategically, she argues for the former. This makes it easier to prevent bullying and aggressive attempts to push the envelope on supposedly limited scope. (It would be useful to discuss the limits on bullying that other countries have, like abuse of rights claims available in some commonwealth countries.)
Ramsey also suggests certain limitations on scope of rights, though. TRIPs provides that we have to protect against unauthorized use of (1) registered trademarks, (2) used in the course of trade, (3) where the signs and goods/services are identical or similar to those of the markholder, and (4) this use of the mark causes a likelihood of confusion. Each of these parts provides potential limits on scope. Speech protective measures could limit protection, or limit remedies, to things outside this core category. Under the double identity rule, for example, courts presume confusion when an identical mark is used on the same goods—but that actually leaves plenty of room for nonidentical marks and nonidentical goods if nations care to use that space. As for likely confusion, we can have rules that protect comparative advertising, keyword use, and the like. We are not required to give likelihood of confusion a broad reading, and Ramsey argues that we should not.
You can tell what her perspective on the protection of well-known marks is, so I won’t detail it. Likewise, she points out that rules on secondary liability have speech implications, and there is plenty of room for states to adopt speech-protective rules on secondary liability.
The overall point is that there is significant flexibility in the international system. The nation that created the exception to trademark rights has the burden under TRIPS Article 17 to prove that (1) the exception is limited, and (2) the exception takes into account the legitimate interests of both the trademark holder and third parties. Given that free speech interests will regularly be significant and legitimate/appropriate for a democratic society, she concludes that any or even all of these proposed limits would be acceptable under TRIPS and related agreements, though in some sense TRIPS is the only one we have to worry about since it’s the only one with enforceable remedies.
Potential barriers worth discussing: can be summarized as politics, though that may be seen as an unfair characterization of the claims of trademark owners who believe that the power of their marks should be reserved for their own exploitation. Politics are often most effective when they have intelligible rights claims behind them; the paper could benefit from more engagement with these rights claims. Why have speech claims seemed unpersuasive to many thus far? This will lead naturally to ways to answer objections to Ramsey’s interpretation of TRIPS as allowing for this much flexibility. One suggestion: discuss property/takings type arguments made by tobacco companies in response to plain packaging and other limits on use of their marks.
Ramsey: property rights aren’t an automatic trump; you still need to consider 1A/free speech: ECJ and national courts have said that fundamental rights need to be balanced.
Rigidity v. flexibility: advocates for more specific defenses. But we need to give courts ability to engage in speech-protective case by case analysis. Want some constraints on judges to prevent overprotection, but there needs to be more flexibility at the international level. TRIPS/Paris Convention have flexible and ambiguous language. Annette Kur etc. have proposed mandatory defenses added to treaties, but because many nations don’t protect speech as well, that list wouldn’t be very protective and might encourage nations to adopt tiny lists.
Q: Cases in the WTO where they use the articles you cite, 7 & 8, for your purposes? Doha takes them into account for pharmaceuticals but it’s not clear where that’s going. Is there case law saying they’re more than hortatory? W/regard to flexibility to nonflexibility: what is the goal? How do you compare harmonization to flexibility?
A: no cases yet; difficult balancing act.
Betsy Rosenblatt: you cite McGeveran’s predictability desire. With a lot of globalized expression, one element of free speech is knowing you can disseminate broadly—predictability is a free speech value. Unauthorized use of TMs in a film can generate risks.
Cultural property: cultural property to whom? Some may be relevant to subcultures, but not to the larger group. What should nations be free to decide?
We in the US don’t put a lot of importance on dignity, but the EU does. What we might see as overriding value might be the opposite in many national philosophies about dignity of markholder, which is different from confusion value.
A: cultural property: valid points about why there might be reasons to give Maori control. But that just makes the point that there are so many different approaches that it’s unrealistic to achieve a goal of clear specific defenses shared by everyone. Reform at a national level is easier to achieve: the more nations have a comparative advertising, the better.
Q: But is Barbie culturally significant?
A: would leave it to nations to decide. Protecting Maori should be sui generis, given free speech implications.
Mark McKenna: will have to tackle sui generis/design rights—some of the issues you discuss would be rendered moot given the existence of adjacent rights. If there are carveouts in TM, people can use unregistered design rights, unless the free speech principle also controls that. Layering.
Q: TRIPS-plus/bilateral agreements as a major recent mode of expansion. US doesn’t seem to be interested in exporting free speech principles in that context—in fact, the US seems happy to raise the bar in IP.
A: European FTA did require descriptive fair use. This is the tension between mandatory limitations—which then prevent regional FTAs contravening int’l obligations—and flexibility.
Justin Hughes: claiming greater flexibility for TM may threaten ability to maintain exceptiosn for copyright and patent under the 3-step test.
A: but article 17 doesn’t require consideration of unreasonable prejudice—puts third parties on par with the markholder—and so it is more flexible than the provisions governing copyright and patent. Wants to protect copyright too!
Hughes: they’re different in wording, but that doesn’t mean they differ in effect. In the Lanham Act, courts have given the same protection to registered and unregistered marks despite wildly different statutory language. Sure the 3-step test looks different, but in the long term we might want to say it offers the same flexibility.
Ramsey also responded to my comments. Re: bullying: model rules could also allow creation of threats action. California’s anti-SLAPP statutes could be another model for protection.
Plain packaging: one argument people make is that there’s a free speech right not to have other people use their marks. There’s no right against impersonation, in her view, but criticism without confusion doesn’t justify protection. Graeme Dinwoodie: we need to consider interests that balance against and justify tolerating a certain amount of confusion.
Timothy Greene, Trademark Hybridity and Brand Protection
Commentator: Deven Desai
TMs as part of language, particuarly language as contextual. Homonymy: words can have unrelated meanings by historical accident—bank (financial institution) and bank (rivers). Paper can be the material but also newspaper, or the entity that runs the newspaper (the paper just fired five journalists). TM law doesn’t deal well with semantic ambiguity.
Word marks are linguistic, and there’s little reason to think they exhibit characteristics distinct from other words, even if processed as proper nouns. Semantic ambiguity has been researched in linguistics, and it helps us. Consumers rarely encounter marks out of context. Even if consumers experience confusion, it’s less likely that the harm would be commercially significant than courts think in many situations. A term shouldn’t be thought of as only generic, or only source-identifying. But all terms, especially TMs, embody multiple meanings and people use context to resolve ambiguity.
The binary view of TM law won’t work. Primary significance test doesn’t make sense because it doesn’t cover the spectrum of possibility. Thermos and Teflon surveys set up false dichotomies between brands and common terms, so they cause more problems.
Detail suggestions: paper suggests likely confusion test be recalibrated to deal with context. This is a real possibility, but paper should explain more how hybridity would build out the contextual analysis and why a judge should consider context equally. Also, what would better empirical work/surveys look like?
Greene: Likely confusion test currently deals mostly with linguistic aspects of mark: sight, sound, and meaning focus on the word, rather than the overall conception of the mark—logo, how displayed to consumers, etc. Can be problematic in registration where similarity in marketing channels is presumed and you can get plain word registration.
Rosenblatt: could the same analysis lead to the conclusion that TM deals better with homonymy than polysemy? We treat words differently in different classes of goods—generic in one class, descriptive in another, arbitrary in another, and TM explicitly anticipates that people can deal with this and still determine whether a word is serving as a mark for goods. It does worse with polysemy when you refer to the cultural meaning of a particular mark.
Genericism: A word doesn’t automatically become generic through sloppy usage: Google, Kleenex regularly used in broad ways—people know that’s not genericism. Does the distinction between knowing and using undermine your claim that we should study how people actually use words? Some of those surveys like Thermos might then have been on the right track.
A: when translated into judicial application, it doesn’t make sense.
Me: Richard Craswell, “Compared to What?”: The Use of Control Ads in Deceptive Advertising Litigation, 65 ANTITRUST L.J. 757 (1997), would be very helpful. Explains in a related context how meaning can differ and how there are tradeoffs when the law endorses one meaning over another. This may provide a way to translate linguistic claims, which judges may be wary of, into more conventional cost-benefit analysis.
Ramsey: more emphasis on problems of registering fancy logo + term and then claiming rights in the term: Park ‘n Fly is the classic case. Also leads to issues where if a term is put in the right place on a T-shirt it’s considered “use as a mark” and if it’s in a different place it’s not, but then if the registration issues the owner will threaten any use anywhere on the shirt.
Greene talked more about surveys: all he’s really clear on yet is that current practices are not reflective of the best linguistic knowledge. Should acknowledge people who see term as both mark and generic and put them in both groups.
Ramsey: surveys are expensive! Would this be more to test than you usually do?
A: doesn’t think it would change the cost, just split up the ways you count the members of the group. May be more useful in registration context—justifies allowing far fewer nonstylized word marks. Context specific look at what the person actually does or intends to do in commerce is really important. Cheap Seats for buses might be different from Cheap Seats for planes.
Q: wouldn’t everything get registration if you could offer a plausible context in which it’s serving as a mark?
Desai adds that he had the same question: Greene is taking this in a low-protectionist direction, but couldn’t it be taken differently?
A: looking for context in which consumers actually deal with marks—is an attempt to look at what is really confusing, and if it is, then it can be stopped.
Q: if meaning can splinter, doesn’t that support dilution theory?
A: that assumption is that meanings can crowd each other out. Psycholinguistic research says no. It’s not crowding out. These meanings all exist together in your mental representation of a given term. Barbie is a doll and an Australian slang term etc. Context distinguishes.