IPSC, Trademark

IPSC – Trademark Law I

Moderator: Mark McKenna 

Trademark Depletion in a Global, Multilingual Economy: Evidence from the European Union (abstract)

Jeanne Fromer and Barton Beebe, NYU School of Law 

Notes on the EU system: runs in parallel with national systems. Must be protectable in all EU nations (in theory); if generic in one nation, unregistrable via EU. Also, EU does not review applications for relative grounds for refusal (confusion). They generally remove diacritical marks for purposes of generating similarity reports. Translational similarity matters: same meaning in different language could bar registration. Not use-based as US is. 5 year grace period. Germany is biggest user of EU system, followed by US. Over 60% of applications contain English words—it’s the dominant language commercially. 

Datasets: EUIPO applications, registrations; gathered opposition data. Compared to corpora in English, French, German, Italian, and Spanish, and translated words into those languages as well as looking for positive/negative meanings. 

Results: nearly all good terms carrying positive affect are already registered. 1000 most frequently used words are 65% registered as single-word marks, covering 69.4% of all word usage, and of the most 20,000 frequently used words, 77.3% of all word usage is a word registered as a single-word marks. Of the unclaimed marks, they’re not great: worst, problem, than, thank, their, said, problems, worse, worry, etc. Some classes like apparel are very crowded; if you look at similarity (small edit distance) then it gets close to 90% coverage. Character, word, and number of word lengths are increasing in applications, suggesting increasing resort to longer/less “shorthandy” marks that may function less well as marks. Trends got really bad about 10 years ago. In other languages: About 95% of French usage is currently claimed. Very high proportions of multi-language words are registered: of the top 1000 English words, 55 are highly intelligible in the other languages, and 54 are registered (not REPUBLICAN). 

Opposition rates are declining, not increasing! 2017: 1 in 10. But those who bother to oppose increasingly win. 70% based on confusing similarity resulted in a refusal either in whole or in part in 2017. 

EUIPO system is very different from US: allowing enormous TM crowding, where lots of similar marks are registered at the same time. Registration rate initially declines w/# of entities opposing the application. Where no opposition, 92% registration. When # of opposers increases to 8 or more, though, the rate increases again: the EUIPO just says: everyone gets a registration; the field is already so crowded that there’s no problem. 

Compare to US: “all or nothing” rule may make it harder to find marks. Is it a good system to have third parties doing the opposing/have no examination on relative grounds? Businesses are arguably in the best position to know what is detrimental. But also businesses might need a sense of what’s potentially confusing, and EUIPO may not give it. Search reports of EUIPO to ping third party businesses are really thin. CREMOLAIT triggered nothing—they appear to identically match whole words in the registration. 

Jeremy Sheff: Disclaimer practice? Does that happen at all where common word is disclaimed?  

Fromer: in the US, disclaimers are everything: if you disclaim, you get your registration. Not clear what’s going on in EUIPO.

Irene Calboli: Applicants (except for largest) are usually just looking at 2-3 countries. With enforcement still country by country, similar rights in other nations aren’t much of a problem. Difficult to factor that in looking just at EUIPO. Licensing/coexistence agreements may also be in place. 

Lisa Ramsey: if European companies aren’t worrying about crowding, should we? Maybe rights should just be very narrow. Companies may think that they can distinguish themselves with packaging etc. 

Beebe: tests our basic assumptions about TM law. Crowding may not cause much consumer confusion; we will probably argue that European example shows that narrow rights are generally fine; businesses aren’t very worried about that when they have to express their preferences by paying. 

Fromer: want to be sure we’re covering large businesses v. small businesses with less sophistication that may not know what’s actually available. Multinationals in the US seem worried about TM depletion as a problem, but not in the EU; legal differences or differences on the ground may matter. 

Endorsing After Death, Andrew Gilden, Willamette University College of Law

Exclusive rights to control post-death endorsements given to families/heirs. While “it’s hard to argue wit h a straight face that a dead person is endorsing a product,” we see Elvis face coverings, Marilyn Monroe perfume, Muhammad Ali quotes selling Gatorade. Ordinarily an individual endorsement involves voluntary association/approval; post-death endorsements don’t fit that model, but TM increasingly supports the idea of posthumous endorsement. Project identifies this trend and identifies serious problems. Proposes narrower framework to better represent interests of decedents, fans, and survivors. 

False endorsement claims: estate/heirs claim that something will cause confusion about whether the decedent or the estate endorsed the product. Registered “selfmarks.” Lanham Act 2(a): false association: ability to block other registrations, gives estate leg up in establishing priority. Verified accounts on Instagram: can get a check mark as authentic presence of a global brand represented by a public figure. Allows estates to post and speak as if they are the decedent. Project brackets the right of publicity, which raises different concerns. 

First major concern: continuity problems. Fails to meaningfully distinguish b/t decedent and successor in interest: courts require plaintiffs to demonstrate confusion over endorsement from decedent or estate. But estate can be practically anyone: estate of Marilyn Monroe is an entity Authentic Brands Group that bought the rights to the Monroe IP from the wife of her acting coach. Endorsement by Monroe means something v. different from endorsement by ABG but courts have refused to distinguish those things. 

Second: autonomy/agency of the deceased. Endorsement signals some desired connection b/t endorser and endorsed. Conscripting the deceased into activities they may not have supported. Marilyn Monroe, Elvis, and Muhammad Ali all “participated” in #blackoutuesday, but did that reflect their views? 

Third: discursive problems: controlling cultural meanings of the celebrity. Official account frames her as “Paris Hilton” of her day, downplaying other issues around addiction and femininity; so too w/Whitney Houston. Pernicious b/c posthumous endorsement gives “official voice” to the estate. 

Might still want to recognize some rights b/c they are material to consumers. Purchasing endorsed products can be a way of mourning/processing a celebrity’s death; may also want to support surviving family members/causes the celebrity cared about. Posthumous endorsements can also push back against exploitation—merchandise sold w/names & images of George Floyd and other victims of police violence. Misrepresentations of official endorsement can be bad. This can also happen politically—Reagan Foundation sent a C&D to the Trump campaign for selling Reagan/Trump merchandise. 

What to do? Proposal: recognize endorsement rights not just on the chain of title (as courts currently do) but require privity and power: a meaningful connection b/t rightsholder & decedent. Should be empowered to make legal decisions on behalf of the decedent—personal representative, trustee—w/corresponding duty to their interests. Consistent with HIPAA, Stored Communications Act, attorney-client privilege. 

Finally, where there’s no entity w/legal authority to act on behalf of the decedent but a risk of deception, false advertising law can pick up the slack. 

Ramsey: Monroe used a different perfume than the one proposed to endorse—should that matter? Literature on compelled speech. 

RT: different perfume example suggests that privity/power is quite complex: (1) naked licensing/assignment in gross issue: what would it even mean for “Monroe” to endorse a perfume; (2) conflating materiality of decedent’s endorsement while alive w/estate’s endorsement—courts make it way too easy by allowing conflation; (3) compelled speech arguments in this context make me nervous b/c they are usually deployed against people who have speech interests of their own. 

A: family endorsements can sometimes be material but in a different way—need analysis. [And of course “family” is not the same thing as “estate.”] Would like law to disaggregate these situations. 

Mark McKenna: really stuck on the label “endorsement” for a dead person. Consumers can’t possibly think that’s happened if they know the person is dead. Many of these examples are more about association or dilution than confusion—this person stood for X and is being used for Y. Where the heir is standing in for the person, there’s an extreme level of circularity—only if heirs have the right to license do consumers think it’s the heirs doing the endorsing. 

Jeremy Sheff: unpleasant flashbacks to NY ROP debate. Once you start dealing w/postmortem rights, you get into complex issues of trust & estates & tax law, which your privity requirement brings to the front. The idea of this being an asset passed from one generation to the next is something those bodies of law have things to say about—taxing on its distribution depends on its worth, which depends on its use before death, which might be different from use after death. Or you could do an objective valuation which would be potentially inconsistent w/decedent’s own desires and/or desires of family. How will you engage w/ those bodies of law—Eva Subotnik has been doing related work in © context. 

A: definitely a goal is to bring TM into dialogue with T&E. Even weirder than ROP and © b/c it’s unclear what needs to be assigned to bring endorsement rights. Nebraska courts say ROP is inalienable so there’s no transferable endorsement right; NY courts said that even though there was no post-ROP there’s still an endorsement right. 

Calboli: compare theories on moral rights postmortem. 

Trademarking “Covid,” “Covid-19” and “Coronavirus”: An Empirical Review and Considerations of a Larger Pandemic, Irene Calboli, Texas A&M University School of Law

Also looking at “social distancing” and “six feet apart”—using these as a measure/example of certain issues in the system. More than 600 applications for these in less than 3 months. Looking at a sample: Medical-related (124), unrelated existing businesses like baby clothes (64), merchandising products/promotional goods (135), slurs (also merchandising, 7).  A considerable number disclaim “COVID” etc. and are combination or design marks. Some could in theory be inherently distinctive given the unrelatedness of the goods. Although she’s ok with the merchandising right for established businesses like Mercedes cars, here there are a bunch of clear examples of failure to function. 

Is there a problem? What are the costs for the PTO given that vast majority will be rejected? Time invested by examiners, including possible office actions/appeals. Costs minus fees generated: hard to determine. 

Cost to possible legitimate business from being cut off from using descriptive terms? Costs from possible litigation—C&Ds?  There’s a backlog in the system—too many applications? A lot of waste to sift through. 

RT: Carol Rose talks about how a property system is in some ways a commons of its own; overuse can impose unexpected costs/externalize harms. Here, one possible immediate result is elevating failure to function as a free floating reason to reject registrations (or even recognition as marks). As with slurs, even when the result might be inherently distinctive according to our usual tools for identifying what is functioning as a mark, the non TM meaning overwhelms the usual signals of what constitutes a mark such as placement on the product. (This breakdown in old standards for what counts as a TM is also happening on the other end of the Abercrombie spectrum with Booking.com telling us that consumer perception is all.) I don’t think that Corona Legal Services is going to work as a mark! The corrosive effect of overclaiming on our usual tools for identifying trademark function is itself a problem, and it also raises the question of what comes next to identify TM function. Crystals & Mud in Property Law offers some ways of thinking about how property systems might react to overuse. 

Ramsey: do other national systems have a similar problem?

Beebe: costs on the system are a real question. PTO examiners might like these b/c they’re so easy to reject—helps their numbers! Saw it as a revenue making thing, but there aren’t that many of these; still, on balance assumed that PTO was making money off of these. Finally, people can waste their money if they want to apply for registrations that won’t succeed—not sure what we could do to stem the flow of these (given right to petition!). 

Alexandra Roberts: consider whether these are deceptive marks too—grabbing attention at a desperate time is not good. Consumer health & safety are also relevant. 

Headaches and Handbags: A Fragility Theory of Trademark Functionality, Matthew Sipe, University of Baltimore School of Law 

Placebo effect is powerful & often mediated by TM. Subjects who took marked Nurofen experienced much more relief than unmarked; subjects who took Nurofen-branded sugar pills even did better than people who took unbranded actual Nurofen. Pink is associated with soothing/care in Western cultures, so pink Pepto-Bismol will work better than red or green. Traffix says that a product feature is functional if it’s essential to use/purpose or if it effects the cost/quality. If so, no need to proceed further. TM names of drugs clearly affect quality, as do the color of drugs. We do have cases finding color functional on these grounds, but not the names. 

Traffix language can’t mean what it says, b/c any TM affects costs by reducing search costs and any TM affects quality by increasing the incentives to preserve quality. So lower courts have added de jure/de facto, aesthetic/utilitarian distinctions. But TM has fumbled towards a consensus on functionality. 

Three themes: (1) preference for bright line heuristics; (2) intentional focus on welfare maximization; (3) conceptual unity b/t marks and dress. 

But Nike golf clubs improve consumer performance versus unbranded or irrelevant branded clubs; 3M branded earplugs improve performance on a math test in a noisy space; Red Bull improves performance on cognitive tests v. Sprite. 

There is also social functionality: conspicuous consumption, conspicuous giving. Organizational functionality: a certification mark allows product manufacturers to avoid costs in warranties, demos, etc. Credibly certifying place of origin allows producers to compel buy in for advertising, research, etc. that would be dissipated by free riding.  And design functionality: granting protection to one shape allows producer to introduce variation w/o losing recognizability and brand prominence: Zippo lighter shape and Funko pop general configuration. Also allows them to avoid tags or other marks. Interlocking functionality: makes other products from the same producers higher quality b/c of match/fit: Lego shapes, or Tiffany blue earrings as only match for Tiffany necklace, or John Deere green accessories being only match for John Deere tractor. 

All these functions are fragile: the use/purpose would be destroyed through unchecked copying. Only nonfragile functionality truly prevents TM eligibility in practice. Pepto pink still soothes even if other producers use pink. But the word mark Pepto-Bismol only works if exclusive and would eventually dissipate if freely used. His proposal: any use, purpose or effect dependent on TM exclusivity itself doesn’t count. But any time there’s a mix of exclusivity-dependent and non-dependent functionality, secondary meaning (fragile functionality) won’t save the matter from ineligibility.  Allows us to dispense with aesthetic functionality b/c test no longer needs it. 

Application: overlapping TM & © protection. Fragility model: overlapping protection shouldn’t be allowed in vast majority of cases. Traffix tells us patented subject matter is only nonfunctional when it’s an arbitrary embellishment. If it’s being used for its innovation, that affects cost/quality. So too w/©. Unless © subject matter is being used for reasons unrelated to creative expression, will affect product in impermissible ways. Of course a Batman backpack is being used for its creativity. Dastar thus follows. 

By contrast, dilution and post-sale confusion: to prevent unwinding of signal, have to protect exclusivity. 

Protecting fragility is welfare-enhancing. Benefits that can be shared are, benefits that can’t be shared aren’t. Leaves space/incentives for branding. But there are distributionally regressive outcomes: healthcare, athletic/academic performance. Inferior products to those who can’t afford them. Potentially anticompetitive. 

Barton Beebe: Metaphysically convoluted. Is SCt approach as expressed in reputation-related functionality v. non reputation related functionality already what you’re getting at? Samara Bros. does seem to get at the distinction you’re making. How is fragility different? 

Different dichotomy: exclusivity related functionality versus non exclusivity related functionality? Not just the high fashion brands that advertise into our heads; all consumer goods participate in economy of exclusivity to some extent, including K-Mart. TM law just talks about the high end brands as a matter of class. Other things to consider: (1) Easterbrook’s approach to functionality: a design that produces a benefit other than source identification is functional. Bright line! (2) Consider definition of useful article in © and recent struggles with that. Similar vibe. 

A: For “non-reputation-related”: two problems. (1) They tell us we don’t ever hit that test if you trigger Inwood first. You can dance around this a little w/a gloss on whether this is a two-step test. (2) It doesn’t deal with all these functionalities. The word “significant” is precisely what starts to look like an antitrust rule of reason. 

Fromer: lots of literature about functionality, including Lunney, Litman, McKenna; grapple explicitly how what you say differs from or works with what they say. 

RT: From the paper: “All trademarks affect cost and quality. Full stop. Trademarks reduce search costs for consumers by providing a “consistent signal,”

--assumes the existence of secondary meaning; b/c search costs aren’t reduced w/o that, this description assumes a de facto functionality that may not exist in mere TM function

--framing elides one of the hardest Qs in functionality: it is good to have a mark, or to have a shape, but does that mean that every mark or every shape is functional? No! That’s just de facto functionality.  (Can see that most clearly in the possible response to my first point: even having an unknown TM signals that “this isn’t the TM you know,” thus reducing search costs in some sense, but of course the actual content of the TM is completely irrelevant to that message. The relevant information is that the product/service is “not-known.”) Some of the paper reads like semantic games: you’ve labeled the credence benefits of trademarks as functional, but I think it doesn’t work without a better account of the level of generality at which you are analyzing functionality. 

I wish I thought this reformulation worked, but because of the level of generality issue I don’t think it does any better than current doctrine. Consider the Apple logo or the design of a Rolex watch: they are beautiful designs; certainly under your view of what constitutes an effect on quality they at least marginally increase the quality of the products simply because of their beauty: even on first exposure with no secondary meaning, they are nice designs. I don’t see how your test helps us with the issue that has confounded every attempt to reformulate functionality doctrine: When does that marginal effect, which rests on generalized features of human perception like our preference for curves, override the effect of the specific trademark meaning?  This is why some courts ask for a substantial non reputation related benefit in the aesthetic functionality test. 

Defense of dilution lacks engagement w/empirical literature where that doesn’t actually seem to happen: maybe the things you say are fragile aren’t actually fragile. 

[[Things I didn’t have time to say: Your argument depends in part on the placebo effect being fragile. [Similar to what McKenna is about to say.] Why not hypothesize that the effect an effect of familiarity and allowing the brand name to become generic would continue that familiarity, as with Pepto pink? My take: Especially likely to persist w/pharmaceuticals b/c those are credence goods; I understand why spitting out the generic Diet Coke in disgust would make me distrust labels on experience goods, but the placebo effect works best on credence goods. Might even work for Nike clubs! 

You try to deal with the placebo effect by looking at the research on counterfeits labeled as counterfeits. But telling the consumer they’re using a counterfeit is different than just putting the label Nurofen on it and saying no more—among other things, the effect of telling someone they’re using counterfeits might well be related to a generalized surprise or distrust—see Sprigman et al on dilution: you might also find that performance using a Bic pen also goes down after experience with a labeled-counterfeit Parker pen b/c consumers are just more suspicious generally. Also, as far as I recall the research isn’t conducted on the set of consumers who prefer to buy (fashion) counterfeits—hard to test, but someone who chose counterfeits because they’re cheaper might have a different reaction. 

Similar Q: if we could empirically show, say by looking at China, that widespread counterfeiting and copying does not diminish the perceived value of brands in certain categories, which is to say that exclusivity remained somehow perhaps through subtler quality signals, would your argument have to change? 

[Again echoing what McKenna says] The charitable signalling & certification mark bits seem to be about false advertising: if the law enforced a rule that everyone using the Toms mark donated a pair of shoes to people in need, the relevant signal would continue; it would just be converted to a certification mark, which is related to the reason that certification marks theoretically need nondiscrimination provisions.]]

Jeremy Sheff: does fragility differ from the economic concepts of congestion or rivalrousness, and if so how? 

Jessica Litman: why fragility and what are the other possible terms? The term is an uncomfortable fit for the phenomenon you’re describing. 

Mike Risch: Doesn’t read cost/quality the same way the paper does. Thinks it’s mostly about cost of manufacture, quality of physical product outside of placebo: if made in the way it’s supposed to be made, how does it work. There is definitely uncertainty, but cost/quality isn’t anything that might affect how good consumers think it is or how much they’re willing to pay for it, or you swallow all TM. 

McKenna: you have to assume quality differences for Nurofen to stop working if freed for all to use. We could make sure that everything labeled Nurofen had the same pharmacological characteristics. It’s really the quality differences that derive dissipation, not mere reproduction [at least for those things that aren’t dependent on excluding the proles].  Separately, it’s at least implicit that it has to affect cost/quality in a way unrelated to reputation. The function has to be distinct from reputation, but can exist along with reputation. Courts just struggle when people want the product for non-source reasons but also want it for source reasons. 

Overlapping and Sequential IP Rights, Lolita Darden, Suffolk University Law School

Patent/TM and ©/TM interface: functionality takes us a long way toward weeding out those designs that might create perpetual monopoly. Trying to expand functionality to be more comprehensive to protect those other systems. Distinguish types of protection available for trademark/trade dress. Using a piece of music whose © has expired as a TM. There may be a narrow opportunity for these things to serve as TMs. Right to copy v. need to copy can be distinguished; functionality test proposed is built on concept of need. 

RT: Paper coming at the end made me think that something unites all these presentations: an increased interest in bright line rules and formalism: most of the presentations seem to be coming if not from classic formalism then from rule utilitarianism—the case by case analysis we have been doing in a lot of these situations has become so costly as to lead us to seek alternatives. For this project: what is the product or service at issue? For the airline example, does it include the advertising (theme music for the airline)? If we confidently say no, then drawing the distinction between protectable and unprotectable uses of music becomes easier—but why then is the Batman backpack not a member of the class of backpacks and Batman just part of its marketing? 

A: paper does discuss Wonder Woman merchandise—may be a hard call. 

Calboli: overlaps of protection also raise issues of overlaps of fair use/defenses. One possibility: redesign the system from scratch and create a “brand right” that would just tell you what you could do.

Sheff: we’ve all gone crosseyed over this problem at one point—previous framing was often one of boundaries. Problem in functionality is that sometimes consumers want things for both reputation and non-reputation related reasons. Ultimately we have to make a choice about what we value more, not necessarily as a matter of grand theory but could be tied to particular cases like the Wonder Woman lunchbox, its markets, and its consumers.