Sixteenth Trademark Scholars’ Roundtable Session 2: part 2

Mid-point discussants:  Rebecca Tushnet

Continuing the theme of offering a series of observations:

Picking up on the relevance of 33(b)/referential use. First, for textualists, it may be true that 33(b) indicates that there is no general use as a mark requirement for infringement otherwise 33(b) would be a null set, but 33(b) also proves to textualists that use as a mark is a real thing. Substantively: 33(b) requires “Use other than as a mark” and solely to describe defendant’s goods/services—if “describe” can include description of defendant’s history or content, then the idea that the First Circuit and the Sixth Circuit have played with that descriptive FU includes many instances of nominative FU is not wrong. There’s no reason we can’t make a use/mention distinction and we’ve seen courts be willing to adopt various linguistic concepts, including synecdoche in genericism and implicature in false advertising law. Maybe we need to start building this out as a concept as part of our efforts to help courts operationalize the intuition that there really are things like TM use.

Keyword ads: WhenU could actually be understood as a “it’s a descriptive/non-speaker-identifying use” case: “Rather, WhenU reproduces 1–800’s website address, «www.1800contacts.com.», which is similar, but not identical, to 1–800’s 1–800CONTACTS trademark.” “The differences between the marks [sic] are quite significant because they transform 1–800’s trademark—which is entitled to protection under the Lanham Act—into a word combination that functions more or less like a public key to 1–800’s website.” [informational functionality!] “Moreover, it is plain that WhenU is using 1–800’s website address precisely because it is a website address, rather than because it bears any resemblance to 1–800’s trademark, because the only place WhenU reproduces the address is in the SaveNow directory” (emphasis added).

Among other things, this provides a potential foundation for the use/mention distinction—there may be typographic overlap but that doesn’t mean there’s TM function overlap.

Rescuecom: Distinguishes WhenU because Google is selling the “mark” – of course those transactions don’t cause any confusion; so it’s one action from column A to find use and then another action from column B to find potentially confusing conduct

But then in the Second Circuit’s recent 1-800 v. Warby Parker keyword ad case, the court finds “dissimilarity of the marks” dispositive—which only makes sense if you think that keyword purchase isn’t “use as a mark” so you’re comparing P’s mark with D’s separate mark.

How to deal with Abitron v. JDI on use? What I’ve been telling students: There are several possibilities: (1) Abitron means Rogers never applies; (2) If there’s source confusion, there’s use as a mark; Rogers is only for other kinds of confusion; (3) Abitron is not relevant, and there can be confusion of any kind without use as a mark; but if there’s no use as a mark, Rogers can be applied (but how do you know what use as a mark?).

My favorite of these: Rogers applies to non-source-confusion theories (side note: maybe possible to distinguish some of the language in 43(a); but 43(a) doesn't even say "mark"! How could Ct say that the definition of use in commerce applies to both?).

Rogers for affiliation/sponsorship/endorsement confusion not source confusion would generally protect the content of a work (though if there was use as a mark then all the theories would be available). Dastar would then deal with attempts to say that the content of the work functions as a mark for itself.

But more generally: A core problem is that 43(a) is as broad as 32 except for counterfeiting and presumption of validity, and that's not useful in speech cases/it removes all the substantive distinctions Congress tried to make.

One fix: insist that 43(a) incorporates a common law style harm requirement (a good idea? give up on the idea that TM infringement should require harm beyond that required by Article III; 32 would become an American version of double identity).

edge cases: Belmora: why did the existence of a registration limit the remedies to which Bayer might have been entitled? A textualist might say, as in Romag, that the statute provides for all of these remedies regardless except counterfeiting. Difference b/t 43 and 32 doesn’t seem to fully explain Belmora—registration preemption? [But the statute also does say subject to the principles of equity, which could reasonably include the difference b/t having a registration and not having one; a disclaimer injunction is an injunction, it’s just not a complete prohibition on use.]

Conclusion: if we can't pry apart 32 and 43(a) in what must be proved beyond validity, then defensive/infringment-side doctrines based on use as a mark will not help

Question re: teaching Abitron: how is it going? Deep confusion!

Leah Grinvald: How to stop litigation early/before it starts—a key concern for small businesses. TM use was supposed to be a limiting doctrine to end confusion early. Lerner & Rowe could have prevented years of litigation for other keyword buy cases, but fact that it was decided on a LOC basis [in the context of courts believing that LOC is always intensely factual instead of intensely factual outside of classes of cases we know quite well] makes that difficult.

Mark McKenna

How much poorer we are for having lost the distinction b/t TM and unfair competition! And the related loss of ontological categories, if TM is not a limited category and anything can be a mark. [Also leaves doctrines like “ideas can’t be marks” sitting out in the cold, looking like anomalies.] Affixation is a formalist concept; formalism is used to delineate the category but there’s a residual role for equity with more limited remedies. Modern version: defendant has to disprove that it’s not using something as a mark, where use as a mark means things much beyond source, and the remedy is likely to be a full injunction. All of this is a reversal of past practice where the focus would be on the defendant’s conduct and not the plaintiff’s (unregistered) right. We’ve adopted a model that’s entirely functionalist in asking whether something is a TM, in a world where consumers are used to seeing marketing efforts of all sorts—broad recognition of services is another thing that changed in the 20^th century and provided the foundation for rights in gross type thinking—celebrities are considered marks for themselves. In re Lizzo: makes you think of a celebrity and is therefore functioning as a TM on the front of a shirt.

The impulse about use as a mark in JDI and Abitron hearkens back to the old central concept of source: who made/stands behind the quality of the product, but even there the SCt couldn’t commit—TM cares about that the most, but the Court says nothing about what else TM cares about and to what degree. To recover TM use, we will need some formalism, and we will need to accept judgments at a wholesale level rather than a retail level. It can’t be the case therefore that secondary meaning can override those rules; they should not be default rules.

Graeme Dinwoodie: 43/32 divide in remedies: there are cases granting more limited relief, like Belmora, Blinded Veterans, Newman’s Ferrari dissent—there are datapoints you can use but their mindset is different from that of Alito’s. The capacity to do that does depend on the methodology the court is willing to adopt.

Jessica Litman: started teaching TM before Two Pesos, which merged 43(a) and 32. Before that, what you had to prove to get relief under 43(a) was close to what RT talked about in terms of proving harm, though you had to prove it in the protectability part of the case. You had to prove entitlement to relief as to what D was doing. There isn’t much reasoning in Two Pesos, and many sensible post-Two Pesos SCt rulings were attempts to undercut it.

Ramsey: “Lifeguard” mark for T-shirts—company has gone after many other uses for front of shirts. Benefit of treating failure to function separately from distinctiveness is that you can acknowledge that “lifeguard” on a T-shirt communicates that the wearer is a lifeguard. It shouldn’t be limited to “widely used” phrases since things like TRUMP TOO SMALL are not widely used.

Lemley: Gringo case rejecting Rogers application: doesn’t make any sense to say that D’s use is TM use b/c P’s use as a title is TM use, so he’s against symmetry in that sense.

Burrell: registration system in Australia can make it more attractive to apply to register a parody mark even if you don’t intend to enforce it against other parodists—given the admin context and the fee-shifting it might make more sense to go through the cheaper registration proceeding if the target decides to oppose, and then if you have a registration you have stronger standing against the parody’s target.

McKenna: need to really insist that whether D makes TM use has nothing to do with whether P makes TM use.

Beebe: 43(c): use in commerce—seems to impose burden on P to show D makes use in commerce to show dilution—that it was source-identifying. How to meld w/exceptions such as fair use other than as designation of source?

Lemley: Dilution clearly has TM use—D’s use of a mark or trade name! But if we’ve concluded that it’s use for confusion purposes, maybe it’s also use for dilution purposes.

Dinwoodie: Rescuecom: Leval says something has to be used or displayed for infringement, but not as a mark b/c he says an infringing D can’t make a “bona fide” use. He wants to keep some part of the statutory definition even for defendants. But thinks G does use/display the mark. Louboutin, then, is a case where the court says there was no use—they weren’t saying there was no TM use, they were saying that there was no use at all. [Dispute over whether that makes any sense, is covert defense-side functionality done by limiting scope of mark, or omits the necessary confusion analysis.]

McKenna: Lettuce Turnip the Beet does the best job harmonizing this by saying that tag use is different than front of shirt use.

Janis: it’s hard for students to figure out what the fuss is about with use until you get to Jack Daniels. [Relates to Lemley’s point that maybe Abitron should be taught with the limiting doctrines—situations that are pulled out of the multifactor LOC test.]

Sheff: we have areas of law where balancing tests are the rule, but also special per se rules for certain situations as well. Example: takings jurisprudence: we do regulatory takings balancing except if there’s a total loss, etc. Attempts in TM have failed—remember the internet troika? They fail by being folded back into the multifactor test. NFU: 2d Circuit folds the factors back into the LOC test. Parody cases: we don’t carve it out but we fold it into our consideration of the multifactor test. Not sure exactly why.

Ramsey: JDI creates an opportunity to use that relatively narrow definition of “use as a mark”—use to identify your own goods or services (so referential use would not be TM use).

Stephen Baird: Louboutin example is a good one b/c the registration was far too broad—the sole didn’t pop without contrast—no highlighting, no use.