Sixteenth Trademark Scholars’ Roundtable Session 3 continued

Midpoint discussant: Laura Heymann

Is the goal consistency? Is the goal limiting principles that can end a case early? Is use the right tool? Is it a proxy? To what extent should we accept the rest of the landscape as fixed, like the merchandising right and students’ intuitive sense that there’s a property right involved?

Context: how much does context matter to the question of use? Should we consider other similar activity by the defendant (VIP’s other products, for example) to figure out whether any particular thing is a “use”? Should we consider other words/symbols on the product to figure out what is “use” versus “nonuse”? Can we test/survey for this?

Context also comes up in the question of whether booking.com is a mark or an address—same lexical unit can serve both as a mark and other things (the use/mention distinction analogy). Descriptive fair use is use, but fair, but what about non-use? Are our main concerns about expressive uses like titles, or technological uses?

JDI says we care about confusion or apply a screen depending on what the defendant is doing—a D making a TM use leads us to care more about confusion.

Jessica Litman: Lemley proposed a deal with the devil: a property merchandising right for a subset of marks; in return we’ll get old TM law back for everything else. Ramsey hoped that failure to function would save us. Some of the work TM use does overlaps with validity doctrines. PTO likes the F2F tool but courts may or may not get on board with it.

Is part of the problem that non-TM experts don’t understand the principle that TMs are supposed to be separate from the product whose source they identify—they’re supposed to be different things! That’s what symbol/sign relationships are! TM claimants want property rights in the thing they sell, but could we disseminate more widely the principle of the Restatement of the Law of Unfair Competition that TM is for that connection & when that connection is absent b/c the TM and the product are unified, there’s no TM claim—that could help courts pick the good guy in the room when deciding cases. TM law is not a good fit for the merchandising right. It’s not supposed to protect intrinsic value. Merchandising right shouldn’t exist—© is sufficient in many cases, and people should be able to signal their fandom for the U of Penn through both licensed and unlicensed merchandise.

Lemley: use of a mark v. use of the identical lexical content of the mark—is that a useful distinction? Why does it help us? All these use arguments are about what purposes the use is serving, and he wants to know about non-word marks in this system. One could use a logo in a non-TM sense, but is there a parallel to the analysis of “identical lexical content”? What about a rotten apple logo to make a comment on Apple?

Heymann: Keyword sales are not about meaning but about matching.

Roberts: lawsuits against UberEats and similar—when they go to UberEats users see logos; nonpartner restaurants object to use of names and logos. Logos seem worse.

Lemley: but see negative matching which does try to target when users are searching using the TM specifically. FTC’s Business Impersonation rule says it’s deceptive to use

RT: Responding to Heymann’s argument that JDI distinguishes b/t situations where we care about confusion & says that Rogers means we don’t care when the D is making an entirely non-TM use in noncommercial speech. You can also, and Rogers itself did, make an empirical claim about the likelihood of deception for creating strong rules justified by error and litigation costs: Dastar, Walmart. That also helps strengthen the Second Circuit’s principle that in non-Rogers cases asserted against noncommercial speech (that is, title v title), First Amendment concerns should lead to a more careful LOC analysis b/c false positives are bad there too—they’re just even worse when the P lacks its own expressive work, justifying a categorical rule like Dastar or Walmart. That’s not confusion-indifferent, it’s confusion-risk-balancing. This is an argument we should keep making.

To Lemley’s Q: Yes, there are NFU image cases. Mattel v. Walking Mountain is NFU of Barbie’s trade dress. Phone reseller case: logos are NFU to identify which phones the reseller accepts. [Kozinski, naming NFU, characterizes it as non-TM use.]

Fromer: can functionality doctrine help us think through this? If something is being used for its intrinsic value, it’s not TM use.

Lemley: this is where VIP’s language can be abused. Walking Mountain: if Mattel says the art advertises itself, then the presence of Barbie advertises/helps sell the art [though I’m not sure that I’d call that a source-identifying function—it seems to be one of the other functions the Europeans talk about].

McKenna: biggest mischaracterization of Rogers by JDI is that it involved an entirely non-TM use with no TM function; Rogers itself was very clear that the title served a “hybrid” function and still needed protection.

LOC test itself: survey procedure for Eveready is built on concerns for a reading test—you ask people “who makes this” and if they answer “Eveready” they could just be reading. You have to ask “what other products does this company make?” Otherwise a survey would find confusion between the multiple Deltas. Which is to say: Using the same word isn’t the same thing as using the same mark, which brings along with it indications of source. The LOC test is asking whether it’s a use of the P’s mark or just the same word. Where the factors break down is where there’s no dispute that the use is referential. Eveready is the opposite of referential use. Merchandising cases invert the argument, so forcing them into LOC multifactor tests doesn’t work well. You need something like NFU or Rogers because the factors are more likely to be wrong.

Stephen Baird: NFU—can you use logos of companies to indicate that you can buy their shares? TM attorneys would be suspicious of the “too much” problem. New Kids footnote: a soft drink competitor can compare to Coke but can’t use Coke’s distinctive lettering, citing a VW case; cf. copyright fair use. That footnote is dicta within dicta and has been overread. Most TM lawyers adopt the view that you can never use someone else’s logo; one of the things that weighs against that was “strength in numbers.” In a car wax ad featuring only a Porsche, Porsche won. But when there were multiple ads w/multiple different luxury vehicles, Porsche didn’t sue—when there’s half a dozen different brands in the ad, it’s hard to say that any one of those is sponsoring/endorsing. [That logic also helps explain the Xfinity case above; I also think about the “collage” reasoning in ETW v. Jireh and Hart v. EA, albeit that’s ROP analysis.]

Burrell: as a person from a registration system, the idea of being able to use the thing that is on the register blows his mind.

McGeveran: “too much” is one of the NFU factors that attempts to be a proxy for real confusion, and it is imperfect.

Dinwoodie: Likes the 3d Circuit version of NFU where you have to reflect the true relationship—it’s a matter of objective truth and that’s an easier thing to evaluate [query whether the 7^th Circuit case about attributing responsibility to architects calls that into doubt].

McKenna pointed out that the 3d Circuit LendingTree case does have a tiny bit of logo use (a photo of a woman with a Coldwell Banker sign partially visible) which the 3d Circuit doesn’t say anything special about & just remands.

Sheff: can we identify where confusion creeps in to the alternative doctrines—functionality, we tolerate confusion; descriptive fair use, we tolerate some confusion. Implicitly the amount of confusion will sometimes be enough to disqualify the defense. With descriptive fair use, the consideration of confusion creeps back in as “use as a mark”—w/enough confusion, the court will find use as a mark. Where will confusion concerns creep in to various defenses? [I don’t think this is how functionality works (aesthetic functionality maybe); and the KP Permanent courts, especially on remand, also think that amount of confusion matters on its own as well as creeping back into use as a mark.]

Lemley: SCt recognizes that defenses aren’t very helpful if confusion defeats them—two ways to do it. KP Permanent remand: hike the standard so 15% isn’t enough. But other defenses like NFU/Rogers can say that confusion just doesn’t matter. Empire’s use of “Empire” for a music company might be confusing to a number of people.

[I would say again that the other way to see it is that in NFU/Rogers cases an individualized determination of confusion just doesn’t matter b/c it’s so likely to be wrong—b/c the only direct evidence, surveys or other reports from consumers who have misconceptions about the legal requirements for parody/commentary is not reliable.]

McGeveran: in an environment of second-rate textualism, calling Rogers a pre-screen as VIP does (and Rogers did not) creates the risk that courts will reject it more than explaining it as a set of rules for more reliably, less false positively determining LOC.

Baird: Declaratory judgment under Rule 57 is equitable—maybe a way to get the judge to focus on invalidity early on?

Lemley: keeping in mind you can’t get a DJ on something that would have to go to a jury.

McKenna: one way to see the defenses is confining LOC to the core—source/passing off. The confusion has to be in the heartland of TM b/c of these other competing interests—descriptive FU, NFU, and Rogers all have these limits. It’s not that there’s no TM right at all, it’s that it’s not triggered without D pretending to be the P. [I’ve characterized this as D speaking as P for noncommercial speech.]

Dinwoodie: EU’s central function doctrine, that D’s use has to interfere with the central function of a TM, is operating as what we might think of as a TM use requirement. The problem is that they then say confusion = interference w/central function. Once you get into the morass of non-central functions, it gets worse.