WIPIP 2022, Session 6 (TM)

Rebecca Tushnet Bad Spaniels, Deceptive Raptors, and Tiny Hands: The Persistence of Commercial Speech as a Category

Jennifer Rothman has done related work, but her focus has been on the different definitions of commerciality across IP regimes; I’m interested in a different question: holding constant the definition of commercial speech as defined by First Amendment jurisprudence, which is basically speech that does no more than merely propose a commercial transaction, does the Lanham Act cover commercial speech? This basic question has three different answers, all regularly used in any given jurisdiction—this is not a matter of circuit splits. The answers are yes, no, and sometimes, a list both comprehensive and dismaying.

In response to the massive expansion of trademark’s scope over the last century, courts have, mostly implicitly, devised a compromise by which trademark is pulled back to a more traditional anti-fraud-like scope when it is applied to noncommercial speech sold in the marketplace, such as movies, newspapers, songs, and visual art, or used as the name of an organization with dues-paying members, such as a political party or congregation. This compromise explains an otherwise surprising feature of the cases: Political speakers and religious speakers can often expect worse outcomes than “commercial” publishers engaged in noncommercial speech, given the kinds of cases brought against them.  The key here is that when I say fraud, I do not mean fraudulent intent, but materially deceptive effect.

Summary of current treatment:

Although courts have often referred to “expressive” or “artistic” works as shorthand for the scope of Rogers, they have applied it to speech that qualifies as noncommercial under the Supreme Court’s First Amendment precedents—speech that does not propose a commercial transaction and is instead the product being offered to the public.  This is unsurprising: Rogers itself was based on an opposition between “artistic expression” and “commercial speech.” 

While Rogers thus supplies the rule for most of what people think of as “speech,” there are caveats: Rogers has not been applied to disputes about the names of noncommercial organizations. In addition, not all circuits have adopted Rogers, though neither has any court of appeals rejected it.  The Sixth Circuit has adopted Rogers but read “artistic relevance” narrowly, so that certain artistic techniques like rap freestyling might not qualify.  The Ninth Circuit has adopted Rogers but tinkered with what counts as “explicit” falsehood, holding that some non-explicit content might qualify as explicitly false if there’s nothing else present that allows consumers to identify the true source of an expressive work. 

Rogers itself may not fully replicate direct First Amendment analysis of a liability claim against noncommercial speech. In general, the content of noncommercial speech may be regulated only to further a compelling government interest, and the regulation must be narrowly tailored and the least restrictive means of accomplishing that compelling interest.  At least in its “explicit falsity” prong, though, Rogers tailors potential liability for noncommercial speakers more closely to classic fraud, excluding most noncommercial speech from trademark liability. But there are several remaining problems. First, requiring artistic relevance wrongly puts the burden of justification on the noncommercial speaker and invites errors such as that made by the Sixth Circuit.  Second, Rogers does not impose a materiality requirement, and it is hard to see how the government has a compelling interest in protecting consumers from confusion they don’t care about.  Finally, Rogers does not require courts to consider a disclosure remedy instead of the full range of Lanham Act remedies including injunctions and damages, suggesting a failure of narrow tailoring.

Rogers is therefore imperfectly matched to the general requirements for regulating noncommercial speech. But by drastically shrinking the set of potentially trademark-infringing noncommercial speech acts, it does eliminate many potential conflicts with the First Amendment. And an explicitly false claim that a biography is authorized by its subject—something that Rogers leaves actionable—would meet the ordinary requirements for fraud. Thus, in Rogers cases, most non-fraud-like conduct has been excluded from the scope of the Lanham Act, making its application to First Amendment noncommercial speech tolerable.

II.        Gripers and Political Critics

Cases involving criticism of ordinary commercial actors provide a fascinating contrast to both noncommercial speech sold in the market and to political and religious conflicts. Courts have increasingly found that critics simply weren’t engaged in commercial speech covered by the Lanham Act even if there was some tenuous connection between the allegedly infringing speech and a distant potential for the defendant to profit.  The 9th Circuit in Bosley v. Kremer, the 4th in Radiance Foundation v. NAACP—these courts very clearly say that trademark law applies to commercial speech, defined as it is in First Amendment case law, and not to noncommercial speech.

 the Sixth Circuit reasoned that, “The Lanham Act is constitutional because it only regulates commercial speech, which is entitled to reduced protections under the First Amendment.”  Other cases cite Representative Kastenmeier’s statement that the law “specifically extends only to false and misleading speech that is encompassed within the ‘commercial speech’ doctrine developed by the United States Supreme Court.”  They thus conflict with the cases in the previous Part, which apply the Lanham Act to noncommercial speech, albeit with a modified test—it would be hard to maintain that Fred and Ginger is “entitled to reduced protections under the First Amendment” or “encompassed within the ‘commercial speech’ doctrine developed by the United States Supreme Court.” In practice, many of the gripe cases say they are following First Amendment precedents, only when they’re confronted by a subset of noncommercial speech—that which does not solicit the purchase of the speech itself.

But the political speech cases don’t fit into that subset, because the speakers are often soliciting monetary support for their speech, if not exactly purchases of speech. [Quote from Nader] [quote from Radiance Foundation] Likewise, these cases held that the Lanham Act did not apply to noncommercial speech.

This line of cases does more than create a conflict with Rogers and its progeny. It also creates a conflict with a third line of cases: The Lanham Act is applied with no adjustment when political or religious plaintiffs, as opposed to ordinary marketplace actors, bring suit against competitors who claim to represent the true ideology behind the trademark. The result is a near-complete division in the case law: when a political or religious plaintiff sues a political or religious defendant for trademark infringement, it can often win, whereas a commercial actor that sues a political actor will often lose on the ground that the Lanham Act simply doesn’t apply to the political actor’s speech.

III.       Schisms: Political and religious organizations in conflict

It may seem odd that political and religious speakers are the noncommercial speakers still losing in this new speech environment. Perhaps they lose because courts are in practice more sensitive to private commercial interests in speech than public political or religious interests, though my hypothesis is that there really are a higher percentage of fraud-like cases litigated under the heading of political speech conflicts than there are in ordinary trademark disputes, given trademark’s general expansion far beyond fraud.

But the formal justification for the results in these cases is not based in an explicit distinction between “ordinary expansive trademark law” and “noncommercial speech that is fraudulent.” Instead, in the political speech cases, courts speak in blanket terms about whether the Lanham Act applies to political speech, answering “yes” or “no,” usually rendering rough justice in the case before them but creating problems for future political speech cases that present different plaintiff/defendant configurations. In religion cases, meanwhile, there is no meaningful division in the cases: breakaway sects are routinely enjoined from using the names that they believe truly reflect their religious commitments.

While courts in “artistic speech” cases have reassessed the weight of First Amendment defenses over the past few decades, they have not done so within religion-v-religion cases, for reasons that are not elucidated in the cases themselves. Unlike Rogers cases, these cases say, for example, that courts are in agreement that there is no “exception to trademark law for religious, political, and cultural expression” —a framing that accepts trademark’s broad coverage of almost all fields of human endeavor, rather than seeing coverage of noncommercial speech as an extension in need of justification.

While Rogers rejected any consideration of whether the speaker had adequate alternatives to using the plaintiff’s trademark because speakers are entitled to choose their own ways of speaking,  the religious cases embrace the concept of adequate alternatives. Thus, it may not even be descriptive fair use to use the name of the religion from which the dissenters have parted.

As long as believers are allowed to use generic terms to identify their “faith,” they may be enjoined from using non-generic terms as the name of their “church,” a distinction that courts are confident they can make.

While I admit to some skepticism about this conclusion,  anti-fraud principles arguably explain why audience interests outweigh schismatic believers’ interests in describing themselves in ways that they believe truthful. The breakaway sects, by all accounts, sincerely believe that they are delivering exactly what they say: the religious services associated with a faith known to them by a particular name.  To disagree with them in the trademark context does seem to require holding that they are wrong: they are not delivering the true faith, at least according to the name by which it is known. 

Some of the work is also done by the idea that trademark control extends only to the name/logo of a congregation and not to other elements of worship.  But trademark law elsewhere extends far past product/service names to things like a building’s layout, slogans, uniforms, and other aspects of the “product” itself; indeed, the PTO has granted registration for NKJV for Bibles, referring to the New King James Version.  If a church adopted a distinctive name for G-d, then general trademark law would, in theory, allow it to prohibit other churches from using that name. Courts’ use of genericity as a boundary in religious cases seems to be designed to limit sects’ trademark control over anything but a church name or logo —which aligns it with an anti-fraud regime, but not with modern trademark law. 

Another factor contributing to the siloing of religious schism cases is that most analysis of their First Amendment implications focuses on free exercise, not freedom of speech.  Courts have said that applying trademark law to schismatics is not a free exercise problem because the governing law is based on neutral principles of who has priority and who has the legal right to control the trademark.  Under Employment Division v. Smith, that would seem to end the question—just as safety codes can be imposed on churches as they are imposed on other buildings, so too with trademark law.  But the treatment of noncommercial nonreligious speech under Rogers suggests that in fact religious schismatics are being treated worse than other noncommercial speakers, at least at the level of the general principle being applied. The Supreme Court has recently begun to suggest that religions are entitled to a kind of most-favored-nation treatment: any exception or limit on a generally applicable law must be extended to religion. At a minimum, therefore, it would seem that schismatics would be entitled to Rogers-style explicit misleadingness analysis, and that small differences in schismatic entities’ names, or the presence of clear disclosures of nonaffiliation, could be enough to avoid explicit misleadingness. 

Compared to the Rogers line of cases, the religious cases reveal a startling inattention to the defendant’s own interest in expressing itself in a way that is truthful to the defendant’s own beliefs and intended meaning.

Registration works like the last set of cases, indifferent to the commerciality of the registrant’s speech.

A.        Lessons for Trademark Law

First, noncommercial speech both is and isn’t covered by the Lanham Act.  Cases like Rogers often seem protective of First Amendment interests in noncommercial speech, but only because the baseline coverage of the Lanham Act has grown so broad. Compared to the reasoning in the gripe cases—which clearly state that the Lanham Act covers only commercial speech as the Supreme Court has defined it in the First Amendment context, that is, invitations to transact in the marketplace—Rogers is fairly weak tea. It holds out, and occasionally delivers, the prospect of liability for noncommercial speech even if consumers don’t care about the source or sponsorship of the speech at issue, but just want to listen to a catchy song or buy a funny greeting card.

One possibility: we should the political v. political cases and religion v. religion cases out of the “trademark infringement” frame and make them show fraud

Explicitly limiting noncommercial trademark infringement claims to situations that meet the fraud pattern—direct competition between the parties plus deception that is material to consumers—could force courts to confront the fact that trademark law now prohibits immaterial confusion among a small percentage of consumers who are not deciding between two competing products. And that admission itself might be uncomfortable, especially as the level of constitutional protection for truthful, nonmisleading commercial speech has been raised in non-trademark contexts.

B.        Lessons for First Amendment Law

Should We Care About Intent or Should We Care About Effect?

The political and religious liability pattern is Fraud minus insincerity. This requires us to consider how serious we are about scienter as a requirement for liability for false speech. The special scienter requirements for defamation perhaps ought to be understood descriptively as specific to the risks of chilling negative speech about others, since we are very willing to suppress core political and religious self-identification when it seems both material and deceptive to outsiders.

In the political and religious trademark cases, courts prioritize effect over intent. While the case results suggest that courts will intervene to prevent fraud-like outcomes, they should be more explicit about why the Lanham Act is being extended to noncommercial speech in situations where such extension can survive strict scrutiny. 

Also suggests something about Rogers: It is still too weak because it doesn’t care about materiality. Explicit deceptiveness might justify an inference of materiality, but artistic relevance is useless in sorting the value of noncommercial speech and should be discarded. A version of Rogers that focused on fraud would be consistent with the political/religious cases and the griper cases, despite their varying answers to the current question: does the Lanham Act cover noncommercial speech? “Yes, but only where consumers are materially harmed in ways that resemble classic fraud” would be the ultimate rule.

Jeanne Fromer: parallel universes. How would you understand megachurch/religion as big business.

Jessica Silbey: Aesthetics, politics, and religion as categories that the Court uses. IP keeps the categories separate for exceptions purposes, but the SCt probably doesn’t. Do you really want to flatten that out in IP?

Betsy Rosenblatt: Whether these categories are doing work: should it always have the same answer—why?  If Lanham Act is about speech versus competition, it seems natural that different kinds of uses would be treated differently, since they’ve led to different potential kinds of confusion. Church thing is weird but leading to something that feels a lot like confusion.

Felix Wu: Courts in political/church cases are trying to get at and see “more source identification” than other uses. May not just be types of cases but types of uses—naming the organization feels more source identifying. Seems more Lanham Act like. [which is true but emphasizes how weird the “core” Lanham Act liability scope has gotten]

Q [Aviv Gaon?]: theory of selection effects makes sense—which cases make it to litigation affect the words the court uses, not just the outcomes.  [Pathway to how selection effects affect the substance of the law.]

Bill McGeveran: soliciting donations comparing to another kind of organization that doesn’t engage in speech in the same way: nonprofit that fights cancer. Some ways of resolving that could have the impact of burdening P markholders in these categories than similarly situated nonprofits that aren’t religious or political.

Julia Lang: court seems concerned w/loss of control over name of a church like “Methodist Episcopal Church”—“not fair”—those words seem at best descriptive. Who is the TM owner? [The key thing is that courts say that TM’s generic/descriptive categories and likely confusion test for scope of rights work w/o modification for religions.]

Mark Lemley & Sari Mazzurco, The Exclusive Right to Customize

Aftermarket customization, sometimes shoes into art and sculpture, sometimes shoes into more decorated shoes. Satan Shoes from Lil Nas X made from modified actual Nikes. Sometimes there are brand partnerships—Nike and Ben & Jerry’s—but sometimes aftermarket customization adds more brands, like Nikes customized w/Amazon Prime logos. And sometimes brands do unauthorized customizations—McDonald’s customized a PS4 with painted McD fries. Is it art, collaboration, or something else? Brands themselves want in on the act, officially sponsoring artists’ modifications. And sometimes the customizers might not start with an authentic shoe.

Aftermarket customizations also can involve repurposing products—taking a face off a watch but keepign the mechanism; taking a Cartier watch and adding jewels to make it look more like a more expensive Cartier watch; furniture made out of FedEx boxes. Replica auto bodies.

We have a lot of doctrines to deal w/this.

Confusion as to source? Unlikely in most cases, but the shoes that look like customized Nikes might cause confusion.

Confusion as to sponsorship? Will people assume co-branding? It’s hard to tell in the modern world. Balenciaga and Gucci have agreed to mashup crossovers that look like graffiti. It’s really hard to know in this world whether McDonald’s and PS have a partnership or whether Amazon and Nike have one.

Post-sale confusion: applied to kit car cases; he thinks it’s quite dubious people will be confused, but it’s possible that low-end Cartier watches passed off as high-end Cartier watches is an actual problem. Not exactly counterfeiting, but feels counterfeit-y.

Reverse passing off—Lil Nas X makes shoes, not Nike—a bit implausible.

Dilution by tarnishment w/o confusion.

I told you 15 years ago that if we killed off TM use it would need to be resurrected in another guise. Here it is: First sale doctrine (material changes to product may defeat this); nominative use (circuit split on how to test for it); expressive use/Rogers. We suggest that Rogers needs to apply to a category of things that is beyond what has traditionally been thought of as an “expressive work”—things w/content inside them—to TM products themselves where the point of the modification is to express an opinion. Applies to Satan Shoes.

Proposal: flow chart for looking at whether customization is commercial; if not (personal use later resold or unquestionably art) then outside TM’s scope. If it is, then is the customization an expressive work? If so, Rogers. Etc.

Why are we here? We wouldn’t be here 40 years ago b/c of the expansion of TM law and b/c sponsorship didn’t used to work this way; mixing has become more significant for brands and for artists, so it’s currently impossible to know what’s sponsored. TM law as norm follower v. norm entrepreneur; tolerating confusion in the interest of expression.

Rosenblatt: Does identify of speaker matter for Rogers? (1) KitchenAid puts out a coffeemaker that says “Range Rover.” That does something more than merely propose a commercial transaction, but not a lot more. (2) Range Rover puts out a coffeemaker that says “Range Rover.” (3) I take my KitchenAid and put “Range Rover” on it to express my belief that coffee gets me through the day.

Mazzuco: a linedrawing problem does exist. We don’t want to make it depend on the speaker or the speaker’s categorization, though that is relevant. W/o a plan to sell it, there’s definitely no commercial use, but if KitchenAid or Range Rover is doing it, it’s more likely to be commercial speech b/c of the roles they play in the commercial market, but the presumption should be different for an individual. We don’t want it to be based on intent—McDonald’s can intend art.

Lemley: Hard to separate artists from art. McDonald’s doing it has a very different feel.

Rosenblatt: Are fan clubs artists or companies?

Lemley: we want very much to not draw lines on who it is. But it’s not obvious where you should draw the line. If we knew what “art” was the problem would be solved. Fan club probably falls on art side of line.

Fromer: doctrinal framework is focused on the customizer, and less on the original TM owner. But some of your bigger questions have much more to do with the TM owner. Nike has changed its business strategy so there are drops every day, collaborations every day, new colors every day. What Nike has done is occupy the space to make it hard for anyone but experts to understand what’s going on. It’s a deliberate business strategy to target collectors and niche audiences. The TM owner bears some responsibility for causing confusion in the marketplace in a way that didn’t exist before.

Lemley: customization is now everywhere—the Google doodle isn’t the same twice. The idea of a brand that changes every day is nutty from the perspective of 40 years ago. Not to punish Nike for it, but we’ve made confusion part of the business model so saying that there is confusion can’t have the effect on artistic uses that it might otherwise.

Jim Gibson: Dastar might be relevant. Reverse passing off as a theory based on the content of the art might be precluded.

Lemley: interesting that there’d be a cause of action if you took the logo off the Satan Shoes and if you didn’t—that’s problematic.

Mazzuco: there were older cases saying it was ok to make big changes if you took the TM logo off, but that’s gone.

Elizabeth Townsend Gard, Just Wanna Trademark Experiment

Building a brand in the quilting space—podcast, Just Wanna Quilt, with logo chosen by vote. Did a book on © and one on TM and sold out of the TM book in an hour at a quilt industry show. Opportunity to teach TM to law students and quilters as well as other entrepreneurs. Not client-based b/c we all talked about everything.

Problems: JELLY ROLL is registered for fabric, but the public uses it generically to describe strips from a fabric line rolled into a roll. What do we do with that?  Born generic: sewcial, for sewing bee in social media age; but SEWCIAL is registered for services and now sending C&D letters to all the other users of the term. What to do next? As expert in the field, people come to her for advice. TM bullying w/ a public domain quilt, claiming rights over “Dear Jane” as quilt/software. Says you can’t write a book w/o her permission. Why does she believe it’s hers? Because she registered a TM for educational services, software and jewelry; she never tells the PTO it’s quilt-based. What do you do to deal with that as a community? Related: © threats.

Next thing that happened: Omaha Quilt Guild copied Just Wanna Quilt’s logo for its show. Interesting to experience the emotional reaction to the appropriation. 4 years of effort!

Colleen Chien: describing community norms would be useful.

A: They’re all afraid of Moda’s claims over JELLY ROLL. It’s not a source identifier, it’s a fear identifier. Dear Jane and Sewcial are also aggressive in perplexing ways. The PTO enabled it by registering generic terms.

Lemley: internet intermediaries: Redbubbles of the world take down content no matter what the ® is for; that is safer for them then keeping it up and getting sued. Need a counterweight to deal w/internet bullies—obligation to ignore threat letters from them.

RT: file cancellations! For Dear Jane, look at the 5^th Circuit bead dog case which finds invalidity of marks when they’re descriptive for images of X on products even if they aren’t descriptive for the product in the abstract. File declaratory judgments; file suit against the senders of takedowns to Redbubble, who cause concrete economic harm; interference w/business expectations—actually available as a cause of action b/c there’s no 512 equivalent for TM. [Also talk to Rebecca Curtin.]

Jess Meirs: SHOP SAFE is going to make this a ton worse. Speak out against it.

Rosenblatt: you can also file letters of protest for pending applications.  You could find someone who is using Dear Jane or Jelly Roll, or you could seek your own mark that competes with theirs and see what happens. Do your own Dear Jane tutorials.

Q: Dear Jane seems like the name of the thing—generic—not merely descriptive and a clear case of not functioning as a mark. Jelly Roll—there are other names, even if people colloquially say jelly roll.

A: but they do that out of deference to Moda. Look at the age of registrations and whether they’re incontestable.

Gibson: if Dear Jane is also bringing bogus © claims, linking those in the same case could be rhetorically powerful. TM may seem thornier.

A: registered © in book but didn’t disclaim photos therein that were PD.