Monday, January 09, 2017

10th Annual Evil Twin Debate,"Free Speech or Free to Reject?: Trademark Registration & the First Amendment”

Sponsored by the University of Richmond School of Law and the AALS IP Section and hosted at Golden Gate University School of Law.   Mark Lemley & I debated, and though I usually try to point out that my summaries may not be complete, here my summary of Lemley's statements is very disjointed, because in this case I was trying to prepare my rebuttal while taking occasional notes, so for a fairer explication of his arguments you should read his amicus brief on behalf of Dykes on Bikes, linked below.  I also hope a recording of the event will be made available.

My presentation: Four big things the parties and amici disagree on, and some thoughts about vagueness and inconsistency in 2(a) refusals:

First: Whether 43(a)(1)(A) incorporates the same policy based limits as registration under section 2.  If 43(a) protection is still available when use of an unregistered symbol causes confusion then the impact of denying registration will be much less.  The problem for those arguing that 43(a) is still available is that the text of the provision leaves us with no real way to distinguish the bars in 2 that we definitely want to apply to 43(a)—like the bar on protecting generic terms as such—from bars that we don’t.  So, you get cases like Belmora in the 4th Circuit, which essentially removed territoriality as a limit on 43(a).  By contrast, the Renna case from New Jersey said that a section 2 registration-barred mark couldn’t be protected by 43(a) and explained why it thought that such a result was sensible in light of the underlying policy behind the Lanham Act and questions of administrability.  Regardless of which way one comes out on Belmora and Renna, the point is that this question has implications far beyond disparagement, and I think that it’s a bad idea to try to answer that broader question merely as a waystation to the First Amendment question about disparagement.

Second: TM’s effects on non-owners or non-registrants. In what way, if at all, should it matter that granting Tam a registration enhances his capacity to suppress the speech of others? Most of those on Tam’s side, Lemley’s brief excepted, don’t say much about this.  Relatedly, and the source of troubling inconsistencies in those on Tam’s side, is a trademark commercial speech? Tam says it isn’t always commercial speech and some amici say it never is because there’s always non-trademark meaning in a mark; if that’s so we’re subjecting infringement cases to far less scrutiny than necessary. Implicit in Tam’s argument seems to be that trademark owners have messages but infringers don’t – except that Tam also says one of his objectives is to prevent a racist band from calling itself the Slants.

From Lemley’s amicus brief: Dykes on Bikes wants to control commercialization of the mark.  Which is to say it wants to prevent other people from using the mark in expressive ways such as on T-shirts and mugs.  Mark Lemley and Stacey Dogan, among others, have argued that this shouldn’t be understood as trademark use, but at the very least it has both commercial and message-delivering aspects.  The difference between the infringer’s desire to speak and Tam’s situation is, supposedly, the harm done by confusion—but as we know, there’s usually no such evidence of harm, and no legislature has made factual findings that confusion itself does harm.  Lemley’s brief gives another example of the benefits of registration by pointing to his client’s ability to use the Sunrise registration process to control a Dykesonbikes.xxx domain  name registration: like it or not, though, that website would have expressed a message, and not one inherently based on confusion with the existing organization.

Third: What framework to analyze 2(a) under: is it government speech, commercial speech under Central Hudson, viewpoint discrimination, unconstitutional conditions, or something else?  One of the problems is our ornate First Amendment doctrine, which grows new protrusions every year.  But another is the problem of applying the First Amendment to government programs that change the structure of the marketplace; trademark registration, not to mention trademark infringement doctrine, does that, and so does the existence of public schools.  When the government is deeply embedded into the structure of an institution, how does that affect our understanding of the government’s freedom to make choices about what to support, as long as it’s not fining people for speech or sending them to jail? I conclude that the best framework is unconstitutional Conditions: the government can’t leverage its program to affect speech outside the program.  And on that framework, it’s clear that there is no effect on Tam’s ability to communicate his message outside of the trademark context. He doesn’t have to give up the name to get trademark protection for the content that includes the name, the way that a prohibition on copyright for disparaging works would require the author to give up on disparagement in order to get protection.

Fourth: Viewpoint versus content discrimination.  The line between the two can be unclear, and I think especially in combination with the previous problem it is very important to understand exactly what the government’s trying to do.  Is the government trying to suppress discussion of racial issues?  Stigmatize abortion?  Or something else?  For better or worse, the government in 2(a) disparagement is trying to limit registrations to those not perceived as disparaging to a targeted group.  To me, that is content discrimination, not viewpoint discrimination.  The government doesn’t care which person or group you target. 

The government also doesn’t care why you chose the symbol at issue—maybe you want to reclaim the number fourteen for non-Nazis, or maybe you want to show your Nazi bona fides; either way, it’s the reaction of the targeted group that matters and not your malign or benign viewpoint.  Of course your choice will express a viewpoint; that’s not the issue.  The issue is whether the overall system, not just the outcome in your case, discriminates against the holders of any coherent class of viewpoints.  One can say that true threats express a viewpoint, but that just highlights the point that it’s possible to reframe almost any content regulation as a viewpoint regulation.  In my opinion, the better path is to ask: why do we care about viewpoint regulation more than content regulation?  As the Supreme Court has previously said, viewpoint discrimination is when the state is “attempting to give one group an advantage over another in the marketplace of ideas,” and that’s just not what the disparagement bar is.

Relatedly: The applicant’s viewpoint about the term or the targeted group matters not at all. The understanding of the targeted group is the key. And many on Tam’s side says that consulting the views of the targeted group constitutes viewpoint discrimination, even though heretofore we’ve used the idea of viewpoint discrimination to identify discrimination based on the speaker’s viewpoint.  You can, I suppose, create a new category of viewpoint discrimination “in the air,” to borrow a phrase from Learned Hand.   But if relying on the reaction of the audience to determine registrability makes disparagement viewpoint based then so are descriptiveness and deceptiveness, also dependent on audience understanding.

Final issue: vagueness.  I think this is frankly the most serious issue for the government, and here I do have to appeal to the nature of registration as a government program in its defense. We’re up to half a million applications a year.  If we are going to have registration as an individualized determination at that scale, we are going to have inconsistencies.  My analogy is to public schools: if we’re going to have them, then some history teachers are going to give better grades to the same essay than other history teachers (which some plagiarists have surely learned to their sorrow).  We cannot have the scale of the government participation in the marketplace without inconsistency; if we want the program to continue to exist—which not everyone does for public schools, but Tam’s supporters definitely want registration to continue to exist—then we have to deal with inconsistency in procedural and not substantive ways. That is, we have to provide mechanisms for making records and challenging adverse decisions.  And we have this, though there’s a reasonable argument to be made that the TTAB should take more account of precedent.

If we decide that disparagement is too vague a standard, then it will be difficult to distinguish almost anything in trademark that we discuss with our students—the line between suggestive and descriptive, which has significant consequences for registrability; so too with the line between descriptiveness and genericity; the standard for confusion refusals under 2(d) as well as the infringement standard under 32 and 43(a); the definition of false advertising; and so on.  Someone like Justice Breyer, who likes to balance, could readily conclude that the game isn’t worth the candle for disparagement but is for these other exclusions, but most of the Justices on the Court are uncomfortable with that kind of ad hoc balancing.  I fear that if they accept the vagueness argument they will just encourage other courts to lie about how vague the other parts of trademark law are, and that makes our law less law-like.

Vagueness & the contradictory Dykes on Bikes rulings, discussed in Lemley’s brief, where the word mark was allowed but a later stylized word mark was refused even though it was just a narrower version of the word mark: Great example because there’s a recent exact match in the genericity space: Churrascos, which was allowed as a standard word mark for restaurants, then rejected for being generic when the registrant tried to register a stylized mark.  Is this dumb?  Yes.  Should we change the rule so that there’s more issue preclusion at the PTO?  Maybe, though I’m concerned given the installed base of decisions, and given the lack of true adversity in many of these cases in which the applicant is the sole source of information about the meaning of the applied-for matter.  But more to the point, I think we should be asking administrative procedure questions about dealing with inconsistencies—if we constitutionalize this process, then we’re inviting the process to grind to a halt, or convert into a filing-based rather than truly examination-based regime.

Mark Lemley: Viewpoint/content doesn’t matter for the outcome because Tam wins either way.  Under any kind of reasonable way, this has to be viewpoint based. Redskins, if it goes down in TM office, goes down b/c it’s offensive to Native Americans. But “Native Americans” is ok.  They want to discourage bad messages and not discourage good ones.  Sorrell: Regulation b/c of disagreement w/message it conveys is subject to heightened scrutiny.  NAACP is registrable but not Nat’l Ass’n for Colored Persons—that’s classic viewpoint discrimination.

At the very least, this is content based restriction on speech. They are applying this new and special law b/c of the topic discussed. That makes it content based.  It is true that the justification is that people are upset by disparagement.  Heckler’s veto.

Our odyssey: the only evidence was that lesbians were perfectly happy w/our use of the term dyke.  What we did see was that the second examiner looked it up and found bad terms including dykes. Counsel said “I’m a dyke,” and you could hear an audible gasp—this examiner was personally shocked.  Even if we say we will apply it neutrally, we use gov’t as listener, gov’t as decisionmaker, and that’s why you get crazy inconsistent decisions.

Can’t pass strict scrutiny under content based. Gov’t has no interest justifying rule other than suppression of speech itself. Less restrictive alternatives. Gov’t can disassociate itself from the speech. Could say on each registration.  Could even probably get away with only not endorsing some of them.  But there’s no way in which they can survive strict scrutiny.

Commercial speech/not doesn’t matter. TM can promote and repress speech. But that fact doesn’t change the fact that what we’re doing when we regulate disparaging speech is regulating things w/real expressive impact.

We’ve all assumed that 43 protection was available w/refusal of registration, but there aren’t any cases saying that. Most of the other aspects of 2 have been incorporated into 43, so it wouldn’t be surprising to have this one either.  If you’re restricting the right to have TM protection at all, that’s well beyond unconstitutional conditions.  This isn’t just a gov’t benefit; it has real significant consequences in enforcement.  It was impossible for a small nonprofit to prevent blatant knockoffs w/o registration: eBay, Etsy, CafePress all want registrations to pull merchandise. If you can’t show them TM registration, they say go away. That’s perfectly reasonable; they don’t know who has common law rights or who’s the first user.  At least TM registration makes them comfortable. Practical effect isn’t just additional remedies.  It’s the difference b/t effective enforcement and lack of effective enforcement.  Even if that weren’t true, unconstitutional conditions cases don’t necessarily apply—just b/c I can impose a harm on you, doesn’t mean I can do it for reasons or purposes themselves directed at the repression of speech.  Even w/ no right to benefit, some reasons on which gov’t may not rely.

Real problem: even if we try to fit this in unconstitutional conditions doctrine, doesn’t avoid scrutiny of purpose.  US v. Playboy: distinction between burden and ban is just a matter of degree; burdens must satisfy the same rigorous scrutiny as bans.

If this is unconstitutional, what else goes down with it?  This is a slippery slope argument. Refuse to do the right thing today for fear of being forced to do the right thing tomorrow.  If we applied rigorous First Amendment, most results would be salutary. We might lose scandalousness.   One way might be to decide this on viewpoint discrimination and then say the rest are just content-based.

RT: It’s not viewpoint discrimination.  Other groups are protected to the same extent as Native Americans. That’s not viewpoint based b/c it doesn’t promote any group over any other.  Content based sure, as is the rest of TM law.

Note all the quotes in Lemley’s presentation about prohibition, and ban, and suppression, and punishment—these quotes make sense when you are talking about sending people to jail or fining them for their speech. But the government deciding not to include specific words in their program is different.  If Tam couldn’t get a registration until he gave up using Slants, everything Lemley said would be correct.  But he can, so it isn’t.

Sorrell: disagreement w/the message = heightened scrutiny.  Consider this statement: The Holocaust happened. Can public schools teach this and give students bad grades if they don’t acknowledge it on tests?  Don’t want to impose Sorrell’s standard when the gov’t is creating the program, and that’s why we have unconstitutional conditions.  Can fund abortion and not contraception or vice versa.  Democracy promotion but not authoritarianism (or perhaps, upcoming, vice versa). 

All of TM law is content based.  That’s the problem. The rest of TM registration can’t survive strict scrutiny either; the less restrictive alternative is always fraud, or even the common law.

Gov’t as decisionmaker: he says that vagueness of disparagement is why you get crazy inconsistent decisions.  But it’s not—Churrascos is the same. The size of the program is why you get crazy inconsistent decisions.

Gov’t interest is nonparticipation in disparagement by gov’t.  Shelly v. Kraemer type rationale.

He says: “what we’re doing when we regulate disparaging speech is regulating things w/real expressive impact.” But that’s the point: because the alleged infringers will have an expressive impact too, there is no neutral position from which to evaluate a rights-granting program.

If we applied rigorous First Amendment, he says most results would be salutary. I like that with infringement, sure, but not so much ITUs and nationwide protection and extension of protection under Paris Convention.   And TM plaintiffs get to yell property rights and not have the same rules apply to them.  Note the majority in Tam said that protecting investment was unquestionably one of the okay purposes of TM law, so I fear that only some parts would get real scrutiny.

Jake Linford: why doesn’t SCt pretend that TM are just like © and find some traditional contour—there’s enough 1A protection in the system already?

Lemley: That would be wrong. There are substantial differences explaining why the court hasn’t said a 1A free ride. You’ve built in more speech protections to © than TM. Fair use might or might not be sufficient, but it’s an effort to protect expressive uses.

The Court is influenced by existence of Patent/© clause in Constitution; couldn’t say that with a straight face re: Commerce Clause.

Linford: recent UHS of Delaware case: pleading standards for 32 and 43 increase the requirements for success under 43: does that shift your perspective on harm to TM owners who can’t get registrations?

RT: No, b/c I think it’s ok not to give 43 protection to disparaging marks.

Jennifer Rothman: Whether it matters who’s filing w/r/t disparagement.  It does seem that perspective is based on who’s filing.  Dykes on Bikes. Even if it’s just content based, not convinced that gov’t interest is sufficient, b/c it’s not gov’t speech.

RT: I don’t think the record bears that out. Democrats and Republican examples, not to mention all the Bitch and anti-woman marks, suggest differently.  They go all different ways.

Gov’t doesn’t have to be speaking to have an interest in not endorsing.  Shelly v. Kraemer is not about public perception but about what positions the gov’t can accept without implicating itself in discrimination, e.g., it's acceptable to call women the c-word.

Irene Calboli: A child wearing Dykes on Bikes would be sent home from public school. That’s suppression of speech. How do you reconcile that?

Lemley: True in elementary school; untrue in college; shouldn’t be true in high school.  If they throw me out of college, I’ve got a slam-dunk First Amendment case.  We pretend that kids aren’t actual people.

Ann Bartow: My sense became that it mattered if someone complained; that can help make sense of inconsistent outcomes.

Lemley: that certainly matters.  In Dykes on Bikes there was one guy who thought lesbianism was bad; lost for lack of standing.  My sense is same as Rothman’s—this is not neutrally applied, but gendered, even as it is supposedly protecting minority groups.  Systematically reinforces some standards.

Bill McG: prediction?

Lemley: affirm.  Viewpoint; or content discrimination with bad justification; may not say anything about the rest of the system. As a practical matter, there’s not much reason to be concerned.  Beyond scandalous marks and maybe emblem or insignia, where there is a plausible confusion based or search cost based justification, courts will just continue as is.

Tyler Ochoa: Does not registering suppress speech?  Granting TM gives Tam a right to suppress others’ speech: on balance registration suppresses more speech than it encourages.

Lemley: always true that there are speech interests on both sides. It’s burdening Tam’s speech to shunt him towards certain types of speech rather than others. That’s problematic under 1A even if the result is more infringement.

RT: The registration will also shunt potential defendants towards certain types of speech rather than others.  Lemley is right that it’s impossible to add this up and that we should probably look to reasons instead.  Then we disagree about whether the reasons are ok.

Q: Institutional counterfactual about vagueness: The PTO is weak; it can’t write substantive rules. In an alternative universe where it were like the EPA or SEC, would a code have helped?


Lemley: it would have helped if you think the problem is vagueness, not an affirmative burden on speech.

No comments: