Sunday, April 18, 2010

Second Trademark Scholars Roundtable, part 3

Session 3: Forms of Defenses. Are the affirmative defenses currently found in trademark law sufficient to vindicate the third party uses that we would like to permit? If not, how best should these be developed? By which institution(s)? What forms should they take? To what extent would the creation of safe harbors be appropriate (especially in the online environment)?

Introduction: Bill McGeveran

Three piles of cases: clearly unlawful, ambiguous, clearly lawful. Discussion of prima facie case was focused on tightening the standards for what’s clearly unlawful, but he’s more interested in making it easier to throw cases on the clearly lawful pile and not waste any more resources on them. Third Circuit’s nominative fair use defense is explicitly to be done after likely confusion, which wastes most of its effect on certainty/resources. The key is that the clearly lawful pile has justifications regardless of whether they are confusing/not confusing or harmful/not harmful to the interests TM is supposed to protect: e.g., functionality. Don’t need to ask about confusion if something is functional. Nominative fair use could work this way as Tushnet suggests, though he has his doubts about whether the third factor can be maintained independent of a confusion inquiry. Maybe there can be streamlined inquiries where we’re very sure where the full inquiry will come out.

Objections—possibility of burden shift. Instead of plaintiff having to prove harm, defendant will have to prove it belongs in that third pile. McGeveran’s not so worried about that; burdens aren’t as important in litigation as we might think. Timing and ease are key: as with §230, the characteristics that qualify a use for the third pile should be easy to identify, and the more complicated it becomes, the less useful it is.

Mission creep: the concern that listing certain things that are certainly lawful invites exclusio unis analysis. McGeveran agrees and would not try to create an exhaustive list, but would use language to warn judges that the list was not exhaustive. He also wants to do work on the front end; a list of exclusions is not the only part of a reform; materiality, trademark use, etc. are also important. Plus we should address incentives to overpolice—see the suggestion of threat actions.

A lot of ways in which we’re thinking about changing the unlawful pile are likely to create new pressure on other categories of liability, dilution and false advertising in particular: litigants will try to move infringement claims to the other boxes, and we will need safeguards there too.

Primary Discussants: Michael Grynberg: Grynberg is skeptical that on-the-fly creation of defenses will work—we have a somewhat more textualist Court with respect to treatment of the Lanham Act. It’s a real problem that we don’t have clean delegating language in the Lanham Act and that common law hasn’t generated doctrines that we can easily expand into broader defenses. Where there is room to innovate, it’s in the cause of action, and materiality can be partially accommodated there, but other screens are difficult to apply at an early/easy stage. The Third Circuit’s nominative fair use test is the best evidence that Grynberg is wrong, but nobody seems to think that approach will spread. We tend to overstate the argument that courts can innovate and understate the extent to which law constrains courts from innovation. This is descriptive, not prescriptive: it’s about how judges currently think of their jobs.

Should we want Congress to take the lead in designing defenses? One’s position may depend on one’s status vis-a-vis TM law. First, perspective of someone regulated by/benefited by TM law. If we think TM law is endangering its own interests, then we want new defenses, but our commitments don’t tell us much about who should be doing the changes. From the perspective of someone engaged in TM litigation, we’d be interested in administrative/enforcement costs. Institutional competence—courts can regulate adjudicative process through elaboration of multifactor test; Congress can put in burden-shifting shortcuts such as functionality. Finally, from the perspective of someone who doesn’t care much about TM one way or the other, and whose concern is about court-statute interactions. This is where battles between textualists, formalists, purposivists occur. Faithfulness to the statute is important—including when statutes contain language delegating to the courts. Respect for statute’s underlying choices; predictability of the law for parties subject to it.

Where courts may be institutionally better: smoothing out conflicts between bodies of law like copyright/TM: Dastar. Similarly, courts may be able to work out statutory/common-law conflicts better: Wal-Mart, where you have various tests and the Supreme Court goes to the common law (though I think Grynberg’s description there is a bit odd—Wal-Mart involved the Supreme Court doing something quite surprising (though welcome) as far as almost everybody in TM was concerned).

FTC analogy: courts can excuse conduct if it benefits consumers more than it helps them. We do want courts to be able to evolve the law. Courts/scholars shouldn’t have to play whack-a-mole as the cause of action expands into the future, but should have authority to deal with them. We need some guidance/direction for courts to scale back potential overreaching: defensive doctrines are important because they remind us that the IP right is a powerful one and is the exception rather than the rule. If defensive doctrines are properly formulated, they can emphasize that our freedoms precede the cause of action. They have to be more than carve-outs.

Mark Lemley: Now he’s profoundly depressed, because if Grynberg’s descriptive claim is right, we’re screwed. We’ve been seeing over the last couple of generations the death of the common law—anything courts do other than mindlessly read statutes is judicial activism. That has taken root in judicial consciousness, making courts less willing to adopt their common-law role. But if so, we’re in trouble. First, none of this legal expansion was created by Congress in the first instance. Second, once Congress gets involved, it almost always ends up reactive, especially where defenses are concerned. Virtually all the defenses are incorporated by the statutes because they were first created by the courts in situations where the defense wasn’t in the statute! Copyright: Fair use, idea/expression, merger, first sale: all created first by courts. TM: descriptive fair use, nominative fair use, functionality, First Amendment/parody, genericide, restrictions on naked licensing/assignments in gross. Congress adopted/vetted them all. Maybe we’ve got all the defenses we need, but that’s not promising.

If we do see new defenses, they are likely to be very cramped, solving a particular case that offended a particular Congressperson—MAI v. Peak problem, which holds that turning on your computer makes a copy and is therefore infringing, and Congress enacts a law making it legal to turn on your computer in order to service it (the particular case) but does not deal with the broader holding. And then we have the exclusio unis problem, where courts decide that Congress blessed the result because it created a narrow statutory exception.

Maybe we could persuade Congress to act more broadly, but he’s seen too much legislation made to believe in that. Even laws with the best intentions are written in ways that create unintended ambiguities; think about the defenses in the TDRA dealing with parody, which don’t cover parody in use as a mark. Not clear what Congress intended (though I bet I know what INTA intended), but Congress definitely isn’t going to revisit the decision no matter what it intended.

So if courts are too timid to play their common-law roles, we’re in trouble. This is bound up with rules v. standards. There is also a tradeoff between predictability and accuracy. Lemley is a standards guy; hates the Federal Circuit’s formalist constructions of patents. TM law is very far on the standards side of the line, though, maybe even too far. Not all TM questions should be survey-based. We need some rules. Could we get them from Congress? Maybe, but also perfectly reasonable to ask courts, because we’ve gotten them from the courts in the past.

We have a lot of ambiguity about being part of the prima facie case/defenses. Fair use in copyright is written as “not an infringement,” but courts treat it as an affirmative defense. TDRA similarly has exclusions that are defined as not dilution, but people talk about them as defenses. This might matter less than we think, but differences do exist. It’s virtually impossible to knock out a TM case on a motion to dismiss if you’re asserting a defense.

From a defendant’s perspective, it’s too easy for courts to wrap likelihood of confusion into an exception that’s part of the definition of the cause of action. Courts that have thought about nominative fair use as pieces of the confusion puzzle have said “of course it doesn’t apply if there’s confusion,” but courts that think of them as defenses—Lending Tree, KP Permanent—set it apart from the confusion analysis.

Grynberg: Many things courts did were interpreting the cause of action, which did require interpretation—Congress didn’t try to specify much about that. Courts can do that, and can balance competing commands (such as the First Amendment). Consider materiality: you can find an argument for it in the cause of action itself. There are Article III reasons for imposing a burden of showing some kind of harm on a TM plaintiff. But a materiality standard of that sort won’t be effective as a screen. So where does materiality come from as a requirement? Where do courts get authority to impose it?

Lemley: Courts have always had authority to interpret the law to serve proper ends in ways that don’t conflict with congressional commands. It’s harder to find materiality in the statute than you might think; it clearly existed at common law, and it’s reasonable for a court to say that a cause of action must involve a showing of harm, but the statute only says confusion.

Grynberg: Then you argue that Congress didn’t explicitly derogate from the common law. But do courts have the ability to say “it’s a good idea to have a defense, because we’ve always interpreted statutes to make them make sense”?

Lemley: Yes. Otherwise we’re taking the view that courts once had this power, and we can still have anything created then, but can’t have, say, prosecution laches in patent, because the problem never came up before.

McKenna: If you think courts have rulemaking authority in cases where the First Amendment is involved, or they’re trying to smooth out edges between statutes like patent and copyright, then McKenna struggles to figure out what, of the cases we’re troubled by/talking about, doesn’t fall within those categories. And the relations with TM are even more complicated than patent/copyright—there’s always been a patent/copyright statute to create the rights, but not a TM statute, and Congress explicitly acknowledged that there’s a back and forth in TM. Arguments for restraint are much weaker in TM. Also important to keep in mind that courts led in creating these broad rights. The idea that they don’t then have the ability to dial it back by limiting doctrines or otherwise seems bizarre.

Grynberg: Courts can definitely create rules that referential uses are not confusing. But he doesn’t think that courts can, without legitimacy concerns, craft broad exceptions. (Faces several questions about whether anyone in the room is asking for anything Grynberg thinks courts can’t do.) It’s the effectiveness of the screen courts create that he doubts: could courts say that keyword ads are beneficial to consumers as a matter of law? He doesn’t think there is a way to do that without confronting confusion. We can’t avoid the consequences of the statute even with the common-law overhang.

McGeveran: Agnostic about which institution could/should generate rules. He understands Grynberg’s argument to be that courts can’t segregate a defense from harm from the definition of the harm, which is where judicial authority lies—courts can say that something isn’t confusing, but not that it’s ok despite being confusing. He disagrees because courts can come up with heuristics that balance various administrative and error costs to figure out what’s confusing/harmful. They have relied on multifactor tests, but they need not. They can create per se tests for cases they are confident are categorically unlikely to confuse. Their selection of mechanism for doing this isn’t constrained by the statute.

Me: This is very obvious in false advertising: look at the development—in recent years, even!—of falsity by necessary implication, puffery, literal falsity, implicit falsity. These categories are, in essence, heuristics. They help manage the costs of litigation and the costs of error. The more up front about that we are, the better we will do.

Bone: he’s not sure that even a common-law court could say “this is beneficial, therefore we allow it”—the justification has to appeal to some principle. Levels of generality matter: look for principles in the statute. Two Pesos is much more wedded to the statutory text than Wal-Mart.

Dinwoodie: Lending Tree shows a process of generalizing from existing law: we know comparative advertising is ok, and that gives us a principle from which we can evaluate things that aren’t exactly comparative advertising but do the same things we want comparative ads to do.

Grynberg: Assuming that we all agree that this is what nominative fair use should be, but a nominative fair use that we are confident is confusing (independent of the test) can courts exclude it from the statute? (General agreement that they can. Much agreement, I think, comes from distrust of confusion measures/worry about error costs/unlikelihood of a real nominative use being confusing.)

Burrell: Historically, legislatures have been very friendly to judicial innovations of this type.

Bone: he thinks the Lanham Act in particular justifies using general principles, rather than confusion, to create rules. Nominative fair use is good for the principles of the Lanham Act, at a high level of generality. When courts confront a new problem Congress didn’t engage, they can reason from other problems Congress did address.

Dinwoodie: similar to asking whether the Lanham Act is based on the common law of 1945, or the common law with its evolutionary principles.

McKenna: Congress does occasionally react—Anti-Monopoly case led to limited response. Nominative fair use: courts have wrongly been convinced that confusion couldn’t coexist with legitimate use; historically, courts were much more understanding that when you use the primary meaning of a word then there is an inherent risk that the secondary meaning will be invoked—it goes back to problems with the quantum of confusion; once we accept that a really low level of confusion is actionable, then there will often seem to be a problem. We could require a bigger showing of confusion.

Litman: we have categories where courts are happy to accept uses—we absolutely tolerate initial interest confusion in comparative advertising. Every court thinks that’s ok, though some don’t call it nominative fair use. Smith v. Chanel was just a holding that this was ok because it was fair competition. When we come to those categories courts are comfortable in recognizing a pattern where the behavior is outside the scope of TM.

McGeveran: wishes that were true, but many courts won’t do that today—still say there’s a need to avoid confusion. More recent example: “There’s a map for that” ad—comparative, but likely would have triggered an extensive multifactor confusion analysis. (Let me correct the record: the lawsuit was pure false advertising. There were no TM claims, which I trust and hope was because the plaintiff’s lawyers properly advised their client (which also engages in comparative advertising on a regular basis). Of course, this goes to the point that the availability of other causes of action—people can and do sue for false advertising based on comparative ads.)

Bently: Judicial development is less likely to do the work we seem to want than Congress—courts are concerned with dishonesty/behavioral norms/taking advantage more than they are concerned with information policy and structural goals. In the context of defenses, the payoff is that they do things like New Kids—references are allowed, but not taking undue advantage/causing excess confusion: these are behavioral norms against free riding. The judicial route necessitates a creeping in of behavioral norms, which will limit the certainty/safe harbors. (Again, I think this is based on later courts’ misreading of New Kids, but that too does a lot to prove Bently’s point. The initial rule did not have much in the way of a confusion component—“do nothing else” is reasonably objective—but later courts thought they needed to add in a full confusion analysis.)


Dinwoodie: But won’t courts do the same thing with legislation?

Bently: you have to create clear safe harbors. (Eric Goldman can probably speak to courts’ attempts to muddy up §230, about as clear as you can get.)

Lemley: Now he’s depressed about both judicial and congressional chances. We have to educate courts about the meaning of TM law, and distinguish consumer/marketplace protection impulses from anti-free riding impulses.

Mid-point summary, Eric Goldman: Basic problem: what are we going to do about all these stupid TM cases we are seeing?


RealSelf, a consumer review site for cosmetic surgery treatments. Consumers provide valuable information to each other—which was good, which they think was a scam. Lifestyle Lift: consumer reactions were not favorable. Lifestyle Lift doesn’t like the negative reviews, and sues RealSelf for TM infringement, on the theory that RealSelf has an economic interest in setting up this forum, so it’s use in commerce, and there are third-party ads on there. The case settled; it’s not frivolous on its face if you look at how courts read the prima facie case. So we need to build a fast lane for cases to get out—a 12 items or less approach.

230 is the prime example of a fast lane, though it’s not well drafted. The way it’s been interpreted is very clear. Most cases in the sweet spot get tossed out on 12(b)(6) or a judgment on the pleadings, before discovery/summary judgment/trial. Ideally we could find a similar TM treatment. We have some work to do on consensus on what qualifies for expedited treatment, but we could probably come up with 3-4 things. We may never get there easily, though, because of the interplay between limitations/defenses and the overall concern with consumer confusion. If confusion infects every doctrinal hook, then you have nothing—you’re always in the same basic situation. We need to be categorical and say we don’t care about confusion in those situations. That’s the problem with Rescuecom: the court wasn’t willing to create a fast lane (even though secondary liability would still be available!).

Cross-elasticity of demand: even if we squeeze TM doctrine, we might push cases into other areas which may or may not be ready. LifeStyle Lift alleged a dozen causes of action. All these doctrines are so soft and malleable, none of them are easy nos.

Much of our discussion has been basic jurisprudential questions: who, how, what. Microcosm of broader issues; need better anchoring in them.

The cloud of C&Ds: another area where we see private adjudication is third-party vendors who set up their own policies and enforce them regardless of the law. Google has a policy: you can’t reference a competitor in ad text if the competitor objects—that’s the law of Google. eBay has VeRo; Twitter has a policy; UDRP. These are adjudicating TM law, or law relating to TMs, and no court has anything to say. CSRs are the TM judges of far more cases than any judge. We can shape those private policies through TM law, but there is no guarantee.

If we all agree, as we do, that the Slip ‘n Slide and Insinkerator cases are bunk, how do we get rid of them on a categorical basis?

Litman: Search for a fast lane is probably futile. If we’ve got a system in which people can claim likely confusion, we can limit it, but the cost of getting sued is huge, not only in TM but all over. Outside of funneling everyone into arbitration, we have a larger problem than TM. If courts reach the right results, that affects settlement, filing decisions; allows best practices policies; but the Realtors can still file suit against the NYT for using the term not as a mark.

Goldman: He agrees that TM gets the right result at the end of the day far more often than not when it’s litigated out. But settlement is so attractive even when you’re in the right. Repeat players are more likely to litigate, but other defendants need a reason to fight back. That’s why the threats action sounds appealing.

Heymann: Lessons from 230? Congressional findings about the intent of the law, which precede 230 in the statute itself, have been used by courts in furtherance of a broad interpretation. So legislative solutions should explcitly incorporate the policies they’re intended to further.

Lemley: Attracted to fees as a possible solution. How does it work in the UK?

Bently: He doesn’t have enough experience in practice to know whether there’s a cloud of C&Ds. And even if he did know, he’d be hard pressed to make relative comparisons. The winner-takes-all approach is regarded as doing something to fuel litigation because it increases incentives to invest in the litigation. Now, conditional fee arrangements are offered by some attorneys.

Dinwoodie: Fee system in UK has existed for so long that litigation culture is different—not as quick to shoot in UK—but it’s not clear that this could be transplanted to the US.

Bone: not a lot of empirical literature; the theoretical literature is ambiguous, and results tend to depend on the information structure of the lawsuit—is the info asymmetric?

Burrell: There’s a thought that you can avoid a threats action by a letter drawing attention to the TM. But most of the threats actions he sees do involve clear threats—while Telstra will understand that a reference to a registered TM is a threat in all but name, if a lawyer is sending a C&D to an individual, the lawyer expects that the individual will not comply unless it is clear that there is a threat of legal actions.

Bently: Threats actions are statutory; they aren’t available in the UK for copyright claims, though they are in Australia.

Bone: it’s a problem of culture, related to the legal background. We have a body of law that credits claims that are overreaching, which means that people feel empowered to make them. If we can change the substantive law, early interdiction isn’t the only benefit—change the culture for lawyers who care, at least a little, about doing the right thing.

McKenna: Clients too like to be persuaded that there’s a good chance they can win. If lawyers can’t show them cases where they can win, they’ll be more hesitant to go forward. The C&D culture depends on a non-negligible chance of winning.

Janis: Too, remedy matters. If all you can get is a disclaimer, that may change incentives.

McKenna: last thing the client wants in the file is a bunch of C&Ds that people ignored, because that looks like sham enforcement. Given that you don’t know what a defendant will do, you need some prospect of winning, though it need not be very high if you think most people will fold. But if the response is “go to Hell,” you have to decide what to do. Now you have a stinkbomb in the file.

Goldman: Doesn’t know of any really good analogue to chillingeffects for TM (though chillingeffects does host such letters); there’s lots going on that would never get onto chillingeffects. North Face lawsuit: changed a lot once the court of public opinion got involved, and (despite a really negative early ruling) it got settled in a way that allowed South Butt to continue doing business. We haven’t seen the same level of scrutiny of stupid TM actions as with copyright.

McGeveran: Best practices, too, has been oriented towards copyright but could add in TM. One story he takes some heart in is what happened during the brief period when dilution had a pretty straightforward fast lane, before the TDRA—Barbie Girl created a fast lane for expressive uses, and a bunch of cases used it in the small category of federal dilution cases. And then Congress fouled it up by larding it up with other considerations that could cycle back to considerations of consumer confusion. (I think this assessment is way too negative. Let’s not forget that the Chewy Vuiton case involved uses that the defendant put “TM” next to—this was not a close case in terms of use as a mark. I think the fast lane for expressive uses remains.)

Lemley: Congress clearly intended to expand the safe harbor, by specifically adding in parody, news reporting, etc. The second thing that happened was a TM use requirement. Then in the course of legislation they got mushed together. This created an issue in Chewy Vuiton.

Hagan: Early on in AdWords, people wanted to complain about all uses of their mark but didn’t want anyone to know. Google said they would explain why to people whose ads were disapproved. That caused some of the complainants to back off and file more specific complaints. So transparency does have a purpose. The problem is when the standard for confusion is so mushy, there will be some lawyers who say that there’s some chance of losing and some chance of winning. The client doesn’t know what the standard means, and is willing to roll the dice. Any sort of enumeration of what’s required will help. (Compare misleadingness, where you need consumer deception evidence.)

She is really worried about opening the Lanham Act up again. INTA etc. are pushing for a DMTA like the DMCA; 1-800-Contacts is trying to make keyword ads illegal in Utah; a can of worms/compromises.

Beebe: TDRA language: can Barbie Girl reasoning still be argued? Has anyone tried it and failed? Kozinski ruled that noncommercial speech was excluded—has that been disapproved by any court? (Nobody knew of a case deciding the issue under the TDRA.) So the fast lane might still be available.

McGeveran: hopes he’s wrong in his fear! The structure of the TDRA changes enough that Kozinski’s analysis might no longer apply (though it might).

Lemley: Exclusion: excludes any nominative/descriptive use other than as a designation of source for the person’s own goods or services, including noncommercial use, parody, news reporting, etc. But the basic definition of dilution is that it’s association between a mark/trade name and a famous mark—so both the defendant and plaintiff have to be making use as a mark.

Me: we are having the debate precisely over whether the existence of exclusions hampers, as a practical matter, the ability of courts to constrain the scope of the main cause of action. If it works well in dilution, maybe that will be a positive sign.

Beebe: Very few cases post TDRA turn on dilution; dilution might not matter.

Grynberg: are we confident that people outside TM think that the Slip ‘n Slide case is a crazy claim? What is the reaction of the public at large when the NFL doesn’t let its logos be used for an Oliver Stone film? This may be different from the public reaction to a small businessperson who gets threatened for selling T-shirts that say “Got Breastmilk?”

Leaffer: negative public reaction comes from oppressing the little guy. Public opinion can be affected with a database of stupid C&Ds; legal change will follow change in public opinion.

Grynberg: the bully narrative does a lot of work. It’s not about TM but about corporate bullying.

Heymann: Prevalence of product placement will affect the answer. Consumers will react differently to North Face/South Butt when the putative defendant is engaging in criticism than to Slip ‘n Slide, where the TM owner’s complaint is essentially that it didn’t get paid for the use. Consumers might just see this as a debate about money instead of about speech. (Though Slip ‘n Slide framed its complaint as one over unsafe use/a use it didn’t like. And note that Slip ‘n Slide would never have gotten paid. It would have paid for placement.)

Lemley: People might not care about this. If we can condition people to think there must be a product placement relationship, does the law have to take that into account? He’s a skeptic.

Dinwoodie: question of deriving principles of commercial ethics. What is it fair or unfair to do? Continental impulse is to ask: who gets paid? If it’s a desert question, that might be important.

McKenna: it’s not negligible that when TM owners pay they exercise control over the content. The free riding impulse is strong among many people, but only in the merchandising cases. In the expressive use cases, there was a much greater willingness to accept that TM owners shouldn’t get control, because it’s clear that control of content was at stake. Importing materiality is only a partial answer. If consumers care about control, we still may need to layer something on top.

Me: Like Lemley, I think this question of what the ordinary person’s reaction will be to the TM owner’s claim of control over non-critical/parodic expressive uses depends on whether we ask about perception in general, or about materiality. Most people might not care who has the right—but they really don’t care who has the right. Even if they believe (which is much less certain) that Microsoft gets script approval over The Vampire Diaries, they do not care—it doesn’t affect their watching decisions, at least not in a positive way.

Burrell: Not confident that bad publicity is a good constraint on big companies like McDonald’s.

McGeveran: the NYT reported straight that a Moneyball movie had problems because Major League Baseball was asserting TM claims to control the script; the reporter treated that as unproblematic. Explicit exclusions are likely to be best for norm entrepreneurship: if you want to teach people that movies are off limits, it’s best to have a law that says explicitly that movies are off limits.

Bone: Norms are not the same as consumer expectations; there is a relationship, but we have to be careful about which we’re trying to effect and what harm we are targeting. Grouping different situations may enable people to go beyond their primitive intuitions--or may convince them that their initial reactions are right.

Eric Goldman said that he’d like to work on some public repository/list of bad TM claims, so contact him/watch his site for more.

Thanks to the organizers, including Vadim Shiffrin of DePaul, who set up a great roundtable.

No comments: