IPSC session 6

Trademark Theory and Empirical Work

Barton Beebe, New York University School of Law, Is the Trademark Office a Rubber Stamp?

What proportion of TM registration applications result in registration? How do they vary by filing basis, type of mark, country of origin, type of applicant, class of goods? What proportion result in publication? What proportion proceed directly to publication without any intervening PTO action? How did applications fail? Dataset from 1884 to 2010, 6.2 million observations with 135 variables. For applications filed before 1981, dataset appears to include only successful applications.

About ¾ of use-based applications made it to registration, 1/3 ITU applications made it to registration. Overall, .68 for use-based applications for certification marks compared to .75 for TMs. Reversed numbers for ITUs: collective TMs/service marks did better than average, but absolute numbers are low.

Country of origin: other countries tend to do better with use-based and ITUs than US entities (maybe related to sophistication of foreign filers and expense of foreign filing). Individuals tend to fare worse (.65 v. .74 for use, .27 v. .37 for ITU). Clothing and telecom did relatively badly, with similar numbers. Overall publication rate for all applications, 1981-2007: .76. This is true both for use-based and ITU-based applications. Uniform level of PTO acceptance.

How many went directly to publication without intervention by the PTO? Use-based applications about 15%; higher for ITUs, 21%. Not sure what’s going on there.

How applications failed: data set doesn’t provide information on grounds for rejection, but does provide final status code. Abandoned: 90% through failure to respond or late response. 1.3% were abandoned after ex parte appeal. 5.2% express abandonment. Use based, 24.1% failed before publication; of the published ones, 2.3% failed before registration suggesting successful opposition but we don’t know why. ITU: 24.0% failed before publication; 51.5% of those published failed before registration, mostly for failure to file statement of use.

Why aren’t applicants filing more borderline applications? Common-law protections; if you fail at the PTO you are truly out of luck, whereas if you’re up against a bad-faith defendant a judge might be more sympathetic.

Q: can you distinguish between inherent and acquired distinctiveness?

A: something is wrong with the coding in the data; 2(f) is not as well coded as he’d like, but he can go back.

Hughes: direct publication—could it be foreign applicants, or families of TMs? If McDonald’s applies for another TM, the examiner might send it directly to publication. Effect of TM attorney?

A: Hasn’t looked at whether 44(e) goes directly to publication more often, but can check. Attorney data is a big mess, but would be useful to look at it.

Brauneis: there’s more to object to in a use-based application—the specimen, the description.

A: will eventually be able to filter out technical objections like that.

Jake Linford, Florida State University College of Law, Controlled Transformation and the Alchemy of Trademark Protection

Change over time: moving in and out of genericity; modernization of a mark that changes how it looks—then there’s a question of when you get priority. Splitting the market: Sunkist for fruit and Sunkist for vegetables agreed that they could coexist—when Sunkist bread comes along, the court thought that the agreement meant that the mark couldn’t be a problem for bread.

Language shifts: Lexeme-->shift in application-->polysemy (multiple different meanings, though perhaps related)-->homonymy (two terms that look and/or sound the same but don’t mean the same thing, no/know). Polysemy is a little slower in terms of comprehension than homonymy—it's easy for consumer to distinguish know and no in context. Natural move of words that are coined is to polysemy—the generic word for X is the market leader; it gets two meanings. In Haiti, klowoks is the name for bleach, etc. The conceptually strongest mark according to Abercrombie thus has the biggest risk of becoming generic.

By contrast, an arbitrary mark like Apple is likely to become homonymous. If this is right, fancifcul marks are more subject to genericide than arbitrary or suggestive marks. Descriptive marks are also more likely to go through a polysemous shift.

Trademark transformation as adverse possession? But we actually treat the mark owner as an adverse possessor of language—the owner reaches out to take ownership, but we require actual use, open and notorious use, hostile (require ownership claims), exclusivity, and continual possession—TM owner as semiotic squatter.

Hughes: One metric you didn’t mention is the length of the TM in relation to the length of the generic category. Xerox is one syllable shorter than photocopy. Aspirin became generic because no one wants to say acetylsalicylic acid. So make sure your mark isn’t too convenient compared to the generic category.

A: makes intuitive sense; would have to go to literature. (FDA regs also affect this for drugs, putting the long form in the public eye, even if the public eye refuses to read it.)

Irina Manta: Stephen Carter talked about adverse possession.

A: some people incl. Carter have talked about junior user as adverse possessor, briefly, but he thinks talking of the mark owner that way might be of use.

Heymann: C&D letters—TM owners are trying to push people off.

A: less comfortable with thinking of public as adverse possessors. Public uses are open and notorious, but a lot of public uses that tend to create genericness are not uses in commerce. Junior uses are not actually the source of much genericity.

William McGeveran, University of Minnesota Law School & Mark P. McKenna, University of Notre Dame Law School, Trademark Defenses That Aren't

Confusion test isn’t working. Often in TM cases other values besides consumer confusion are on the line. Primarily: competition and free speech. It should be clear that these are enough to allow uses of a mark despite the presence of otherwise actionable confusion: KP Permanent explicitly so holds. But we know how infrequently a court actually makes that ruling. Courts are more comfortable declaring that the facts don’t constitute confusion, by fiat, rather than saying “confusion notwithstanding.” Consider the KP Permanent remand in which the 9th Circuit rams confusion back into the analysis. Similarly with Grupo Gigante: court wants to adhere to territoriality principle, but can’t stomach the resulting potential confusion so develops its own famous mark exception.

TM lacks true independent defensive doctrines equivalent to necessity in tort. Perhaps functionality/territoriality qualify, but normally defensive doctrines turn into a different lens to examine confusion: genericism, nominative fair use, first sale, aesthetic functionality.

Why does that happen? Possible explanations: Courts have difficulty comparing two difficult-to-quantify abstractions. Property ownership mindset buttresses courts’ confusion focus. Plaintiff faces a greater disadvantage from losing, where defendants are stand-ins for general more diffuse values. Courts don’t like to innovate in common-law fashion as they used to, so they want a multifactor test. They worry about cancelling a mark, which they consider a radical step. Confusion runs through other doctrines with other rationales and so they become accustomed to them.

What might we do? Courts get less entangled in consumer perception when they look at cases in more general terms—rather than considering whether this particular d needed this particular use when they ask more categorically whether this type of dispute implicates particular value. Significant improvement in First Amendment cases over past decade by shift from “alternative avenues” question to Rogers v. Grimaldi. Don’t ask about need to use the mark, but about whether this is the sort of use that qualifies for special First Amendment solicitude.

Functionality: ask whether the feature is abstractly functional. Nominative fair use: Tabari, where Kozinski comes back and says that the third factor about sponsorship or affiliation needs to be downgraded—the first thing to do is to protect the defendant where use is nominative. Ask courts to look at the alternative value first before running to the confusion inquiry. As a bonus, this can reduce costs because these inquiries are likely to be cheaper.

Remedies: sometimes the defendant should just win. In reality, even the strongest trump doesn’t/shouldn’t result in triumph no matter what kind of confusion results. Core passing off injury v. penumbral post-sale, initial interest, etc. confusion. Remedies short of a complete ban on using the mark would be more appropriate for the latter. Blinded Veterans case: genericism is a defense, but there may be disclosure remedies for common law passing off. Tabari: the real importance of the third factor is the opportunity for the plaintiff to show core passing-off confusion, subject to possible injunctive remedies short of a full ban. Wallace Silversmith case on functionality: there’s a finding of functionality but a recognition that unfair competition could still have a role.

Grimmelman: courts have been unmoored from traditional confusion values (dilution, initial interest confusion, etc.)—can we do something about that?

McKenna: what’s happened is that courts have hid behind calling it confusion. If courts were better at bounding confusion, defenses would be less necessary.

Lemley: sounds like we need a TM use doctrine! Boundary-setting: these things don’t go into confusion analysis. Whether it’s psychological or burden of proof-oriented, it could lead people to rethink this. Is this unique to TM law or is it something about defenses? Fair use gets a decent amount of play in copyright infringement; trade secret law has reverse engineering—how do we think about these things when we’ve determined that you’re culpable in the first place.

McGeveran: yes, and necessity outside IP. (Ellerth/Faragher defense to sexual harassment cases comes to mind too.)

McKenna: structurally, the way to advance the agenda is to get courts to think about these issues separately.

Q: sounds like an attempt to generate more common-law defenses but the common-law cause of action has been statutorily appropriated. If you could get the court to think “tort” instead of “property,” maybe that would help.

McKenna: maybe we’re going to try to come up with new defenses. Right now we have defenses that don’t work; we’re not asking courts to create them out of thin air. Courts have not been afraid to innovate in expanding rights. Congress came in and layered on, but courts have innovated in only one direction.

Ramsey: empirical work by Neil Netanel: a judge’s consideration of transformative use made a finding of fair use more likely. Even raising the issue and forcing the court to think about it can help.

McKenna: courts recognize the values, but almost always find them swamped by confusion—even after Traffix, you find courts wanting to rely on competitive need. They usually conclude that there’s a confusion harm first and then face pressure to stop it.

Lisa P. Ramsey, University of San Diego School of Law, What the United States Can Learn From Australian Trade Mark Law

Taking speech into account in actual legislation. Australia has a TM use requirement and no dilution; categorical exclusion for comparative advertising; something like our descriptive fair use; and they have a threats action, allowing suit over a groundless threat of TM infringement.

Hesitate before we make all laws the same; use nations as laboratories: is dilution working in the US? Is there a big problem of blurring/tarnishment in Australia? Attorneys’ fees go to the winner in Australia; there is document discovery but not depositions/interrogatories; may be differences in litigation culture that affect lawsuits. But we can still try to find lessons.

Use as a TM means use of a sign as a badge of origin/designation of source. There are some cases to the contrary—one case found that repeated references to a mark might serve as a designation of source. Use as a mark would have some certainty at the extremes. Some contextual analysis may be required. “Chill out” on a T-shirt was an exhortation to relax, not a use as a mark. But some uses on a T-shirt might be use as a mark.

Some say that Australia does have a dilution provision (well-known marks) but she disagrees.

Q: is Australia moving towards a US system of stretching the rationales of confusion to get to results that favor plaintiffs? Recent case involving Paul Hogan.

A: Bilateral trade agreement with US: Australia agreed to move towards harmony, but isn’t required to add dilution as some other trade agreements provide. There is going to be influence, but should the default be to protect TM owners or the public interest? One argument against TM use is that it might protect future bad uses; but Ramsey would say that then the legislature can come in and cover the bad activities.