Usual disclaimers: my summaries are partial; I’m missing many great presentations (including ones I’ve seen in earlier forms elsewhere and patent stuff; also, this year I tended to pick copyright over trademark but was often forced to make tough choices).
Christopher Sprigman, Measuring the Incentive Effects of Innovation Thresholds in Intellectual Property (with Chris Buccafusco, Jeannie Fromer, grad students)
Thresholds in patent and copyright differ. Doesn’t agree that there’s only one way to build a water pump but a million ways to write about love, but take this as given: are the thresholds doing what they’re supposed to do? Hypotheses: patent’s higher threshold will encourage people to try harder (get a bonus for high performance); v. focusing on thresholds can be a distraction, which undermines performance (Dan Ariely; crowding out of intrinsic motivation). Offer a creativity task: Oregon Trail problem. People have to figure out whether to take bullets or hardtack, etc.
Three conditions: copyright condition, patent relative condition, patent absolute condition. Copyright: there’s a prize, and your chance of winning depends on how well you do in achieving a percentage of correct answers. As long as you provide any kind of solution at all, you’re in the game with a weighted lottery ticket. Small act is enough to get in the game, though you can improve your chances. Patent relative: only the top 10% get into the contest at all. High threshold for entry. Patent absolute: all who reach a certain score get in the contest. Participants were English speakers using Mechanical Turk.
Results: copyright participants scored reliably lower than the two patent conditions. Classic economic account of threshold seems to win over the behavioral economics account: high thresholds motivate behavior. Statistically powerful distinction. Copyright subjects overload the wagons more and fail the attention check questions (do you understand what we’re asking) more often: give less care overall to the task.
Broke patent conditions into high and low thresholds—from 10% to 20%, etc. Tweaking the numbers didn’t make much difference. All the patent conditions performed about the same, better than copyright.
Other subjects: recruited Techdirt readers. Tended to be interested in tech/computers. For them, copyright and patent relative subjects did about the same. Patent absolute participants did significantly better. Why? No strong thoughts. (Competition versus a standard compared to competition versus people?)
Want to look at other tasks, such as verbal creativity.
Q: is a Mechanical Turk participant someone who gets intrinsic rewards at all? They don’t seem to be there for that. Innately more mapped to pure economic rationality? Whereas Techdirt people might get more of a kick from doing well at a math problem in and of itself.
A: That’s why we wanted them. Doesn’t explain why patent absolute still motivated them more. Maybe relative caused them to focus on how good everyone else was, while absolute caused them to focus on themselves.
Rebecca Tushnet, Make Me Walk, Make Me Talk, Do Whatever You Please: Barbie and Exceptions
Barbie’s marketplace success, and her ability to represent multiple overlapping and even contradictory meanings, makes her a useful embodiment of some key exceptions to copyright and trademark law.
Mattel attempted to use both copyright and trademark to control the meaning of Barbie, reflecting a trend towards such overlapping claims: creative material becomes a brand, and brands have so many creative elements that copyright’s low originality standard recognizes copyrightable subject matter therein.
Likely confusion, the core of trademark infringement, now has a multifactor test in every circuit, though the test can be modified or substituted in certain circumstances, such as for use of a trademark as a search keyword to trigger competitive advertising (in some circuits) or in cases of what courts have come to call nominative fair use.
Notably, legislative as well as judicial complexity has increased over time. Just as new copyright provisions tend to be complex and reticulated, the relatively recent trademark dilution law and its subsequent revision added a number of multifactor tests and multistep exemptions, instead of creating a general prohibition on dilution similar to the prohibition on confusion; anti-cybersquatting law is similar in this regard. American judges have not been shy about adding further multifactor tests even to dilution, and the tendency of the law to become more complex over time has significant costs, especially for defendants who find their potential defenses turned into a confusing maze of overlapping but differently defined concepts.
Comparing trademark defenses to copyright fair use, copyright’s vagueness starts to look a little better. In copyright, courts have been less likely to add new elements defendants must satisfy in order to qualify for fair use, though at the same time each individual factor remains highly manipulable to accord with a judge’s view of the equities.
Among the lessons that Barbie has for us is that IP owners can see the virtues and vices of the defenses the same as IP users can. Thus, copyright owners like Mattel, and now Disney and even ordinary authors of novels, are increasingly turning to trademark to assert claims:
Disney, for example, asserted copyright and trademark claims against the creators of a John Carter comic. This reflects both the fact that much of the underlying material is in the public domain, and the fact that it is arguably easy to create a well-known mark very quickly in today’s market if one has either luck or a massive marketing campaign.
The traditional trademark businesses are thinking along the same lines. INTA ran a panel this year about how to assert copyright in one’s marks to get more rights than would otherwise be available by mere confusion or even dilution alone. Recent case, denied motion to dismiss copyright and TM claims about packaging. Most noticeable there is the ongoing Costco v. Omega dispute, where Omega put a tiny insignia on the back of its uncopyrightable watch and attempted to use it to prevent importation of grey market copies, a right allowed it under copyright but not under trademark. The district court found copyright misuse, but that won’t always be available, especially if the copyrighted trademark at issue is prominent on the product.
But why talk about this using Barbie? Well, part of it is that she can mean many things to many people. Amy Richards and Jennifer Baumgartner: “Barbie didn’t so much influence us as she was a blank screen on which to project what was happening in our heads.” And this is both why she’s so fruitful for unauthorized uses and why Mattel might have thought that it had claims worth pursuing: to attack Barbie is to project meaning onto her as well as to find meaning in her, and courts have regularly though not uniformly suggested that a proper fair use defense requires targeting something that’s already in the work.
Barbie’s multiplicity starts with her outfits, as Michael Cornelius says: she’s consumptive capitalism at its finest:
Barbie is perhaps the ultimate representation of this ability to market sameness in form.… Barbie herself is not her physique, not her plastic curves and molded-over vagina; rather, she is the hair and clothing that cover these aspects of her personage. Her names suggest as much, since Barbie alters identity every time she switches wardrobe: differing personas like Ballerina Barbie, Western Barbie, and Pan Am Stewardess Barbie are differentiated only by raiment, hairstyle, and accoutrement. Her identity is designed to be discerned by how her form is both covered and represented by these accessories.
But children’s play often strips her of those outfits and mutilates her body, as Suzanne Pitt did in her dungeon dolls.
Mattel sued Pitt for copyright infringement. The court found fair use in part because there is no line of “S&M” Barbie—notice how we’re back to the outfits again. The court didn’t mention the physical alteration of the Barbies.
What is the commentary on Barbie here? That she’s already sexual, because as Pitt argued she was initially based on a German sex doll? Or is it that Barbie is not sexual enough and needs to be given additional holes, against her will? One of the lessons of Barbie is about heterogeneity of meaning, both inside and outside fair use.
Yochai Benkler used her as an example of the multiplicity of meaning of cultural objects that, while it always existed, was exposed in new and more salient ways by the rise of the internet and, in particular, search engines.
- Mattel’s official site;
- the official collectors’ site;
- AdiosBarbie.com (a critical site);
- a Barbie-collectible magazine; a quiz, If You Were a Barbie, Which Messed Up Version Would You Be?;
- the Visible Barbie Project (dissected dolls);
- Barbie: The Image of Us All (an undergraduate paper on the cultural history of Barbie);
- a Barbie and Ken sex animation;
- a Barbie dressed as a suicide bomber;
- Barbies dressed and painted as countercultural images.
This multiplicity also has implications for fair use: Ordinarily, judges are asked to produce definitive answers about the meanings of texts, and there are both good reasons for them to do so and a number of tools that, though highly manipulable in many cases, are well-established ways of fixing meaning. For example: When a contract promised to deliver “chickens,” what did that mean? What did the parties mean by the words “the ship Peerless”? Courts’ job is generally assumed to be to produce answers about meaning.
Following this pattern, even a defendant-favorable fair use case tends to fix one meaning to the plaintiff’s work, another meaning or purpose to the defendant’s work, and then declare them different enough that the defendant’s use is transformative and therefore fair. When the defendant loses, the court tends to determine that the meaning of the works is the same, taking a universalist perspective that denies that different observers might generate different meanings from the same view.
I think this is a mistake. The district court in Salinger v. Colting, for example, found that the book 60 Years Later was a mere infringing sequel to The Catcher in the Rye, denying fair use because it considered that a sophisticated reader would have automatically understood the defendant’s alleged critique without any need for the defendant’s copying: “Holden Caulfield as delineated by Salinger was already often ‘miserable’ and ‘unconnected’ as well as frequently ‘absurd[ ]’ and ‘ridiculous,’ as Colting says of his elderly version of the character. . . . [T]hose effects were already thoroughly depicted and apparent in Salinger’s own narrative about Caulfield.”
The court here intervenes in a literary dispute. I’m inclined to agree that Holden was always pathetic, and certainly this response strikes me as well within the range of understandable critical reaction to The Catcher in the Rye. But the court was unable to see the ways in which other people might think differently. The disagreement over the reception of the original itself demonstrates that a version of Holden Caulfield focused on his pitiable and failed life is a critical intervention into an ongoing debate.
The history of parody and similar forms is one in which defenders of a targeted work always have available the responses, “that’s not really in there; you’re seeing things!” and “that’s obvious; your so-called parody is not subverting anything.” As for the former, Mattel submitted a consumer survey in the Walking Mountain case, which I’ll discuss in a moment, purporting to show that most consumers didn’t perceive a parody in the naked Barbies.
Converted into legal reasoning, it turns out that both these responses—your commentary is obvious and your commentary is obviously untrue—can underlie a finding that a new work was not transformative of an old one, and yet the interaction between those responses can demonstrate how the new work is transformative by making parts of the old more salient (and more uncomfortable for its fans). Without recognizing that works mean different things to different people, transformativeness as a concept is at war with itself.
So: Barbie is already sexual; Barbie is desexualized. Barbie is materialistic and shallow; Barbie encourages girls to run for president. Barbie sued MCA for trademark infringement and dilution over the song “Barbie Girl,” then after it lost the case licensed the song and altered the lyrics to be purely celebratory of Barbie: a great example of reappropriation and refashioning of meaning.
Barbie already covers all the bases. But people still get to make unauthorized uses of her! That’s the ultimate lesson of the Mattel cases, including Pitt and Walking Mountain where Barbie is again stripped and threatened with imminent dismemberment.
And what about trademark? So far I’ve been talking about copyright and the troubles courts have focusing on multiple meanings. The problems are different in trademark, also at issue in Walking Mountain. The result in that case: In order to hold that the use of Barbie’s image, not just her name, was protected by nominative fair use, the court expanded the nominative fair use test. Initially the test was framed as providing a defense for use of the minimum semantically required unit—imagined to be the name of the trademarked good or service. After Walking Mountain, the test is now whether the amount used was reasonable in light of the purpose: here an artistic portrayal of Barbie. This isn’t fully reflected in the court’s articulation of the test, of course, so matters are still unpredictable and confusing. And it’s not surprising to have complex and overlapping multifactor trademark defenses; the district court in Walking Mountain had carried out a completely different analysis in coming to the same conclusion.
But do defenses matter at all in a world where most people can’t afford lawyers? Recently, norm entrepreneurs like Peter Jaszi and Patricia Aufderheide have vigorously argued that most fair use determinations can be made by paying attention to context, norms of fair practice, and the statutory fair use factors (primarily the nature of the use). The protected categories approach reinforces that.
Questions tended to be: Talk to me more about the rule of reason. A: I like Pam Samuelson etc.’s projects of identifying categories of cases that are fair uses with different justifications and boundaries; this could be done with TM instead of adding a new multifactor test every time. Example: Recently, in the Rosetta Stone case, the Fourth Circuit suggested both that it would recognize 9th Circuit nominative fair use and that it was adding a good faith requirement to the Ninth Circuit formulation.
Q: Isn’t it encouraging that there are a bunch of tm defenses, when Salinger has only one shot?
A: Practically no, because courts rarely reach different results when a well-represented defendant raises the First Amendment, nominative fair use, and Rogers v. Grimaldi; it’s just a slog that makes cases more expensive. Whereas an ordinary commercial lawyer may pick the wrong one for the court/circuit and then we’re back in likely confusion land.
Bonus content: my necklace, by Margaux Lange:
Paul Ohm, Branding Privacy (nothing he says expresses the view of the FTC)
We should have a new form of remedy for certain types of privacy invasions.
Systematically: abrupt changes/privacy lurches. You went to bed thinking you had a certain level of privacy and you wake up differently private. Who can see what information by default? Visualization of Facebook, and each category represents a different type of info. FB sometimes flips the defaults on what other people can see.
Problem with notice and choice: inattentive users; too many disclosures that are too long; cognitive biases prevent understanding. The problem is worse during a lurch. Inconsistency: when you first log in that’s standard, but every company does the lurch differently. Doublespeak: NebuAd, went to cable companies and asked them to look at customers’ surfing. Letter to consumers: ads will better reflect the interests you express. Google did the same thing earlier this year breaking down the walls between its services: this will be better for you.
Instead of Google’s “this stuff matters,” we should use TM to indicate the changes. Not arguing that we should use theories of traditional TM doctrine that are about TMs representing nothing more than source. New TM: TMs symbolize a complex set of meanings, only some of which are controlled by the TM holder. Invert the focus: wield TMs/brands as weapons against TM holders to protect consumer interests.
TM should require selecting a type of privacy at the outset. Facebook is a low privacy company. If company decides it’s abandoning a core privacy commitment, it should be forced to rebrand—Google to Google Omni, giant conspicuous logo at the top of the page. Goes double for mobile, where screen real estate is so limited that small = gone. This is no longer the product you thought you bought.
Lemley: why privacy? TM allows people to integrate information into brand perception over time.
A: why not labor practices? Privacy is rapidly becoming core quality of concern and may be killing different sources of information. Regulatory history of notice and choice not working. But Doug Kysar’s paper suggests: maybe we could expand the concept. But happy to stay a privacy exceptionalist. Google Buzz, FB Beacon: the name was important in identifying the problem and focusing debate.
Q: why the TM and not at the level of the interface?
A: Doesn’t have to be either/or.
Q: so what if FB changes? FB changes the way we think about privacy. Isn’t this rigid in not allowing for change?
A: This is a regulatory call for action (not the FTC’s). I own up to wanting to slow/stop this change. There are significant harms that are being stored in the database of ruin. Privacy will cost some dynamic glory.