Mark McKenna, Dastar’s Next Stand
Recent cases involving some attempt to use TM writ broadly to control creative content. Claim by Teller who sued another magician for copyright infringement and false designation of origin for copying a magic tricks. Edgar Rice Burroughs trust sues John Carter of Mars for TM infringement. Most famously: Betty Boop on T-shirts.
Second set of cases, such as Louis Vuitton/Hangover Part II, more conventionally arising TMs but targeting use in a creative work itself. Also the Moore/Alabama case; ETW v. Jireh; etc.
Using TM to control the content of someone else’s creative work is the goal. Both, first more strongly, implicate the TM/copyright boundary; one would therefore hope that Dastar would provide clear guidance as Traffix did at the patent/TM boundary. Not so! Dastar has been used to dismiss false advertising cases (even though the Court left that open), but not much in the cases with the greatest possibility for conflict.
Lack of clear understanding of scope of decision, due to fairly unique facts and critical reaction to Dastar. These criticisms are wrong; courts should read Dastar broadly.
Decision is rooted in old style sense of the domain of TM. But courts have been eviscerating the domain except as defined by confusion writ broadly. Dastar is throwback: confusion has to be confusion of a particular kind, among particular people, for particular reasons. Because confusion has been so watered down, other courts resist Dastar. Dastar: There might be harm associated with misrepresentation of authorship, that’s not TM’s problem. Maybe you want moral rights, maybe there’s false advertising, but not TM.
The Court isn’t creating a special rule about authorship. To the contrary, it’s resisting the effort to create a special rule just because consumers care about authorship. If I go to the store and buy Coke and pour it into cans of Mark’s Cola, that’s easy reverse passing off. If I buy Coke and reverse engineer the formula and create a new soda that tastes just like Coke and call it Mark’s Cola, Coke has no claim even if consumers think it tastes just like Coke. That’s what the Court is saying in Dastar. TM protects statements “I made this product” and that’s it, not authorship/inventorship. In Dastar, Dastar did make those videos, just as Mark made his cola in the second hypo.
What does this tell us about the scope of the case? You can read it narrowly as reverse passing off and we could say it has nothing to tell us about cases in which the P does have secondary meaning. McKenna thinks that’s wrong and ignores the Court’s statutory construction. Any claim alleging confusion about source of content is out, whether forward or reverse.
But can you replead to allege confusion about goods in a really insignificant sense. What if Steamboat Willy were to fall into the public domain? Someone copies and sells DVDs. Disney alleges confusion about the source of the DVDs, existing only because of the content of the DVDs? This is the form of many of the cases in the intro: John Carter, Alabama v. Moore: confusion that can come only from the presence of the content but alleged to be about the source of the goods.
This can’t be right. Court was thinking of the intersection of TM and copyright. Virtually anyone could evade Dastar by repleading carefully. What should the scope be? Assume the movie is in the public domain; if we agree that Disney shouldn’t be able to extend the life of the copyright by simply alleging that the content causes confusion as to source. Possibilities: (1) bar any claim of confusion based on copyrighted works (strong form of channeling); (2) bar any claim of confusion arising out of content. (1) is too broad in a world with such broad subject matter of copyright. This is also a good reason to think hard about exclusions from copyright, channeling from copyright to TM. (2) is an extension of Dastar: rule out claims in which confusion arises from the content of an expressive work, whether alleged as confusion about tangible good or not, as long as it’s the content that’s allegedly causing the confusion. That would get rid of the John Carter case along with the Louis Vuitton and Moore cases.
Many of these claims could be dealt with through other doctrines: Rogers v. Grimaldi, nominative fair use. Those doctrines stink because you don’t know which one applies and it takes 8 years to litigate before you win. Dastar is cleaner, doesn’t require balancing, doesn’t require consideration of confusion. These claims don’t belong in TM.
Two difficult questions remain: if excluded claims are based on those arising from content, what counts as an expressive work? We do have experience with that with Rogers v. Grimaldi; can do through common-law development. Also requires some ability to determine whether alleged confusion derives from something in the work or outside the work, potential can of worms, but same thing. Second: what about services? Inclined to say the basic point holds: if confusion arises out of the content, it’s barred.
Justin Hughes: This sounds like an alternative to Dastar, not an application, because it wouldn’t have the Lanham Act/moral rights implications for rights of attribution. Wouldn’t necessarily interfere with a right of attribution based on the Lanham Act.
A: depends on what you mean by right of attribution. If it’s omission, then it will be barred. There are potentially valid false advertising claims for express misstatements. It’s the explicit statements that cause all the problems.
Von Houweling: Channeling: SCt seemed more enthusiastic about channeling than some parts of the law, with overlapping copyright/patent protection for computer programs—things get into the public domain and then stay there as one benefit. Others?
A: I like that; prospect of overlapping claims frustrates the public domain. Also has lots of ancillary costs in TM. Lots of trouble in TM can be explained by attempts to treat as TM problem any problem of confusion: false advertising, fraud, etc. People strategically use TM because it’s so easy to tell a TM confusion story. Should think hard about what belongs in TM and what belongs in other areas; consumer harm story doesn’t mean a problem belongs in TM.
Sandeen: courts are saying “we think this is wrong.”
McKenna: that’s Traffix: the Court is saying that competition is a basic principle and if it’s not patentable the default is that it’s copiable.
Ramsey: regular TM case—if I create a fake NYT, it’s my content and thus not a copyright problem. If I produce content and slap a fake source on it, different than if I include a LV handbag in my film. Explicit v. (arguably) implicit suggestion of source/sponsorship by inclusion in content of movie or title of movie. Different issues when you’re talking about impersonating a mark holder and claiming the mark holder is providing the whole speech.
A: the more we go down that road the less clean it is. Categorical approach is easier than figuring out what counts as “explicit” (since the LV handbag was “explicitly” in the film).
Ramsey: TM use/noncommercial speech.
A: (famously) not wild about that.
Jeremy N. Sheff, Selling Information
When there is secondary meaning in information, but that info also has value in itself. The Louboutin case, and most cases with status goods/social brands. Because of popularity/cachet/identifying ability, the brand is desired not because it says something about the product but because it is the product. Tend to be among the most contentious at the TM/other IP interface. Not new: any merchandising rights question/Two Pesos/Anti-Monopoly. Info that’s the subject of the property right also has value in itself as a subject of exchange.
Is any property right in valuable info anticompetitive? Or is the potential for confusion enough regardless of the competitive consequences? Because there’s often outright copying in these cases, courts tend to resolve these issues on grounds of protectability rather than liability: see Louboutin.
Room for nuance: McKenna’s suggestions; limiting remedies to proven damages/disclaimers. None of these doctrines require us to choose between pro-competitive and anti-competitive effects.
None of these doctrinal levels have the same deterrent effect against a potential plaintiff than a potential finding of invalidity will have. Relying on these, you might think, would increase threats of litigation by well-funded incumbents. A genuine concern. But if the doctrine makes clear that the worst outcome wouldn’t preclude the defendant from doing its business, the defendant might be more willing to litigate. (Anti-SLAPP? Rule 68 fee-shifting from an offer of judgment? His reaction: Interesting, but not clear how it works mechanically.)
Second objection: putting the distinction into practice (value in itself v. value as signal) requires us to look at subjective determinants of demand and that’s a tricky thing. Requires us to look at purchaser motivation.
Lunney: we all imagine a great past for TM, but there’s a structural flaw in the litigation process that led to this evolution. We can put a patch on it, but we’re still stuck with the process. This is why the SCt steps in every 30 years for Kellogg, Sears, Traffix/Wal-Mart. Rule 68: if defendant wins outright, doesn’t get its fees. So there’s an inherent weakness. (I think Rule 68 might actually work interestingly with a disclaimer if that’s all the P gets, though then we’d fight about whether the P got a better disclaimer than it was initially offered.) D only cares about individual result, but Ps care about doctrine: evolution away from correct balance.
McKenna: Coke tells you who made it but also affects your enjoyment. Maybe the word “primarily” can help. If you look to the binary, you’ll never find one or the other.
Copyright, Orphan Works and Fair Use
Jennifer Urban, How Fair Use Can Help Solve the Orphan Works Problem
Legislative focus has detracted from in-depth consideration of how library and archive preservation projects could benefit from fair use as a solution to orphan works. Is fair use a viable basis for digitization? What about for making the works available to patrons? More generally, can fair use do work where the copyrighted work is specifically an orphan?
The nature of the work: usually underused/doesn’t drive fair use decisions, but could play an important role along with the usual creative/noncreative and published/unpublished distinctions. Courts sometimes consider the availability of the work under this factor. S. Report for 1976 Act: if a work is out of print/unavailable through normal channels, user may have more justification for reproducing it. Maxtone-Graham, Williams & Wilkins, Hofheinz also use this consideration: all involved works unavailable on the market, though different result when the owner purposely withheld the work from the market. Interacts with the other factors: Nature of the work can also give guidance about whether allowing a use would damage incentives: grocery lists v. Hollywood films. Leval and Kasunic have written on this.
A true orphan represents a complete market failure. There’s simply an entity missing on one side of the transaction; no matter how much money you throw, no deal. Meets even Wendy Gordon’s most exacting test. NB: for this to be true, has to be unfindable through reasonable search.
Licensing system costs would be misallocated to highly inefficient ends: most times administrative costs would be not worth bearing, and fees would go to someone other than the owner. Building a licensing market that looks like a market would also have circular effects. Important to avoid building licensing systems where they are both inefficient and create a perception of a market—a ghost market—undermining fair use. Inhibit access without countervailing benefit, as per Justice Stevens.
Purpose of the use: libraries/archives are favored. Using the whole work: initial copies at least; other uses might involve only parts. Modern trend asks whether the taking is reasonable in light of the purpose, and given the digitization purpose whole copies are likely to be acceptable.
Wild card: Authors’ Guild argument in HathiTrust case arguing that §107 shouldn’t apply to libraries. Wrongheaded.
Grimmelmann: some purposes are unambiguously legal under §108 and §121. Given that, you can produce a digital scan to make a permitted use (e.g., for blind students on campus). Once the scan exists, does that justify other uses given that the scan is there and can be used costlessly?
A: I think so.
Lemley: copyright cases generally don’t have a fruit of the poisonous tree doctrine. Connectix: copying is ok if it’s intermediate even if it leads to a market-harming noninfringing product.
Market factor: Is there an actual market? Is there evidence that there could be a market? Is there a bad market? He thinks Urban’s argument is the last: there’s a market, but it’s not good/efficient.
A: between two and three. Nobody on the other side of the transaction. You could have a scheme with a probabalistic payout; maybe that’s just a bad market because inefficient. So few actual owners that there’s nobody on the other side.
Lemley: but what if someone wants to go into business finding them?
A: a misallocation given that we could instead have fair use.
Greg Lastowka, Nominative Fair Use Still Doesn't Make Sense
We started with consumer confusion as to source, and added to that—taken from the shrinking territory “not infringing”—confusion over sponsorship or affiliation; dilution; other confusions such as initial interest and post-sale. Then there’s branding/labeling use v. expressive use, mark used as an element of expression. Character carrying a Louis Vuitton handbag: not an indication of source but part of expressive work. That could be called a non-TM use.
Nominative fair use should have carved out, in this expanding zone, cases in which defendants should have been saved from liability. That’s not how it worked out. New Kids had a copyrightish understanding of what was going on in TM: cited a lot of copyright law; plaintiffs’ TM deserved protection against “copycats” and false claims of affiliation, but not newspaper articles, conversations, polls, and comparative advertising. True, but existing doctrines didn’t say that conversations and comparative ads were generally deemed infringing. District court deemed the claim barred by the First Amendment. Invented and invoked nominative fair use anyway, and now part of TM law generally.
Made no sense originally because defendant need to prove no implied sponsorship/endorsement, so it essentially had the D disprove the standard likelihood of confusion and disprove more (need to use mark, use only so much as necessary, no distinctive fonts)—a defense that gave D a heavier burden. An additional type of liability. Makes additional less sense as an exclusion in federal dilution law, because dilution D needs to show no confusion to satisfy the test: a confusion principle in a dilution test.
Tabari still makes no sense because the test replaces likely confusion and plaintiff has to disprove: D’s need, D’s use of the minimum needed, or D’s implied sponsorship or endorsement; so P still has an edge. Playpen case shows that the First Amendment defense is needed to back up nominative fair use.
Proposed fix: look backwards to confusion regarding source. Before New Kids there were ways to defend referential fair use: nonconfusing, descriptive fair use, aftermarket use, comparative advertising. Kozinski lumped it all together. Using fair use too broadly as Tushnet advocates would destabilize TM law, with too many Type 1 and Type 2 errors including replacement of traditional doctrines. Need sensible infringement standard instead.
Lemley: I get what you’re saying, but is that really how the 9th Circuit standard goes: third part: must do nothing that would, in conjunction with the mark, suggest sponsorship or endorsement. One way to read that is as you do. Another way is to say confusion caused by the use of the mark itself doesn’t count; only other things you do beyond that could disentitle you to the defense. Reading it that way gets you out of the confusion problem.
A: if the referential use causes confusion, I’d want that to be infringing. That’s the test I’d prefer because it’s the use causing the confusion, not the other things.
Lemley: but then why is it a defense? If it’s not a defense if it causes confusion, then it’s not a defense.
A: not a defense any more.
We had a back-and-forth: I think his beef is really with the multifactor likelihood of confusion test; a lot of times we shouldn’t look at a bunch of different contextual factors (we agree here). I see nominative fair use as a way of getting out of the multifactor test in situations where it’s expensive and troubling, and he sees nominative fair use as insufficient. I am willing to vote for the abandonment of the multifactor test!
Harry Surden, Fair Use, Transaction Costs, and Computable Contracts
Legal scope depends on transaction cost levels. Legal scope can change even when law is constant. Computable contract terms: expressed in forms computers can process. Data-oriented contracting: contract terms not expressed in natural language (option will expire on Jan. 18, 2015) but computable data (
). Done in finance so trading
systems can pick them up. State of the
art: computers can’t understand natural language at the level of a
sophisticated literate person. Semantic
contract terms: parties translate contract terms into computer rules. (Do not
allow contract to be executed if current date is after Jan. 18, 2015). Automated assessment: parties provide
computer with data relevant to performance or compliance with terms. Contract to pay $100 Austl. on Jan. 1, 2015;
provide computer with verifiable data about currency transfers between parties;
can make a prima facie assessment of conformance.
Transaction costs prevent professors from showing snippets in class, even setting the benefits aside.
Q: one danger is how we define transaction costs. Set of other transaction costs; externalities. Not all can be solved. Snippets in the classroom: social value of criticism/learning.
A: this only lowers some transaction costs. It will just impact the doctrine which is tied to bargaining and enforcement costs.
Sag: I see how this works with foreseeable uses, but what happens when no one contemplated the use at the time? E.g., library digitization.
Q: is this different from older discussions of transaction costs/DRM? There was a lot of talk about rights expression languages.
A: main difference is new tech makes it possible to be very specific about your contracting terms. DRM is a blunt version of this. Tech is making it possible to be more specific. (Examples? What I saw at DMCA hearings, where the movie folks were trying very hard to convince us that this would be possible, was about pure content consumptions; snippets weren’t a designed feature though they could occasionally be achieved by strategic pausing, but you still had to pay for the whole piece of content.)
Goldman: microtransactions literature too.
A: could be bad for fair use’s scope.