IPSC: Copyright and Trademark

Panel 17 – Copyright Substantial Similarity

Crossprogrammed with my panel; I came here first because I had more experience with the first paper in the other panel.

Clark D. Asay, An Empirical Study of Copyright Law’s Substantial Similarity Test

1005 cases found. Significant rise in substantial similarity litigation starting in 2006—tripled. Possibly related to internet adoption. Share of 9^th Circuit opinions also exploded in 2006. Courts don’t do really prong one beyond assessing access (not independently assessing copying). 50% of opinions use access as the sole means to resolve prong one, 80% rely on it alone or in combination with some other factor or subtest. Only about 25% of cases assess similarity under prong one, whether striking, substantial, probative, or otherwise. Experts only used under prong one in 3.8% of opinions. Under prong two, that’s 11.5%. Direct evidence, which court says is quite rare, is used in 17% of opinions. Rare, but not much rarer than assessing similarity under prong one. Ps win prong one around 77% of the time.

Prong two is a mess. Ordinary observer is most popular, 28% (47% in 2d Cir). Extrinsic/intrinsic in 24%, 64% w/in 9^th. Abstraction/filtration/comparison only 11%.  No test at all, nearly 24%; total concept and feel 21%, striking similarity, 18.5%.

A little over 65% of opinions rely on multiple subtests. No dominant subtest, even w/in circuits. Courts frequently rely on multiple conflicting subtests in the same opinion. Ordinary observer and reasonable observer language are used inconsistently.

Experts are used more frequently under prong two than under prong one.

Defendants win substantial similarity decisions slightly more than Ps.

Whether a court engages with copyright limitations in the decision appears to make a big difference about who wins the case. 60% invoke some limitation; P wins 2/3 of the time when court doesn’t mention limitations; D wins 2/3 of the time when court does do so.

Annemarie Bridy: reproduction right? Derivative works?

A: we coded for which right; substantial similarity applies to any right according to treatises.

Matt Sag: What looks like an increase over time may be an artifact of the way cases became available on WL post-2002 or so. Break out reported/unreported for all your analysis to deal w/that.

Bruce Boyden: 9^th Circuit screws this up. Is probative similarity prong one? Is it extrinsic similarity? Pre-Rentmeester, all you did under prong one was examine access and then jump to extrinsic/intrinsic.

A: we saw that sometimes. We coded extrinsic analysis as prong two when it focused on protectable expression. Earlier 9^th Circuit opinions did often use “extrinsic” as prong one; we tried to avoid interpretation and rely on what it said it was using; did code some as unclear. 

Panel 18 – Trademark Liability:

Jake Linford, Justin Sevier, Allyson Willis, Sex, Drugs, and Trademark Tarnishmyths

Sex/drugs inherently tarnishing? Courts take that as a truism. Prior attempts to measure tarnishment: Buccafusco et al 2016 focused on porn parodies, found burnishing except among very conservative consumers.

Bedi and Reibstein 2020 found marginal evidence of tarnishment in a single-impression, centrally processed experience (they’re paying attention); stronger tarnishment with multi impression, peripheral processing (they’re not paying attention).

Their studies: we might see tarnishment in the brand context, and stronger among conservatives.

Asked consumers to rate and pick among Reese’s and Snickers. Test subjects saw sexy M&M and Reefer’s Cup. Other tests: McDonald’s sexualization with M changed to woman’s spread legs; Titleist/Titties; Enjoy Cocaine; Sour Patch/marijuana .

Control to tarnishment—They rated the brand as stronger if they saw the tarnishing stimuli. Target choice—those who picked the tarnished brand increased significantly over control. Change in brand strength mediates purchase choice. Perception of brand strength leads to preference.

If you split conservatives and liberals, that’s marginally or not significant for conservatives and significant for liberals. For Reese’s and Titleist, burnishment for conservatives, for Coca-Cola and McDonald’s, there’s tarnishment among conservatives; liberals show burnishment among all but most for Titleist.

In the control group, men were more disposed to pick target brand than women; in treatment, women shifted dramatically enough to make them look like the source of statistical significance.

Second proposed experiment: Stoner Skittles and Satan Shoes. There is some evidence on social media that people were blaming Nike (then again, people on social media blame Biden for the Delta variant, so).

Previous experiment: exposed to 4 ads for CHICKS FILL A, people saw a drop in rating for tastiness and wholesomeness of Chik-Fil-A, though it recovered after 8 ads (if I am understanding the graph).

Tarnishment, they concluded, takes multiple impressions; and reading a news story removes focus from tarnishing stiumuli; consumers better control effect when their attention is drawn to it.

Do too many people already know the Satan Shoes were unsponsored, and if so, does that make the instrument a bad one?

Nicholson Price: Is there a theory justifying the use of multiple hypothesis testing?

Chris Sprigman: Did you cut by age, not just gender? With 57% women panel, it’s possible there’s an age skew in the panel in the way people look at ads/internalize messages.

Felix Wu: If some people are confused, then you’re mixing up people who are confused and people who experience what the law calls dilution.

Alexandra Roberts: Agreed, there are complicated first sale issues with Satan Shoes given that they are Nike shoes, but customized. So there are issues about what even constitutes confusion or dilution!

Luminita Olteanu, The ‘trade-mark-law-as-innovation-catalyst’ trap: why it would be wise to

conceptualise innovation outside the realms of branding and dilution protection

Broader PhD project: reconceptualized dilution test for EU. Current approach is unpredictable/not rigorous when there’s no evidence of actual harm. Relaxation of proof requirement of harm or unfair advantage conceals other normative goals. But, the other potential normative goals, including innovation incentivization, are unconvincing/unsound. Marketing literature may show appropriate methods to evaluate trademark reputation, or potential harms or unfair advantages.

Dev Gangjee has written about how this new justification is sneaking in—EC says “The mark works in this case as an engine of innovation: the necessity to keep it relevant promotes investments in R&D”; WIPO says branding helps firms recover investment in innovation, providing them a further incentive to innovate; diluting may reduce economic rents. It’s creeping into EU law, e.g., AG Opinion in Google France.

They’re using innovation in a very broad way. Not all innovations are good or desirable: innovative advertising may target consumers in vulnerable positions; innovative branding can be employed for discriminatory/racist ad targeting. Also innovation is ill defined.

Literature claims: strong TM protection is likely to lead to investment into strengthening the mark, not to innovation in product; and TMs are more valuable for incremental innovation than basic, so that might induce overinvestment in existing tech rather than new and untried tech. Empirical research showed reduced R&D spending following the FTDA.

The indicators used by the claimants don’t measure innovation; TM registration counts aren’t innovation. Since brands can be used anticompetitively or to promote the wrong type of innovation, skepticism is required in applying an innovation rationale.

Glynn Lunney: Many aspects that we now call dilution showed up as expanded likely confusion issues. That literature may be of use. Arnold Plant said the case for monopoly can’t be justified on the basis that the profits of monopoly will go to desireable things.

Sprigman: never understood the economics of this [me neither]. If you raise expected returns to innovation for incumbents, it’s possible that they will innovate more (and also that they won’t). But you’re also raising market entry barriers, so new entrants will innovate less. So how would you know? It’s related to the question of whether market power promotes innovation. That’s quite a bet with a lot of downsides. Confused as to how we got where we are [though the larger defense of monopolies from the Chicago School does provide some clues, I think].

McKenna: European economists did try to demonstrate empirically that TM registrations were associated with firms that they thought were innovative; they made those claims in causal terms, though they were probably showing that successful firms with new products often registered TMs. Did that literature peter out and now we are getting a new wave of less empirical, more theoretical literature, or is the current discussion derived from that?

Linford: do we think of certain types of marks as innovative? Fanciful marks are a kind of innovation, though maybe not what you mean by innovation. Exploring that could be fruitful.

A: one element of this is that patent is our IP system for innovation; TM excludes functional features, so the claims about “innovation” writ broad are untrue.

Wu: Good to make clear that descriptive claims are really cover for normative claims about value of branding. Innovation is being defined as including brand value, and once you define it that way then it looks like TM protects innovation. No need for causal claims!

RT: maybe design rights are causing the crossover here because they are also accustoming people to think both “everything is an innovation” and “everything is protectable.”

Robert W. Woods and Derek E. Bambauer, Is the Bloom Off the “Tea Rose”? Reevaluating the Tea Rose Doctrine for the 21st Century

Does the internet age change things by making it very easy to spread your reputation and business around the world? There’s a circuit split on whether good faith in a remote junior use means absence of knowledge or also incorporates an intent to trade upon the senior user’s mark. In theory, there’s no consumer confusion in a remote area, so why would that matter? [And why would knowledge mean bad faith, if you also knew they were remote?] Natural zone of expansion theory also is accepted in some courts and not in others. Some courts follow the common law and some courts rely on the federal statute—so there is a doctrinal soup, and also the world has changed around the doctrine, which is having trouble adapting. Maybe there are two different subsystems, one for registered and one for unregistered. Institutional competition among courts and legislatures; also sometimes the Lanham Act controls state law and sometimes it doesn’t. Theory and doctrine seem to fit least well together in this doctrinal area.

Questions: how does this doctrine interact with concurrent registrations? What are courts actually doing? Some troubling internet exceptionalism here. Notice to competitors and consumer protection might both indicate that Tea Rose is a bad idea. Registration providing nationwide rights is a fiction, but one upon which the Lanham Act is founded.

Most controversial possibility: propose that Congress eliminate unregistered marks at state/federal levels. State registration could allow pockets, reducing burdens on small businesses/startups, but limit part of 43(a) and state protections for unregistered marks. The thought is that there’s really no remoteness left.

We could also allow the likely confusion analysis to do all the work for us. We could just build in geographic considerations from Tea Rose and Dawn Donut into the LOC test. We could expand on 33(b) and limit exceptions to registered marks. But that might lead to some gaming of the system by different states.

Lunney: this would be the last thing on his list to fix in TM. The doctrine benefits small businesses, lots of which are purely local and don’t want to expand. Requiring them to get a TM registration is unrealistic. And consumers adjust.

A: Fair, and consumers do adjust, but that’s a burden that we often want to take away from them.

Rosenblatt: Strength of the mark is tied into this. There are hundreds of Broadway Pizzas—this is not an obsolete doctrine for many kinds of marks.

Jennifer Rothman: are you suggesting eliminating all TM protection for unregistered marks? That would be extreme and raise distributive justice issues. Why disfavor limited area markholders? State registries are also somewhat problematic given that they offer virtually no review—not clear what we gain by sending people to them.

A: agrees that state registry quality is important; they’re ministerial generally but can solve the notice problem. Recognize that it’s a steep hill to climb.

McKenna: the original Lanham Act was not supposed to cover unregistered marks; it was supposed to be common law protection, and courts just created that. So it wouldn’t be radical [as long as state common law wasn’t preempted].

Linford: consider Lady Antebellum/Lady A case—the earlier artist didn’t have a TM registration. If you do your system, that’s a shift from first to use to first to file.

Would your proposal also have implications for famous foreign marks?

A: yes.

Bita Amani: TMs in Transition, with Carys Craig—argues that the shift to first to file was not good for Canada. Slippery slope for removing the use criterion for purposes of protection in the registry. The US is the last bastion of use as the basis for protection, and that should stand.

Wu: Question: are you proposing preemption or merely that there’d be no federal protection? Tea Rose originally was not about federal law at all. [Federal law has to have some preemptive effect, I would think.]

A: strong and weak version of proposal. Paper defends strong version: unregistered marks cause problems. Lesser options may at least ameliorate the problems. 

Rebecca Tushnet, House Brands: The History of an Idea

Interested in this area of the law since I was a baby law professor getting followed in a Wal-Mart and kicked out of a Walgreens for taking pictures.

Partially a descriptive project: the case law, as well as the people producing empirical literature, aren’t as favorable to national retail stores as a trip to the CVS would seem to suggest. Very common for shampoo, tampons, cereal, soda—basically any core grocery/pharmacy product.

Empirical literature mixed/plaintiff favorable [my doubts about its solidity since the confusion stuff often makes assumptions about confusion or defines it in ways that many lawyers would not; interesting antitrust-ish claims with some literature saying it’s anticompetitive free riding on national brands by distributor chains which has resonance with current arguments about Amazon.

Cases mixed at best [Splenda: court reached split results, finding too-close similarity in some versions and enough difference in others.]

Practice nonetheless entrenched for major brands of household basics (courts are more likely to find actionable with third-party copiers and perfume) (contrast to the Amazon practice which applies to anything Amazon sees is selling)

(1)  Why? What combination of profit and incentive to litigate on both sides generates a practice that is far more favorable to copiers than the blackletter law might seem to indicate?

(2)  Is this what unfair competition would look like as a general rule for trade dress?

Dennis Crouch: In many situations traditionally some retailers communicate and push back against the mark holder: if you want shelf space, you have to allow us to sell this. Amazon might disrupt this tradition and spur more litigation.

Lunney: you’d like to have price or market share data as generic comes closer or further apart. [But very hard to get; hard to think that consumers make the distinctions about small variations in the packages the way the Splenda court assumed. As I think about it, the fact that all the different store brands—Giant and Food Lion &tc—lined up against each other in that case may well have affected the court’s decision to split the baby, even though consumers would never see them that way.]

Sprigman: those colors are quasi-functional—different sweeteners branded very powerfully w/colors, and likewise soda flavors. Is any harm transitory because consumers learn? What message do consumers learn? Is it that supermarkets and CVS use this, but not bodegas? Shaping competition in a way that favors major players. [Competition considerations go not only CVS v. bodega but CVS v. J&J]

Laura Heymann: On the shelf versus there’s no referent—in a dollar store, you’re likely not to see the major brand comparator on the shelf. Does that affect the analysis? Also how does the confusion get operationalize: physically grabbing the wrong product even though intellectually the consumer knows that the store and national brand are distinct.  [those things cut against each other: if it’s not paired on the shelf there’s no risk of pure accidental grabs; but maybe being on the same shelf increases the likelihood of such accidents based on peripheral cues]

Rothman: Intent also matters. House brand producers are viewed as good actors providing consumers with a meaningful choice. Knockoffs are seen as targeting a particular product and usurp its value. [here’s the weird thing: the case law isn’t that favorable. Maybe the law doesn’t actually shape perceptions. The Splenda case is an example: it’s a mixed result, though the intent is the same throughout.]

Chris Buccafusco: The world we have seems like the perfect response to fair v. unfair competition. There are two different sets of source identifiers being used: the word mark, which mostly is doing a really good job of minimizing confusion, and trade dress, which is doing a really good job of signalling a genre of products to consumers. So isn’t that the best of both worlds?

[But why only allow this for CVS and not for a third party copier? And there probably is a tradeoff of increased accidents for the not attentive shopper and increased benefits from the shopper who uses trade dress to make a simple comparison]

McKenna: Power of the brand name as opposed to retailer—British Brands Group is extremely upset about products in grocery store/pharmacy that look similar to national brands. They were not allowed, as a matter of law, to negotiate over retail placement or slotting fees, so the backstory is about retailers/competition policy.

The good cases say you can’t get so close that the package is confusing about identity, but we aren’t interested in sponsorship or affiliation confusion. Consumers may know that brands sometimes produce house brands, so what would sponsorship or affiliation confusion even look like?

RT: when the national brands do make house brands, the packages tend to look completely different. But consumers may not know or care about that, and this version of unfair competition may require a notional reasonable consumer, not an empirical average consumer, to draw its lines.