Epistemology of the diaper

Procter & Gamble Co. v. Kimberly-Clark Corp., -- F. Supp. 2d --, 2008 WL 3103758 (E.D. Wis.)

Kimberly-Clark makes Huggies Natural Fit diapers, with sides contoured towards the center. Huggies’ global brand director testified that consumers would view this feature positively, allowing Huggies to advertise a better fit and greater comfort. Huggies tested size 4 diapers, under the theory that size 4 was the largest seller and spanned a large number of baby weight ranges. Single-size tests are supposedly common in the diaper industry, and size 4-wearing babies generally engage in the standard baby activities of crawling, lying down, etc. The study allegedly supported the “natural fit” claim, though not a claim of less bulk.

Huggies used ads emphasizing natural fit across multiple media. A big part of the campaign involves claims that Huggies are designed for “babies of the human variety” while the on-screen image suggests that other diapers are designed to fit inanimate bricks. Huggies’ own study concluded that the ad was very persuasive.

P&G was less impressed and filed suit, arguing that the “fits more naturally” claim is false.

The court found that “natural fit” and “fits more naturally” were puffery, not factual claims. The court identified two kinds of puffery. The first, exaggeration, is a general superiority claim—no one expects that consumers will take claims for “lowest” prices or “best” products seriously. Here, the puffery is different. Instead of exaggerating product quality, the puffing claim is immeasurable. Huggies’ claims about its diapers are “so vague and subjective that they are neither provable nor disprovable.”

First, claims about natural fit are vague “by their very nature.” Neither of the parties were able to identify “a fixed set of criteria” for what “fit more naturally” might mean. There are better and worse fits in general, but the court asked: “which garment fits a woman more “naturally”-a tailored English business suit, a Japanese kimono, or an Indian sari?” Indeed, testimony suggested that consumers’ standards for measuring fit vary widely because fit is subjective, and even depends on the diapers’ appearance.

Despite extensive testing, P&G still doesn’t know how consumers decide that a given product “fits naturally.” (Would this matter if the issue were “tastes better”? Overall taste test claims are considered verifiable, even though individual tastes are subjective and difficult to explain.) In fact, P&G’s designers were once convinced, using quantifiable measures, that a particular diaper would be an excellent fit, but consumers disagreed and “won.”

Given that the benchmark concept is incapable of concrete definition, “fits more naturally” and “natural fit” could neither be proved nor disproved, and thus they could not be falsified. In fact, even if we could all agree on what it meant for a diaper to “fit more naturally,” our individual conclusions on which diaper did so “would be based on purely subjective judgments.” Such “opinion puffery” can’t be proved.

At the end of the day, “the only way of measuring natural fit is through consumer testing.” P&G’s expert noted that different consumer populations use different diaper brands and have different fit goals. So even if P&G internally considers Pampers Cruisers its best-fitting brand, consumers might not agree.

P&G argued, reasonably enough in my opinion, that because consumer testing can substantiate a diaper fit claim, such claims are verifiable through consumer tests. But the court concluded that even if consumer testing provides companies with useful market information, that doesn’t make it a legitimate way of assessing truth or falsity for Lanham Act purposes. The court was guided by the Pizza Hut decision, in which the Fifth Circuit rejected a falsity claim for “Better Ingredients. Better Pizza.” “Better Pizza” was typical puffery, and in context so was “Better Ingredients.” (This is something of a simplification: the court of appeals held that “Better Ingredients” might, in context of slightly more specific claims about quality sauce or dough, have been actionable, except that there was no survey evidence of materiality. Given the apparent evidence of materiality here, in the context of ads showing other diapers fitting bricks better than babies, I would think the analogy would support potential liability.)

The court held that, just as it’s impossible to prove which pizza is better, it’s impossible to demonstrate “conclusively” whether one diaper “fits more naturally.” “[A]n assertion whose truth depends solely on consumer opinion is an inactionable one.” Thus, taste tests showing that Pizza Hut was preferred by 75% of consumers over Papa John’s still wouldn’t have disproved the “Better Pizza” claim. (This is to say that the court is unwilling to believe that consumers perceive an implied representation of a factual basis for a claim as general as “Better.” Depending on the claim, however, consumers may well perceive an implied representation of a factual basis. Imagine, for example, if the diaper claim was “gentler”—which says something to me, as a frequent diaper consumer, about the effects on a baby’s skin—or, perhaps for another product, “safer.”)

Thus, “[w]henever subjective preference is the arbiter of a claim--whether of pizza or natural fitting diapers--resort to consumer studies is of limited value in the Lanham Act context and only underscores the inherent difficulty in disproving such claims.” In fact, the court held, the study results presented to it demonstrated that consumers couldn’t agree on which diapers fit more naturally. For several sizes, 51% of consumers found that Pampers fit more naturally, while 45% preferred Huggies and 4% had no preference. The court called this a “coin toss” based on individual preference, not objective truth. (Hmm. It seems to me that the objective truth is that you are just about as likely to prefer the fit of Huggies as the fit of Pampers, which is probably good news for competition. The court is shifting back and forth between individual preference—subjective—and population results—not subjective. It may get to the right result, but on this evidence it doesn’t seem right to accuse Pampers of being designed to fit bricks.)

Finally, the court pointed out that the studies weren’t just of subjective impressions. They were the subjective impressions of the consumer, not the actual wearer, who ought to be the judge of “fit.” The court was struck by the fact that the “aesthetics of the diaper” play a role in “mothers’” [sic] perception of natural fit. (I’m pretty sure that aesthetics play a significant role in everyone’s perception of fit, even with their own clothes.)

Even if the “fit more naturally” claim were more than puffery, the court concluded that P&G had not disproved it. First, the court pointed out that (in the absence of some signal in the ad itself that there is substantiation) the Lanham Act doesn’t require testing to support a claim’s truth. Thus, challenging the Huggies study does not in itself show falsity: when test evidence isn’t incorporated into an ad, a Lanham Act plaintiff can’t show falsity merely by attacking “any tests that might happen to exist.” It is the difference between “that claim is false” (actionable) and “you have no proof” (not). (Note that the result may be different in courts, like the Third Circuit, that hold that consumers reasonably expect a minimum level of substantiation from factual claims—that is, consumers don’t assume that advertisers pick their claims by throwing darts at a board marked with possibilities. Even such courts, however, are likely to attempt to cabin this substantiation doctrine to cases in which the claims are reasonably specific, so I can’t imagine this rule changing the result here.)

P&G argued, however, that its criticisms and its own tests could actually show falsity. The court disagreed, persuaded by Huggies’ expert that its test was generally reliable, despite minor flaws (the possibility of a “halo effect” leading to positive answers on multiple attributes given that the survey asked 26 questions; the limitation of the test to only size 4 diapers). Huggies’ expert had an interesting mirror-image criticism of P&G’s survey, which only asked about “fit.” He testified that the results indicated “dumping,” meaning that consumers answered the “fit” question even though a “fit attribute” wasn’t driving their opinion, because it was the only way to express a preference for one diaper over another. The court found this “persuasive in some limited sense,” but not enough to make the survey unreliable—or to skew the results.

In sum, the court found that consumer preference as to natural fit was a “close call.” Both sides had decent but not perfect surveys. The court concluded that the experts’ criticisms, in particular the focus on the “halo effect,” pointed in the direction of finding all the surveys unreliable: “[I]f consumers are so fickle that their own survey responses are swayed by such phenomena as dumping and halo effect, how can a Lanham Act falsity claim ever be premised on consumer preference surveys?” (Well, that statement’s never going to be applied out of the context of a vague product claim! I had thought it was well understood that surveys can readily be used to manipulate even apparently objective reactions.)

In the end, P&G’s claim failed because of an “epistemological problem”: how one can “know,” much less “prove,” whether certain diapers fit more naturally than others.

The court also found that there was no evidence of irreparable harm. P&G moved for preliminary relief 9 months after filing suit, and Huggies submitted credible testimony that ads like the “Brick Baby” ad have a limited lifespan. Moreover, P&G’s own brand manager testified that damages are calculable now. This is a good reminder that a large company needs to act fast on false advertising: the larger you are, the more closely you ought to be monitoring the market unless you wish to be charged with relief-defeating delay, and the more evidence you'll have establishing your own damages, making irreparable harm harder to show.

It’s important to be clear about the meaning of this result, and others: When a court determines that a claim is puffery, then even evidence of materiality—effects on consumer purchasing decisions—will be irrelevant. As others have, I find this somewhat paradoxical given that the core stated justification for the puffery doctrine is that consumers aren’t influenced by mere puffing. See Ivan L. Preston, Puffery And Other "Loophole" Claims: How The Law's "Don't Ask, Don't Tell" Policy Condones Fraudulent Falsity In Advertising, 18 J.L. & Com. 49 (1998).

However, the doctrine does fit into the American approach to advertising regulation: you can fool people into making decisions based on irrelevant criteria as long as you don’t actually deceive them. What’s the difference between being deceived and being fooled? It’s in the specifics of the claim. Not a very stable division. But the alternative, which is more European, would be to say that you can only influence consumer decisions based on (a) brand value or (b) objectively verifiable claims, and that’s not a particularly easy row to hoe either.

Colbert and branding

Stephen Colbert, himself something of a branding genius, interviews author Lucas Conley on branding mistakes—the intro made it sound like Conley was anti-branding, but the interview was more Conley admonishing marketers for self-dilution, trying to port brand attributes to incompatible extensions.

false marking, not false advertising

Forest Group, Inc. v. Bon Tool Co., 2008 WL 2962206 (S.D. Tex.)

Forest sued Bon Tool for patent infringement. Bon Tool counterclaimed for false marking under 35 U.S.C. § 292 and false advertising under the Lanham Act, along with invalidity. Bon Tool won summary judgment on noninfringement.

The patent at issue, No. 5,645,515, covers a stilt used in construction to allow people working on suspending ceiling structures to walk around more freely. In this and a separate patent case against another party, the courts ruled after Markman hearings that the patent was not infringed by products that lacked a resiliently lined yoke. However, Forest’s actual product also lacked a resiliently lined yoke until after the courts ruled on the patent. Beginning in February 2008, Forest’s product now has a resilient lining.

Marking an unpatented article with “patent” or a patent number for the purpose of deceiving the public violates the patent law. The intent requirement is that a party must act with sufficient knowledge that it is not telling the truth and that the recipient will be misled. Intent to deceive can be established by objective evidence; false marking plus proof of knowledge of falsity is enough to create an inference of deceptive intent. A plaintiff must show that the false marking defendant did not have a reasonable belief that the articles were properly marked.

§ 292 means that a patent holder can’t mark its product with a patent number when the product doesn’t conform to the patent claims. The court cautioned, however, that this obligation “contrasts and sometimes conflicts with” the obligation to mark a product consistently and continuously to provide constructive notice of the patent; notice is necessary to entitle a patentee to damages. So, if a patent holder stops marking its product, even only temporarily while a lawsuit is pending or after an interlocutory adverse Markman ruling, it will lose the ability to recover damages from infringers, even if it ultimately succeeds in establishing that the asserted patent covers the product. Thus, the “tension” between § 292 and the notice provision puts the patentee in a “precarious” position.

(The court cited a 1966 6^th Circuit case for the proposition that suspension of marking after an adverse interlocutory ruling destroys the ability to recover damages from infringers; it seems wrong and unlikely to be current law for precisely the reasons the court identifies. The solution is not to read the false marking statute incredibly narrowly, but to hold that stopping marking a product that a court has told you shouldn’t be marked doesn’t keep you from the benefits of constructive notice, at least if marking resumes when the adverse decision is reversed. Not to mention that actual notice will still be another route to getting damages, so the peril is not necessarily as great as all that. But then, patent law may have nuances I’m missing.)

Anyway, §292 needs strict construction, especially when the false marking allegedly occurred during pending litigation. A single district court adverse claim construction shouldn’t suffice to establish false marking.

In this case, however, the court found “unusual circumstances”; Forest was liable when it could no longer reasonably believe that the articles it was making were covered by the ‘515 patent. Forest’s principals genuinely believed that the stilts were covered by the patent, which was written by experienced patent counsel who possessed an exemplar of the stilt.

But after November 2007, Forest engaged in false marking. At that point, Forest had been sued for a finding of noninfringement by another company; it had received an opinion of counsel in 2005 advising that a stilt without a resiliently-lined yoke didn’t literally infringe the patent; it had two separate adverse Markman results and two summary judgment rulings; and it had retained experienced patent counsel who advised it to modify its stilts. At that point, Forest had no reasonable belief that its stilts were properly marked. Forest’s principal is not a native English speaker or experienced in fields that require interpreting patent language; it was not until November 2007 that he realized the serious implications of the 2005 legal opinion—that his own product wasn’t covered by the patent. Moreover, the decisions’ lack of finality counsels against an intent finding, but that factor was overcome by the existence of four rulings from two separate district courts on a straightforward patent. The court was also influenced by the fact that, at trial, none of Forest’s witnesses could explain how the lining of their stilt was “resilient” in any way.

The court didn’t credit Forest’s testimony that it directed its manufacturer to stop marking the old stilts with the patent number. There was no documentary evidence of such an order, and the court thought an international businessman should have immediately and carefully checked the new batches for compliance. The fact that, a week before trial, the stilts were still arriving marked with the number suggests that Forest did not try to stop such marking. In fact, its failure to notice the continued presence of the markings was strong circumstantial evidence of intent to deceive.

All this, for what? False marking is penalized at $500 per offense, one-half to the claimant and one-half to the US. Cases are divided on how to count “offenses.” The court rejected a per-stilt calculation. The issue is whether there was a single, continuous act or change in the time and circumstances of the false marking sufficient to separate the offenses. Separate orders placed by the patentee for falsely marked articles, for example, might show separate offenses. The court rejected a time-based approach (e.g., one offense each week) as unsupported by precedent or statutory language. Here, Forest made only one separate, distinct decision to mark its stilts after it knew they weren’t covered. $500 in damages.

Bon Tool also alleged that Forest’s advertising that the stilts were covered by the ‘515 patent constituted false advertising. In the patent context, the elements of a false advertising claim must be made out by clear and convincing evidence. Falsity was undisputed and, because the statement was false rather than misleading, the court assumed consumer deception.

What about my good friend materiality? Bon Tool had no evidence that the false claim of patent protection was likely to influence purchasing decisions by consumers. Likewise, Bon Tool couldn’t establish that it had suffered injury; even with falsity by a competitor, there’s no presumption of causation and harm in the absence of a comparative falsehood.

Both parties’ fee requests were denied.

IPSC 2008: Closing plenary

Copyright Through a Liberty Lens
Abstract | Paper
Jennifer Rothman
Loyola Law School

Proposes a fundamental shift in thinking about copyright uses, away from free speech as an outer limit on copyright. This dominant approach has been a virtual failure in facial and as-applied challenges. Why do we keep beating the First Amendment drum when it’s not working? (Comment: Fred Schauer has something to say about this.)

Lawrence v. Texas has a lot to say for a small subset of uses: identity-based uses unrelated to political or cultural dialogues.

Why did the 1A fail? Copyright as a free speech exception; copyright as already incorporating speech protections; copyright as the engine of free expression. Even if it worked, it would be highly constrained in defending identity-based uses. The vast majority of 1A scholarship on copyright has focused on the democratic civil society, valuing the political over the personal, ideas and facts over expression, and transformative over non-transformative uses.

What does she mean about identity? Our understanding of ourselves; important life-altering experiences; our beliefs and values. These all become intertwined with copyrighted works. Religious works: one’s entire belief/value system could be entwined with copyrighted works. Or a particular song could have been playing during a vital life event, as for example a woman who was raped while a particular song was playing. Or a photographer could have taken a picture of Lindsey Lohan kissing Samantha Ronson, and Ronson could put the picture up on her personal website. This is a heartland identity use because it documents something that happened to her and it doesn’t matter that she wasn’t the one who took the photo.

Why liberty is better at this: Substantive due process connects up better with our understanding of the freedom at issue than First Amendment concepts. This is more a feel argument than a doctrinal argument. Lawrence: “Liberty presumes an autonomy of self that includes freedom of thought, belief, expression, and certain intimate conduct.” Copyrighted works are important to self-definition. So we should focus less on freedom of speech and more on freedom of a person.

If we try users’ rights as speech rights, copyright owner will likely prevail—copyright is supposed to generate more speech overall. Or, if copyright is property, there’s no right to “make someone else’s speeches.” If a user’s interest is a liberty interest, liberty can trump property—the Civil Rights Acts weren’t viewed as takings (comment: note that this was a legislative enactment; the courts didn’t find a liberty right). If you call the copyright owner’s interest a speech one, speech has to be balanced with user’s liberty (and there’s no issue of compelled speech); likewise if you call the copyright owner’s interest.

Implications: no permission should be required; no payment should be required; entire works may be copied. The liberty based approach asks for the reason for a use, not whether copying the entire thing was necessary. (Comment: I don’t see how this is consistent with interest balancing. Can’t the user who was raped while a certain song was playing express that by playing thirty seconds of the song, in recognition of the copyright owner’s competing interests?)

Likewise, the private/public nature and the commercial/noncommercial divides would not be important. And it wouldn’t be necessary for the use to have a role in political dialogue or public debate. A lot of the work would be done through as-applied challenges.

Comments: This seems to me essentially a variant of fair use, with a constitutional candy coating. I’m happy with anything that gets us to think that there are a bunch of different possible fair use configurations, including self-expressive use. Amour-propre also compels me to note here that I’ve written on these issues, both standard First Amendment doctrine and an extended defense of nontransformative copying done in the service of personal commitments or meanings.

Q: A rule providing that no payment is required disturbs me where the loss is unambiguous. If you took someone else’s personal property without permission just because it was iconic for you (your foreclosed house, your repossessed car), we wouldn’t allow that. Imagine if I license my work for $10, but you refuse to pay. Why can’t I get my $10?

A: This is a concern about line-drawing. It would be a limited defense—the work has to be integrated into your life. The court would have to look at motivation.

My question: Where is “the right to communicate yourself to others” in liberty? See, e.g., Boy Scouts v. Dale, don’t ask don’t tell—you have no interesting right to communicate core aspects of your identity, at least not if that core is your gayness. As applied here: Why should you be able to play a song that’s important to you to other people online who don’t know you? If you should be able to play it, should you be required to take reasonable measures to prevent downloading?

A: One could argue that the 1A and free speech is really contained in liberty. This has an interaction with the autonomy-based, self-expressive features of the First Amendment; this is just a way to hook up liberty principles. W/r/t downloading, your identity interest doesn’t extend to others’ downloads, so you’d have to find a way to limit access. (I guess I have trouble seeing why your liberty interest extends just far enough to let them hear it while they’re visiting your page, but no farther. What interest do you have in their hearing it? How is it a liberty and not an expression interest?)

Wendy Gordon: Why do you think liberty will be any more successful? What cases are your models?

A: Lawrence is a fundamental broadening of our contemporary understanding of substantive due process, from privacy to autonomy.

Samuelson: This seems like fair use.

A: She is not trying to revitalize fair use. Her perception is that the 1A has failed as an outer limit, and she’s trying to find something else. She wants to remind courts that there are real people involved in using these works, and if that affected fair use that would be great.

Justin Hughes: A right of private performance would be the type of thing barred by your framework.

A: Even on a facial challenge. (Really? Because while Integral is a song vital to my self-concept, Boy with a Coin isn’t; if we really require investigation of motive, then a private performance right would generally be okay except where there was a specific reason to reject control over a particular person’s use of a particular work.)

Patent and Antitrust: Differing Shades of Meaning
Abstract | Paper
Robin Feldman
U.C. Hastings

The trend has been to harmonize the bodies of law by folding antitrust into patent, but it hasn’t worked well. The divide is deeper than courts/commentators realized—it’s not just that one encourages monopoly and the other restricts it, but that the fields use the same terms to mean different things, and that makes reconciling them very hard.

Two examples: The concept of exclusivity and the definition of product. Antitrust: exclusivity means one party can do something to the exclusion of all others. Using that language, people analyze patent rights as if there was a right to exclude everyone from the sphere defined by the patent, but that’s not always true. A patent doesn’t grant the right to make, use and sell the invention. She’s not talking about the availability of competing substitutes, but something more fundamental. Suppose a patent holder has a patent on use of a substance as an industrial cleaner, and the patent says “all uses.” If someone else discovers that the same substance treats breast cancer; she can patent that use. Anyone who wants to use that substance to treat breast cancer will have to negotiate with both of them, as with an improvement patent. That’s far less powerful than complete control.

The misperception that the patent holder is like a monolithic power in a particular space overestimates the power of the patent grant, encouraging courts to make improper assumptions about patent markets and be overly lax in defining patent markets.

Second example: what’s a product? The patent shouldn’t be looked at as a product in itself, because patent power often comes from interaction with other patents. E.g., defensive patenting. Biotech firms may engage in selective suppression, holding patents on multiple paths and only working on one at a time. This might be ok or not, but if you look only at the single patent you won’t be able to analyze it properly. Likewise, pharmaceuticals can have interactions that antitrust doesn’t quite notice, like the difference between a chewable and a nonchewable form of the same medication. Or selling two medications together could be tying, but putting them in the same pill has the same anticompetitive effect but is only one product. Or look at Roundup and Roundup Ready grain—is that one product or two?

Antitrust law’s conception of a monopolist has an explicit conception that some consumers enjoy the product or at least some substitute. Patent doesn’t assume that consumers get anything during the term, allowing perfect suppression. So treating the patentee like a monopolist underestimates the amount of competitive damage accepted under patent law.

Pragmatism, Knowledge, Copyright
Abstract
David McGowan
University of San Diego

This is in part a descriptive model of academic discourse about IP, and in part a normative model, with an acknowledgement of limitations. Arthur Leff’s classic Some Realism about Nominalism: we like to talk about things, but the critique of knowledge has been so powerful that we have few tools left. The economic analysis of law was a way to keep talking without saying anything. Conversation requires common vocabulary, and Posner’s law and econ has drained the appearance of conflict out of our definitions. We’ve stripped out the things motivating us and speak of costs and benefits.

Deirdre McCloskely pointed out that much of what passes for economic analysis is actually the product of introspection; we rarely have data, even crude data. So we introspect about what a rational actor would do, according to us, and then we call that consequentialist/utilitarian analysis. And we never measure or net it out; our graphs don’t even have numbers.

How people argue about IP: his introspection suggests that we retreat from extreme reflexive positions until we find safe ground within the range of plausible consequential arguments. Law profs are practical and like to talk about consequentialism, but all arguments within the space are structurally/logically sound, and choice between them is a matter of psychology. McGowan would like to shrink the area of “plausible consequential arguments” in order to make it less likely that we will be talking past one another.

Facts shrink the area. So our copyright policy should adduce facts, rather than suppress their production. We tend in copyright to opt for the suppression of facts. We substitute moralism for utilitarianism—Grokster. This calls into question the extent to which copyright is utilitarian.

Examples: (1) Parody/satire. Walking Mountain case: Mattel had a mall intercept survey about what people perceived when looking at photos of Cactus Barbie and the like. He doesn’t believe surveys are probative in the slightest, but he still thinks that the survey should have been considered in the copyright half of the case (hunh?). People were asked, along with sponsorship, about the message/meaning. 97% of people didn’t mention parody. OTOH, 25% of people surveyed in a mall intercept thought that Joan of Arc was Noah’s wife.

The 9^th Circuit refused to consider the survey on the copyright side. That doesn’t mean that there will be no parody/satire analysis. It means that the panel will be the one to do it. Given our socioeconomic status and our verbal agility in this room, we’re all going to want the hoi polloi to be kept out of it—they might miss the point and suppress speech. A utilitarian has to reject that, because if people miss the point, than the utility of the parody is not what we introspect it to be.

One of the things the free speech argument in copyright does is make facts irrelevant: that’s the point of a right. It doesn’t matter if porn leads to rape: you have a right.

Second example: Perfect 10 v. Visa: Perfect 10 argued that the ability of credit card payment increased the availability of porn/infringing porn. One could measure the incremental effect of access to credit on consumption of porn, but the 9^th Circuit said that inducement was not a matter of actually causing incremental infringement. Likewise, Sony is a fact-suppressing doctrine because it doesn’t examine how much infringement actually goes on; if there are potentially significant noninfringing uses the case is over.

His proposition: get facts. Inquiry is costly, and his proposal is costly. Limitations: Perfect 10 v. Amazon. The court says that proof of market harm is required. And yet there’s a standard Texaco circularity problem: how do you know there’s harm before knowing there’s a right? Demanding harm is unrealistic because you need to know the scope of the right first. That is a limit on his thesis—you can’t reduce all uncertainty with facts.

Second, is tailored pricing working? He likes tailored pricing, pay per use. But there are positive externalities from use and pricing is costly. The question whether the pricing mechanism is worth it drives a lot of copyright discourse, and it’s very difficult to get a handle on. We should focus less on questions of law—this is a contrast to Rothman’s paper—liberty, freedom, and speech are too diffuse and cut off inquiry.

Lemley: You’d add property to the list of those concepts, right?

A: Yes.

Lemley: The level of generality at which facts are valuable to us. How many people find this to be a parody? Or something else might be relevant—not this particular work, but the broader effects that having a category of parody has on innovation and creativity. How do you think about that?

A: As a function of a cost of inquiry/uncertainty. It’s rule/act utilitarianism.

Gordon: Yes, we all love facts, but we can’t jump to the position that there are no rights arguments or principled arguments worth making. Just because we’re a practical bunch doesn’t mean we’ll all follow you on this—we look at particular facts because they plug into particular values.

A: He doesn’t mean to deny framing based on priors, or the relevance of moral debate. But where is it going?

IPSC 2008: "Other"

Trade Secrets, Counterfeits, False Advertising

The Suprising Virtues of Treating Trade Secrets As IP Rights
Abstract | Paper
Mark Lemley
Stanford Law School

Nobody seems to agree why we have trade secret law. The dominant theory in the 19^th century: tort theory. But if acts are independently wrongful, why do we need trade secret law? The remaining justification, free riding, is ultimately empty because it doesn’t tell us what is ok and what isn’t. Standards of commercial morality: ultimately no more helpful in giving us a standard of commercial morality. And it adds the layering problem of judicial anthropology. Contract, too, can only explain part of trade secret law—why do we need a secret to have a contractual issue? And it doesn’t explain why we treat trade secret as a crime.

What about property? Well, let’s ask what the justification is for the creation of new property rights. In the end, trade secret has a standard IP justification: it provides an incentive to innovate. That extends beyond contract, so you don’t need an agreement; it extends beyond patent, so you don’t need the subject matter; and it’s easier/faster to get than patents. For startups/fast-moving industries, trade secret is better.

But trade secret seems strange as IP. Why would we want to encourage secrecy? His answer: trade secret law discourages secrecy.

Law can substitute for what would otherwise be investment in secrecy. Without trade secret law, you’d still try to protect valuable secrets, but you might do so in inefficient ways: this is an explanation of the classic DuPont case, where we don’t want DuPont to have to invest in building a roof over the facility while it’s under construction. Evidence from countries that lack strong trade secret protection suggest inefficient investments in secrecy: they’re less likely to enter into business relations with strangers, and have a tendency to hire as employees people who are trustworthy but less competent, notably relatives. There’s also evidence in the breakfast cereal industry that, fearing ineffective protection, companies won’t outsource. (Raises the question: do people know about trade secret law? Why isn’t it working in breakfast cereal?)

Trade secret lets you spend money only in the circumstance in which misappropriation actually happens.

Arrow’s information paradox: you need to have a legal right over which to bargain if you want to enter into a transaction. How does a person evaluate whether an idea is worth $1 million without knowing the idea? Disclosing the idea destroys its value unless there is a legal ability to constrain the behavior of the person given the idea. Contract could do a lot of that, but all. Trade secret does.

If trade secret law encourages disclosure, then we ask: why require secrecy at all? Why not just a legal right over information? The answer: secrecy serves a channeling function, creating a legal right only when it’s needed. The wheel (the secret is that it’s round) v. the formula for Coca-Cola. Trade secret law won’t change people’s behavior w/r/t the wheel. If I want to make money from it, I have to sell it, and I disclose it to the world. W/r/t Coca-Cola, in the absence of trade secret law, people will invest too much in secrecy. So trade secret law creates incentives only in the case of non-self-disclosing information.

Implications: (1) Secrecy is an important, not an accidental, requirement in trade secret law. Tort-law cases regularly dispense with proof of secrecy when there’s a bad act. That’s backwards. But if the info could have come from public sources, then there isn’t property—the question IP forces you to ask is “what right do you have?” (2) Reminds us that balance is everything. Incentives aren’t pure goods. Creation of rights comes at a cost. Trade secret law should preempt and substitute for other state laws that would otherwise create freestanding rights without similar limits, such as unfair competition. We don’t want to protect the wheels of the world. Treating trade secret as IP doesn’t define the scope of preemption, but it at least allows us to tap into a doctrine that has been thinking about these questions in the right way for decades.

Ann Bartow: Look at how easy it is to get injunctions in IP cases, as you’ve written, Mark Lemley (with Volokh)—enjoining average people who work in the factory will become easier; enforcement of noncompete clauses will be made easier.

Lemley: Relative to what? He thinks the answer is the opposite. In a world without trade secrets, parties will have even more stringent employment agreements. He sees trade secret as a substitute for noncompete clauses. It’s true that treating it as IP v. tort might incline a court to an injunction rather than damages. But eBay helps moderate that—the move is to make injunctions less automatic. Also the availability of criminal relief for trade secret theft is important.

Q: At heart, trade secret is about competition; calling it IP might lose that mooring.

Lemley: If he believed that the alternative to IP was pro-competition, he’d agree, but he doesn’t. The trade secret cases talking about competition often restrict it in the name of “good” behavior. The more we rely on the general statement of bad acts in tort, the easier it is to treat unfair competition as redundant.

Q: What are the other implications of your approach?

Lemley: Dubious of the “reasonable efforts” requirement—if we care about disclosures, those efforts are largely wasted, though they may serve as proof of secrecy. Also, other IP rights have expiration dates. Maybe at some point we say your incentive has been realized.

Q: What about security-led innovation?

Lemley: Certainly, without legal protection, people would invest in TPMs. Innovation in secrecy has collateral consequences for the info kept secret, which swamps the innovation of TPMs.

Trade Secret Prices and High-Tech Devices: How Medical Device Manufacturers are Seeking to Sustain Profits by Propertizing Prices
Abstract | Paper
Annemarie Bridy
University of Idaho College of Law

Two lawsuits involving a medical device manufacturer: implantable devices for use in cardiology. First case: Guidant sued Aspen, a healthcare consultant, claiming trade secret for its strategic pricing process, its sales contracts with hospitals, and the prices and terms in each individual sales contract. The contracts all provided that the hospitals wouldn’t disclose the prices. ECRI, a database, then sued after threats by Guidant over ECRI’s database of prices paid by hospitals for Guidant devices. Guidant counterclaimed, seeking protection for the pricing process and also for the actual price paid for every device. This is an ambitious claim. The endgame: maintain negotiating leverage with customers by preventing benchmarking and other forms of price comparison, chilling discussion among hospitals and among healthcare consultants and other third parties.

Such a claim didn’t used to be viable, under the common law as expressed in Restatement §757. But it could be now under the Uniform Trade Secrets Act. There are good policy reasons for excluding prices paid for medical devices, and other goods, from the scope of trade secret.

Early evolution: protection was given for processes and formulae, also tools. Early days: courts said there was no protection for private but ordinary business information that was not the product of special ingenuity. The Restatement required that trade secret be information that was more than ephemeral and was in continuous use in the business. It preserved the separate treatment of trade secrets and private business information.

1970s: UTSA is drafted with no reference to the earlier limits. UTSA’s definition is information “including” all these things. The drafters thought they weren’t opening it up too much—they wanted to protect information that was generally but not presently in use by the plaintiff—but the words of the statute don’t reflect that.

Under the early common law cases and the Restatement, the price paid is not the product of special ingenuity, is ephemeral, and is not for continuous use, though the formula for determining the price might be protected. What would be a question of law under the common law becomes factual under UTSA—what measures were taken to keep it secret? Does it have economic value?

Both cases settled before trial, though Guidant avoided summary judgment in each. Neither opinion discusses any limits on subject matter. This shift from a bounded to unbounded definition is generally a challenge to the public domain. And markets need info, including comparative price info, to function efficiently. The medical device market in particular is highly inefficient in ways that hurt buyers.

Courts can read and some have read the UTSA as a codification of common law, rather than an expansion. Most courts continue to rely on §757 when interpreting the UTSA, and it was intended to codify the common law except in very limited circumstances.

Pam Samuelson: This is a regulatory issue!

A: There is a bill on transparency in medical devices, but the drafting was so leery of takings issues in light of Monsanto that it’s incredibly narrow.

Counterfeits, Copying, Class and Confusion
Abstract
Ann Bartow
University of South Carolina School of Law

Grew up in a working-class family, and when she was growing up designer jeans were coming into their own. She was a nonconformist, but that would have been more fun if she’d had a choice, and that experience left her very dubious about the value of TMs. TMs are manipulative, and coerce people into buying more and paying more than they need. Posner & Landes say people reap value from paying more, and she doesn’t believe that’s a conscious choice. Posner & Landes say it’s ok to pay more for branded laundry detergent. Michelin makes 20,000 types of tires, but they’re all the same tire, just priced and branded differently. But if most people knew it was all from the same factory, then they wouldn’t spend the extra money.

Counterfeiting and copying: everyone knows that high-end designers are looking for paracopyright. But plan B is to use TM, by conflating counterfeiting with knockoffs. Knockoffs are not counterfeits, and are not a TM violation. She won’t defend counterfeiting, but how do we think about it as a scholarly community?

Pharma counterfeiting ought to be a crime. But why is it with clothes and sneakers? We ask about crime in copyright law. Ask how many sales are lost? Very few people are actually fooled. The buyers are less likely to buy the real thing.

Who goes to jail? The person sitting on the blanket/at the flea market. The people making money may occasionally pay the fine, but poor people are the ones who get rounded up and taken to jail.

Is buying counterfeit goods worse than downloading? TM owners want to offload enforcement costs on other people, including swap meet owners. Microsoft: it can be hard to detect software counterfeits, even for experts. One of the amazing things about the eBay/Tiffany’s case was that even Tiffany’s had trouble telling whether 22% of the merchandise was real or counterfeit.

Why is imposing liability on intermediaries important? Because it makes second-hand items very difficult to have at swap meets. It’s great for Tiffany to kill the secondary market, but it’s terrible for people who like Tiffany and don’t have much money. The eBay approach should be applied to physical swap meets.

Copying is endemic in clothing—partly from the creative process and partly because the market requires it. If consumers are asking for it, don’t we need to take that into account in our balancing of interests and assessment of welfare effects?

TM owners denigrate copying and consumers who buy copies. They are tacky scumbags trying to fool people—an instrumental muddying of counterfeit and knockoff. The dominant discourse deploys class signifiers—you need to buy the “authentic” thing. But protecting design rather than TM is not a legitimate use of authenticity. If you like a dress but don’t have $6000, it’s not morally wrong for you to buy a similar dress at a lower price. Ultimate agenda: the designers want to sell the same clothes at lower price points under “stealth” TMs; they want exclusive rights in the lower-end market.

Any new legal regime should be unambiguously linked to either copyrights or design patents. Illegitimate prestige is not a TM problem. Judges have contempt for shoppers whose purchases are driven by style and not TM: they seem to think that if you aren’t paying attn to TM you are defective as a consumer. That’s wrong.

McKenna: You don’t like trade dress, but you’re willing to protect the Polo pony logo?

Bartow: Yes.

Lemley: If you buy a shirt with the Polo logo or a Chanel handbag, are those serving TM functions? You seem to draw a sharp distinction between depicting TM and not that doesn’t necessarily map to what consumers do. E.g., Burberry plaid.

B: When the TM is the design feature, she agrees that law should protect it as TM. But shape, color, rhinestones, fabrics—they should be excluded. She thinks the line is maintainable.

Pasquale: Think about the worries about domain names, the pressure of which was reduced by the rise of search engines. If you can verify provenance in the label, it’s less important what the overall design is. (This is the standard rationale for providing less protection for trade dress when house marks can distinguish the products.)

Lisa Ramsey: Is knockoff the right term? Some knockoffs do violate TM. Counterfeiting is where the mark is identical.

B: It’s the closest term; she wouldn’t advocate limiting protection to identical marks.

Some discussion by various commenters about functionality as another way to say this—if you can’t protect it with copyright or with patent, you can’t bar copying in itself.

Running the Gamut from A to B: Federal Trademark and False Advertising Law
Abstract
Rebecca Tushnet
Georgetown Law

TM was born as false advertising and remains tightly connected. For preliminary relief, remedies, and survey evidence, courts treat the cases interchangeably. Elsewhere, there’s been more divergence, and that’s bad. Today I focus on materiality.

In false advertising law, for a competitor to win a case, a falsehood must be material: it must be likely to affect the purchasing decisions of a substantial number of ordinary consumers. Older cases sometimes spoke of this in terms of the elements being (1) falsity plus (2) harm, because harm requires causation and materiality shows causation. A five-factor test breaking materiality out specifically also developed and now dominates, because law tends to become more elaborate over time.

False advertising law has some basic theories about materiality. We can presume materiality for (1) health claims, (2) safety claims, and (3) central ad/product claims. But there’s lots of room for specifics: a tax preparer’s claim to offer instant “refunds” and not “loans” was material given evidence of consumer fear of “loans” and desire for “refunds”; a representation of EPA approval was material because consumers care about the environment. False advertising law also recognizes immateriality: misstating the number of real estate transactions at issue by 4%; statements about technical aspects of a product which was consumed by people who didn’t care about the technical aspects; statements that sports scores were updated “from the arena” when they were really updated from the radio; etc.

In false advertising cases, there’s been a regrettable trend to require overly specific evidence of materiality, starting with the Fifth Circuit’s Pizza Hut v. Papa John’s case, in which the court held that the false claim “Better Ingredients” wasn’t actionable because there was no survey evidence of materiality, mangling the rule that surveys or other response evidence is required when a claim is only implicitly, not explicitly, false. Because of this heightened attention to materiality, we see things like the Pom Wonderful court carefully going through the evidence that consumers actually cared that a product called “100% Pomegranate Juice” had 100% pomegranate juice.

Materiality remains, though, a matter of consumer preference. The public is entitled to get what it chooses, even if it is choosing by caprice, fashion or ignorance: FTC v. Colgate-Palmolive, 380 U.S. 374 (1965). The Supreme Court explicitly linked TM and false advertising, holding that infringement and false advertising are barred even if the product’s actual quality is high and consumers would be satisfied with it. The seller may reason that when the “annoying or irrational habit” (preferring a particular brand, preferring verification of a product claim) is broken consumers will be happy. But a misrepresentation cannot be used to “break the habit.” I focus on this language because it presumes materiality—an immaterial claim would not break a reasonable consumer’s habit because it wouldn’t change her behavior.

Now, though, we’ve abandoned materiality in TM; confusion is king, with no consideration of whether it matters. And it’s worse than it seems: abandoning materiality has feedback effects on infringement liability, because our multifactor test tells us that consumers are more readily confused when they aren’t paying much attention, and consumers don’t pay much attention to things they don’t care about, so confusion over an immaterial matter is automatically more likely, simply because courts have detached confusion from its moorings.

A couple of traces of materiality remain in TM at the margins, for example in discussions surrounding expressive works.

We should return to this conception of TM infringement, in which materiality is an integral part. We have a semi-natural experiment of what that would look like in the fallout from Dastar. Pre-Dastar, people sometimes brought false attribution claims as false advertising claims; they basically always lost. Bringing them as §43(a)(1)(A) claims proved much more successful, until Dastar, which explicitly said that §43(a)(1)(B) claims remained. But it turns out to be very hard to replead false attribution to be false advertising, and materiality is one reason.

How could materiality help TM? Certain people have devoted a lot of energy to finding a TM use requirement in the structure and history of TM law. But Mark McKenna responds that the question of whether a defendant is using a term “as a mark” can be collapsed into whether consumers are confused about source or sponsorship. Materiality allows us to cut at least some of the loops of that Gordian knot. And it has a substantial history and current presence in the Lanham Act.

Possible counterargument: materiality shouldn’t be required because false perceptions of connection can damage the value of the plaintiff’s brand. Mark McKenna is rebutting this empirically. Moreover, materiality can help us force TM owners to explain this story concretely rather than allowing it to stay as cliched doctrine. If we unpack the assumptions here, it becomes clear just how badly this brand-value rationale fits the modern sponsorship/permission cases. Even if people believe that Google needs GEICO’s permission to run ads in response to a search for GEICO, and even if those people become disillusioned with Google as a search engine, it is bizarre to suggest that they’d conclude there was something wrong with GEICO. It’s bizarre, frankly, to suggest that they’d even remember half-consciously thinking that GEICO gave permission to Google.

Presumptions can help us manage materiality in TM, like the health/safety presumptions in false advertising more generally. Thus, core “source” use can be presumed material, and we can identify source the same way we identify core general advertising claims: what’s the focus of the ad? What’s the biggest word on the product? Is there a special font drawing your attention to it? Etc.

Materiality won’t solve all our problems, nor should it. But it can answer some vexing questions and contribute to restoring some sense to the TM side of the Lanham Act.

IPSC 2008: Copyright/First Amendment and Trademarks/Reputation

Copyright and the First Amendment: Comrades, Combatants or Uneasy Allies?
Abstract
Joseph Bauer
Notre Dame Law School

Standard reactions to copyright/speech conflicts: minimizing the conflict by harmonization (same goals); internal mechanisms of copyright (idea/expression, fair use, limited duration, merger/scenes a faire, others). The growth in the scope and duration of copyright makes that a harder sell. There are various reasons why each of these doctrines fail. He is particularly concerned with the chilling effects of fuzzy boundaries and limits on copyright.

There is still a problem, despite claims by copyright maximalists. He thinks that true compelling interest analysis should be applied at least in some cases. Antitrust, like the First Amendment, offers ways of thinking about what alternatives a defendant had to particular infringing content. Would the alternatives have been effective in conveying the speaker’s viewpoint without copying expression?

This perhaps turns into a question of debating and defining “traditional contours” after Eldred.

Trademarks/Reputation

Marks of Rectitude: Fair Trade, Brand-based Regulation and New Global Governance
Abstract
Margaret Chon
Seattle University School of Law

(Came in a bit late, unfortunately.) There’s a signifcant fair trade premium for coffee sold in the US and EU. There are a lot of middlepeople between the coffee grower and the drinker—for every $3 spent on a latte, only $.02 gets to the farmer. (Coffee is the second-largest traded commodity in the world, after oil.)

Fairtrade Labelling Organizations International is a nonprofit, multi-stakeholder organization uniting 20 labelling initatives in 21 countries. They aspire to set fair trade standards worldwide. They set several generic standards for producers and traders. The standards cover farmers and laborers. FLO only works with associations or cooperatives, though, not individual farmers, which is a source of critique.

Chon is interested in competition among intermediary certifiers. There’s a competing coffee certification process focused on environmental issues and provenance, and another one focused on ensuring a higher fair trade premium for farmers. There are vast differences in the kinds of standards they care about, and that consumers might or might not know about—the use of GMOs; whether the certifier is independent of the industry; etc.

Dinwoodie: there are two different modes, passive and active construction of consumer understanding of marks. There is market competition for standards; what does the consumer understand from all these certification marks? There’s also a genericism issue attached to the term “Fair Trade”—we might have a feel-good association attached to the term, but we don’t know what it represents.

Certification marks are much less specific than traditional “source or origin” functions of TMs. They imply more objectivity: but we don’t know how consumers understand that. OTOH, certification marks are arguably much more specific than merchandise licensing, esp. in terms of quality control. The solidity of meaning is contextual.

Consumers may not be aware of specific standards, but they may have different awareness of the certification process or outcomes. That awareness may also be low, but it could be affected differently. US law, she thinks, insufficiently regulates the stringency/honesty of certification processes.

Similarities between GIs and fair trade: they guide consumers to different consumption paradigms, and they make a more specific connection to local culture and local values. They can also link with traditional knowledge. Do they form an alternative to the hegemonic trade system? Given the percentage of fair trade coffee traded, it’s not a huge impact.

Third party certifiers are regulatory entrepreneurs, doing work necessary in a fragmented global structure. Development hasn’t focused on TMs as a source of development. Certification marks can provide institutional standards promoting tech transfer, which may have occurred with forestry certifications.

Eric Goldman: Chicken-and-egg problem—certification marks are a supply-side solution, used to further normative goals. But we can’t ignore the demand side—there isn’t necessarily an organic demand for fair trade coffee. But then the certification agency has to market to consumers to convince them of the need to demand fair trade.

Chon: That goes to the passive/active construction of marks. Most of these agencies devote most of their resources to producers, and have limited marketing budgets. And they have an ideological dilemma: are they marketers, or are they helping farmers? “Organic” is a good example of where consumers have taken up a term without much marketing by USDA.

Q: The Idaho potato organization has done a lot of marketing to convince suppliers and consumers of the value of the certification mark.

Chon: Consumer education is a huge deal. Middlepeople can also help—Starbucks can sell its coffee as fair trade coffee.

Justin Hughes: A difference between GIs and fair trade marks is that there isn’t a clutter problem for GIs. We expect tens of thousands of GIs and it doesn’t feel like clutter. Will the market cause fair trade marks to shake out a few dominant ones?

Chon: Democratic, grassroots approaches aren’t efficient, and the groups are trying to respond to decentralized constituencies.

"Smithers, release the hounds": Adopting a new normative framework and analysis (safe harbor?) for dealing with copyright infringement by electronic agents
Abstract
Eran Kahana
Datacard Group

It’s clear that courts are much more likely to enforce browsewrap in a B2B situation than in a consumer-to-business situation, so his analysis is confined to B2B. The “hounds” in his title are bots that are released and return to their bot masters with the content “duck,” and then the masters post the duck on their own site.

He argues that human-centric analysis of bots is flawed. The bad cases include discussions of things like “awareness” on the part of a bot, or the location of a Terms of Use, or the color and font thereof. His argument: you should look at whether (1) a bot was used (2) in a wrong way, with “wrong” to be the appropriate subject of debate.

Internet Archive v. Shell: IA argued that Shell failed to state a claim for breach of contract because it only learned of the ToU after it copied the pages and no human was ever aware of the ToU—he thinks those are weak arguments. Field v. Google is better in how it treats the bot, but it is still human-centric in relying on fair use.

Remedy: dispose of notions of consent and fairness to the bot master. The solution: adopt technical specs for safe harbors for bot designers. Roommates.com provides a useful guide. If a bot designer adheres to the standards, then the designer doesn’t need to monitor each site visited. Reciprocally, drafters of ToU must follow drafting conventions. A compliant bot encountering noncompliant ToU would still be immune from suit for browsing, while a noncompliant bot would not be immune.

The drafting conventions must mirror the terms the bot recognizes, like “do not archive.” Future bots might recognize Boolean logic, synonyms, and antonyms. The more sophisticated the intelligent bot, the more flexibility there is in drafting. Drafters could allow some indexing and not others.

Pasquale: Niva Elkin-Koren argued “let the crawlers crawl.” Let the Crawlers Crawl: On Virtual Gatekeepers and the Right to Exclude Indexing, 26 Dayton Law Review 180 (2001). Smaller entities need to be indexed, and larger ones don’t be. A person trying to build a new search engine is stymied by Google’s prohibition on crawling its site.

Goldman: robots.txt is much of the solution that’s necessary. There are some automated robots, but when people configure scrapers, there’s almost always a human factor. Is it fair to say that they couldn’t review every user agreement?

Kahana: Bad actors will never use this standard—e.g., competitors who don’t want to configure scrapers to comply with ToU. Robots.txt is a good beginning, but it won’t help with your desire to ban competitors from scraping while allowing Google. It could be made much richer.

The Economics of Reputational Information
Abstract
Eric Goldman
Santa Clara University School of Law

How do we help consumers make better decisions? Goldman defines reputational info as info about an actor’s past behavior that will help predict the actor’s future performance. Examples: word of mouth, recommendation letters and references, job evaluations and student evaluations—these are unmediated, going directly from giver to receiver. Mediated: credit scores, investment ratings, GPAs, product reviews and ratings on Amazon, ratemyprofessor.com, possibly voting systems like Digg and PageRank.

Hypotheses: (1) Anomalies in reputational information supply and demand, whether over or under, hinder markets. (2) Inconsistent regulation of reputational information should be examined for unwarranted dichotomies. Credit scores are heavily regulated, but GPAs are not, nor are Google searches. There may be good reasons for difference, but examining the family of reputational information might give us good lessons for changes.

One specific topic: the undersupply of reputational information. People have first-hand reputational information that remains non-public—our individual opinions of movies, banks, etc. Every single person has some such info. Why? Costs in time, vendor retribution, norms against public criticism, privacy, legal risks. Thus, the market doesn’t work as well as it could.

His solutions are designed to be provocative: We could make more information about consumer decisions public information—automate disclosure of when Mark Lemley switches banks, publish customer lists. Or, we could increase channels for anonymous distribution of reputational information. (He earlier referred to the problem of pollution of the data stream; I expect that will come up as to this solution.) We could recalibrate the legal consequences of sharing reputational info—make it harder for plaintiffs to win; make it easier for evaluatees to fight back; we could protect intermediaries facilitating production of reputational information, as §230 does. Government funding of reputational information production: his least favorite solution.

Q: Undersupply is an instance of the general problem that a market in information is inherently imperfect.

Goldman: Absolutely, which puts gov’t involvement on the table.

Lemley: Wants to push on the premise of undersupply. Striking to offer that argument when we now have more info than we ever have. Maybe the market worked much worse in the past, but our personal behavioral decisions on the basis of nonpublic info may be a sufficient proxy for public info. If enough people leave the bank/don’t go see the movie, the aggregate effect of that sends a signal the market can internalize.

Goldman: Yes—there is feedback just from choice. His concern is that’s not enough. Lemley’s Facebook page gives much better info about movies than Lemley’s movie attendance.

Me: This is getting into the tension between analyzing behavior in terms of incentive and analyzing it in terms of taste that is also hampering copyright analysis. Because producing such info has costs, we sometimes ask what the barriers are—if we want more, we try to lower the costs or increase the benefits. But some people just don’t want to talk: should we try to change their tastes?

Goldman: And that structures the question—do we have enough people providing the information, or not enough?

Elkin-Koren: These proposed solutions will affect the other problems you haven’t discussed—corruption and manipulation of information. Every solution you offer may increase that problem.

Goldman: Absolutely. Fixing undersupply is linked to avoiding/correcting oversupply.

IPSC 2008: IP and Government, with YouTube detour

The Political Economy of Intellectual Property
Abstract
Clarisa Long
Columbia Law School

She collected all the amicus briefs in the Federal Circuit since 1982 (only 385 of them) looking for correlations between support and outcome. She’s done the same thing with proposed bills in Congress on copyright and patent since 1952 (about 6000) and will be running regressions.

Copyright law’s rate of evolution has been much faster than patent’s. To some degree this can be seen in statutory length—copyright is 2x as long and tends to be much more industry-specific. One conclusion: one of patent law’s successes is its uniform nature. The relative paucity of industry-specific provisions has slowed down change and kept interest groups from getting a toehold into the law. Copyright doesn’t have a uniformity norm.

Copyright changes that get passed: have uniform industry support, are industry specific or are about a small number of organized industries, and are in copyright-only bills. Patent changes that get passed: it helps to be related to tax. Death knell: patent bills that stand on their own and are only about patent law; patent bills that don’t have unified industry support.

Who supports whom before the courts? Bottom line: nobody likes big pharma; nobody cares about big manufacturing whether to support or oppose it; one of the most powerful interest groups is “big finance,” a group that flies further under the radar than she would have expected. She also wants to look at the effects of law professor support (laughter from the audience).

In questions: Long discussed the hydraulics of regulation—if you can’t get a result from Congress, maybe from a regulator; if not patent, maybe the FDA; etc. This is very hard to take into account.

Pam Samuelson: Big pharma’s biggest success in the past decade has been TRIPs, and other work with the USTR, so looking for success there might also be of interest.

Long: one of the ways she measured success was the number of bills proposed, who they benefited, and whether they got passed. Big pharma’s success rates there are quite low, but the key may be output. (Samuelson pointed out that stopping unfavorable bills may also be success.)

Other comment: private term extensions for patents are also an indicator of success.

Long: These were very popular until the mid-70s, and they almost always failed; since then, the number has dropped off sharply. The success rate has gone up a little though.

Q: What about effects on international treaties, creating pressure on domestic governments to change law?

Long: Yes, she should also take that into account.

Comment: It’s really easy to introduce a bill. The story of a piece of legislation is qualitative, not quantitative. Interview people about particular bills. So much happens behind closed doors, and involves exchanges that go beyond individual bills.

Mediating Public Choice Problems in the DMCA Safe Harbors
Jason Schultz
U.C. Berkeley School of Law

Public choice theory says that producers have more influence than consumers, and private interests have more influence than public ones.

Given the insights of public choice/interest group theory, should we treat statutory interpretation like contract interpretation? That is, unrepresented parties should be favored in term construction, whereas interested parties who participated in drafting should have terms construed against them. If necessary, they can go back and get an amendment, whereas unorganized/unrepresented parties are unable to do so.

Special note with tech law: there may be nascent public benefits, emerging and undefined, which means that existing users are advantaged over future innovators, users, and consumers. The amount of user-generated content in 1998 when the DMCA was passed was very different than it is today; the economy of UGC has changed and the level of investment has risen.

Who was unrepresented with §512 of the DMCA? Large copyright owners served as proxy for small copyright owner interest, while telecom carriers and ISPs served as proxies for small users. Question: how well did they do? The vast majority of §512 deals with service provider and copyright owner interests. User rights are only represented in counternotificaiton provisions and the cause of action for takedown notice abuse.

Look at the requirements of notice v. counternotice. Notice requires substantial compliance; counternotice requires strict compliance. Notice requires reasonably sufficient contact information; counternotice requires full name, address, and phone number. Notice: no consent to jurisdiction or service; counternotice: mandatory consent to jurisdiction and service. Notice: takedown mandatory and must persist 10-14 days even if user files declaratory judgement; counternotice: putback not mandated. Notice: filing suit voids putback; counternotice: filing suit has no effect on putback.

Injunctions: injunctive relief only requires notice to the ISP, not the user.

§512(f), allowing suit against knowing misrepresentations of infringement in a DMCA takedown notice. A number of cases: Diebold takes down embarrassing emails; Ric Silver takes down videos of people incorrectly dancing the electric slide; Viacom takes down Colbert Parody; Uri Geller takes down a critique of him; etc.

Two arguably ambiguous terms in the statute: (1) “knowingly” materially misrepresent and (2) “infringing.” In Diebold, the N.D. Cal. applied a pro-user stance, finding “knowing” when no reasonable copyright owner could have believed in infringement. Rossi v. MPAA: pro-copyright holder, applying a subjective good faith standard.

Currently: Lenz v. Universal: Prince argues that even an unreasonable belief that 29 seconds of a dancing baby is infringing is enough to defeat a §512(f) suit. Public choice says no: we should interpret the ambiguous provision in a pro-user way, because Prince was well-represented in the drafting process and the mom who filmed the music video and put it up on YouTube was not. Universal is also arguing that “infringement” means a prima facie violation of §106, regardless of whether there are any applicable defenses or exclusions. Schultz argues that this is wrong—copyright owners could have drafted a version that spoke only of violations of §106 rights; users should get the benefit of the doubt, and copyright owners should have a duty to consider defenses like fair use before sending a takedown.

Susan Crawford: Does this create perverse incentives to stay out of drafting? How long does a sympathetic reading for nonparticipants last—100 years? And sometimes courts are more sympathetic to expansive readings because they don’t know anything about drafting.

Schultz: Staying out is a high-risk strategy. This theory only applies to ambiguous provisions, and so if the drafter is working alone she can just write the most unambiguously favorable provision she can think of. Courts like to read the statute as if the language were plain; but we all know that something else is going on.

Q: Where do legislators fit in?

Schultz: There’s a big debate in the literature.

Q: Incentive to draft as carefully and specifically as possible, which is the point, but the result would be to limit judicial discretion—are we worried about that?

Schultz: Yes, we’ve seen a patent/copyright difference with courts’ ability to jigger patent doctrine as appropriate.

Sag: Sometimes Congress is deliberately vague.

Schultz: Yes, that shows that the interests were equally opposed, and that’s useful info for the judge to know.

Sag: Where does constitutional theory come in?

Schultz: The majority of the public choice literature focuses on constitutional rights, arguing that courts should be more sensitive to interest group manipulation.

YouTube and Sharing: Competing Business and Cultural Models in the Digital Era
Abstract | Paper
Olufunmilayo Arewa
Northwestern University School of Law

There are a wide range of behaviors: copying verbatim, reuses and recreations, transformative uses—all are enabled by the digital era. We often assume that cultural practices are homogenous, but they’re not.

Sharing is not new. Lessig: one way to see Disney is as a major remixer of existing cultural works. And the Brothers Grimm from which Disney borrowed were also remixes. Still, the high culture model of the isolated genius in the attic retains a lot of influence in our discourse. There are extensive interactions between high, medium, and low culture. High borrowed from folklore; low borrowed from high and popular; low is redefined over time to become high—classical music is canonized, though at the time it was written it was popular music. Many classical composers were persistent borrowers.

19^th century: folklorists ran around collecting peasant folklore, because they assumed it was going to die out; they were also worried that people who came to the city stopped doing peasant dances and started hanging out in bars. But folklore is persistent and can give us an alternate model of how people create, by sharing. Riddles, jokes, proverbs, urban legends. There are multiple variants, shared cultural elements. People are using and partaking of them in everyday life—a model of Web 2.0 creation.

What do we care about—Culture or culture? Culture is perfected, while culture is evolving—Matthew Arnold and Andrew Keen like Culture. She’s not sure Keen is entirely wrong about the narcissism of Web 2.0, but it’s useful to think about what people who aren’t “authorized” to create culture actually create. The peasants who came to new urban enviroments behaved differently—not like they were supposed to, going to dance halls etc.

Cultural studies treats copying as something very different than copyright law. In musicology, there are dozens of words to talk about copying, from borrowing to homage, and we in law don’t have similarly rich tools. Folklore could be a good model for thinking about culture.

From a business perspective, the competing visions of culture are (1) vibrant, where existing works are building blocks and (2) valuable asset, where copyright enables the protection of revenue streams from existing works and strains towards a pay-per-use model. They aren’t mutually exclusive, but they do tug in different directions. She sees parallels with the canonization of classical music in the 19^th century. The canon was untouchable, had to be performed as-is, and classical music died as a vibrant, living cultural form.

Q: What about the economic motivations of creators? Bach stole from Vivaldi because they were both being paid by the church.

Arewa: True, there have been significant structural changes in entertainment industries. Before the 19^th century, it was hard to print sheet music. But the uses were very different: ordinary people routinely sang at a level we don’t reach now. It’s not economic returns that she is focused on, but culture and the appropriate scope of copyright rights.

Q: How do you deal with the tension between the two cultural models? And how do we put interdisciplinary research like this into the classroom?

Arewa: She teaches music and copyright. Students aren’t aware of precedents—music out of context, even 10 or 15 years, loses the context of creation. With copyright getting so long, we have no idea what the original work sounded like. So something that sounds original to us now may have sounded completely unoriginal to its initial audience. Students think that great composers were total originals, and they’re just wrong.

As for dealing with the tension, we reconceive copyright as being about culture as well as assets. Compensating creators is good (but standing recording contracts don’t do it that well), but we also have to think about enabling reuse.

Privacy, Free Speech and "Blurry-Edged" Social Networks
Abstract
Lauren Gelman
Stanford Law School

Focus of her research: the conflict between privacy and free speech. And the effects of disintermediation on privacy; work in copyright may have insights for privacy.

The basic question: why do people post content on a medium available to the whole world that is not intended for the whole world. (See also: context collapse.) Clay Shirky responded to critiques that most blog content isn’t interesting, because it’s just trivia about people’s lives: It isn’t for you. It is interesting to people it’s for.

People have undefined networks. Users are calculating that they can’t identify in advance all the people they are intending to reach with their posts because their social network is undefined. So they make their baby pictures available to the world on Flickr, and don’t need to expend energy figuring out who to send the pictures. Old-world analogy: do you want to make your phone number public so that anyone who wants to can find you, or will you keep it private and only tell people you know will want to contact you? This cost-benefit calculation is dependent on how well we assume we’ll be able to identify good contacts in advance. (Comment: also on costs.)

What if you have been diagnosed with cancer? Your social network includes people who know you, but it may also include people with the same diagnosis. Before the internet, you might have had to seek out a hospital with a support group; now, you can start posting your story, and find people who are in a different physical space. This creates enormous value and new communities of interest. By posting publicly, the cancer patient takes advantages of the blurry edges of her social network.

You have to disclose in order to get this value. This is the cost of Facebook: you say what high school you went to, and get access to that network. Right now, our privacy regime finds that once you disclose something you have no more privacy in it. What she’s concerned about is that if we continue, we are forcing people back into a world of not disclosing and we won’t get all the speech benefits of being able to share information without fear that this information will be completely uncontrolled and used against you in any way possible.

Why should we care? First, our stories are rarely about ourselves. Even if you say people take their own risks, they’re often taking risks with their friends, parents, etc. Second, users make mistaken assumptions about access, storage, and search, though it might be changing over time. Third, we will get less speech if people stop using the medium for this purpose. The speech argument against privacy has to be weighed against the speech-encouraging benefit of protecting privacy.

Legal history: why do we think that disclosure of any kind destroys privacy? There’s been a move from intermediated institutional organizations to non-intermediated individual publication. With newspapers, if you wanted to tell your story, you needed to go through an intermediary that made determinations about newsworthiness. This created a binary notion of information: either it should be disclosed or it shouldn’t be.

Technically, we started with message boards, which were targeted at limited groups and felt more private. WWW was public to the world, but info there was much different than the info on message boards—more commercial, more government, etc. Web 2.0 combines both of the above.

Copyright analogy: my ability to protect my privacy may affect your ability to tell your life story. She thinks that the greater policing burden on copyright owners is a good thing overall, but as a privacy theorist she is worried about the social costs of eliminating intermediaries.

Q: Suggests further specification of harms. What are the problems of having people reveal information partially? Chillingeffects gets a bunch of DMCA notices that are ostensibly about protecting copyright but are really about people trying to hide youthful indiscretions. What are the implications of continuing down our current path?

Sprigman: What’s the relevance of Facebook’s ToS? And new tools to control partial access?

Gelman: Maybe Facebook can solve much of this problem. She’s nervous about proscriptive analysis that endorses how Facebook does it and ignoring the rest of the internet. And she’s not totally happy with Facebook—it’s relatively hard to protect information; you have to make your Wall available to allow other people to write on it. Market solutions have to be a part of the solution, but we’re not there yet.