Thursday, August 07, 2008

IPSC 2008

Stanford Intellectual Property Scholars Conference 2008

Usual caveats: these are my notes on works in progress; my attendance at panels is shaped by very idiosyncratic interests and I will miss a lot of great things by necessity; I don’t do patent law so don’t expect much from me there.

In the welcome session, Mark Lemley presented on the forthcoming Stanford IP Litigation Clearinghouse, which collects data about IP litigation across the country and allows you to play with the data in interesting ways, including looking at geographic and time trends, win rates, and damage amounts. They’ve done a lot of work to harmonize party names, for example, so you can find all the cases filed by a particular plaintiff even though there are multiple names used. Academics can get access to the Clearinghouse before its formal launch by emailing the Clearinghouse—contact the executive director.

Modern Trademark Law and the Right to Make Derivative Works
Mark McKenna
Notre Dame Law School

The core of trademark protects entities from direct competition; he raises no serious questions about that. Around the core, there can be confusion from noncompeting goods; then further out are outlier doctrines like initial interest confusion and dilution. People focus on the outside ring. Beneath all that criticism, there’s a widespread acceptance of the second ring, confusion over noncompeting goods, and that’s where all the real junky doctrine is as well as most of the junky cases. We would be better off spending more time on confusion over noncompeting goods and less on outlier doctrines.

His focus: claims by trademark owners against others using the mark/similar marks for noncompeting goods. Justifications for such claims: (1) Reputational feedback—bad junior products will harm the senior mark; (2) control—we don’t require proof on (1) because we believe that without control, at some point quality may decline; (3) market preemption—this is a key argument at the time TM expands to noncompeting goods: if we don’t protect a mark owner against noncompeting goods, the mark owner won’t be able to expand into that market; and (4) free riding. (3) and (4) suggest that the TM owner ought to have a superior right to expand into new markets, e.g. Coca-Cola into snack chips, because the junior user isn’t doing anything valuable.

These are all about producer interests, not consumers. And they’re all claims about brand dilution—not about present harms, but about TM owner’s ability to operate in the future.

Can any of these claims be tested against empirical evidence? Brand extension and brand alliance (e.g., Intel Inside on a Dell) literatures. The brand extension literature approximates true source confusion: consumers think that the product comes from the senior user, because they’re told so in these studies. His question: Even when they’re confused, what’s the harm that flows from that? Likewise, the brand alliance literature approximates sponsorship or affiliation confusion.

The studies look for forward (spillover) effects on the new product, as well as reciprocal (feedback) effects on the original, including overall assessments (quality) and specific brand attributes (Neutrogena is mild). Most studies don’t make global claims, but McKenna sees certain themes emerging.

For extensions, forward spillovers to the new product depend on whether the original brand is a good fit for the new product. Backwards effects on perceptions of global brand quality: none. unsuccessful extensions may impact evaluation of future extensions, at least for lower quality core brands. Effects of extensions on specific brand attributes: this is complicated, but they can be affected by congruence and motivation; if Neutrogena began selling sandpaper, consumers might think Neutrogena was less mild as an overall brand, even though their opinion of the hand soap remained the same. Congruence has to do with whether brand attributes fit together. Congruent extensions benefit core brands; incongruent extensions hurt only if familiarity is high (though in that case consumers are less likely to be confused).

Alliances: They just don’t affect perceptions of the core brands. This casts real doubt on the harms from sponsorship/affiliation confusion. Even true source confusion doesn’t cause much feedback harm, and its costs may be offset by benefits of reminding consumers about the brand.

Thus, we aren’t talking about robust evidence of present harm. This is about the ability to make “derivative works.” This is more like copyright/other forms of IP because it is about allocating markets between producers, not anything about consumers.

Lemley: Does this depend on the marketing literature? Should TM be a norm-follower?

McKenna: He’d be happy to talk about norm entrepreneurship, but the doctrine makes empirical claims, and those claims don’t hold up. (Comment: and this is true even given that the studies took place in a context in which the law already protected TM owners against use on dissimilar goods—it’s not the belief in the connection that’s at issue; it’s the effects of that belief, which apparently haven’t been shaped by law.)

Jennifer Rothman: There’s still a role for consumers to be concerned about those uses, in particular with source confusion. It will influence purchasing choices, consciously or unconsciously, if a company starts calling its cars Sony cars. This is a harm to consumers.

McKenna: His goal is to get us to think about both pieces of the puzzle. They are tradeoffs. Where producer interests are really strong, we might be less concerned about consumers. Where producer interests are low, we need a compelling consumer-based story to maintain those rights. The extension literature suggests that consumers are much more nuanced and complicated than we give them credit for. In some cases, they will rely on the brand. But not in nearly as many cases as we assume, so let’s look for the factors that do prompt such reliance. (Comment: consumers might have a cause of action, but is there a reason Sony needs to sue? It’s GM that has the problem in Rothman’s hypothetical.)

Kathy Strandburg: What about brands as identity, community, etc.? People like having “Apple” stuff because they like the computer.

McKenna: It’s real, and poorly accounted for in current TM law. It’s hard to imagine retrofitting TM law to account for it. Sociological impact is true; but it’s also true that there are lots of “apples” out there. So he doesn’t know how we’d rebuild TM law from the ground up, looking at sociological meaning instead of product information, though dilution law may be pressing in the direction of accounting for soft meaning. If we don’t just give an exclusive right to “apple,” it’s very hard to figure out which “apples” are going to interfere with meaning.

Mark Janis: Does the literature tell you anything about “famous” marks?

McKenna: Marketing people don’t treat fame in the same way we do. They focus on marks as compared to other marks in their categories—dominance.

Rethinking the Patent System's Early Filing Doctrine
Abstract | Paper
Christopher Cotropia
University of Richmond School of Law

Remember that patent caveat above? It applies.

Various rules remove barriers to, or incentivize, early filing. There are numerous articulated benefits to this, from ending patent races to earlier patent expiration.

Technology goes through cycles, beginning with many technical solutions that are winnowed before eventual commercializations. An early filing date encourages filing before a lot of market information is present.

Costs of the “File early, file often” mentality in which people err on the side of filing: (1) More applications, more patents; always have a continuation on file, and have a lot of follow-on filing, negating the benefit of early filing for earlier expiration, since the Patent Office gives you credit for the time they take. (2) Underdevelopment of issued patents (referencing an article by Michael Abramowicz) because people wait for more market information and don’t develop the technology, and because of the relatively high cost of commercialization versus the cost of obtaining the early patent—an early patent turns into a cheap option with a relatively high exercise cost. (3) Encourages trolling, since people don’t have to invest in commercializing the technology; they can just file suit. (4) Leads to unclear patent boundaries.

Proposal: eliminate constructive reduction to practice and require actual reduction to practice, including observation that the invention works for its intended purpose, and this would need to be included in the specification. That pushes the filing date forward.

Comments focused on the effects on smaller firms. Cotropia made the point that in certain industries, such as pharmaceuticals, huge resource investments are already required to get to the patenting stage to meet other requirements such as utility. His proposal aims to develop additional market-relevant information.

Lemley: The underlying assumption is that there’s not net social value in having bunches of people come up with ideas and never commercialize them. Semiconductor industries, for example, have a lot of failure to reduce to practice. Do we benefit, in a theory of the firm sense, from being able to separate out invention and purchase it from people outside who are just able to come up with ideas?

Cotropia: Those ideas may be beneficial, but patents might not be the right way to do it. Contracts could monetize the relationship between the idea person and the prototype person. (So now Cotropia is walking towards trade secrets.)

'Gift Failure' versus 'Market Failure'
Wendy Gordon
Boston University School of Law

We know our current copyright systems are wasteful, but we have thought that they were less wasteful than the practical alternatives. Is it time to look at alternative models, and not wait for market failure before looking outside the market? One possible default is the commons, a starting point for figuring out where social and economic relations should go. With the growth of the internet, the failure of patent, the importance of lead time even in the absence of IP, and the development of Creative Comments and the GPL, it seems time to start looking at nonmarket models more systematically.

Gordon here focuses on gift. For 3 kinds of intangible products, gifts should be the starting point. Nobody imagines that a perfect market exists; nobody imagines that a perfect gift exists. But that concept can be a good place to start.

Her 3: (1) High culture/high art; (2) peer science; (3) software programming. (1) and (2) persistently involve people who do the work saying that they’re producing gifts—the scientists and artists involved constantly talk about gifts, from the sense of a muse giving them ideas they wouldn’t otherwise have to the desire to share and the experience of pleasure in sharing with others. (1) and (2) are areas in which doing the work is the point of doing the work, not to make money or get famous.

They’re not indifferent to money or prestige, of course. But direct payments can hurt quality—the work of Teresa Amabile shows that direct payments for art degrade performance while increasing quantity. See also Punished by Rewards. Nonetheless, we need to recognize the need of the producers to eat, so we sometimes need rewards, direct or indirect. The research suggests that it makes a huge difference how the rewards are provided—directly or indirectly; slowly or fast; as a return gift—these things can sustain the artist’s creativity without weighing it down.

Community matters, especially to the impecunious poet or struggling writer. Some people need to respect a bohemian, less profitable lifestyle and freely critique the work, or other people will not be able to stay in that lifestyle long enough to become successful as artists (or economically). What encourages and inspires poets/scientists is the work of other poets/scientists. Create a community, and art and science flow from it. (And people in the community may have psychological reasons to reject commercial success.)

Why these fields and not others? Because some situations induce gift failure, and if you know in advance there’s going to be gift failure then there’s no point in using gift.

Definition of gift: there’s no tit-for-tat, bargained-for remuneration. It fits the need for a reward that isn’t exactly a reward.

Is this all hypocrisy? Any contradiction can be explained away—scientists repress their need for prestige/credit. But there is some bad faith. Jeff Koons is interesting because he takes advantage of the culture of modesty in a way that is openly exploitative.

The big problem with gift is reciprocity, the pressure to pay back. The GPL says “we don’t care about reciprocity, just use our stuff.” But in gift economies generally, the gift must circulate, and we still need to give our creators the means to eat. Gordon’s model: soft reciprocity. Jean Schroeder has written that contract is superior as a meeting of equals, while gift is a relationship of subordination. In all anthropological studies where gift is exchange, people have things they value and gift is about getting rid of what you don’t want by tricking the other party.

But gift can also be “keeping while giving”—especially when it comes to IP, because IP is intangible and inexhaustible. Through attribution and memory, people know where the gift came from.

Should we eliminate patent and copyright entirely? No, that’s an example of gift failure. A gift should be voluntary for the giver and the receiver, and have soft reciprocity to take the sting out.

What about the GPL? That’s a conditional gift: you can copy my software if you do certain things. Reciprocity doesn’t naturally arise in the recipient; it forces the recipient to put his own stuff in the gift world. But that’s a tolerable distortion—it doesn’t force destructive, Veblenian competition on the givers.

Eric Goldman: What about the broader relationship between giver and donee? He thinks of gifts in a personal context. How can this work in a commercial context? Reciprocity may break down if there are few social contacts between the parties.

Gordon: She’s not looking for the true or essential gift, but some model that will create and sustain creative communities. The personal gift is a way to take the sting out of reciprocity—the perfect gift to someone you know well shows that you are taking account of them as a person, recognizing them. When you generalize, the gift is no longer tailor-made, but there are other ways to eliminate the sting.

In a truly commercial realm, there is likely to be gift failure.

Q: What about managing boundaries, as with commercial and noncommercial scientists?

Gordon: Gifts are a way of managing boundaries. If you accept this gift, you are part of our community, and for the GPL that means a commitment to certain kinds of noncommercial use.

Niva Elkin-Koren: Why describe all these as gifts? You assume ownership of a gift, and the right to withhold it.

Gordon: It can be a labor relationship, not an in rem relationship.

E-K: What do you gain by calling it gift instead of “communicate, share, care”? Gift implies something that is exclusively mine that I can possess or withhold. When I share my opinion, it’s not a gift. When I share my scientific discovery, it’s not a gift because there’s no give and take.

Gordon: This relates to E-K’s critique of Creative Commons as overemphasizing copyright/ownership. Gordon doesn’t see a divide between “share” and “gift.” There is a danger of making people too conscious of “property” when gift might have been natural/organic without the legal structure. With the GPL, a gift success, programmers were already very conscious of the IP regime. For fine artists, who knows?

Bartow: Look at gifts to universities as gifts that are badly motivated—publicity, help the grandkids get in.

Gordon: If all a gift does is hurt other people in a prestige competition, that’s gift failure. If it ends up with a new lab built, that can be a gift success. Prestige/publicity isn’t a bad thing, especially if people aren’t compelled to give.

Reputation Regulation: Rationalizing Internet Intermediary Responsibility
Abstract | Paper
Frank Pasquale
Seton Hall

This paper could be called “the law of Google, eBay and Facebook.” We love Google, but what if Google decides to make Google Books like Lexis or Westlaw, so that you have to pay to get access? What if Google manually changes rankings by downgrading sites that sue it, or upranking sites of business partners? Should Google at least have to disclose this? Should eBay be able to favor certain sellers, like Disney, and disable search functions for competitors’ merchandise? Should law be part of making eBay a level playing field? Should Facebook be able to kick members off without due process?

Intermediary markets are very concentrated: eBay, social networking by country, Google in search. The new neutralities: net nondiscrimination, devices (iPods), operating systems, and intermediaries. Regulate intermediaries when competition is unlikely to develop, especially search engines and auction sites. It is important to recognize social and cultural effects, not just economic analysis.

Dominant search engines and social networks are becoming like other entities who control other layers—the concerns are for common carriage of data; discrimination based on the source of data; transparent routing; and vertical integration—“will Comcast charge more to let you watch YouTube?” is the same sort of question as “will YouTube cut special deals with Universal and not with Fox?” There are also political issues with control of substantial methods of communication by dominant entities. Pasquale’s overriding point is that net neutrality arguments map really well onto Google and other dominant intermediaries at the content level.

So, for example, there should be due process for banning on social networks, and data portability to make it easier to leave. Auction platforms are less troublesome; note that eBay has added substantial elements of community and democracy, to moderate the control imposed by being a “company town.”

Intermediaries are cultural voting machines. If we’re concerned about transparency in voting machines, we should have the same concerns about online intermediaries. We shouldn’t rely just on the market and the black box of algorithms subject to hidden manipulation.

Lemley: His worries about regulation cover more than just expertise. Capture is a huge problem in telecom. Restrictions on innovation are also troubling. The more you advocate a model that is beyond openness/data portability, but allows the government to determine what you put on your site and in what order (e.g., TM owners allowed to attach info to the search results) the more you affect innovation. Think about levels of intrusiveness.

Pasquale: Interoperability is so hard to arrange—like Google and AdSense data portability—how hard is it to reveal that data? It might be an iterative process rather than a one-off mandate.

Q: Ten years ago the referent would have been Microsoft, not Google. Fun shift to note? Separately: the thing about network effects is that replicating structures can be wasteful, so regulation is a better substitute. But that looks like the costs of all software—a declining marginal cost is a characteristic of software generally, not just market dominant software. So what does that mean for regulation? Open access reinforces network effects and reinforces the dominance of the existing platform.

Pasquale: That’s true. But he doesn’t want to accept a purely economic framework for social networks and search engines. Search engines have an advantage over legacy networks like credit reporting bureaus; should we level the regulatory playing field as between them?

Susan Crawford: The carriers have successfully convinced us they’re content companies, and that’s quite a rhetorical move, because for 100 years we treated them as common carriers. There could be social reasons to treat general-purpose, physical infrastructure differently from everything else. But she doesn’t agree that these internet intermediaries are the same as the telephone or the telegraph. Carriers are carriers, and net neutrality is about pushing them back into the appropriate category, which is carriers and not content providers.

Pasquale: He doesn’t want to apply common carrier rules point for point, but to think about commonalities.


  1. Matthew Sag2:53 PM

    I find it hard to accept the premise that there is no effective competition in the social networking field. Check out the list of such sites on Wikipedia

    If we tightly regulate the format, content and processes of things like facebook, how do we expect to get innovation in those areas. it would be better in my view to concentrate on making sure technologies that would enable switching are not unduly hampered by the DMCA or a narrow understanding of the fair use doctrine.

    Matthew Sag

  2. Matt: I propose a sliding scale of regulation--to the extent competition is unlikely, regulation is more warranted. I have not said that there is "no effective competition" in SN's--though a) it's easy to see how tipping points can be reached and b) I think the data on concentration in SN markets is revealing: see the charts in David S. Evans, Antitrust Issues Raised By The Emerging Global Internet Economy, 102 Nw. U. L. Rev. Colloquy 285 (April 28, 2008) citing comScore, MyMetrix Key Measures Report (Dec. 2007); comScore, My-Metrix qSearch 2.0 Key Measures Report (Dec. 2007).

    My only proposals at the presentation with respect to dominant SNs were:

    1) Due process for the “banned”
    2) Public knowledge of how salience is allocated (i.e., restrictions on stealth marketing)
    2) Data portability: Here the model may be Zittrain’s proposals for device neutrality. There is a parallel between the right to transport your music and to transport your social graph.

    I am not proposing tight restrictions on the "format, content and processes of things like facebook." And data portability may be our only shot at innovation--unless you subscribe to the naive Schumpeter/Rand view I mentioned in the talk.