Scarcity of Attention in a World without IP
We impose artificial scarcity so authors can pick up sunk costs. But costs of creation have fallen; perhaps © should be narrowed as a result. Wrinkle: attention scarcity. Cheaper dissemination = information gluts. Imposes costs on consumers. The scarce resource today is attention. People compete for attention by riffing on cultural artifacts. Some of the competition is zero sum. Internet time has shifted 2008-2013, from chat and news sites to social media and video.
Trademark is a way of dealing w/scarce attention, but © is different. Music played in restaurants: it draws in consumers. Attracting attention w/© expression so they can charge you. [Is attracting customers for food the same thing as attracting attention?]
If we had less copyright protection [compared to what?], then what? Perhaps: Less entry; less glut. Narrower protectionàlower expression costs, lower costs of creating around, lower barriers to entry from building off of existing ©. Retain protectionàmore costs of inventing around. Less creative/distinct entry. Things that are more similar. Abramowicz/Yoo have talked about this as rent dissipation/product differentiation: more resources dissipated racing to capture first mover advantage.
Fishman’s work: © may impose useful restraint on entry. [I have written a response to Fishman here.] Constraints on type of creativity: sonnet constrains output more than free verse, which can make work more creative/better by several metrics. George Lucas wanted to do a Flash Gordon remake; he couldn’t get permission so he gave us Star Wars—wouldn’t have been the same cultural phenomenon. [Because reboots of existing franchises, like Batman, never become cultural phenomena!] Reducing © protection is therefore premature.
Mark Lemley: If you think we’re creating too much derivative stuff, ok, but challenges premise that attention scarcity causes that. Attention scarcity = difficulty sorting is the theory, but people have worried about disintermediation for a long time, and crowdsourcing is at least as good and maybe better than expert selection in allowing us to find good/bad. If 20 people make a derivative Star Wars movie, but the crowd is good at finding the right one.
Linford: if the crowd tells us the first Star Wars movie is good, 20 people rush to follow, and Lucas has to rush out his sequel. There is also intermediation via FB, Google, and Spotify, and maybe they’re not sorting in our interests. [But that is not the same thing at all as causing an overload problem.] We’re not getting the full picture. [We never were.]
Q: why allow the public domain then? Pride and Prejudice and Zombies.
Linford: that’s a strong form of the argument—doing something with a work that I don’t like is wrong b/c I have a strong connection to the argument. Different question about length v. breadth while in place. People share that intuition, but the evidence isn’t strong enough.
Betsy Rosenblatt: we live in a world with trends. The solution is that people tire of sameness. Doesn’t seem to be a need to do anything other than let the trend run out.
Linford: The more you narrow ©, the more similar the horror movies are.
Rosenblatt: Why is that bad?
Linford: negative quality. Bigger rush to move sequels through the process. What we really want is new takes on the trend, and © provides some discipline to make things at the margin more different than they would be. [This is not what the literary scholars think. See David Roh’s book. More production around a type by people trying to work with what audiences liked about the original is how you get the kind of variation that creates whole new genres and subgenres. Consider Twilight and its progeny; you may not like them, but there’s no question that Twilight, a sexualized but not explicit supernatural romance, was part of a causal chain creating a billion-dollar “New Adult” genre.]
Linford: reputation matters in part b/c of incentive to produce the next installment. [This account works ok for movies if you ignore that people care about who the actors/directors are, but terribly for singers. Very few albums are “sequels,” even Adele’s.]
Q: it’s an empirical Q whether variety makes people better off, or whether things that are more different from each other are actually better b/c of that. It’s complicated. You need to support that at the level of a particular genre or market.
Linford: genres might matter, but we do a lot of thinking about theoretical work/modeling w/o empirics.
Q: but the models don’t predict that variety makes people better off. You need an empirical analysis of the genre/market.
Linford: disagree w/©.
[I think the argument conflates lack of scarcity and diversity except when it equivocates about them. We also impose artificial scarcity to incentivize dissemination. The idea is that we get “more” copyrighted works overall. There are two kinds of “more,” creating the classic incentives/access tension. But you have to define your “more”—is it lots of Batman merchandise/Frozen merchandise, which © incentivizes and lack of © incentivizes, or is it lots of comic books? To the extent that ©’s idea/expression distinction allows close cousins, it also allows you to manage your attention very easily: I tune out ads for horror movies; I don’t tune in to Fox News.]
What's In vs. What's Out: How IP’s Boundary Rules Shape Innovation
Mark McKenna & Christopher Sprigman
McKenna: Every IP rule defines itself in some way through negative relation to patent. TM doesn’t allow utilitarian functionality. © is the same. Design patent focuses nominally on ornamentality, defined as inverse of functionality; concern was of people avoiding rigor of utility patent standard by getting design patent instead. Doctrine is noticeably weaker in exclusion than TM and ©, but in all cases the matter isn’t excluded b/c it doesn’t meet ordinary standards for protection in ©, TM, etc. but rather b/c this matter is for utility patent: to avoid undermining utility patent supremacy.
This only works if other systems have a reasonably clear and stable sense of what kinds of things belong to utility patent. IP sense of patent’s domain is mostly reductive and intuitive; undertheorization of utility patent’s boundaries is one reason courts have struggled so much to channel. Utility patent law itself has been quite inconsistent on its domain. This is more than a doctrinal problem. If you believe that IP exists to create incentives, you can’t be ambivalent about this point. IP rights as substitutes for each other: create incentives that create types of production. You also still need an explanation for the subject matter of our different rules even if you don’t care about incentives.
Sprigman: Utility patent focuses on technological innovation, but what does that mean? Historical sense of meaning of Useful Arts, working from machines, chemical compositions, etc. Most important to this is nonobviousness. Patent law has been inconsistent: look at utility patent for mixed cut gemstone which enables appreciation of desirable characteristics of a diamond in ways that prior cuts don’t allow. Is that functionality? Design patent for gemstone cut was also granted; court says that’s just pretty, not functional. Patent on method for assessing character: what kind of utility is this? Not technological. Utility patent is thus inconsistent about its own grounding.
McKenna: Conundrum: widespread belief in utility = tech innovation, but in practice inconsistently enforced, yet every other are of IP acts as if tech innovation is where they should defer to utility patent. But in many cases TM courts struggle to understand with what it means to be functional—and then there’s aesthetic functionality, which has always been controversial b/c some people consider it an oxymoron. Even though the SCt endorses the idea, it applies different rules—competitive need rather than Inwood/Traffix rules. Undertheorization of functionality thus means courts don’t even recognize that they’re using functionality in a particular way. Features can be useful as marketing tools, to make things look better, etc.; tech utility is a normative choice, which doesn’t make it wrong but does make it worth talking about.
Star Athletica panel agreed that the design should be evaluated by asking what the functions were and then seeing if they were separable. For majority, function was to cover the body. For dissent, the purpose of the design was inescapably to identify the wearer as a cheerleader, making chevrons etc. inseparable. Case turns on court’s understanding of what should be kept out of © as functional.
Sprigman: how to fix? One approach: if patent is open to aesthetic utility, then exclude much more from other fields. Or we could explicitly limit utility patent to tech innovation and be rigorous about it w/in patent too.
Q: Ted Sichelman: a lot of this goes back to historical differentiation of guilds, dividing tech and fine arts. Based on institutional concerns, modes of production; do changes in these things justify rethinking the boundaries?
Sprigman: the kind of education of people who enter into the patent system is very different; part of difficulty w/design patents is that the PTO doesn’t attract people w/aesthetic education to understand trends and prior art.
McKenna: Perhaps the features of the patent system are suited to the kinds of innovation it generally covered. If that’s the case, then we might want to avoid growing the system.
Q: what’s the right answer?
McKenna: collective effort is required. The paper tries to show that IP has a problem as a system and that it needs to be addressed. Makes it easier for people to use one right to protect that which other rights want to leave free for general use—overprotection. Most areas of IP haven’t internalized the systemic costs, not accounted for overlapping rights, thus haven’t calibrated right.
Sprigman: if pill efficacy depends in part on color b/c of placebo effect: is utility patent going to treat that as patentable? Might be within patent subject matter but not patentable because obvious. If so, TM will kick it out. Is that good? If utility patent doesn’t accept pill color, then will TM accept it? Or will TM start thinking about competitive need for pill color? If you think placebos respond to incentives, we might want to exclude it from TM.
Q: If all regimes have delineations, why do we need to worry?
McKenna: Not everything will fit somewhere; we also need to be concerned b/c TM etc. make claims about what utility patent is/does and if those are wrong we have issues.
We don’t know much about what’s in patent claims. They’re supposed to point out and distinctly claim the invention. Many empirical studies about patent litigation, few about claims. Textual analysis of 25,000 patents. Are claims really only about the invention? Test: does claim language appear in the specification? Result: about 25% of claim language appears rarely or never in the specification.
Invention in claims must be discussed in the specification; if not discussed, unlikely to be new/used in a new way/combined in a new way/etc. “Non-inventive language.” E.g., list of pharmaceuticals that are known to be delivery devices—not really what the invention is about. Average of 9.4% of claim words are nowhere in the specification. About 90% of noninventive language is in dependent claims. Highest in chemistry and pharma (27%), computer lowest (21%). Patents w/more noninventive language roughly correlates w/fewer forward citations and lower renewal rates.
Limitations: probably underestimates noninventive language b/c a claim term can be mentioned many times in specification and not be about the invention. Can’t account for drawings.
Why? Signaling: patents are marketing documents. Decoy function: if you claim 100 different things, competitor doesn’t know which you’ll choose. To avoid rejection during prosecution. Insurance against future commercial uncertainty or litigation—to make it more likely that your claim covers the commercial embodiment.
Implications: may not satisfy §112. Clarity. Have to be one sentence long; sometimes noninventive elements make a claim 4 pages long. Words by definition aren’t defined in specification, and if they’re vague that increases the challenge. But exemplars can increase clarity of general terms. Prosecution: examiners struggle w/noninventive language; can be part of strategy to run down examiner; creates difficulty of search by generating false positives; creates impresssion of patent thicket, but if term is part of noninventive language, patents may not be relevant.
Claims are doing more than just describing an invention. Is there a heart of the invention? What does the ideal claim look like? What is its purpose? May not be possible for claims to be only about the invention.
Jeanne Fromer: amended claims are written by lawyers, not inventors. Hard to make comparisons.
Freilich: synonyms are discouraged; prosecuting attys are careful about trying to satisfy written description/enablement. Great idea to look at amended claims. About half of patents add noninventive language during prosecution, apparently in response to examiner rejections. 25% take out noninventive language during prosecution, and the remainder stay the same.
What We Buy When We “Buy Now”
Aaron Perzanowski & Chris Hoofnagle
What does it mean to own or buy something in the digital economy? Amazon’s remote deletion of copies of 1984 was a particularly notable incident, but happens quite often. Empirical testing of different consumer options.
Takeaway: consumers are confused about what rights they acquire. We asked: can you lend this book to a friend, resell it, give it away, leave it to someone in your will, keep it as long as you want? Most people believed of digital goods that they could own, keep, have it on any device; 1/3 think they can leave ebooks in will, give them away; ½ think they can lend them. Resale and copy were the only ones where about 15% thought they could do it. “License now” instead of “buy now” created many fewer responses that they “owned” the book but the other operations like lend, give in will, etc. remained very similar. Short notice about rights for ebooks did a bit better in terms of reducing beliefs in right to lend, gift, resell (though many percentages were still above the #s that often indicate deception in reasonable consumers). Women and people over 65 tended to be more correct than younger white men, who tended to have a sense of entitlement over what they could do with their “purchases.” [Which is a reminder that entitlement has good and bad aspects.] On average, people who saw “license now” and “buy now” got less than ½ of questions about rights correct; short notice increased correct answers by one, whereas “buy now” for hard copy were mostly right, though not all thought they had the rights they did. Preferences in terms of rights were the same for ebooks and hard copy; 54% were willing to pay more for at least one of three control rights, often $1-3 for products ranging from $6-12. Lack of rights makes them more likely to download illegally and to stream.
Justin Hughes: did you adequately separate digital file from device in Q about lending/gift?
A: we tried to make it clear that you were asked about lending a digital file, not a device, but that is definitely a concern.
Eric Goldman: Another story to be told from your data: it’s really hard to inform consumers. Many still aren’t getting the message w/the short notice.
A: They had one exposure to the short notice—I would think that repeated exposure might increase results. [Probably not, which he then notes as a potential for ignoring the standard notice.] We aren’t pro designers/ UIX designers—a professional might well do a much better job. What this tells me is that this is a promising avenue for improving consumer understanding. [Richard Craswell has done excellent work on the CBA aspects of this—how much we should invest to inform those consumers who can be informed, even if we can’t fix everything.]
Q: re change over time.
A: there’s a level of dissatisfaction w/inability to get kinds of rights they expect. The only thing growing as quickly as streaming is vinyl records: people are willing to pay for permanence, but the sellers don’t want to sell. [And of course there’s the classic xkcd making this point, not to mention Calibre and its easily available plugins for your ebook backup needs.]