As Henry and Jim both note in their comments, there are doctrines within intellectual property law that are entangled with the servitude issues that I discuss in The New Servitudes. Specifically, both the first sale doctrine embodied in section 109 of the Copyright Act and the software-specific provisions of Section 117 that Jim mentions allow owners of copies of copyrighted works to do specified things with those copies notwithstanding the exclusive rights of copyright holders. In the patent context the analogous judge-made doctrine is patent “exhaustion”—the meaning of which the Supreme Court is currently considering in the LG v. Quanta case to which Henry refers. I’m anxious to discuss Quanta. But first let me give a little background (familiar to those who have read the paper) about the connections I see between first sale, exhaustion, and more general concerns with servitudes.
I argue in the paper that the early twentieth century cases in which these IP doctrines developed reflect an anxiety about enforcing servitudes attached to items of personal property (an anxiety that is in turn explained by the concerns with notice, “the problem of the future,” and externalities that I described in my initial post).
For example, in Bobbs-Merrill Co. v. Straus, the case that established the (later codified) first sale doctrine, a book publisher attempted to enforce a restriction printed inside books that purported to limit the price at they could be resold. The plaintiff copyright owner argued that by distributing books subject to restrictive terms it was granting purchasers only a conditional license to exercise the vending right that the Copyright Act otherwise granted exclusively to it; vending outside the terms of that license (that is, selling books for more than the minimum price) therefore amounted to copyright infringement. The Copyright Act, according to this logic, provides a mechanism for imposing running restrictions on chattels that the common law lacks. The Second Circuit rejected that argument, and the Supreme Court affirmed. 210 U.S. 339 (1908).
The Court’s opinion purports merely to interpret the language of the Copyright Act, concluding that the right to “vend” granted in the Act is exhausted as to a given copy of a copyrighted work once that copy is sold. But this interpretation seems to have been motivated in part by the more general notion that chattels should not be burdened with running restrictions. Now, to be sure, the statutory rights granted to copyright holders supersede that hostility to some extent: there are some things that the owner of a copyright-embodying chattel is not permitted to do with it (for example, reproduce each of its pages) on account of the non-possessory intellectual property rights created by copyright (which have, as I noted in my earlier post, a servitude-like character). But Bobbs-Merrill articulates limits on even a copyright holder’s power to impose running restrictions on personal property. In the patent context, a similar logic seems to motivate cases like Bauer v. O’Donnell, 229 U.S. 1 (1913), in which the Court refused to enforce a retail price restriction printed on packages of patented medicine because “a patentee who has parted with a patented machine by passing title to a purchaser has placed the article beyond the limits of the monopoly secured by the patent act.”
These cases (and other similar cases including Straus v. Victor Talking Machine and Motion Picture Patents Co. v. Universal Film Manufacturing Co.) are complicated by the fact that the substance of the restrictions was arguably anti-competitive--a fact that has led several commentators to suggest that the general hostility toward chattel servitudes (and the specific IP doctrines of first sale and exhaustion) could be replaced with particularized antitrust scrutiny--especially now that antitrust law is more doctrinally mature than it was when these cases were decided. Richard’s post suggests that he would endorse this view. And it is also reflected in the Federal Circuit’s 1992 decision in Mallinckrodt, Inc. v. Medipart, 976 F.2d 700 (1991).
In Mallinckrodt the Federal Circuit characterized the early patent exhaustion cases as “establish[ing] that price-fixing and tying restrictions accompanying the sale of patented goods were per se illegal. These cases did not hold, and it did not follow, that all restrictions accompanying the sale of patented goods were deemed illegal.” The court, therefore, refused to interpret those cases as necessarily invalidating a use restriction that neither fixed prices nor tied multiple products but rather limited a patented medical device to “single use only.” The court suggested that so long as the condition did not violate the antitrust laws, it was enforceable against a purchaser of the medical device through a suit for patent infringement. The Federal Circuit reaffirmed this logic in LG v. Quanta, in which the Supreme Court recently heard oral argument.
To my mind, the Federal Circuit has adopted an overly narrow view of the purpose of the exhaustion doctrine. Safeguarding competition through antitrust might deal with some types of negative externalities, but it does not eliminate every problem associated with chattel servitudes. Problems with notice and information costs, and with unjustified but inescapable burdens on future activity (the problem of the future) can still loom large.
There are some interesting allusions to these problems in the LG v. Quanta oral argument. Justice Breyer suggested that the “doctrine that you cannot impose equitable servitudes upon chattel” might help to decide the case. And he returned several times to a hypothetical in which a patent holder tries to enforce restrictions on a consumer’s use of patented bicycle pedals:
“Imagine that I want to buy some bicycle pedals, so I go to the bicycle shop….The inventor has licensed somebody to make them, and he sold them to the shop. I go and buy the pedals. I put it in my bicycle. I start pedaling down the road. Now, we don’t want 19 patent inspectors chasing me….”
Breyer apparently finds this a troubling scenario even if the pedals come with some sort express notice of the use restrictions. He asked whether
“if … I go in the bicycle shop and I buy the pedals and then they give me … one of these pieces of paper that has all of the 42,000 words on it and there are in these 42,000 words it says, and now you are put on notice that once you put it in your bicycle and you pedal away, they’re going to get you and you’re going to be hauled into Patent Court, then--then that’s ok?”
I don’t want to read too much into the tea leaves of this oral argument transcript. But Breyer seems to be suggesting (in part relying on the traditional rule against chattel servitudes) that consumers would be justifiably surprised to find themselves “hauled into Patent Court” for merely using things they had lawfully acquired, even where the restriction was imposed by a patent holder by means of a written notice attached to the good. As Henry's post suggests, the problem here is not exactly lack of notice--the information is available in the “42,000 words” in Breyer’s hypothetical. But it may not be reasonable or even desirable to expect people (at least not retail pedal purchasers) to absorb that information. The Court expressed this same anxiety back in Straus v. Victor Talking Machine Co., in which it refused to enforce a restriction printed on plates attached to record players after noting that “not one purchaser in many would read such a notice, and that not one in a much greater number, if he did read it, could understand its involved and intricate phraseology.”
This gets us back to Lee’s question: don’t the problems that I associate with running servitudes also plague restrictions that don’t run--e.g. plain vanilla contracts (which are also often unread, misunderstood, and not fully contemplated)? What is especially problematic about servitudes? That is a critical question to which I will return in my next post.
Molly,
It's easy right now to see the notice and information costs and problems of the future (to which Breyer's hypos allude).
What's harder to see is how (and how fast) these will decrease over the next ten years thanks to new technology. In the patent arena, the transaction costs that you've mapped out have already dropped (and are dropping at an accelerating rate) thanks to the wider availability of patent disclosures and search algorithms.
Could you say something about how we could build some empirical release valve into any rule that we come up with that takes account of decreasing transaction costs?
Posted by: Michael Martin | February 07, 2008 at 11:15 AM