Should patent infringement require proof of copying?

• Author: Lemley, Mark A.
• Position: Vermont

TABLE OF CONTENTS I. THE BENEFITS OF AN INDEPENDENT INVENTION DEFENSE 1526 II. RISKS OF AN INDEPENDENT INVENTION DEFENSE III. ALTERNATIVES TO INDEPENDENT INVENTION A. Willfulness B. Prior User Rights C. Secondary Considerations of Nonobviousness D. Injunctive Relief IV. CONCLUSION Patent infringement is a strict liability offense. Patent law gives patent owners not just the right to prevent others from copying their ideas, but the power to control the use of their idea--even by those who independently develop a technology with no knowledge of the patent or the patentee. This is a power that exists nowhere else in intellectual property (IP) or real property law, (1) but it is a one that patentees have had, with rare exceptions, since the inception of the Republic. In an important paper in the Michigan Law Review, Samson Vermont seeks to change this, arguing that independent invention should be a defense to patent infringement, just as it would be in copyright or trade secret cases. (2) In an era in which both the Supreme Court and Congress are showing a nearly unprecedented interest in patent law and in which there is a general sense that the patent system is out of whack, Vermont's idea may be one whose time has come.

1. THE BENEFITS OF AN INDEPENDENT INVENTION DEFENSE

   An independent invention defense has much to recommend it. The most significant problem facing the patent system today is the rise of so-called "patent trolls"--entities that do not manufacture products or transfer technology, but instead assert patents against successful companies that independently develop and manufacture technology without knowledge of those patents. Patent trolls have a number of tools at theft disposal. They can file continuation applications and modify their invention to track changes in the defendant's product. (3) They can threaten injunctive relief and seek damages out of proportion to the importance of the invention, allowing them to capture significantly more than just the value of their technical contribution. (4) In the information technology industries, it sometimes seems as though the overwhelming majority of patent suits are not brought against people who copied a technology, but against those who developed it independently. (5)

   An independent invention defense would eliminate the troll problem. (6) It also comports with our sense of equity. Those not schooled in patent law would likely find it odd that a patent not only prevents the imitation of the patentee's technology but also limits the ability of inventors to develop and market their own technologies. And while an independent invention defense would be new to U.S. patent law, it is hardly unprecedented. Not only do other U.S. IP laws have such a defense, many other countries provide a "prior user right defense," which, as we will see, is akin to independent invention. (7)

   Vermont analyzes and dissects a number of common objections to an independent invention scheme for patent law. He responds persuasively, for example, to the concern that people would exploit an independent invention defense by falsifying evidence of independent invention. As he notes, outright fraud and perjury are rare in the patent system, even when more is at stake than proving an independent invention defense. (8) If people won't lie about when they invented to get patent rights, it seems unlikely they will lie to get the lesser benefit of a defense to infringement.

   The heart of Vermont's article is a response to the concern that an independent invention defense will weaken patent rights and therefore undermine incentives to invent. (9) Vermont uses Bayesian analysis to argue that while an independent invention defense will weaken patent rights, it will not significantly affect invention incentives. He points out that independent invention will succeed as a defense only when an invention is made by multiple people in near simultaneous fashion. Vermont's core insight is that the fact that many people succeed in making an invention provides persuasive evidence that a monopoly right is an excessive reward because the invention would have been made even without granting exclusive rights to one of those inventors. (10) Since the point of patent law is to encourage invention, not to satisfy some moral entitlement, (11) we should limit a patent owner's rights--and therefore facilitate competition--in any case where a stronger right isn't necessary to induce invention.

2. RISKS OF AN INDEPENDENT INVENTION DEFENSE

   In short, Vermont offers persuasive justifications for an independent invention defense. There are good reasons to think that the defense would solve the problem of excessive litigation by patent trolls, and do so without significant cost to innovation incentives. Nonetheless, I have concerns. These concerns do not completely undermine the defense, but they are significant enough that we should think carefully before making as significant a change as Vermont recommends.

   To begin, the stakes are quite high. While we tend to glorify the individual inventor who makes a significant leap forward, most of the important inventions in U.S. history were made independently by multiple inventors, or at least were built on a solid base of prior work by others. Vermont identifies a number of such simultaneous inventions, including the light bulb, the telephone, the telegraph, the telescope, and the integrated circuit. (12) We might reasonably add to his list the steamboat, which was patented by different inventors in different states; (13) the airplane, which was first patented by the Wrights but independently developed and significantly improved upon by Glenn Curtis and others; (14) the laser, which was the subject of patent applications by two different groups; (15) and polypropylene, which was the subject of a 30-year interference between competing inventors. (16) The fact that so many important advances in technology involved independent invention by multiple parties underscores the significance of Vermont's argument. But it also means that if Vermont is wrong, and an independent invention defense would significantly reduce the incentives to innovate, the potential losses for society are substantial.

   Vermont counters that simultaneous invention is evidence that a right to control not just copiers but independent inventors is excessive. Underlying Vermont's reasoning is the idea that if the prospect of a huge reward tempts many people to invest in research and they all succeed, the prospect of a more modest reward will likely motivate at least one person to make the same investment. And because the fact of independent invention demonstrates that those remaining inventors are still likely to succeed, society will still obtain the invention.

   This reasoning makes considerable sense if the barrier to invention is the uncertainty of research outcomes. Researchers will either think an outcome is certain or that it isn't. If they expect an outcome is certain, they will anticipate independent invention by others, but they will also believe that they can reach the outcome and therefore that engaging in research will be relatively cheap. (17) If, on the other hand, a researcher thinks a particular line of inquiry is a long shot, she will not expect others to independently develop the invention, and so won't worry as much about competition.

   This reasoning breaks down if the barrier to research is not uncertainty but cost. Suppose that potential researchers know that they can develop a new drug for $500 million and that a full right to exclude will offer potential rewards well in excess of the cost. If the researchers face competition from one or more independent inventors, the researchers may not be able to recoup their investment. In that case, an independent invention defense may undermine incentives to develop the invention at all. In theory, this case shouldn't differ from the uncertainty case. If a researcher knows that lots of others will race to a particular outcome, she should discount for her reduced odds of being first. But in practice that is not how invention works. Researchers may not...