Efficient Uncertainty in Patent Interpretation

Washington and Lee Law Review, Aug 2018

Research suggests that widespread uncertainty over the scopes of issued patents creates significant costs for third-party firms and may decrease innovation. This Article addresses the scope uncertainty issue from a theoretical perspective by creating a model of patent claim scope uncertainty. It is often difficult for third parties to determine the legal coverage of issued patents. Scope underdetermination exists when the words of a patent claim are capable of a broad range of plausible scopes ex ante in light of the procedures for interpreting patents. Underdetermination creates uncertainty about claim coverage because a lay interpreter cannot know which interpretation will ultimately be elected and employed by a judge or jury in a patent infringement proceeding. This Article models this uncertainty problem by the set of interpretations that are plausible for a patent-claim element in light of constraints that restrict meaning, internal and external to the patent document. The model suggests generalizable properties against which we can critically evaluate patent interpretive rules and procedures. On this basis, the Article develops an approach to improving the ex ante scope precision of any given patent claim. The general approach is to reduce the set of interpretative scopes that patent claim words can plausibly obtain. By increasing explicit, scope-defining information in the public patent record, it is possible to improve scope precision by ex ante clarifying scope coverage and exclusion in foreseeable scope uncertainty scenarios.

A PDF file should load here. If you do not see its contents the file may be temporarily unavailable at the journal website or you do not have a PDF plug-in installed and enabled in your browser.

Alternatively, you can download the file locally and open with any standalone PDF reader:


Efficient Uncertainty in Patent Interpretation

Efic ient Uncertainty in Patent Interpretation 0 Thi s Article is brought to you for free and open access by the Law School Journals at Washington & Lee University School of Law Scholarly Commons. It has been accepted for inclusion in Washington and Lee Law Review by an authorized administrator of Washington & Lee University School of Law Scholarly Commons. For more information , please contact , USA - Efficient Uncertainty in Patent Interpretation Harry Surden∗ Research suggests that widespread uncertainty over the scopes of issued patents creates significant costs for third-party firms and may decrease innovation. This Article addresses the scope uncertainty issue from a theoretical perspective by creating a model of patent claim scope uncertainty. It is often difficult for third parties to determine the legal coverage of issued patents. Scope underdetermination exists when the words of a patent claim are capable of a broad range of plausible scopes ex ante in light of the procedures for interpreting patents. Underdetermination creates uncertainty about claim coverage because a lay interpreter cannot know which interpretation will ultimately be elected and employed by a judge or jury in a patent infringement proceeding. This Article models this uncertainty problem by the set of interpretations that are plausible for a patent-claim element in light of constraints that restrict meaning, internal and external to the patent document. The model suggests generalizable properties against which we can critically evaluate patent interpretive rules and procedures. On this basis, the Article develops an approach to improving the ex ante scope precision of any given patent claim. The general approach is to reduce the set of interpretative scopes that patent claim words can plausibly obtain. By increasing explicit, scope-defining information in the public patent record, it is possible to improve scope ∗ Associate Professor of Law, University of Colorado Law School; B.A. Cornell University, J.D. Stanford University. Many thanks to the University of Colorado Law School for supporting this work. I am also grateful to Paul Ohm, Deb Cantrell, Sarah Krakoff, Ahmed White, Andrew Schwartz, Phil Weiser, and the rest of my colleagues at the University of Colorado for their helpful comments. Also, many thanks to Seema Shah, Andrew Coan, Joe Miller, Lydia Loren, Tomas Gomez-Arostegui, Andy Johnson-Laird, Kevin Ross, Viva Moffett, Bernard Chao, Danny Sherwinter, Andy Hessick, Carissa Hessick, Michael Sousa, Will Hubbard, and Ann Cammett for their insightful comments. Thanks to Ashley Boothby for her excellent research assistance. precision by ex ante clarifying scope coverage and exclusion in foreseeable scope uncertainty scenarios. I. Introduction ................................................................................1740 The Patent and Trademark Office routinely issues patents of uncertain scope.1 Patents give legal rights over inventions, and patent holders can exclude others from making, selling, or using products2 encompassed by a patent’s claims.3 A patent claim has an uncertain scope if it is difficult to determine which products infringe that claim with any degree of legal certainty. Scope uncertainty is problematic because it can reduce net innovation in particular fields of endeavor.4 Ambiguous patent claims often provide private benefits to the holder but impose external costs upon others.5 Uncertainly scoped patent claims can increase transaction costs for third-party firms that compete in the patent’s space.6 When firms cannot easily determine claim boundaries they 1. It is difficult to quantify the number of patent claims whose scope would be considered highly ex ante uncertain. However, the overall level of ex ante legal uncertainty in patent scope is suggested by the difficulty in obtaining patent infringement insurance. See JAMES BESSEN & MICHAEL J. MEURER, PATENT FAILURE: HOW JUDGES, BUREAUCRATS, AND LAWYERS PUT INNOVATORS AT RISK 51 (2008) ("[U]ncertainty about [a patent’s] scope and validity undermine the market for patent enforcement insurance."); see also Michael Fitzpatrick, Battles of the Tech Intellects, RISK & INS. (Sept. 15, 2010), http://www.riskandinsurance.com/story.jsp?storyId=531905766 (last visited Nov. 18, 2011) ("Because of the severe monetary risks and uncertainty associated with patent infringement, the insurance market for this risk has been slow to develop.") (on file with the Washington and Lee Law Review). 2. To avoid verbosity, I will primarily refer to patent "product" claims rather than the more cumbersome "products or process claims." Patents can and often do claim "processes" or "methods" which are "an act, or a series of acts, performed upon the subject-matter to be transformed and reduced to a different state or thing." Gottschalk v. Benson, 409 U.S. 63, 70 (1972). The principles developed herein for product claims apply to process claims as well. 3. Patent law also allows holders to exclude others from "offering for sale or selling" those products covered by a claim. 35 U.S.C. § 154(a)(1) (2006). 4. See BESSEN & MEURER, supra note 1, at 146 ("[O]ur evidence implies that patents place a drag on innovation."). 5. A patent claim drafted with scope uncertainty makes it more likely that invalid patent claims—or claims with a range of scopes, some of which are invalid—will issue. During prosecution, patent examiners search for prior art to determine if the patent meets the requirements of non-obviousness and novelty. Variability in the scope of claim words complicates the finding invalidating prior art. Examiners operate in a world of limited time and search resources. Multiple, plausible interpretations mean a wider range of potentially applicable prior art. If there are multiple interpretations, examiners are instructed to employ "the broadest reasonable interpretation." See U.S. PATENT & TRADEMARK OFFICE, MPEP § 2111 (8th ed. Rev. 1, July 2010) [hereinafter MPEP]. The tendency is to think about the costs that patents impose on third party may be unable to mitigate legal risk by altering their behavior or "inventing around" known claims to avoid infringement.7 More generally, simply operating in contexts of endemic patent uncertainty can depress innovation and can raise the risks and costs of patent infringement.8 When these thirdparty effects are taken into account, it appears that uncertain patent claims generate substantial costs from a society-wide perspective. These costs are often exacerbated when patent holders opportunistically exploit uncertainty in actual, or threatened, litigation.9 Accordingly, the reduction of some patent scope uncertainty would likely lead to greater overall efficiency. The conventional view from the academic literature is that many problems of patent scope uncertainty are simply inherent and unavoidable.10 For example, patents often issue in rapidly changing technological contexts. Precise claim scope delineation may be difficult when technological progression is unpredictable and unforeseeable technologies arise. Similarly, patent law requires applicants to employ written language to competitors—firms in the same field as the patent holder—which make and sell competing products or processes. But patent rights also cover the end users of potentially infringing products. The cost to this class of third-party users (e.g., consumers or purchasing firms) is often overlooked. This is explored in Part II. 7. The outcome of patent litigation is often unpredictable. That can result in expenses of $5 million or more per side. DAN L. BURK & MARK A. LEMLEY, THE PATENT CRISIS AND HOW THE COURTS CAN SOLVE IT 18 (2009). 8. Carl Shapiro, Patent System Reform: Economic Analysis and Critique, 19 BERKELEY TECH. L.J. 1017, 1018 (2004) ("Although these companies rely on patents to protect their inventions, they feel strongly that defects in the patent system are raising costs, [and] imposing uncertainty . . . Ultimately, the companies claim these defects hinder their ability to compete, to innovate, and to contribute to economic growth."). 9. Some patent holders opportunistically exploit the ex ante uncertainty of claim language by suing (or threatening to sue) through the entire range of a given claim’s scope. Once issued, patent claims receive a legal presumption of validity. It is often quite expensive to invalidate an issued patent claim. Some strategically take advantage of the high cost of invalidation by invoking overly broad, but potentially invalid, interpretations. For pioneering work on the "notice externalities" of uncertain patent claims, see Peter Menell and Michael Meurer, Notice Externalities (unpublished research) (on file with the Washington and Lee Law Review); see also BESSEN & MEURER, supra note 1, at 5 (discussing the E-Data patent infringement case). 10. See, e.g., ROBIN FELDMAN, RETHINKING PATENT LAW 1–10 (forthcoming 2012); William R. Hubbard, Efficient Definition and Communication of Patent Rights: The Importance of Ex Post Delineation, 25 SANTA CLARA COMPUTER & HIGH TECH. L.J. 327, 330–31 (2009) (arguing that increased certainty in the patent application process can lead to inefficiency). delimit their legal rights within patent claims.11 Words have well-known limitations in capturing intangible concepts, and inventive ideas are often quintessentially abstract. Language can be imprecise, ambiguous, openended, or inadequate in describing relevant future conduct.12 Even acknowledging these limitations, this Article argues that it is possible to broadly reduce patent scope uncertainty. This is because many uncertainty issues do not arise for the intractable reasons just described. Rather, numerous problems of patent claim uncertainty result from scope underdetermination. Underdetermination refers to scenarios where a patentee could have supplied information (e.g., in the patent document) that would have clarified the patent’s scope for a significant portion of relevant contexts but did not do so.13 Although there has been some recognition of an underdetermination problem in patent law, without a meaningful theoretical framework, the issue has been difficult to address. To that end, this Article offers a model of uncertainty in patent scope. This model suggests classes of efficient, well-targeted types of information that can reduce the external costs of patent scope uncertainty.14 11. Strictly speaking, the legal scope of a patent emanates from the words of a patent claim, and not the supporting drawings or enabling descriptions that accompany claims. See 35 U.S.C. § 112 (2006) ("The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention."); see also Giles S. Rich, Extent of Protection and Interpretation of Claims—American Perspectives, 21 INT’L REV. INDUS. PROP. & COPYRIGHT L. 497, 499 (1990) ("To coin a phrase, the name of the game is the claim."). Interpretive doctrine, however, allows for the usage of the non-claims portion of the patent document as evidence of meaning. See, e.g., Phillips v. AWH Corp., 415 F.3d 1303, 1312–19 (Fed. Cir. 2008) (en banc) (discussing evidence outside the language of a claim that courts use when construing a claim). 12. See Festo Corp. v. Shoketsu Kinzoku Kogyo Kabushiki Co., 535 U.S. 722, 731 (2002) ("Unfortunately, the nature of language makes it impossible to capture the essence of a thing in a patent application."); Dan L. Burk & Mark A. Lemley, Fence Posts or Sign Posts? Rethinking Patent Claim Construction, 157 U. PA. L. REV. 1743, 1744 (2009) ("Literally every [patent] case involves a fight over the meaning of multiple terms, and not just the complex technical ones."); H.L.A. Hart, Positivism and the Separation of Law and Morals, 71 HARV. L. REV. 593, 607 (1958) ("Human invention and natural processes continually throw up . . . variants on the familiar . . . ."). 13. For an excellent description of the concept of "underdetermination," see Lawrence B. Solum, On the Indeterminacy Crisis: Critiquing Critical Dogma, 54 U. CHI. L. REV. 462, 473–75 (1987) (describing underdetermination as a decision, the results of which are a subset of all results that were imaginable at the time of the decision). 14. For a good high-level approach to improving patent certainty by requiring more information during patent prosecution, see Lee Petherbridge, Positive Examination, 46 IDEA 173, 174 (2006). In Part II of this Article, I examine the evidence that widespread legal uncertainty in the existing patent system creates significant economic and social costs. Scholars usually evaluate the patent system using a utilitarian framework—comparing the net social benefits from exclusive patent rights against their net social costs. Uncertainly scoped patents appear to disproportionately tilt this calculus, increasing net cost. Some researchers have argued that the external costs to third parties of operating in environments of uncertainly scoped patents may be leading to broad reductions in net innovation.15 If correct, this would appear to undermine a primary purpose of patent law, which is to promote net inventive activity.16 Part III explores the theory of patent scope uncertainty more rigorously as a foundation for developing mitigation strategies. In particular, I frame the issue of patent scope underdetermination as a special case of "legal uncertainty" generally by drawing from the jurisprudence literature. This approach frames legal uncertainty in terms of ex ante17 constraints on "legal officials" (e.g., judges, juries, etc.).18 Uncertainty results when such ex ante constraints do not meaningfully limit (and hence render unpredictable) the range of ex post decisional outcomes available to an official.19 15. See BESSEN & MEURER, supra note 1, at 5–10 (discussing the inherent difficulties of equating patents to property, one of which is the inability to define the boundaries of patent rights). 16. See Kewanee Oil Co. v. Bicron Corp., 416 U.S. 470, 480 (1974) ("The [purpose of the] patent laws [is to] promote . . . progress by offering a right of exclusion for a limited period as an incentive to inventors to risk the often enormous costs in terms of time, research, and development."). "The productive effort thereby fostered will have a positive effect on society through the introduction of new products and processes of manufacture into the economy . . . ." Id. Scholars have noted other important purposes of patent law, including incentivizing disclosure and commercialization of technology. 17. The "ex ante" period that I refer to throughout this Article is the period consisting of any time up until a judge (or other legal official) makes an authoritative legal determination as to the meaning of a claim word during a patent infringement lawsuit. Thus, because most patents are not litigated, and even fewer reach the claim construction phase of the lawsuit, the vast majority of patents will remain in the ex ante stage, never to have their claim words authoritatively determined. See, e.g., John R. Allison, Mark A. Lemley & Joshua Walker, Extreme Value or Trolls on Top? The Characteristics of the Most-Litigated Patents, 158 U. PA. L. REV. 1, 4 (2009) (noting that only 1.5% of patents are litigated). 18. I use the term "legal official" broadly, referring not just to judges, but to any legal position that has the capacity to authoritatively resolve an issue of law or fact. This term encompasses juries and administrative officials, in addition to trial and appellate judges. 19. See Duncan Kennedy, A Semiotics of Critique, 22 CARDOZO L. REV. 1147, 1158 (2001) (stating that choices made by social actors are often underdetermined by "available principles" and influenced by "strategic behaviors"); see also H.L.A. HART, THE CONCEPT OF This view maps well onto patent scope uncertainty scenarios. Third parties cannot easily determine the boundaries of a patent claim when it has a high degree of ex ante interpretive variability.20 A claim scope is variable if it has a wide range of plausible interpretations even after applying the internal and external constraints to the patent document that limit coverage. For example, consider a claim with the word "encryption," which can be reasonably interpreted broadly as "a cipher"21 or narrowly as "computerized encoding" read in context of the patent document.22 Such interpretive variability leads to scope uncertainty because a third party cannot know with any degree of legal certainty, ex ante, which of the plausible scopes will be adopted by a judge ex post (during an infringement lawsuit).23 If the application of one plausible interpretation over another would produce a diametrically opposite outcome such that a party would be liable for LAW 132 (2d ed. 1994) (noting the considerable discretion allowed courts in making determinations); Frederick Schauer, Exceptions, 58 U. CHI. L. REV. 871, 873 (1991) (finding that the power to create "exceptions" to defined rules "turns out to be the power both to change the rules and to avoid their constraints"); Joseph William Singer, The Player and the Cards: Nihilism and Legal Theory, 94 YALE L.J. 1, 11–13 (1984) (characterizing the indeterminacy of legal doctrine as desirable and inevitable to an extent). 20. A patent claim that has a high range of ex ante interpretive variability can more easily straddle the issues infringement of patent validity. Ex ante variation in scope range permits a wide range of sometimes shifting meanings during litigation. For some examples of shifting meanings of claim terms during infringement proceedings, see Highmark, Inc. v. Allcare Health Mgmt. Sys., Inc., 732 F. Supp. 2d 653, 664 (N.D. Tex. 2010). ("The Court also cited Allcare’s shifting claim construction as a basis for sanctioning Hill. . . . Hill now explains that it is ‘not unusual’ in patent litigation for a party to adjust its claim construction during the course of litigation."). 21. See Cipher, OXFORD WORLD DICTIONARY ONLINE (2011) ("A disguised way of writing; a code."). Cipher codes predate computers and do not require computer technology to implement. 23. A patent claim drafted with a fair amount of scope uncertainty probably makes it more likely that invalid patent claims—or patent claims with a range of scopes, some of which are invalid—will issue. During prosecution, patent examiners search for prior art to determine if the patent meets the statutory requirements of non-obviousness and novelty. Variability in the scope of patent claim words complicates the process of finding invalidating prior art. Patent examiners operate in a world of limited time and search resources. Multiple plausible interpretations means a wider range of potentially applicable prior art. If there are multiple plausible interpretations, patent examiners are instructed to employ "the broadest reasonable interpretation." See MPEP, supra note 5, § 2111. infringement under interpretation "A" but not under interpretation "B," the claim will be uncertainly scoped relative to that product.24 Part IV develops a formal model of claim scope uncertainty that builds upon the earlier model. This formal model’s goal is to provide a more rigorous structure for understanding the problem of scope underdetermination in order to develop mitigation strategies. This Part frames the salient features of scope uncertainty using the mathematical frameworks of set theory and constraint satisfaction.25 This model highlights a marginal approach to reducing scope uncertainty. We can view scope uncertainty as a relational inquiry from the perspective of a third party seeking to avoid infringement. One can measure the adequacy of claim scope definition by contrasting the claim against a relevant set of products. This set consists of all products that third parties would like to determine, ex ante, whether they do or do not infringe a particular claim. Given that a claim’s scope is underdetermined, the marginal approach queries: What additional information is likely to increase the certainty of such infringement analyses? Requiring patentees to provide clarifying, scope-defining information ex ante can reduce scope uncertainty. The way to increase scope certainty is to create more precise claim boundaries. To make the boundaries of a claim element more precise ex ante, one must prospectively exclude ex post scope coverage. 24. For example, imagine an accused product that employed non-computerized encryption in the product. If "encryption" is interpreted broadly by a judge during claim construction, it may literally cover a non-computerized version of encryption. By contrast, if the term is endowed with a narrower meaning restricted to computers, the accused product may not literally infringe. The interpretation and hence scope may be hard to predict ex ante. 25. This constraint-satisfaction model serves as a useful analogy to the way in which patent claim interpretation operates. The patent document serves to establish the boundaries of the patent holder’s exclusive rights. See 35 U.S.C. § 154(a)(1) (2006) (granting patentees various rights to use, and restrict the use of, the patented invention or process). In interpreting patent claims, judges look to hone in on the meaning of the claim words by consulting evidence from multiple sources. See Vitronics Corp. v. Conceptronic, Inc., 90 F.3d 1576, 1582 (Fed. Cir. 1996) (noting that the court may utilize a number of sources). These indicia of likely meaning come from within the patent document itself (intrinsic evidence), as well as evidence of meaning external to the patent document such as dictionaries (extrinsic evidence). Id. at 1582–83 (discussing the proper weighing of intrinsic and extrinsic evidence). Judges typically use these indications to restrict the possible meanings of claim words to their most reasonable interpretation in light of the sources of evidence. Id. at 1582 (describing clear and unambiguous language as highly relevant). In Part V, guided by the formal model, I suggest policy changes to improve overall claim scope certainty.26 The model largely supports the Federal Circuit’s current patent interpretive framework but proffers important modifications. For example, the model reveals that aspects of the doctrine, such as reliance on an "ordinary meaning in the art" default interpretive standard, often provide illusory constraints on claim scope.27 A more theoretically justified approach requires patentees to provide classes of low-cost but targeted scope-defining information in the patent record. Mindful of efficiency, this Part focuses upon disclosures that are limited yet likely to yield improvements. For example, scope uncertainty problems often arise when different "embodiments"28 included in a patent share common physical design attributes. A common issue of scope ambiguity concerns whether these shared characteristics should be read as implicit limitations on the scope of a given claim during an infringement proceeding.29 The model suggests that the patent system can efficiently reduce this and other predictable uncertainty issues by requiring patentees to explicitly disambiguate foreseeably problematic boundary lines ex ante on the public patent record.30 26. Timothy R. Holbrook, Substantive Versus Process-Based Formalism in Claim Construction, 9 LEWIS & CLARK L. REV 123, 127–33 (2005) (discussing different standards of "formalism" used by courts in their efforts to give patent law more certainty); Kelly Casey Mullally, Patent Hermeneutics: Form and Substance in Claim Construction, 59 FLA. L. REV. 333, 365–66 (2007) (describing problems with formal patent rules in not incorporating relevant contextual information). 29 Consider a patent claim with the word "motor," which only discloses internal combustion motors in the specification, but which is being extended to electric motors. There is a fundamental tension in patent law between the embodiments (or exemplars of technology) that are included in a patent document and the claim scope. The scope of a patent claim is not supposed to be limited simply to the included embodiments. Rather, it is supposed to be expansive beyond whatever particular physical form is disclosed as an example. On the other hand, common features among included embodiments often suggest genuine limitations on the intended scope of a patent claim. 30. The public patent record generally includes the patent document itself as well as the prosecution history—the record of the patent prosecution. Such a targeted approach can be efficient and complementary to other ex post approaches to handling problematic patent claims. See, e.g., BURK & LEMLEY, supra note 7, at 45; Hubbard, supra note 10, at 331. II. The Problem of Uncertainty in Patent Scope Many patents granted by the Patent and Trademark Office have claims of ambiguous legal scope.31 The aggregation of so many uncertain patents creates an environment of ambiguity as to legal coverage, leading to significant social costs. Before presenting a model of scope uncertainty, it is helpful to have an understanding of why uncertainly scoped patents may partially undercut patent law’s economic justification. In this subpart I will survey some costs of uncertainly scoped patent claims. A. The U.S. Patent System Is Justified if It Produces Net Social Benefits The classic justification for the U.S. patent system is utilitarian.32 Under this view, the net benefits of a system of exclusive patent rights must outweigh the net costs from a society-wide perspective.33 Primarily, we grant private patent rights to induce technological inventions that would not otherwise be produced absent such incentives.34 31. See Mark A. Lemley, The Changing Meaning of Patent Claim Terms, 104 MICH. L. REV. 101, 102 (2005) ("Claim construction occurs in every patent case during a ‘Markman hearing.’"). For various reasons, it is difficult to quantify the percentage of indeterminate claims. One admittedly imperfect metric is the claim construction rate during litigation. Nearly every patent infringement lawsuit requires a judge to choose between different, but reasonably plausible, interpretations of patent claims. Id. Because claim construction occurs in every or nearly every case, one reasonable interpretation is that, for that subset of claims litigated, they are of ex ante uncertain scope. Of course, this is only a rough metric of systemic uncertainty. One would presume that there is selection bias because those claims that are litigated, rather than settled, are probably those that are more uncertain. 32. See U.S. CONST. art. I, § 8, cl. 8 (granting Congress the power "[t]o promote the Progress of Science and useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right to their respective Writings and Discoveries"). 33. See Lab. Corp. of Am. Holdings v. Metabolite Labs., Inc., 548 U.S. 124, 126–27 (2006) (Breyer, J., dissenting, joined by Stevens and Souter, JJ.) ("[T]he reason for the exclusion is that sometimes too much patent protection can impede rather than ‘promote the Progress of Science and useful Arts,’ the constitutional objective of patent and copyright protection." (citing U.S. CONST. art. I, § 8, cl. 8)). 34. Beyond the development of inventions, patent rights are also thought to be important in incentivizing the commercialization of inventions, and the disclosure and dissemination of new technological information to the public. See generally WILLIAM LANDES & RICHARD POSNER, THE ECONOMIC STRUCTURE OF INTELLECTUAL PROPERTY LAW 294, 308 (2003). According to the standard economic account, patent rights are designed to overcome the "public goods" problem that leads to the underproduction of inventions.35 We can think of an invention as "information" because the value of an innovation consists largely in the underlying knowledge about how to make and use the underlying technology. Because an invention is information, it has the economic characteristics of a "public good" (i.e., non-rivalry and non-excludability). In the marketplace for physical goods, excludability and rivalry are the properties that allow creators to sell copies of their product to each purchaser willing and able to pay. Ideas, as intangibles, lack these properties. It is thus difficult for inventors to use the traditional market to capture the value of their inventive contributions. Inventors have no practical way of selling their inventive ideas to only those who pay for them while excluding others who do not. Because traditional market mechanisms are inadequate for intangible ideas, the belief is that, absent exclusive patent rights, inventions will be underproduced relative to a socially desirable level.36 Inventive ideas are often expensive for an inventor to develop, but once created, often inexpensive for competitors to copy.37 An inventor must recoup the 35. Patent laws as formulated exist to rectify the "public goods" problem of inventions. This problem is best understood by contrasting physical, tangible goods, and the "informational" nature of inventions. Physical goods are, by their physical nature, rivalrous and excludable. The ability to exclude those who do not pay for physical goods from using those goods allows sellers of those goods the ability to appropriate the value of those physical goods in the marketplace. This ability to sell and profit from goods in the marketplace establishes a sufficient incentive to encourage producers of physical goods to invest in the means of producing those goods, as they have a mechanism to recoup their investment. Inventions are, at their base, non-tangible ideas, and are thus non-rivalrous and non-excludable. Thus, absent some intervention, inventors will not be able to appropriate the full value of their inventions. Without the ability to appropriate the full value of their inventions, inventors will not invest in research and development to create new inventions. 36. The notion of underproduction of inventions refers to two distinct ideas: the ultimate creation of inventions and the pace of inventive activity. The first refers to the idea that some inventions may never be produced at all absent adequate incentives. The second refers to the idea that some inventions may still ultimately be produced absent patent incentives, but that the speed of innovation will be diminished and such technologies will arrive at a socially undesirable slower pace. See LANDES & POSNER, supra note 34, at 302– 07 (discussing rules that reduce the social cost of patent protection). 37. See BURK & LEMLEY, supra note 7, at 80–81 (stating that the ratio of inventor cost to imitator cost is large, for example, in the pharmaceutical industry). Although inventions are often costly to develop because they are essentially information, they are easy to copy by competitors who did not incur the same development costs as the inventor. Id. Competitors will be therefore be able to price the product lower than the inventor because the inventor research costs of developing the invention, but copying competitors will be able to "free ride" off an inventor’s investment without having borne the development costs.38 The concern is that putative inventors will not invest in developing inventions in the first instance if they believe they will not later be able to recoup their investment due to underpricing by copying competitors. Patent rights are a government intervention designed to overcome the problem of underproduction of inventions in the normal competitive marketplace. The legal system artificially endows inventive ideas with legal analogs to the excludability and rivalry properties that naturally allow the value of physical goods to be appropriable in the market. Time-limited, exclusive patent rights are thus designed to incentivize private inventors to invest in, develop, and commercialize technologies. Exclusive patent rights permit supra-competitive marketplace profits for a limited period to permit inventors to recoup research and development costs and appropriate the value of their inventive contribution.39 From a utilitarian perspective, one weighs the net social benefits of exclusive patent rights against their net social costs. All intellectual property regimes have costs, including dead-weight economic losses from monopoly pricing, administrative costs, third-party transaction costs, and increased downstream-development costs.40 Exclusive patent rights are has to price the product at a higher level to recoup its development costs; whereas competitors did not incur such development costs and can afford to charge a lower price. Id. Under conditions of perfect competition, competitors will drive the price of inventions down to their marginal cost. RICHARD A. POSNER, ECONOMIC ANALYSIS OF LAW 38–40 (4th ed. 1992). Inventors will therefore be at a disadvantage because they incur the fixed costs of development in addition to the marginal cost of producing each product; whereas copiers do not incur the fixed development costs. See id. (discussing the economics of intellectual property law). 38. See LANDES & POSNER, supra note 34, at 294–95 (discussing the tension between the costs of keeping trade secrets and the patent disclosure rules). 39. See In re Ciprofloxacin Hydrochloride Antitrust Litig., 363 F. Supp. 2d 514, 540 (E.D.N.Y. 2005) (agreeing that the "exclusion of competitors and charging of supracompetitive prices are at the core of the patentee’s rights, and are legitimate rewards of the patent monopoly." (quoting United States v. Studiengesellschaft Kohle, m.b.H., 670 F.2d 1122, 1128 (D.C. Cir. 1981))). 40. See LANDES & POSNER, supra note 34, at 16–21 (discussing the costs of granting intellectual property rights). For a good discussion of downstream costs imposed by early patents, see Charles W. Adams, Allocating Patent Rights Between Earlier and Later Inventions, 54 ST. LOUIS U. L.J. 55, 56–60 (2009) (explaining the problems that arise when allocating intellectual property rights to an initial inventor and a later improver). primarily justified to the extent they incentivize the introduction of technologies that would not have otherwise been produced. The additional public welfare contributed by these incentivized technologies largely provides the benefits that justify any costs of the system. Recent empirical research suggests that the existing patent system’s costs might outweigh its benefits.41 According to evidence from James Bessen and Michael Meurer, a significant source of these costs stems from the relative indeterminacy of patent coverage.42 If true, the net external costs imposed by uncertain patents may undermine the utilitarian justification for patent law. Since patents of uncertain scope appear to disproportionately shift the cost/benefit calculus, it is worth particularly focusing on the certainty of patent scope as a problem within patent law. B. The Social Costs of Systemic Uncertainty in Patent Scope In this subpart I will explore some of the external costs of patent scope uncertainty. Before doing so, it is helpful to have an intuitive view of what it typically means for a patent claim to be ex ante uncertain in scope. 1. Patent Claim Scope Uncertainty: An Intuitive View A patent gives the holder a time-limited right to exclude all others from making, selling, or using products or processes that fall under the scope of the patent’s claims. The claims, found at the end of a patent document, define the legal scope and boundaries of "the subject matter that 41. See BESSEN & MEURER, supra note 1, at 120–44 (finding that, for most industries in the late 1990s, patents likely provided a net disincentive to innovate due to the costs of patent litigation). Bessen and Meurer provide evidence that for the average firm outside of the pharmaceutical and chemical industries, the private costs of the patent system outweigh the private benefits. Id. This, of course, is a different measure than net social welfare—how the patent system benefits society as a whole. Net social welfare is difficult, if not impossible, to measure empirically. True, their results are not conclusive about net social welfare. Nonetheless, their research provides some strong indirect evidence that the costs of the patent system probably outweigh its benefits. If, as Bessen and Meurer argue, for the average firm, the incentives provided by the patent system are negative, it seems unlikely that the patent system will induce overall increased innovation, thereby increasing net social welfare. 42. See id. at 147–48 (concluding that notice failure and inadvertent infringement are major factors contributing to the surge in patent litigation). the applicant regards as his invention."43 Patent claims are composed of language—by law, a single sentence that describes the technology or invention that forms the basis of the right to exclude.44 Many instances of scope uncertainty involve claims with a high degree of ex ante interpretive variability. The process of patent infringement largely involves comparing the words of the patent claim to the product or process that is accused of infringing that claim.45 The scope of a patent claim—which technologies it covers—will often turn on the interpretation of claim terms.46 A well-scoped claim will have, ex ante, a well-defined, bounded, and definite range of meanings. In an uncertainly scoped patent claim, the range of plausible scopes is often variable.47 In other words, the legally authoritative meanings of most of the words of the claim are not definitively knowable ex ante, but rather, exist in a probabilistic range of possible scopes. Patent claims with a wide range of interpretive variability often create scope uncertainty because it is difficult to know objectively, and ex ante, which is the legally determinative scope. The scope of such a claim may be an uncertain, probabilistic assessment, because during a patent infringement lawsuit, a judge (or jury) might elect any one of the plausible interpretations. Ex ante, third parties cannot tell what is, and what is not, covered by such patent claims with any degree of legal certainty. With one interpretation, a competitor’s product could be held to infringe and with another interpretation, not. Precision of scope therefore refers to the degree 43. 35 U.S.C. § 112 (2006). 44. Id.; see also MPEP, supra note 5, § 608.01(m) ("Each claim begins with a capital letter and ends with a period. Periods may not be used elsewhere in the claims except for abbreviations."). 45. There are two steps to the infringement analysis—literal infringement, and infringement under the doctrine of equivalents. 4 R. CARL MOY, MOY’S WALKER ON PATENTS § 13:1 (4th ed. 2010). For the moment, I focus on the former, and equate the prima facie scope of a patent with literal infringement of the claim. 46. See In re Van Geuns, 988 F.2d 1181, 1184 (Fed. Cir. 1993) (stating that "it is axiomatic that claims define the invention which an applicant believes is patentable"). 47. Thus, when we talk about an uncertainly scoped patent we are not concerned with a broadly but well-defined scope (e.g., clearly "All motor vehicles"). Rather, we are concerned with scenarios where the scope of the patent may be broad, narrow, or somewhere in between—we are just unable to determine which (e.g., unclear if "All motor vehicles," "electric-motor vehicles," "plug-in hybrid cars"). For reasons explained later, this Article is not concerned with broadly scoped patents but uncertainly scoped patents, whether broad or narrow. to which the patent adequately conveys to third parties which technologies are, and which are not, covered by a given patent claim.48 2. External Costs to Third Parties of Uncertainly Scoped Patents The primary problem with uncertainly scoped patents is that they can impose significant costs on third parties. Any issued U.S. patent is potentially valid against all parties within the jurisdiction of the United States.49 In any field of endeavor, a firm must take into account the total risk of patent infringement based upon their activities or products. This risk and associated expenses represent costs that firms internalize into their investment and production decisions. All patents impose some cost on third parties, but uncertainly scoped patents impose disproportionately greater costs. When a patent claim is uncertain because of a wide range of interpretive variability, a third party cannot easily determine its boundaries.50 This increases the information transaction costs for a firm of estimating and assessing infringement liability risk.51 It also raises the cost of proactively reducing risk by reorienting production activities, or attempting to "invent around" a particular claim by creating products that do not infringe. Such legal risk reduction strategies are more difficult because, in many cases, a firm cannot know with any level of certainty what the ex post boundaries of a given claim will be held to be. In many cases where a patent claim exhibits a wide scope variance, it might be impossible or costly to restructure a product outside of its scope. Whether or not firms routinely clear patent rights prospectively, they do often become aware of patents in their field and must make investment and planning decisions accordingly.52 For example, inventing companies 48. Mullally, supra note 26, at 333 ("The claims inform us of the boundaries delineating the subject matter over which the inventor holds an exclusive right."). 49. See 35 U.S.C. § 271 (2006) (discussing patent infringement). 50. See BESSEN & MEURER, supra note 1, at 147 ("[T]he boundaries created by patents are hidden, unclear, or too costly to determine."). 51. See id. at 147–50 (suggesting that patent scope uncertainty increases patent litigation risk, which acts as a disincentive to innovation). 52. See Mark A. Lemley, Ignoring Patents, 2008 MICH. ST. L. REV. 19, 21–22 (explaining how inventors in patent law commonly adopt a wait and see approach for patent infringement rather than conducting research into the patent rights landscape prior to invention). routinely receive "cease and desist" letters that alert them to particular patents, or are actually sued for infringing particular patent claims.53 When the claims at issue are uncertain, risk-mitigation strategies will be reduced because of the inability to fully determine conduct firmly outside of that claim’s scope. Such claims likely also increase the costs of patent litigation. The inability to unambiguously determine coverage contributes to the typical costs of patent-infringement litigation, ranging from $2–$5 million per side.54 Legal uncertainty typically makes settlement less likely and pre- and post-litigation costs higher. These higher litigation costs will cause a firm to redirect money that could have been directed to more socially productive uses.55 Uncertainly scoped patents also increase firms’ costs of confronting potentially invalid patent claims.56 For a patent to issue, it must meet the statutory validity requirements of the patent act, including novelty, nonobviousness, and enablement. A patent claim that has a high range of ex ante interpretive variability can more easily straddle the issue of patent validity. Some plausible interpretations may meet the statutory validity requirements while others may not. Not only does this variable scope complicate the job of the patent examiner in discovering invalidating prior art during prosecution (likely leading to more claims with only probabilistic validity), it requires third-party firms to contend with more such claims that simultaneously have both valid and invalid prima facie interpretive scopes. These external costs become significant from a society-wide perspective when we consider the number of firms potentially affected and the number of patents potentially at issue. Given the size of the United 53. Also, company researchers sometimes come across or passively gain knowledge of existing patents. Finally, in a minority of technological areas, such as in pharmaceuticals, there are well-known patents that are widely known by industry participants, and inventing entities frequently do actually attempt to do prospective searches of patent rights. Id. at 29. Lemley notes that in the pharmaceutical industry, companies are required to list all of the patents covering each FDA approved drug. This allows third parties to survey the patent coverage much more succinctly in that domain. 54. See AMERICAN INTELLECTUAL PROP. LAW ASS’N, 2009 REPORT OF THE ECONOMIC SURVEY (2009) (showing that the cost of patent-infringement litigation ranges from $1 million to $25 million depending on the size of the firm and the stage of the case). 55. Id. 56. See BESSEN & MEURER, supra note 1, at 6 (discussing the E-Data patent infringement case). The original invention covered by the E-data patent was a kiosk that produced digital audio tapes in the retail settings. It was later asserted in patent litigation to cover broad swathes of e-commerce activities. States patent system, even a small amount of systemic patent scope uncertainty can result in dramatic costs. There are approximately two million patents in force at any given time.57 Each patent, in turn, contains approximately ten claims, with each claim bestowing its own separate legal right to exclude technologies.58 Thus, nearly every organization or individual has to incorporate into their planning the probability of infringing roughly twenty million active patent claims. Even if we exclude active but unrelated patents, the number of claims that should be theoretically included in the expected cost valuation numbers is in the still unmanageable tens of thousands in many industries.59 Only a small percentage of issued patents are economically viable and ever litigated or licensed.60 However, even issued patents that have not yet been litigated may increase a firm’s internal expected costs if there is a non-trivial possibility of litigation. Moreover, even if a small percentage of the viable claims are uncertainly scoped, given the large numbers of patents generally, the net costs can be quite significant. In theory at least, a third-party firm making investment decisions factors in the expected cost of patent infringement for some percentage of these patent claims.61 The Supreme Court has noted that even a small 57. See WORLD INTELLECTUAL PROP. ORG., WORLD PATENT REPORT: A STATISTICAL REVIEW 23 (2008) (estimating the number of patents in force as of 2006 at 1.8 million). 58. See Dennis Crouch, Dropping: Average Number of Claims per Patent, PATENTLYO BLOG, http://www.patentlyo.com/patent/2010/01/dropping-average-number-of-claims-perpatent.html (last visited Nov. 18, 2011) (providing a rough estimate of the number of claims per patent) (on file with the Washington and Lee Law Review). According to Crouch’s random sample of 22,000 patents, the average number of claims per patent has increased over the last twenty years, dropping slightly from its peak. Id. The number of claims per patent increased from roughly ten claims per patent in 1980 to over fifteen claims per patent by 2009. Id. 59. BESSEN & MEURER, supra note 1, at 8–9. The authors estimate that for a firm selling products on the Internet, the firm would have to take into consideration over 4,000 potentially applicable patents. Id. With an average of ten claims per patent, that translates into over 40,000 claims whose expected infringement cost might be incorporated by a firm. Firms may be able to filter out the vast majority of active patent claims as they are likely to be outside of any particular field of endeavor. 60. Mark A. Lemley, Rational Ignorance at the Patent Office, 95 NW. U. L. REV. 1495, 1507 (2001) ("As we have seen, only about 1.5% of patents are litigated at all. The total number of patents licensed for royalties without even a complaint being filed is likely somewhat higher, but I suspect the total number of patents litigated or licensed for a royalty (as opposed to a cross-license) is on the order of five percent of issued patents."). 61. See Richard A. Posner, Efficient Responses to Catastrophic Risk, 6 CHI. J. INT’L L., 511, 523 (2006) (presenting the use of a cost-benefit in situations of extreme uncertainty). amount of systemic legal uncertainty in an inventive area can depress development. As the Court said, "‘[T]he limits of a patent must be known’ . . . otherwise, a zone of uncertainty which enterprise and experimentation may enter only at the risk of infringement claims would discourage invention only a little less than unequivocal foreclosure of the field."62 In other words, even without knowledge of particular uncertain patent claims, simply operating in an environment of suspected endemic patent uncertainty can encumber innovation.63 C. Relationship Between Clear Legal Boundaries and Growth While patent uncertainty seems to dampen investment, by contrast, economic research suggests a general relationship between clear legal boundaries and increased economic growth.64 Scholars have suggested that research to this effect from the real property realm can be helpful as a (limited) analogy to the scope uncertainty issues in the intellectual property realm. It is generally thought that in property-rights systems, extensive uncertainty as to scope of legal rights leads to increased transaction costs and decreased investment.65 In the real property domain, much has been written about the importance of public, certain, and easily accessible titles and deeds in facilitating economic development.66 With respect to land, the boundaries of real property are precisely described by the "metes and bounds" of the certainty or uncertainty. From a rule consequentialist perspective we can examine properties of patent procedures in light of their general propensity to marginally increase or decrease legal uncertainty in particular situations. 3. Metrics for Notice-Enhancing Constraints The set theory/constraint satisfaction model is thus useful because it allows us to identify general theoretical properties for constraints in terms of their ability to improve patent scope certainty. On the basis of these properties, we can then evaluate specific patent interpretive procedures. If improving public notice is about narrowing the search space of potential meanings ex post, there are three general theoretical properties that will be desirable in constraints: (1) A constraint must be "set narrowing"; (2) the content of the constraint rule must be ex ante determinable; and (3) the interpretive information-defining criteria of the constraint must be sufficiently precise to classify the universe of accused products as infringing or not. More generally, a patent claim will have improved scope certainty to the extent it is bivalent with respect to unambiguously classifying, ex ante, the set of potential accused products into two sets: those that literally infringe the claim and those that do not. The first desirable property in a constraint is ex ante determinability. If we consider a constraint as a classification-rule then ex ante determinability means that we are able to determine ahead of time the precise content of the rule itself. If the content of a constraint is itself ex ante contestable, then the constraint cannot have a narrowing effect because we do not know which version of a filtering rule will ultimately be applied to narrow meaning. This is the primary problem encountered earlier with implicit constraints inferred contextually from usage in the specification. In our set theory/constraint model we can now understand the problem of implicit constraints. Claim scope certainty is improved when an ex ante constraint definitively removes from plausibility some possible interpretations for a claim word. An implicit constraint is typically not able to perform such an ex ante filtering function. This is because the content and scope of the constraint itself is open to contestation. A lay interpreter cannot rely on such an implicit constraint to definitively narrow the set of possible scope interpretations because it is not clear that a legal official, ex post, will either adopt that same constraint or apply it at the same interpretive scope. The multiple meanings pattern of underdetermination provides another example of an ineffective ex ante constraint. Consider the use of a dictionary definition as a constraint to resolve the meaning of a claim term in litigation. Suppose that there are five respected dictionaries, each with its own set of definitions. For each dictionary, the included definition has multiple meanings for the word. If we were to say that the meaning of a patent claim term should be resolved by reference to "external dictionaries," we will not meaningfully narrow the "solution space" of possible interpretations ex ante. We do not know, as a threshold matter, which dictionary will be used and which numbered meaning within that dictionary will ultimately be selected. Rather, the content of the constraint is probabilistic and is not ex ante determinable with certainty. Contrast an implicit constraint on meaning—the way the patentee uses a word in context in the patent document—with an explicit constraint—a formal, explicit, patentee-provided definition for a word. The content of the latter explicit constraint is ex ante determinable as compared to the former. The content of an explicit definition included by a patentee in the four corners of the patent document has the possibility of being ex ante determinable. At the very least, if the rules require the interpreter to use the patentee-supplied definition, we know, ex ante, the specific version of the interpretive rule that must be applied ex post. The fact that the patentee has supplied an explicit definition in the four corners of the document does not by itself guarantee precision in claim scope. After all, patentees can always supply definitions that are vague or ambiguous on their face. But as I will argue, such a scenario is preferable to the scenario without any explicit definitions. If an explicit, unitary definition is vague or ambiguous, there is at least a possibility that the patent office can require specific additions to make it more clear and precise. When a definition exists implicitly in a probabilistic range of possible background definitions—the set of plausible definitions according to a person of ordinary skill in the art—a patent examiner cannot know with precision the ultimate definition that will be employed, or may overlook such an implicit constraint altogether. In sum, the set theory constraint satisfaction model tells us to aim for a theoretical target of bivalence for claim scope: a well-scoped claim will classify the set of relevant objects as literally infringing or not. Constraints that are explicit and ex ante determinable in form tend to be superior to constraints whose content is probabilistic. Moreover, the model gives us a target for improving scope certainty with a baseline patent claim: Given a patent claim as currently scoped, what additional pieces of scope-defining information are likely to exclude foreseeably relevant claim word interpretations, moving incrementally closer to bivalence? The set-theory/constraint model is clear: Scope certainty is generally improved via explicit constraints that reduce the set of plausible interpretive scopes that patent claim elements can have. For example, consider that a patent often includes implicit limitations that can be arguably inferred from the specification—particularly features that are common across all included embodiments. Scope certainty can be improved by an explicit constraint that definitively indicates whether such a constraint should or should not apply. As I will suggest in the next section, patent examiners can efficiently reduce scope uncertainty by requiring explicit articulation of such boundary lines—effectively turning ambiguous implicit constraints into explicit, determinable, set-narrowing constraints. This will make the claim’s boundary more certain relative to the class of products that seek to invent around the claim. The model also suggests an efficiency limitation—when the marginal benefit of additional interpretive information will not meaningfully narrow the set further. The next Part will explore the policy and efficiency implications of the model. V. Policy Changes A. Implications of the Model on Patent Rules and Practices The model has important policy implications toward improving overall patent scope certainty. The first section of this Part will explain why the model provides some limited justification for the Federal Circuit’s patent interpretive doctrine. The second section includes suggested modifications to the patent system based upon the model. These generally seek to improve claim scope certainty by requiring patentees to include clarifying information in the ex ante public record. The target will be disclosures that disambiguate uncertainty in foreseeable190 scope uncertainty scenarios. 190. There are common, repeated patterns in which patent scope is commonly contested. For example, defendants often seek to impute limitations from the embodiments. Such ambiguity can be resolved ex ante by requiring scope clarifications by the patentee on the record. 1. Theoretical Justification and Patent Interpretation Doctrine The Federal Circuit’s claim interpretation jurisprudence identifies different sources that act to constrain the scope of meaning of patent claim words (e.g., dictionary definitions, plain and ordinary meaning standard, specification). This Part critically evaluates these different constraints in light of ex ante scope definition and this Article’s formal model. Recall that the boundaries of a patent claim are often uncertain ex ante when claim words are capable of multiple, plausible scope interpretations. A third party cannot easily know which scope will ultimately be chosen until it is determined by a judge or jury at litigation. The model represented this scope uncertainty problem by the set of plausible interpretative scopes that any given patent claim word can have ex ante after applying the interpretive rules that restrict the potential meanings of the scopes of words. The model highlighted three general properties of constraints associated with increased scope certainty. An interpretive constraint that provides ex ante notice as to boundaries should: (1) be ex ante determinable (e.g., explicit); (2) provide distinct criteria with enough precision such that a lay interpreter can exclude particular interpretive scopes; and (3) be bivalent—classify the relevant set of accused products as either infringing or not with some degree of legal certainty. Constraints that exhibit these properties are more likely to restrict the set of plausible interpretive scopes, excluding relevant products from the literal scope ex ante and ex post. We can use these properties to normatively evaluate patent law’s various interpretive constraints. The rules for interpreting claim words are largely established by the Federal Circuit’s claim construction (interpretation) doctrine.191 During the process of claim construction, a presiding judge often seeks objective guidance on the appropriate scope to give the contested words of the claim.192 According to doctrine, there are two main categories of sources upon which interpretive decisions are to be made—intrinsic (e.g., evidence internal to the patent document and prosecution history) and extrinsic (e.g., external sources such as dictionaries, expert testimony, and treaties). In 2005, the Federal Circuit in 191. Some portion of the claim terms (primarily non-technical terms) may go directly to the jury for infringement determination without first being construed by the judge. In that instance, the jury implicitly decides what the scope of a claim term is in conducting the infringement analysis. See Menell et al., supra note 80, at 731. 192. As indicated previously, claim construction is the process by which the presiding judge fixes the meaning of a claim term during the course of an infringement lawsuit. Phillips v. AWH Corp.193 instructed district court judges to give "intrinsic" sources of evidence considerably more weight than "extrinsic" sources of evidence in making claim interpretations.194 Did the Phillips court reach the desired outcome in terms of improving claim scope certainty? This Article’s model suggests that the court’s approach is generally sound. However, I will argue that the courts "intrinsic/extrinsic" classification, while useful, is undertheorized. The important question is not whether various sources of interpretive constraints are classified as "intrinsic" or "extrinsic" in the court’s typology, but whether they are consistent with the theoretical properties of ex ante determinability, bivalence, and set-narrowing precision. To understand this distinction, consider the benefits of intrinsic evidence for scope purposes. A primary characteristic distinguishing intrinsic from extrinsic evidence appears to be ex ante determinability. The label "intrinsic" primarily refers to evidence from the "four corners" of the patent document and prosecution history. Extrinsic evidence is characterized by contrast to intrinsic, consisting of "all evidence external to the patent and prosecution history, including expert and inventor testimony, dictionary, and learned treatises."195 An important feature of intrinsic evidence for scope defining purposes is that it is part of the ex ante publically accessible record. To a lay interpreter attempting to discern boundaries, intrinsic evidence is relatively more determinable ex ante than extrinsic evidence. The interpreter knows exactly where to look for intrinsic evidence and what its primary content will be. By contrast, extrinsic evidence is largely undetermined until litigation. Consider reliance upon expert testimony ex post to establish claim scope. The specific testimony of particular expert witnesses cannot be known with precision until established during the course of an actual infringement suit, except at perhaps a very broad level. The court’s prioritization of intrinsic evidence—publically available information—is consistent with the model’s determinability property. Ex 193. Phillips v. AWH Corp., 415 F.3d 1303, 1317–22 (Fed. Cir. 2005) (en banc) (reestablishing the priority of intrinsic evidence over extrinsic evidence, and stating, "extrinsic evidence may be useful to the court, but it is unlikely to result in a reliable interpretation of patent claim scope"). 194. Id. 195. Markman v. Westview Instruments, Inc., 52 F.3d 967, 980 (Fed. Cir. 1995), aff’d, 517 U.S. 370 (1996). ante claim scope boundaries are enhanced when a lay interpreter can know with certainty the form and substance of the restrictions on possible claim meanings that will later be employed ex post. Intrinsic evidence is fixed to some degree in the written record. The problem with extrinsic evidence generally is that it is not typically a static, but a probabilistic, proposition with a wide range of variance, until determined at litigation. Because of this, lay interpreters attempting to discern literal scope boundaries typically have less legal certainty about those boundaries. They will ultimately be determined by reference to such probabilistic, extrinsic evidence. This is further illustrated by a prototypical extrinsic source—a dictionary. If a claim term is not defined, doctrine requires judges to construe it from the "plain and ordinary meaning" of the term to one of ordinary skill in the art.196 For courts seeking objective indicia of contemporaneous technical understandings of a term, one natural source is a technical dictionary.197 In a typical case, there are several such dictionaries that might plausibly be used.198 For each dictionary, the relevant entry may have multiple meanings or scopes for a given word.199 If extrinsic dictionaries are used ex post to resolve the scope of a word, the scope-delimiting rule that will be employed during litigation cannot be firmly established ex ante. A third party cannot know, for instance, which dictionary will be used, which definition in that dictionary will control, or what the ultimate scope the definition will be held to consist of will be. 196. The patent document serves to establish the boundaries of the patent holder’s exclusive rights. 35 U.S.C. § 154 (2006). In interpreting patent claims, judges look to hone in on the meaning of the claim words by consulting evidence from multiple sources. See Vitronics Corp. v. Conceptronic, Inc., 90 F.3d 1576, 1582 (Fed. Cir. 1996) (noting that the court may utilize a number of sources). These indicia of likely meaning come from within the patent document itself (intrinsic evidence) but also include evidence of meaning external to the patent document such as dictionaries (extrinsic evidence). See Vitronics, 90 F.3d at 1582–83 (discussing the proper weighing of intrinsic and extrinsic evidence). Judges typically use these indications to restrict possible meanings of claims to their most reasonable interpretation in light of various sources of evidence. See Vitronics, 90 F.3d at 1582–83 (characterizing unambiguous language as highly relevant). 197. Other sources that might be used to refine that meaning include expert testimony, technical books, or contemporary scholarly or operational literature from the field. 198. See, e.g., L & P Property v. JTM, LLC, 578 F. Supp. 2d 318, 326 (D. Mass. 2008) ("Defendants take issue with L & P’s definition of ‘encryption’ by citing numerous definitions from technical dictionaries."). 199. See Mullally, supra note 26, at 369 ("The unfettered discretion that judges had in choosing among a considerable number of dictionaries without explanation and then selecting a particular definition within a given entry gave rise to a highly uncertain environment for claim construction."). Such an uncertain constraint will often not provide needed certainty to define ex ante boundaries. For this reason, the default "plain and ordinary meaning" standard, resolved via extrinsic sources like dictionaries, often provides illusory scope restrictions. It represents, not one, but a probabilistic set of possible constraints. By contrast, intrinsic evidence typically fares better than extrinsic evidence in terms of scope-defining properties (although it is by no means perfect). This is illustrated by the paragon of scope-defining intrinsic evidence—the patentee—supplied definition. Patent applicants are permitted (but not required) to explicitly define their patent claim terms in the specification.200 Such explicit definitions have the potential to maximize ex ante notice because they will meet at least two desired constraint properties. First, because a definition is embodied in a fixed, reified, textual form in the specification there are not threshold contestability issues about the primary content of the definitional rule to be applied. Contestation is limited to refinements of meaning. Second, if properly linked to a specific claim word, there will not be normalization problems. Moreover, third parties can know if the definition, on its face, contains ambiguities, open-texture, or explicit discretion. The Federal Circuit has also touted intrinsic evidence as more reliable because the patentee created it contextually at the time of drafting.201 This contextual benefit of intrinsic evidence can be seen in terms of increased ex ante determinability relative to typical extrinsic evidence. There is a greater chance that the patentee deliberately intended to employ a constraint of which they were potentially in control during drafting, as compared to a constraint provided by a completely extrinsic source, of which the applicant may not actually have ever known about. 2. Federal Circuit’s Approach: Both Over and Underinclusive The Federal Circuit’s approach, while mostly supported by the model, is both over and underinclusive. There are examples of intrinsic evidence defining scope poorly and of extrinsic evidence providing adequate scope guidance. For this reason, it is helpful to consider constraints in terms of 200. See MPEP, supra note 5, § 2173.05(a). 201. Phillips v. AWH Corp., 415 F.3d 1303, 1312–14 (Fed. Cir. 2005) (en banc). underlying theoretical properties rather than in the Federal Circuit’s intrinsic/extrinsic typology, which fails to guide in several cases. Some extrinsic evidence sufficiently narrows the scope of possible meaning ex ante at the level of precision needed for infringement distinctions. Examples come from research areas with technical words that are inherently well-bounded202 or which have well-defined, widely agreedupon meanings. This is exemplified by certain aspects of the biological and chemical domains. In some instances, references to molecule formations or other technical terms are sufficiently well-bounded in the art so as to make further, intrinsic, refinement unnecessary. In that case, a lay interpreter can actually determine extrinsically defined boundaries ex ante. Extrinsic references, such as technical treatises, may share common meanings. Such examples of scope-defining extrinsic evidence can be ex ante determinable, precise, and bivalent. Similarly, simply because evidence is intrinsic does not mean that it will clarify scope boundaries ex ante. Part III provided examples of implicit constraints that are often inferred from intrinsic evidence. When intrinsic evidence is implicit, there are typically problems of ex ante determinability of scope, content, and applicability of the constraint similar to extrinsic evidence. For example, defendants often infer constraints from the specification based upon limitations shared across embodiments. Such implicit constraints are often ambiguous and contestable and tend not reduce scope uncertainty ex ante. Similarly, even explicit definitions in the specification employ imprecise words leading to similar scope uncertainty issues.203 Moreover, intrinsic constraints may effectively turn into extrinsic constraints if a judge resorts to external sources (e.g., dictionaries) to further define them. Certain intrinsic evidence is endowed with all of the extrinsic limitations just described. This suggests that we should evaluate constraints specifically in terms of actual content, explicitness, and precision. 202. Clarisa Long, Information Costs in Patent and Copyright, 90 VA. L. REV. 465, 476 (2004) ("[A]ttributes that are easy to recognize and define make it easier for observers to cognize the boundaries of the good."). 203. Scope disputes often concern words that are framed at an open-ended or high level of abstraction. It is easy to imagine an explicit patentee-supplied definition in the specification that did not improve scope certainty simply because it employed imprecise criteria. In this respect, the mere addition of a definition in the specification does not necessarily improve precision if the included definition is poorly or broadly defined. The fact that considerable scope underdetermination is common may partially explain the Federal Circuit’s high reversal rate in district court claim-construction decisions.204 In such underdetermined contexts there are simply multiple, plausible interpretive scopes available in light of the existing intrinsic and extrinsic constraints on meaning. Thus a high reversal rate would be expected under de novo review. A reversal may not indicate that a district court necessarily arrived at an objectively wrong (i.e., implausible) interpretation. Rather, it may be that the Federal Circuit has simply elected a different but reasonable interpretive outcome given the latitude of options permitted by intrinsic and extrinsic constraints. B. Changes to the Patent Drafting Process Suggested by the Model In this section, I suggest changes to the patent system indicated by the model. They are largely centered on strategically increasing scope-defining information disclosed by patentees ex ante. The model’s principle of bivalence provides a useful theoretical guide: Does a patent record provide sufficient constraints on claim scope so that a lay interpreter could determine whether relevant products infringe or not? If the answer is no, then there may be foreseeable boundary lines that a patent examiner can require a patentee to clarify during the application process. I will analyze these suggestions in terms of efficiency, noting that they are compatible with ex post approaches to managing patent uncertainty.205 1. Patentees Should Be Required to Be Their Own Lexicographers One theoretically supported approach is to require patentees to include definitions in the specification for all claim words. Currently, patent applicants have the option of including claim word definitions in their specification if they wish to act as "their own lexicographer."206 Since the patentee-supplied definition is the paragon of a scope-defining constraint, a rule requiring patentees to include definitions could increase overall scope certainty for issued patents. Putting net efficiency concerns aside for the moment, consider why definitions are desirable in this regard. From a theory standpoint, a patentee-supplied definition has the potential to maximize scope notice. Because such definitions are fixed in the public record, several points of contestability, and hence, ex post uncertainty, tend not to apply. There will be less uncertainty ex post about whether the definition should be interpreted as a scope restriction and less uncertainty about the general substance of that constraint. In short, such explicit definitions are ex ante determinable. Second, such definitions can potentially provide precise scopedefining criteria. The goal of a well-scoped constraint is to refine the contours of the claim boundaries to more precisely distinguish ex ante literal coverage. According to the set-theoretic model, the approach is to provide incremental, additional explicit clarifications in the record that include or exclude particular scope interpretations (i.e., narrow the set of plausible interpretations). In this way, these explicit, additional declarations interpreted together as a set of constraints that must be met more finely demarcate boundary lines. This allows for more reliable ex ante distinctions relative to potentially infringing products. This principle is illustrated by the earlier example involving the claim term "board." A refined set of constraints might distinguish ex ante the types of boards the claim is meant to cover in terms of its component material. Such an incremental clarifying definition might add that board covers "only wood." Precise claim scope contouring is achieved by adding successive "classification rules" that provide additional criteria likely to classify external products as literally infringing or not. One might imagine that patentees would routinely pick the broadest interpretation to ensure maximum future coverage. But even a broadly scoped declaration of "all materials" is preferential to an ambiguously scoped claim because its broad scope is explicit on the public record. At a minimum, third parties can know that the claim is broad on its face and can more easily identify prior art or enablement issues that limit its validity. In 206. See Vitronics Corp. v. Conceptronic, Inc., 90 F.3d 1576, 1582 (Fed. Cir. 1996) ("Although words in a claim are generally given their ordinary and customary meaning, a patentee may choose to be his own lexicographer and use terms in a manner other than their ordinary meaning."). principle, ex ante scope delineations could be more complex than the example, to more finely approach the actual boundary contours of the patentee’s invention. A variation of this logic is already at work in patent law. Patentees make explicit disavowals of claim scope on the public prosecution history record to avoid invalidating prior art. Such explicit disavowals have the effect of demarcating scope boundaries ex ante (and ex post) with greater precision.207 Ex ante clarifying declarations make contrary ex post scope arguments vastly less plausible. The constraint satisfaction model also addresses problems of "infinite regress"—that arguments simply shift to "definitions defining definitions."208 A well-specified and properly interpreted patentee definition should not suffer from this issue, as all additional scope constraints should be applied additively. This means if we have a patenteeincluded definition, and we are using an external dictionary to further define it, the external definition should be interpreted with an "AND" to the existing definition. All constraints must be interpreted together to successively narrow scope. An additional constraint on meaning, should rarely209 serve to broaden the scope beyond that which is already specified. It should instead provide another classification rule for refining scope. The problem of infinite regress (definitions of definitions) occurs when courts mistakenly interpret additional external definitions in isolation from existing constraints. To the extent a new external definition contradicts or supplants existing constraints on meaning, it is improper. There are a few caveats to the suggestion for patentee-supplied definitions. First, intrinsic definitions themselves do not necessarily 207. During the patent application process, patent applicants have to contend with patent examiner rejections due to prior art which already covers the applied-for invention. Patentees will often amend claims to narrow the scope to avoid the prior art. The doctrine of prosecution history estoppel is a limitation on infringement under the doctrine of equivalents. There is a presumption that patentees cannot cover equivalents that would fall under the scope that they have previously surrendered during prosecution. See Festo Corp. v. Shoketsu Kinzoku Kogyo Kabushiki Co., 535 U.S. 722, 732–33 (2002) (addressing the issues stemming from the uncertainty of a patent claim’s extent); see also Holbrook supra note 143, at 13 ("The prosecution history will act as an estoppel in the literal infringement context just as it will in . . . infringement under the doctrine of equivalents."). 208. See, e.g., Kristen Osenga, Linguistics and Patent Claim Construction, 38 RUTGERS L.J. 6, 69–70 (2006) (suggesting that defining claims using additional words simply generates "meta" debates over the meaning of the words used to define). 209. Constraints that are disjunctive or contain an "or" may not be set-narrowing. Thus, adding a constraint that says "X or Y" might expand the set of possible interpretations. improve scope if not done correctly. As suggested, even explicit definitions in a specification can be ambiguous on their face or provide inadequately precise criteria. Scope definition likely has to be an interactive process, with patent examiners taking an active role in requiring clarifications regarding the scope of supplied definitions and with such clarifications appearing on the public record to aid ex ante interpretation.210 Second, scope uncertainty is legal uncertainty and is ultimately related to interpretive procedures ex post by judges in infringement proceedings or on appeal. Thus, any improvement in clarifying formal boundary lines ex ante, must be accompanied by commitments by judges, ex post, to generally follow those boundary lines when the constraints are on point. Notice requires a policy of ex post forbearance on the part of judges in crafting deviations or shifting uncertainty downstream to the equivalents analysis. Third, there could be exceptions for areas of inventions in which the terms are already extrinsically well-bounded and have a well-understood meaning with a low range of variability (e.g., pharmaceuticals). Finally, there are efficiency concerns when increasing demands on patent applicants and examiners with more stringent requirements. I will explore some of these concerns below and suggest some more limited approaches. 2. Efficiency and Increased Ex Ante Disclosure Requirements It is important to consider the net costs of any mitigating strategies requiring more ex ante information from patentees. When applying for a patent, it is difficult to predict the ultimate commercial value of the protected technology. Some researchers estimate that only 5% of issued patents actually turn out to be valuable enough to be worth contesting after issuance.211 Rules that require increased application disclosures broadly create private costs that are borne by all patent applicants, whether or not the patents turn out to be economically viable. From a society-wide perspective, it might be inefficient to require applicants to expend resources 210. For a good approach to such active examination by patent examiners, see Petherbridge, supra note 14, at 174. I will presently discuss how to strategically limit expenditure of examiner resources. 211. Lemley, supra note 60, at 1507 ("As we have seen, only about 1.5% of patents are litigated at all. The total number of patents licensed for royalties without even a complaint being filed is likely somewhat higher, but I suspect the total number of patents litigated or licensed for a royalty (as opposed to a cross-license) is on the order of five percent of issued patents."). refining the scope of all patents, when only 5% of those patents end up being consequential. The resources expended refining the remaining 95% appear to be a social loss and would not be efficient if that cost outweighs the scope improvement benefits gained for the 5% of consequential patents. If patent rules require too much additional ex ante information, the net costs may outweigh the net benefits from a society-wide perspective. Because of the small percentage of patents ultimately of value, some scholars have advocated ex post solutions focused on active, judicial filtering at litigation. William Hubbard and others have argued that it only makes sense to increase the costs of improving patent certainty if the marginal costs of such ex ante investments are outweighed by the net social benefits.212 I agree with this position. However, as Part II suggests, current cost-benefit analysis is likely underestimating the external costs imposed by even the 95% of patents that remain unused on third-party firms and also on end-users. Firms cannot know which patents will be asserted and, to some degree, must internalize the risk of all patents and not just the 5% that ultimately are asserted in some way. Similarly, a much larger percentage of patents are the subject of threatened litigation, which can have a similar cost-inducing effect.213 Thus, the net external costs of uncertainly scoped patents are likely to be somewhat larger than the percent that are worth asserting, in the cost-benefit analysis. However, Hubbard’s marginal benefit/cost approach is ultimately correct. There is some efficient balance between requiring enough information from patentees to provide reasonable scope notice, and not so much information ex ante so as to broadly waste resources that could be put to more productive uses. Existing patent requirements already cause applicants to incur considerable costs. Application rules currently require a considerable level of detailed information on an application (e.g., specification, claims, statutory requirements). Providing such information is often difficult and costly for a patentee. In the spectrum of "too much ex 212. See Hubbard, supra note 10, at 359–60 ("[D]rafting more explicit patents and patent claims requires additional investment by patentees and may require additional work on the part of patent examiners as well. These additional costs will not be efficient unless they are outweighed by the benefits they produce."). 213. It is a very low-cost strategy for a patent holder to mail threatening "cease and desist" letters and to demand relatively low licensing fees. Simply receiving such a letter is enough to cause most firms to expend resources analyzing and perhaps internalizing the expected risk of infringement. ante information" and "not enough ex ante information," it is difficult to know where the current balance lies relative to some theoretical optimum. However, the evidence in this Article suggests that we can recalibrate the rules incrementally towards the ex ante end of the spectrum to achieve net benefits in improving overall claim–scope, given its external costs. Rules that create more certain claim scope boundary lines also have costs in terms of patent value. As suggested in Part II, there is some public benefit to patent scope uncertainty because it allows judges, ex post, to flexibly adapt claim construction scope to reflect the patentee’s inventive contribution. An interpreting judge can give an uncertain claim a wide scope if he believes that coverage is merited given the accused product and inventive contribution. Similarly, a judge can elect a narrower interpretive scope to ensure that an accused product falls outside of the literal range. If patent scope boundaries are "too" rigid ex ante and ex post interpretations are "too" constrained, the scope of meritorious patent claims may not adequately cover unpredictable after-arising technologies or avoid claim boundaries through insubstantial changes. If rigid ex ante patent requirements create the perception among inventors that the patent system leads to inadequate coverage and value for their claims, inventors may not be incentivized to develop inventions in the first instance.214 There is an inherent tension between ex ante notice of boundary lines and ex post flexibility in claim coverage that is common throughout law.215 On the spectrum of ex ante formality and ex post discretion, this Article suggests a moderate recalibration toward the ex ante spectrum. Patent law has a policy that the infringement coverage of a patent claim should be proportional to the patentee’s inventive contribution.216 Formal and precise 214. See, e.g., Autogiro Co. of Am. v. United States, 384 F.2d 391, 396–97 (1967) The inability of words to achieve precision is none the less extant with patent claims than it is with statutes. The problem is likely more acute with claims. Statutes by definition are the reduction of ideas to print. Since the ability to verbalize is crucial in statutory enactment, legislators develop a facility with words not equally developed in inventors. An invention exists most importantly as a tangible structure or a series of drawings. This conversion of machine to words allows for unintended idea gaps which cannot be satisfactorily filled. 215. See, e.g., Louis Kaplow, Rules Versus Standards: An Economic Analysis, 42 DUKE L.J. 557 (1992) (offering "an economic analysis of the extent to which legal commands should be promulgated as rules or standards"). 216. Indeed this policy is represented by the enablement requirement from 35 U.S.C. § 112. See, e.g., In re Fisher, 427 F.2d 833, 839 (C.C.P.A. 1970) ("[Section 112] requires that the scope of the claims must bear a reasonable correlation to the scope of enablement provided by the specification to persons of ordinary skill in the art."). patent boundary requirements risk that a patent holder’s coverage will be underinclusive relative to its inventive contribution. However, existing patent drafting rules already have a significant amount of formality, and already tolerate the risk of underinclusion. Patent applicants are required to specify their patent claims with a significant level of precision, dividing them into descriptive elements. There are most surely cases where this formality requirement results in ex post underinclusion. Again, it is difficult to know where current patent drafting rules are on the optimal ex ante, ex post spectrum relative to underinclusion of coverage. As indicated, a possibly beneficial artifact of an uncertainly scoped claim is that it permits courts ex post flexibility to correct for underinclusion and undercompensation. However, a more straightforward approach would work better. Significant cases of underinclusion are likely rarer compared to marginal scope disputes. In the presumably rare cases of significant undercompensation due to rigid claim boundary lines, judges should aggressively employ the doctrine of equivalents.217 This has the benefit of achieving the same effect more transparently without the negative externalities of scope uncertainty applying in the larger remainder of the patent pool in which underinclusion is not a problem. In the (comparatively small) instances of ex ante delineation that are unavoidable—as described in Part III—judges should be responsive ex post.218 Patent applicants are also likely to be the least cost avoiders for such scope coverage and uncertainty issues.219 Placing a greater ex ante scopedefining burden on applicants may properly align incentives to reduce net uncertainty costs. Defining patent boundaries precisely ex ante is often costly, difficult, and prone to error.220 In one view, the boundary-defining problem can be reframed in terms of risk: Who bears the ex ante risk of over or underinclusion of patent boundaries, the patentee or the public? Patent holders often benefit from scope uncertainty because they can draft their different claims at successively narrower levels of precision and 217. See John R. Allison & Mark A. Lemley, The (Unnoticed) Demise of the Doctrine of Equivalents, 59 STAN. L. REV. 955, 956–58 (2007) (discussing the "demise" of the doctrine of equivalents). 218. For an approach suggesting stronger ex post involvement of judges in mitigating patent problems, see BURK & LEMLEY, supra note 7. 219. Menell et al., supra note 80, at 748. 220. See Long, supra note 202, at 477. coverage and can often selectively change the asserted scope ex post to dynamically reflect limitations, such as invalidating prior art. Patent law has issues concerning uncertainty that are unique compared to other, superficially similar areas. Patent drafting is a hybrid between public law and contracting. Like contract law, patent drafting allows one to be a "private lawmaker" by allowing private citizens to be in control of drafting legal obligations.221 Unlike contract law (and like public law) the scope of the privately drafted obligations are not limited to the contracting parties, but imposed on the entire public.222 Given that broad patent rights depend upon words chosen by applicant, and that patentees benefit from ambiguity, it is somewhat surprising that applicants are not required to bear this risk. Most drafting attorneys deliberately employ as much ambiguity in patent drafting as will benefit their client and still survive the examination process.223 Uncertainty is often beneficial to patentees because it allows them leeway in ex post coverage to prevent underinclusion in infringement. Indeed, canons of construction dictate that judges interpret claims ex post to preserve validity. Ex ante, patent applicants can assert claim scopes through the entire valid and invalid range of plausible interpretation. Rules that permit uncertain claims to issue simply shift the risks of over and underinclusion away from the patent holder and onto third parties who bear the costs of such uncertain boundaries in internalizing infringement risk. Requiring more precise and rigid ex ante boundary delineation by patentees, and thereby increasing the risk of ex post underinclusion of a claim’s literal coverage, may actually reduce net transaction costs. The patent applicant is usually the party in the best position to know the actual contours and inventive contribution of her invention. Thus, to the extent a patent claim can be precisely defined ex ante to better match the inventive contribution and cover relevant products, the patentee is clearly the best party to do so. Given that patentees may not currently suffer the consequence of uncertainty, increasing ex ante disclosures may reduce the level of deliberate uncertainty and better align incentives to better reflect 221. Of course, this "control" is somewhat limited, given that the patent drafter has to proceed through the patent examination process. It may be more accurate to say that the patent is partially drafted by the applicant (or the applicant’s attorney) and partially by the patent examiner through rejections or allowances that shape its scope. 222. There is admittedly a limited scope of these privately drafted obligations. Unlike lawmakers, patentees cannot broadly legislate. Rather, their privately drafted legislation is strictly confined to delineating the boundaries of their exclusive rights. 223. See generally FABER, supra note 119. actual scope. A similar effect may occur through interpretive "penalty default" rules that incentivize clarification.224 C. Foreseeable Interventions: Strategic Uncertainty Reduction One might be concerned about burdens on patent applicants and examiners if adopting a broad requirement that definitions be supplied for every claim word. This section suggests a series of more limited approaches that might be more feasible. Because they are guided by the model, these targeted policy changes can still produce significant benefits in reducing scope uncertainty with comparatively small costs. 1. Internalization of External Disclosures "Disclosure internalization" is one approach aimed at improving scope certainty while reducing up-front application burden. This method involves referencing particular extrinsic sources within the text of the patent document. This can be thought of as "converting" extrinsic evidence into intrinsic evidence by bringing external sources of meaning partially within the four corners of the document. In the previous section, I detailed the benefits of patentee-supplied definitions in terms of ex ante determinability of scope criteria in the public record. Drafting specific definitions for each patent claim involves significant information costs above and beyond the drafting burdens already required. Internalization occurs when patentees explicitly identify, in the specification, specific intrinsic sources of meaning to be used later in interpretation. Internalization of an extrinsic source can be direct, through copying the text of a source into the specification, or indirect, through explicit reference to a particular external source. There are several variations on this approach. The most cost-intensive change would be to require, for each patent word, a copy of a specific definition from a treatise within the four corners of the patent document. A slightly less cost-intensive approach would be to permit a reference to a 224. See Ian Ayres & Robert Gertner, Filling Gaps in Incomplete Contracts: An Economic Theory of Default Rules, 99 YALE L.J. 87, 96 (1989) (describing penalty default rules which penalize ex ante non-clarification). One version involves construing ambiguities against the drafter. Id. at 104 n.80. specific version of a dictionary or treatise, a definition within that edition, and a specific meaning within the chosen definition. Finally, the least burdensome variation on this theme would be to specify one or more specific editions of dictionaries and treatises that would be used to resolve questions of meanings within the document. Joseph Miller and James Hilsenteger have previously advocated for the increased use of dictionaries by reference in the patent document.225 Others have suggested that the Patent and Trademark Office create an official default technical dictionary with specific definitions for common claim words.226 A patent applicant could explicitly incorporate these standard definitions deliberately by reference. Internalizing extrinsic sources of information can help reduce scope uncertainty at minimal up-front cost. The most tangible benefit of internalizing external sources of meaning is that it potentially reduces the ex ante determinability problems typical of extrinsic sources. For example, if using an extrinsic dictionary ex post to define scope, a lay interpreter cannot predict ahead of time which dictionary or which definition will ultimately be used. Internalization mitigates this problem if we require reference to a particular dictionary/treatise and one specific meaning for a claim word. During prosecution, a patent examiner can reference the external source both to know the intended scope, and to seek clarification if the elected scope is not precise enough. Similarly, lay interpreters can reference specific meanings, thereby reducing, but not necessarily eliminating, one source of uncertainty. Such external references must be readily available to the public for the technique to be useful.227 In this way, extrinsic sources become part of the ex ante determinable public record. Internalizing extrinsic disclosures also addresses another problem with extrinsic sources—lack of contextual integrity within the scope of the patent process. The Phillips court considered extrinsic evidence less reliable because it "is not part of the patent and does not have the 225. See Joseph Scott Miller & James A. Hilsenteger, The Proven Key: Roles and Rules for Dictionaries at the Patent Office and the Courts, 54 AM. U. L. REV. 829, 886–87 (2005) (advocating for the patent office to require inventors to indicate a preferred dictionary in patent applications). 226. Joseph Miller has suggested this very promising idea. 227. Alternatively, the meanings could be copied (or paraphrased) directly into the text of the specification to avoid this problem. There may be copyright issues with extensive copying. specification’s virtue of being created at the time of patent prosecution."228 Internalizing extrinsic sources partially mitigates this problem. The patent applicant has the ability to examine the extrinsically defined meaning and explicitly adopt that meaning by reference in the specification. Finally, internalization of extrinsic sources can help overcome the temporal problem of changing meaning. By reference to specific meanings in specific editions of past dictionaries we are able to conceptually make meaning static to a greater degree by fixing past understandings of words in the interpretive context of the various constraints at the time. Although partial internalization is a compromise approach it could provide a meaningful, marginal improvement in certainty with the least amount of cost tradeoff. 2. Foreseeable Boundary Clarification A slightly different approach focuses upon clarifying foreseeable claim boundary issues in the public record. There are several, common patterns of scope uncertainty that arise in patent litigation. Many of these are predictable and can suggest a strategic use of limited patent examiner resources during examination to efficiently reduce uncertainty issues.229 At a general level, rather than requiring definitions for every word, examiners could require explicit clarification only for particular claim words that are poorly scoped.230 Highly abstract or open-ended criteria are likely to be particularly problematic in this regard. Examiners can quickly identify problematic terms by adopting a theoretical frame based upon the model’s principle of bivalence. Given a claim word "X," is there sufficient interpretive information to distinguish a product with X from one without X? If not, then the patent applicant should likely be required to provide some additional scope-refining disclosure on the record. Examiners can ask two highly relevant scope-defining questions for the record: "What is an example or definition of an X, and why?" And importantly, "What is a 228. Phillips v. AWH Corp., 415 F.3d 1303, 1318 (Fed. Cir. 2005) (en banc). 229. This approach would involve adopting patent examiner incentives and requirements to aim at improving the scope certainty of issued claims. 230. In areas that involve inherently well-scoped terms, or terms with widely agreed upon meanings, such as molecular structures, examiners could use their discretion and assess that ex post extrinsic analysis would still provide sufficient scope benefits. definition of something that is not an X, and why?" It is this latter, excluding aspect of scope-limiting disclosures that bivalence highlights. There are also specific areas that are efficient targets for examiners’ focus. Although claims are usually made up of multiple elements, not all elements are equally important. Usually there is a subset of the elements that capture the invention’s unique contribution over the prior art. The remainder of the elements often consists of structures or combinations well known in the prior art and potentially less likely to be of future controversy. With limited resources, examiners can focus scope-defining resources on those novel elements likely to be the source of controversy. Other predictable scope issues concern embodiments. Defendants often argue that features that are common to all embodiments in the specification should be read as implicit scope limitations. An examiner could easily spot such a limitation by scanning the embodiments and require an applicant to explicitly clarify this scope ambiguity, whether the implicit limitation is intended or not, on the prosecution record. Similar scope issues surround common points of debate: whether the scope is limited to materials that the parts of the embodiments are made out of, the location and positioning of key parts of the embodiments and whether they matter, whether the invention is limited to the inventive context displayed by the embodiments, timing, which claim elements are meant to refer to what parts of the invention, and so on.231 As previously argued, there is justification in requiring explicit ex ante clarification from the applicant. Finally, the adequacy of the scope of a term is relative to the set of potentially accused products. Examiners could compare abstract claim text against relevant competitive products to determine if scope criteria are adequate. These are some examples in which problems of scope underdetermination may be efficiently reduced ex ante. Not all uncertainty issues are foreseeable or tractable, but some may be efficiently reduced with only relatively minimal clarification costs. In this way, patent examiners will surely not remove all scope uncertainty ex ante, but they can efficiently head off particularly foreseeable uncertainty issues and reduce external costs to competitors. 231. Scope defining questions can forestall certain but not all scope issues. Most scope issues can be reduced to "all" or "some" questions. For example, does the term "motor vehicles" apply to all motors or just internal combustion motors? VI. Conclusion It is often difficult to discern, with any degree of certainty, the coverage of patents issued by the Patent and Trademark Office. Given the broad liability possible under patent law, such scope uncertainty can result in significant external costs to firms and users. This Article develops a model of one common source of scope uncertainty—scope underdetermination. Scope underdetermination exists when the words of a patent claim are capable of a broad range of plausible interpretations in light of the patent document, under the interpretive rules. Underdetermination creates ex ante uncertainty about patent scope because a lay interpreter cannot know with certainty which interpretation will ultimately be used in an infringement proceeding. The choice of one interpretive scope over another could be dispositive on the issue of literal infringement. This Article models the underdetermination issue by the set of possible interpretations plausible after restricting the scope based upon interpretive constraints internal and external to the patent document. These constraints limit the range (and set) of possible meanings to varying degrees of certainty. The set-theoretical model suggests an approach to marginally increasing the precision of a claim by adding targeted, clarifying constraints aimed at distinguishing potentially accused products or processes as literally infringing or not. These additional pieces of scope-clarifying information can increase scope certainty to a lay interpreter. The Article concludes with policy modifications suggested by the model. Generally, scope can be improved by requiring more precise scope delineation ex ante from patent applicants. One approach is to require explicit definitions for claim words in the specification. This has the benefit of reducing uncertainty by placing claim boundary lines in the determinable public record. Patent examiners can require that these definitions be precise enough to make literal infringement determinations. A more limited approach enables patent examiners to focus upon requiring patent applicants to clarify foreseeably problematic claim boundary lines. These include disambiguating implicit scope limitations that might be inferred from included embodiments and similar predictable scenarios. II. The Problem of Uncertainty in Patent Scope .............................1747 A. The U.S. Patent System Is Justified if It Produces Net Social Benefits..............................................................1747 B. The Social Costs of Systemic Uncertainty in Patent Scope ........................................................................ 1750 1. Patent Claim Scope Uncertainty: An Intuitive View ... 1750 2. External Costs to Third Parties of Uncertainly Scoped Patents ..............................................................1752 C. Relationship Between Clear Legal Boundaries and Growth..........................................................................1755 D. Opportunistic Behavior from Patentees ..............................1757 E. Some Benefits of Scope Uncertainty...................................1758 F. A Theoretical Model Can Improve Scope Certainty...........1759 III. Patent Scope and Legal Uncertainty ..........................................1761 A. Claim Scope Uncertainty: A More Rigorous View............ 1761 1 . Literal Scope Versus Non-Literal Scope of Patent Claims ................................................................ 1762 2. Scope Is About Boundaries and Function..................... 1764 B. Defining Claim Scope Uncertainty .....................................1766 1 . A Theoretical Metric: Perfect Scope Certainty............ 1766 2 . Characterizing Literal Claim Scope Uncertainty..........1767 C. Legal Uncertainty Through the Lens of a Legal Official ...................................................................... 1767 1 . One View of Legal Uncertainty.................................... 1768 D. Sources of Patent Scope Uncertainty ..................................1772 1 . A Patent Claim Is a Classification Rule........................ 1772 2 . Underdetermination and Scope Certainty.....................1774 E. Underdetermination: Precision Inadequacy .......................1780 F. Underdetermination: Element Divisions ............................1780 G. Scope Uncertainty Issues Difficult to Avoid.......................1781 IV. Legal Uncertainty and Scope-A More Formal Approach ....................................................................................1784 A. Set Theory as a Model of Legal Scope and Uncertainty ................................................................... 1784 1 . Set Theory Overview .................................................... 1785 2. Justifying the Use of Set Theory to Characterize Legal Concepts ............................................................. 1786 3 . Legal Bivalence and Legal Scope.................................1787 B. Formally Defining and Patent Scope................................... 1788 1. Formally Defining Patent Scope and Certainty ............ 1788 2 . Patent Scope Uncertainty Using Set Theory.................1790 C. Formalizing Scope Restrictions........................................... 1791 1 . Constraint Satisfaction and Boundary Delimitation .................................................................. 1791 2. Origins of Constraint Satisfaction as a Framework ................................................................. 1792 3. Constraints on Meaning in Patent Law.........................1794 D. Patent Scope Uncertainty and Marginal Analysis ............... 1796 1 . Principle of Comparative Legal Uncertainty ................ 1798 2 . Marginal Reduction of Legal Uncertainty .................... 1799 3. Metrics for Notice-Enhancing Constraints ...................1801 V. Policy Changes...........................................................................1803 A. Implications of the Model on Patent Rules and Practices........................................................................ 1803 1. Theoretical Justification and Patent Interpretation Doctrine.................................................. 1804 2. Federal Circuit's Approach: Both Over and Underinclusive.............................................................. 1807 B. Changes to the Patent Drafting Process Suggested by the Model........................................................................ 1809 1. Patentees Should Be Required to Be Their Own Lexicographers..................................................... 1809 2. Efficiency and Increased Ex Ante Disclosure Requirements ................................................................1812 C. Foreseeable Interventions: Strategic Uncertainty Reduction ............................................................................ 1817 1. Internalization of External Disclosures......................... 1817 2. Foreseeable Boundary Clarification ............................. 1819 22. For a similar case, see L & P Prop . v. JTM, LLC, 578 F. Supp . 2d 318 , 326 (D. Mass. 2008 ) ("Defendants take issue with L & P's definition of 'encryption' by citing numerous definitions from technical dictionaries.") . 27 . See Bell Atl. Network Servs., Inc. v. Covad Commc'ns Grp ., Inc., 262 F.3d 1258 , 1267 (Fed . Cir. 2001 ) ("As a starting point, we give claim terms their ordinary and accustomed meaning as understood by one of ordinary skill in the art . ") . 28 . An "embodiment" is "a specific, disclosed example of how an inventive concept, INTELLECTUAL PROP. ORG ., INT'L PATENT CLASSIFICATION GUIDE 33 ( 2009 ). 62 . Markman v. Westview Instruments , Inc., 517 U.S. 370 , 390 ( 1996 ) (quoting Gen. Elec. Co. v. Wabash Appliance Corp., 304 U.S. 364 , 369 ( 1938 ) and United Carbon Co . v. Binney & Smith Co ., 317 U.S. 228 , 236 ( 1942 )). 63 . See Michael A. Heller & Rebecca S. Eisenberg , Can Patents Deter Innovation? The Anticommons in Biomedical Research, 280 SCI. 698 , 698 - 99 ( 1998 ) (discussing the invention) . 64 . See Tamar Frankel , The Legal Infrastructure of Markets: The Role of Contract and Property Law , 73 B.U. L. REV. 389 , 395 ( 1993 ) ("The policies of both contract and relationships.") . 65. Id. at 395-97 . 66 . See , e.g., Arthur R. Gaudio , Electronic Real Estate Records: A Model for Action, 24 W. NEW ENG. L. REV. 271 , 289 ( 2002 ) (stating that recording promotes efficiency and economic development) . 204 . See , e.g., Kimberly Moore , Markman Eight Years Later: Is Claim Construction More Predictable ?, 9 LEWIS & CLARK L. REV . 231 ( 2005 ) (finding a claim construction reversal rate of roughly 34.5% from 1996 to 2003 ); David L. Schwartz , Practice Makes Perfect? An Empirical Study of Claim Construction Reversal Rates in Patent Cases, 107 MICH. L. REV . 223 ( 2008 ) (citing reversal rates of between 30% and 40% in various years from 1995 to 2005 ). 205 . See BURK & LEMLEY, supra note 7, at 18; Hubbard, supra note 10, at 327.

This is a preview of a remote PDF: https://scholarlycommons.law.wlu.edu/cgi/viewcontent.cgi?article=3306&context=wlulr

Harry Surden. Efficient Uncertainty in Patent Interpretation, Washington and Lee Law Review, 2018,