Constructing Commons in the Cultural Environment

and innovation, in particular, and information and knowledge production, conservation, and consumption, generally. We propose an initial framework for evaluating and comparing the contours of different commons arrangements. The framework will allow us to develop an inventory of structural similarities and differences among cultural commons in different industries, disciplines, and knowledge domains and shed light on the underlying contextual reasons for such differences. Structured inquiry into a series of case studies will provide a basis from developing theories to explain the emergence, form, and stability of the observed variety of cultural commons and, eventually, to design models to explicate and inform institutional design.

The proposed approach would draw upon case studies from a wide range of disciplines. Among other things, we argue that theoretical approaches to constructed cultural commons should consider rules pertaining to membership criteria, contribution and use of pooled resources, internal licensing conditions, management of external relationships, and institutional forms, along with the degree of collaboration among members, sharing of human capital, degrees of integration among participants, and any specified purpose to the arrangement.

The Maine lobster fishery has been recognized as a successful example of a managed natural resource commons. To ensure an ongoing supply of lobster in the face of threats to the fishery from unregulated overfishing, over a period of years Maine lobster fishermen crafted a set of formal and informal rules to determine "who gets the lobster." By design, the product of their efforts is a commons, a managed-access property regime that allows both lobsters and the lobster industry to flourish. ' This Article confronts the theoretical challenge of understanding the governance of what we refer to as constructed commons in the cultural environment, in which the resources to be produced, conserved, and consumed are not crustaceans but pieces of information: copyrighted works of authorship, patented inventions, and other forms of information and knowledge that may, but need not, be aligned with formal systems of intellectual property (IP) law. The phrase "constructed cultural commons," as we use it, refers to environments for developing and distributing cultural and scientific knowledge through institutions that support pooling and sharing that knowledge in a managed way, much as a natural resource commons refers to the type of managed sharing environment for natural resources that the Maine lobster fishery represents. These environments are designed and managed with limitations tailored to the character of those resources and the communities involved rather than left to evolve via market transactions grounded solely in traditional proprietary rights. Just as research on the Maine lobster fishery is grounded in the case study approach used by Elinor Ostrom and her colleagues, this Article develops and argues for the use of a theoretical framework to systematize case study-based research exploring the construction of the cultural commons. We borrow from Ostrom in developing our framework, but we necessarily adapt, extend, and distinguish her approach to account for important differences between constructed cultural commons and natural resource commons.

Our approach provides a framework to systematize and unify analyses of various constructed cultural commons that until now have been regarded largely as separate species. We do not claim that all cultural commons work in exactly the same way or that they all solve exactly the same problems or that they all produce exactly the same benefits (or costs). Our claim is precisely the opposite: by aligning case studies of related but distinct commons phenomena, over time we will be able to identify those features of commons that are more and less significant to the success and failure of a commons enterprise. Aligning the empirical study of constructed cultural commons via a framework of structured inquiry will facilitate the development of and choice among theories to explain the existence and structure of the commons currently in existence and provide insights into the interplay between IP law and commons arrangements. Some initial illustrations make the point more vivid.

Intellectual property pools. A patent pool is an agreement by two or more patent holders to aggregate and share their patents by crosslicensing. 2 The patents in question typically relate to complementary technologies, where one holder's exercise of patent rights "blocks" a different holder's exercise of related rights. Pooled patents are typically available to all members of the pool and are available to nonmembers on standard licensing terms. 3 A well-known example of an early patent pool in the United States is the Manufacturer's Aircraft Association (MAA), which formed in 1917 and encompassed nearly all American aircraft manufacturers. 4 The Wright Company and Curtiss Company held major patents on aircraft technology, but Wright and Curtiss did not hold all the relevant patents, and for any given manufacturer the cost of licensing a single needed patent from a competitor might have made manufacturing an airplane prohibitively expensive. During World War I, the United States government needed airplanes at reasonable costs and in a short time. As a result, the government facilitated the implementation of the MAA, a private corporation. The MAA entered into an agreement with airplane manufacturers, through which the manufacturers pooled their patents and their potential claims for exploitation of the patents by rivals and 119, 127-28, 134 (Adam B. Jaffe et al. eds., 2000).

[Vol. 95:657 agreed to cross-licensing of the patents to one another on what was, essentially, a royalty-free basis. 5 Largely because of this functioning commons of patented inventions, airplanes were built, and the war was won.

Open source software. The Linux operating system, an alternative to Windows and Mac OS (the Macintosh operating system), was produced and is still maintained by a volunteer collaborative of individual programmers. 6 The Linux collaborative is linked loosely by communications technologies, by members' voluntary allegiance to the project, and by the terms of an open source license. Unlike proprietary computer programs, which are distributed to users in object code or executable format only, open source programs such as Linux are made available in source code form so that members of the community may modify their copies and, under the terms of the governing license, publish their modifications for use by others. Members of the community may also volunteer their modifications for inclusion in the standard Linux code base. Thus, each member of the Linux community may use material in the Linux commons and may contribute material back to the Linux commons. Each individual member of the community contributes code to the accumulated archive of the Linux kernel, which is the core of the operating system. The rules governing the use of open source material and contributions to the open source commons are partly formal and partly informal. Formally, the software is governed by copyright law, and its use is managed by the terms of the General Public License. Informally, the integrity of Linux as an identifiable and stable program depends on a thin hierarchy of informal authority, which extends from Linus Torvalds at the top to the body of individual developers at the bottom. 7 The result is a complete, complex, and successful industrial product that is built and maintained not by a traditional, hierarchical, industrial firm, but by a loose-knit community.

5 See Mfrs. Aircraft Ass 'n, Inc. v. United States, 77 Ct. Cl. 481, 483-87 (1933); Harry T. Dykman, Patent Licensing Within the Manufacturer's Aircraft Association (MAA), 46 J. PAT. OFF. Soc'y 646, 648-50 (1964); Robert P. Merges, Contracting into Liability Rules: Intellectual Property Rights and Collective Rights Organizations, 84 CAL. L. REv. 1293, 1343-46 (1996).

6 Some of these programmers contribute to the Linux project at the behest of their corporate employers and thus are not volunteers in the strict sense. They are volunteers with respect to the Linux project, however. Their relationship to the Linux community is governed by the same rules as apply to those who participate for other reasons. For an exploration of the governance issues involved in the relationship between open source software projects and commercial firms, see Siobhdn O'Mahony & Beth A. Bechky, Boundary Organizations: Enabling Collaboration Among Unexpected Allies, 53 ADMIN. Sci. Q. 422, 440 (2008).

Wikipedia. This free, online encyclopedia is widely read and cited.

It resembles an open source software project in many respects. Volunteer authors create and edit Wikipedia entries; anyone with Internet access can read and use the contents of Wikipedia.foot_0 Wikipedia is not the product of unregulated, potentially chaotic openness, however.foot_1 A governance structure exists among 'Wikipedians" that modulates the openness of the project and operates as a kind of law. 10 For example, not all additions and edits to Wikipedia are automatically added to the site. Moreover, the GNU Free Documentation License, the copyright license that governs the contents of Wikipedia, restricts the use of the contents of the site. 1 ' Wikipedia also has a dispute-resolution system that plays an important role in sustaining the commons. 12 The site is open, but with limits.

The Associated Press. For more than a century, the Associated Press (AP) has been the leading American wire service for newspapers. 13 It offers a compelling example of a constructed cultural commons that is not grounded in formal IP rights. As factual material, the news itself cannot be copyrighted (though there is an important but narrow "hot news misappropriation" tort rule1 4 ). Local newspapers could not afford to cover all of the stories that their readers wanted to read, yet the ease with which news stories can be appropriated served as a disincentive to invest in reporting-a classic free-rider dilemma. The solution was a not-for-profit cooperative, owned by the participant news organizations, which partnered originally with Western Union. t 5 Co- the years stemmed from simple disbelief: 'That can't work.'" CLAY SHIRKY, HERE COMES EVERYBODY. THE POWER OF ORGANIZING WITHOUT ORGANIZATIONS 115 (2008). As David Hoffman and Salil Mehra note, "Wikipedia is 'the canonical bee that flies despite scientists' skepticism that the aerodynamics add up.'" David A. Hoffman & Salil Mehra, Wikitruth Through Wikiorder, 59 EMORY L.J. 151, 155 (2009) 10 See ZITrRAIN, supra note 9, at 143-46 (2008).

See Wikipedia: Text of the GNU Free Documentation License, http://en.wikipedia. org/wiki/Wikipedia:Text_of.theGNUFreeDocumentationLicense (last visited Mar. 9, 2010).

12 See Hoffman & Mehra, supra note 9, at 163-75 (examining the Wiki dispute-resolution system).

See REPORTERS OF THE ASSOCIATED PRESS, BREAKING NEWS: How THE ASSOCIATED PRESS HAS COVERED WAR, PEACE, AND EVERYTHING ELSE 18 (2007).

14 See Int'l News Serv. v. Associated Press, 248 U.S. 215, 239-43 (1918); see alsoWendy J. Gordon, Harmless Use: Gleaning from Fields of Copyrighted Works, 77 FoRDHAM L. REv. 2411, 2421-23 (2009) (discussing the structure of the tort doctrine).

operative members could both upload to the wire service material that they originated locally and download from the wire service material that other members produced. 1 6 Local papers were able to carry AP reports on national and international news that they otherwise could not have afforded to produce. The AP itself was a structured commons managed by its members.

Jamband fan communities. Musical groups known as jambands 'jam," or improvise heavily, during live performances. Beginning with fans of the first and best-known jamband, the Grateful Dead, jamband fan communities have long been encouraged by the artists themselves to produce and share their own concert recordings. These recordings initially were shared via physical media and now are shared using online archives (organized via the website and organization etree.org) .17

The bands encourage this sharing, provided that the fans comply with informal rules that are set by the bands and honored and policed by the fan communities themselves. 8 For example, as Mark Schultz describes in his detailed case study of the jamband phenomenon, fan communities generally undertake not to interfere with commercial exploitation of the bands' own concert recordings. 19 The commons of jamband recordings is structured not merely by fan expectations that norms will be honored but also by file sharing and archiving technologies that reinforce the commercial/noncommercial distinction, by intermediary institutions that hostjamband archives, and by the bands, which cooperate with and nurture their fan communities. 20

For histories of the AP, see REPORTERS OF THE ASSOCIATED PRESS, supra note 13; Shmanske, supra note 15; The Associated Press, AP History/Corporate Archives, http:// www.ap.org/pages/about/history/history.html (last visited Mar. 11, 2010). The thumbnail sketch presented above necessarily omits important details about the history of the AP, including allegations that its content was biased politically and that it behaved monopolistically, as well as information about its current, troubled state. On its history, see L.J. 651, 674, 679 n.116 (2006); see also etree.org, http://www.etree.org (last visited Mar. 18, 2010).

These examples are illustrative of constructed cultural commons phenomena studied in isolation from each other. 21 At first glance, these examples may appear to be disparate and unrelated-like comparing apples to oranges to plums to pears, and so forth. Standing alone, each example is important and worth studying. Nonetheless, a systematic, comprehensive, and theoretically informed research framework offers significant potential for learning within and across these commons phenomena. 22 The framework described in this Article provides a means to investigate the social role and significance of cultural commons institutions. This investigation is relevant to property law, in particular, and social ordering, more generally. The conventional view of property scholars, particularly those with interests in IP law, is that resource production and consumption are-and ought to be-characterized primarily by entitlements to individual resource units, held individually and allocated via market mechanisms. 23 To the extent that those market mechanisms are inadequate to optimize the welfare of society-in other words, in the event of market failure-government intervention may be appropriate. Intellectual property rights themselves are generally justified on precisely this basis. 24 Creative works and new inventions are characterized as public goods, whose intangibility prevents their originators from excluding potential users and thus recouping their investments via pricing. 25 Copyright and patent laws create artificial but legally sanctioned forms of exclusion, restoring a measure of market control to creators and innovators. erally regarded as exceptional and dependent upon preexisting property entitlements.

The framework for collecting and analyzing case studies of constructed cultural commons across a wide range of domains that we describe below offers a method for assessing the validity of this property-focused narrative. We suspect that over time the constructed cultural commons framework will yield a far larger and richer set of commons cases in the cultural context than one might discover by focusing only on patent law or scientific research or software development. We anticipate that social ordering both depends on and generates a wide variety of formal and informal institutional arrangements, and that the logical and normative priority assigned to proprietary rights and government intervention may turn out to be misplaced.

Part I of this Article describes prior approaches to theoretical understanding of intellectual production based on juxtaposition of IP regimes with a conception of the public domain. It highlights how recent case studies demonstrate the need for a more textured theoretical approach. Part II introduces the Institutional Analysis and Development framework pioneered by Elinor Ostrom in the natural resources context and presents our proposed adaptation of the framework to constructed cultural commons.

This Part reviews our motivation to produce the constructed cultural commons framework and highlights the key theoretical regimes to which the framework connects. We explain why a research framework is needed in this domain and justify the type of framework and case study-driven approach we develop in Part II. We begin with a discussion of IP law and the theoretical problems and solutions that typically characterize IP regimes. In particular, we explain the conventional functionalist account of intellectual property, highlight some of its flaws, and suggest that a more nuanced functional account is needed to better understand the cultural environment. Next, we explain why the functionalist approach of Ostrom and her colleagues, developed in the context of the natural environment, provides a useful starting point but needs adjustment to account for differences between the natural environment and the cultural environment. Finally, we suggest that narrative and metaphorical accounts of property institutions and social ordering might complement a functional approach and enrich the constructed-cultural-commons framework we develop in Part II.

Intellectual property law scholarship has typically viewed invention, creative expression, innovation, and related or subsidiary activities (such as research and development) as a special set of practices for which extra encouragement is warranted. Despite considerable variation and nuance, these activities all can be understood to present a simple core problem: as public goods, the "output" from these activities-whether described as information, expression, invention, innovation, research, ideas, or otherwise-is naturally nonrivalrous, meaning that consumption of the resource does not deplete the amount available to other users, and nonexcludable, meaning that knowledge resources are not naturally defined by boundaries that permit exclusion of users. 27 As a result, such resources face the well-known supply-side problem, common to public goods: the inability to (cheaply) exclude competitors and nonpaying consumers (free riders) presents a risk to investors perceived ex ante (prior to production of the good), and this risk may lead to undersupply. Essentially, in the absence of [some institutional solution], there would be a significant underinvestment in (some types of) [intellectual resources] because of the risk that competitors would appropriate the value of the [resources].28 Two standard solutions to this problem are IP rights and government subsidies. These solutions heavily influence the framing and perception of the cultural environment.

As a result, the conventional functionalist approach to information law and policy divides the information environment into two domains. First, there is the domain of exclusion, in which producers of 27 On the public-good nature of intellectual resources (ideas, information, expression, and knowledge), see Brett M. Frischmann & Mark A. Lemley, Spillovers, 107 COLUM. L. REv. 257, 272-73 (2007) CLUB GoODs 40-43 (2d ed. 1996) (describing externalities as the absence of markets and the effect of externalities on property rights).

creative and innovative things employ proprietary rights sanctioned by law to control their development, distribution, and exploitation. Private rights and private market exchange serve to limit, by law, the natural shareability of knowledge and innovation. At the core of IP law, as traditionally conceived, is the right to exclude, without which it is assumed that some producers would abandon their efforts for fear of free riding (unlicensed sharing) by competitors. 29 Without exclusion, competition facilitated by sharing would undermine incentives to invest in the production, development and/or dissemination of some resources in the first place. Intellectual property law constructs and assigns these exclusive rights and encourages their exploitation through market exchange.

Second, there is the public domain, a vast collection of openly accessible resources for which exclusion is disallowed under existing intellectual property systems (for example, due to express exclusion from the system 30 or expiration of rights 31 ) or not practiced (for example, because potential owners dedicate their resources to the public or because exclusion is practically impossible).32 For some types of resources, overconsumption and underinvestment problems associated with unlicensed sharing are solved through direct or indirect provisioning by the public sector using a combination of grants to researchers; tax credits or subsidies to researchers and enterprises that employ them; prizes; and production and distribution of knowledge and innovation by the government itself, either by organizing research enterprises or by purchasing and distributing private research. These alternative solutions often, though not always, supply resources to the public domain.

The conventional functionalist paradigm is woefully inadequate as a descriptive matter. 34 Essentially, it is a caricature-an oversimplified account that roughly describes some cultural practices and productive activities but leaves much too much out of frame, unexamined, and unaccounted for. (Consider, for a moment, how many creative and inventive activities involve much more than "public goods production" or how often participation in such activities is outside of and inexplicable in terms of IP incentives or government subsidies.) We further explain this view below, but first we must emphasize how much rests on the descriptive account: it shapes-if not determines-the set of baseline premises that undergird the legal and social institutions that structure the cultural environment and shape normative outcomes. Although our undertaking is emphatically positive and descriptive, we believe significant normative implications should flow from a better understanding of the cultural environment.

Over the last decade, scholars increasingly have recognized that many of the most interesting and important aspects of the information environment exist in the area between these private and public extremes precisely because of what Brett Frischmann and Mark Lemley characterize as "spillovers": social benefits that flow from uses and reuses of information resources and sustain the dynamic character of the information environment. 36 The information environment is riddled with so-called "semicommons" 37 -complex combinations of private rights and commons, some of which are constructed at the "macro, system level" of law, 38 and some of which are constructed at the "micro, contextual level" of cultural commons.

At the macro level, the rights of exclusion that comprise the default regimes of patent and copyright law are, by design, not absolute. Intellectual property regimes moderate their exclusionary principles with limitations and exceptions. In part, those limitations and exceptions are designed to construct a public domain of resources that are 131 (2000). Henry Smith gives the example of a highway, which is a commons in that its most significant aspect is its openness to all users-yet the individual driver has private rights with respect to the moving portion that his vehicle occupies. Id. at 132.

Frischmann and Lemley explain how copyright and patent law are semicommons. See Frischmann & Lemley, supra note 27, at 282, 284-85, 290-91. open and freely available to all, 3 9 and in part, they support public capabilities to use resources in ways that generate spillovers. 40 In addition to these macro-constructions, intellectual property rules also are used at a micro-level, in conjunction with contracts and social norms, to construct a wide variety of semicommons or limited commons of cultural resources that are partly open and partly closed, usable by members and sometimes by the public at large, though not always on a purely "free" basis. Examples include such diverse institutions as public lending libraries and open source software programs. Default rules of intellectual property thus may be combined with licenses and contracts, with social norms, and with cultural and other institutional forms to construct these cultural commons, which depend on-but are built alongside and on top of-the basic forms of knowledge and culture, on the one hand, and intellectual property rules, on the other hand.

Our focus on constructed cultural commons emerges from the proposition that cultural production is an inherently social phenomenon, taking place over a wide range of scales and within a complex, overlapping variety of formal and informal institutional structures. 4 1 Indeed, social production of cultural goods has become more salient and more economically important as a result of globalization and the communications revolution symbolized by the Internet. 42 We are thus beginning to grapple with the realization that legal facilitation of innovation and creative production is not-and cannot-be confined to a simple set of property rules to incentivize individual innovative and creative efforts. Innovation and creativity are matters of governance of a highly social cultural environment. The question for both public policy and legal theory becomes how best to use legal and other tools 39 On the public domain, see generally Samuelson, supra note 32 (exploring various conceptions of the public domain). J. 391, 409 (2005) ("Fair use is fair, after all, because (we assume) that it generates social benefits that the market can't otherwise produce.").

Cultural works and information goods have always been socially constructed in many senses, meaning that they do not arise "naturally" or inevitably but, instead, are the products of human actions and interactions with each other, social institutions and systems, and the physical environment. The traditional, economically inspired realm of production around which intellectual property protection is designed cannot be treated as independent of the larger cultural environment within which it is situated. On the legal construction of cultural things, see MichaelJ. Madison, Law as Design: Objects, Concepts, and Digital Things, 56 CASE W. REs. L. REv. 381, 447-63 (2005). to encourage the growth and persistence of creative, sustainable, and equitable cultural environments. 4 3 Much of the scholarly debate in IP law has pitted proponents of privatization as a means of incentivizing production of intellectual goods against proponents of a widely available public domain upon which cultural goods can be built. The discussion has often devolved into a disagreement over the relative importance of incentives and access for production of ideas and creative expression. 44 As technology has evolved to facilitate an increasingly extensive, varied landscape of social and cooperative projects that enable creativity and innovation, a third perspective has emerged. Books, articles, and scholarly discussion of such projects, of which open source software has become the poster child, increasingly extol community production as a solution to the free-rider problems of cultural production. 45 The danger is that the amorphous idea of "community production" will become the new one-size-fits-all panacea approach in rivalry with privatization, public subsidy, and the public domain.

Fortunately, ours is not the first scholarly enterprise to confront the realization that common-property regimes are more complex and various than had been appreciated. A group of scholars of commons regimes in the natural environment, spearheaded by Ostrom and her collaborators, has eschewed a simplistic approach and developed a framework for synthesizing studies of various and diverse natural resource commons. We argue that now is the time to recognize that the lessons learned by these scholars of the natural environment caution against an overly simplistic view of community cultural production. 46 The primary lesson of their years of case studies of commons regimes, which culminated in Ostrom receiving the 2009 Nobel Prize in Economics, 4 7 is that the devil is in the details. Complex environments 43 We use the term "cultural environment" advisedly, following the work of James Boyle. SeeJames Boyle, A Politics of Intellectual Property: Environmentalism for the Net?, 47 DuKE L.J. 87, 108-12 (1997). The environmental metaphor should not lead scholars to rely uncritically on the assumption that there is anything "natural" about the cultural environment. As we discuss, one of the most important differences between natural resource commons and cultural commons is the degree to which cultural resources are manufactured, both by humans and by law. Natural resources, typically, are given. Yet we also caution against going too far in the other direction. Although the natural environment is given and not made by humans, it is continuously and unavoidably affected by humans and, in a sense, made and remade and unmade with irreversible consequences through those interactions. And although the cultural environment is always made by humans, it is also inherited, subject to considerable path dependencies that can have irreversible consequences, and contingent on human interactions with the physical environment. demand a more contextual empirical and theoretical approach. 48 The complexity of natural and cultural environments and the futility of analyzing them with one-size-fits-all theories or legal approaches create the challenge of developing appropriate conceptual understanding and policy prescriptions. Here is where and why we think that an adaptation of Ostrom's approach to the systematized study of natural resource commons may be particularly helpful for intellectual property scholarship.

As we describe in detail in Part II.A, Ostrom's approach represents a starting point, but one that requires significant modifications. Those modifications, discussed in Part II.B, are woven into the baskets of questions we propose to use to interrogate constructed cultural commons. Those baskets of questions reflect an intentional reliance on two approaches to intellectual property that are sometimes thought to conflict: a functionalist account and a metaphorical or narrative approach.

The functionalist aspect of our proposed framework mirrors the functionalist approach that Ostrom and her colleagues take with respect to regimes governing the sharing and exploitation of natural resources. Despite the many virtues of functionalist accounts, their focus on a handful of variables or factors often limits them. A focus on excludability as a solution to public goods problems can lead to isolated analysis of boundary problems.

In a recent article, Frischmann argued that a myopic focus on excludability leads to biased analyses:

[P]ossible free riding drives analysts to focus on supply-side considerations, and more specifically, to correct market-driven supply problems by designing property-based institutions to lessen the costs of exclusion and minimize free riding. As I have argued elsewhere, nonexcludability is not a necessary condition for market failure; markets may fail for many other reasons. Nor, however, does exclusion fix all market failures. In fact, exclusion may aggravate other failures of the market. Even if an owner can exclude users from a nonrival resource and therefore meter use by charging a fee, dynamic inefficiencies still may abound. Simply put, property rights and other institutions that lessen the costs of exclusion and facilitate market-driven provision of nonrival goods are no panacea. Brett M. Frischmann, An Economic Theory of Infrastructure and Commons Management, 89 MINN. L. REV. 917, 948-49 (2005) (footnotes omitted).

formation-forcing and limiting mechanisms. Such approaches can effectively highlight strengths and weaknesses of policy prescriptions in their respective domains, but these approaches frequently fail to deal adequately with systemic issues or with dynamic changes. Ostrom's approach to natural resource commons addresses the limitations of functionalism by expanding the number of variables considered.

In the context of the cultural environment, the sharing/exclusion and cooperation/competition dichotomies present especially interesting and challenging puzzles to a functionalist approach. This is so for three reasons.

First, those who create, invent, innovate, and participate in similar intellectually driven, productive activities necessarily borrow from or share with others. 50 It is impossible to divest oneself of that to which one has been exposed. Inevitably, the intellectual products of past and contemporary "producers" (a term that we use as a shorthand to refer to creators, inventors, innovators, thinkers, and the like) serve as inputs into each of our own productive activities. We necessarily borrow and share. Second, as discussed above, the resources that shape the cultural environment are by their nature naturally nonrivalrous and nonexcludable, meaning that knowledge resources are not naturally defined by boundaries that permit exclusion of users. 5 1 Third, unlike resources in the natural world, resources of information and expression must be created before they can be shared. Because of the public goods character of these resources, a cultural commons must manage both use and production of cultural resources.

These features of cultural resources, practices, and environment add to the range of variables that must be considered when studying commons institutions. Moreover, although natural resource management often focuses on sustainability, in the cultural commons the goal generally goes beyond sustainability to innovation, growth, and progress. 52 This shift in focus means that in assessing any particular constructed-cultural-commons arrangement, we must expand the 50 Frischmann & Lemley, supra note 27, at 258-61. 51 On the point that these resources have no natural boundaries and that, as a result, boundaries are necessarily constructed, see Frischmann & Lemley, supra note 27, at 274; Madison, supra note 41, at 401-04.

52 Progress in the cultural environment should not be equated with innovation or economic growth. Progress may encompass much more. See Margaret Chon, Postmodern "Progress": Reconsidering the Copyright and Patent Power, 43 DEPAUL L. REv. 97, 101-03 (1993) (arguing that the definition of "progress" is socially constructed and should be shaped by social values and human priorities); Frischmann, supra note 35, at 1095-96 (arguing that 'we know too little about that which we seek to promote" with intellectual property, that the functionalist approach to intellectual property follows from an impoverished view of the cultural environment, and that "[w]e might, for example, imagine Progress as measured by the degree of participation in creative and inventive activities").

[Vol. 95:657 framework used in studying natural resources to consider not only how resources are managed and shared within the community but also how and if resources are created within and transferred outside of the community. Dynamic effects are central to this analysis. The constructed-cultural-commons framework we propose in Part II reflects these functionalist distinctions between Ostrom's aipproach to natural resource commons and our modified approach to intellectual-resource institutions.

A second and equally important way to address the limitations of functionalist theory in the cultural commons context is to examine a system of social ordering in expressive terms rather than purely functional terms, looking to the construction and evolution of meaning in the system, as reflected in symbol and narrative. In proper proportion, a humanistic and metaphorical inquiry into information policy, on one hand, and a functional approach grounded in social science models, on the other hand, are complementary and can be effectively unified in a framework of research questions to yield accurate descriptive summaries of commons phenomena and eventual policy prescriptions.

We draw on narrative and metaphorical approaches to legal and sociological questions, specifically by examining the metaphorical dimensions of the information "environment" and the knowledge "commons." 53 The environmental metaphor for information law and 53 Scholars of many stripes have focused increased attention over the last decade on the role of language and metaphor in structuring analysis of legal and policy problems, both in connection with IP law and otherwise. For a recent review of the Law and Literature movement, see Bernadette Meyler, The Myth of Law and Literature, 8 LEGAL ETHICS 318 (2005) (reviewing THANE ROSENBAUM, THE MYrH OF MORAL JUSTICE: WHY OUR LEGAL SYS-TEM FAILS TO Do WHAT'S RIGHT ( 2004)) (suggesting that Law and Literature should focus on the role of literature in understanding the institutions of law). Other scholars have emphasized connections between language and metaphor, on the one hand, and cognitive processes that drive behavior and experience, on the other hand. See GEORGE LAKOFF & MARKJOHNSON, METAPHORS WE LIVE By 3-6 (1980). James Boyle focuses on the rhetoric of authorship and invention metaphors in order to expose the political character of property law. See Boyle, supra note 43, at 87-91. Carol Rose focuses on narratives of property law to demonstrate the essentially social character of the law. See CAROL M. ROSE, PROPERTY AND PERSUASION: ESSAYS ON THE HISTORY, THEORY, AND RHETORIC OF OWNERSHIP 35-42 (1994). A number of intellectual property scholars draw on environmental and spatial metaphors in their discussions of information law and policy. See, e.g., Frischmann, supra note 35, at 1084-85, 1088-97 (reviewing the use of cultural environmentalism as a metaphor in information policy discussions and arguing that, to be analytically useful, "we need to know more about what the cultural environment is, how it works, how it is constructed, how we interact with it, how we change it, how we are a part of it, and how it relates and interacts with other environments"); David Lange, Recognizing the Public Domain, 44 LAW & CoNrEMP. PROBs. 147, 158-59 (1981) policy-focusing on cultural and knowledge resources rather than physical or natural resources-offers an especially illuminating and useful starting point for our project. Relying on this metaphor offers the ability to explore connections within and between interconnected and interdependent natural and constructed resource systems; to differentiate growth and progress from stewardship, conservation, and sustainability; to describe the differences between natural and constructed environments and the differences between open and closed or "gated" or "managed" environments; to describe different versions of concepts based on adjacent metaphors, such as the public domain and the commons; to identify and describe important patterned behaviors that correspond to different kinds of environments; and to draw lessons from a variety of regulatory and governance structures in other environmental contexts (public and private, legislative and administrative, oriented toward individual entitlements and collectivist, and so forth).

The environmental metaphor also illuminates the importance of the nesting process that Ostrom identifies as a salient feature of commons. 54 By "nesting," we mean that a particular commons phenomenon might be analyzed at many levels; these levels may interact strongly with one another. One of the issues that must be resolved in any particular inquiry is the appropriate level of complexity at which a particular commons should be studied. Ostrom analogizes this nested analysis to a set of maps at different levels of detail, 5 5 such as one sees when using the zoom function in Google Maps. 5 6 All of these maps are accurate, but the usefulness of a particular map depends on the question one seeks to answer. Moreover, some questions can be answered by focusing only at street level, while others may require zooming back and forth to different levels. Similarly, analyzing a commons institution may require more or less detailed knowledge of the larger cultural institutions within which it resides.

This Article begins to explore how this nesting process can and should be examined in the cultural context. Our combined functionalist and expressive framework is deliberately intended for application to the cultural commons in environments that are structured not only by IP law but also by other legal rules, such as the rules of contract and license, and by informal cultural institutions and social practices. A structured inquiry is needed to make progress in understanding the complex and diverse commons arrangements that may be constructed in the cultural environment. This Part sets out our proposed framework for analyzing constructed cultural commons using a case study approach.

A. Learning and Building from Ostrom's Institutional Analysis and Development Framework

As discussed above, we base our proposed framework on the work of the political scientist Ostrom and her colleagues, who have for decades been studying commons in the natural resource environment. 57 Examples include not only the lobster fishery with which we began the Article but also numerous instances of grazing pastures, forests, and irrigation systems. 58 In each case, a similar underlying problem was diagnosed: what property scholars, following Garrett Hardin, refer to as a "tragedy of the commons." 5 9 Given a pool of physical resources that is presumptively open to all comers, such as a meadow for grazing sheep, and absent a mechanism for coordinating the actions of resource users, that is, the owners of the sheep, resources are likely to be overconsumed and underproduced. 60 Eventually, the pool will collapse under the weight of individuals acting rationally in pursuit of their own self-interest, without regard for the costs imposed on other existing and future resource users. The tragedy of the commons illustrates a standard externality problem that manifests a failure of collective action. allocation of property rights via markets is government intervention and regulation. 63 A key insight of Ostrom's approach to the natural environment was recognition of the important role for institutions intermediate between private property and the state in solving problems of collective action. 64 These intermediate institutions are sometimes called "common property" or "limited commons" and generally are collective (but not necessarily governmental or even formal) means for sharing and making productive and sustainable use of resources such as fish, water, trees, and so forth. 65 The research of Ostrom and other scholars demonstrates that solutions to these resource-sharing problems are various and highly contextual. Standard theoretic models, whether or not grounded in the presumption that a tragedy of the commons is present, can therefore be only the beginning of a much more complex analysis. The temptation to seek out regulatory panaceas based on universal models, whether through private property, state action, or even notions of community, must be resisted in favor of a more nuanced approach. 66 In response to the inadequacy of one-size-fits-all approaches, Ostrom's method for systematizing the investigation of commons regimes focuses on the nested, multitier character of the natural resource commons. 67 Her Institutional Analysis and Development (LAD) framework is used to structure a common set of research questions to apply in diverse contexts, with the eventual goal of coming to some conclusions about the significance and interactions of various factors in facilitating effective management of common resources. 68 Our objective is to do the same in the cultural context.

To implement the LAD framework, Ostrom and her collaborators employed a three-pronged approach: " First, they engaged in a broad range of case studies of natural resource commons to form a basis for a bottom-up practice-based taxonomy of successful and unsuccessful approaches to resource management. " Second, they developed a framework, based on the initial case studies, for identifying the variables that are significant in deter- [Vol. 95:657 mining the success or failure of a commons enterprise and of the viability of institutions in particular contexts. * Third, they preserved flexibility in their framework so that it could be adapted and revised in response to learning derived from the case studies. 69 Based on the information obtained by applying their framework to structured case studies, they developed theories and models for particular commons situations, designed experiments to test those theories, and advanced a set of design principles for successful natural resource commons.

The framework approach recognizes the crucial importance of the interplay between the characteristics of a commons resource and the social, political, economic, and institutional arrangements for its governance in which it is embedded. 71 This approach also walks the difficult line between overly simplistic theoretical models, which paper over important complexities, and a fragmented taxonomy of diverse situations. As Ostrom explains, frameworks, theories, and models have different roles to play:

The development and use of a general framework helps to identify the elements (and the relationships among these elements) that one needs to consider for institutional analysis. They attempt to identify the universal elements that any relevant theory would need to include.... The development and use of theories enable the analyst to specify which components of a framework are relevant for certain kinds of questions and to make broad working assumptions about these elements. Thus, theories focus on parts of a framework and make specific assumptions that are necessary for an analyst to diagnose a phenomenon, explain its processes, and predict outcomes ... Microeconomic theory, game theory, transaction cost theory, social choice theory, public choice, constitutional and covenantal theory, and theories of public goods and common-pool resources are all compatible with the LAD framework discussed in this book.

The development and use of models make precise assumptions about a limited set of parameters and variables. Logic, mathematics, game theory models, experimentation and simulation, and other means are used to explore the consequences of these assumptions systematically on a limited set of outcomes. Multiple models are compatible with most theories . 69 See id. One needs a common framework and family of theories in order to address questions of reforms and transitions. Particular models then help the analyst to deduce specific predictions about likely outcomes of highly simplified structures. Models are useful in policy analysis when they are well-tailored to the particular problem at hand. Models are used inappropriately when applied to the study of problematic situations that do not closely fit the assumptions of the model . ... 72

In the cultural environment context, the need for a framework to systematize theoretical and empirical research is readily apparent. The conventional functionalist explanation of public goods leads to a rather constrained set of theories and models and a consequently constrained set of prescriptions and knowledge. The theories and models work well for some observed phenomena but not for many other phenomena. Yet identifying or developing other relevant theories or models (or even knowing when to look outside the proverbial box) can be difficult. 7 3 A framework can help researchers to develop and choose among different theories. For example, club theory may be helpful in understanding the dynamics of patent pools but poorly suited to understanding Wikipedia. Patent pools are less open in terms of membership, and they manage shared resources in a fashion that is much less focused on sustaining joint production. Wikipedia, by contrast, is quite open in terms of membership, contributions, and participation in various aspects of the project largely because it focuses on sustainingjoint production. Consequently, club theory is not likely to be particularly helpful for a researcher studying Wikipedia. Other theories of cooperation might be better suited to the task. A research framework such as ours aims to systematize the investigation, facilitate a more rigorous evaluation by matching and testing of theories and models with observed phenomena, and, most generally, enable learning over time. By design, the IAD framework remains a work in progress in the natural resource domain, 7 4 which is one of its strengths. We follow that tradition.

The LAD framework for natural resource commons is illustrated in Figure 1.75 It divides the investigation of a commons regime into a set of inquiries into groups of underlying factors. The first group, including biophysical characteristics, community attributes, and what Ostrom and her collaborators denote "rules-in-use," or governance 72 OSTROM, supra note 54, at 28-29.

Recent study of peer production in online environments provides a promising example of efforts to move beyond the conventional set of theories. See supra notes 9, 35, 42, and accompanying text. mechanisms, are denoted by Ostrom as "exogenous variables." These attributes of a particular commons are fixed, at least with respect to the analysis of a particular situation. 7 6 In the case of the lobster fishery, for example, these attributes might include the relevant biological characteristics of lobsters, such as the rates at which they age and reproduce; attributes of the community of fishermen, such as the proximity in which they live to others, the existence of familial relationships, and the skill sets needed for lobster fishing; and the rulesexplicit or informal-that govern fishing. 7 7 The second group focuses on what Ostrom calls the "action arena." An action arena "refers to the social space where participants with diverse preferences interact, exchange goods and services, solve problems, dominate one another, or fight (among the many things that individuals do in action arenas)"v 78 -in other words, the place at which the exogenous variables combine in particular instances, leading over time to observed patterns of interactions and outcomes.

Ostrom's work emphasizes the diversity of possible combinations of exogenous variables that determine what actually happens in particular instances and, hence, the outcomes that result. 79 The rules governing lobster fishing contribute to the activity's long-term sustainability, for example, but the patterns of interaction actually observed depend on the richness of the particular environment for lobsters, the degree to which rules are actually enforced, seasonal factors such as weather, and interaction with outside influences such as pollution 8 ° and the state of the larger economy. 8 ' Understanding the observed success or failure of a commons enterprise such as a lobster fishery may require accounting for all of these factors, even though it may turn out that outcomes are relatively impervious to some of them.

A simple illustration of the framework might be a soccer league. The formal rules of soccer are fixed, but the rules-in-use clearly vary somewhat between a professional league and a recreational league, between a children's league and an adult league, and so forth. A specific soccer league is also characterized by the relationships among the players (whether neighbors, professional competitors, or friends), by the attributes of the fields on which games are played, and even by the 76 Of course, as Ostrom herself recognizes, variables that are treated as exogenous or fixed for the analysis of a particular institution (laws relating to a fishery, for example) may be the outcomes of another (a legislature, for example). See OSTROM, supra note 54, at 13.

climate of the places where the games take place. The action arenasoccer games-depends on complex and specific interactions among all of these characteristics. Nonetheless, the outcomes over time in a particular league are the patterns of interaction that are clearly identifiable as "professional soccer," "friendly weekend game," "children's soccer league," and the like. Moreover, some leagues will be successful, lasting for years even as players come and go, while others will fail. The goal of applying the Ostrom framework of analysis in this context would be to use studies and analysis of many different soccer leagues to come to an understanding of success and failure as a function of specific context. The foundation of the Ostrom analysis is asking related questions in clusters about the exogenous variables, the action arena, and the patterns of interactions and outcomes. We think that the LAD concept has proven sufficiently fruitful to make it worth adapting for our purposes. 84 The nested, multi-tiered character of sustainable cultural environments and the diversity of attributes that contribute to successful governance regimes are keys to understanding cultural commons as mechanisms for knowledge production, collection, and distribution in the context of modern information and IP law regimes.

Ostrom and her colleagues have taken preliminary steps toward understanding how these methods might be relevant to investigating certain cultural commons. 8 5 Ostrom and Hess have analyzed the management of digital collections of existing knowledge resources, an admirable first step that signals the need for and plausibility of extending the LAD framework to the cultural environment. 8 6 As our discussion of the categories of questions below reveals, we argue that the LAD framework must be extended to account adequately for the wide variety of constructed commons in the cultural environment. 8 7 Most importantly, unlike commons in the natural resource environment, cultural commons arrangements usually must create a governance structure within which participants not only share existing resources but also engage in producing those resources. 88 This characteristic of cultural commons produces a more intertwined set of exogenous variables because separating the managed resources from the attributes and rules-in-use of the community that produces them is impossible. Cultural commons are also nested within and interact with more complex systems of natural and socially constructed environments.

To see the point, consider a copyright or patent pool, such as the MAA described in the Introduction, to which IP rights holders agree to contribute patents or copyrights that those same holders may exploit on standardized terms specified as part of the construction of the pool. 8 9 This arrangement creates an environment for pool members that facilitates sharing and use internally and interacts simultaneously with the external environment and shapes relationships with nonmembers. In other words, patent and copyright laws construct particular environments with default boundaries governing access to and 84 See, e.g., Commons Sense, ECONOMIST, Aug. 2, 2008, at 76 (paying tribute to the commons research of Ostrom and her colleagues and emphasizing the need for a study of"new commons" of the sort we focus on here). use of certain forms of knowledge. Commons arrangements grounded in those laws involve contextually specific deviations from the default given by IP law. These constructed cultural commons may lead to creativity, innovation, and improvement that would not be attainable either in the so-called "natural" state of information without intellectual property protection or in the state of information with "full" intellectual property protection.

Our proposed modified framework for constructed cultural commons, which reflects the iterative and constructed character of the commons community, its cultural resources, and its governing "rulesin-use," is illustrated in Figure 2. Figure 2 also reflects a second modification to Ostrom's AD framework, the collapse owlee distinction between outcomes and "patterns of interaction" that results from the intersection between the commons "action arena" and community attributes, resource attributes, and rules-in-use. We argue that given the role of both formal law and informal rule systems in commons governance, these patterns of interactions are inseparable from the outcomes of commons systems. How people interact with rules, resources, and each other, in other words, is itself an outcome that is inextricably linked with the form and content of the knowledge or informational output of the commons. In an open source software project, for example, the existence and operation of the open source development collaborative, the identity of the dynamic thing called the open source software program, and the existence and operation of the relevant open source software license are constitutive of one another. 9° Because of the more complex relationships among resources, participants, and governance structures in cultural commons, relevant attributes may not divide as neatly into categories as they do when one is describing a pooling of natural resources. Nonetheless, as we describe in the next Section, we believe that a structured interrogation of specific examples of constructed cultural commons is both possible and likely to be fruitful.

This Section describes in greater detail our proposed framework for approaching case studies of particular constructed cultural commons, building on the concepts reflected in Figure 2, above, and focusing on the character of the questions that should guide any specific investigation. We begin in Section 1 with some comments on important distinctions between our framework and the LAD framework for natural resource commons, namely that the cultural environment itself differs in certain fundamental respects from the natural resource environment. In Section 2, we describe how those distinctions drive the appropriate set of inquiries into basic characteristics of the commons. This section is the heart of our analysis. In Section 3, we briefly discuss how the expansion of the framework affects the analysis of patterns of interaction and outcomes of the action arenas arising in culturally constructed commons.

When seeking to apply the Ostrom approach to constructed cultural commons, we immediately confront a conceptual challenge. Ostrom's inquiry begins by asking questions about the "biophysical characteristics" of the resources involved in the limited commons in question. This inquiry assumes, implicitly, a natural environment containing natural resources that are to be shared and managed. In describing a constructed cultural commons, we must take a step back before describing the relevant characteristics of the shared resources to ask how we should define the environmental backdrop against which a commons is constructed. As is generally true for understanding constructed cultural commons, there may not be one correct answer to this question. There is no clean way to separate a particular cultural commons from the "natural" cultural background, because cultural activity is always grounded in human social interaction, the material environment, laws, and norms. Although there may not be one right answer, it is important to choose a starting point for investigation in a particular case. Asking the question ensures the salience of the choice of the background against which further description is 2010] made. Importantly, that choice frames the larger environment within which a particular commons and related institutions and practices are nested, leading to a better description of the sources and significance of the social, political, and economic aspects of commons. 9 '

We identify two reasonable points of "natural environment" reference for the investigation of constructed cultural commons: a "natural" cultural environment without intellectual property and a "default" intellectual-property-based cultural environment. These two starting points correspond roughly to the public domain and to a propertized environment, respectively. Which starting point is most appropriate to use for a particular inquiry will depend on which most closely approximates the cultural commons in a particular case. In a context such as a patent pool 92 or the jamband phenomenon that emerged around the recorded performances of the Grateful Dead, 93 for example, describing a constructed cultural commons according to how it deviates from the default IP regime may be the most useful approach. In other contexts, such as the shared journalism resource constructed by the AP, 94 describing a cultural commons according to its differences from a completely open public domain may be the most useful approach. In still other contexts, such as the sharing of magic tricks among the community of magicians, 9 5 secrecy may provide the most natural backdrop. Here we comment briefly on the "natural" cultural environment and on the copyright and patent law default propertized environments before moving on to suggest additional questions that should be pursued in analogy to Ostrom's framework.

Despite what appears to be the expanding scope of IP law, a significant range of activities, practices, and intellectual resources remain outside the intended scope of even the most expansive IP regimes. When cultural commons are constructed in these arenas, the most appropriate choice for the "natural environment" is a cultural environment unmediated by rights of exclusion or other regulation. This "natural" cultural environment may also be the appropriate starting point for discussing constructed cultural commons in which IP rights, though available, play a marginal role.

See supra notes 54-55 and accompanying text (noting the role of nesting in Ostrom's framework).

See supra notes 17-20 and accompanying text. [Vol. 95:657

The contours of the "natural" cultural environment are not uncontested. The major IP regimes exclude many different types of intellectual resources based on many different criteria and doctrines. Some would describe the complete set of nonenclosed resources as the public domain, including not only matter excluded on subjectmatter grounds but also matter subject to rights of fair use or fair dealing, 96 or as to which IP rights have expired. 9 7 The "natural" environment then can be seen as a vast pool of cultural resources that are openly accessible and openly usable by anyone without the permission of anyone else. 9 8

Julie Cohen has argued that a purely natural resources conception of the public domain, and one that relies on the distinction between permitted use and use that does not require a rights owner's permission, may lead to a misleading follow-on analysis too closely tied to geographic concepts-that is, a conception of the public domain as a separate place. 9 9 Cohen argues persuasively for a more contextual understanding of the "common in culture," a cultural landscape that is informed and shaped by cultural practices. 1 0 0 Our conception of the "natural" environment relates to Cohen's cultural landscape model as it similarly integrates a more dynamic and contextual understanding of intellectual resources. We might say that the "natural" cultural environment encompasses all that we inherit and experience. We inherit the natural physical environment-live within, use, interact with, and change it-and we pass it on to future generations. Similarly, we inherit, live within, use, interact with, change, and pass on an intellectual and cultural environment, which is itself comprised of many overlapping sub-environments of science and art, among other things. Experience constitutes an important intellectual resource that simultaneously relates human beings to their inherited and evolving environment(s) and constitutes a resource that may shape the intellectual environment. Experience (or perception or observation) is not enclosed within IP regimes except when expressed and embodied in a particular qualifying form, such as a copyrightable work of authorship or a patentable invention. that the cultural environment might be seen as a garden and the processes of inheritance and experience as cultivation. Nevertheless, he cautioned against taking the notion of a

In sum, the natural intellectual environment consists of a vast pool of open intellectual resources within which and with which we experience life and engage in a wide variety of activities and practices. The salience of specific features of a "natural" background will depend on the context of the inquiry. In many cases, constructed cultural commons arrangements build directly on this nonpropertized "natural" background. Examples of constructed cultural commons for which the "natural" environment is the most appropriate baseline likely include the commons of scientific research results and tools in the basic sciences, the facts and related information that are compiled not only by journalists via the AP but also in online creations such as Wikipedia, and devices invented by and shared among sports enthusiasts.

The "natural" environment may be the most appropriate baseline for viewing a cultural commons even if intellectual property protection is available for the resources contributed to the commons, and even if IP law plays some role in its construction. For example, the AP, which was constructed initially as a means of managing the sharing of an intellectual resource (news) that was not protected by the standard forms of IP law (the First Amendment generally precludes copyright protection for facts) was the plaintiff in the leading case that established an intellectual-property-like "hot news" doctrine regulating a very specific type of misappropriation of factual information important to the AP constructed commons. 1 0 3 The importance of a constructed cultural commons analysis is that it recognizes that creative environments are constructed by deviating from both the purely "natural" and the purely propertized extremes. Indeed, once we have identified the background environment and shared resources of a particular cultural commons, the bulk of the analysis will focus on the institutions that are constructed to govern deviations from that background structure.

The two principal regimes of IP law-patent and copyright laware the most salient alternatives to the "natural" environmental baseline described above.

Patent grants are justified generally on the ground that the natural shareability of technological ideas undermines incentives to pro-"natural" cultural environment too far. Our discussion of choosing an appropriate baseline should make clear that we have heeded his advice.

102 See, e.g., Nikolaus Franke & Sonali Shah, How Communities Support Innovative Activities: An Exploration of Assistance and Sharing Among End-Users, 32 Rxs. PoL 'Y 157, 169 (2003) (studying how individuals in a community of sports enthusiasts share information freely).

duce and distribute more and superior forms of innovation. 10 4 This basic conception highlights an important difference between constructed cultural commons and commons in the natural resource context. Constructed cultural commons must be concerned not only with managing and sustaining existing resources but also with providing institutions to encourage the creation of new resources. Patent rules vary somewhat from country to country, but generally time-limited patent rights are granted to the developers of an "invention" after an appropriate government agency examines a patent application. 1 0 5 The applicant must demonstrate to the satisfaction of the patent examiner that the innovation represented by the invention is new (or "novel," in the language of patent law) in that no one has invented the device before; useful; nonobvious (in the language of American patent law) or possessing an "inventive step" (in most European systems), such that the invention represents a technical advance over the existing art;' 0 6 and adequately described in the application for the benefit of future adopters and adapters of the technology. 1 0 7 The holder of a valid patent possesses a statutory right to exclude all others from producing or selling the invention, 1 0 8 subject to extremely limited exceptions for experimentation and research on the subject matter of the patent. 10 9 Notably, however, patent rights expire after a relatively short term, typically twenty years. 1 10 The material covered by the patent then passes into the public domain. An example of a constructed cultural commons for which a patented environment serves as an appropriate baseline is a patent pool.

Copyright law departs from the "natural environment" norm for the cultural environment in ways that resemble patent law, and for the same reasons, but with respect to material forms of artistic and creative cultural expression rather than technological and technical innovation. As with patent law, copyright statutes vary in their details from country to country yet generally embody a set of core principles. 1 1 1 The author of an "original" or creative work is granted a statutory entitlement to exclude others from reproducing, adapting, performing or distributing copies of that work to the public. Unlike patent law, copy-

right generally embeds a broad range of exceptions and exclusions, including exclusions of subject matter that is functional rather than expressive (and therefore the subject of patent law) or that is too broad or abstract to be identified clearly as the specific product of a specific author. In the United States, the copyright holder is subject to a user's power to engage in "fair use" of copyrighted material." 1 2 In the Commonwealth countries, a copyright typically is subject to a somewhat more limited "fair dealing" exception.

The next step after choosing an appropriate characterization of the "natural" environment in which a particular cultural commons resides is to identify basic characteristics relevant to the function of that commons in producing, managing, and disseminating intellectual goods. Here we suggest, as a starting point, a series of nested inquiries that we hope, when applied to and refined by a series of case studies, will assist researchers to identify the attributes that define successful and sustainable cultural-commons regimes and distinguish them from unsuccessful regimes.

112 See 17 U.S.C. § 107 (2006). 113 See, e.g., Copyright, Designs and Patents Act, 1988, c. 48, § 29 (Eng.) [Vol. 95:657

By analogy to Ostrom's inquiries, we propose and discuss in this Section the following initial inquiries, which we refer to as "baskets" of research questions: * particular subject matter, resources pooled and so forth; " particular activities undertaken and the actors who perform them; * goals and objectives of the cultural commons; " degree of "openness" of the cultural commons; and * governance or "rules-in-use" of the cultural commons.

These inquiries are related to the Resource Characteristics, Attributes of Community, and Rules-in-Use section of Figure 2.118

After choosing an appropriate baseline environment, the next step in investigating a constructed cultural commons is to identify the set of resources being pooled and the relevant community of actors. The resources might appear to be obvious, such as patents in a patent pool, news items for a news service, recordings for a database of music, or recipes shared within the community of French chefs. 19 Or, it may take some consideration to identify the most salient description of the relevant resources. What resources are pooled and shared in an open source software community? Ideas? Code? Coding expertise? Debugging opportunities? In many contexts, and perhaps even in patent pools and French cooking, multiple types of resources are being shared within a community. The dynamic character of intellectual and knowledge resources means that shared resources may not be fully independent of one another both in the sense that the value of one shared resource may depend on its relationship to other shared resources, and in the sense that any shared resource may arise from or be developed from a different shared resource. The multiple contributions to a single open source software project demonstrate both features. The durability of shared resources must be considered. Patents and copyrights expire; tangible objects may wear out or be consumed. Our framework aims to be inclusive and thus aware of the variety of resources collected in cultural commons. We avoid a myopic focus on intellectual property assets.

We refer to these as "basic characteristics" rather than Ostrom's "exogenous variables" because, as discussed above, here, much more than in the natural resources context, it is unreasonable to regard important characteristics of the resources themselves as "exogenous" to the commons structure. See OSTROM, supra note 54, at 13. 119 See Christopher J. Buccafusco, On the Legal Consequences of Sauces: Should Thomas Keller's Recipes Be Per Se Copyrightable?, 24 CARDOZO ARTS & ENr. L. J. 1121, 1151-55 (2007); Emmanuelle Fauchart & Eric von Hippel, Norms-Based Intellectual Property Systems: The Case of French Chefs, 19 ORG. Sci. 187 (2008).

Similarly, the identity of community members may be clear-as it is in a patent pool-or questions may exist about how the community is constituted. Does the open source software community consist of programmers alone? Does it include users of the code? People who submit comments or assist with debugging? Entrepreneurs who initiate meetings and dialogue or organize the community? People who develop, disseminate, and manage the relevant licenses? Those who monitor compliance with license terms? People who develop tools, host websites, and otherwise support the community? 120 Rather than seeking a single answer, our inquiry leads us to identify each of these constituencies and describe their specific roles. Critically, asking who is part of a particular constructed cultural commons sharpens the inquiry and helps pave the way for inquiries into institutions, governance, and sustainability.

Cultural commons exist on a spectrum. At one end of this spectrum are collective organizations that manage copyrights or patent pools. This subset is useful to work with because the set of pooled resources is easily identifiable, as is the relevant community of actors. Specifically, the set of resources is comprised of rather discrete, propertized, intellectual works, such as patented inventions, and the community is comprised of those who own those works and wish to reuse them. 1 2 ' At the other end of the spectrum are more complex examples. For example, the sharing and development of ideas, skills, tacit knowledge, and even the intellectual and cultural components of social capital within a university research community constitutes a constructed cultural commons. 22 This example invites significant variation among case studies based on the particular resources and subset of the university community targeted for study. Depending on the resources under investigation, the relevant community may be defined broadly in terms of the university as an institution, more narrowly in terms of a particular university or academic discipline, or most narrowly in terms of a specific academic unit, department, or school. Even something as seemingly mundane as an academic conference or scholarly presentation or an e-mail listserv among colleagues in a specific discipline might be usefully analyzed using our constructed cultural commons framework. [Vol. 95:657

Before describing the "rules-in-use" or governance structure of a constructed cultural commons, it is important to identify the particular problem or problems that a given commons is constructed to address. In the natural resource context, this question does not often come to the fore because common-pool resources are defined by the problem of subtractability or rivalrousness (e.g., removing lobsters from the pool results in fewer lobsters for everyone else) and the risk that a common-pool resource will be exhausted by uncoordinated selfinterested activity (e.g., unmanaged harvesting may jeopardize the sustainability of the lobster population) . 123 Intellectual commons address different problems, which are defined initially by the fact that, as public goods, intellectual resources are not rivalrously consumed. 124 A copyrighted work or patented invention can be "used" simultaneously by many people while it is part of a commons without diminishing its availability for others. For example, the news remains available on the ALP wire even after a particular member newspaper publishes an AP-generated story. The various problems that cultural commons institutions solve are not merely, or even primarily, problems of overuse. The problems addressed by cultural commons include the production of intellectual goods to be shared, the overcoming of transaction costs leading to bargaining breakdown among different actors interested in exploiting the intellectual resource, the production of commonly useful platforms for further creativity, and so forth. 125 More generally, we can distinguish among different types of cultural commons based on their core purposes. Some such commons arise as solutions to collective action, coordination, or transactions cost problems that exist apart from IP rights (and perhaps would not be solvable without these rights). These solutions might involve instances of cooperative behavior where members construct an open environment to pool resources and use those resources themselves for some specific purpose. Open source software projects, mediated by formal free and open source licenses and by informal communal structures for determining what code becomes part of the "authorized" code base, are examples of this type. 126 Standard-setting enter-prises also fit into this category, 12 7 as do joint ventures for scientific research and development. 128 These cultural commons depend on each member's possessing certain intellectual property interests as a facilitator of participation.

A second type of commons or pooling arrangement is intended to solve collective action, coordination, or transactions cost problems that exist only because of the IP rights themselves. 29 In some of these cases, a commons is constructed as a defense against potential privatization of commonly useful resources. Examples of such arrangements might include constructed commons for basic biological building blocks such as the Single Nucleotide Polymorphism (SNP) consortium 1 3 0 or the publicly available databases of genomic sequences that are part of the Human Genome Project. 13 ' Formal licenses and related agreements assure that participants in the commons become part of what amounts to a mutual nonaggression pact that is necessary precisely because of the possibility that intellectual resources may be propertized. So long as the resource is in the commons, it can be shared among commons members, and neither commons members nor outsiders are able to appropriate that resource, patent it, and then assert a patent claim against a commons member. Within the commons, research proceeds more or less as it otherwise would, according to informal disciplinary norms and free of (or at least, less burdened by) undue anxiety about propertization and potential holdup.

A third type of cultural commons may be designed to mediate among communities with different default norms. Technology transfer institutions, which enable universities and other nonprofit research enterprises to deliver information resources (such as the technical knowledge described in patent specifications) to the private market, are examples of this type of commons. 132 The cultural environment inside the university is typically characterized by information sharing not governed by IP rights, even if IP rights are present as matters of form.' 33 The environment outside the university is governed largely by IP rights. Technology transfer institutions may constitute an institutional pool or commons that mediates these two regimes. 13 4 Similarly, open source projects have developed "boundary organizations" to mediate their relations with commercial firms. 135 By specifying these distinct types of cultural commons, we are probably setting up a more sharply delineated field of institutions than is really obtained in practice. In any given commons, it may be the case-and may even likely be the case-that the motivation for the pool arises from a variety of considerations, that is, some do not arise from the character of intellectual property interests themselves, and some do.

Pooling arrangements also may exist for less socially salutary reasons. Most obvious is the case of members colluding to restrict competition, and it is certainly within the purview of our approach that commons should be evaluated in part by reference to the possibility of anticompetitive behavior and other possible costs. But by requiring as an initial matter that an intellectual commons operate via sharing of intellectual resources, we distinguish this project from investigations of cartels as such, which operate by sharing price and output information and which pose significant and obvious risks of anticompetitive behavior without offsetting welfare benefits. 136 The functional purpose of cartels is different from the cultural commons arrangements noted above; that is, cartels are not designed to create an open environment within which resources may be shared and productively used by members or to sustain individual members. But just as the line between different types of intellectual commons may be difficult to draw consistently, the line between commons and cartels similarly may be difficult to draw. Antitrust regulators have long faced the challenge of identifying illegitimate cartels disguised as legitimate patent pools and other knowledge-sharing institutions. 13 7 c. Degrees of Openness and the Character of Control Commons regimes are defined both by the degree of openness and control that they exhibit with respect to contributors, users, and resources, and by the assignment of control, or custody of the power to administer access. The rules-in-use of a constructed cultural commons will delineate its degree of openness, particularly with respect to use of the resources by outsiders who do not contribute to resource creation. Again, this inquiry is less relevant to natural resource commons arrangements. Natural resources generally are finite, potentially rivalrous in consumption, often congested, and subject to tragic overconsumption. 1 38 Consequently, it is often necessary to limit access to a common-pool resource to a defined community. The boundaries of the community sharing a resource tend to be coextensive with the boundaries of commons self-governance. 1 3 9 Thus, in many cases, the commons is open to members and closed to everyone else, and that is the end of the story. Intellectual resources, by contrast, are not subject to the same natural constraints and are naturally shareable without a risk of congestion or overconsumption. Rarely does "too much information" diminish the value of individual items of information.' 40 137 See HOVENKAMP ET AL., supra note 127, § § 30.4, 34.2.

Complications arise here because of distinctions between "natural resources" like oil, wood, and other purely rivalrous resources that are necessarily depleted and nonrenewable, and "natural resources" like forests, lakes, or other common-pool resources that are not necessarily depleted and are potentially renewable if managed sustainably. See RUHL ET AL., supra note 48, at 52, 64-65 (describing a "threshold of irreversibility," noting that "once thresholds are crossed, it can take enormous spans of time to rebuild natural capital through ecological processes," and describing the risk of congestion); Brett M. Frischmann, Environmental Infrastructure, 35 ECOLOGY L.Q. 151, 155, 166-68 (2008) (examining differences in the degree of rivalrousness for different natural resources and distinguishing between rival natural resources like oil and partially (non)rival natural resources like the atmosphere). It is entirely possible and desirable for a community to produce and/ or manage a cluster of cultural goods that is accessible to outsiders. 1 4 One of the measures of the social benefit of a constructed cultural commons may be the degree to which it disseminates the intellectual goods it produces to a wider audience.

What do we mean by openness? Little ambiguity exists in most everyday contexts (e.g., an open door), but openness can be a confusing concept when used to describe an intellectual resource. Openness describes our capacity to relate to a resource by accessing and using it. In other words, openness describes the extent to which there are barriers to possession or use. Openness varies according to the costs of surmounting barriers (in terms of money, conditions, or other restrictions) to exploitation. Openness in this sense may encompass joint or shared access to and use of the resource. 1 4 2 Barriers to possession or use of a resource may be natural or constructed. A resource may be open naturally because its characteristics prevent it from being possessed, owned, or controlled by anyone. 4 " Frischmann provides one example:

[F]or most of the earth's history, the oceans and the atmosphere were natural commons . .. [b]ecause, for example, exercising dominion over such resources was beyond the ability of human beings or simply was unnecessary because there was no indication of scarcity.

A resource also may be open.. . as the result of social construction. That is, laws or rules may prohibit ownership or ensure [a certain degree of openness] .... 144 For example, copyright law grants protection over creative expression but excludes protection for ideas in order to maintain open access and use of ideas. Patent law likewise excludes abstract ideas from patentability. Openness may arise through norms and customs among owners and users and through institutional design. 45 Openness and the vesting of control over openness are related. In part, both concepts may simply reflect choices regarding how best to manage resources. In the context of intellectual property pools, for example, management of the pooled resources may be vested in a central institution created specifically for that purpose, or it may be decentralized and vested in the hands of individual IP rights holders.t 46 Openness and the sources of control also reflect power and its distribution among potential possessors and users. Openness may be measured by the degree of control over the terms of access and use of a specific resource. Such control is exercised by human beings on human beings. Openness is relational, and it relies on social institutions.

In sum, openness is a functional variable that describes the degree to which possession and use of a resource is controlled, and it is a relational variable that describes the structure of relationships among potential resource users.

ii. Openness as Applied to a Community As a resource or set of resources may have an open character, so may a community. As openness is applied to resources, openness of a community is defined partly in functional terms, by natural and constructed attributes that define membership in the community, and partly in terms of power and other bases for relations among participants.' 47 Above, we defined the cultural environment as a set of interdependent and interconnected systems and resources. 1 48 As with openness applied to resources, openness with regard to a community describes our capacity to relate to that community as a contributor or user of resources that comprise in part the cultural commons. Thus, openness describes the extent to which there are criteria for or barriers -to membership or participation in the creative or innovative processes that the cultural commons is intended to support. Openness also describes the extent to which a particular community is accessible to and interconnected with related context, institutions, and social practices.

Openness with respect to a community has an internal dimension as well as an external one, as it reflects the degree to which participants in the cultural commons collaborate with one another or otherwise share human capital as well as (or rather than) resources. For 146 See supra notes 2-5 and accompanying text.

147 Accordingly, we focus much less on whether some social context is or is not a "community" according to predefined criteria and much more on the functional characteristics of that context.

[Vol. 95:657 example, the participants in an intellectual property pool may specify rules regarding how resources are contributed to and withdrawn from the pool. The General Public License for open source computer programs specifies that membership in the community defined by users of the program is open to anyone. 149 Anyone may add to, use, or redistribute the licensed program. Redistributors, however, must abide by the license term that they make the full source code of the program accessible to further users of the program. Moreover, in most open source software projects, only certain contributions are accepted into "official" versions of the code. 1 50 Thus, although use and modification of the code for personal use are open to anyone, the ability to contribute to the shared resource is regulated.

In describing and assessing the degrees of openness and control that characterize a cultural commons or pool, one should bear in mind more than just the conventional producer perspective by which information and knowledge shareability problems often are analyzed. Hardin's "tragedy of the commons" 1 5 1 is typically understood as challenging markets and governments to offer ways to supply resources in the face of cooperation and competition problems. 1 52 In analyzing openness with respect to resources and communities, accordingly, it is tempting to limit the analysis to openness with respect to actual and potential resource producers.

In information and knowledge environments, those resources are "naturally" given only in part. The cumulative and aggregative character of knowledge is fundamental to human culture. Producers of knowledge and culture resources are therefore simultaneously users and consumers. In analyzing openness, it is important to consider the degree to which openness expresses the interests of users, as matters of both function and relation. In particular, a constructed commons in the cultural environment may function as infrastructure. 15 3 In the cultural environment, the tragedy of the commons that Hardin described may refer not to an undersupply of a resource prompted by overconsumption but instead to an undersupply prompted by the failure of the private market to aggregate user or consumer preferences for certain fundamental or "infrastructural" resources. This situation occurs, for example, in the context of basic research conducted within and across universities. 154 To the extent that the Internet itself constitutes a commons, it is likely better characterized as an infrastructural resource that solves certain problems of consumption rather than problems of production.

Having identified a cultural commons, chosen an appropriate description of the background environment within which the commons is nested, and assessed the characteristics of associated resources and populations, goals and objectives, and the degree and character of openness and control, the next task is to investigate more specifically the characteristics of the cultural commons that relate to its governance and the way it functions. Here we identify several relevant clusters or "buckets" of variables that will be important to explore: 155 " history and narrative; " entitlement structures and resource provisions; * institutional setting (including markets and related firm and collective structures, social structures that describe the roles and interests of individual actors in the commons, and boundary organizations or mechanisms mediating internal governance of the commons with external markets, the public domain, and other institutions); " legal structures (including intellectual property rules, subsidies, contract and licensing law, antitrust provisions); and * governance mechanisms of the commons (membership rules, resource contribution or extraction standards and requirements, conflict resolution mechanisms, sanctions for rule violation).

What is the relevant history and narrative of a given commons? Above, we noted the importance of language and metaphor in understanding the information environment. 1 5 6 Any given knowledge pool depends in an important sense on its creation narrative. That narrative depends in turn on a variety of linguistic and metaphor resources: the vocabulary and syntax that participants and observers use in describing the construct are keys to unlocking its origins, its operation, and even its future. Carol Rose has written of property as a 154 See Frischmann, supra note 28; KatherineJ. Strandburg, Curiosity-Driven Research and University Technology Transfer, in 16 UNIvFRsrrY ENTREPRENEURSHIP AND TECHNOLOGY TRANS- FER: PROCESS, DESIGN, AND IrNTrELLECTUAL PROPERTY 93, 97-99 (2005); sources cited supra note 33; supra note 122 and accompanying text.

The clusters of questions that follow are analogous to Ostrom's inquiries into the descriptive characteristics of a commons regime. See Ostrom & Hess, supra note 66. story. 15 7 Michael Madison 158 andJessica Silbey 159 have both described the creation myths that accompany default regimes of intellectual property. The very phrase "patent pool," for example, itself has come to signify a specific set of legal expectations and criticisms. If one says "patent pool," an informed commentator thinks immediately of (i) a self-governing arrangement and (ii) antitrust considerations, rather than intellectual property problems and solutions. (In part, we aim to realign that point of view.) Calling something a "knowledge commons," or recharacterizing certain patent pools as solutions to "anticommons" problems, triggers a different set of expectations. The rhetorical frame shifts primarily to dynamic problems in information and information property, rather than to largely static output concerns. A commons is a rhetorically open place. A "pool" emphasizes the resources themselves and how those resources are bounded.

Explicitly giving attention to creation narratives also encourages attention to evolutionary processes. Changes in the narrative over time, or conflicts embedded within a narrative, can illustrate debates over purpose, which can illuminate the normative foundations of commons and highlight points of conflict. Looking at history can help answer the question of how the commons changes and adapts over time, in light of changes in firm structure, market structure, and resources (such as emergent legal structures, background legal entitlements, and temporal features like limitations on duration) that are built into those entitlements. History and narrative also emphasize the importance of contextual details that are ignored or marginalized in an overly rationalist account of institutional design. For example, within this cluster, we would like to uncover details concerning the influence of power, politics, and personalities that are often necessary to understanding commons. This cluster of questions provides an important link back to the earlier discussion of community and objectives. 160

In any resource pool, the resources that are part of the commons have to come from somewhere. This cluster is intended to capture an array of questions concerning the boundaries around the resources themselves and how those boundaries are socially constructed. Thus, 157 See CAROL M. ROSE, Property as Storytelling: Perspectives from Game Theory, Narrative Theory, Feminist Theory, in PROPERTY AND PERSUASION: ESSAYS ON THE HISTORY, THEORY, AND RHETORIC OF OWNERSHIP 25, 25-45 (1994).

See MichaelJ. Madison, Comment: Where Does Creativity Come from? and Other Stories of Copyright, 53 CASE W. RES. L. REv. 747 (2003).

159 SeeJessica Silbey, The Mythical Beginnings of Intellectual Property, 15 GEO. MASON L. REv. 319 (2008). moving beyond, or at least complicating, the initial set of questions concerning what resources are contributed and subject to the commons arrangement, we would like to better understand how the resources are delineated and how they are contributed and made part of the commons.

The "natural" information environment contains an abundance of raw information resources, including inherited and experienced knowledge. Those resources often become information "works" and, therefore, resources in the pool via some cultural construct, such as the default copyright or patent law systems, via some other institution, such as a publishing industry producing books, films, or songs, or via some combination of these and other things, such as cultural practices or norms. Understanding the construction of cultural commons therefore requires understanding the mechanisms by which resources are provisioned to the commons, whether via legal entitlements or otherwise, and the nature of entitlements to use and consume those resources while they are part of that commons. A patent pool offers an obvious example. The patents themselves are resources constructed via rights of exclusion offered by patent law. As pool members develop follow-on inventions based on the pooled resources, the agreement by which the pool is constituted may obligate members to contribute patents covering those inventions to the pool. 161 As with some natural resource pools that (when suitably managed) are self-sustaining and thus supply their own resources, in the cultural context the commons resources themselves may be sources of additional resources. The follow-on invention is an obvious example. An essential attribute of the governance of a cultural commons, therefore, is the way in which it allocates resources as they are produced dynamically.

Boundaries in an information environment are likewise more obviously culturally constructed than their counterparts in the field of natural resources. Oceans, lakes, and rivers have beds and shores; forests yield to fields. Boundary maintenance is an important part of commons management in natural resources, but the maintenance question often has a reference point in naturally occurring boundaries. In the information environment, all boundaries ultimately depend on cultural constructs. 1 62 Accordingly, this cluster is intended to help flesh out the connections between the construction of both resources and commons governance systems and the situatedness (or [Vol. 95:657 nestedness) of both resources and commons in broader systems. The cluster provides an important link back to the earlier discussion of baselines and degrees of openness. 163 iii. Institutional Setting Pools and commons in the cultural environment are functional entities; they often serve markets, industries, and firms. It is important to understand the identities and roles of those institutions and how their own functions relate to the pool and its members. What are those markets, and how do they relate to the pool? The MAA, an example of an early, well-known patent pool, was organized in large part to facilitate the production of aircraft for military use during World War 1.164 The institutional and social setting of a cultural commons may include related collectivist enterprises. Members of a pool may be part of a network structure that extends to related collectives, firms, individuals, groups, and social structures, including disciplines and social norms. 165 Research scientists may be organized formally into pools or commons structures within firms and other formal institutions, such as universities. Their functional network will include both members of their own technical art and related arts and other researchers in different arts who share a related but distinct set of social norms related to sharing of information and knowledge. Networks in not-for-profit or educational research settings will overlap to a degree with related networks in commercial environments. Researchers in university science departments will be interested in sharing information resources with researchers in corporate research-and-development groups. 166 participation in a particular project. 6 The range of institutional settings likely to be observed in case studies of cultural commons is thus likely to be considerably broader than the range observed in natural resource commons. The institutional setting will, of course, constrain and determine the ways in which rules for production and sharing of resources are both put into place and applied.

Although industry, market, and networked institutional structures are essential reference points for many knowledge commons, positive law and direct government involvement with a particular cultural commons are likewise keys to understanding it. We distinguish between law that creates and enforces the entitlements that create and sustain information works, on the one hand, and law that is specifically addressed to cultural commons themselves, on the other hand. Here, legislators and judges often find that law can reinforce and sustain a pool that is determined to be welfare-enhancing. An exemption or more deferential treatment from antitrust scrutiny for parties engaged in a form of concerted activity, or intended to engage in concerted activity, may be adopted. 169 Market conditions or technologies may develop to the point where observers recognize that some kind of information collective would be useful, but fear of prosecution under antitrust law or relevant IP law may be a barrier to the emergence of the pool. A safe harbor of a sort may emerge, either via legislation or judicial decision. The 1984 judgment of the Supreme Court of the United States in Sony Corp. of America v. Universal City Studios, Inc.,' 170 upholding the legality of distributing videotape recorders over the objection that they facilitated copyright infringement, may be characterized as creating a form of judicial safe harbor for innovative technologies for reproducing and distributing copyrighted works.

Legal rules may create subsidies or safe harbors in ways other than relieving at-risk parties from potential liability. For example, income-tax regimes may permit (or limit) the deductibility of research expenses by firms, nonprofit enterprises, and/or research collectives. In the United States patent statute, the section that bars patenting inventions that are "obvious" in light of prior art in the relevant technical field includes a subsection that suspends the rule if the inventor and the producer of the relevant prior art are part of a common 'joint research agreement." 71 Laws designed for one purpose may contrib- [Vol. 95:657 ute differently to promoting collaborations or collectives in ways not intended by the drafters of the law. Such a rule becomes part of the constitution of a commons, even if it was not designed to do so in the first place. Jessica Litman uses this proposition to analyze the persistence of a legal regime subsidizing jukeboxes in American copyright law.' 72 A compulsory license permitting owners of coin-operated record players to use copyrighted American music was initially incorporated into the copyright statute to prevent holders of those copyrights from monopolizing adjacent markets for performances. 173 Over time, the rationale for the subsidy became less significant, but the statute was retained because a new collective emerged to support its continued existence-companies that manufactured and distributed jukeboxes.

A constructed commons is an alternative to proprietary exclusion and to direct government intervention as a means of addressing market failures associated with public goods, externalities, and the tragedy of the commons. Accordingly, understanding the mechanism of governance of a particular commons, in the context of its legal and institutional setting, is at the heart of the analysis. In Ostrom's work, the degree of se/fgovernance is an important characteristic of a resource pool.' 74 Members have rights not only to contribute to and extract from the pool but also to govern themselves by adopting and modifying the relevant rules of participation.

The attributes to be considered here overlap to some extent with those addressed in the context of determining the scope of the openness of the pool. The focus shifts, however, from access to the resources of the commons to participation in decision making about how the resources will be produced and managed. Who is a member, and who decides who may be a member? J. 337, 351-53 (2002). the programs, constitutes assent to its terms.' 75 Violation of those terms, such as onward distribution of a copy of a program without including a copy of the program's source code, constitutes a license violation and automatically terminates that membership. 17 6 Actual enforcement of that regime, however, typically is not pursued by individual contributors to the open source commons but instead by an independent entity, the Free Software Foundation, which operates as a freestanding nonprofit organization dedicated to advocacy on behalf of "free" software and accompanying open source license terms. 1 77

Research on natural resource pools emphasizes that effective selfgovernance typically requires formal access to public sanctioning and/or enforcement mechanisms. Without the threat of seizure or attachment or injunction, community-based or purely norm-based mechanisms may lack sufficient bite to sustain the pool. In the context of cultural commons, effective connections between self-governing collectives and formal sanctioning authorities have not yet been identified. In the open source computer software area, only recently have courts begun to consider the enforceability of the licenses. 178 Conflict resolution mechanisms within a pool depend on monitoring mechanisms. Before the emergence of the Internet, research on self-governing communities emphasized size and distance as key variables in a monitoring system. As Benkler 179 and Cohen 8 0 each argue, networking technology offers not only the potential for community development and resource aggregation but also the potential for monitoring and enforcement. Examination of a pool should include assessment of whether and how it is embedded in network technologies that perform some or all of the pool's governance functions.

Finally, as when analyzing natural resource commons, the analysis should include an inquiry into patterns of interactions and outcomes, including: 178 See, e.g., Jacobsen v. Katzer, 535 F.3d 1373, 1381-82 (Fed. Cir. 2008). In Jacobsen, the open source collective was represented by one of its members, Jacobsen, as assignee of copyright interests in the open source program at issue. See id. at 1375-76. [Vol. 95:657 " solutions to the underlying collective action problem and benefits delivered by the commons (including innovations and creative output, and production, sharing, and dissemination of those innovations and output to a broader audience), along with the dynamics of the social practices that emerge from the interaction of commons resources and commons participants; and * costs and risks associated with the commons, including, for example, any negative externalities.

A cultural commons should be assessed not only in light of its ostensible purposes but also in light of its consequences. This aspect of the case study approach should both identify the consequences and describe relevant criteria for evaluating them.

The consequences themselves typically will take at least two forms, which in a particular case often will be inextricably linked. With respect to intellectual resources, a cultural commons usually will produce some intellectual or knowledge-related (or material) output. The MAA enabled the production of airplanes. The Linux open source project supports the Linux computer program. Wikipedia produces Wikipedia.org. The AP enables the production of newspapers.

In the case of most cultural-commons examples, moreover, the social patterns that emerge from the construction and governance of the commons may themselves constitute ongoing, constantly refreshed commons outcomes. Many of the companies that were parties to the original MAA agreement combined via merger and acquisition by 1929 to form the Curtiss-Wright Corporation, which is still a significant defense contractor today. 18 1 The Linux project and Wikipedia are notable not only for the production of complex industrial-scale products but also for the governance of networks of loosely aligned contributors, distributed broadly in space and time. 18 2 The AP and other wire services have cultivated and retained identities as distinct and productive enterprises in their own right despite the fact that much of what they publish is created by their members. 1 83 The jamband community is a recognized community that defines itself partly via its practices of archiving and sharing jamband performances.' 8 4 a.

We defined constructed commons in the cultural environment as solutions to collective action or other transactions cost problems not arising from the character of intellectual property entitlements them- selves, as solutions to problems that do arise from those entitlements, as solutions to boundary-spanning dilemmas, and as reactions to an "infrastructure"-type problem-the market's inability to aggregate individual demand for standards or platform resources-that is the inverse of the standard tragedy-of-the-commons diagnosis.

For any specific cultural commons, the questions involve not only the type of problem that it appears to be designed to solve and precisely how the combination of legal rules and other "openness" constructions propose to solve it, but also the success of the commons in sustaining and generating spillovers and a dynamic cultural environment. Quantifying or otherwise documenting that success is particularly difficult in the cultural environment, because the desired benefits often accrue to populations other than those in direct producer/consumer relationships. Commons can enable spillovers, 18 5 the dynamic benefits that an information environment can be designed to enable whether in its "natural" state, via the "default" variations on that state as described earlier, 186 or via some pool or other constructed environment. One beneficiary of the MAA patent pool was the American government, which purchased aircraft during World War I. 187 The beneficiaries of an online jamband archive may include ordinary music lovers, who are able to listen to jamband recordings even though they may not count themselves among the jamband community itself. 18 8 Under some circumstances, the very persistence of an institution may be evidence of the success of a commons regime. 18 9

Any cultural commons may engender a tradeoff between the benefits anticipated from the commons in terms of dynamic welfare enhancements and the costs and risks associated with that commons. In conventional law-and-economics terms, these costs and risks are fairly well understood (and, importantly, they are generally better understood and easier to describe and quantify in many instances than the downstream benefits that commons may supply). Enabling collaboration and cooperation among firms sharing access to pooled information resources facilitates cooperation along lines that may be anticompetitive and therefore socially harmful: agreements to raise [Vol. 95:657

and fix costs and agreements to reduce output. 190 Pools, like other collective arrangements, also involve administrative costs associated with constructing, monitoring, and enforcing compliance with the rules of the pool. From a welfare standpoint, the level of those costs must be compared to the level of administrative costs associated with a system that provisions information resources in the absence of the pool.

In sum, we offer the framework described in this Part as a template for ongoing case study investigations of constructed cultural commons across a broad variety of domains. In many contexts, the existing scholarly literature has undertaken case study investigations of phenomena grounded in social norms, 191 transactions cost economics, 19 2 and even history and anthropology,' 93 all of which may be profitably aggregated and recast as examples of cultural commons. One step in that direction is an application of the framework to analyze universities as cultural commons. 194 Collecting and reconstructing this literature using the clusters of questions listed above will, in our estimation, yield new insights into the emergence and effective functioning of cultural commons.

Going forward, we anticipate developing an inventory of new commons case studies, perhaps beginning with more detailed investigations of some of the examples discussed in the Introduction. We also hope other scholars will consider using this framework as part of their own work. Over time, we anticipate that the results of further case studies will yield not only better information regarding commons themselves but also refinements to the constructed cultural commons framework and to the above baskets of questions. In a real sense, the study of commons is itself a constructed cultural commons, and our own three-part collaboration is a nested commons within the scholarly community that studies commons.

The theoretical discussion of intellectual property policy has been myopically focused on extremes of exclusion and open access, ignoring a wide range of cultural commons that persist between the extremes. It is often divorced from empirical studies of creative and inventive communities. To the extent that case studies are undertaken, they tend to be done in isolated areas (such as open source software or academic publishing) and consider a limited number of descriptive variables. This narrow focus makes integration and learning from a body of case studies quite difficult, which in turn discourages people from pursuing further case studies. Scholars appear to be aware of the need for a more nuanced and structured approach to these questions but have not yet developed a framework for studying them.

This Article offers precisely such a framework. Applying the environmental metaphor that is increasingly common in studies of information and intellectual property policy, we analogize information and knowledge resources in the cultural environment to physical resources in the natural environment. We identify a set of constructed cultural commons, or pools of information resources, that serve functions in the cultural environment similar to the functions provided by common-pool resources in the natural environment. Those functions consist largely of serving as alternatives to purely private rights of exclusion and to government intervention in solving underproduction and overconsumption problems associated with an unmanaged or "natural" resource. Although constructed commons in the cultural environment exist for a variety of purposes, in general we hypothesize that they are often welfare-enhancing in regard to promoting valuable spillovers of information and knowledge distribution.

Borrowing from Ostrom, we argue that understanding the origins and operation of beneficial cultural commons requires detailed assessments that recognize that they operate simultaneously at several levels, each nested in a level above, and that each level entails a variety of possible attributes that cannot, at this stage of the inquiry, be specified in detail in advance. We suggest a set of buckets or clusters of issues that should guide further inquiry, including the ways in which information resources and resource commons are structured by default rules of exclusion, and the ways in which members of these pools manage participation in the collection, production, preservation, and extraction of information resources. Case studies across disciplines and reviews of existing literature that addresses cultural commons will help specify relevant attributes within each cluster. These variables will help scholars and eventually policymakers assess the level of openness associated with a given commons and determine the extent to which "openness" is, as we hypothesize, associated with pools that are welfare-enhancing.

Beyond our proposal of a framework for studying them, our consideration of constructed cultural commons has highlighted a number of points that are important in the study of intellectual property going forward. Considering constructed cultural commons helps to elevate collective, intermediate solutions to their possible place of significance in accounts of property regimes and should diminish the skepti-cism of many scholars that collective, norm-driven solutions can work beyond narrowly defined situations. Case studies will also call our attention to the constructed, designed character of both the cultural and the legal environment in regard to knowledge and information policy problems. Finally, as they have done in the study of natural resource management, systematic analyses of cultural commons across a wide range of collected case studies should lead us to doubt panacea prescriptions drawn from overly simplistic models.