Part I: Introduction

“Openness” has become a watchword for higher education in Canada.¹ From open-access textbooks and free online courses to a range of initiatives around open data and publishing in research, calls for more openness are everywhere on university campuses.² This “open” movement is moored in genuine optimism about the potential for digital media to improve access and affordability, especially for students at a time when they face soaring costs to post-secondary education.³ In smaller measure, the move toward “open” is also driven by a shared enthusiasm for new modes of knowledge production, based on the belief that authors–freed from copyright constraints on sharing and re-use–will collaborate to produce works that rival the quality and coverage of those created through the dominant commercial publishing houses.

But Canadian law teachers have not yet, by and large, embraced openness in that traditional centerpiece of the law school curriculum: the legal casebook. Nor have they gone very far in collaborating with one another toward that end. In this article we try to explain why, then propose a new conceptual and practical approach to creating open casebooks in the future.

The current lack of openness in the materials produced by and for Canadian law teachers might seem surprising–not only in light of the open-access movement’s broader successes, but also because legal casebooks are in some ways ideal candidates for openness. Most source materials used in conventional law school courses–statutes, regulations, case law and the like–are free from significant copyright restrictions,⁴ enabling creators to avoid the high costs of negotiating special licences or locating open alternatives to build their courseware. Law teachers have for many years produced their own self-published “course packs” for their students precisely because most of the raw materials for legal education are relatively easy to access and remix. But while the idea of open casebooks has received some recent attention in legal education circles,⁵ in practice Canadian law teachers will still find it difficult to locate a single open casebook suitable to assign or adapt for most of their courses.

To be sure, there are a few excellent examples of open legal casebooks available in Canada. But even among those examples, most if not all are produced and maintained by one author or a small group of authors using what appear to be standard methods and technologies for co-authorship, rather than by alternative modes of “peer production” in which groups of contributors flexibly build, share, and remix materials as part of broader collaborative projects.⁶ In other words, while some existing open casebooks are of high quality, the process of creating them seems to track that of closed casebook authorship in most if not all significant respects.

Focusing on the question of how open casebooks are created may have significant implications for the sustainability of openness in the future of legal education. There has been a great deal of excitement about the potential for alternative economies of commons-based peer production in computing, science and technology. But according to the standard theory, legal casebooks–like other educational texts–lack some of the key structural features of other knowledge goods, like software or encyclopedia entries, that make them amendable to production in peer networks.⁷ For casebooks, these missing features define the incentive structure of production in a way that makes it difficult, if not impossible, to build them collaboratively among any large or loosely organized group.

Or so the conventional story goes. In this paper, we offer an alternative account of the challenges and possibilities for building open casebooks collaboratively. Our account tests the basic assumption that good peer-produced texts can only be achieved through a process designed to reach and sustain group consensus about the final product of collaboration. Much of the existing literature on peer production assumes that the end goal of any collaborative project is to create a unified “knowledge object” (e.g., a book or computer application) from the shared contributions of many participants. Our alternative approach eschews this basic goal of consensus around core controversies in casebook design–What does this case stand for? How should that doctrine reconcile competing values?–and makes space for disagreement, dissent and ultimately disloyalty without the need to abandon a shared commitment to the collective project represented by the casebook as a whole. As it turns out, this is exactly how some prominent examples of commons-based peer production actually work in “classic” open domains like open-source software.

In Part II of this paper, we aim to do four things. First, we apply insights from research on knowledge commons to describe why openness matters for legal casebooks and to detail the link between commons-based and collaborative peer production that has largely been ignored in the literature on openness in education. Second, we report some early data on the incidence and form of open casebooks available for use in Canadian law school classrooms. Third, we recount the conventional story among scholars of peer production as to why educational texts are difficult to sustain using non-market forms of content creation. Fourth, we describe why this conventional story fails fully to explain the lack of peer-produced open casebooks and we sketch our theory addressing this gap using the two concepts of “exit” and “voice” in collaborative casebook design.

Against this background, in Part III we explore an approach to open casebook development that supports “disloyalty” to group consensus while sustaining commitments to the broader common project. In order to illustrate, we turn to the world of internet browser software as a prominent example of the “forking” model that open-source developers have embraced to balance exit and voice and to sustain several versions or flavours of a shared endeavour. We then briefly describe some practical and technological considerations for implementing our proposed model of collaborative open casebook authorship, using a first-year property casebook we are developing to illustrate.

Part II: Openness in Canadian Legal Casebooks

Our empirical claims in this paper are (1) that we have seen a relatively small number of open casebooks produced for use in Canada to date and (2) that we can point to few, if any, examples of casebooks created through a collaborative process of peer production. These observations raise critical questions about the future of openness in legal education and about whether or not we should pay closer attention to the role of authors’ motivations and incentives to engage in non-market forms of casebook production that do not rely on exclusionary intellectual property rights. We do not claim to answer those questions here. Instead, we aim to situate open casebooks in the broader discussion about commons-based peer production and evaluate how well existing theories in that field explain the available data. Finding some blind spots in the current theory, we go on sketch what we think is one possible missing piece of the puzzle.

Casebooks as Commons-Based Production

Open legal casebooks are a subset of open educational resources (OER) produced specifically for use in law school courses. In broad terms, OER are digital learning materials made available by their authors to others for free, under a licence that permits copying, remixing and redistributing the content (with varying rules about when attribution to the original author or authors is required). Accordingly, open casebooks are comprised of three essential components: (1) the learning content itself (such as cases, statutes, commentary, exercises, etc), (2) the digital tools used to develop and share that content (from simple text files to complex content management and web-publishing systems) and (3) the legal resources needed for implementation–namely, “copyleft” licences, like Creative Commons,⁸ that modify the default copyright restrictions placed on original works.⁹

Growing interest in open casebooks is part of a broader global movement to build resources and capacity for publishing, sharing and using OER in a whole range of educational settings. In 2019, the General Conference of the United Nations Educational, Scientific and Cultural Organization (UNESCO) emphasized the benefits of OER in a recommendation observing that:¹⁰

the application of open licenses to educational materials introduces significant opportunities for more cost-effective creation, access, re-use, re-purpose, adaptation, redistribution, curation, and quality assurance of those materials, including, but not limited to translation, adaptation to different learning and cultural contexts, development of gender-sensitive materials, and the creation of alternative and accessible formats of materials for learners with special educational needs.

As the UNESCO recommendation suggests, the primary case for supporting OER within the open access movement rests on their potential to lower or eliminate the cost of learning materials for users and thereby increase affordability and strengthen equality of access. As the growing cost of learning materials outpaces even the rising costs of education overall,¹¹ OER may help students to cut their expenses substantially and therefore improve financial access to learning environments that for many are falling out of reach.¹²

In the language of those who study how scientific and cultural information is generated, the benefits of OER in terms of access and affordability are a consequence of their production in commons-based system. Commons-based knowledge production draws its inputs from the public domain and returns its outputs back to the commons.¹³ This model–when it can be sustained–may overcome a fundamental problem with the market-based production of knowledge objects, which relies heavily on exclusive intellectual property (IP) rights. As modern IP scholarship now recognizes, information is a public good–meaning that it has the inherent quality of being both “non-excludable” and “non-rival”.¹⁴ In other words, once new information has been created, there is little or no marginal cost of reproducing or sharing that information among all potential users.¹⁵ While producing and distributing conventional print casebooks entails costs to the publisher arising from material inputs, labour, shipping and so forth, there are often minimal costs associated with sharing digital books through a digital network. From the vantage point of neoclassical economics–where, in a competitive market equilibrium, price equals marginal cost–the optimal price for reproducing and sharing information that has already been created is effectively zero.

Based on this logic, which OER advocates have long emphasized,¹⁶ commons-based production of educational materials avoids the “deadweight loss” to society that arises whenever private IP rights are afforded to information producers. But the worry with this result, and the very premise of modern IP, is that at zero price there is no material incentive for prospective creators of intangible goods to produce them in the first pace. While law students–and society as whole–would be better off if existing casebooks were open to all, casebook authors may be less inclined to write and update those books without some monetary return to their time and effort. Contemporary IP law’s response to this worry is to afford creators a set of private property rights (such as copyrights and patents) in the information they produce. Those rights allow creators to exclude others from using their original works of cultural or scientific production unless users pay some positive price to do so.¹⁷

Market-based production of educational materials therefore raises a fundamental conflict of aims. While IP rights incentivize creation of materials by giving authors the right to charge a positive price, the seemingly inevitable trade-off is the deadweight social loss arising from foregone opportunities to access the information at its marginal cost. While advocates have rightly drawn attention to the benefits of commons-based production of OER when freed from market-based IP rights, they have less frequently engaged with the incentives problem that arises for authors in the absence of a legal basis to charge for their work. The ultimate consequence of this incentives problem may well be the under-production of open casebooks and other forms of OER.¹⁸

While OER advocates have underplayed this problem of incentives, nearly a generation of legal scholars and economists–following Yochai Benkler’s early work in this area–have addressed the challenge at a broader level by exploring “peer production” as an alternative model of economic organization. Peer production refers to the creation of informational goods within decentralized and often (though not exclusively) large networks of participants who individually contribute small amounts of time and effort toward a collective project. Taking open source software (OSS), Wikipedia and various digital citizen science initiatives as longstanding examples, advocates of peer production claim that it represents a new structure for coordination in the digital environment–one that can sustain commons-based models for creating information, even in the absence of the material incentives IP rights make possible.

The potential for peer production in domains like OSS is now well established. But whether or not OER–and legal casebooks in particular–are likewise amendable to peer production and therefore sustainable over the long run is a crucial question. Unfortunately, the answer offered by research on peer production to date has not been an especially optimistic one. In the next section we examine the available evidence on the number of open legal casebooks that currently exist in the Canadian context and on the process by which those casebooks were produced. In the following section we look at the standard theory used to explain our results, which predicts that OER are not generally amenable to peer production. Finally, we explain why we need to extend this theory, along with our understanding of the challenges facing open casebooks, in order to move toward truly sustainable models of collaborative peer production in the legal domain.

Evidence of Open Casebooks in Canada

The task of locating relevant OER is a well-documented problem in the world of open education.¹⁹ Our attempts to build a database of open legal casebooks in Canada reflected this challenge. There are several digital platforms available that might potentially host an open casebook but there is no single, centralized repository for easily locating these materials. Moreover, existing repositories for OER often lack adequate search and filtering tools for identifying legal casebooks specific to the Canadian context. To our knowledge, no existing studies have attempted to compile these data from online repositories or to produce survey or other primary data on open casebooks in Canada for analysis.

In light of these challenges, we sought to compile an initial data set of publicly-available information on open casebooks in Canada using two concurrent strategies. First, we searched more than a dozen OER repositories that we identified as capable of hosting open content for use in Canadian law schools. Out of that initial search, we identified one or more completed or partially completed casebooks in each of five repositories: the Canadian Legal Information Institute (CanLII), Pressbooks institutional repositories, H2O, BCCampus’ Open Textbook Collection, and Wikibooks.²⁰ Second, we circulated a request, through the Canadian Association of Law Librarians listserv, to law school libraries across Canada for public information on open casebooks produced by teachers at each school. Using these two strategies, we gathered basic bibliographic metadata for each open casebook identified in our search. Our resulting data set is therefore only a preliminary one and we do not claim that it is comprehensive of all open casebooks currently available in Canada. We note, however, that by definition open casebooks must be publicly accessible, so any existing materials not identified through our relatively intensive search methods might de facto fall outside that definition. It is also important to clarify that our dataset does not include other open educational materials, such as extended syllabi or instructor guides. Our search was for material which could be used as the primary textbook in a Canadian law course.

Figure 1: Incidence of Open Legal Casebooks in Canada

Our final data set includes seven completed and sixteen incomplete open legal casebooks. The completed casebooks addressed a wide range of subject areas, from general (tort law, administrative law) to specific (global corruption). Casebook authors were faculty members at the University of Ottawa, the University of Victoria, the University of Windsor, Dalhousie University, and the University of British Columbia, although not all authors had a clear institutional affiliation.²⁴

Figure 2: Completed Casebook Metadata

In our sample, we found little to no evidence of peer production or other forms of collaboration or remixing among the casebooks in our data set. Six of the seven completed casebooks were produced by a single author. Even on the Wikibooks platform–which is specifically designed for small decentralized contributions similar to Wikipedia–in 12 of the 14 Wikibooks, 90 percent of the text was added by one user.²⁸ We found no evidence that any casebook in our data set, complete or incomplete, had been remixed (copied, adapted, and republished) or was itself a “remix” of a preexisting book.

Relevant to this lack of collaborative authorship may be both the technical capabilities of the publication platform and the licences under which the materials are published. Among the completed casebooks in our data set, those published on CanLII and BCCampus are available only online or in PDF or ePub formats. Since CanLII and BCCampus are only repositories, not open textbook creation platforms, textbooks cannot be remixed and republished on the site itself.

By comparison, Wikibooks, Pressbooks, and H2O are all book creation platforms as well as repositories, and therefore come with specific functionality that make remixing and collaboration more or less possible (see Figure 3). Wikibooks is designed for collaborative production of textbooks based on the Wiki model: anyone can contribute to or edit an existing book at any time. This model is designed to produce a single, constantly-evolving iteration of a textbook; as such, it has no in-platform features that permit remixing or publishing remixed versions of textbooks, although Wikibooks can be downloaded as PDFs or XML files. Pressbooks is designed for the creation of educational books and hosts a general repository of over 3,000 texts. In addition, many different institutions (including eCampus Ontario, AtlanticOER, Dalhousie University, Toronto Metropolitan University, and the University of Saskatchewan) have standalone Pressbooks networks with their own repositories. Pressbooks features collaborative authorship tools as well as a “cloning” tool that allows authors to copy, adapt, and republish existing books in-platform, provided those books are published publicly and openly licensed. However, cloning is only available within institutional Pressbooks networks–and a single author license to use such a network is $500 a year.²⁹

Harvard’s H2O is the only law-specific platform and repository we searched that hosted Canadian casebooks. Designed by the Library Innovation Lab at the Harvard Law School Library, H2O is built to create, share, and remix open legal materials, including casebooks.³⁰ Accordingly, users can freely and easily create and remix casebooks within the H2O platform. H2O also includes tools to facilitate the collaborative creation of open materials. The different types of licensing arrangements for all casebooks in our data set are described in Figure 4.

Figure 3: Collaboration Tools by Platform

Figure 4: Open Casebook by License Type

In summary, our data backs up our claim that there are few extant examples of Canadian open legal casebooks, especially completed casebooks ready for classroom use. The scarcity of co-authored casebooks and casebook remixing also supports our second claim: the majority of existing open casebooks were not created through collaborative peer production. The broader literature on OER reflects similar findings when it comes to the collaborative production of open materials. A review of the literature in 2014 found “there is little empirical evidence” that people actually revise, remix, or redistribute OER.³¹ More recent literature suggests the rate of OER remixing, including textbook remixing, remains low.³²

Explaining Barriers to Peer Production

The data described above are consistent with the now-standard story about why commons-based peer production is a viable alternative to market-based models for some informational goods (like computer software) but not for others (like educational texts). That story turns on two core components. The first deals directly with the underlying theory of motivations and incentives, while the second addresses connected issues surrounding the practical structure of peer-produced projects in terms of three characteristics: modularity, granularity and the costs of integration.

As many scholars of peer production have pointed out, it seems obvious from a practical standpoint that the vast majority of contributors to successful projects like Wikipedia or Slashdot³³ are willing to participate even in the absence of monetary incentives made possible by exclusionary property rights. Why then do these contributors participate at all? Benkler theorizes that contributors to peer production are motivated by a diverse set of non-monetary rewards. Some of these may come in the form of personal satisfaction–for example, the satisfaction of neatly describing a legal argument or the warm feeling generated by a well-copy-edited text (which is unquestionably its own form of reward). Other motivations may be social-psychological in nature, derived from social goods like status and reputation. This latter category includes rewards that are often imbued with cultural meaning and therefore specific to cultural context.³⁴

Empirical research over the last few decades has largely confirmed the idea that participants are motivated by a number of possible non-monetary rewards. These range from “intrinsic” rewards such as enjoyment, fun, ideology and kinship, to “extrinsic” (including “internalized extrinsic”) rewards such as own-use value (personal need), learning new skills or knowledge, reciprocity and reputation.³⁵ A related branch of this literature has suggested that participation in peer production is largely part of a “social practice” that develops and is reinforced by social norms.³⁶

The second component of peer production and a key focus of research on these processes concerns the structure of peer-produced projects themselves. Certain structural characteristics are thought to be necessary preconditions to peer production because they enable and support the range of non-monetary incentives described above. Three structural conditions have been identified and emphasized: modularity, granularity, and costs to integration. The premise behind these conditions is that, while non-monetary rewards may be central to motivating participation in peer production, such rewards are unlikely to be very large for any given contribution. This means that, for each contribution, the costs to participation must be relatively low. Those costs are determined, in turn, by the combined effect of modularity, granularity and integration on individual incentives to participate.

Modularity means that the project can be broken down into smaller components or modules that are feasible to produce independently. This characteristic supports the basic nature of decentralized peer production, in which participants independently select tasks to work on without the need for centralized oversight and coordination. The second property, granularity, refers to the size of the modules, which in turn determines the both the relative costs of individual contributions (in terms of time and effort) and, potentially, their relative benefits (depending on whether more value is ascribed to larger contributions or to a larger number of smaller contributions). In general, heterogeneous granularity may be desirable to allow contributors with varying levels of motivation to participate as much or as little as they like. Once individual tasks are completed, they will–to a greater or lesser degree–need to be integrated back into the project as a whole, raising the possibility of integration costs. Those costs may depend on the content of the project (e.g. narrative text vs computer code) as well as the granularity of the modules, with finer granularity potentially raising the costs associated with integrating a large number of discrete pieces.

The early focus on these three properties has been confirmed by subsequent research on peer production in a wide range of domains.³⁷ When it comes to open educational texts, Benkler notes that modularity and granularity pose special barriers to peer production.³⁸ Educational texts are not as easy to modularize as other knowledge objects, like encyclopedias, where topics or entries are relatively discrete and free from the need for an overarching conceptual framework or unifying narrative–though this problem may vary to a degree between, say, a history text and a legal casebook. A tension also exists between making modules fine-grained enough to attract more contributors and ensuring the coherency of the text as a whole, which raises the costs of integration.³⁹ As Gary Matkin notes, more granularity in OER production can be an obstacle “because it destroys the learning context—the logical sequencing of carefully selected learning objects”.⁴⁰ Similar concerns have been raised in related text-based projects such as the online open library Project Gutenberg, where fine granularity in distributed proofreading tasks led to quality problems with the books as a whole.⁴¹

Overall, the theory described above predicts that open legal casebooks are unlikely candidates for peer production in most circumstances. That prediction seems consistent with the empirical data we presented above. One or a small group of law teachers might be motivated to produce an open casebook in much the same way and based on similar motivations applicable to other forms of scholarly output in academia. But more diverse and decentralized collaborative casebook projects run directly into problems of modularity, granularity and integration, ultimately leading them to be under-produced from a social perspective. There could, of course, be other reasons for the relatively slow pace of open casebook development by Canadian law teachers and the non-emergence of peer production in those casebooks in particular. For example, despite a growing number of new platforms available for open publication, access to suitable technology for the peer production of open casebooks has not caught up with the needs of authors. Authors may also have limited information about the options available. Some observers have expressed apprehension about the credibility and quality of open publications compared to those produced by conventional publishers.⁴² Others have noted a lack of institutional support and recognition for open casebooks).⁴³ Inertia and convention might also play a role, creating high switching costs for teachers that prevent them from adapting courses long reliant on traditional casebook offerings from the major publishers.⁴⁴ Despite these other possible explanations, however, the barriers posed by the need for modularity, granularity and low costs to integration still seem considerable.

At the same time, the standard story provides an incomplete explanation of the data on open casebooks in Canada for at least two reasons. First, while the problems of modularity, granularity and integration will be particularly acute for large-scale, highly-decentralized casebook projects where fine-grained modules are needed to attract a large number of contributors, it is not as clear that these dynamics are prohibitive for more modest-scale projects with a smaller number of contributors and certain degree of centralized administration. Second, because these problems have been well recognized in the OER context, several technical platforms have emerged over the past decade to start to address them, making it easier to achieve a certain level of modularity and reducing integration costs.⁴⁵ Research on the use of these platforms has begun to develop evaluative frameworks that emphasize a technology’s ability to address three interrelated challenges in the context of OER.⁴⁶

Voice and Exit in Casebook Development

Explaining the current status of open legal casebooks in Canada requires going beyond the conventional story about peer production and its particular emphasis on the costs of modularity, granularity and integration. Our goal in this section is to propose an extension of the conventional story that may help to explain why so few open casebooks are being produced collaboratively. Building on Albert Hirschman’s work, this story underscores what we call “voice” as a significant but underappreciated factor motivating law teachers to contribute to a collective casebook, along with the interrelationship between voice and possibilities for “exit” from the project in the face of irreconcilable disagreements or a lack of group consensus about core aspects of the project.⁴⁷

The concept of “voice” in open casebook production arises from the problem that most commercial casebooks are highly generic with respect to core questions of pedagogical design. To be sure, it is possible to glean certain choices by casebook authors about their conceptual approach to a course or about why they included certain materials based on overall organization, more or less explicit notes in the text, or perhaps an accompanying teaching guide. But the very constraints of market-based production under exclusive IP rights dictate that commercial casebooks, like commercial textbooks in general, must appeal to teachers with a broad range of preferences and perspectives.⁴⁸ While different casebooks may compete on different coverage of subject matter, doctrine or case selection, they rarely offer conceptual or pedagogical approaches directed explicitly to particular audiences. The way around this problem for most law teachers is to publish a detailed syllabus that curates content from an assigned casebook, supplemented by other materials available either online or by way of fair dealing exceptions to copyrighted material. For example, in a study of Canadian contract law professors, Sandomierski found that over one-third of those surveyed assigned supplementary materials in their courses. He argues that this shows a perception among a significant proportion of instructors that “[commercial] casebooks…focus on a dogmatic core at the expense of critical perspectives”.⁴⁹

The generic nature of most conventional casebooks readily masks the very real choices that law teachers necessarily make in designing and delivering their courses. Most, if not all, law teachers will likely have deeply-held beliefs about what material should be taught in their courses and about how that material should be addressed.⁵⁰ What often goes unnoticed in a world dominated by commercial casebooks is that most of those beliefs are only minimally and imperfectly expressed by the choice of which casebook to adopt. This point is of course well appreciated by those teachers who devote much time and effort to creating their own custom “course packs”. These teacher-authors implicitly recognize the value associated with having one’s voice–meaning the full expression of one’s personal, political and pedagogical goals and beliefs–clearly reflected in the materials that comprise the core of the courses they teach.

Emerging research on OER use in general supports this point. In a study of incentives and barriers to OER adoption in the United States, faculty cited the ability to adapt OER materials to suit their particular teaching needs as one perceived benefit of using OER.⁵¹ Another study found that the ability to customize and (re)design course materials can improve teachers’ pedagogy by shifting them “from serving as deliverers of publisher-designed content to embracing their own authority as curators of open content that they have discovered, organized, and/or created with specific needs in mind”.⁵² Similarly, a study of OER use in Columbia found that “teachers create and use OER more effectively” when they receive support which allows them to have “an active role in the creation and recreation of knowledge”.⁵³

Unfortunately, the costs of producing a legal casebook that fully expresses one’s voice but which rivals commercial offerings in quality and coverage imposes high costs that are likely insurmountable for many. Those costs highlight the potential of collaborative open casebook authorship and opportunities for participating authors to exercise a voice in casebook design and creation without bearing the full burden of developing the project on their own.

The concept of “voice” therefore captures an important motivation for participating in open casebook authorship beyond either material compensation or other forms of non-material reward described in the literature on peer production. We suggest that this idea has largely been overlooked by those who study peer production because of the emphasis on what Haythornthwaite has called “light-weight peer production”, in which relatively small contributions to a project are perceived as one-offs and therefore mostly preclude any deeper, sustained commitment to having one’s voice reflected in the final product.⁵⁴ By contrast, “heavy-weight peer production” occurs when group members invest significant time and resources in the success of the project on an ongoing basis.

If voice is indeed an important motivation for collaborative casebook authors, it must be understood in relationship to the relative costs of “exit” from a collective project, by which we mean the costs associated with ceasing to take part in the group collaboration after making some level of investment in the project. In Hirschman’s popular formulation, exit and voice serve as alternative but interrelated options for responding to dissatisfaction among members of a group (whether they are participants in a specific organization or project, or customers of a particular firm).⁵⁵ Exit and voice can serve as substitutes for one another: for example, when consumers have alternative options to either complain about poor product quality (voice) or to switch to a competing product (exit). But Hirschman went on to observe that these options can, in some circumstances, also be mutually reinforcing. When low-cost exit is available, firm managers may face greater incentives to respond to the needs of their consumers.

Translated to the context of collaborative casebook authorship, Hirschman’s insight implies that the costs of exit from a group project can dramatically influence the consequences of disagreement among members of the group. At one extreme, some casebook projects offer no opportunity for exit. This is, in effect, the situation with respect to conventional commercial casebooks. Co-authors of those casebooks cannot generally take their contributions to the project and leave in the face of any disagreement: they must either pre-commit on all issues of potential contention (assuming those could be known in advance), or they must resolve those issues by mutual consensus as they occur. This problem is diminished to some extent in the production of open casebooks where–by definition–the collective work product of the group is governed by an open license, such that any contributor who wishes to exit the group can take the casebook with them and continue to modify or update it as suits their needs and aims [Matthew Bodie⁵⁶, 74-5 refers to this as the “uniform/varigated” dichotomy in peer production; Bodie recognized that open legal casebooks might be produced without consensus about the final product but he does not address the problem of exit costs.]. However, while this mode of exit is costless at the point of exit, it raises increasing costs over time as the dissenting author is now in the position of having to update and maintain their particular version of the project without the help of other contributors. We might–recalling the significance of integration costs in research on peer production–call these disintegration costs facing the exiting author.⁵⁷

When such disintegration costs are too high it may be difficult to elicit participation from would-be contributors. If contributors are unable to exit the project in a way that allows them to continue capturing the value of their and others’ contributions in the future, they will face significant disincentives to participate in the first place. In this sense, as Hirschman observed, exit and voice are mutually reinforcing. While the nature of open casebooks means that exit costs among participants are lower than for conventional closed casebooks, these costs may still be high enough to deter active participation in peer-produced projects.

On the other hand, Hirschman also recognized that the costs of exit from a group might conceivably be too low. When exit cost are minimal–for example, when switching to a new product is relatively easy–contributors may have little incentive to exercise voice in the first place. This situation may characterize reasonably well Haythornthwaite’s category of light-weight peer production, in which contributors feel little attachment to the collective project, engage mainly on a one-off basis and bear little cost if they decide to stop participating, and therefore have minimal motivation to exercise or value voice in this context.

The theory we have sketched above extends the conventional story about when peer production may be difficult in the context of open casebooks and other forms of OER. Whereas the conventional story emphasizes the problems of modularity, granularity and integration as a set of “entry” conditions for participating in peer production, our perspective urges more attention to the “exit” conditions that face prospective contributors down the road, when collaborating authors inevitably confront disagreement and conflict in the face of strongly held beliefs about various aspects of law school pedagogy. At this point, we have said little about the variables that might influence these costs of exit and, reciprocally, the benefits of voice. We discuss some of these variables in Part III, where we explore the application of our theory in practice.

Part III: Consensus and Disloyalty in Open Casebooks

If we are right about the role of voice and exit in structuring incentives to peer produce open casebooks, is there a model of casebook development that can help to make exit and voice mutually reinforcing in this context? In other words, is there a way to reduce the costs of exit enough to circumvent the need for consensus, without lowering them so much that there is little incentive to sustain the collective project at all?

To answer this question we revisit the world of open source software (OSS) that has so often served as a archetype for peer production. But whereas the standard analysis of peer production has emphasized OSS mainly for its ability to overcome the challenges of modularity, granularity and integration, we turn to OSS in order to explore the ways in which software developers have sometimes achieved an equilibrium between voice and exit and thereby realised what we call “disloyalty” to certain projects that persist even in the absence of a consensus about some of its core components and aims.

Forking, Disloyalty and the Contemporary Web Browser

Web browsers–the software used to navigate the visual internet–are a ubiquitous part of everyday digital life. While many internet users probably give little thought to which browser they use, for those who care to dwell on this question there is a remarkable range of choice among the several free alternatives that currently dominate the browser market. Those different alternatives address a range of different priorities for users. Some popular browsers emphasize privacy, while others focus on speed. Still others promote their ability to integrate with other software applications or their flexibility toward user customization and experience.

The diversity in both design and the driving philosophies behind the internet’s most popular web browsers represents an undeniable lack of consensus among those who produce browser software about what constitutes a “good” product. This situation alone would be unremarkable if today’s leading browsers were simply competing products in the broader web browser market. But the process of producing this software does not resemble normal market production at all. Instead, nearly all popular web browsers today are what software developers call a “fork”–that is, an offshoot or branch of the central, collaborative open software initiative known as the “Chromium Projects.”

The Chromium Projects began in 2008, when Google released their new web browser called Chrome–and at the same time, publicly released all the code for the browser as an open source project called Chromium. At the time of this release, Google stated that they made the code open source both to allow other people to improve Chromium’s code and to facilitate the creation of other projects based on that code.⁵⁸ Since 2008, the Chromium source code has remained available to anyone to copy, remix and redistribute for free. And, as Google intended, Chromium was quickly used for other projects, becoming the base for most of Chrome’s leading web browser competitors including Opera, Brave, Vivaldi, and Microsoft Edge.⁵⁹ As of May 2022, Chromium-based browsers made up over two-thirds of the browser market share worldwide.⁶⁰

The structure of modern web browser development is therefore a prominent example of peer production, but one with an important twist. Chromium itself operates along the lines of “normal” OSS development, with a range of different software developers contributing to building, reviewing and updating the project’s source code. What makes the project significant is that, from a larger perspective, it encompasses a community of developers and users who both contribute to that core project (or “trunk”) while sustaining long-term dissent about some of its most important features in the form of a practice called “forking” (or “branching”). In the realm of open source software, forking occurs when a software’s source code is copied, modified, and then used as the basis for an alternative version of the original.⁶¹ The new fork “begins life as a copy…and moves on from there, generating its own history”.⁶² Because the code is open-source, creating a new fork does not require the original developer’s permission or consent.

The practice of forking effectively offers anyone who disagrees with certain features of an open source project the opportunity for low-cost exit. In the most extreme case, a developer might create a new fork and simply exit the original community of developers altogether–leaving them free to realize their divergent vision.⁶³ But in other cases, creating a fork is a means of only partial exit from the original project, with dissenters maintaining the ability to “update” their forks with subsequent changes and improvements made to the trunk, as well as the option to contribute their own innovations back to the original project. This process essentially describes the way that much of contemporary web browser development works.

By copying and then modifying Chromium’s code, different development teams create unique code forks. When Chromium’s code is updated, developers can integrate the updated code into their own fork to ensure the merged code runs smoothly and their own software stays current.⁶⁴ These forks result in different browser softwares and thus different user experiences: Vivaldi touts its full customization to users needs,⁶⁵ Microsoft Edge calls itself “the best browser for Windows”,⁶⁶ and Brave states that it’s “on a mission to protect your privacy online”.⁶⁷ The existence of multiple Chromium forks means there is no need for software developers (or users) to come to a consensus on what makes the “best” web browser. This model of peer production stands in stark contrast to other open projects—like Wikipedia–that resolve conflicts by building and enforcing group consensus.⁶⁸

What Chromium-based developers appear to have achieved, if imperfectly, is a model of disloyalty in which the option for exit serves to reinforce rather than detract from the expression of voice. Changes to the Chromium code base come from developers at Google but they are also contributed from individuals outside the company,⁶⁹ including by developers who work on other Chromium-based browsers. For instance, Microsoft (whose browser Microsoft Edge is a direct competitor for Google Chrome) accounted for 35% of non-Google contributions to Chromium in 2020.⁷⁰ Vivaldi developers chose to use Chromium’s code base in part because of these outside contributions, which allowed them to “benefit from code improvements many developers–including ourselves–report and upstream [i.e., contribute back to the Chromium’s code]”.⁷¹

Practicing Disloyalty in Open Casebooks

A key question for law teachers interested in the future of “open” is whether the model of disloyalty in web browsers that we have just described can be applied or adapted to build open casebooks collaboratively. We think that it can. In this section, we move beyond our theoretical discussion to describe how we are developing the disloyalty model in practice to build and revise an open-source property law casebook for use in Canadian first-year law school courses, entitled Land and Property in Canada’s Political Economy.⁷²

The motivation for one of us (Baxter) to begin writing an open property law casebook is well captured by the concept of voice described above. Several commercial casebooks on Canadian property law are comprehensive (at least with respect to Anglo-Canadian legal sources), but each lacks a substantial and explicit conceptual foundation that makes it possible for students to develop critical understandings of the material in a comprehensive way. Baxter’s research interests in both the history and the political economy of legal institutions motivated a first experimental version of Land and Property, created during a sabbatical year, for use in his first-year property course in 2020-21.⁷³ The result was a product of sole authorship, not peer production. A subsequent version of the casebook, now under development, is being redesigned by the authors of this paper explicitly to address the challenges and possibilities for collaborative development we previously described. Among other things, this redesign aims to ensure that anyone can quickly and easily edit, add, remove, or reorder the casebook’s modularized content to produce their own version of Land and Property by creating a fork of the original. These forks could produce versions in which the difference is as small as a reworded sentence or a new case, or as large as an entire chapter–all while sustaining the possibility of integrating future changes to the original into the new fork, or from the fork back into the original.

Whether or not the Land and Property project will ultimately attract other collaborators among those who teach property law in Canada and achieve a form of sustainable commons-based peer production like that seen in modern web browsers remains to be seen. What the project offers at this stage is not a successful example of collaborative development, but an evolving assemblage of ideas and tools for practicing disloyalty in casebook design. In this section, we will defer a thick description of the technical details of that project to a separate article and focus on briefly explaining why the ideas and tools we are adopting make the casebook uniquely positioned to support collaborative authorship in line with the theory described in Part II.

The central approach behind the Land and Property project can be summed up in the adage that we advocate writing “casebooks as code”. For the most part, this means that we have abandoned the longstanding practice of drawing high-level analogies between OER production and open software development. Instead, we have started to adopt more of the actual, everyday practices and technologies of software developers themselves.

This idea of law professors with no experience in the computer sciences writing “code” is not as far-fetched as it might sound. In fact, it mirrors a broader trend among non-programmer writers within the OSS sector itself who produce technical documentation for software projects (such as the FAQs and help files that explain to users how to use a particular program). For most of computing history, these “docs” were created using standard tools and processes for technical documentation, such that manuals for open software were written in much the same way as manuals for toasters. But over the last decade or so, a small group of writers began to realize that they could employ some of the very same tools and processes in their work as those used by their programmer colleagues who develop the software itself. This approach to authoring documentation became known as writing “docs as code” and has led to a flourishing ecosystem of collaborative technical authorship.⁷⁴

The details of these tools and processes adopted by technical writers are beyond the scope of this paper, but in general they rest on two foundations of all contemporary OSS development that make the disloyalty model we described above achievable in practice. The first is writing in source code. Source code is essentially a set of instructions and associated commentary written by programmers in one of several possible text-based computer languages. All collaborations between OSS developers happen at the level of source code, its most important feature being that it is understandable not only by a computer (which “compiles” the code into a working application), but also by any human being who is proficient in the language.⁷⁵ Source code is a unified means of communication, within a computing system, between all participants in that system, both human and machine.

Casebook “source code” looks much different than computer code (and much less confusing to the lay reader). The Land and Property casebook is essentially a collection of plain text files, with minimal markup, that anyone can read and understand having learned a few basic rules. This approach is strictly content-focused: casebook authors work directly on the source code content but do not need to worry about formatting or design. The source code is then “compiled”, at a later stage, for output to particular media (e.g., digital, print, or both).

In creating Land and Property, we therefore abandoned now-standard practices of working with visual website builders or conventional desktop word processing applications and returned to the basics of writing text. While this will appear to many a banal, technical detail, it is difficult to overstate the importance of such a shift for realizing our model of disloyalty. The move to writing casebook source code makes possible the second foundational element of “casebooks as code”: source control. Those who have been involved in multi-author writing projects will be familiar with the process of co-producing a document using “track changes” in Microsoft Word or Google Docs. While useful for limited works with a small number of authors, these tools are clearly inadequate for managing complex projects with many decentralized contributors and offer no way to sustain the durable “forks” that are essential to the disloyalty model used in modern web browsers. OSS developer communities working with computer source code have created vastly more sophisticated systems for version control that make Chromium–and indeed, all open source software–possible. These systems enable a mode of collaboration facilitated by a multi-versional representation of content as a range of snapshots. These snapshots are updated every time a user saves a different state of source code".⁷⁶ Imagine, in other words, that every change to a casebook–no matter how small–is captured and remembered by the system, making it possible to branch off from the existing content at any point in its history.

Because casebook and computer source code share the essential feature of being text-based, the same system for version control can–in theory–be used to manage either type of complex collaboration. Most current platforms available for open casebook production store information in inaccessible databases that make sharing and remixing difficult or nearly impossible. The casebooks-as-code approach dramatically simplifies how this information is stored and thereby makes accessing and using it for real collaboration eminently more feasible.

To illustrate, consider a brief example. Suppose that a dozen property law professors across the country express an interest in developing the original content of the Land and Property. Using the system of version control described above, each contributor could make a copy of the existing content, make changes or additions, then suggest those back to original. Such an exercise would of course require some decision-making or governance structure to determine which changes to accept or reject. On its own, this process might be manageable using conventional tools for casebook authorship (though likely not as the number of participants increases). But now imagine that one of the contributors, who’s suggested change has been rejected by the group, decides to create their own fork of the Land and Property project. As with modern web browsers, the system for version control that we have described enables this dissenter to pursue their own changes, while sustaining the ability to automatically “update” their version with subsequent changes to other aspects of the original. The disloyal fork carves out its own distinctive identity from group consensus, while maintaining an ongoing relationship as that group consensus itself evolves.⁷⁷

None of this is to suggest that law professors interested in open authorship will or should need to go to great lengths to learn new tools and processes for creating their casebooks. Indeed, the spectre of such high fixed costs to participation would surely outweigh any benefits of voice or otherwise. Reducing those costs has been a significant feature of our current work on the Land and Property project. But we are comforted by the fact that our optimism about the potential for casebooks as code is shared by technologists who think broadly about the future of content generation and management on the web as a whole:⁷⁸

Imagine going to a website and spotting a typo, clicking an edit button, fixing the issue, and submitting the change to the website owners for approval. Or to take it a step further: the same workflow, but you’re writing a guest blog post or generating entirely new pages of content that you feel is missing. These workflows won’t just transform content creation on the web. They can also apply to writing a book, academic writing, print media … the list goes on.

We imagine a future where multiple, constantly-evolving forks of collaborative casebook projects exist, created by law professors across the country (or around the world) who have the tools to continually adapt their versions to suit their unique voices, all while continuing to contribute to the core project. At the same time, we acknowledge that the practical tools to support such collaborations are still in their early days. Working at the edge of technological and conceptual development in this area comes with a certain level of uncertainty and it will not be for everyone, at least for now. But for some, the potential both for collective action and for realizing the voices of diverse authors–aligned with the increased affordability and accessibility promised by open casebooks–will be well worth the experimental effort, with all that this entails for a more equitable distribution of power and resources in legal education.

\pagebreak