Are Digital Knowledge Commons Real Commons?

2.1 Proximity Between Peer-to-Peer Production and “Digital Commons”

To the best of our knowledge, the first scholars to discuss these digital knowledge commons were close to the fields of law, science, and technology (For a presentation of these early works, see Benkler, 2017; Benkler et al., 2015). They saw the phenomenon as a way for actors to regain control of innovation and its collective management, or “peer production” of innovation (ibid). The initial question was how intellectual property rights, which were seen as preventing users from accessing innovations such as software source code, could instead be used to “create opportunities for individuals to draw on resources without connections, permissions, or access granted by others” (Lessig, 2002, p. 85).

This is why the Free Software movement has been created by Richard Stallman (see his original manifesto), as the Foundation he initiated, the Free Software Foundation (See the FSF’s about page). Lessig has been a close companion of that movement, in addition to being one of the creators of the Creative Commons licenses (the Creative Commons website details L. Lessig’s role in that creation). Research later emphasized that the resources are managed by a group in a radically decentralized, collaborative and non-proprietary way (Kapczynski, 2010, p. 35). This field of research, with its focus on the individuals, has helped better understand why actors participate in these projects and how they organize to do so.

For instance, the open innovation research field, analyzed in terms of value, incentives, and opportunities (Chesbrough, 2023; Trzeciak et al., 2022), symbolized by the “user as innovator” model (von Hippel, 2002), put forward a vision of autonomous actors who cooperate because they are incentivized to do so, primarily because the resulting innovation benefits them or is less costly to produce.

Rapidly, these studies invalidated its initial claim (Benkler & Nissenbaum, 2006) that the cooperation between individual actors means the absence of a hierarchical organization. There is a hierarchy in these projects, and the artifact, the digital infrastructure helps built it (Shaw & Hill, 2014).

However, they did not make the connection with other commons, which would have helped consolidate the field of study on collectively managed organizations. This discussion is at the heart of a second group of researchers who are closer to unorthodox economics or social sciences, such as Bollier and Helfrich (2012), Dardot and Laval (2019), and Pazaitis et al. (2022).

They aim to study whether the commons, as collective movements, propose new social constructs capable of changing our social organization and transcending the current state-market duopoly in producing collective actions and public goods, as well as managing rival resources. The success of digital collective action, coupled with the failure of traditional institutions to protect shared natural resources (air, climate, fisheries, etc.), makes such an agenda necessary.

From this perspective, digital collective projects seem to provide an answer to the theory of the knowledge economy (Foray, 2004) which has long emphasized the limits of the market’s and the state’s capacity to produce accessible and useful knowledge goods.

More fundamentally, as explained by Bollier and Helfrich (2019), Caffentzis and Federici (2014), and Terranova (2022), studying these collectives allows us to discuss how people can build cooperation. This approach challenges the dominant model of non-cooperative Homo economicus, as presented in classical economic theory and the capitalist model. It enables the study and the conceptualization of alternative social models (Bollier & Helfrich, 2019) and to propose an alternative to the digital world dominated by corporate platforms (Terranova, 2022).

Let’s put it right: the goal is to study where, and how this new organizational system can be such an alternative, not to claim that it can always replace these institutions, as in the study of classic commons. E. Ostrom’s analysis has consistently emphasized that classic commons are organizations that interact with other organizations, especially classic authorities. Her “principles,” or “rules,” are summarized in 8 main “rules” for managing the commons; Her fourth principle states, “Make sure the rule-making rights of community members are respected by outside authorities.” And the digital collective projects, as stressed by the first group or research mentioned: FOSS projects are deeply embedded in the market (Fitzgerald, 2006) or in public (university) production, which has initiated many FOSS projects that were initially considered a specific type of academic research project (Bezroukov, 1999).

From this perspective, a pertinent question regarding online collective knowledge production is whether these groups adhere to similar organizational structures and cooperative principles to establish shared control over resources that benefit both individuals and groups, to paraphrase Bollier and Helfrich (2019, pp. 69–72).

Research has shown that the governance structures of these projects are very similar to those of classic commons.⁴ In terms of the selection of participants, control (as decentralized as possible), hierarchical organization or system of sanctions, they echo Ostrom’s basic principles of such collective regulation (the already mentioned 8 main “rules” for managing the commons). As exemplified in a study of FOSS projects by Schweik and English (2013), this lens helps us to better understand how digital knowledge production projects organize themselves.

We would have a governance organization for these projects of online collective knowledge production that would be close to that found in classic commons. If so, is it enough for an organization to be governed in the same way as traditional commons in order to be called a commons? Is everything that is collectively managed a commons, even exclusive goods? Is FOSS controlled and developed solely by a company a commons? Is software developed and used only by a club of actors a commons? Is Facebook a commons, as some sub-group organize themselves on this platform? This is the argument defended by Mindel et al. (2018) with their concept of “information commons”. This last article illustrates the idea that a new, better model for meeting collective needs in a more collective way has emerged and should become dominant (for an example of such a discourse, see Rifkin, 2014). However, it does not convince us of the value in labeling everything a commons or of the soundness of the analysis. It ignores, for example, the fact that a commons must have collective governance at its core.

The proposition to name some digital collective projects “commons”, is an “analogical gesture […] has found scientific legitimacy in the work of economists, jurists, and political scientists, the most famous of whom is undoubtedly Elinor Ostrom”* (Laval, 2016). We would add “some scientific legitimacy” and argue that more research is necessary to improve our understanding of how people can establish and maintain cooperation online. Framing the question as such make us closer to institutional economics, from a theoretical point of view, but also to the first line of research, as the former’s results can be mobilized by the later to discuss if those structures of organization are similar with those found in other commons, and can strengthen a theory of the commons. Following Laval’s suggestion, we propose revisiting the origins of these works: Ostrom’s discussion of the definition of collective action commons and the challenges posed to this definition by digital collective actions.

2.2 What is a Commons?

The analyses we cited ground their approach to the commons in theoretical debates about collective actions and human nature.

At the heart of the debate and the interest of the commons are two elements pointed out by classical economics and summarized by Olson (1965). First, in alignment with the homo-economicus model, Olson emphasized the tendency of individuals to avoid participating in collective efforts if they can get away with it, a phenomenon known as “free-riding.” Second, for a collective project to work a mechanism must exist to ensure that everyone contributes to the effort, either by contributing or by being compensated to do so. Consequently, he stressed that participants must be able to monitor each other so that defection, or “free-riding,” can be quickly detected and punished, thus rendering it ineffective. He also claimed that the larger the group, the more difficult monitoring would be.

Ostrom’s work proved Olson wrong in the sense that she showed that certain types of resources did not require such mechanisms of compensation to be collectively managed. “Attributes of the goods produced and allocated, as well as the rules used for their production and allocation, affect the different incentives faced by participants. The first attribute that scholars should examine is whether the cost of excluding beneficiaries is high or low. If exclusion is costly, potential beneficiaries face a collective action problem. Second, it is necessary to determine whether consumption is subtractive or not. Potential beneficiaries whose consumption is subtractive face a common pool resource type collective action problem” (Ostrom, 2003, p. 248).

However, what Olson said about the control and the size of the group remained. Ostrom demonstrated that, in order to sustain these collective, cooperative actions, it is necessary to define rules that specify what can be done and by whom, as well as to monitor people’s activities. In physical commons, these rules and the capacity to control others’ activities stem from laws (e.g., property rights), usage norms, practices, and technology control (e.g., barriers and monitoring systems). These institutional arrangements establish common ground by creating bundles of rights and duties for those who exploit the resource and control others’ actions. These “bundles of rights” establish roles and a hierarchy among participants (Schlager & Ostrom, 1992), giving some participants the ability to control what others do and even sanction them in case of wrongdoing. Second is the size of the group. Ostrom showed that these bundles of rights allow small groups to effectively organize cooperation and regulatory mechanisms. For example, if a large population of users has the right to enjoy a resource, such as walking in a park or forest, then the privilege to consume the resource, such as cutting wood, belongs to a smaller group of consumers. Monitoring consumption rights and issuing sanctions are additional rights granted only to a subsection of consumers: the “policymakers.” This small group’s ability to effectively monitor consumers’ actions is essential to the sustainability of the commons (Ostrom, 2009). Ostrom’s designation of these groups as “privileged” aligns with Olson’s argument that the larger the group, the more difficult it is to sustainably manage a substractive resource collectively.

Why these results? In the management of subtractive resources, it is true that individual interests can work against collective interests. Consumption is sustainable if a group defines a level of resource consumption and distributes it among its members, and if everyone respects their allocation. However, even a small amount of overuse of the authorized withdrawal by one person can have disastrous consequences for the group as a whole. It may be in one’s own interest to exceed the limit because one can earn more in the short term. However, this does not mean that financial mechanisms are the most effective. Departing from the standard model of a rational and selfish individual, one strand of economic literature has, in connection with social and evolutionary biologists, suggested that prosocial behaviors have evolved to sustain cooperation, a key feature of the survival of the human species (Gächter, 2014; Gintis et al., 2003; Seabright, 2012). Established in long term relationships, these cooperative behaviors often become norms. Commons may be a proof of, and a model for such cooperative organizations (see, for a development Bollier & Helfrich, 2019; Caffentzis & Federici, 2014; Chatterton & Pusey, 2020).

Even without accepting this argument, the commons’ situation differs from Olson’s concept of “free-riding,” which is using something without contributing or making an effort. In the commons, people exert effort (e.g., cutting down a tree or fishing), and problems arise when they exert too much effort. In short, in response to neoclassical economics in general and to Olson in particular, managing the commons means defining maximum effort levels for accessing subtractive resources and preventing overuse through collective governance. It does not mean identifying those who contribute nothing and forcing them to act.

This rapid theoretical detour helps to better frame the challenges of qualifying digital knowledge production and collective actions as commons.

2.3 What is the substractive resource in the digital commons?

At first glance, the so-called “digital knowledge commons” do not seem to fit the description of the collective management of a substractive resource. While preserving and perpetuating physical resources is understandable, pieces of knowledge can be shared without substractivity and at low cost with digital technology. Non-exclusivity and non-substractivity are the defining characteristics of what economists call a “public good.” According to classic economic theory, users of public goods have no individual incentive to produce these goods (Foray, 2004). This is similar to Olson’s free riding problem, in which the risk is that not enough people produce knowledge while waiting for others to do so. However, it differs from Ostrom’s commons situation, which concerns preventing people from doing too much.

In the first case, this would reinforce the idea that there is a fairly constant model of collective governance of a collective action, the wider application of which would be beneficial to society (for a defense of such an approach, see Dulong de Rosnay & Stalder, 2020). The risk is that commons would become a loose definition of any collectively managed collective action, failing to capture the specificity of such organization and favoring implementation where it does not work.

In the second case, there is a need to better understand the effort required to access a substractive resource that must be collectively governed. This may reduce the scope of the digital commons because the effort may not always be or may not be collectively governed. However, it may also provide a more solid basis for understanding the nature of the commons in terms of the governance framework for collective action and the nature of the collective (community) that governs it.

We will defend the latter view by demonstrating that there is no free-riding problem between users and producers (contributors) of knowledge in online, collective knowledge production projects. In Section 3, we will generalize the argument of Sen et al. (2022), who claimed in their introduction that the FOSS system manages a common-pool resource –the development effort to provide a software solution – and that the resulting solution is a public good. However, the size of the group may be another difference that can be explained. The largest digital projects, such as Linux and Wikipedia, coordinate hundreds or even thousands of contributors who may only be identified by a pseudonym that can be changed. Section 4 will discuss this point alongside other differences and similarities between digital and classic knowledge commons.

3.1 The access to scientific publication as a subtractive resource

In the “Republic of Science,” Polanyi (1962) claimed that the scientific validation of research requires its acceptance by a journal and subsequent endorsement by the scientific community (p. 5). The written trace formalizes the scientific proposal (Latour, 1987), while concurrently asserting a “property” on that discovery (Merton, 1957). Scientists must publish in journals, and the more visible the journal is the more impacting is the acceptance for a scientist.

However, due to the scarcity of reviewers, a journal can only publish a limited number of articles per year. In the past, limited paper space was also an issue. But, again today, what they propose must be different from what has already been published, while conforming to the journal’s rules regarding the length and scope of articles. Several journals may be in the same scientific field, and access to the limited pool of reviewers may be more or less difficult for a given journal. Conversely, access to publication may be more or less difficult for contributors. Managing this access is at the core of editing work. Prominent journals may have fewer problems finding reviewers, but the more visible they are, the more appealing they are to contributors, and the greater the volume of contributions they must manage. Finally, sitting on one of the journal’s committees (e.g., the editorial board or editorial office) and being responsible for a subsection or journal means controlling its editorial line and, thus, the evolution of science. Thus, governing a scientific journal provides specific rewards. In return, it increases the need to regulate access to these positions.

Note that we are considering a collectively managed scientific journal here, where an editorial board usually decides what to publish. Not all scientific journals work this way. We will come back to this later, but for now, it’s important to note that not all scientific journals can be considered collectively managed knowledge production projects.

The following elements are found in online knowledge production collective projects: access to a publication space and the peer-review system; the necessity to manage access because both resources are subtractive; and a system of control that provides extra powers and benefits.

3.2 Why do contributors contribute?

Online knowledge-producing projects are intended to organize the production of knowledge, or more precisely, the “codification” of knowledge elements (software functionalities, encyclopedia articles, scientific articles…) articulated between them. The more the “stock” of knowledge elements grows, the richer and more interesting it becomes for users. However, as research on the knowledge economy has shown (Foray, 2004), the production of new knowledge –which we will call the “flow”– is constantly needed to improve and update the stock, and even to maintain it this stock as needs, scientific knowledge, and technologies evolve.

In the classic knowledge economy, producers invest in producing the flow because they anticipate the remuneration they will receive from charging a toll for access to the stock. Why are contributors willing to contribute to the flow of knowledge without being compensated by the access toll?

There is a rich literature on the subject, albeit a bit old, probably because the findings are quite solid and consistent and thus less of a research question: they contribute because they have direct benefits.

In software development projects, the intellectual challenge of solving complex problems and comparing oneself with peers, as well as the need to modify the software to meet one’s own needs, explain the motivation to contribute (Hertel et al., 2003; Shah, 2006; Zhou & Mockus, 2014). Actually, for a long time, the majority of participants in FOSS projects have been software professional and an important part have been paid to contribute for their employers (Lakhani & Wolf, 2005). In addition to the intellectual challenge, motivations for contributing to Wikipedia include accessing peers to deepen one’s knowledge or having the opportunity to publish work on such a popular platform (Bateman et al., 2011; Crowston & Fagnot, 2018; Forte & Bruckman, 2008; Kamata et al., 2010). Online forums provide access to expertise, and a platform for organizing one’s knowledge, in addition to intellectual challenges. Sometimes, the visibility in an arena related to a potential professional market (i.e., a community of professional practices) is also valued (Bateman et al., 2011).

Finally, it appears that attachment to the project (Crowston & Fagnot, 2018) and social interactions (Butler et al., 2007) –something akin to what Terranova (2012) called “paying attention to the other”– are at the heart of the motivation of the most active and longest-standing participants, especially those involved in regulating contributions.

In other words, and aligned with Deci and Ryan (2002)’s model, intrinsic motivations (effectiveness, challenging and stimulating activities, and peer recognition) are sufficient to encourage contributions, though these motivations may differ depending on the task and the individual.

In short, for the projects where there are several contributors, the problem is not so much encouraging people to contribute as it is encouraging them to contribute well, and coordinating their efforts to maintain and develop the stock of knowledge. It is not a free-rider problem but an effort regulation problem. We will now show that the effort is made to access a substractive resource, and that this access is managed by a hierarchy.

3.3 Contribution as the managed subtractive resource by a hierarchy

Seeing one’s work being published, but also the access to evaluation of one’s work, and to social interactions are accessing to a substractive resource, the attention (Goldhaber, 1997), as explained by (Terranova, 2012).

First, there is competition for publication space. In Wikipedia, topics and article names are unique, as are package or software layouts in the Debian operating system distribution. Wikipedia has only one article for the title “World War II,” even though more specialized articles on related subjects may exist elsewhere. Similarly, there is only one “Firefox package” in Debian. While fledgling projects may have difficulty attracting contributors, successful projects have difficulty managing them and arbitrating between their conflicting interests (Greenstein et al., 2016). Some of the questions that must be addressed include who should respond in a forum and which edits or commits should be accepted. Dealing with this regulations requires the access to a second subtractive resource: the evaluators’ time, which is limited at any given moment. In Wikipedia and other FOSS projects, the project must constantly assess the adherence of contributions to publication rules to ensure they improve the body of knowledge. In forums, questions must be managed so experts can focus on unanswered questions and those presenting a greater intellectual challenge.

To do so, all projects have contribution regulation structures (Carillo et al., 2017; Halfaker et al., 2013; Steinmacher et al., 2015). In FOSS projects, the code, but also these systems of validation of a contribution, creates different layers of control over the contributions and over the project (Palazzi et al., 2019; Shaikh & Henfridsson, 2017).

Wikipedia is less explicitly structured, as there is nobody who validates each contribution (or “edit”). Mechanisms of control do exist, by the rules, but also the activity, and the technical tools (Safner, 2016). Rules regarding what is acceptable have increased over time, making harder for contributors to see their work last on the platform (Butler et al., 2008). Ownership practices, i.e. the monitoring of an article and the systematic revert of edits by a contributor, have been identified (Halfaker et al., 2009). These rules are also enforced by (elected) contributors with special powers, such as the administrators (Liang et al., 2008), who can forbid new edits on an article, or ban contributors, or delete an article among other powers.⁵ Some contributors specialized in what is called “Recent changes patrol”, to track “bad edits”. Finally, automate tools, developed by certain contributors and validated by seasoned contributors, named “bots”, help all these contributors with special powers to supervise all the contributions, making Wikipedia an algorithmically managed project (Clément & Guitton, 2015; Geiger, 2014; Zheng et al., 2019).

As said in the review of literature, these activities, rules, and tools build hierarchical roles (Table 1).

3.4 Synthesis: online knowledge production collective projects are commons

In short, online collective projects for producing knowledge build a shared body of knowledge, or a “stock of knowledge,” which is maintained and enriched through contributions from members in the form of a flow of codified submissions. The project must coordinate this process by filtering and regulating these contributions. Due to limited publication space and validation time, it is effortful to access this subtractive resources. The regulation is carried out by a subset of members who have the specific right to do so. These rights are granted by law, the licenses chosen by the project initiators, participants’ activity, and the project’s rules. These rights are also enforced by software systems connected to the digital platform hosting the project.

All these elements are those of a commons –that is, “resources that are collectively managed by a community that establishes rules and governance with the aim of preserving and perpetuating these resources.”*⁶ We will only emphasize that the “substractiveness” is implied by the terms “preservation” and especially “perpetuation”.

The shift in regulation between physical and digital commons, from the right to access a natural resource to the right to access publication spaces and peers, is due to a change in consumption patterns. However, this is still a form of regulation that controls people’s access to a subtractive resource (what we call the knowledge flow production), in order to preserve the quality of the global resource (what we call the knowledge stock) for all.

As for the classic commons, the precise implementation of collective governance varies from project to project. However, they all establish sets of rights that determine what each person has access to and the power that some have over others’ actions.

One discrepancy seems to remain: the size of the group. How do digital knowledge commons succeed in managing large groups? We will argue that digital tools explain what is unique about digital knowledge commons. Now that we have established the similarities, these specificities are easier to describe.

4.1 Initiating a Digital Collective Action

A commons is based on a group of people with a shared interest, something facilitated by digital technologies (Alstyne & Brynjolfsson, 1996). For example, it is easier to reach a group of potential contributors interested in a software project (Schweik & English, 2013). Linus Torvalds did just that when he initiated Linux and recruited contributors from a Usenet forum list.

The incremental aspect of such projects also helps (Marwell & Oliver, 1993). Each Wikipedia article and each forum question-and-answer thread has value in itself, as does each scientific article. The project does not need to reach a certain level of knowledge before it can start offering value. Therefore, it is not necessary to convince a certain number of contributors in advance, raise a certain amount of money, or wait for the “complete” realization of the project (as with a bridge, for example). The most interested contributors begin, and as the project grows, it attracts new contributors, generating an increasing return on adoption. The situation is slightly different for software that requires a minimum level of functionality to be usable. However, this suits the user-developers behind FLOSS projects, who are willing to collaborate to develop a new product that better suits their needs (von Hippel & von Krogh, 2003).

According to Marwell and Oliver (1993) model, if the project succeeds in attracting the initial “adopters,” it enters an expansion phase. As a consequence, as previously explained, access to the contribution space, i.e., the ability to consume it, can become increasingly competitive, thereby reinforcing the need for regulatory systems. Digital has also developed original responses to this challenge.

4.2 The Architecture of participation and control

The general project can be divided into subprojects, such as articles, portals, software files, and discussion forums. This modularity has several advantages (Benkler, 2002). It makes it easier for new contributors to join (Bessen, 2006), since they can focus on the area that interests them the most and that is the right level of difficulty for them.⁷ This “participation architecture” (MacCormack et al., 2006) also makes control easier, as illustrated by Table 1. Each module can have a person in charge of controlling the contributors, as close as possible to the task (Forte et al. (2009), in the case of Wikipedia, O’Mahony (2007) in the case of Free Software). This is not unique to digital commons. Indeed, Ostrom’s eighth design principle for sustainable organizations states that “appropriation, provision, monitoring, enforcement, conflict resolution, and governance activities are organized in multiple layers of nested enterprises” (Ostrom, 1999, p. 7). Our claim, based on Bessen (2006) and MacCormack et al. (2006), is that digital elements facilitate the construction of these nested enterprises by promoting modularity.

More specific to the digital world is the a posteriori regulation of subtractive resource access. In order to benefit from feedback from peers and ultimately from contributions, one must make the effort to codify and propose the flow of knowledge one wishes to see accepted (see the piece of knowledge incorporated into the existing stock). In traditional commons, individuals with the right to consume the resource are selected a priori.

Because production is digitally codified, contributions can be monitored and various sub-projects can be coordinated in digital commons (i.e., the bots we spoke about) in addition to the infrastructure that structures how people contribute. Examples of this infrastructure include submission platforms for scientific journals and Wikipedia’s MediaWiki content management system. Carlsen et al. (2020), inspired by Foucault’s theory 1982, named this “power-over” (in this case, policymakers over contributors).

This algorithmic control is probably an organizational necessity in terms of the volume of work (van Knippenberg et al., 2015). This, and the modular organization explain how these projects are able overcome the problem of managing large groups. This approach has some drawbacks, but also advantages for contributors and project governance, i.e., the connection and dialogue between contributors and policymakers.

4.3 Connecting Policymakers and Contributors

The main drawback of automated management is that it is probably more difficult than in traditional commons to understand how to access the contribution. As explained by de Vaujany et al. (2018), digital management generally makes it more difficult for contributors to determine which rules they must follow for their contributions to be accepted and which rules have been infringed upon when a contribution is refused. This issue has been highlighted in the context of Wikipedia and FOSS (see, for example, Halfaker et al., 2011; Steinmacher et al., 2015).

However, there is another side to the story. For contributors, interacting with a machine means they can learn the rules through trial and error, which reduces the risk of being stigmatized or excluded by angry moderators who see the same mistake repeated. For this reason, online knowledge production may be more welcoming to a diversity of profiles than other commons projects, in which contributors are selected a priori,⁸ but, again, at the cost of a large effort to become and remain a contributor.

Another difference from a privately managed platform, but also from classic commons is that contributors who learn and respect the rules gain recognition as valued participants and may gain access to policymaking. This opportunity is open to anyone. In the other hand, an individual’s position in a group may also be more “contestable” than in classic commons because it is subject to constant negotiation or renegotiation and is based on performance. This position can be stabilized in social roles, referring to models specific to a function or position in a community (Rocher, 2012, p. 43). However, one can benefit from one’s status only by taking action. For example, one is a moderator only if one moderates; a person responsible for a file in software development must manage contributions to remain responsible. The acquired position partly protects individual contributors but also constrains them since renown is a social process that generates “reputation control” (Shinn & Ragouet, 2005): this individual reputation is under collective supervision for the collective interests of the project.

A consequence is that entries, but also exists, are more difficult to define in this digital (knowledge) commons. Hence, the contours of Ostrom’s “community” of policymakers are also more difficult to define, both for the contributors and the observer.

The traditional definition of a community (Bowles & Gintis, 2002) as a closed group that interacts with the outside world —with a boundary between those who participate and are recognized within it and those who are not— imperfectly describes these social arrangements.

As far as online communities are concerned, research has shown that people define themselves as belonging to a “community” when they adhere to the project’s goal, something they acquire by interacting with their fellow contributors, and when they are recognized as such by their fellow contributors (in addition to the articles already mentioned in Section 3, see Pentzold, 2011, on Wikipedia). This aligns with the definition of Georgiou and Arenas (2023) of “a logic of creating and maintaining boundaries within which members form ’thick’ relationships and work together in pursuit of interdependent goals.” It echoes what Follett (1919) defined a long time ago as “power to do with” activities that foster a sense of belonging: the discussions about the project and its rules, creates. This also echoes the concept of the “group” by Mullins (1972) or the “reference group” by Merton (1957), which emphasizes that the “community” is perceived by the members, not necessarily by outsiders.

In this sense, participants in power-with activities can be analyzed as forming a “community,” i.e., a closed group that interacts with the outside world. The boundary between this group and the “simple” contributors is as much a boundary of knowledge —knowing that such a group exists— as it is classic recognition of membership by others. The precise mechanisms of this concrete process of recognition and alignment warrant further study. In the conclusion, we will argue that our definition of digital knowledge commons and their communities helps facilitate such studies.

5.1 What a digital knowledge commons is

Making knowledge freely available, for example under a free license, is thus not enough to make it a commons. There has to be collective management of that production. For example, the MySQL project, where a single actor, the company MySQL-Oracle, has the right to regulate the contribution, is not a commons. Crowdsourcing platforms, collective innovation, or question-and-answer forums run by a private company, such as Quora or Stack Overflow, are also not commons because there is no collective governance of access management to the substractive resource: the filtering policy is centralized by the owner of the platform.

Pooling and managing resources within a close group is not enough either. It is the management of a non-substractive but excludable good, a “club good” (swimming pools, theaters…) Innovation groups, named by Potts (2018) “innovative commons”, is such example that is, according to our analysis, misnamed as a “commons”. Other examples of the collectively management of a club good are 3D printers for hackerspaces, or scientific equipment, or, in the digital realm, MOOCs and guilds in online video games.

This does not mean that the commons organization have no role to play in the production or negotiation of innovation. For example, some standardization systems or institutions seem to fit the definition: on the one hand, there is a clear interest for some actors to participate in these systems in order to promote their technology, and the published standard is accessible to all; on the other hand, standards management can be collective, for example for the Internet, with the Internet Engineering Task Force IETF).⁹

However, naming them as commons requires identifying both the collectively managed subtractive resource and the public resource produced for all users. Becoming a commons is a choice, meaning it is opened up to others, but it is also an indicator of a project’s success. A project becomes a commons because it succeeds in attracting contributors. Finally, being a commons once does not guarantee that it will always remain so. A project is built through the negotiation of rules, collective discussion, and the continuous evolution of these rules (power with over power over). The set of rules of a commons is constructed through a socio-technical environment, which can also evolve as techniques, incentive systems, and the legal environment evolve. Problems with governance can also lead to the disappearance of collective governance. For instance, the algorithmic management that seems essential for major projects also has the consequence of creating a boundary between users and consumers (contributors), as well as between consumers and policymakers.

This highlights the fundamentally dynamic nature of digital knowledge commons: in the resource they manage, the flow of knowledge, and in the status of the participants. While this dynamic also exists in the resource managed in traditional commons, it may be less applicable to the evolution of the group of contributors and those who benefit from the commons, which Ostrom referred to as the “community.” Here, no one is excluded from consumption a priori, even though the human and technical selection of contributions becomes more difficult as the project becomes more successful. Belonging to the community seems to be a journey or “career,” as Becker (1963) defined it, where recognition and achievement of the project’s goals and value are closely intertwined. The values and rules of projects are learned through contribution because contributors learn them by interacting with other participants (Dang Nguyen et al., 2017).

This journey has been studied for a long time in other contexts (e.g. Borzillo et al., 2011, regarding communities of practice within firms), as well as in some digital commons (e.g., for Wikipedia Arazy et al., 2015; Rullani & Haefliger, 2013, for free open source projects). However, an overarching theory has yet to be proposed.

This opens the discussion of what our text brings to such a building, but also what remains to be explored.

5.2 Open research questions

Like classic commons, digital knowledge commons deal with the collective management of a subtractive resource. They structure this management by creating a hierarchy whose powers derive from rights controlled by norms, technology, or the institutional framework. We were able to show that by using the institutional analysis and development framework (Ostrom, 2005, 2011). We suggest to use it to study the digital knowledge commons, and we will use it also to frame the research questions.

Recognizing the similarities, we hypothesized that Ostrom’s eight design principles (1990, 2008) hold true. This article has stared a discussion with some of them on how digital knowledge commons make them evolve: clear boundaries, collective-Choice arrangements, Monitoring, Nested Enterprises, (Proportional Equivalence) Benefits and Costs.

For instance, we have showed that some elements, such as the boundaries between the community of policy makers and the mere contributors are less “clear” than in her presentation. The digital aspect, and specifically the modularity of the production, conjugated with the use of digital supervision tools, explains the possibility of creating commons with hundreds to thousands of consumers. More research is needed to understand how the nested enterprises are coordinated, how are the different layers of management organized for the most complex projects, but also how the community controls “simple” contributors, their careers, and thus, the community’s steps. This includes a better understanding of how the law is used to establish the global framework, as well as the various contracts (licenses, contributor agreements, etc.), which, along with the technological structure, instantiate the project and create a hierarchy beyond the global structure presented in Table 1. In general, the research agenda may prioritize discussions about how digital commons organizations can contribute to organizational innovation with regard to collective control over algorithms.

Further research is also needed on the similarities and specificities of the remaining principles. For example, without claiming exhaustivity: how are sanctions organized in these spaces (Graduated Sanctions principle)? How are conflicts managed (Conflict-Resolution Mechanisms principle)? How do commons interact with other institutions (Minimal Recognition of Rights to Organize)? For example, what responsibility does Wikipedia have as a repository of knowledge? Should it be held responsible due to its quasi-monopoly on encyclopedic knowledge? On the other side of the hierarchy, the role of the “owner,” or manager of alienation, and its characteristics deserve further study. For example, one could establish a community-controlled foundation (or one controlled by policymakers), but one could also question whether a company could manage a commons. Today, most foundations, such as the Wikimedia Foundation, the Linux Foundation, the Apache Foundation, and the Eclipse Foundation, are nonprofits controlled by a board whose members are not always, or sometimes not at all, members of the community of policymakers. Whether a legal entity is necessary to ensure the long-term sustainability of the project remains an open question. One of the main database systems and a FOSS project, PostgreSQL, does not have one). This point is true for any commons, as illustrated by the different regimes discussed for local and global commons in McGinnis and Ostrom (1996). Is there a digital specificity?

We hope that our clarification of what a digital knowledge commons is will help us address these questions and, more importantly, improve our global understanding of commons and collective governance.