1. The Problem of ‘Best Practice’: Embedded Infrastructure, Situated Governance

Neither governance nor infrastructure is generally seen as an exciting, enthralling topic. Susan Leigh Star’s foundational essay in infrastructure studies describes asking people to pay attention to infrastructure as ‘a call to study boring things’ (Star, 1999: 377). Similarly, the perception of governance as dull is often noted: a recent blog post on its importance for charitable associations, for example, is titled ‘Why is Governance soooo Boring?!’ (Chamberlain, 2020). Both infrastructure and governance tend to be noticed only when something goes wrong. As long as both are functioning reasonably smoothly, most of us are happy to ignore them and focus on the things they are supposed to facilitate (Ross-Hellauer, 2016, Bilder, 2015).

Yet both governance and infrastructure (and their ongoing maintenance) fundamentally shape the conditions in which we work and operate. Both are, as mentioned in the introduction, profoundly political, and—partly for this reason—both often come to serve the interests of a narrow clique rather than the people and communities they are designed to support. Infrastructure, as Leslie Chan has noted, ‘concerns the power (otherwise hidden) to: set agendas and decisions – which are never neutral but embedded with ideological assumptions and biases; mobilize and accumulate resources; set standards and norms; set boundaries of participation; discriminate – or not, hopefully; and control what gets built, what’s possible’ (Chan 2019, cited in Hall 2021: 31).1 One of COPIM’s workshop participant’s ad hoc definition of governance, meanwhile, sums up its deeply political significance: governance is ‘how things get decided’. Within this context of power and control, community-led organisations, and movements towards various forms of ‘openness’ — whether open access publishing, open source software, open scholarship, or any one of a number of others — can be seen as reactions to appropriations and failings of existing publishing infrastructure and governance.

Our initial intention in planning this report was to document best practice in the governance of open source community-owned infrastructure projects—as a useful resource both to others and to ourselves—as well as to provide a record of our own governance practices and how we arrived at them. Two parallel developments, however, one internal to COPIM and one external, have combined to make us revise and supplement this original plan.

On the one hand, as the wider literature on the governance of community-led, open scholarly infrastructure projects has developed, a substantial body of governance best practice guides and toolkits has emerged over the last few years. Introduced in more detail below, these already offer extremely useful practical advice and resources for establishing or revising governance structures and procedures. Often written or overseen by prominent figures in community-led open infrastructure projects, they are generally concise, intelligent, well informed and researched, and thoughtfully presented. We have taken the view here that to attempt to further condense, summarise, or synthesize their recommendations would be a redundant exercise: rather, we recommend that readers involved in establishing or revising community-led governance structures and procedures consult them directly.

If these guides and toolkits have a common shortcoming, however, it is perhaps that in focusing (for very good reasons) on the more procedural practicalities of governance and providing widely applicable guidelines, they tend to risk decontextualising, de-situating, and de-historicising governance, and thereby playing down its fundamentally political character. This in turn risks eliding some of the most intransigent governance issues open infrastructure organisations are likely to confront, as well as some of the opportunities. This is rarely if ever an explicit failing but rather a cumulative effect of the main foci of this whole body of literature. This report therefore looks to pay particular attention to these aspects of governance.

As both COPIM’s own projects and governance research proceeded, it also became increasingly evident to us that the idea of a widely transferable ‘best practice’ is an illusion. Since its inception COPIM has produced a number of blog posts and reports on open source and community-led project governance (these can all be read here), and documented the development of its own governance structures and those of the projects it has developed—as we will outline further in this report. Foremost among our research, as has already been noted, is Samuel Moore’s extended report on Exploring Models for Community Governance, published in May 2021, which this report looks to build on and extend. Exploring Models presented our initial work on the various structures that already exist for community governance and discussed their pros and cons for COPIM’s open access monograph infrastructures, particularly with regard to one of our main deliverables, the Open Book Collective (OBC). This is a platform and intermediary organisation that allows open-access academic publishers and infrastructure providers (‘initiatives’) to promote their publishing activities and to sell and manage their funding schemes. Moore’s report offers a landscape study of forms of governance within scholarly communication alongside an exploratory study of the theoretical literature on the various forms of governance actually deployed within community-led organisations.

Perhaps Exploring Models’ most important finding is the extent to which community governance is situated: in other words, any good model of governance is invariably ‘highly specific to the resource and community in question’ (Moore, 2021a: 5). This means that it cannot serve as a ‘model’ at all, if by that is understood a template that might be straightforwardly lifted and applied in a different context or situation. As Moore concludes,

there is no real ‘off the shelf’ approach that will work for complex organisations, not least because so much of governance refers to specific values, principles and norms dictating how participants interact with one another. Other considerations such as size, history, culture and the kind of resource(s) governed all play a factor in the sorts of accountability models that shape a project (Moore, 2021: 31).

COPIM’s identification of the situatedness of governance as its primary characteristic mirrors Star’s foregrounding of infrastructure’s embeddedness. For Star, infrastructure is ‘sunk into and inside of other structures’: it ‘does not grow de novo’, but ‘wrestles with the inertia of the installed base and inherits strengths and limitations from that base’ (Star, 1999: 382).

The combination of infrastructure’s embeddedness and governance’s situatedness has profound implications for any practical attempt to establish community-led, open infrastructure projects. It also makes it even more difficult to establish governance ‘best practice’ in the case of such projects than it might be for other kinds of venture. In the face of these difficulties, it can be tempting to abandon any attempt to establish basic principles or widely applicable guidelines for better practice. If it is impossible to identify a single preferred method of governance, however, or even a single reliable method for choosing from among and customising the many models of community governance available, this report does attempt to trace a number of pathways. It does this in part by mapping the roads that COPIM and the infrastructure projects whose development it has supported have laid down for themselves with regards to governance. As will become evident, this is better characterised as a series of diversions, remappings, backtrackings, and reroutings than a simple highway. One lesson that emerges from this, though, is the indispensability of this process. Taken together, the embeddedness of infrastructure and the situatedness of governance mean that governance design for such projects is of necessity iterative, ongoing, continuously to be revisited and revised.

Yet because of this, learning from others’ practices and insights becomes more rather than less important. A number of immensely useful guides to governance have been produced in the last few years for community-led and open organisations. Among the most notable of these are the Governance Toolkit and accompanying Guidebook produced under the aegis of It Takes a Village, a project funded by the Institute of Museum and Library Services in the USA to establish models of collaboration and sustainability for open source software projects. The Educopia Institute has also developed a very helpful portfolio of governance design tools: these include Governance in Formation: Identifying Priorities for Action and Making Decisions, a facilitator’s guide which also contains helpful templates and a plethora of links to further resources (also available as a Google Doc). Katherine Skinner’s Community Cultivation — A Field Guide, also from Educopia, ranges well beyond issues of governance, but does explicitly address implementing or updating governance structures, identifying governance as one of five interlinked key growth areas for a community (the others being ‘vision’, ‘infrastructure’, ‘engagement’, and ‘finances and HR’). The long chapter dedicated to project and community governance in The Open Source Way 2.0 (the full guide is also available in PDF, EPUB, MOBI, and Kindle Fire EPUB formats ) is another invaluable resource, and one of the most alert to the practical implications of the conflicts and tensions that community-driven projects have to deal with. The University of Kansas’s Community Toolbox, meanwhile, includes practical sections on organisational structure, writing bylaws, and other governance issues (see the Table of Contents for a full list). Similarly, the Open Source Alliance for Open Scholarship Handbook includes a dedicated section on ‘Governance in Open Source for Scholarship’. This sets out some useful initial questions that any project should ask itself about its governance, aimed particularly at smaller scale projects, including those initiated by a single individual. The OSAOS Handbook also includes a handy governance ‘Resource List’.

Each of these documents sets out invaluable guidance, recommendations, and best practices for the governance of community-led and/or ‘open’ projects and organisations. Together they offer what is in many ways an excellent starting point for any practical attempt to establish or review the governance of any given infrastructure project. All published within the last five years or so, they are indicative of a recent growth in such projects, especially within academia, and of the slow but unmistakable growth in funding and interest that makes such reports possible. They also reflect a widespread recognition that the neglect of governance (which has sometimes shaded into outright hostility) has been a real hindrance to many open projects, and in some cases led to painful fractures within and between the communities they seek to serve. This report is indebted to this literature and draws upon the examples and reading it references, and one of its purposes is to highlight its key points and elements of particular relevance to the governance of open infrastructure for scholarly communications.

The Politics of Governance

Yet in performing the necessary and invaluable labour of providing practical frameworks and tools for establishing governance structures, the collective effect of this proliferation of reports, guidebooks, toolkits, worksheets, and field guides, mostly produced from within Europe and the Anglophone world, also runs the risk of reducing governance to a predominantly managerialised, procedural affair.2 While acknowledging the importance of establishing and articulating a project’s values and principles, this can lead to governance being presented as occurring within a vacuum. One thing that we therefore would like to add to the existing governance guidance emerging mainly from the Anglophone and European ecosystems, is that many community-driven projects committed to values of openness find themselves embedded in quite inhospitable environments, which makes it hard to apply an off-the-shelf approach to governance. A range of powerful pressures either actively work against—or create conditions in many ways extremely hostile to—open-source and community-led organisations. Whether we group these forces, irreducible to just another ‘risk factor’, under the shorthand of the neoliberal university, big tech or simply capitalism, their gravitational pull is likely to be the single most disruptive influence upon attempts to sustain alternative infrastructure projects that have an impact on the intellectual communication ecosystem and that actually function. Establishing governance without a clear-eyed, conscious, practical reckoning with these forces may result in a compromised organisation whose ‘open’, ‘community’ credentials are mainly window-dressing, or in a bad utopia liable to shatter or crumble at its first impact with brute reality.

Foregrounding the search (or the need) for a more explicitly political mode of community-led governance has the potential to be deeply discomfiting. For open infrastructure organisations constrained to operate in inhospitable environments and built on potentially fragile coalitions, there are good reasons to shy away from the disruption and antagonism that any mention of the political can provoke. But governance, like infrastructure, can never be objective or neutral: and in a radically inequitable environment, anything that casts it as such, even if unintentionally, is liable to reinforce inequity. The same holds for scholarly practice and publication too. As Connelly and Joseph-Salisbury note, while ‘myths of objectivity and neutrality characterise traditional academic practice, more radical work, particularly within feminist and anti-racist traditions, has insisted on the importance of rejecting neutrality’ (2021: 12). We follow Connelly and Joseph-Salisbury in arguing that ‘rather than undermining academic rigour, the explicitly political and partisan nature of anti-racist scholar-activism offers a higher level of integrity and honesty than scholarship that purports to be objective,’ making clear the assumptions that underpin not only scholarship but also, we would emphasize, its publication, exchange, and circulation (Joseph-Salisbury & Connelly, 2021: 13).

The widespread ‘illusion that we can finally dispense with the notion of antagonism’ has been diagnosed by Chantal Mouffe (among others) as the ‘evasion of the political’, a symptom of the ‘rationalist, universalist and individualist’ underpinnings to a great deal of the current (liberal) democratic thinking on which the current infrastructural norms that open infrastructures challenge are also founded. The main shortcoming of this evasion, Mouffe argues, is that it ‘cannot but remain blind to the specificity of the political in its dimension of conflict/decision, and that it cannot perceive the constitutive role of antagonism in social life’ (2005: 2). Open infrastructure governance has lessons still to learn (and, perhaps, also to teach) to other more explicitly political projects about how to situate the antagonism and the ‘conflict/decision’ at its centre in a creative and productive (rather than wholly destructive) manner.3

In Revolutions in Reverse, David Graeber tells a story that serves as a useful parable. Graeber is describing his involvement in the Direct Action Network (DAN) in New York, a decentralised group at the forefront of the global justice movement in the city, ‘operating on principles of direct democracy according to an elaborate, but strikingly effective, form of consensus process’. DAN ‘existed in a purely political space’, with ‘no concrete resources, not even a significant treasury, to administer’ (Graeber, 2011: 43):

Then one day someone gave DAN a car. This caused a minor, but ongoing, crisis. We soon discovered that legally, it is impossible for a decentralized network to own a car. Cars can be owned by individuals, or they can be owned by corporations, which are fictive individuals. Governments can also own cars. But they cannot be owned by networks. Unless we were willing to incorporate ourselves as a nonprofit corporation (which would have required a complete reorganization and abandoning most of our egalitarian principles) the only expedient was to find a volunteer willing to claim to be the owner for legal purposes. But then that person was held responsible for all outstanding fines, insurance fees, and had to provide written permission to allow anyone else to drive the car out of state; and, of course, only he could retrieve the car if it were impounded. Before long the DAN car had become such a perennial problem that we abandoned it (Graeber, 2011: 43-44).

This relatively minor story struck Graeber as having something important to tell us about why projects aiming to democratise society are ‘so often perceived as idle dreams that melt away as soon as they encounter hard material reality’. Contrary to one common narrative, this has nothing to do with any supposed inefficiency of radical, community-led forms of organisation: in the case of DAN, ‘police chiefs across the country had called us the best organized force they’d ever had to deal with’ (Graeber, 2011: 44). Instead, Graeber writes of the ‘reality effect (if one may call it that)’ that occurs when radical projects engage with the embedded infrastructure of the status quo. For Graeber, this effect is particularly associated with the moment at which such projects ‘enter into the world of large, heavy objects’. This, he suggests, ‘is not because these objects are somehow intrinsically difficult to administer democratically – history is full of communities that successfully engage in the democratic administration of common resources – it’s because, like the DAN car, they are surrounded by endless government regulation, and effectively impossible to hide from the government’s armed representatives’ (Graeber, 2011: 44). It is when radical community-driven networks and organisations come up against this reality effect that they ‘tend to founder, or at least become endlessly difficult’.

The experiences of movements for open scholarship suggest that today it is not only (or even principally) large heavy objects that trigger the effect Graeber describes. Any attempt to exercise influence over anything that might be commodified or otherwise turned to profit is liable to do so. Practically all open access and community-led projects have their own stories of running up against such regulations. Yet because infrastructure is by definition not simply a single instance of work but the very systems and services that make possible the continued functioning of the whole, any attempt to reform, challenge, or set up alternatives to established infrastructures is likely to come up especially quickly against the kind of structural violence Graeber outlines, even if that violence sometimes remains hard to see. Infrastructural changes, by their very nature, have system-wide implications. The fate of Aaron Swartz should serve as a reminder, however, of the deadly consequences of structural violence for those who explore solutions that existing infrastructures consider too radical to the way scholarly knowledge ‘is increasingly […] locked up by a handful of private corporations’ (Swartz, 2008).4

Infrastructure projects also sometimes pull together a more diverse range of actors and organisations than other kinds of open projects, creating a broader ‘community of communities’. This can be a source of strength, but it can also mean that the alliances and coalitions involved are more fragile and fraught. In particular, because they ‘encourage system-wide changes’ (as Educopia describes its mission) such projects arguably have to bridge the gap between a radical, alternative, oppositional ‘undercommons’ (from which many of the various movements for ‘openness’ emerged) and those embedded within existing infrastructures who recognise their flaws and the need for reform, and who often have most power to transform them, but may be hesitant or reluctant to disturb the established status quo. For these reasons, infrastructure projects may find themselves having to engage more immediately and directly with existing, deeply embedded structures and regulations than do other (open) projects, for example as part of incorporation processes. Measures such as legal incorporation may be the best or only way to provide legal protection to community members driving a project forward, for example, or to enable them to apply for much-needed grants. But these concessions and compromises need to be recognised as such, and their costs weighed.5 The initial, significant challenge of governance then becomes how to guard against surrendering communitarian values and principles of openness against external more vertically and hierarchically established structures.

The scholarly communications landscape is littered with examples of how ostensibly community-driven projects attempting to open up the existing infrastructure have succumbed to just the kind of structural violence of which Graeber writes. Only a few months ago, COPIM responded in a blog post to the acquisition of Knowledge Unlatched (KU) by Wiley (global profits in 2021 $1,941,500,000). This acquisition followed KU’s opaque transition in 2016 from a UK Community Interest Company (a non-profit organisation) into a German GmbH (roughly equivalent to a UK PLC, i.e., a for-profit company).6 As COPIM’s blog post noted, this move by Wiley was only one of several recent takeovers of open access (OA) infrastructure by large commercial organisations. Other examples include the acquisitions of Mendeley, the Social Science Research Network (now SSRN), and of bepress by Elsevier, in 2013, 2016 and 2017 respectively, and of F1000 Research by Taylor & Francis in 2020. These takeovers reflect ‘an ongoing consolidation of research infrastructure by major publishing corporations, and in particular the increasing attempts to monetise and, potentially, monopolise the infrastructures of open knowledge dissemination’ (COPIM, 2021). They often fit within a broader programme of ‘openwashing’, ‘deployed as an intentional marketing tactic […] to co-opt the values of Open for profit’ and ‘mobilized through subtle methods such as appropriating the language of Open to redefine its meaning’ (Waugh & Carlisle-Johnston, 2022). Each of these cases has its own distinct history, but the general pattern is all too evident. Other fields of open, community-led endeavour outside academia have their own stories to tell.

Under such circumstances, the probable outcomes are that any open infrastructure organisation for openness will either be radicalized or will ‘sink into a morass of compromises that absorb it back into the social order that it once sought to change’ (Bookchin, 2015: 18).7 The broader socio-economic and political structures that make such acquisitions common are precisely those that existing infrastructures of scholarly communication are deeply ‘sunk into and inside of’ (to reprise Susan Leigh Star’s phrase), and within which projects that would provide new, open infrastructures must find a way of operating as they look to establish alternatives. In such circumstances, good governance is perhaps best thought of as an important tool in negotiating that difficult relationship. Good governance may require above all a clear-eyed alertness to the reality of the landscape in which an infrastructure project is embedded and its governance situated, as well as of the nature of alliances that are being entered into and of the likely costs and benefits of compromises made. Only such an awareness will lead to governance structures and mechanisms that provide meaningful defences against co-optation, commercial appropriation, or takeover. Most open source, community-led projects are either explicitly inspired by or implicitly sympathetic towards at least some part of ‘that anti-authoritarian tradition that wishes to replace bureaucratized forms of violence with bottom-up, directly democratic structures’ (Graeber, 2017). There do remain ‘pockets of contradiction and possibility’ within all forms of hegemonic infrastructure, whether within the contemporary university, the publishing world or beyond, and ultimately negotiation and engagement with existing infrastructures is not only worthwhile but in most cases all but unavoidable for open, community-driven projects (Joseph-Salisbury & Connelly, 2021: 3). This report therefore begins by urging that ultimately, this is what governance is – or must be – for the kinds of projects it discusses: a means of exploiting to the maximum the specific situated affordances of both the cracks and nooks in existing hegemonic structures and regulations, and—through an alternative ethos of mutual aid and collaboration—the possibilities offered by an emergent community-led ecology, and its capacity to scale small.