One natural way to explain why the illegal sharing of academic research—and unilateral remedial action more generally—calls for moral justification is to appeal to the fact that they breach a general duty to obey the law. This explanation is familiar from discussions of political disobedience, and it may seem to apply naturally to cases of unilateral remedial action. The basic thought is that in reasonably just democratic societies, laws enacted by the political majority or its representatives carry a presumption of authority and, as such, generate a (defeasible) duty to obey: one that is typically understood as a universally applicable, pro tanto, and content-independent moral obligation to comply with all laws within a given legal system.24 Under this view, obedience serves as a means to “respond correctly” to the values endorsed by the political majority.25 Acts that involve unilateral lawbreaking—such as sharing academic research illegally—are counter-majoritarian insofar as they involve defiance of those democratically enacted rules. Hence, they face a justificatory burden: those who engage in them must show that the duty to obey is outweighed by competing moral considerations.

This view, however, provides at best a partial explanation of why illegal sharing calls for justification. This is because it is structured around the existence of a duty to obey the law, and it is well known that such a duty is highly susceptible to objections (many of which we think are plausible, if not decisive).26 However, even if such a duty were granted, it would not by itself settle the moral questions raised by illegal sharing (and unilateral remedial action more generally), since this practice raises further concerns (for instance, about the values underpinning specific laws, the interests of third parties, or the stability of systems of coordination). A more promising (and less theory-laden) explanation, as we will show, appeals instead to a variety of considerations tied to the specific laws that such practices violate.

Illegal sharing of academic research breaches copyright laws that purport to protect the interests of authors and publishers. Complying with these laws may be valuable and pro tanto obligatory even in the absence of a general duty to obey the law. Hence, still somewhat inspired by the earlier view, we can say that illegal sharing is pro tanto wrong because those who engage in this practice are not “responding correctly” to considerations that underlie a duty to comply with copyright laws. This shift from a general duty to obey the law to reasons for complying with particular laws reflects an account of the moral concerns raised by illegal sharing that is both more circumscribed to the practice and less dependent on highly controversial theses.27 Of course, the plausibility of this view hangs on there being good reasons to regard compliance with existing copyright regimes as valuable and pro tanto obligatory.

There are, indeed, good reasons to support complying with copyright laws. At least four mainstream theories help articulate these reasons. According to the personality theory, the product of an author’s labor represents an expression of their inner self and creative freedom, which is intrinsically valuable.28 Copyright, on this view, is necessary to safeguard something intrinsically valuable. A second view, the Lockean theory, stresses the fairness of allowing authors to retain rights and control over the fruits of their labour. Here, copyright ensures a just allocation of the value created by individual effort.29 The welfare theory, by contrast, is less concerned with individual rights and more with promoting the overall good. It justifies copyright on the grounds that certain restrictions are indispensable and conducive to greater creation, wider availability, “widespread public enjoyment,” and greater production of knowledge and cultural goods.30 Finally, the culture theory maintains that copyright law fosters not welfare, but a “just and attractive” culture of creative and intellectual innovation.31 It supports mechanisms that protect attribution, prevent misappropriation, and ensure that both credit and material benefits are fairly distributed.

We need not determine which of these theories is most plausible, nor treat them as mutually exclusive. For present purposes, it is sufficient to acknowledge that each highlight independent moral considerations that support maintaining copyright protections. At their core, these accounts converge on the view that authors’ intellectual creations (i.e., their ideas and expressions) are prima facie valuable. Protecting them, and regulating their dissemination and use, is therefore also valuable to the same extent. This does not imply that all existing copyright regimes are justified, nor that they strike the appropriate balance between competing values; it merely establishes that, in the absence of countervailing considerations, protecting authors’ intellectual creations give us some good reasons to maintain and comply with a system of copyright laws.

There are also domain-specific reasons for why copyright laws matter in academic contexts. They help secure credit for scholarly contributions, define responsibilities towards copyright-holders (e.g., authors, publishers, or research institutions), clarify the conditions for fair use (e.g., citation practices, commercial restrictions, and royalty arrangements), and allocate responsibility for the content of published research. To some extent, copyright laws serve the interests of academic authors by protecting their work from plagiarism and by assigning both credit and responsibility for their intellectual output. In this respect, they give institutional expression to the values identified by the personality theory, safeguarding the author’s creative autonomy while ensuring proper attribution and recognition within the scholarly community. Copyright law also establishes a framework for distributing the financial fruits of authorship, particularly through royalties. Although such financial incentives play a more limited role in academic publishing than in other creative industries, they nonetheless illustrate how copyright regimes aim to regulate the allocation of economic rewards associated with intellectual labour.

These considerations suggest that copyright law protects morally significant interests and that practices which are inconsistent with them may be morally problematic. If the illegal sharing of academic research is to be justified as a form of unilateral remedial action, it must overcome these considerations.

Copyright in academic publishing is typically mediated through contractual agreements between authors and publishers. When submitting a manuscript for publication, authors are commonly required to assign or license certain intellectual property rights in exchange for dissemination, editorial services, and, in some cases, financial compensation. These agreements may regulate royalties, citation and attribution practices, permissions for use, reproduction, and distribution. By structuring the allocation and exercise of these rights, publishing contracts give concrete institutional expression to the legal protections afforded by copyright law.

One common way to account for the moral significance of such agreements is to understand them as promises. On the promissory view of contract, voluntarily undertaken agreements generate pro tanto obligations that bind the parties who enter into them.32 If this account is correct,33 then breaching the terms of a publishing contract constitutes a pro tanto wrong insofar as it involves the violation of a promise freely made.

This has important implications for evaluating the illegal sharing of academic research. To the extent that authors transfer or license rights to publishers, they may incur obligations not to disseminate their work in ways that contravene the agreed terms. Certain forms of “small-scale” sharing like sending photocopies of their books to colleagues or uploading their work to online research repositories in breach of contractual restrictions may therefore be morally problematic. While such practices are often regarded as benign or even commendable within academic culture, they may nonetheless involve the violation of promissory obligations. In this respect, they illustrate how even “small-scale” sharing, as it were, may be objectionable.

Shadow libraries, however, are not themselves parties to these agreements. As such, they are not bound by the promises made between authors and publishers. This creates an important asymmetry: while authors may incur promissory obligations not to disseminate their work in violation of contractual terms, third parties who were not party to those agreements cannot straightforwardly be said to breach them. At most, their actions may be characterised as interfering with contractual arrangements and with the interest parties may have in fulfilling those arrangements.

The contractual framework of academic publishing thus introduces a distinct source of moral concern. While authors may bear promissory obligations arising from the agreements they voluntarily enter into, the case against shadow libraries rests on the more contested claim that third parties wrongfully interfere with those arrangements. This asymmetry suggests that contractual considerations cannot fully explain the moral status of illegal sharing. At the same time, it reveals a potentially counterintuitive conclusion: authors may bear more stringent obligations than shadow libraries themselves, and certain forms of small-scale illegal sharing may be more morally problematic than common sense would suggest.

A further source of concern arises from the normative force of imperfect conventions. Even when legal regimes are flawed or suboptimal, they may nevertheless generate moral duties by structuring expectations and coordinating social life. A recent defence of this line of thought has been advanced by Aaron Salomon, who proposed what he calls the “switch interpretation.”34 According to this view, when assessing the justification for departing from current norms or conventions—as in debates about reform, disobedience, or unilateral remedial action—we should not evaluate alternative candidate principles as if they had always been in place, but we should instead consider the moral weight of the transition costs and reliance losses that would result from switching to them from our current social arrangements. The view, therefore, highlights the moral relevance of transition costs and the reliance interests generated by established conventions. It recognises that such conventions often function as stable “rules of the game,” enabling individuals and institutions to coordinate their behavior and plan their activities over time.35

Property regimes (including intellectual property regimes) and social conventions more generally “allo[w] us to plan our lives around a shared set of expectations concerning what we can and cannot do in our interactions with each other.”36 Abrupt departures from existing practices may therefore impose significant harms or costs on those who have organised their activities in reliance upon them. Of course, this does not entail blind obedience to imperfect conventions. Even though reliance and expectation-based considerations can generate moral duties, there are limits to the moral force of imperfect conventions. On Salomon’s account, for example, the moral force of imperfect conventions depends on a balancing of interests, including especially the costs of transition and the benefits of reforming existing practices.37 Thus, when the interests protected by reform or transitioning outweigh the reliance interests sustaining the status quo, existing conventions may be justifiably overridden.

Applied to the case at hand, this suggests that copyright regimes—as imperfect as they may be—may generate pro tanto duties of compliance insofar as they enable coordination, protect reliance, and sustain the infrastructure of academic research. Academic authors, publishers, universities, libraries, and funding bodies organise their activities against the backdrop of existing copyright norms. Scholars enter publishing agreements with expectations concerning attribution, dissemination, and professional recognition; publishers invest in editorial processes, peer review, and digital platforms in reliance on legally enforceable rights; and universities and libraries allocate resources, negotiate licences, and maintain subscription systems that ensure access to scholarly materials. Even open-access initiatives, such as institutional repositories and transformative agreements, operate within the existing copyright framework. These practices reflect good-faith reliance on a system that, though imperfect, structures the production and circulation of academic knowledge.

Widespread disregard for copyright regimes may therefore carry significant costs. Unilateral departures from established norms risk disrupting expectations, undermining licensing arrangements, and destabilising the institutional structures that support research dissemination. If generalised, such practices could erode incentives for investment in publishing infrastructure, compromise mechanisms for quality control and attribution, and generate uncertainty about the lawful dissemination of research. These concerns resonate with broader consequentialist objections, including the fear that uncoordinated illegal sharing could lead academic research to an anarchic “grab what you can” regime marked by legal uncertainty, uneven access, and diminished institutional support for scholarly production and publication. While such risks should not be overstated, they illustrate how uncoordinated lawbreaking may impose significant coordination and reliance costs on others.

Taken together, these considerations suggest that existing copyright regimes, albeit imperfect, may carry normative force. They help explain why practices that breach them—such as the illegal sharing of academic research—are not morally trivial, but the kind of unilateral remedial action that call for moral justification.

Members of the open access movement and shadow libraries typically take themselves to be responding to injustice. They also use the language of rights and duties to motivate their activities. However, not every perceived injustice is a real injustice. By examining existing publishing practices, we suggest that restricting access to academic knowledge gives rise to at least three real injustices, namely: (i) violation of the right to access knowledge; (ii) marginalization of research communities; and (iii) epistemic injustice.

A distinctive feature of unilateral remedial action is its remedial character. Unlike civil or uncivil disobedience, which typically aim to communicate dissent, persuade the political majority, or bring about institutional reform, unilateral remedial action seeks to directly rectify injustice. Whereas disobedience is often communicative and forward-looking, unilateral remedial action is backward-looking and restorative. Hence, to qualify as remedial, an action must at least aim to provide or restore that of which individuals have been unjustly deprived.

This remedial orientation imposes a justificatory constraint: for unilateral remedial action to be justified, it must minimally remedy the injustices it seeks to address. Actions that, despite having remedial aims, are merely symbolic, futile, or counterproductive fail to satisfy this condition. Nor is it sufficient that an intervention merely possesses a speculative chance of success. Given the moral risks and wrongs that unilateral lawbreaking generates, its justification requires more than a remote possibility of remedying injustices. From the agent’s perspective (i.e., someone who is considering engaging in unilateral remedial action), the action must be reasonably expected to deliver a minimal and non-trivial remedy; from the standpoint of evaluating the practice, it must in fact succeed in doing so. The point is not that success must be complete or guaranteed: many injustices cannot be fully remedied. Rather, the point is that the action must at least provide a genuine and meaningful alleviation of the relevant injustice. The more effectively it remedies the injustice at stake, the stronger its moral justification.

To take a familiar example, acts of redistributive theft—often associated with the “Robin Hood” trope—may themselves constitute instances of unilateral remedial action. However, their justification is typically harder to establish than that of illegally sharing academic research. While redistributive theft aims to redress economic inequality, it deprives others of fungible resources, and such deprivation may itself be unjust depending on the circumstances. Moreover, although empirical evidence suggests that, for example, unconditional direct cash transfers can effectively alleviate poverty and hardship, their success is highly context-dependent and sensitive to design and implementation features.77 Importantly, this evidence concerns institutional redistribution programmes rather than unilateral acts of theft. The remedial effects of unilateral acts of redistributive theft remain contested and uncertain. The remedial condition thus cautions against overgeneralization. That illegal sharing of academic research may be justified, despite certain similarities to redistributive theft, does not entail that the latter is likewise permissible. Their justificatory statuses depend on whether they genuinely mitigate injustice rather than merely express opposition to it.

A further comparison helps clarify the importance of the remedial condition. The unauthorised sharing of music, films, or other forms of entertainment may resemble the illegal sharing of academic research. Yet, unlike the latter, such practices typically fail to remedy injustices. Deprivation of entertainment does not ordinarily constitute an injustice, and its illicit distribution does not generally serve to rectify a wrongful exclusion from an essential good or the deprivation of a right. Consequently, while both practices involve similar methods, their normative profiles differ.

The illegal sharing of academic research clearly meets the remedial condition. Shadow libraries do not merely protest or denounce the injustice of restricted access to academic knowledge; they directly alleviate it. By providing immediate and unrestricted access to scholarly materials, they restore a good that is unjustly withheld from researchers, students, institutions, and the public at large. In doing so, they enable participation in global research communities, facilitate education and scientific progress, and directly mitigate real injustices. The number of testimonies from researchers who express gratitude to shadow libraries for allowing them to continue their research activities is evidence of that.

The “necessity” condition is often framed as a corollary of the idea that lawbreaking must be a “last resort”—that it is permissible only where available “legal means of redress have proved of no avail.”78 While this formulation emphasises the unavailability of legal channels, the core principle is broader: lawbreaking is justifiable only if there is no less harmful or disruptive means of addressing the relevant injustices. This conception of necessity is not uncontroversial.79 Nevertheless, with some qualifications, it serves our purposes.

One such qualification concerns the nature of alternative means. Any less harmful or disruptive alternative must also be feasible. This is not to suggest that individuals should be able to pursue alternatives at no cost or inconvenience, but rather that the associated burdens must be both bearable and not more harmful in themselves than the injustice the alternative seeks to avert. On this view, unilateral remedial action may be necessary even where alternative routes exist, if those alternatives are prohibitively costly or unrealistic. This is often the case for marginalised communities, for whom poor access to the ordinary channels of contestation is constitutive of their very social exclusion or oppression. In such contexts, requiring adherence to these formal mechanisms in lieu of unilateral remedial action would be morally perverse.

For similar reasons, alternatives to unilateral remedial action must also be just as remedial. If there is little reasonable expectation that petitions, protests, or legal challenges will secure change—or if these measures merely postpone actual remedial action—it is again morally problematic to insist on their use. Even Rawls, a proponent of the “last resort” criterion, acknowledged that one need not insist on lawful contestation when the political majority is “immovable or apathetic” to the cause.80

With these considerations in mind, to show that sharing research illegally is necessary we must compare illegal sharing with plausible remedial alternatives. In doing so, we must assess whether these alternatives are not only less harmful overall, but also feasible and sufficiently remedial. We can consider a few alternatives that, at face value, may fit these requirements:

Making research open access via the “golden route” involves making published work freely available on the publisher’s website. To this end, publishers charge authors “article processing charges” (APCs), typically ranging from USD 500 to USD 11,000 per article.81 The rationale behind this model was that journals could be sustained entirely through APCs, thereby allowing subscription fees to be significantly reduced and eventually phased out. This shift, however, has yet to materialise. At present, universities and libraries pay APCs in addition to existing subscription costs.

APCs are generally covered by the researcher’s institution or research funders—organizations that have often already financed the research itself. Yet, as previously noted, current publication costs are becoming prohibitive even for well-resourced institutions and remain entirely inaccessible for most independent scholars and researchers in LMICs. The problem is even more acute for academic books. The inequities generated by APCs are now well documented, revealing a pattern of global disparity in academic output.82 As discussed, this not only impoverishes the collective knowledge base but also exacerbates the marginalization of low-income and epistemically disadvantaged researchers.

Some journals have introduced APC waivers for scholars in least developed countries. However, those based in lower-middle-income nations or unaffiliated with research institutions—including activists and grassroots researchers who rely on academic work to inform their efforts—are still excluded. Moreover, the golden route relies on someone’s willingness and ability to pay for the APCs. While this may not present an insurmountable barrier for researchers affiliated with wealthy institutions, it poses a serious problem for those not in this position. Who, for instance, will pay the APCs for unaffiliated researchers? And who will cover the costs of making research published fifteen or fifty years ago freely accessible?

Despite its prominence and backing by major publishers, the golden route remains limited in terms of feasibility. More troublingly, it risks reinforcing marginalization and epistemic injustice by ensuring that only research funded by those able to pay APCs is made open access. Unless this model is significantly reformed—through widespread APC waivers and retroactive policies to render paywalled research open access—it cannot serve as a viable solution to the restricted access to academic knowledge.

An alternative to the golden route is the so-called “green route,” whereby authors achieve open access by depositing versions of their manuscripts in designated repositories. A prominent example of this model is the Rights Retention Strategy (RRS), promoted by cOAlition S.83 Under this approach, authors are instructed to apply a Creative Commons licence to their accepted manuscript, thereby retaining the right to share it immediately upon publication, even when publishing in subscription-based journals. The stated aim is to achieve universal access without prohibitive costs or limitations.

In practice, however, the green route is subject to a range of constraints. Publishers that accept this route frequently impose embargo periods—typically between six to thirty-six months—which delay dissemination and undermine the model’s capacity to deliver timely access. In fast-paced fields these delays may significantly diminish the relevance or impact of the research. Besides, despite having advantages over the golden route, its implementation relies heavily on individual authors asserting contested rights in contexts of pronounced power imbalance. Publishers routinely disregard such assertions, insisting on exclusive rights or embargo terms as a condition of publication. Although they sometimes lack a clear legal entitlement to do so, the practical reality is that many authors—wary of compromising their chances of publishing in high-prestige journals—choose not to contest publishers’ demands. In this way, the green route effectively shifts the burden of contestation onto researchers, many of whom lack the institutional backing or bargaining power necessary to challenge entrenched publishing norms.84 As a result, the green route, while attractive, falls short of offering a sufficiently feasible or effective solution to restricted access.

Alternative (b) involves accessing academic research through public libraries and inter-library agreements. The central problem with this option mirrors that of the golden route: just as research centres and universities—particularly in LMICs—lack the funds to subscribe to journals or pay APCs, public libraries are similarly under-resourced when it comes to accessing paywalled academic publications. Where access is available, its scope is determined by the financial capacity of the library in question. In addition, most public libraries require users to be physically present to access online databases and resources, creating further obstacles for researchers living at a distance.

How about inter-library agreements? As a rule, these agreements apply only to physical bibliographic collections and typically exclude most online databases and e-books due to licensing restrictions. While such cooperative arrangements do exist in many LMICs, they tend to be less effective than those in high-income countries, owing to limited financial support and lack of institutional commitment.85 In their current form, these mechanisms do not offer an adequate solution to the problem of restricted access to academic research.

Alternative (c) is the dissemination of pre-prints via private academic networks. Increasingly, researchers use for-profit platforms such as ResearchGate, Academia.edu, and SSRN to share pre-prints of their work. Many authors, though not all, do so in part to bypass paywalls and broaden the reach of their research. However, while some of these platforms have become targets for ghost journals,86 fraudulent publishers,87 and manipulated impact metrics,88 others accept a wide range of content—including non-peer-reviewed papers, working drafts, and incomplete versions of peer-reviewed articles. This means that accessibility often comes at the expense of quality control.

Recent analyses reveal that despite a degree of geographical diversity among uploaded research, search algorithms on these platforms tend to prioritise work from institutions in high-income countries, displaying it more prominently than other material.89 This likely contributes to the entrenched epistemic credibility accorded to researchers from particular locations.90 In addition, the copyright status of pre-prints is frequently ambiguous, and several of these platforms have faced legal challenges from major publishers over alleged copyright infringements.91

Another issue is that while such networks can be relatively effective in distributing pre-prints of recent journal articles, they are ill-suited for disseminating academic books or work by authors unable or unwilling to upload their research—whether due to copyright constraints, greed, illness, or death. As such, this avenue offers, at best, a partial and inconsistent remedy to the problem of limited access.

If none of the alternatives considered so far prove adequate, one might argue for prioritising radical reform of academic publishing. We are sympathetic to this proposal, and there have indeed been a number of movements aimed at reforming the current model. The open access movement of 2012—dubbed the “Academic Spring”—led many journals to adopt the golden open access route.92 Unfortunately, as previously discussed, this model has serious limitations and unintended consequences.

Other journals have adopted the so-called “diamond open access” model, under which articles are published as fully open access without imposing processing charges on authors. This model has existed since at least the late 1980s and has been widely embraced in Latin America.93 While diamond open access journals tend to have lower average output than their commercial counterparts, they succeed not only in providing free access to academic knowledge for both authors and readers, but also in shielding editorial decisions from the profit-driven pressures of commercial publishers.94 These advantages have recently prompted a few philosophy journals to sever ties with major publishing houses and transition to the diamond model.95 Nevertheless, the shift towards diamond open access has been slow. A key factor for this is that most diamond journals operate with limited infrastructure, minimal staffing, and unstable funding which restrict their capacity for growth. Many rely on volunteer labor and lack dedicated support for technical operations, licensing, and archiving, which larger publishers typically provide.96 In the context of growing financial strain on universities, the question of sustainable funding thus remains unresolved. Furthermore, while the model appears increasingly viable for academic journals, it remains more constrained in its application to academic books—arguably due, at least in part, to institutional incentives that lead academics to prioritise publishing with established commercial presses over independent open access alternatives.

Although this and similar reform initiatives are encouraging, comprehensive reform of academic publishing is likely to be a slow process. Even if, in the long term, reform offers a feasible and less harmful alternative to illegal sharing, it cannot resolve access issues in the short term. To expect researchers in LMICs to wait indefinitely—enduring ongoing injustices and suspending their scholarly activities while reform progresses—is to offer an ineffective and morally inadequate solution.

Having reviewed these prominent alternatives and their limitations, it becomes clear that even the most promising—i.e., reform—pales in comparison to what shadow libraries have achieved. Immediate and unrestricted access to academic books, peer-reviewed articles, and materials spanning authors and decades represents a substantially more effective remedy for the lack of access to knowledge. As for feasibility, the continued existence of shadow libraries—despite facing numerous lawsuits and operating without institutional support—speaks volumes.

While shadow libraries clearly outperform existing alternatives in terms of feasibility and remedial effectiveness, we must also consider the potential harms and disruptions associated with their activities. This is where the question of proportionality emerges.

We should distinguish between two forms of proportionality: narrow and wide.97 Narrow proportionality concerns the harm caused to those directly responsible for the injustice—such as large publishing conglomerates and states that enforce restrictive copyright laws and sanction exploitative publishing practices. Wide proportionality, by contrast, concerns harm to those not responsible, or only marginally so. The rationale here is straightforward: we should be more cautious about inflicting harm on those who are not clearly culpable than on those who are.

Attributing responsibility to commercial publishers and state actors that enable the current copyright and licensing model, particularly given their roles in enforcing monopolistic pricing and relying on unpaid academic labour, should be less controversial.98 A more complex issue arises in assessing the responsibility of authors and academics who publish in, review for, or serve on the editorial boards of these publishers. Many of these individuals find themselves in a double bind, caught between career or job-market incentives and the systemic flaws of the publishing model. While a fuller analysis of their moral responsibility lies beyond the scope of this paper, we suggest—at minimum—that we distinguish between harms inflicted on clear perpetrators and those whose complicity is partial, constrained, or excusable.

Starting with the potential harms that illegal sharing may inflict on researchers. Sharing someone else’s work without consent typically constitutes a breach of copyright and contractual interests. Nonetheless, unauthorised sharing might undermine financial interests linked to copyright, particularly through reduced royalties. In practice, however, most academics earn little from royalties. Moreover, a significant proportion of academic research is publicly funded, and even where it is financed by private universities or funders, the public often bears indirect costs. Most academics are salaried, and the production of research forms part of their paid role. While academic salaries may not always reflect the full value of scholarly labour, the fact that authors are already compensated typically weakens the weight of their financial claims when balanced against the epistemic and moral benefits of broad dissemination.

A further consideration concerns contractual obligations arising from academic publishing agreements. As discussed in Section III.C, authors often undertake commitments with publishers that restrict the dissemination of their work. As third parties, however, shadow libraries are not bound by these commitments and therefore do not incur the relevant obligations. Any wrongdoing on their part consists, at most, in interfering with contractual expectations of the contracting parties. The justification of such interference is thus subsumed within the broader assessment of harms to authors and publishers undertaken in this section.

By contrast, authors who illegally share their own work may violate commitments they have undertaken, and these commitments carry additional moral weight. While the binding force of such agreements is not beyond question (particularly where they arise within structurally unequal and potentially exploitative arrangements), authors may nonetheless have stronger pro tanto reasons not to share their research illegally than shadow libraries. This helps us shed light on an important aspect of “small-scale” violations of access to knowledge. Such violations—for example, an author illegally sharing a copy of their own work through an open repository or with a LMIC-based colleague who lacks access—may be harder to justify. This is not only because of the weight of contractual obligations, but also because authors face greater personal risks in engaging in illegal sharing, including legal sanctions and professional repercussions—risks from which shadow libraries are partly insulated by their infrastructure and anonymity.

That said, wide proportionality considerations do pose challenges to certain aspects of shadow libraries’ activities. Some repositories include trade books and works of fiction by professional authors whose livelihoods depend meaningfully on royalties or sales. Unlike academic research, such works are often intended primarily for entertainment. As a result, their unrestricted dissemination is less directly connected to the epistemic values commonly invoked in defence of the right to access knowledge—values such as truth-seeking, error reduction, scientific advancement, and understanding. This is not to deny the value of making literature and artistic works widely accessible. Indeed, there are important cultural and educational goods that may be served by open dissemination of such materials. However, the normative case for doing so is distinct, and lies beyond the scope of this paper.

Illegal sharing of academic research seems to fare well when wide proportionality is concerned. Beyond the potential infringement of copyright, interference with (morally suspicious) contracts, and some minor impact on the financial interests of academics—interests which, as noted, are generally modest—such sharing rarely harms them. On the contrary, empirical evidence suggests that research made available through illicit channels often receives a boost in citations, thereby advancing the professional interests of the authors concerned.99 Indeed, to the extent that copyright law aims to promote the dissemination of knowledge, secure attribution, and foster intellectual progress, the widespread circulation of academic research through shadow libraries may, in some cases, advance rather than undermine the values underlying copyright. In this respect, illegal sharing is not incompatible with and does not subvert important values underlying copyright (i.e., those discussed in section III.B). In fact, it can realise some of these values more effectively than existing publishing practices. After all, it is difficult to see how locking research behind paywalls or pursuing litigation against private networks and shadow libraries contributes to social welfare or fosters a vibrant culture of ideas or creativity.

Turning to narrow proportionality, it is unclear whether illegal sharing meaningfully harms actors responsible for enforcing copyright and licensing regimes. As for publishers, such practices may affect their financial interests, potentially reducing revenue streams or threatening the viability of certain journals. In some cases, this can lead to unemployment or to cause publishers to go out of business.100 However, these harms have not proven significantly disruptive to the academic publishing industry. The empirical record also does not aid the concern that illegal sharing would usher in an anarchic “grab what you can” knowledge regime. As discussed in Section III.D, copyright may function as an imperfect but coordinating convention. Yet despite the proliferation of shadow libraries over the past fifteen years, academic publishing has not collapsed into disorder. On the contrary, despite the growth and continued presence of shadow libraries, commercial publishers have reported record profits in recent years—a development that is hardly surprising given their low operating costs, reliance on unpaid academic labour, and oligopolistic control of the market.

Overall, illegal sharing is a proportionate remedial response. It grants marginalised research communities free and immediate access to scholarly work, it is often compatible with values underlying copyright, and promotes key epistemic values associated with academic inquiry. Shadow libraries also serve a vital archival role, preserving academic research across a range of servers and physical media—a task that commercial publishers rarely undertake.101 These significant benefits are achieved without imposing serious harms on researchers, publishers, the stability of copyright conventions, or the state.