In Open Access’s Long Shadow – A view from the Humanities

Historians have been in recent years among the most vocal critics against open access to scientific literature. Discussing the controversies they have triggered in Europe and in the USA, we argue that research on open access should be broadened chronologically and thematically. The first section recalls the very first debate on open access that took place among library professionals at the turn of the XXth century and points similarities with the present situation. The second section reviews the criticisms levelled by humanities disciplines against mandatory regulations on open access. The third section argues that the potential of open access for science democratization and knowledge dissemination may not be taken for granted and need further empirical assessment.


Introduction
Open access was formalized internationally in 2002-2003 by the Budapest, Bethesda, and Berlin declarations as financial, legal, and technical barrier free Internet access to scientific information for everyone. 1 The object of open access varies across the declarations, although journal articles and monographs are the primary focus, as increasingly reflected in the requirements of governments and funding agencies.
The formal definition does not exhaust, however, the full spectrum of what open access has become in the modern higher education landscape. In the twenty years since its creation open access has grown to be a central tenet of science policy, a cause to embrace for free knowledge activist, as well as a commercial strategy for major scientific publishers. As a consequence, the term "open access" has turned into a word with many hats, a plurisemantical portmanteau that conceals partially irreconcilable concepts of scholarly publishing.
The polymorphic nature of open access creates difficulties, notably among humanities scholars. In fact, parts of the humanities, especially those that are long established as academic disciplines, have in recent years repeatedly expressed criticisms against some of open access principles and implementations. Among other scientific fields, by contrast, there is a widely shared endorsement of open access, as part of a broader commitment to make research output as widely accessible as possible.
The present article looks at open access in the humanities through the particular lens of historical sciences. Historians and historical learned societies in various countries have voiced in recent years strong criticism against some aspects of open access. Each of these statements sparked numerous reactions, in form of commentaries, position statements, studies, and surveys, which form the work's principal source corpus.
The object of this work is not to give points among the contestants of this controversy. Much more, it uses the debates it sparked as a key to "deterritorialise" open access -to borrow the concept of Deleuze and Guattari (Deleuze and Guattari 1977) -i.e. to take a critical look at some of the layers that have coalesced into the open access movement in the twenty years since its modern reinvention.
Taken as an object of study on its own, open access has the potential of producing crucial insights of how the Internet and web technologies, ideologies, public and private interests, and professional practices interact to shape the transformation process of scholarly communication.
The article is divided into four sections. The first one is an attempt to put the contemporary debate in a broader historical perspective by recalling an earlier debate on "open access" that took place at the turn of the 20 th Century.

"The battle of the books"
"Open access is the Home Rule question 2 of the library world, and has provoked much warm discussion without any appearance of terminating in a decision which shall satisfy both sides. The controversy, indeed, may almost be said to have evolved this curious featurethat it appears well-nigh incapable of logical, unbiased, and generous argument" (Doubleday 1899:187). This statement originates from an article on "The open access question" published in 1899 in The Library journal. The title refers to a debate that took place between library professionals at the turn of the 20 th Century. In its original meaning, open access designated the practice of admitting readers to the stack-rooms of public libraries and allowing them to browse among the books on the shelves. This represented a radical novelty for the time and one where library professionals passionately debated the associated risks and opportunities for more than fifty years.
The above open access principle was first introduced in England by James Duff Brown in The Library journal in an article bearing the title "A Plea for Liberty" to Readers to Help Themselves (Brown 1892). Brown had visited the Unites States where he had seen the open access system in operation. A few years later, he implemented the system in the Clerckenwell library in London. By 1899, fifteen English, public-rate supported libraries had adopted the open access system.
The advocates of the new system were enthusiastic about it and tended to endorse it radically; a first point mirroring the situation today. Open access was "the system of the future" (Jast 1904:140). "That inevitableness and simplicity which distinguish every absolutely true principle struck at one in his system" (Moore 1899:52). Other library systems were to be abandoned as a consequence. "The most immaculate of systems, we readers should cast aside for the most rudimentary, if the latter opened out to us sources of knowledge closed to us by the former" (Moore 1899:58).
The new system seemed to be in line with the Victorian ideal at the basis of the Public Library Act of 1848: to educate large portions of the population to higher moral standards. "The mere fact of being able to go to a shelf, see the books, handle them, take them down and look into them, enables him [the reader] finally to find something that he really wants and something which will encourage him to come again and again until he becomes a steady and regular reader and a person who is gradually developing and improving his mind" (American Library Association 1899:52).
Another argument made in favour of open access and evoking the present situation was that the public, as the collective owner of public libraries, had a right to access books without restriction of any kind. "We have in the public library the people's book, paid for by their money and deposited in libraries for their use. This use should not be restricted in any way which is not clearly necessary to guard the people's interest." No argument could be made in favour of the "principle of imprisoning books" (Moore 1899:53).
Further major advantages were credited to the system, most noteworthy that the circulation of books would be increased dramatically and the cost of library staff would be reduced.
But some librarians wouldn't accept those arguments unchallenged, and attacked the rhetoric used by open access advocates. "At first blush a scheme that provides for the direct admission of the public to the books upon the shelves may seem so obviously the best as to admit no discussion. It is ideal! (...) But there are others who doubt the solidity of its advantages; and its superiority at once becomes a mere matter of opinion" (Doubleday 1899:188).
Critics warned about a number of risks that required further examination before consideration of adopting the new system. Librarians should "carefully consider the question, otherwise they may embark on what they think is a smooth sea, but will afterwards find out, to their own regret, that in consequence of the storm and the choppy billow there will be great difficulty in keeping afloat" (Chorton 1898:282).
Several major risks were identified. One of them was the theft of books. Accounts of stolen books attributed to open access were published by critics in journals and newspapers, only to be subsequently contested or explained by other causes by defendants. The CC-BY 27 reduction of running costs was also a disputed issue. The open access system would require more labour costs, not less (Doubleday 1899:192). Misplacement of books on the shelves would lead to further additional work for library staff. More worrying still, books would also suffer extensive damages resulting from excessive wear and tear, thus compromising the preservation of the books and the long-term sustainability of the library (Chorton 1898:13).
Another concern was the difficulty in producing reliable statistics on how many books were actually being consulted by patrons. This argument was related to another criticism of open access, the fact that users would not use the library catalogue any more. "The readers as a body having direct access to the shelves, do not use the catalogue nearly so much as in libraries where other systems prevail, while many do not use them at all" (Cotgreave 1898:284).
Lastly, some considered that the educational value of the system had been overstated. "Too much is probably made of the educational value of permitting the public to rummage the shelves" (Doubleday 1899:193). The argument about the educational benefit of open access was even turned on its head. "The open access system not only leads to disorder but to dishonesty. It is a system which expects too much from weak humanity" (Cox 1899).
The debate became heated over its course, eventually leading to a point where "not merely divergent but diametrically opposing views are honestly held and earnestly maintained" (Doubleday 1899:187). Anonymous pamphlets were circulated, one of them with the title "Open Chaos alias Open Access" (Johansen 2003:81). The controversy was referred to as "The Battle of the Books" in the newspapers, manifestly in reference to Jonathan Swift's 1704 homonymous novel on the respective merits on the Ancient and Modern writers (Johansen 2003:75). These circumstances lead to the emergence of a series of simplistic oppositions, such as "semi-lunatic optimists" against "fearful" critics, "inertia" against "progress", or the "lazy librarian" against the "genuine reader".
The fact that the librarians who opposed open access were generally also committed to the enhancement of public education and the circulation of knowledge, got lost among the hard feelings.
In reality, open access critics mostly ended up admitting the benefits of the new system, provided that certain safeguards would be guaranteed.
"Surveying the varying aspects of the question, (...) it is clear that unless open access is thoroughly safeguarded it must infallibly lead to anarchy and waste. (...) With sufficient safeguards, which ought not to be obtrusive or otherwise vexatious, and in a building adapted to its peculiar necessities, the plan in many respects is excellent" (Doubleday 1899:194).
In the following decades the system of open access to stacks was gradually adopted as the new standard for public libraries, soon followed by some research libraries (Johansen 2003:82

CC-BY 28
Like his predecessors sceptical about open access, Ratcliffe concludes that "there is a great deal to be said in favour of Open Access in large university libraries. My object is to (...) remove some of the complacency and fanciful thinking that surrounds it. I am convinced that simply to consider its advantages and ignore the real problems which go with them is a disservice to us all" (Ratcliffe 1968:110). The problem with most of the literature on open access in libraries, he went on, is that it "illustrates so classically the all too frequent gap between theory and our experience of practice."

From old to new open access
We are today in a situation similar to the past debates on open physical access to library stacks and, long before that, access to the libraries themselves. Modern open access, focused on financial, legal, and technological aspects, is regarded by almost everyone as having some good principles built into it, but opinions diverge on how to implement them. While its proponents are calling for a swift adoption in the name of the greater good or the democratization of knowledge and tend to see the new position of openness as inhabiting a sort of moral high ground, critics tend to discard the discourse of the former as rhetorical, and call for empirical assessments of the challenges and for the possibility of implementing safeguards.
Some arguments have remained the same: the new system will achieve a significant increase in the circulation of scientific literature while reducing costs at the same time; the taxpayer should have a right to access research he or she has been paying for through taxes. Conversely, some of the criticisms are also similar: open access is posing a threat to the established organisation of knowledge; it opens spaces for abuses like theft and predatory publishing; it undermines certain information literacy practices, posing new challenges for the discoverability of knowledge.
Both debates revolve ultimately around the problematic gap between expected benefices and the practical outcomes resulting from its implementations. Jean-Luc Guédon, one of the founding fathers of "modern" open access, acknowledged the problem in a recent article looking back at 20 years of open access: "The advent of computers and networks made clear what the solution could look like. But when the moment came to implement the vision (or the dream), pragmatic difficulties quickly became obvious" (Guédon 2017:7). The third section below discusses extensively those "pragmatic difficulties".

Genealogy of modern open access
Although still young, the history of modern open access has been sketched out by various proponents such as Peter Suber and Martin Paul Eve (Eve 2014;Guédon 2017;Suber 2012). Some of its important milestones can be summarised as follows.
In the late 1970's, the equilibrium of the system of scientific publishing maintained since World War Two started vacillating. Reduced public funding increasingly exposed the higher education sector to the whims of the market and pressure rose to increase the economic returns of scientific research (Donoghue 2008). In the following decade, scientific publishers, especially in the natural and hard sciences, evolved into highly concentrated global industries, starting to adopt monopolistic strategies and raise their prices drastically. Today, four major scientific publishers alone account for more than 30% of the market of scientific journals, which was estimated at 25.2 billion dollars in 2013 (STM 2015:23). In the last twenty years, publishers started to licence access to their content via online web platforms, thus increasing reliance on their services. Research libraries have been struggling to cope with journal prices, and, as a consequence, have been forced to reduce other acquisitions, notably monographs.
Paralleling these developments, the creation of Richard Stallman's Free Software Foundation in 1985 and successively of the Open Source Initiative in 1998 spread among software developers and computer literate scholars the principles of free software and the idea of their possible extension to other types of materials (Pomerantz and Peek 2016).
Almost immediately after Tim Berners Lee released the World Wide Web standards in 1991, a community of scholars set up an online database to share drafts of scientific articles. This scholar-led initiative was thought up as a solution to allow scholarly communication to keep pace with the acceleration of scientific research and publication in the experimental sciences. ArXiv, as it was named, is often cited as the first open access repository and is to this day one of the most successful open access publishing platform (Ginsparg 2011). During the 1990s, a series of similar scholarly-led initiatives emerged, some of them in the Humanities, which offered free online access to their research outputs (Open Access Directory 2015).
A few years later, other parties, most notably libraries, started to get interested in the free access to scholarly literature on the Internet. They had been going through the computerization of the library environment and were willing to experiment with new ways of sharing knowledge through the World Wide Web (Borgman 1997). They were motivated in doing so by the disproportionate rise in the costs of scientific journals, known as the "serial crisis", which prompted them to look for ways to reduce the cost of access to scholarly literature.
By the early 2000s, the label "open access" had emerged to identify an agenda aiming to make scientific literature available online for free. The Budapest Open Access Initiative can be credited for coining the term and coming up with the first definition: "By 'open access' in this literature", it reads, "we mean its free availability on the public Internet, permitting any users to read, download, copy, distribute, print, search, or link to the full texts of these articles, crawl them for indexing, pass them as data to software, or use them for any other lawful purpose, without financial, legal, or technical barriers other than those inseparable from gaining access to the Internet itself" (Open Society Institute 2002).
The authors of the Budapest declaration had like their nineteenth century predecessors great expectations about the expected outcomes of their movement. Open access would promote "unprecedented public good": "removing access barriers to this literature will accelerate research, enrich education, share the learning of the rich with the poor and the poor with the rich, make this literature as useful as it can be, and lay the foundation for uniting humanity in a common intellectual conversation and quest for knowledge" (Open Society Institute 2002).
To As we have just seen, different actors came to be involved in the open access movement at different times and driven by different motivations. Pioneering scholar-led initiatives first appeared in the early 1990s, followed by libraries and interest groups by the end of the 1990s and finally universities and research funding organisations from the early 2000s. Those different timings also included different agendas, which soon led to frictions in the humanities about implementation of open access mandates, as we will see in the next section.

Open chaos in the humanities
When the first open access mandates were being issued some disciplines of the humanities felt that those policies were failing to take into account their specific needs and were putting some of their basic principles at risk. The "traditional" and "closed" world of scientific publishing was opposed to the "open" and "free" world of open access (Fitzpatrick 2013). Resistances against open access were criticized as "locking up ideas" (Cohen 2013). In the American case, the hashtag #AHAgate emerged on Twitter like a modern-day pamphlet of yore.
Digital humanities practitioners have been among the most vocal critics of historians and their learned societies, accusing them of "always looking backwards" (Leon 2013), "turning the clock back" (Fister 2013), and being "myopic and rebarbative" (Ramsay 2013). Attacks against historians culminated in the following statement by digital humanities professor Andrew Prescott. 027.7 Zeitschrift für Bibliothekskultur 6(1), S. 25-47. ISSN: 2296-0597 "The only time that the corduroyed Colonel Blimps of the British historical establishment have grudgingly bestirred themselves from their deep slumber to engage extensively with digital matters was when they belatedly realized that changes in open access might upset the cosy financial arrangements that provided a life-support system for ailing learned societies, and a hasty rearguard action was mounted to try to preserve the status quo." (Prescott 2014:340) Conversely, some historians assumed radical positions against open access and rejected the project altogether. The Swiss historian Michael Hagner stigmatised it as a dangerous utopia that combined "salvation expectations, neo-liberal accelerationist phantasies and antiintellectual populism" (Hagner, 2014).
Others found that open access in general was a project that made "no sense" (Osborne 2013).
The crux of the argument by the historical establishment was that open access policies were designed to address first and foremost the STEM disciplines and were therefore out of touch with the publishing standards of the humanities. As in the case of the open shelving debate, they were warning against adopting a new system without thorough assessment of the risks and consequences involved for the humanities.
The centrality of the monograph in the humanities, for instance, had been overlooked by open access policies, and introducing open access for monograph publishing would require a different set of considerations than in journal publishing, it was argued.
According to the Royal Historical Society, humanities journals, despite the fact that they require more work to copy-edit and to perform peer-review, feature longer articles and include commissioned content, are comparatively quite inexpensive and do not represent an intolerable burden for libraries as their STEM counterparts do (Royal Historical Society 2014).
Also, the half-live of articles -a measure that describes the span in which half of the total consultation of an article is going to happen -is much longer in the humanities, which means that they retain their scientific value for a long time. Studies suggest that a journal article in humanities has a half-life of 3.5 years, while the half-life of articles on the JSTOR database, which contains retro-digitized collections of printed humanities journals, is as long as 20 years. 6 This reflects the specific cumulative nature of humanities research, which produces new interpretations on previous research, but hardly replaces it or makes it obsolete.
In addition, scholarly publications in the humanities tend to be published not only by a handful of major scientific publishers, but by a variety of stakeholders. Those include academic presses, learned societies, small publishers, mass market publishers, and initiatives led by scholars. Many of these have fragile business models and are vulnerable to disruption. Larivière has evidence of the on-going loss of diversity of the publishing landscape in the social sciences and the humanities since the 2000s (Larivière, Haustein, and Mongeon 2015:4). Therefore, historians have argued, safeguards are needed in order to avoid that open access mandates result in driving existing publishers into bankruptcy and depleting diversity in the publishing landscape of the humanities.
The main safeguard identified is that of embargo periods, within which publications are exempt from open access release. Embargo periods, history journal editors have claimed, allow publishers to retain sales monopoly for a sufficient period of time to cover their cost before the articles are released in open access.
Over the following section, we will report some of the criticisms that have been levelled against current implementations of open access. The sources used stem largely from the historical community, but other commentaries from academics from across the humanities are also taken into account. Three topics will be addressed Initial criticism from the humanities in the UK was addressed to the preference for the gold route, as for most humanities scholars it is extremely difficult to access APC funds. The dispatch method of APC funds induces a centralising effect that favours big research-intensive institutions against smaller or teaching-oriented ones. Moreover, it gives the university administrators in charge of dispatching these funds unprecedented power to decide how to allocate these funds. This was lamented as a further expansion of administrative control, already considered as invasive by many scholars (Boffey 2013). More generally, independent scholars, as well as academics moving between institutions or doing research outside the framework of research grants wouldn't have access to APC funds, putting their capacity to publish at risk.
Additionally, the mandatory nature of open access mandates implies that only specific journals would count as compliant with the open access policy. A journal having no open access policy, or offering longer embargo periods than requested by mandates would therefore be considered non-compliant, and would not be admissible as publication venue by science funding bodies. Researchers would therefore be limited in their choice of the most appropriate publication venue, or run the risk to be penalized by their institutions. This would also endanger international cooperation, as in the UK for instance, more than 30% of articles in history are published abroad in journals that may not be compliant to the UK open access mandate (History UK (HE), 2013). For those reasons, historians have cited these issues as a "threat to 7 For a specific discussion of copyright in the humanities, see Adema (2015), Berry and Hitchcock (2017), Emmott (2013), and Feather (2010). For the position of UK historians on the subject, see Crossick (2015), Mandler (2013), and Social History Society (2013). 8 The section title is borrowed from Harnad (2012). academic freedom", and the gold route model including author's fees is generally considered inoperable in the humanities (Boffey 2013;Collins 2013;Waltham 2010).
Other perverse effects are attributed to gold open access with APCs, like the fact that it inevitably incentivises quantity over quality, because a publisher's revenue relies on the charges collected from authors. This further exacerbates the "publish-or-perish" imperative, at the expense of the quality of the research (Ball 2012). In parallel, a number of rogue scientific publishing houses have emerged, that basically publish anything that is submitted to them provided that the author pays a fee (Beall 2015).
To make things worse, in conjunction with the rise of gold open access publishing, scientific publishers have been focusing on so-called "hybrid journals", where authors can optionally pay an APC to have their article immediately released in open access, while the rest of the content is accessible through subscription (Pinfield, Salter, and Bath 2015;Research Councils UK 2015). Hybrid publishing thus forces libraries to maintain journal subscriptions while at the same time authors pay publishing fees. This commercial strategy has been criticised as "double-dipping", because the tax payer therefore ends up paying twice. 9 Overall, with the emergence of APCs, it started to become clear that any promise open access had shown in terms of reducing the cost of scientific publications was ill-founded. A significant number of studies looking at the costs of the transition to open access have been commissioned. They have faced significant methodological challenges and sometimes came up with conflicting results. It became generally accepted, though, that if the transition to open access could bring long-term benefits, in the middle term it will be at best cost neutral, or would likely imply a rise in the total costs of scientific publication (Cambridge Economic Policy Associates 2017; Houghton, Rasmussen, and Sheenan 2009;Pinfield, Salter, and Bath 2015). Like in the previous open shelving debate, the presupposition that open access would help reducing running costs did not stand to the facts.

Green technocracy
Green open access -i.e. the self-deposit by researchers of their own research output in disciplinary or institutional repositories in parallel with publication in a recognized journal (either with or without restrictions of version or timing of deposit) -was identified from the outset of the open access movement as one of the two routes likely to bear fruit. The main infrastructure required to operate green open access are repositories. Repositories are basically online databases where the scholarly materials are stored. From the early 2000s, most higher education institutions started operating their own repositories. They generally use a common protocol for metadata harvesting that enables their interoperability and allows searching (Open Archives Initiative 2001).
The response of academics in general and of scholars in the humanities in particular has been very slow however. As a British historian puts it: "institutional repositories are there, ready, waiting -and empty" (Matthews-Jones 2013). Taking account of this situation, several models emerged to push academics to deposit their work. The most frequently cited models are those of the University of Liège, where only publications deposited in the institutional repository are taken into account for the promotion of faculty, and the Harvard model, where academics are required to grant to the university by default the right to publish and distribute their publications through the institutional repository. In the UK, a project inspired by the Harvard model is under study.
In the humanities, as we have seen, the debate about green open access focussed on the length of embargo periods. The embargo periods allowed by research funders varies. In the UK, Research Councils UK, the national scientific policy coordinator, accepts a 12 months' embargo for articles, while the Higher Education Funding Council for England, funding agency, tolerates 24 months' embargo (HEFCE 2014;Research Councils UK 2012). In Switzerland, the duration is 6 month for journal articles and 12 month for monographs (Swiss National Science Foundation 2014a). The Horizon 2020 open access European policy sets 12 month for journal articles in the humanities (European Commission 2017).
Against this background, the aforementioned letter by history journal editors stated that the minimum embargo length to preserve journal sustainability was 36 months. There is no easy way to settle this issue. Both the effects on sales of repository publishing and the length of embargo periods remain heavily debated issues. A recent study in Switzerland could demonstrate that free online availability of scholarly monographs doesn't impact sales, but overall, empirical evidence to sustain differing views remain scarce (Ferwerda et al. 2018;Slauter and Wulf 2014).
There have been other issues raised against green open access, mainly dealing with institutional repositories. Repositories are generally considered unsuitable venues for academic publishing as they focus more on storage than use (Fitzpatrick 2011). The fact that they usually host a diverse range of content that includes research reports, dissertations, and administrative documents further contributes to this negative perception. Moreover, many academics consider repositories a part of the administrative apparatus of the university and see the obligation to deposit a copy of their publications as an additional burden imposed upon them. Finally, the costs and forms of labour required to maintain the infrastructure for storage and availability of digital objects (metadata production, software development, software migration) tends to be overseen and represents an additional risk for the sustainability of scholarly communication (Eve 2017).
As many publishers deny academics the possibility to deposit the published version of their work in repositories, researchers are forced to deposit a version of their manuscripts instead. As a consequence, more than one authoritative version of a given piece of work may be available and page numbers or layout may vary between different versions, which is regarded as having an adverse effect on the stability of publications. Stability embodies the accreditation function of publishing and allows citation, and is therefore crucial for scientific literature. Permanent identifiers and technologies allowing the versioning of articles are currently being implemented to deal with those issues, but for Jean-Claude Guédon, digital publications are still in an age of "digital incunabula", where adequate solutions are still lacking (Guédon 2017).

"Good books need time and paper"
Monographs remain a central, if not the most important vector of scholarly communication in the humanities. Publishing a scholarly monograph through the established commercial or academic press is "for the humanities, what Nature and Science represent for the natural and life sciences: the most important stage both within and outside the academy" (Hirschi 2014). In the competition for scientific authority that stands at the centre of the scientific field, the monograph represents the most prestigious achievement in the humanities (Bourdieu 1975).
Although monographs are not yet included in UK open access mandates, the push towards open access for monographs is getting stronger. In Switzerland, the Swiss National Science Foundation introduced an open access mandate for subsidised monographs with a maximum two-years embargo in 2014 (Swiss National Science Foundation 2014c). On a similar note, most universities require PhD students to deposit their thesis in their institutional repository. In 2013 the American Historical Association (AHA), worrying about the fact that these policies could make it harder for students to publish their work as monographs, released a statement urging universities to allow a 6 year embargo for dissertations (American Historical Association 2013). According to its director, "free open-access policies for online dissertations will be one more nail in the coffin of book writing, book publishing, and book reading as we know them" (Cronon 2013).
There is widespread evidence indeed that the business of publishing monographs has become increasingly difficult. Academic presses, traditionally responsible for publishing scholarly monographs, are facing since the 1980's declining sales figures combined with shrinking library budgets (Fitzpatrick 2011). This has led publishers to be increasingly selective and reduce print runs. Nowadays, print runs for scholarly monographs in the humanities are as low as 200-250 copies in the UK, and around 600 copies in the US (Eve 2014:15;Straumsheim 2016).
What the debate sparked by the AHA statement has clearly shown is not the threat of the imminent death of the scholarly monograph -a prospect which is contradicted by evidence (Crossick 2015:5) -but rather the fact that the physical book printed on paper is still largely considered as a proxy for quality and a must-have to pursue a career in the academy (Antley 2013). That "good books need time and paper", as stated by a Swiss history professor, is a widely-held view within the humanities (Hagner 2014).
Thus the major obstacle in the way of open access for monographs lies arguably in the mechanisms of tenure and career advancement. As digital publishing is still considered by many in the humanities to be less prestigious than print, scholars tend to be reluctant to engage with it (Crossick 2015:6;Dalton 2008;Harley et al. 2010). As these sources show, humanities scholars generally assume that in order to get tenure, having a printed monograph published by a university press is a much stronger scholarly asset than having a digital-only publication. Although researchers routinely read publications -including monographs -in digital formats, they are still largely reluctant to engage with digital-only publishing when it comes to publishing their own books as a matter of course (Bulger et al. 2011). This leads to a paradoxical situation where humanities researchers favour more and more digital-based publication and mostly agree that the print monograph publication system is broken but are still committed to engage with it and to identify themselves as staunch defenders of this tradi-tion. As Jean-Claude Guédon puts it, "scholars and scientists are essentially schizophrenic beings; as authors, let me call them Dr. Jekyll; as readers, they become Mr. Hyde" (Guédon 2001).
Other scholars still consider that digital publishing technology is not mature enough for monographs. Most digital monographs are still merely PDF files, failing to harvest the potential of digital media, and the experience of reading on screen is seen as a poor substitute for reading on paper (OAPENUK 2014).
Some humanities scholars tend to see the opportunities offered by digital monographs for embedding more than text, include non-narrative arguments and embed books in virtual research environment. Others, among them many historians, tend to stick to the importance of traditional book writing as a central feature of their discipline (Crossick 2015:13-16). Nonetheless, opinion converges in considering open access monograph publishing as a good expedient to grant access to publication for young scholars and to expand the dissemination of scholarly monographs.

Is open access synonymous with public good?
There is no question that the "open" movements have been very powerful in challenging the status quo of scholarly communication and bringing together an array of different actors. But in their rallying power rests their ambiguous nature, as they tend to flatten all differences between disciplines and between implementation contexts, and to make disappear the diversity of particular interests that they bring together. Therefore, we should be very prudent in considering open access a perfect expression of science ethos in the digital age and a one-size-fits-all solution for scholarly communication.
In 1985, Daryl E. Chubin, a social scientist with a career at the top levels of American science institutions, who was presumably very well acquainted with the complex logics of science funding, made a clear point about the science ethos in an article called "Open and Closed Science; tradeoffs for a democracy".
"The road to scientific virtue is paved with diverse motives. The norms, which sanction and shield those motives may not be 'best' for all research communities nor for all time. Scientists, like other professionals, have their own interests and values. They must defend their autonomy while eliciting investments in research from public and private patrons. Thus, they cannot be relied upon to guarantee an open, universalistic science: there is no such thing. (...) Openness is an interest-bearing idea; it cannot be settled with recourse to facts or logic. It is a matter for political debate, not scientific judgment alone." (Chubin 1985:80) In this quote Chubin argues against what we could call a neo-platonician understanding of openness in science, which would consider "openness" to have a static and almost transcendental meaning, in favour of a situated understanding of openness, which acknowledges its fundamentally political nature. Mertonian ethics of science (Merton 1973), while at the same time critics are resisting against open access policies in the name of the same set of ethical values. This leads to a point where, once again, "not merely divergent but diametrically opposing views are honestly held and earnestly maintained" (Doubleday 1899:187).
In order to move beyond those infertile oppositions, we would like to point at a couple of issues that have been raised between the lines of the debate but rarely directly addressed. These relate to common assumptions on the evolution of knowledge regimes in the age of the Internet and ubiquitous computing.
Three of such assumptions will be outlined in the next section: the idea that technological progress will necessarily lead the way to an open science ecosystem, the idea that business models based on free access are inherently more fair than others, and the assumption that contents that are freely accessible online are having an enhanced impact of the wider public and society at large.

Technological determinism
Technological determinism is the reductive assumption that technological innovation is the dominant and most important factor that leads to changes in established social practices. From an epistemological viewpoint it is biased, because it obscures the fact that the adoption of new technologies by social groups is in no way determined by the simple availability of a new technology, but is rather the result of interacting economic, societal, cultural, and technological factors (Rogers 1971).
Open access provides an interesting case study into this matter for at least two reasons. First, discourses making the case for open access have often been imbued with inherent technological determinism. The Budapest declaration, for instance, begins with the following line: "An old tradition and a new technology have converged to make possible an unprecedented public good" (Open Society Institute 2002). The open access presentation on the website of the Scholarly Publishing and Academic Resources Coalition (SPARC) goes further: "This fundamental mismatch between what is possible with digital technology -an open system for communicating research results in which anyone, anywhere can contribute -and our outdated publishing system has led to the call for Open Access" (SPARC 2018). As recently acknowledged by Jean-Luc Guédon, this kind of formulation confers "a form of historical necessity to Open Access" (Guédon 2017:1).
Secondly, as we have seen in previous sections, the mere availability of a series of new technologies that could enable the establishment of a comprehensive open access environment was not nearly enough to convince humanities scholar to abandon their "outdated" practices, nor did they succeed in forcing global scientific publishers to discontinue their dominant position. On the contrary, as discussed in the next section, the success of open access can be attributed to the fact that is promotes a vision of the relationship between science and society that is favourable to private actors in the market economy.

Neoliberalism
The term "neoliberalism" keeps coming back in the critique of open access. Authors on both side of the barricade note that open access fits disturbingly too well into a neoliberal view of science, where the immediate free flow of scientific information is seen first and foremost as a strategy to fuel economic growth.
For the historian Peter Mandler, the reason why open access is being pushed forward is primarily to serve the interests of the economy, and not for the benefit of the public good (Mandler 2013:553). This also explains why policy makers explicitly favour gold open access and licences that allow for commercial and derivative uses. Many policy statements in favour of open access explicitly acknowledge this ambition (Finch 2012).
For John Holmwood this is problematic because this process takes place in a context of commercial enclosure. Scientific literature and data ought to be given out for free, while knowledge produced under patents, or subject to commercial exploitation, is exempt from the requirements of open science. The fruits of scientific research are thus provided for free to businesses, which can then draw from it to develop commercial products that will be brought to the market in a re-enclosed form (Eve 2014:23;Holmwood 2013).
Even in the humanities, whose research results are not as attractive to businesses, David Golumbia considers that open access is undermining the value of intellectual labour and dispossessing academics of their work. In his Marxist critique, Golumbia writes that "the very point of OA, despite what its advocates claim, is to entirely brand the labour of intellectuals as unproductive vis-à-vis capital, and therefore to make available for exploitation that labour by everyone but the labourer" (Golumbia 2016:101).
Whatever one may think about this statement, it draws attention to the fact that open access bears disturbing similarity to the "free" Internet economy, where contents and data generated by users are given out for free in exchange for access and services. The data is then privately exploited for their corporate interest by a handful of dominant players with massive computing power. As Casey Brienza notes, "companies like Google, Amazon, Microsoft, and, yes, Wikipedia are positioning themselves as new gatekeepers to global knowledge" (Brienza 2012:166-167).
This global context does not exclude the possibility to benefit occasionally from those trends to streamline scholarly communication in the humanities. But the question of the overall political economy of open access publishing, as well as the question of the rise of new gatekeepers securing monopolies over the discoverability of freely accessible scholarly content, should not be overlooked.

The wider public
Lastly, the most commonly cited argument in favour of open access is its outreach to the wider public. But who and where is exactly this "public"? The notion of the public that should benefit from open access to scientific literature is not often questioned, and, some argue, is understood in a naïve way.

CC-BY
This question seems at first to call for an obvious answer, but has not been empirically settled yet. A recent research survey finds that there is surprisingly little evidence available, and that the societal impact of open access still needs to be systematically investigated (ElSabry, 2017). Data on this topic is difficult to gather, although computers make it easy to monitor traffic, because the metrics of online access -downloads, sessions, page views -still poorly reflect the realities of knowledge dissemination (Piker 2017;Townsend 2013).
Therefore, Casey Brienza considers that the value of open access for the public good is consistently overstated. "Too many scholars seem to assume that if an article or a book is 'open access', it must necessarily then be available to all people equally and therefore in perfect alignment with the greater public good" (Brienza 2012:166). The effective dissemination of knowledge requires more than the sheer online availability of scientific literature (Allington 2013).
For the historian Robin Osborne the linguistic and disciplinary boundaries in which humanities research happens are a further barrier for their accessibility by a broader readership. Scientific literature, he argues, is firstly intended for peers and requires specific training to be understood (Osborne 2013).
For Michael Dieter this unsatisfactory situation is reflected by the fact that today the vast majority of research articles in the humanities never gets a single citation (Dieter and Lovink 2017). In this context, making an article open access could mean merely placing it in online databases for machine reading. "The only person who will read your article", he claims, "is a robot" (Dieter and Lovink 2017).
These criticisms do not deny the benefits in terms of accessibility of scholarly outputs that can arise with open access. Nor do they ignore the crucial need to enable developing countries to access and to participate to scholarly communication, or disregard the venerable practice of citizen science -transposed on digital networks by a number of successful crowdsourcing projects -or the uncountable individuals that are, yesterday as today, genuinely interested to help out in scientific research. What these critics rather point at, are the risks involved in taking for granted the benefits of online availability of scholarly literature without assessing the effective dissemination of knowledge happening through them.

Conclusions
The open shelving system, after being debated for more than fifty years, has become commonplace in public and research libraries, although some already think that digital access will soon make it obsolete (Barclay 2010). Open shelving is generally implemented according to specific library profiles and the system is widely accepted, although some of its acknowledged weaknesses, book losses for instance, remain an issue (Kean 2017). Significantly, what was once presented as a two-front controversy has evolved into a diverse spread of methodologies running the gamut from libraries that only make a few shelves of reference materials available through to libraries granting free access to the entirety of their stacks.
As for modern open access, the system has not yet reached this same level of maturity.
Although open access has now been around for almost 30 years, it is still controversial CC-BY 40 and there is every reason to believe it will remain so. Recent open access mandates, now being complemented by new policies regarding the publication of research data, are directly impacting humanities scholars, and are increasingly forcing them to engage with this brave new world.
Indeed a range of new initiatives endorsing a more differentiated approach to open access has sparked in recent years in the humanities, trying to overcome the shortcomings of classical open access models. Among them are publication services infrastructures offering software solutions, "consortial funding" schemes that rely on upfront financing by libraries, and New University Presses (NUPs) focusing primarily on offering a publication venue for in-house academics.
These initiatives all put a strong focus on traditional peer-review, with the explicit intent to confer the best scholarly standards to their publications and to counter prejudices against the scientific quality of digital-based publications. Many of them also offer commercial print-on-demand additionally to open access, thus incorporating the pro-print bias of the humanities into their business model.
On the economic side, most initiatives mentioned above rely on mixed revenue streams including several types of public and private funding. Interestingly, the idea that scholarly publishing has ever been an economically sustainable business at all is also being challenged in recent years. Fitzpatrick and Adema find that scholarly monograph publishing has always partially relied on some form of subsidy, and in parallel voices are being raised asking for a non-profit scholarly communication ecosystem (Pooley 2017).
These are very positive developments, we would argue. The majority of humanities scholars were largely uninterested in the dynamics of scientific publication until recently. Their lack of interest is reflected in their low rate of contribution to institutional repositories. To this end it is desirable that humanities scholars critically engage with open access, including by calling into question both its key principles and implementations. Although the humanities occupy a minor position in the open access policy making, it can make a strong contribution to the debate by providing an in-depth reflection on its wide-ranging implications (Kember 2014