The changing copyright landscape in academic publishing

Key points This article marks the publication of the Tenth edition Clark's Publishing Agreements: A book of precedents. The growing complexity of traditional assignment agreements is accompanied by researchers’ misunderstanding about their rights. There is a shifting balance between assignment and licensing of rights in an author's work. Academics continue to agree standard assignment contracts even where they disagree with the principles. Publishers are loosening control over copyright in response to demands from funders and a growing minority of researchers. There is a complex range of researchers’ attitudes to copyright issues.


INTRODUCTION
The publication of the tenth edition of Clark's Publishing Agreements: A book of precedents (Owen, 2017) provides an opportunity to reflect on the changing copyright landscape in academic publishing. Some of the tensions and forces at work are highlighted by the editor, Lynette Owen, in the preface to the new edition: '…copyright has remained under the microscope at both national and international level, and it is all the more important to achieve a fair balance between the interests of creators, the publishers who invest in bringing their works to market, and users of those works, many of whom continue to feel that content should be freely available' (Owen, 2017).
This article discusses the following themes: • The growing range of terms covered by author-publisher contracts.
• The shifting balance between assignment and licensing of rights in an author's work.
• Researchers' attitudes to the changing copyright landscape.
• Future trends in the field.

THE CHANGING SCOPE OF CONTRACTS
Contracts have always been at the heart of the publishing process.
An established model in the academic sector has been for the author to assign rights in a work to a publisher to exploit for their mutual benefit, with the emphasis on the publisher acquiring rights 'wide and long', both to maximise opportunities for revenue and to allow publishers to defend those rights if they are infringed (Owen, 2017). This arrangement is in contrast to trade publishing, where authors are more likely to retain rights as they may constitute important sources of additional income (e.g. serialization, film, or television rights). Because the primary goal of many academic authors has traditionally been publication in a prestigious journal, they have been more likely to agree to standard assignment terms to secure publication of their work.
Such agreements have become more complex over time given factors such as the increasingly diverse network of channels to market and the wide range of means to derive value from a product (such as subsidiary rights). As an example, the potential terms of a standard author agreement for a scientific, technical, and medical (STM) book (Precedent 2) in the new edition of Clark's now run to 17 pages (plus explanatory notes). A more recent development has been the electronic forms of a product and the ways of exploiting these new forms (e.g. accounting for income from site licenses for multiple access to a single work, re-purposing a work as an 'app' for use on mobile phones, or mining text for data). These developments are covered in a helpful Introduction to Electronic Precedents in the tenth edition as well as seven separate precedents covering electronic use (Precedents 18-24). Indeed, the tenth edition exemplifies the new digital environment with its own provision of an online resource of customizable precedents for publishers to adapt for their own use (replacing the CD, which accompanied previous editions).

THE SHIFTING BALANCE BETWEEN ASSIGNMENT AND LICENSING
Whilst alternatives to the model of assignment have always existed, perhaps the most important recent trend in this area has been the open access (OA) movement, which has encouraged a greater shift towards the retention of rights in a work by the author and towards granting a license by the author, which controls the publisher's use of a work. This reflects the moves by funding bodies and others to exercise greater control over the research they sponsor to ensure the widest possible readership.
Research bodies have become more proactive in encouraging and sometimes requiring researchers to retain rights, for example, by making funding conditional on publishing research results in an OA forum of some kind. As an example, one of the major funders of science research in the UK, the Wellcome Trust, states its support for 'unrestricted access to the published output of research as a fundamental part of its charitable mission and as a public benefit to be encouraged wherever possible' (Wellcome Trust, 2017). This policy is enforced, for example, by requiring all that published research funded in whole or in part by the Trust be made freely available on the OA platforms PubMedCentral (PMC) or Europe PMC. The National Institutes of Health (NIH), a major funder of medical research in the USA, has a similar policy (National Institutes of Health, 2017). These policies are consistent with both main OA models: 'green' OA, where published work can be made freely available after an embargo period (e.g. 6 months), and 'gold' OA, where published work is made immediately available, with the cost of publishing covered by an article processing charge (APC).
Some research bodies provide model agreements or addenda that authors can use to replace or modify standard publisher agreements that might limit meeting OA requirements. An example in the author's area of publishing is the guidance provided by the CGIAR (formerly the Consultative Group for International Agricultural Research), which coordinates global research in agricultural science research (CGIAR, 2017). Indeed, it has been said that the OA policies of some funders and employers require researchers to make OA work 'even when they publish in a nonopen-access journal, even when the non-open-access journal does not give standing permission for open access, and even when faculty members have not negotiated special access terms or permissions with their publishers' (Suber, 2013).
The new edition of Clark's reflects this shift, for example, in Precedent 8, which covers contracts for authors of journal articles. This includes three different model agreements: • A license to publish.
• An assignment of copyright.
• A Creative Commons license.
As the inclusion of the latter indicates, a key element in building a contractual framework for OA has been the development of a set of standardized and widely used Creative Commons licenses to govern OA use (summarized in a new Appendix G in Clark's). Developed by the Creative Commons organization, founded in 2001, these licenses set out standard terms governing the use of an author's work by others. The four main types offer varying degrees of control over how a work can be used, covering attribution, types of use, and adaptation. They include: • These can be compared with the model OA licenses developed by the STM, the principal body representing STM publishers. These licenses cover issues such as definitions of permitted private use and public access, commercial and noncommercial use, translations, and data mining (STM, 2017a).
A recent initiative has been the development by a UK group of senior university librarians and academics of a standard Scholarly Communications Licence. This is a model OA policy with a standard set of licence terms designed for use by universities to supersede any further restrictive terms offered by publishers (http://ukscl.ac.uk/). It favours the CC-BY-NC type of Creative Commons license, allowing copying, redistribution, adaptation, and non-commercial reuse. This initiative has been criticized by publishers on a number of grounds, including the emphasis on immediate non-commercial use, which undermines the embargo period that underpins the green OA model (Publishers Association, 2017).
Over time, publishers have adapted to pressures for greater freedom of access to and circulation of research by ceding some of the intellectual property rights they have traditionally exercised. This process has ranged from allowing some limited free use of material in publishers' copyrighted works by other researchers (through rules first developed by the STM in 1979 (STM, 2014) to allowing authors greater freedom to re-use material from their previous publications in new works for other publishers. Occasionally, the industry has had to highlight how far publishers have moved in this respect in the face of accusations of restricting author freedom (STM, 2008). The current wide range of permitted uses of journal articles by their authors can be seen, for example, in Elsevier's policy on author and user rights (Elsevier, 2017).
Reflecting the OA requirements of funding bodies and others, many academic publishers now have policies recognizing the right of authors to post versions of works (typically the accepted article prior to production) in institutional repositories for others to access freely. The SHERPA RoMEO database, which lists publishers' policies in this area, shows that 80% of the more than 2,400 publishers on the database support archiving either the pre-print (the version before peer review) or the post-print version (the accepted article after peer review) in an institutional repository (SHERPA RoMEO, 2017). Most recently, the STM has allowed a further loosening of copyright by agreeing on a policy permitting limited circulation of published journal articles within scholarly collaboration networks (SCNs), whether or not they are covered by OA licenses (McKay, 2017;STM, 2015).

RESEARCHERS' ATTITUDES TO COPYRIGHT
The shift towards a more diverse copyright landscape, with a wider range of agreements to reflect differing degrees of control over work between authors and publishers, has been further complicated by the attitudes of the researchers themselves. The research community appears to be divided by the moves by governments and funding agencies to allow broader access to and use of researchers' work. Several studies have found that researchers still value journal reputation, target audience, quality of peer review, and impact over whether a journal is OA, with a similar picture for monographs (Fyffe et al., 2017;Jubb, 2017; Van den Eynden, Knight, & Vlad, 2016). Many also show a limited interest in depositing their work in an institutional repository to allow wider access, a key element in OA (Collins & Milloy, 2016;Harjuniemi, 2012;Nicholas et al., 2016). Recent studies show that even early career researchers (under 35), who might have been thought to be more open to change, remain relatively conservative. One study has concluded that they 'have limited opportunity to change as they are shackled to a reputational system that promotes, above all else, publication record and citation scores' (Nicholas et al., 2016).
As this study implies, whilst researchers may feel obliged to continue to assign copyright in their work to publishers as a precondition to ensure publication in the right journal, this does not mean they are necessarily happy with the terms set by publishers (Harjuniemi, 2012;Lee, 2017). In many cases, researchers seem to agree to publishers' terms without fully investigating or understanding them (Rapple, 2017). A senior academic scholar in the UK has commented: 'I do wish that researchers would take more interest in the conditions of knowledge production and dissemination in the academy. Most researchers I know never really even read or query the contracts they are given by publishers' (Anonymous, 2017a). His comments are supported by a recent study that found 'a high degree of confusion and uncertainty related to copyright' and significant misunderstanding about the restrictions on depositing work in SCNs (Lovett, Rathemacher, Boukai, & Lang, 2017). There is similar confusion about the broader issue of sharing data that underpins published research as well as about intellectual property in general (Anonymous, 2016;Lyubetskaya, 2017;Soetendorp, Haberman, & Smith, 2016). Some of the challenges of understanding copyright and 'fair use' in an American context, and in recording copyright accurately, are discussed from the point of view of university librarians by McCormick (2014) and Schlosser (2016).
Where researchers do read publisher agreements, attitudes differ. One recent study suggests that, whilst most researchers agree publishers' copyright should be respected, opinion is divided over whether publisher terms are fair or unfair, with over 40% feeling copyright policies are restrictive (Rapple, 2017).
Other research suggests that up to 49% of researchers assign copyright reluctantly because they feel they have little choice (Gadd, Oppenheim, & Probets, 2003). A recent analysis discusses the tension between 'scholarly' and 'academic' cultures and the resulting ambivalence in academics' attitudes to ceding copyright to publishers (Gadd, 2017). Whilst it is not the purpose of the new edition of Clark's to discuss the rights and wrongs of any particular type of agreement, the book does include the Publishers Association Code of Practice on Author Contracts, which seeks to set out 'an equitable framework for author-publisher contracts' (Appendix J).
Many researchers emphasize the importance of having the freedom to circulate work amongst colleagues as part of the process of scholarly debate, including uploading work to SCNs (Dylla, 2016;Van Noorden, 2014 researchgate.net/). Studies suggest as many as 50-60% of academics may be using ResearchGate to load their own publications, access work by colleagues, or facilitate collaboration (Borrego, 2017;Harseim & Goodey, 2017;Lovett et al., 2017;Van Noorden, 2014). On occasion, this desire of researchers to share work with colleagues can lead authors to contravene the terms of their agreements with publishers, which limit the sharing of their work. One study by Jamali (2017) found that 50% of articles loaded on a leading SCN, ResearchGate, were in contravention of publisher agreements, a practice that has been called 'black' OA (Bjork, 2017). This research concluded that a key reason was confusion over the fact that, as has been noted above, many publishers do allow archiving (and thus free access to) of pre-print and post-print versions of journal articles, even if they try to restrict circulation of the published version of record to protect revenue streams such as journal subscriptions or book sales.
The potential seriousness of this situation can be seen in the recent decision by a number of leading publishers, including Elsevier and Wiley, to take action requiring ResearchGate to remove papers that infringe the terms of agreements with publishers (Matthews, 2017 However, whilst many researchers want greater freedom to share their research with colleagues, there are divergent views on how that research can be used by others. Whilst some academics are relaxed about how their work is used in an OA environment, putting emphasis on the widest possible circulation of their work, others are more concerned about protecting the integrity of their work by restricting adaptation or limiting use for commercial purposes. Some authors also wish to retain a degree of control over the data they have invested significant time and resources in acquiring and which may have value in terms of future publications or commercial applications. Some studies suggest that many academics across both the sciences and humanities are opposed to commercial re-use, adaptations, or inclusion of their work in anthologies (a particular aspect of humanities publishing), whilst there are mixed views about allowing data mining of their work (Rowley, Johnson, Sbaffi, Frass, & Devine, 2017). An example of these contrasting views is the bioRxiv site, which hosts pre-print papers in biology. A study of the site by McKenzie (2017) found that over a third of authors had selected the most restrictive Creative Commons (CC-BY-NC-ND) license, which bars commercial use and 'derivative' works, including translations and annotations. Another 29% had not selected any licence that, by default, reserved all rights in the work, requiring permission from the author for copying and re-use. These reservations about the benefits of OA can also be seen in attitudes to the Scholarly Communications Licence, which has been criticized by academics as being too inflexible (Wulf, 2017). This criticism highlights problems such as how such an open license affects terms governing quoting material from other copyright holders as well as the financial viability of smaller professional societies that still publish a number of journals, particularly in the humanities.

MANAGING THE COPYRIGHT LANDSCAPE
The academic publishing sector is thus facing a situation where there is both significant disagreement and confusion over the degree of copyright control that should be exercised by authors or publishers over research publications, as well as their use by others. To add to this confusion, there is also growing debate about the roles of differing versions of a work. Concerns about the speed and reliability of peer review, for example, have led to moves away from reliance on the published version of record as the gold standard to circulation of draft versions for new types of open, post-publication publication review (Nature, 2017). In addition, the development of annotation tools has made it easier for readers to add review and other comments to an article, effectively creating new, amended versions of a work (Anderson, 2017). As one senior publishing professional has commented, '… the editorial and publishing workflow has become more transparent, more measurable and more fragmented. There are increasing numbers of workflow solutions and tools for discovery, assessment, analysis, publishing and outreach' (Thomas, 2017). The potential result is a proliferation of different versions of works covered by an increasingly diverse set of agreements and terms. This can only add to the existing confusion amongst researchers about their rights and responsibilities in the publication and dissemination of research.
Attempts are being made to clarify this situation, for example, by publishers simplifying and standardizing policies on how content can be shared as well as helping users to navigate the process of sharing research legally. An example is the STM's new 'How can I share it' website (STM, 2017b). The 'How can I share it' site uses a tool based on an article's unique digital object identifier (DOI) to allow users to check its contractual status as well as provide a repository of publisher policies on sharing. Another key development is the move to tag material more consistently and clearly to show its status and the terms governing its use (McKay, 2017).
These developments include initiatives by publishers to make better use of article tagging standards such as the Journal Article Tagging (JAT) Document Type Definition (DTD) Suite, developed by the US National Information Standards Organization (NISO), which has been updated to accommodate OA agreements (Donohoe, Sherman, & Mistry, 2015;Lizzi, 2017;NISO, 2015). These developments will make it easier for researchers to understand their contractual rights and responsibilities in sharing published work. Other initiatives include CrossMark from CrossRef, which adds an icon to a published piece of work to confirm its status, including whether any updates or corrections have been added (www. crossref.org/services/crossmark/).
Whilst the copyright landscape has become more diverse, it is also important to note the continued strength of more traditional types of agreement governing the use of researchers' work. In some respects, the divisions within the research community about the merits of OA may be to the advantage of academic publishers who have adapted quickly to the new OA environment, migrating income from subscriptions to APCs paid to allow immediate access under the 'gold' OA model (Dodds, 2017). The fact that researchers are divided in their views about OA, and the merits of allowing readers unrestricted access to and re-use of their work, has contributed to limiting the impact of OA publication, which still remains a small percentage of the overall STM market. As an example, the Open Access Scholarly Publishers Association (OASPA) recorded just under 190,000 OA articles published by their members in 2016 (OASPA, 2017). This is comparable to estimates of between 1.8 and 2.5 million articles published annually (Bornmann & Mutz, 2014;Plume & van Weijen, 2014). Analysis by the STM of the various kinds of OA output suggests 12% of all articles are 'gold' OA, whilst a further 5% are 'green' OA, and 10-12% are archived by authors in repositories in compliance with employer and/or funder policies (Ware & Mabe, 2015). Other figures suggest that as much as 17% of articles are 'gold' OA and 6% 'green' OA (Tickell, 2017). This still leaves the great majority of articles published through conventional subscription-based channels and presumably, in many cases, with standard assignment agreements. As the STM figures suggest, authors' use of institutional repositories to archive their publications also remains low (Borrego, 2017;Lovett et al., 2017;Poynder, 2016).

CONCLUSION
Given some remaining confusion over copyright, ambivalent attitudes to OA, and a continued (if sometimes reluctant) commitment to the status quo, it looks as if academics are still likely to continue to agree to standard assignment contracts in many cases. This has been reinforced by the flexibility of publishers in loosening control over copyright sufficiently to allow researchers greater freedom to develop and share new research.
Whether because of greater awareness of copyright issues, funder requirements, or institutional policies, a growing minority of researchers will opt for licenses from publishers, giving them greater control over their work. This will include a broad range of types of OA agreement where, as responses to the Scholarly Communications Licence suggest, academics prefer flexibility in deciding how far they want their work to be shared.
This increasingly diverse and complex copyright environment makes the new edition of Clark's all the more important.
This latest edition captures the key developments in publishing contracts and maintains the reputation of Clark's Publishing Agreements as the standard work on drafting robust and comprehensive publishing contracts. It should be widely used by publishers and all those with an interest in publishers' agreements and copyright issues.