This article originally appeared in Against the Grain, vol. 16, no. 3 (June 2004). The published version is not open access and I haven’t seen it. This is the version of the text that I submitted to the editor. – Peter Suber
A Primer on Open Access to Science and Scholarship
The scientific journal was invented in 1665. For readers, the new form of publication surpassed books for learning quickly about the recent work of others. For authors, it surpassed books for sharing new work quickly with the wider world and, above all, for establishing priority over other scientists working on the same problem. Because authors were rewarded in these strong, intangible ways, they accepted the fact that journals couldn’t afford to pay them. Over time, journal revenue grew but authors continued in the tradition of writing articles for impact, not for money. Scholars may write books and software for money, but for journal articles they are amply paid by advancing knowledge and advancing their careers.
The tradition that started in 1665 continues today and makes the scientific or scholarly journal article nearly unique in the landscape of intellectual property. It’s not rare for authors to earn nothing from their work. But it is rare and peculiar for authors to earn nothing regardless of the sales of their work. It’s rare and peculiar for authors to submit their work voluntarily, even eagerly, to publishers who will not buy it or pay royalties. One measure of its rarity is that our copyright law is designed for literature that will produce royalties; it provides a temporary monopoly on distribution, and limits fair use, primarily in order to protect royalties. Even content that users get for free, such as TV and radio shows supported by advertising, generates royalties for its creators.
Scientific and scholarly journal articles are among the most significant examples of this rare and peculiar breed, which could be called royalty-free literature. The only other examples of comparable in significance are the statutes and judicial opinions of public law. But in most countries, the texts of public law are uncopyrightable. This makes the research journal literature even more rare and peculiar: it is royalty-free and still copyrightable. Researchers relinquish revenue for their journal articles but do not necessarily relinquish intellectual property rights.
Open access may come one day to royalty-producing bodies of literature, like books, and other categories of royalty-producing content, like music and movies. But the open-access movement is focused on royalty-free literature. To understand why, note two important consequences that follow from the fact that this literature generates no royalties. First, authors of journal articles can consent to open access without losing revenue. Second, journal publishers who charge subscription fees and then limit access to paying customers are violating the interests of these authors, not advancing them. Authors of scientific and scholarly journal articles want to disseminate their work as widely as possible, so that it will be noticed, read, taken up, built upon, applied, used, and cited. Publishers who stand between authors and readers, and charge access fees, negate the authors’ sacrifice, shrink the authors’ audience, and reduce the authors’ impact.
In the age of print, journals had significant expenses that could only be recovered through subscription fees. Price was a barrier for readers seeking access and for authors seeking readers, but the economics of print left no alternative. Moreover, until the 1970s or so, the price barrier was fairly low. But since the 1970s, journal prices have risen faster than inflation, and since 1986 they have risen four times faster. Libraries now speak of a pricing crisis and cope with exorbitant price increases by canceling subscriptions and cutting into their book budgets. Today the price barrier is so high that even the wealthiest research institutions in the world are canceling journals by the hundreds and issuing public statements condemning the price increases, licensing terms, bundling policies, and negotiating tactics of the major publishers.
The explosive growth of published knowledge will only intensify the problem. Adhering to business models that charge for access (subscription fees, licensing fees, or pay-per-view) will guarantee that over time researchers will have access to a rapidly shrinking percentage of the published knowledge in their field.
Against this background, the internet emerged in the 1990’s as a kind of miracle. For the first time, it became physically and economically possible to distribute perfect copies to a worldwide audience at virtually no cost. For the first time, it became physically and economically possible to connect authors, who want to give away their work, with readers who want to read and build on it. This new form of distribution —digital, online, free of charge, and free of most licensing restrictions— is now called open access.
Only the creators of royalty-producing content had any reason to hold back from taking full advantage of the new medium. The creators of royalty-free content could finally share their work in a way that matched their interests.
The internet is not just a faster, cheaper, searchable alternative to print. By widening distribution and reducing expenses at the same time, it frees us to adopt new business models. We are not limited to the business models that evolved for print. We can seriously explore models that dispense with subscriptions and other price barriers, and use new and better ways to recover the (newly reduced) expenses of a peer-reviewed journal. We can dispense not only with subscriptions, which are costs for users, but for many costs at the journal end as well: the costs of soliciting and renewing subscribers, the costs of maintaining their addresses or authentication data, the costs of blocking online access to non-subscribers, and the costs of drafting and enforcing licensing agreements. We can seize rather than fear the opportunities created by new medium for making copies and sharing them with others. We can serve the research interests of researchers first, rather than subordinate them to the economic interests of intermediaries. We can no longer say, as we could say in the age of print, that the access barriers attached to traditional journals are an unavoidable consequence of the best available method of distribution.
Open access is compatible with copyright. Authors hold copyright on their articles until and unless they transfer copyright to someone else. If authors consent to open access while they still hold copyright, then open access is fully authorized and lawful. The fact that most musicians and movie-makers do not consent to open access should not make us pessimistic about open access to science and scholarship. Musicians and movie-makers create royalty-producing content and have understandable concerns that open access will diminish their revenue (even if there is growing evidence that it could actually enhance it). Authors of royalty-free literature do no share this concern, and have everything to gain and nothing to lose by consenting to open access.
Open access is compatible with print. Users who prefer to read printed text can print any open-access file they like. Libraries and publishers can use print for long-term preservation. Both BioMed Central and the Public Library of Science, the two leading open-access journal publishers, offer print editions at cost for those who want them. As long as journals offer an open-access edition, then priced, printed, or enhanced editions do not interfere in any way. As long as we make print copies and deposit them in libraries, then open-access literature has at least the longevity of print literature.
Open access is compatible with peer review. In fact, all the major open-access projects and campaigns —the Budapest Open Access Initiative, BioMed Central, the Public Library of Science, SPARC, the Bethesda group, the Berlin Declaration— insist on the importance of peer review. Open access to science and scholarship is not about using the internet to bypass peer review. It’s about removing the barrier of price, not the filter of quality control.
Peer review is a mix of professional and clerical tasks. The professional task is editorial judgment exercised by subject-matter experts. In most journals in most fields, the editors and referees who exercise editorial judgment donate their labor, just like the authors. The clerical tasks are increasingly being taken over by software -assigning the files to referees, distributing files, monitoring progress, nagging dawdlers, collecting comments, sharing comments with the right people, tracking revisions, and collecting data. Because the professional labor in this process is largely donated, the cost is already low. Because the clerical labor is being automated, even by open-source software, the overall cost of peer review continues to drop.
But even low expenses must be recovered if open access is to be sustainable. Open access repositories, which do not perform peer review, have negligible expenses, are built on open-source software, and are supported by the institutions that benefit from increasing the visibility and impact of their faculty.
Open access journals, which do perform peer review, are usually supported by processing fees on accepted articles paid by the author’s sponsor rather than access fees paid by the reader’s sponsor. These processing fees are closely related to the costs of peer review, manuscript preparation, and hosting. When they cover all of a journal’s expenses in producing and distributing an article, then the journal is on sound financial footing to provide free online access to the article for all connected users worldwide. This model is similar to the economic model of television or radio, in which some viewers pay for everyone, or advertisers pay the costs of production and distribution upfront so that the public needn’t pay anything for access.
One reason to think that this model is economically sustainable is that it works in an industry, broadcasting, with much higher costs and none of the advantages arising from the use of royalty-free intellectual property. A more specific and persuasive reason is that the costs of vetting and disseminating articles online are much lower that the prices currently charged by publishers, and paid by libraries, for access to them.
Some traditional publishers charge that this business model endangers the integrity of peer review. The objection is that charging processing fees on accepted articles gives a journal an incentive to accept more articles regardless of their quality. This objection seems strong at first, but its strength disappears as we look more closely. In fact, the objection takes a turn and starts to raise suspicions about the business model of conventional of subscription-based journals. (1) First we must note that accepting more excellent submissions does not compromise peer review. A journal would only compromise peer review if it lowered its standard and accepted more weak submissions. (2) Conventional journals that justify price increases by pointing the growing number of articles they publish have exactly the same incentive that the objection attributes to open-access journals. (3) But conventional journals must fill their allotted space and give subscribers their money’s worth. Hence, when they face a temporary dearth of excellent submissions, their business model gives them an incentive to lower their standards in order to fill the issue. This incentive does not exist for open-access journals, which have no allotted space to fill and no subscribers who paid for a certain volume of content. If open-access journals face a large number of excellent submissions, they can publish a large issue. If they face a small number of excellent submissions, they can publish a small issue.
Moreover, (4) the processing fees charged by open-access journals barely cover their costs. So if these fees do create an incentive to lower standards, it is a weak force. Essentially all the financial gain of increasing the acceptance rate would be offset by the costs of publishing the additional articles. But subscription-based journals that tie their prices to their volume, raise prices faster than inflation, and realize high profit margins, have much more to gain by lowering their standard, accepting more papers, and increasing their subscription price.
Finally, we may add that conventional journals with low submission and rejection rates have higher profit margins than conventional journals with high submission and rejection rates. This gives their publishers and incentive to lower standards until the margins peak. It also gives them an incentive to bundle the lower-quality journals with higher-quality journals so that it is costly for libraries to cancel them. These incentives do not arise for open-access journals.
In any case, the upfront funding model is not the only business model for open-access journals. It works best in fields like biomedicine where most research is funded and where the major funders are already on the record as willing to pay the upfront fees. But in less prosperous fields, including the humanities, one attractive model is for university libraries to publish open-access journals. The Philosophers’ Imprint, for example, is a peer-reviewed, open-access journal published by the University of Michigan. Its motto is, “Edited by philosophers, published by librarians.” Because the philosophers and librarians are already on the university payroll, the journal needn’t charge processing fees. Documenta Mathematica is another peer-reviewed, open-access journal that charges no processing fees. It covers its expenses through the sale of printed volumes. What’s important is that there is not just one way to cover the expenses of a peer-reviewed, open-access journal, and we have a long way to go before we can say that we’ve exhausted our cleverness and imagination.
There are many reasons why open access is moving more slowly in the humanities than the STM fields (science, technology, and medicine). One is the relatively low level of research funding. This reduces the money available to subsidize open-access dissemination of research results. Because there is much less government funding of the humanities than the STM fields (to match the lower level of private funding), humanists get less traction than scientists with the “taxpayer argument” for open access, or the argument that taxpayers shouldn’t have to pay a second fee for access to publicly funded research.
A major difference is that journal prices in the STM fields are much higher than in the humanities, drawing urgent attention to open access as an obvious method of relief and giving institutions a motive to join scholars in promoting it. Open-access preprint exchanges thrive in the STM fields, where researchers need to put a time-stamp on their results in order to establish their priority. By contrast, open-access preprint exchanges are rare in the humanities. Conversely, demand for journal articles in the humanities persists longer than in the STM fields, which means that humanities journals have a greater fear than STM journals that offering open access even after some delay (say, six months after publication) will undercut subscriptions and reduce revenue. Similarly, the average rejection rate is higher at humanities journals than at STM journals, increasing the cost of peer review per accepted article, and making open access harder to subsidize through processing fees on accepted articles. Humanities journals often want to reprint poems or illustrations that require permission from a copyright holder, and this permission rarely extends to open-access publication.
Behind these specific reasons is a more general one. Journal articles are the primary literature in the sciences, but in the humanities they tend to report on the history and interpretation of the primary literature, which lies in books. But as we’ve seen, journal articles are royalty-free while books are royalty-producing. The logic of open access applies much better to journal-based fields than to book-based fields.
All these are reasons why we should not expect to make progress toward open access in all disciplines at the same rate. They are also reasons to expect that different business models will evolve in different fields to cover the costs of open-access journals. (By contrast, the costs of open-access repositories are already low and essentially constant across disciplines.)
Open access is within reach of scientists and scholars today. They can launch an open-access repository whenever they like, at essentially no cost, and more and more universities and disciplines are doing so. With a bit more planning and investment, scholars can launch an open-access journal. Scholars themselves decide whether to submit their work to open-access journals, whether to deposit it in open-access repositories, and whether to transfer copyright.
Conventional journals can experiment with open access to their back runs, to back issues after a certain embargo period, to all new articles, or to selected new articles, in order to learn the methods and economics of open-access publishing. But scholars needn’t wait for conventional journals to make these experiments, and needn’t persuade them to accept open access as a superior, or even desirable, alternative. The internet has already given scholars a chance to reclaim control of scholarly communication. For the first time since the journal appeared on the scene in 1665, price needn’t be an access barrier to this critical body of royalty-free literature. For the first time since the rise of the commercial publishing of scholarly journals, scholarly communication can be in the hands of scholars, who answer to one another, rather than corporations, who answer to shareholders. The only question is when scholars will fully seize this beautiful opportunity.
Portions of this article appeared in the InfoPaper for the World Summit on the Information Society for November 11, 2003, and the SPARC Open Access Newsletter for November 2, 2003, and February 2, 2004.
1. Claude Guédon, “In Oldenburg’s Long Shadow: Librarians, Research Scientists, Publishers, and the Control of Scientific Publishing,” ARL Proceedings, May 2001.
2. For more on royalty-free literature, its peculiarities, and its connection to open access, see my “Creating an Intellectual Commons through Open Access,” a preprint based on a presentation at the Workshop on Scholarly Communication as a Commons, Workshop in Political Theory and Policy Analysis, Indiana University, Bloomington, Indiana, April 1, 2004.
3. See the data of the Association of Research Libraries, “Monograph and Serial Costs in ARL Libraries, 1986-2002.”http://www.arl.org/stats/arlstat/graphs/2002/2002t2.html
4. For example, Harvard, Stanford, MIT, Cornell, and Duke are in this category. See my list of “University actions against high journal prices,” first published in the SPARC Open Access Newsletter, April 2, 2004, and then put on the web for future updating. The list includes substantial excerpts from the public statements that accompanied the cancellation decisions.
Updated web version
5. See my “The scaling argument,” SPARC Open Access Newsletter, March 2, 2004.
6. See the Budapest Open Access Initiative (February 2002), which made “open access” the term of art for this kind of free online availability.
7. See the SPARC list of journal management software (both open and closed source).
8. A 2002 survey of the literature put the cost of peer review at about $400 per accepted article. See Fytton Rowland, “The peer-review process,” Learned Publishing, 15, 4 (October 2002) pp. 247-58, <http://makeashorterlink.com/?E2CA26BF7>. This is the amount that an average journal spends on peer review per accepted article, and therefore includes the cost of reviewing the average number of rejected articles per accepted article. Rowland’s survey was published one month before the launch of Open Journal Systems, <http://www.pkp.ubc.ca/ojs/>, the first open-source journal-management software. I’m not aware of more recent surveys that take new and open-source software into account.
9. See Raym Crow, “A Guide to Institutional Repository Software v 2.0,” Open Society Institute, January 2004. Crow’s guide is limited to open-source software.
10. For a more detailed exposition of this argument, see my “Objection-reply: Whether the upfront payment model corrupts peer review at open-access journals,” SPARC Open Access Newsletter, March 2, 2004.
11. The largest private funder of medical research in the United States, the Howard Hughes Medical Institute, and the largest in Britain, the Wellcome Trust, have adopted this policy. In June 2003, they and other stakeholders issued the Bethesda Statement on Open Access Publishing, calling on others to follow suit.
Also see BioMed Central’s list of funding agencies willing allow grantees to use grant funds to cover article processing charges.
12. Philosophers’ Imprint,
Another example is the Journal of Insect Science, , published by the library of the University of Arizona at Tucson. For details on the arrangement, see Henry Hagedorn et al., “Publishing by the American Library,” <http://www.arl.org/sparc/meetings/Henry_Hagedorn.htm>, a conference presentation from January 2004, and Eulalia Roel, “Electronic journal publication: A new library contribution to scholarly communication,” College & Research Libraries News, January 2004,
13. Documenta Mathematica,
Also see Ulf Rehmann, “Documenta Mathematica: A Community-Driven Scientific Journal,” High Energy Physics Libraries Webzine, October 2003.
14. See my “The taxpayer argument for open access,” SPARC Open Access Newsletter, September 4, 2003.
15. For a longer discussion, see my “Promoting Open Access in the Humanities,” a preprint based on a talk to the American Philological Association, January 3, 2004.
16. For more, see my list, “Disciplinary differences relevant to open access.”
17. See my list, “What you can do to help the cause of open access.”