The debate over whether research results should be freely accessible has always been fraught. Having given a lot of ground, journal publishers are determined to hang on to one last bastion: their rights to the published version of scientific articles. Now librarians and open-access advocates have set their sights on that final prize — by encouraging researchers to demand the right to distribute the published versions freely and immediately.

Funding agencies are increasingly adopting policies to make the results of the research they fund free for all. Both the US National Institutes of Health (NIH) and Britain's Wellcome Trust, for example, encourage this practice. They ask that the version of a manuscript accepted for publication be put in an open-access library such as PubMed Central within 6–12 months of it coming out. The Wellcome Trust's policy will become compulsory on 1 October, and legislation that would make the NIH policy mandatory is pending in the US Congress (see Nature 441, 915; 2006).

Ann Wolpert, director of libraries at the Massachusetts Institute of Technology (MIT) in Cambridge, Massachusetts, has launched an initiative that she says will clearly assign rights to the author in a way that would satisfy funders. Wolpert has drawn up a document that researchers can add on to the rights agreement the publisher gives them to sign. Similar agreements have been drafted by the Scholarly Publishing and Academic Resources Coalition and the MIT-affiliated Science Commons.

“I look at it as responding to a request by faculty members to simplify their lives,” says Wolpert. “They say 'it is crazy that we are supposed to read and understand these publishers' agreements. Give me something that I can just staple to any agreement, so I can comply with NIH or Wellcome Trust policy'.”

This isn't a balance of rights. This is giving MIT all the relevant rights.

But publishers' groups argue that the agreements being drafted go much further than is necessary to comply with current policies. Wolpert's document, for example, would allow authors to publish the final, formatted version of their paper anywhere on the Internet, as many times as they like, immediately after publication. “This isn't a balance of rights. This is giving MIT all the relevant rights,” says David Hoole, head of brand marketing at Nature Publishing Group.

Publishers point out that most journals already allow authors to post the accepted version of their paper online, as required by the NIH and the Wellcome Trust. Such versions have been peer reviewed, but aren't copy-edited, formatted or paginated. But giving authors rights to the final versions, they say, could make it impossible for journals to earn a living.

Jerry Cowhig, managing director of the publishing arm of the Institute of Physics, says that the institute provides articles free online for 30 days after publication, and that he is happy for authors to post the accepted versions of their papers. But he is not in favour of making the final, edited version of a paper freely available everywhere. “That would be a real threat to the continuation of established journals, and the eventual outcome would be to damage scholarship,” he says.

Sally Morris, chief executive of the Association of Learned and Professional Society Publishers, agrees. “The final version is where publishers add value,” she says.

On 27 June, Morris's group, along with the International Association of Scientific, Technical and Medical Publishers, wrote to Wolpert outlining their concerns and proposing a meeting. A similar letter went out from the professional and scholarly publishing division of the Association of American Publishers on 7 July.

But would giving authors such rights really damage journals? After all, many authors already post the final version of their paper on the web regardless of what their rights agreement says. A study by Jonathan Wren, a bio-informatician at the University of Oklahoma in Norman (J. D. Wren Br. Med. J. 330, 1128–1131; 2005), revealed that the final versions of more than one-third of articles in high-impact journals are freely available online (see Fig 1).

Figure 1: More than a third of authors already post final versions of their science papers online, a BMJ study found.
figure 1

SOURCE: J. D. WREN BR. MED. J. 330, 1128–1131 (2005)

SOURCE: J. D. WREN BR. MED. J. 330, 1128–1131 (2005)

Full size image

Ted Bergstrom, an economist at the University of California, Santa Barbara, argues that for libraries and other users, the convenience and authority of journal subscriptions will still be preferable to searching out free versions of papers individually. So agreements such as Wolpert's shouldn't affect the bottom line of any but the most overpriced publications, he says.

Wolpert adds that the value of journals isn't just in locating and reading individual papers, but in browsing and the 'serendipity factor'. “There is value behind a collection of articles judged worth your attention by smart people in your discipline,” she says.

John Cox, a consultant to publishers and academic societies who is based in Chichester, UK, says that the value of papers as they appear on journal websites is often underestimated. “It's not just the copy-editing, but the infrastructure that is provided: the linking to citations, indexing, alerting services, the presentation of the product on the screen to the reader.”

But he argues that the desire to post final versions across the web is misguided because the version published on the journal website will always be definitive. “It becomes, if you like, part of the minutes of science,” he says. “That is deeply embedded in the scientific research culture.”