Last week, Macmillan's Nature Publishing Group, a major for-profit publisher of scientific journals, cracked open its paywall a little. Or more precisely, Macmillan created a way for people to tunnel under the paywall: Journal subscribers, along with some blogs and media organizations, will now be able to share links to free, read-only copies of paywalled scientific papers. On the surface, this sounds like a good thing—as the publisher put it in its announcement, the change "will make it easy for readers to share an unprecedented wealth of scientific knowledge instantly with researchers and scientists across the globe." The policy change is also supposed to benefit the public, by letting bloggers and journalists offer their readers direct links to technical articles discussed in their stories.
While the policy seems like an improvement for everyone, many scientists aren't impressed. They view the policy as a cynical move by the publisher to avoid making more substantial changes that would benefit the scientific community. Whether you believe that characterization or not, the discussion around Macmillan’s policy change is a symptom of the increasingly dysfunctional relationship between scientists and their journals.
To understand what's going wrong, we need to consider why scientific journals exist in the first place. Ever since they first appeared in the 17th century, scientific journals have filled three critical functions: First, they are the primary means of distributing new scientific knowledge. Second, they carry out quality control of that knowledge by organizing peer-review. And third, they help determine how career rewards are doled out among scientists, by documenting a scientist’s research successes. As they carry out these functions, scientific journals are part of an important set of institutions and community norms that make science work by creating incentives for people to produce shared knowledge. As the economist Partha Dasgupta put it, one of the reasons the Scientific Revolution was so revolutionary was because "it created institutions that enabled the production, dissemination, and use of knowledge ... to be transferred from small elites to the public at large." In other words, scientific journals are supposed to help solve the problem of ensuring that new knowledge is shared.
Sharing papers has been central to science for centuries now. Scientists eagerly pass their own manuscripts as well as the published work of others to each other. If a colleague asks you for a copy of a paper she doesn't have access to, you don't refuse. That is why many scientists were upset by the language in Macmillan's announcement, which kept referring to the new policy as a "legitimate" way to share papers. In the release, Timo Hannay, directing manager of the Macmillan division that is developing the new read-only platform for viewing paywalled papers, said: "We know researchers are already sharing content, often in hidden corners of the Internet or using clumsy, time-consuming practices." Nature's new system will be "a convenient, legitimate alternative that allows researchers to access the information they need ... from the definitive, original source." The clear implication is that much of the sharing scientists do now is illegitimate.
Researchers see this as part of an ongoing pattern by journal publishers to prevent them from sharing their publicly funded research. Last year, some researchers who had posted their papers on the academic social media site Academia.edu received requests to take them down from Elsevier, a major for-profit publisher. Scientists who blog have had publishers threaten them with legal action for including published charts or data in their blog posts. Most importantly, major publishers and their trade associations have pushed back hard against government policies, in the United States and elsewhere, that require published results of government-funded research be made freely available to anyone within a certain time period.
To be fair to the publishers, carrying out the three critical functions of a scientific journal costs money. Journals need some source of revenue, which has typically come from institutional library subscriptions. Before the Internet, sharing physical copies of published papers could be a laborious process, sometimes requiring a photocopier and a trek into the depths of the university library. Sharing papers didn't constitute much of a threat to a journal's revenue. But with email and the Web, sharing is obviously possible on a much larger scale. As university libraries cut back on subscriptions due to their staggering costs, there is a strong incentive for scientists to share. Some publishers are now worried that widespread sharing will kill their business.
And they should be worried about their business, because there is an ongoing debate over how well journals are carrying out any of their three traditional functions. In addition to concerns that some journals are preventing scientists from sharing, researchers are also questioning how well journals are carrying out their quality control and career rewards functions. The biomedical sciences are increasingly plagued by published research that is not reproducible. Many scientists are arguing that the community needs to put more emphasis on post-publication peer review—which, it's important to emphasize, can only work well when papers are freely available. Researchers are also concerned that some journals aggressively promote their brands and reputations in a way that distorts their roles as arbiters of career rewards. The result is that cachet in science is often outsourced to professional journal editors, who are not currently part of the community of active research scientists, and whose decisions often reflect the best interests of the journal, but not necessarily those of the scientific community.
Scientists aren't ready to give up on journals just yet. But the growing size of the scientific community, together with rapidly changing new technologies for communicating knowledge, pose a huge challenge to scientists and their journals. To fix the strained relationship between them, publishers need to focus on the core functions of their journals and figure out how to best execute those functions today.