Björn Brembs and nine co-authors, in a sweeping paper posted to Zenodo last fall:
There is now a very real threat of a single (or few) corporations effectively owning all scientific data, both research data and user data, on top of their share of the scholarly literature. […] The break will not be technological, as all the technology for the disruption already exists, but with regard to governance. In general, there is broad agreement on the goal for a modern scholarly digital infrastructure: it needs to replace traditional journals with a decentralized, resilient, evolvable network that is interconnected by open standards, that allow seamlessly moving from one provider to another, under the governance of the scholarly community, de-centering the journal article as the sole scientific output that “counts”.
It’s an important paper, with a sharp diagnosis of the oligopolist stranglehold—a “public good in private hands,” as the authors put it.
There’s lots to unpack in the Brembsian alternative proposed here. One cornerstone is the adoption of open standards that—as best I understand it—would enable university repositories and nonprofit, community-led platforms like Open Library of Humanities (OLH) to form a kind of global, interoperable library. A second cornerstone is a regulated market for services. In an open procurement process, publishers and other firms—nonprofit or otherwise—would submit bids for peer review services, for example, or for copy editing or even writing software. The idea is that a regulated marketplace will, through competition enabled by open standards, discipline the overall system’s cost.
It’s a fascinating proposal, one that—as the paper notes—could be implemented with existing technologies. The problem is the lever of change. The incumbent publishers’ entrenched position, Brembs et al explain, renders a first move by libraries or scholars impractical. That leaves funders, whose updated rules and review criteria could, the paper argues, tip the incentive structure in the direction of an open, journal-free alternative.
There’s a lot to chew on in what is an exciting, radical blueprint. One facet of the current system that the paper underplays, however, is the journal system’s prestige lock-in. It’s not just outdated procurement rules that stand in the way of the journal’s demise, but also entrenched features of the scholarly reward system. That system, to some large extent, exists in scholars’ heads, as reinforced and reproduced by the tenure and promotion process. Maybe funder rules can erode those norms, but work to reform the reward system should happen in tandem with funder-driven incentives—along the lines that Juan Alperin and colleagues are suggesting.
A second friendly annotation to the Brembs paper has to do with its implicit orientation to nomothetic disciplines. The paper designates the replication crisis as one of three driving problems in the current system. Likewise, in the paper’s imagined future, there’s a discussion of hypotheses generated by machine learning models, which, the authors write, “will help mitigate researchers’ all too human tendencies to cut corners, tell stories or cheat.” Hmmm. The contrast between machine and human is too tightly drawn, and the premise of AI value-freedom is hard to defend. More to the point, replication and hypothesis testing make no sense—that is, they do not apply—to idiographic fields, many of which reside in the humanities. There’s a hidden parochialism in the paper’s assumptions about what academic knowledge is.
Quibbles aside, the paper is a thrilling template for an alternative future.
A final thought: I hope that Brembs et al are in contact with the Confederation of Open Access Repositories (COAR), which just received a $4 million grant from Arcadia for its ongoing project to build a “standard, interoperable, and decentralised approach” to a global network of networks.