Andrew Elfenbein, University of Minnesota
What exactly does “open” mean?
“Open science” is a tricky phrase because of the adjective, “open.” It implies the existence of an evil Doppelgänger, “closed science.” “Open” has happy connotations of being accessible, welcoming, available, and comprehensible, while “closed” is the grim-faced opposite: inaccessible, forbidding, unavailable, and opaque. I instinctively side with the underdog “closed science” when faced with such oppositions because I suspect that those on the “closed science” side of the debate would prefer a different term. “Open science” has a slightly bullying tone, as if opposing it, or even questioning it, automatically makes you an elitist creep.
In the history of European science, the stark opposition of “open” and “closed” has rarely captured the full status of science. At times, the same things that have made science open have also made it closed. For example, for hundreds of years, scientific publication in much of Europe was in Latin. Since Latin was a dead language during those centuries, scientific Latinity looks like a concerted, successful attempt to keep science the exclusive province of an elite. But that is not the whole story. Latin as a lingua franca allowed works in Latin to have a wider international reach than they ever would have if they had been written in a vernacular: they were immediately available, without having to go through the barrier of translation. Science was the province of an elite, but one not confined by national boundaries in Europe; the same linguistic fact that closed science to most guaranteed that it was open to all with comparable educational backgrounds, regardless of nation or religion.
What happens when science becomes professional?
In England, the late eighteenth and early nineteenth centuries were notable for the commitment to making science available. Humphry Davy’s public lectures at the Royal Society drew the cream of London society to hear talks on such topics as galvanism and the isolation of boron as an element. The scientific publications of the day worked against specialization: there was something for everybody. For example, the first five articles in the Royal Society’s 1818 Journal of Science and the Arts are “Of the Dissemination of Plants,” “On the Influence of Corporeal Impressions in Producing Change of Function in the Living Body” “On the Theory of Spherical Atoms,” “Suggestion of a New Principle for the Register Thermometer,” and “A Short Account of Horizontal Water Wheels.” Findings in botany, anatomy, chemistry, and engineering all live comfortably together in these pages.
As with any gesture of openness, this one had limits: such science was open only to those who could read and could afford either to buy periodicals or to join a library that might carry them. Yet by the end of the century, such openness had constricted. American and European science became the property of university research and corporations, and sharp divisions between disciplines emerged. With that specialization came a professionalization of language that assumed that specialists wrote for other specialists. Even as ideals of writing cultivated by universities insisted on clarity and brevity (as opposed to older early modern ideals of copia and amplificatio), those ideals ran against the cornerstone of specialization: discipline-specific vocabulary. In the case of science, the discourse of statistics further estranged that vocabulary from the everyday, so that science required multiple knowledges and specialties to be comprehensible. The result, which remains today, is a modern version of Latin: a language close to all except a few, but a language available cross-nationally to specialists. Such a professional idiolect both gives science prestige as a province of experts and makes it vulnerable to anti-intellectual bias that dismisses expertise as the province of academics out of touch with reality.
Even the most open of open sciences will remain closed to most because of the barriers I have described. Although those barriers are hardly a reason to dismiss efforts at open science, most of which I support, they do reveal that the adjective “open” might point to a fantasy rather different from the reality. Most practically, even if all hypotheses are preregistered, if all data are made public, and all publications made accessible on the web, the audience for such material is still minuscule relative to the population: it remains fundamentally closed, for all the efforts at openness.
This was fun to read because it focuses on an entirely different aspect of “Open Science” than what, I at least, usually think of (e.g., pre-registration, data and publications freely available. etc.). The notion of shared language, and especially a shared, specialized language is really important to bring out because it highlights the idea that– while science may be open, not every person who accesses any given scientific piece is qualified to adequately consume it. Case in point– my biggest discomfort associated with gestures of Open Science is the practice of posting and promoting non-peer-reviewed research. I have personally seen very prestigious authors post pre-prints that later did not pass peer-review because of fundamental flaws that could only have been identified by someone intimately familiar with the associated literature and previous findings. Yet, the pre-print can be accessed and consumed by individuals who will take it as Truth sheerly on the basis of the author’s reputation and seniority. As a community, we should really be diligent to ensure that Openness doesn’t end up completely negating the peer review process. The end result of that can only be a diluted science that advances at a halting pace.
Some will reply that on the contrary, open access to data can only support peer review, but the expectation that reviewers should or could be re-running analyses to verify results is also misplaced. In my experience as an AE, I have seen the process of securing reviewers become more and more difficult over the past 10 years, and that was even before COVID. When reviewers are already asked to spend significant hours doing non-compensated work, it isn’t realistic to expect that they will involve themselves with re-running analyses (although some certainly will!)
So- in the end, we’re sort of stuck. How do we pursue Openness when progress in any particular domain is inherently dependent on a the select group of experts who work in that field? Maybe one answer is to recognize that Open Science means Open in some ways (i.e., FREE access to data and papers AFTER the Experts have given their stamp of approval) and closed in others (not sharing until the experts say you can?). This doesn’t feel very open, I admit, and let alone that at least in the US it breaches our 1st amendment rights, but perhaps some professional protocol like this is necessary to ensure integrity in our field. Would it be at all helpful to just focus on the notion of Free Science or Accessible Science instead of Open Science?