Uploaded February 2026 | Updated September 2026, 2 weeks ago
Title: “Datasets and dead assets: The missing archives of science”
Abstract: According to lore, the Library of Alexandria was created to compile every manuscript its librarians could get their hands on. Science is not like that. It is a system of filters that limits what is acceptable and accessible. Science communicators have increasingly understood it to be their responsibility to guide science consumers towards trust in prevailing views by promulgating authoritative datasets that restrict what can be considered reasonable questions and disagreements. Databases, journals, conferences, and institutions keep gates narrow. Science aims for consensus, and pillories anything that isn’t deemed mainstream. But mainstream science includes many well-known biases, and is also prone to flawed or dubious practices. For example, recent work suggests data from clinical trials published in top journals is regularly flawed or faked. While the call for consensus is variously justified—for instance, on the basis of concern for public health—these arguments often underestimate the value of dissenting views and reflect the impact of political polarization. Yet efforts to create consensus archives continue. This talk presents a counterpoint. We have much to gain from an Alexandrian approach that emphasizes inclusion. We could create archives of science that reflect the full spectrum of views, including unpopular views, views of nonprofessionals and outsiders. What would such an archive look like? How would it work? These are the kinds of questions we should be asking. Here I motivate the Alexandrian approach and offer suggestions about its implementation.
Title: “Datasets and dead assets: The missing archives of science”
Abstract: According to lore, the Library of Alexandria was created to compile every manuscript its librarians could get their hands on. Science is not like that. It is a system of filters that limits what is acceptable and accessible. Science communicators have increasingly understood it to be their responsibility to guide science consumers towards trust in prevailing views by promulgating authoritative datasets that restrict what can be considered reasonable questions and disagreements. Databases, journals, conferences, and institutions keep gates narrow. Science aims for consensus, and pillories anything that isn’t deemed mainstream. But mainstream science includes many well-known biases, and is also prone to flawed or dubious practices. For example, recent work suggests data from clinical trials published in top journals is regularly flawed or faked. While the call for consensus is variously justified—for instance, on the basis of concern for public health—these arguments often underestimate the value of dissenting views and reflect the impact of political polarization. Yet efforts to create consensus archives continue. This talk presents a counterpoint. We have much to gain from an Alexandrian approach that emphasizes inclusion. We could create archives of science that reflect the full spectrum of views, including unpopular views, views of nonprofessionals and outsiders. What would such an archive look like? How would it work? These are the kinds of questions we should be asking. Here I motivate the Alexandrian approach and offer suggestions about its implementation.
