An article by “Science” reports on a new statistical method that evaluates the number of medieval novels we have no more trace. A promising model, according to specialists.
by
Usually, when I publish a scientific article, I receive an email from a colleague, and that’s all. It’s much more intense this time … “Remco Sleiderink, a medieval Dutch literature professor at the University of Antwerp (Belgium), has just made an experience to which his job did not seem to exhibit: February 17 , with eight other researchers, he published a Article in Science , one of the world’s largest scientific journals, “Forgotten Books: The Application of Unseen Species Models to The Survival of Culture” (“Forgotten Books: The Application of Invisible Species Models to The survival of culture “). Since then, the reactions abound, whether to wonder, be enthusiastic or, already, sketch a sequel to the site that this publication opened with a smash.
because the case is important. Fruit of three years of research, the article asserts to evaluate the number of medieval chivalry novels which we no longer have trace. It even gives numbers: 32% of these works (in the immaterial sense) would have been lost, and more than 90% of the manuscripts – the difference explaining by the fact that the works, before printing, circulated using the aid. Multiple copies, so that a small number of manuscripts is enough to make us a greater number of works. But, precisely, what texts were the 90% missing manuscripts? On the one hand, texts become “invisible”, “says Remco Sleiderink.
species to books
The question of invisibility is a constant concern in ecology, where we have established so-called “invisible species” models, to give themselves the means to measure the number of unknown plant and animal species . This was, in 2019, the intuition of Mike Kestemont, Professor of Computational Science at the University of Antwerp, and Folgert Karsdorp, a specialist in cultural evolution at the Meertens Institute (Amsterdam): what would the application give of these models in the field of literary works? Point tests have been launched to test existing methods. One of them quickly imposed itself as the most robust, the Chao1 biodiversity estimator, created by the great Taiwanese statistician Anne Chao, which finally joined the team – it is one of the signatories of the Article.
Specialists from various European languages, Tel Remco Sleiderink for French, were then responsible for searching different corpus. They observed strong contrasts between works copied in a hundred manuscripts preserved, and others found in one or two only. A contrast that draws a kind of scarcity curve, that the mathematical formula of the Chao1 model makes it possible to extend in the area of the “zeros” (or manuscript or work), to measure its extent. And this is how the invisible is counted.
You have 49.44% of this article to read. The rest is reserved for subscribers.