Estimating the Loss of Medieval Literature with an Unseen Species model from Ecodiversity (software and data)

This software and data accompanies a short paper for the first Computational Humanities Research Workshop (18-20 november 2020; to be held online):

M. Kestemont & F. Karsdorp, 'Estimating the Loss of Medieval Literature with an Unseen Species model from Ecodiversity'. Computational Humanities Research Workshop. Amsterdam [online], 18-20 november 2020.

This is an automatically captured archival dump of the releases of our Github repository (https://github.com/mikekestemont/minsampling) that contains all the code and software (with some documentation interwoven) necessary to replicate our findings.

Abstract: The century-long loss of documents is one of the major impediments to the study of historic literature. Here we focus on Middle Dutch chivalric epics (ca. 1200-1450), a genre for which little archival records exist that shed light on the survival rates of works and documents. We cast the quantitative estimation of these survival rates as a variant of the unseen species problem from ecodiversity. We apply an established non-parametric method (Chao1) and compare it to a number of common alternatives on simulated data. Finally, we discuss the implications of our results for conventional philology: our numbers suggest that the losses sustained on the level of works seem to be more dramatic than previously imagined, whereas those at the document-level align surprisingly well with existing estimates in book history, although these were based on completely different data sources.