Bibliographic bias and information-density sampling

In the present paper, we discuss the bibliographical limits for commonplace typological studies and address how to estimate the resources available for an in-depth study using a full-text corpus of grammatical descriptions, considering different metalanguages, temporal stages of description, theoretical perspectives, and quality of grammatical descriptions. In a case study on motion, we illustrate the above perspectives and show how computer-assisted sampling using large-scale keyword searches for information-dense descriptions is a time-saving resource for the linguistic researcher to create genealogically independent samples. The measures discussed in this study allow for a better appraisal of the state of existing information for typological studies, but the problem of wider access to rare publications remains a significant challenge.