The selection of the codices to be included in the Innovating Knowledge database was governed by the following three criteria:
- Manuscripts included in the database belong to the early Middle Ages, dating principally before the year 1000. However, because the turn of the centuries does not provide a particularly useful benchmark for the development of script and manuscript culture, the database also includes manuscripts dated up to the mid-eleventh century, as far as their features place them in the early Middle Ages.
- Manuscripts included in the database are direct witnesses of the Etymologiae, rather than containing citations and paraphrases from Isidore’s work or drawing on it. The database, therefore, excludes manuscripts of works based on the Etymologiae or making use of large parts of Isidore’s encyclopaedia, which are, however, transmitted in their own right (e.g. the Liber glossarum) or attributed to a different author (e.g. early medieval grammatical texts heavily dependent on the first book of the Etymologiae). Manuscripts containing certain complicated types of texts transmitting bits and pieces of the Etymologiae, such as computistic collections, grammatical compilations, and pastoral handbooks, are, nevertheless, included. Their inclusion was motivated by their unique character as manuscript objects (rather than as new texts). It is, moreover, likely that similar collections would not make it to other handlists due to their anonymous and pragmatic nature.
- Manuscripts included in the database must contain a minimal amount of text of the Etymologiae. This minimum is constituted by at least a single section in the edition, or, in the case of very long sections (e.g. Etym. I 27.1), a larger part of one section. Furthermore, the text of the Etymologiae in the manuscripts is presented in a continuous manner in the same fashion as in the edition, not departing from the general structure and tenor of the Etymologiae. For this reason, all glossaries, glossographic texts and thematic lists based on or containing definitions from the Etymologiae were excluded from the consideration.
The database includes fully preserved manuscripts, partially damaged manuscripts, and fragments (including two palimpsests). Included codices transmit the entire text of the Etymologiae and parts of this work, including excerpts and additions in the blank spaces and margins of manuscripts. As far as these manuscripts are described in Anspach’s notes, the database also accounts for manuscripts destroyed in the nineteenth and twentieth centuries.