Corpus di letteratura come corpus di lingua: il caso del Medioevo

The implementation of a corpus of a historically determined variety of a language poses very relevant methodological problems. The principle one: is it possible to create a corpus actually representative of a linguistic variety of which we do not know, nor can we know, the extension, and in which the weight of literary texts necessarily is more conspicuous than in a real language? The answer from the point of view of corpus linguistics is certainly negative. But numerous tools are available today (first of all the Corpus TLIO and the Corpus OVI dell'Italiano antico) that would seem to show the opposite. Starting from the OVI experience, some methodological reflections on the creation of corpus of historical varieties of language will be proposed, showing how, although it is impossible to apply criteria of balance and representativeness stricto sensu, a series of corrections can be applied (in in particular by selecting the texts on the basis of the discursive traditions) which allow to have tools that effectively meet the needs of linguistic research