The volume proposes a corpus-driven study on the 14,000 occurrences of the verb “essere” (to be) in a corpus of Italian speech C-ORAL-ROM, in the light of the main proposals for the semantic interpretation of the verb. Thus, the very possibility of carrying out research that is strongly characterised in a theoretical sense on a large amount of oral data, generally not very inclined to a systematic taxonomic treatment, is put to the test. The observable datum therefore defines the explicandum of linguistic reflection and the analysis on corpus guarantees the observational adequacy with respect to the actual production data. The proposed results show that the wide semantic variation in the use of the verb is structured in broadly comprehensive macro-classes: the copula (attributive) uses, the identifying uses and the predicative uses with a locative basis. Within each of these classes, primary, consistent and productive variation lines are found.