The paper presents the first results of a terminological study concerningthe knowledge management of the Area 10 National University Council (cun) research domains .Sciences of antiquity, philological–literary and historical–artistic.. In the present research, we created a corpus of institutional texts in order to extract the terms which could be candidates to automatically populate a machine–readable controlled vocabulary able to describe the topics under examination. In the paper we will describe the statistically and linguistically motivated strategies we adopted to process the corpora and the evidences that emerged from the text analyses.