A new project has just been launched to write a synchronic, descriptive grammar of Niçois, the Occitan dialect of Nice. In this article, I define the corpus of the research. To do so, I first review written production from the Middle Ages to the present. I then analyze the linguistic features of Niçois over time, in order to determine the precise starting point of the current language state. But because of reinforced normativism and the decreasing social use of Niçois among the educated population, written language after WWII became artificial and does not really correspond to recordings made in the field. The corpus will thus be composed of writings from the 1820’s to WWII and recordings from the last few decades.