Abstract
The aim of our paper is to illustrate an exploratory analysis on the vowel system in Italian spoken in Bolzano by both native and non-native speakers (i.e. Tyrolean German speakers). Data for the analysis come from DIA (Dialogic ItAlian) corpus, a speech corpus of spontaneous Italian spoken in Bolzano and based on dialogic interactions. After a short illustration of the studies on Italian vowels and a phonetic description of the Italian speech variety of Bolzano, we provide some information on the construction of DIA corpus. The central part of the contribution is devoted to the discussion of the specific methodological problems that arise in the acoustic analysis of vowels extracted from speech corpora. To address these problems, we carry out a spectral analysis based on DCT coefficients and we test the hypothesis that there are two vowel sub-systems using linear mixed models.