The IPIC resource and a cross-linguistic analysis of information structure in Italian and Brazilian Portuguese
We present a multi-level XML online database, DB-IPIC, designed specifically for the study of linear relations among information units in spoken corpora. DB-IPIC adopts Language into Act Theory (L-ACT) as the basis for spoken language modeling. According to L-ACT, information is coded at the prosodic level: prosodic units relate to information units. By exploiting data from DB-IPIC, we produced a comparative study on information structure in Italian and Brazilian Portuguese. Language samples derive from C-ORAL-ROM Italian and C-ORAL-BRASIL corpora and received prosodic boundary annotation and information tagging according to the information functions proposed by L-ACT. We highlight the frequency of occurrence of information units and information patterns in Italian and Brazilian Portuguese.