The VARSUL Database (Urban Language Variation in the South of the Country) assembles speech data from social and culturally significant locations of the three southern states of Brazil. This article presents some considerations on the use, -both in the field of language variation and in several other fields of research, especially as regards to phonological phenomena - of data collections gathered according to the Labovian approach. The importance and advantages obtained by corpora analyses of spontaneous spoken language are addressed. We also discuss the possibilities that collections of sociolinguistic interviews provide for interface with other areas of research. Some features of the constitution of the VARSUL Database are presented, such as data collection methodology, criteria for informants selection, transcription protocols, as well as search devices developed for this Database. Finally, a survey of researches conducted on these data involving phonetic-phonological variables is presented to illustrate the scope of phenomena that have been studied and to show the possibility of generalization of results.
VARSUL; Speech Database; Phonological Variation
