Assessing language vitality in multilingual contexts: ethnographies versus computacional models

Abstract

This paper aims at discussing the import of studies on language vitality in multilingual ecologies, carried out within the framework of complexity science by means of computer models. One of these studies will be described and compared with ethnographic reports which illustrate the dynamics of the socioeconomic and political relations which characterize some multilingual ecologies, in particular the Upper Rio Negro in Brazilian Amazonia. Considering that assuming language as a complex, open, dynamic and self-organized system is a promising path for linguistic investigations, this work suggests that ethnographies and sociolinguistic descriptions seem to capture the dynamicity and complexity of language as a social phenomenon better than computational models which simplify the characterization of the system, and which are based on factors chosen by the researchers a priori, rather than factors which actually define the system.