IIn this article it was studied the scientific collaboration networks formed from a group of researchers linked to the Programa de Pós-Graduação em Geociências da Universidade Federal do Rio Grande do Sul (PPGGeo/UFRGS), based on a theoretical-methodological approach known as social network analysis (SNA), considering some concepts from Pierre Bourdieu's Theory of Practice, and focusing on the identification, characterization and the structural evolution of scientific co-authorship networks. Three networks were built based on data from the CAPES' Cadernos de Indicadores related to the intervals 1998-2000, 2001-2003 and 2004-2006. The results characterized the structure of the networks, comparing them, focusing on their connectivity and centrality, taking into consideration the research lines, projects and groups of actors. The networks of 1998-2000, 2001-2003 and 2004-2006 had, respectively, 524 actors and 11.296 links; 576 actors and 14.674 links; 741 actors and 14.188 links. It was found that the central/dominant group of actors in the networks is formed mainly by teachers and, in general, this group tends to remain in the central/dominant positions over the years; it was also found that there are partnerships recurrences in the scientific knowledge production in the three networks, culminating in a social reproduction of the co-authorship network structure.
Social network analysis; Scientific co-authorship; Scientific knowledge production; Geosciences
