Integrating scientometric indicators with linguistic data mining to enhance international research collaboration

Although scientometric literature provides valuable insights into international collaborative networks, approaches that integrate interdisciplinary perspectives to study international co-authorship practices remain uncommon. To fill this gap, we enhance co-authorship network analysis by incorporating Fleck’s concepts of ‘thought-collective’ (or communities that have a shared body of knowledge, epistemology and practices) and ‘thought-style’ (particular ways of understanding concepts and ideas) as expressed through language. Specifically, we use network metrics and community detection algorithms to describe and interpret the configuration of networks, identify distinctive structural features and evaluate the significance and influence of four leading researchers in several subdisciplinary fields. Additionally, we use linguistic data mining techniques to explore how the thought-styles of the networks, as thought-collectives, are expressed semantically and discoursally in their co-authored articles. Our findings show that enhancing network analysis with linguistic analysis provides a better understanding of how information circulates in collaborative networks and how collective thinking is expressed through language in articles published in impact-factor journals. We also tentatively suggest how collective styles can influence researchers’ relevance and impact. These novel findings could inform researcher training programmes raising awareness of successful social interactions in international collaborative ecosystems.