Quantifying simultaneous innovations in evolutionary medicine

Abstract

To what extent do simultaneous innovations occur and are independently from each other? In this paper we use a novel persistent keyword framework to systematically identify innovations in a large corpus containing academic papers in evolutionary medicine between 2007 and 2011. We examine whether innovative papers occurring simultaneously are independent from each other by evaluating the citation and co-authorship information gathered from the corpus metadata. We find that 19 out of 22 simultaneous innovative papers do, in fact, occur independently from each other. In particular, co-authors of simultaneous innovative papers are no more geographically concentrated than the co-authors of similar non-innovative papers in the field. Our result suggests producing innovative work draws from a collective knowledge pool, rather than from knowledge circulating in distinct localized collaboration networks. Therefore, new ideas can appear at multiple locations and with geographically dispersed co-authorship networks. Our findings support the perspective that simultaneous innovations are the outcome of collective behavior.

Computational History: Challenges and Opportunities of Formal Approaches

“…Applying a range of methods from computational linguistics and network analysis to text corpora in the history of science, we could model the growth and differentiation of these discourses [51] , distinguish individual discourses through disambiguation techniques, identify sub-clusters of concepts that represent incipient specialization or novel dimensions of discourses, and identify the origin of novelties within these corpora. Time series analyses [52] of these developments give us a dynamic view of change on the semantic or idea level. But we can also link these contentbased analyses with the underlying social networks that represent the actual historical actors.…”