Network structure explains intellectual discourse across human history

The production of knowledge is a collective endeavor. Scientific discovery, for instance, reflects not only the insights of individual scientists but also interactions among scientists. This is often taken to reflect social influences on collective epistemic vitality, the capacity to generate and synthesize new ideas. However, in science, it is difficult to separate intellectual influence from access to material resources, such as equipment and grant funding, since networks of collaboration influence the circulation of both intellectual and non-intellectual resources. Here, as a strict test of how social structure shapes intellectual discourse, we use the three-thousand-year history of a human debate in communities that relied on intellectual argumentation (rather than, say, empirical experiments). Building on the work of historians and sociologists, we digitized and quantified the time-evolving network structure of interaction among intellectuals (N = 3187), broadly construed, from religious debate in ancient India (c. 800 BCE) to 20th century debates about the logical foundations of mathematics in Europe and North America. We find that the production or preservation of knowledge by a community is explained by its network structure but not with overall levels of antagonism, suggesting that how communities are organized matters more for intellectual progress than how contentious they are. Extending tools from collective intelligence to intellectual history, we call for an integration of the science of science, the philosophy of science, and the history of ideas to forge a comprehensive understanding of the social dynamics of knowledge.