Network Gravity in Intellectual History

Network gravity models intellectual proximity as inverse-square attraction proportional to node masses (influence, output) and inverse to distance (geographic, confessional). In early modern Republic of Letters, it infers locations from associate clusters, treating scholars as masses in socio-epistemic space.

Definition

Gravity G_ij = (m_i m_j) / d_ij^2, where m is node mass (e.g., publication count, correspondence volume), d distance (km or affiliation divergence). Bayesian update: P(location_i | associates) ∝ ∑ G_ij. Priors from baselines (Six Degrees of Francis Bacon datasets); masses weighted by centrality (PageRank).

Workflow

Build graph from prosopographies (e.g., CBDB for Chinese parallels, EMLO for Europe). Segment temporally (decades). Infer position: Maximize ∑ G for unobserved nodes. Validate via negative evidence (e.g., silence in high-gravity clusters implies absence).

Implementation: NetworkX/PyMC for inference; Gephi for visualization.

Applications

Rank diffusion hubs (Sulzbach EMS covariate). Comparative: Christian Knorr von Rosenroth (high local gravity) vs. Franciscus Mercurius van Helmont (dispersed). Predict archival yields in low-gravity peripheries.

Example: Van Helmont

Ragley 1670s: High gravity from Conway/More (m=0.8), pulls P=0.95 despite gaps. Sulzbach 1667: Knorr mass elevates posterior from 0.5 to 0.85.

Related Concepts

Esoteric Maneuverability Score, Methodological Foundations for Probabilistic Geo-Temporal Timelines, Socio-Epistemic Networks.