Coevolution within political cultures is a credible but undertheorized mechanism for persistence and divergence. Cultures are the weight of social knowledge, informing about how to act; cultures culminate from information held across social networks. Focusing on political networks, we investigate the coevolution of political decision-making in medieval Europe. We apply network methods to relational data on over 800 medieval cities and over 500 governing elites, alongside their political institutions. We use a community detection algorithm to cluster cities into political communities based on elite networks and then explore the coherence of political behavior within communities as well as their divergence from sample-wide trends. Some political communities evidence distinctive governing practices. Bayesian regressions for policy choice evidence that cities take cues from peers within their networks moreso than general trends. Local political cultures contributed to variation in political development.