On networks, trees and traits: evaluating explanatory power of network models in cultural evolutionary research

Abstract In this paper, I assess the explanatory power of four graph‑ and network‑based models recently used in the domain of Cultural Evolutionary Theory (CET): (a) Phylogenetic Trees, (b) Death–Birth Graphs, (c) Interaction Graphs, and (d) Trait Networks. First, I show that Interaction Graphs, Trait Networks and (empirically‑calibrated) Death–Birth Graphs can function as genuine mechanistic explanations, while Phylogenies remain largely at the level of correlational evidence. I then broaden the analysis to two non‑mechanistic accounts: Kitcher’s unificatory perspective and Kostić’s counterfactual theory of topological explanation. Interaction graphs excel mechanistically and topologically, death–birth graphs yield strong topological explanations, phylogenetic trees lead in unification, and trait networks offer mixed, moderate strengths. Because no single model or mode dominates, explanatory strength in CET is to a large extent question relative. I argue for a pluralistic approach in CET: rather than competing, these three explanatory axes (“modes of explanation”) complement one another and suggest a roadmap for future hybrid network models that could combine these complementary virtues and invite closer collaboration between CET modellers and philosophers of science.