HIGHER-ORDER NETWORK DYNAMICS REVEALS THE SEMI-NESTED HIERARCHICAL STRUCTURE OF NATURAL AND SOCIAL SYSTEMS

The notion that biological hierarchies comprise nested entities underlies research in macroevolution. However, as it has been observed in hierarchical structures connecting humans, the hierarchical spatial and temporal structuring of biodiversity may not follow a strictly nested pattern. Here we use higher-order Markov dynamics to explicitly describe higher-order interactions in biogeographic network (Middle Devonian Brachiopods from the Old-World Realm) and colexification networks (i.e., networks of concepts linked to each other if they are expressed by the same word). This approach allows to capture the larger-scale modular structure of bioregions and semantic communities at different scales, especially when modeling one-step dynamics (first-order Markov models) do not reveal a hierarchical pattern. The multi-scale community structure delineated using the Map Equation framework indicates that both bioregions and semantic communities do not have a unique level of description but multiple levels reflecting a semi-nested hierarchical organization of biodiversity and cross-linguistic colexification patterns.