Graphs are focal hypergraphs: strict containment in higher-order interaction dynamics

Researcher Elkaïoum M. Moutuou has published a foundational paper on arXiv that fundamentally reclassifies the relationship between graphs and hypergraphs in modeling complex systems. The work introduces a critical taxonomy distinguishing between 'focal' interactions (centered on a designated node, like traditional graph neighborhoods) and 'non-focal' interactions (where all participants are equivalent). Moutuou proves that every graph is canonically a 'focal hypergraph' through its closed neighborhood structure, meaning all graph dynamical models are a special case of hypergraph models. This establishes a strict, three-level containment hierarchy: graph models are a proper subset of focal hypergraph models, which are themselves a proper subset of general hypergraph models.

The paper's technical core shows that closed graph neighborhoods are precisely focal hyperedges, and that hyperedges generalize graph neighborhoods by removing the focal constraint. This reveals that while graph models do encode genuine many-body interactions within a node's neighborhood, they remain incapable of representing the full spectrum of non-focal, higher-order interactions captured by general hypergraphs. The implications are significant for AI and complex systems science: it provides a formal principle of 'representational alignment,' arguing that model choice should be driven by interaction type rather than formalism preference. This directly impacts fields using Graph Neural Networks (GNNs), agent-based modeling, and systems biology, where misaligned models may miss critical dynamics.