Interaction matrices as unifying tools for navigating ecological complexity

Interaction matrices summarize pairwise species impacts within ecological communities into an analytically tractable format. Despite their contribution to linking subdisciplines and advancing our understanding of ecosystem dynamics, they have faced recurring criticism for oversimplifying ecological complexity. Extensions developed to address specific shortcomings often come with trade-offs, including increased data demands and analytical complexity. Here we review the foundations of the interaction matrix framework and examine the state of the art of approaches developed to tackle its challenges. We argue that the unifying role of interaction matrices could be improved by embracing, rather than fighting, their inherent nature as uncertain and transient abstractions. We explore how such an approach could bring substantial benefits and incorporate advances from wider complexity sciences. As we enter the second century of interaction matrices within ecology, such a shift will reinforce their place as a valuable tool to unlock the complex dynamics of ecological communities.