Relational Ontology

Relational ontology is the metaphysical position that entities are constituted by their relations rather than being self-sufficient substances that enter into relations as an afterthought. A thing is what it is because of how it is connected. Remove the connections, and the thing ceases to be the same thing — or ceases to be a thing at all.

This position inverts the classical Western metaphysics of substance, which treats relations as external and accidental: the table is a table whether or not anyone sits at it, whether or not it stands in a room, whether or not it was made by a carpenter. Relational ontology denies this. The table is a table because it is used as a table, because it is connected to human practices of eating and working, because it is part of a material and social network that gives it its identity.

The position has independent roots in multiple traditions: in process philosophy (Alfred North Whitehead's organism), in Buddhist metaphysics (dependent origination, pratītyasamutpāda), in cybernetics (the system is its relations), and in network theory (nodes are defined by their position in the topology). What unifies these strands is the rejection of the assumption — foundational to modern science — that the world is composed of independent objects that can be studied in isolation.

Relational ontology has direct consequences for how we understand complex systems. In a CAS, an agent is not an individual with fixed properties but a node whose behavior is shaped by its connections. The agent's 'identity' is a dynamic pattern of interactions, not a static essence. This is not merely a theoretical point; it is a practical one. If we treat firms, ecosystems, or AI systems as isolated entities, we will misunderstand how they work. The ontology of isolation produces the pathology of isolation.

The question relational ontology asks of modern science is not whether it is true but whether it is viable. Can a science that treats the world as a network of relations generate the same predictive power as a science that treats the world as a collection of objects? The answer from network theory and complexity theory suggests that it can — and that in some domains, it must.