CONCEPT

The Santiago Theory of Cognition

Maturana and Varela's radical claim — adopted and extended by Capra — that cognition is the process of life itself, and that every living system cognizes by the fact of maintaining its own organization.

The Santiago theory of cognition, developed by Humberto Maturana and Francisco Varela in the 1970s at the University of Chile in Santiago, proposes that cognition and life are the same process viewed from different angles. A living system is an autopoietic system — one that continuously produces itself through its own operations — and the process of self-production requires continuous sensing, responding, and adjusting to maintain organizational integrity against perturbation. This sensing-and-responding, Maturana and Varela argued, is what cognition is. It does not require a brain. A bacterium swimming up a nutrient gradient is cognizing. A plant tracking the sun is cognizing. An immune system distinguishing self from non-self is cognizing. Capra made this thesis central to his framework, and it carries sharp implications for the AI question: if cognition is inseparable from the autopoietic process of life, then AI systems — which are not alive in this strict sense — do not cognize in the same way, regardless of their behavioral capabilities.

In the AI Story

Hedcut illustration for The Santiago Theory of Cognition — The Santiago Theory of Cognition

The Santiago theory dissolves the assumption — pervasive in both popular and academic discourse — that cognition is a special trick performed by brains. Under the framework, brains are not the seat of cognition; they are elaborate organs that extend cognitive capacities a cell-sized organism already possesses. The bacterium's gradient-climbing, the plant's phototropism, and the human's philosophical reflection are points along a continuum of the same process: living systems sensing and responding to their environments in ways that maintain their organization.

For Capra, this continuum is not a metaphor. It is a claim about what cognition actually is at the organizational level. What appears in humans as sophisticated conscious thought is, structurally, a high-complexity instance of the same process that operates in every living cell. The implication is philosophically radical: consciousness is not the threshold that distinguishes cognitive systems from non-cognitive ones. Life is the threshold. Cognition is what life does.

Applied to AI, the Santiago theory produces a careful distinction. Large language models process language, generate outputs, and participate in networks of human communication. They do not, however, maintain themselves through their own operations in the autopoietic sense. They are allopoietic — produced by external agents (engineers, training processes) to perform functions other than self-maintenance. On the Santiago view, this makes their cognitive status categorically different from biological cognition, not merely quantitatively reduced. Evan Thompson has pressed this argument vigorously in recent years, insisting that AI does not cognize because it does not live.

Capra's own position is more nuanced. In his 2025 interview, he distinguished sharply between 'living intelligence' and 'artificial intelligence' while simultaneously acknowledging that networks including both biological and artificial nodes generate emergent properties irreducible to either. The tension is productive: the Santiago theory insists that the artificial node is not cognizing in the biological sense, and Capra's network framework insists that what matters for the network's properties is the quality of the interactions, not the categorical status of each participant.

Origin

Maturana and Varela developed the theory through a series of books beginning with Autopoiesis and Cognition (1980) and culminating in The Tree of Knowledge (1987). Capra adopted the framework centrally in The Web of Life (1996).

Key Ideas

Cognition is life. The process by which a living system maintains its organization is the process of cognition; they are not two separable phenomena.

Brains extend cognition; they do not constitute it. Cognitive capacity exists wherever autopoiesis exists, and brains are specialized organs that elaborate capacities already present in single cells.

The observer is part of the system. What counts as a perturbation, a response, or a relevant distinction depends on the organization of the cognizing system, not on external objective features.

Cognition is embodied. There is no cognition without the body whose organization the cognition maintains.

Machines do not cognize in the biological sense. Systems that do not maintain themselves through their own operations are not autopoietic and therefore, on the Santiago view, not cognizing.

Debates & Critiques

The Santiago theory is contested on multiple fronts. Computationalists argue that cognition is substrate-independent — that whatever matters about cognition can in principle be implemented in non-biological systems. Enactivists influenced by Maturana and Varela argue the opposite — that cognition requires the specific organizational closure of autopoiesis. Capra occupies a middle position, insisting on the categorical distinction while acknowledging that networks incorporating both biological and artificial nodes generate novel emergent properties.

Appears in the Orange Pill Cycle

Fritjof Capra — On AI