Enactivism

Enactivism is the radical thesis that cognition is not the representation of a pre-given world by a pre-given mind but is rather the enactment of a world and a mind through the activity of a living body. The thesis was developed by Francisco Varela, Evan Thompson, and Eleanor Rosch in their 1991 book The Embodied Mind, drawing on Maurice Merleau-Ponty's phenomenology and Humberto Maturana's biology.

The core claim is that perception and cognition are not processes of information pickup and processing but of structural coupling between an organism and its environment. A living system does not passively receive information about the world; it actively brings forth a world through its sensorimotor engagement. The world that appears is not independent of the organism's structure of perception and action — it is enacted by that structure.

This has radical implications for the philosophy of mind. If cognition enacts rather than represents, then the classical problems of intentionality — how mental states come to be about the world — dissolve and are replaced by a different problem: how a living system maintains its identity through ongoing engagement with an environment. The mind is not a representational system but an autonomous, self-producing system that maintains its own boundaries and generates its own domain of significance.

The challenge for enactivism is to explain systematic, abstract cognition — mathematics, logic, language — without appealing to internal representations. Proponents argue that these capacities are constituted by embodied practices and social interaction, not computed by symbolic engines. Critics argue that enactivism has not provided a positive account of systematic cognition and that its rejection of representation is premature.

From the perspective of AI safety, enactivism raises a disturbing possibility: if genuine cognition requires embodiment and autonomy, then disembodied artificial systems may not be cognitive in the relevant sense. They may be powerful optimization processes without the kind of understanding that would make alignment possible. The alignment problem, on this view, is not merely technical but ontological: we are trying to align a system that may not have the conceptual resources to be aligned.\n\n